United States
Environmental Protection
Agency
Policy, Planning,
And Evaluation
(PM-221)
EPA-230-05-89-053
June 1989
SEPA
The Potential Effects
Of Global Climate Change
On The United States
Appendix C
Agriculture
Volume 2
Printed on Recycled Paper
-------
THE POTENTIAL EFFECTS OF GLOBAL CLIMATE CHANGE
^ ON THE UNITED STATES:
APPENDIX C - AGRICULTURE
Editors: Joel B. Smith and Dennis A. Tirpak
OFFICE OF POLICY, PLANNING AND EVALUATION
VS. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
MAY 1989
-------
TABLE OF CONTENTS
APPENDIX C-l: AGRICULTURE
PREFACE iii
EFFECT OF GLOBAL CLIMATE CHANGE ON AGRICULTURE
GREAT LAKES REGION 1-1
J.T. Ritchie, B.D. Baer, and T.Y. Chou
IMPACT OF CLIMATE CHANGE ON CROP YIELD IN THE SOUTHEASTERN
USA: A SIMULATION STUDY 2-1
Robert M. Peart, J.W. Jones, R. Bruce Curry, Ken Boote, and
L. Hartwell Allen, Jr.
POTENTIAL EFFECTS OF CLIMATE CHANGE ON AGRICULTURAL
PRODUCTION IN THE GREAT PLAINS: A SIMULATION STUDY 3-1
Cynthia Rosenzweig
THE ECONOMIC EFFECTS OF CLIMATE CHANGE ON US. AGRICULTURE:
A PRELIMINARY ASSESSMENT 4-1
Richard M. Adams, J. David Glyer, and Bruce A. McCarl
CLIMATE CHANGE IMPACTS UPON AGRICULTURE AND RESOURCES:
A CASE STUDY OF CALIFORNIA 5-1
Daniel J. Dudek
EFFECTS OF PROJECT CO2-INDUCED CLIMATIC CHANGES ON IRRIGATION
WATER REQUIREMENTS IN THE GREAT PLAINS STATES (TEXAS, OKLAHOMA,
KANSAS, AND NEBRASKA. 6-1
Richard G. Allen and Francis N. Gichuki
APPENDIX C-i AGRICULTURE
DIRECT (PHYSIOLOGICAL) EFFECTS OF INCREASING CO, ON CROP PLANTS
AND THEIR INTERACTIONS WITH INDIRECT (CLIMATIC) EFFECTS 7-1
EliseRose
POTENTIAL EFFECTS OF CLIMATE CHANGE ON PLANT-PEST INTERACTIONS.. 8-1
Benjamin R. Stinner, Robin AJ. Taylor, Ronald B. Hammond,
Foster F. Purrington, and David A. McCartney
IMPACTS OF CLIMATE CHANGE ON THE TRANSPORT OF AGRICULTURAL
CHEMICALS ACROSS THE USA GREAT PLAINS AND CENTRAL PRAIRIE 9-1
Howard L. Johnson, Ellen J. Cooler, and Robert J. Sladewski
FARM-LEVEL ADJUSTMENTS BY ILLINOIS CORN PRODUCERS TO
CLIMATE CHANGE 10-1
William E. Easterling
ii
-------
TABLE OF CONTENTS (continued)
Page
APPENDIX C-i AGRICULTURE (continued)
CHANGING ANIMAL DISEASE PATTERNS INDUCED BY THE
GREENHOUSE EFFECT 11-1
Edgar Stem, Gregory A. Mertz, J. Dirck Stryker, and Monika Huppi
EFFECT OF CLIMATIC WARMING ON POPULATIONS OF THE HORN FLY,
WITH ASSOCIATED IMPACT ON WEIGHT GAIN AND MILK PRODUCTION
IN CATTLE 12-1
E.T. Schmidtmann and J A. Miller
AGRICULTURAL POLICIES FOR CLIMATE CHANGES INDUCED BY
GREENHOUSE GASES 13-1
G. Edward Schuh
111
-------
PREFACE
The ecological and economic implications or the greenhouse effect have been the subject of discussion within
the scientific community for the past three decades. In recent years, members of Congress have held hearings
on the greenhouse effect and have begun to examine its implications for public policy. This interest was
accentuated during a series of hearings held in June 1986 by the Subcommittee on Pollution of the Senate
Environment and Public Works Committee. Following the hearings, committee members sent a formal request
to the EPA Administrator, asking the Agency to undertake two studies on climate change due to the greenhouse
effect.
One of the studies we are requesting should examine the potential health and environmental
effects of climate change. This study should include, but not be limited to, the potential impacts
on agriculture, forests, wetlands, human health, rivers, lakes, and estuaries, as well as other
ecosystems and societal impacts. This study should be designed to include original analyses, to
identify and fill in where important research gaps exist, and to solicit the opinions of
knowledgeable people throughout the country through a process of public hearings and
meetings.
To meet this request, EPA produced the report entitled The Potential Effects of Global Climate Outage on the
United States. For that report, EPA commissioned fifty-five studies by academic and government scientists on
the potential effects of global climate change. Each study was reviewed by at least two peer reviewers. The
Effects Report summarizes the results of all of those studies. The complete results of each study are contained
in Appendices A through J.
Appendix Subject
A Water Resources
B Sea Level Rise
C Agriculture
D Forests
E Aquatic Resources
F Air Quality
G Health
H Infrastructure
I Variability
J Policy
GOAL
The goal of the Effects Report was to try to give a sense of the possible direction of changes from a global
warming as well as a sense of the magnitude. Specifically, we examined the following issues:
o sensitivities of systems to changes in climate (since we cannot predict regional climate change, we
can only identify sensitivities to changes in climate factors)
o the range of effects under different wanning scenarios
o regional differences among effects
o interactions among effects on a regional level
iv
-------
o national effects
o uncertainties
o policy implications
o research needs
The four regions chosen for the studies were California, the Great Lakes, the Southeast, and the Great Plains.
Many studies focused on impacts in a single region, while others examined potential impacts on a national scale.
SCENARIOS USED FOR THE EFFECTS REPORT STUDIES
The Effects Report studies used several scenarios to examine the sensitivities of various systems to changes in
climate. The scenarios used are plausible sets of circumstances although none of them should be considered to
be predictions of regional climate change. The most common scenario used was the doubled CO2 scenario
(2XCO2X which examined the effects of climate under a doubling of atmospheric carbon dioxide concentrations.
This doubling is estimated to raise average global temperatures by 1.5 to 45°C by the latter half of the 21st
century. Transient scenarios, which estimate how climate may change over time in response to a steady increase
in greenhouse gases, were also used. In addition, analog scenarios of past warm periods, such as the 1930s, were
used.
The scenarios combined average monthly climate change estimates for regional grid boxes from General
Circulation Models (GCMs) with 1951-80 climate observations from sites in the respective grid boxes. GCMs
are dynamic models that simulate the physical processes of the atmosphere and oceans to estimate global climate
under different conditions, such as increasing concentrations of greenhouse gases (e.g., 2XCO2).
The scenarios and GCMs used in the studies have certain limitations. The scenarios used for the studies assume
that temporal and spatial variability do not change from current conditions. The first of two major limitations
related to the GCMs is their low spatial resolution. GCMs use rather large grid boxes where climate is averaged
for the whole grid box, while in fact climate may be quite variable within a grid box. The second limitation is
the simplified way that GCMs treat physical factors such as clouds, oceans, albedo, and land surface hydrology.
Because of these limitations, GCMs often disagree with each other on estimates of regional climate change (as
well as the magnitude of global changes) and should not be considered to be predictions.
To obtain a range of scenarios, EPA asked the researchers to use output from the following GCMs:
o Goddard Institute for Space Studies (GISS)
o Geophysical Fluid Dynamics Laboratory (GFDL)
o Oregon State University (OSU)
Figure 1 shows the temperature change from current climate to a climate with a doubling of CO, levels, as
modeled by the three GCMs. The figure includes the GCM estimates for the four regions. Precipitation
changes are shown in Figure 1 Note the disagreement in the GCM estimates concerning the direction of
change of regional and seasonal precipitation and the agreement concerning increasing temperatures.
Two transient scenarios from the GISS model were also used, and the average decadal temperature changes
are shown in Figure 3.
-------
FIGURE 1. TEMPERATURE SCENARIOS
GCM Estimated Change in Temperature from 1xCO2 to 2xCO2
Great Southeast Great California United
Lakes Plains States'
Great Southeast Great California United
Lakes Plains States'
Great Southeast Great California United
Lakes Plains States*
GIS'3
GFDL
OSU
* Lower 48 Stetes
-------
FIGURE 2. PRECIPITATION SCENARIOS
GCM Estimated Change in Precipitation from 1xCC>2 to 2xCO2
i.u-
>- 0.8
0 0.6-
g? 0.4-
LU
ts °-2'
5 0.0-
d -0.2-
-0.4-
ANNUAL
„„! 1 |j-|
I.U
0.8
0.6
0.4
0.2
0.0
-0.2-
-0.4-
-0.6 J
WINTER
i
W
*
|
1
r
i JL
\
i 1
I..H1
i.u-
0.8-
0.6-
0.4-
0.2-
o.o-
-0.2-
-0.4-
-0.6-
No
Change
1
1
s
i
GB
^
1
i
SUMMER
i
n
^
i
K
i
n! 1
•""I— i
1 n
BS
i
Great Southeast
Lakes
Great California United
Plains States'
Great Southeast Great California United
Lakes Plains States*
Great SoutheaM Great California United
Lakes Plains States'
GISS
GFDL
OSU
* Lower 48 States
-------
O
g __
UJ
DC
<
oc
UJ
1
UJ
4
3.5
3
1.5
1
0.5
0
3.72
2.99
2.47
1.72
1.36
0.70
0.88
- 0.30
1980s 1990s 2000s 2010s 2020s 2030s 2040s 2050s
TRANSIENT SCENARIO A
*
3.5
3
O
w 2.5
oc
3
•- 2
< *
oc
£ 1.5
S
UJ
^v 4
•^ 1
0.5
0
—
—
—
_
-
1.02
^ f f f
0.59 '///
0.1 a //// '//s
v ; / y j x X / x / / x .
\s / / A V//A '///, ///
1.26
s / s /
J * s s s
/ '///
/ y///
y ////
1980s 1990s 2000s 2010s 2020s
TRANSIENT SCENARIO B
FIGURE 3.
GISS TRANSIENTS "A" AND "B11 AVERAGE
TEMPERATURE CHANGE FOR LOWER 48 STATES
GRID POINTS.
VUl
-------
EPA specified that researchers were to use three doubled CO, scenarios, two transient scenarios, and an analog
scenario in their studies. Many researchers, however, did not have sufficient time or resources to use all of the
scenarios. EPA asked the researchers to run the scenarios in the following order, going as far through the list
as time and resources allowed:
1. GISS doubled CO2
2. GFDL doubled CO2
3. GISS transient A
4. OSU doubled CO2
5. Analog (1930 to 1939)
6. GISS transient B
ABOUT THESE APPENDICES
The studies contained in these appendices appear in the form that the researchers submitted them to EPA.
These reports do not necessarily reflect the official position of the US. Environmental Protection Agency.
Mention of trade names does not constitute an endorsement.
IX
-------
DIRECT (PHYSIOLOGICAL) EFFECTS OF INCREASING CO, ON CROP
PLANTS AND THEIR INTERACTIONS WITH INDIRECT
(CLIMATIC) EFFECTS
by
EliseRose
502 S. Baldwin
Madison, WI 53703-3707
Contract No. 68-01-7489
-------
CONTENTS
Page
FINDINGS 7-1
CHAPTER 1: INTRODUCTION 7-2
STATUS OF RESEARCH 7-2
RECOMMENDATIONS FOR RESEARCH 7-3
CHAPTER* PLANT PROCESSES AFFECTED BY INCREASING
CARBON DIOXIDE 7-4
YIELD 7-4
PHOTOSYNTHESIS 7-5
FEEDBACK INHIBITION 7-5
OTHER CARBON-EXCHANGE REACTIONS 7-7
STOMATAL RESPONSES 7-8
TRANSPIRATION AND WATER USE 7-8
LEAF TEMPERATURE 7-9
TEMPERATURE INTERACTIONS 7-9
DROUGHT STRESS 7-10
WATER-USE EFFICIENCY 7-11
CHAPTER 3: SOYBEAN 7-12
YIELD 7-12
PHOTOSYNTHESIS 7-13
FEEDBACK INHIBITION 7-14
STOMATAL RESPONSES 7-15
TRANSPIRATION AND WATER USE 7-15
LEAF TEMPERATURES 7-16
TEMPERATURE INTERACTIONS 7-16
DROUGHT STRESS 7-16
NITROGEN FIXATION AND NUTRIENTS 7-17
WATER-USE EFFICIENCY 7-17
INSECT FEEDING 7-17
CHAPTER 4: WHEAT 7-19
YIELD 7-19
PHOTOSYNTHESIS 7-20
FEEDBACK INHIBITION 7-20
OTHER CARBON-EXCHANGE REACTIONS 7-21
STOMATAL RESPONSES 7-21
TRANSPIRATION 7-21
LEAF TEMPERATURE 7-22
TEMPERATURE INTERACTIONS 7-22
DROUGHT STRESS 7-23
WATER-USE EFFICIENCY 7-23
NUTRITION 7-24
CHAPTER 5: COTTON 7-25
YIELD 7-25
PHOTOSYNTHESIS 7-25
FEEDBACK INHIBITION 7-26
TEMPERATURE INTERACTIONS 7-27
STOMATAL RESPONSES 7-27
TRANSPIRATION, WATER USE, AND WATER-USE EFFICIENCY 7-28
LEAF TEMPERATURE 7-28
DROUGHT STRESS 7-29
• •
ii
-------
CONTENTS (continued)
Page
CHAPTER 6: CORN 7-30
PHOTOSYNTHESIS 7-30
YIELD 7-30
PHENOLOGY 7-30
STOMATAL RESPONSE, TRANSPIRATION, AND WATER-USE EFFICIENCY 7-31
LEAF TEMPERATURE 7-31
DROUGHT STRESS 7-31
REFERENCES 7-32
ui
-------
Rose
FINDINGS1
At moderate temperatures, most crops will probably show biomass increases and possibly yield increases
as carbon dioxide concentrations rise.
Leaf temperatures in all species are expected to rise even more than air temperatures, which may
inhibit plant processes sensitive to high temperatures.
Photosynthesis will increase in certain species and in some environments. If under certain conditions
photosynthetic products accumulate, plants may regulate the level of photosynthetic CO2 uptake rates through
feedback inhibition of some enzymes or enhanced export of photosynthetic products out of the source leaves.
Drought-stressed plants exposed to high partial pressures of CO2 will be more able to cope with water
deficits. Plants lacking adequate moisture will experience higher rates of photosynthesis and higher yields at
enriched levels of carbon dioxide than at current CO2 concentrations.
Increasing air temperatures predicted by global climate change models will interact with all of the direct
effects of enhanced concentrations of carbon dioxide, and will influence the direct effects over the life cycle of
plants in ways that are not predictable with certainty at this time. The few exceptions to this uncertainty are the
few. areas where temperature and elevated concentrations of CO2 have been studied simultaneously over the
annual cycle or the whole life cycle of plants. Even these limited studies need to be verified with a range of crops
and cultivars. Those studies indicate that the positive effects of carbon dioxide enrichment may be reduced or
even eliminated at high temperatures.
It is not clear whether plants' water use throughout their life cycles will increase, decrease, or remain
the same as now. Stomata (leaf pores) in some environments dose as levels of carbon dioxide rise. However,
other environmental factors such as humidity, temperature, light, and soil moisture also affect stomata and other
facets of water use. These effects may be more important than the amount of carbon dioxide in determining
water use.
Future research should include a range of higher temperatures similar to those predicted by the global
climate change models, as well as higher partial pressures of carbon dioxide. Although more complex to conduct,
experimental water-deficit scenarios also should be included in some future experiments.
Experimental research should be integrated with efforts to model plant processes.
'Although the information in this report has been funded wholly or partly by the U.S. Environmental
Protection Agency under Contract No. 68-01-7489, it does not necessarily reflect the Agency's views, and no
official endorsement should be inferred from it.
7-1
-------
Rose
CHAPTER 1
INTRODUCTION
Predicted global increases in atmospheric carbon dioxide are expected to influence plant metabolism,
growth, and development in two ways. Elevated concentrations of CO2 directly affect plant processes. Changes
in the level of CO, are also expected to influence plants indirectly through predicted increases in temperature
and changes in other climate variables such as precipitation and atmospheric humidity.
This paper reviews recent reports of experiments on the direct effects of increasing atmospheric carbon
dioxide on crop plants. First, the general status of research on effects of increasing levels of carbon dioxide on
plant behavior will be summarized, with remarks about different experimental techniques being used and current
research needs. Then, general effects of CO, on particular plant processes across many species will be outlined.
Finally, because each species responds to carbon dioxide in ways that are peculiar to its phenology, growth habit,
and adaptation, the review will explore findings specific to four major U.S. crops. Literature since 1980 on
soybean, wheat, cotton, and corn will be examined, with particular attention to interactions between the effects
of elevated partial pressures of carbon dioxide, rising temperatures, and other environmental stresses. Plant
growth models, which are useful research tools to explore the possible effects of rising levels of CO2 but which
differ substantially from an experimental approach, are not reviewed here.
STATUS OF RESEARCH
Research on the direct effects of CO- enrichment is most comprehensive for soybean, and photosynthesis
is the plant process researched most extensively. In other species, research has concentrated on the influence
of elevated amounts of carbon dioxide on yield, photosynthesis, stomatal responses, and water use, with little
attention to other direct effects such as effects on leaf temperature, crop phenology, water potential, dark
respiration, and photorespiration.
Investigations of the interactions between projected increases in global air temperature and direct effects
of CO2 on plant processes, including effects on economic yield, biomass accumulation, phenology, photosynthesis,
stomatal responses, water use, water-use efficiency, the rise in leaf temperatures seen in all species,
photorespiration, dark respiration, and nitrogen fixation are very limited. There have been a few studies relating
the interactive effects of high levels of carbon dioxide and temperature on field and pot-grown soybean plants,
three on wheat, two on forage grasses, and no temperature-controlled studies on other species.
Research to this point has usually provided windows of knowledge about certain crops at specific stages
of their life cycles. Both direct and climate change effects of high amounts of CO2 are probably quite different
at different stages of development.
Experimental conditions and facilities. The literature is replete with studies lacking proper experimental
design, particularly with regard to replication and randomization. Plants in the same field chamber, indoor
growth chamber, or phytotron at the same time are not replicates of each other, but subsamples of the same
replicate (Lee and Rawlings, 1982). This statistical safeguard was violated more often than it was observed.
Also, very few investigators used proper randomization when assigning treatments in the open field or in
chambers. Lee and Rawlings (1982) found great, nonrandom heterogeneity within and between chambers, which
greatly affected soybean growth (accounted for 50 to 70% of the plant-to-plant variation) unless strict
randomization was observed
Any growth chamber or phytotron results should be examined with respect to light intensity, which is
usually less than that of ambient conditions. Researchers have sometimes assumed that chamber experiments
with longer, less intensive photoperiods are equivalent to shorter times in high-intensity outdoor environments
as long as total photosynthetic energy (photosynthetic photon flux density or PPFD) is equal. Because of the
7-2
-------
Rose
interactions between light intensity and the photosynthetic response to COj, this assumption cannot be made for
CO2 studies.
Field chamber experiments in general lack this light intensity problem, and Havelka et aL (1984b) found
no difference in yield or plant growth between field-grown or field chamber-grown wheat plants. But because
there is no wind, natural air mixing is sometimes not replicated in closed and even some open-top chambers.
Any experiments using pots probably can not be viewed as representative of field soil results because of the
strong effects of root restriction on overall growth and even on the physiological processes involved. (Sionit et
al., 1984; Rogers et al., 1986). Low-density canopies (individual plants) also do not represent typical field
conditions where intraspecific competition takes place.
RECOMMENDATIONS FOR RESEARCH
Future research should be more comprehensive and multidisdplinaiy, integrated with modeling efforts
and the experiments of other research teams. To account for interactions with temperature increases predicted
by global climate change models, experiments should investigate the effects of carbon dioxide enrichment over
a range of temperatures, including high temperatures. To minimize experimental artifacts, techniques should
emphasize field research using randomly placed open-top chambers with crops planted directly in the field soil.
Calculations indicate that such techniques are physically and economically feasible.
Major food crops that have had few experimental results published, including rice, corn, wheat, potatoes
and sorghum, should be investigated for response to both elevated CO2 concentrations and rising temperatures.
Research on plant processes, which has concentrated on photosynthesis, should be broadened to include other
metabolic events that affect yield.
Researchers should study the interactive effects of elevated carbon dioxide concentration and temperature
over the whole life cycle of the plant, with plants grown under different water and nutritional regimes rather than
with only "pampered" plants under optimal conditions. Then more realistic predictions of crop response to the
combined climatic and physiological effects of rising CO2 may be made.
With more data about these processes and better understanding of them, we can give a firmer foundation
to crop growth models. These models in turn can integrate knowledge about CO2 enrichment and global
warming and refine their predictions of plant responses to the changing atmosphere.
7-3
-------
Rose
CHAPTER 2
PLANT PROCESSES AFFECTED BY INCREASING CARBON DIOXIDE
YIELD
Yield can be expressed either as biomass accumulation (increases in total plant dry weight) or as economic
yield (weight of seeds, fruits, cotton lint, soybean oil, wood, etc.). Yield is affected by dozens of different plant
processes, including but not limited to photosynthesis, development of vegetative or reproductive organs (the
latter being called yield components), water relations, and nutrient uptake, as well as the effect of environmental
stresses on such processes and the length of the life cycle.
Economic yield and biomass accumulation. Most research groups have found increases in economic
yield when plants are placed in enriched atmospheres of carbon dioxide at moderate temperatures. One reviewer
who surveyed 70 prior reports calculated an average yield increase for C3 plants of 33% with a doubling of
atmospheric CO2 (Kimball, 1983). Plant dry weight has increased in most species that use the C3 photosynthetic
pathway because of greater photosynthesis (see individual species sections on soybean, wheat, and cotton). Leaf
area increased (Sionit et al., 1981b; Rogers et al., 1983a; Jones et al., 1984,198Sb; Kimball et al., 1984; Marc and
Gifford, 1984; Morison and Gifford, 1984a; Chaudhuri et al., 1987; Wall and Baker, 1987) as did leaf thickness
(Rogers et al., 1983b; Cure et al., 1986,1987). Plants using the C4 pathway of photosynthesis also had higher
yields, but apparently because of more favorable water relations rather than because of effects on carbon fixation
(Carlson and Bazzaz, 1980; Rogers et al., 1983b; King and Greer, 1986). CO, enrichment also appeared to have
little effect on plants using the crassulacean acid metabolism (CAM) photosymhetic pathway (Szarek et al., 1987),
but only one notable crop species, pineapple, uses this pathway.
Yield components. C3 species exposed to high levels of CO, responded by producing more vegetative
or reproductive organs, as well as by enhancing organ size. Increasing CO2 concentration often had a positive
effect on initiation rate of vegetative organs such as leaves in many species (Chaudhuri et al., 1986; Schonfeld,
1987), as well as. on the rate of leaf expansion. The resultant increase in leaf area per plant also further
enhanced overall photosynthate production. Plants also initiated more reproductive structures. The numbers
of flowers per cotton plant (Kimball et al., 1986, 1985, 1984, 1983), of tillers in wheat, which determines the
number of potential heads (Marc and Gifford, 1984; Schonfeld, 1987), and of soybean seeds per plant (Rogers
et al., 1984a; Cure et al., 1986; Sionit et al., 1987) and pods per plant (Ackerson et al., 1984) increased with
exposure to high carbon dioxide concentrations.
Phenology. Crop yield depends in part on the rate of filling pods, seeds, grains, or bolls with products of
photosynthesis and other synthetic pathways, times the duration of the filling periods. Higher temperatures cause
plant phenology (the progress through its life cycle) to speed up, leaving less time in these filling stages for
carbon and nitrogen accumulation. With the rate of net photosynthesis increasing, but the duration of grainfill
shortened, yields may not differ from present levels. Determinant plants, with a fixed length of life cycle, would
not benefit from accelerated phenology; but indeterminant plants, which can continue reproducing indefinitely
if provided with enough nutrients, would be able to produce more if their life cycles were speeded up. Ackerson
et al. (1984) found indeterminate soybeans were more responsive to CO2 enrichment than were determinate
varieties.
King and Greer (1986) reported accelerated phenology in corn was correlated with carbon dioxide
enrichment at moderate temperatures. Floral initiation was advanced in wheat and sunflower, but the rate of
floral differentiation was retarded slightly hi wheat and significantly in sunflower exposed to high concentrations
of CO2 (Marc and Gifford, 1984).
7-4
-------
Rose
PHOTOSYNTHESIS
Biochemically, yield increases depend in part on increases in photosynthesis, the process by which green
plants use light energy to form sugar molecules from carbon dioxide and water. Photosynthesis is usually
expressed as the net CO2 assimilation rate, or the rate at which the gas is extracted from the ah* and added to
the plant mass, minus any CO, lost through respiration or photorespiration.
If the concentration of carbon dioxide in the air rises, so does the diffusion gradient between the outside
of the leaf, which is rich in COj, and the inside of the leaf, where the molecules are continually being used up
by the reaction. As a result, more carbon dioxide will diffuse into the leaf.
We might think of photosynthesis as an assembly line. Since carbon dioxide is one substrate or raw
material for this reaction, the more CO- provided to the plant machinery, the more products (sugars and other
carbohydrates such as starch) could be made, and the more the plant could grow or yield. Accordingly, some
research supports the contention that photosynthesis is expected to increase at moderate to high temperatures
for most C3 plants (see Kimball review, 1983). Because global wanning is expected as atmospheric CO2
increases, this enhancement of photosynthesis may pertain in many environments.
FEEDBACK INHIBITION
However, a number of groups observed that plants acclimated to higher substrate levels and eventually
drifted down to a photosynthetic rate similar to that in plants grown in existing concentrations of carbon dioxide
at low temperatures (Azcon-Bieto, 1983; Coyne and Bradford, 1984; Idso et al., 1987b) and moderate
temperatures (Clough et al., 1981; Azcon-Bieto, 1983; Coyne and Bradford, 1984; Feet, 1984; Sionit et al., 1984;
DeLucia et al, 1985; Sasek et al, 1985; Breen et al., 1986; Rogers et al., 1986). Several mechanisms could have
been involved in the acclimation process.
Carbohydrate accumulation. The carbohydrate end products of the reaction may have built up to a higher
level than could be processed by the plants' metabolizing enzymes, and beyond the ability of the plants to
transport them away from the site of photosynthesis. This would clog the machinery and slow down the process.
Sucrose is the major end product of photosynthesis in most leaves and the form of carbohydrate most
often exported to the rest of the plant (Stitt et al. 1984a). Sucrose can be converted to starch, an insoluble end
product used primarily for growth of shoots (stems and source leaves) in the dark. So carbon partitioning
between the two compounds affects carbon allocation and growth of the various plant parts (Huber, 1983).
The concentration of sucrose in the leaf helps control the rate of translocation from source leaves (Huber,
1983; Huber et al., 1984; Stitt et al., 1984b). Export rises with leaf sucrose level until the latter reaches 12 mg
per dm2, when export levels off (Huber et al., 1984). High leaf sucrose levels, such as were found in plants
grown in high concentrations of CO2 (Ackerson et al., 1984; Gent, 1984; Huber et al., 1984), inhibited the enzyme
that controls its synthesis, sucrose phosphate synthase.
High concentrations of leaf sucrose also promoted starch synthesis. Thus more photosynthate was retained
in the leaf and converted to starch, which cannot be exported from the chloroplast (Huber, 1983; Huber et al.,
1984; Stitt et al., 1984b). Huber et al. (1984) found that essentially all the extra carbon fixed by soybean plants
under carbon dioxide enrichment was partitioned into starch; sucrose export did not increase over plants in
ambient CO2.
The accumulation of carbohydrates has been associated with feedback inhibition of photosynthesis at
various species-dependent temperatures (Clough et al., 1981; Azcon-Bieto, 1983; Peet, 1984; DeLucia et al.,
1985; Sasek et al., 1985; Breen et al., 1986; Sage and Sharkey, 1987). This end-product inhibition could even
cause carbon fixation to decrease relative to current levels in some species and climates as global levels of
carbon dioxide rise.
7-5
-------
Rose
Source-sink relationships. The strength of sinks (growing organs that receive the products of
photosynthesis) has been shown to affect photosynthesis and growth of plants in elevated atmospheres of carbon
dioxide (Clough et al., 1981; Feet, 1984; Sasek et aL, 1985; Cure et aL, 1987). In each case, plants with greater
sink strength had greater photosynthesis. Apparently plants need plenty of places to send their newly synthesized
carbohydrates before they will respond photosynthetically to rising levels of COy wnich 'K Probab'y linked to
carbohydrate accumulation discussed above.
Ackerson et al. (1984) placed soybean plants in enriched atmospheres of carbon dioxide at different stages
of their life cycle. Plants showed seed yield increases only when given extra CO2 at times when there was a
heavy sink load, especially at podfdl. These results suggest that breeders may want to select plants with greater
sink capacity to cope with and to take better advantage of increasing carbon dioxide.
Azcon-Bieto (1983) found that treatments which reduced the rate of translocation of carbohydrates from
mature wheat source leaves to sinks (such as lowering the ambient temperature or chilling the base of the leaf)
produced marked reductions in carbon uptake at CO2 partial pressures of 340, 700, or 825 /jbars.
The source-sink ratio affected enzymes that control the concentrations of sucrose and starch in the source
leaves (Rufty and Huber, 1983), which in turn could control photosynthesis and translocation (see above).
Carbon dioxide enrichment also reduced the sucrose uptake capacity of field-grown sugar beet taproots, which
could limit carbon flux from the source to this sink (Wyse and Saftner, 1982), again causing carbohydrate
accumulation in the source leaves. However, translocation from source leaves in two ecotypes of barnyard grass
(a C4 plant) was enhanced by high levels of CO2. CO, enrichment also helped the plants overcome inhibition
of export at cool temperatures. High rates of export, wnich reduced leaf sucrose concentration, were associated
with higher net photosynthesis (Potvin et al., 1984).
Experimental conditions. The buildup of carbohydrates in storage organs (such as roots, tubers, or bulbs)
differs in field-grown and pot-grown plants, and affects feedback inhibition of photosynthesis. There are also
other differences, such as plant reactions to soil strength or nutrients and presence of intraspecific competition,
between the two conditions. In two careful experiments, field soil-grown soybeans did not show the same
increase in photosynthesis under very high levels of CO, as seen in plants grown outdoors in pots (Sionit et al.,
1984; Rogers et al., 1986). Therefore care must be taken to express uncertainty when extrapolating from
experiments done in pots to projected results in the field.
Photosynthesis is also known to be affected by transpiration (the loss of water by the plant), which in turn
depends on several environmental factors, such as humidity, wind speed, temperature, soil water, and solar
radiation. Although evapotranspiration rates in closed chambers have been measured as comparable to those
in the open field (Jones et aL, 1984; 1985a,b,c,d), experiments done in closed chambers usually do not duplicate
outside wind speed and probably do not have internally uniform temperature and humidity (Lee and Rawlings,
1982), and therefore must also be interpreted with caution.
Photosvnthetic regulation. Alternatively, just because they have more raw material to work with, the very
sensitive machinery of photosynthesis (enzymes) may not be able to work faster, especially at temperatures that
are too low or too high for their optimum functioning (Hofstra and Hesketh, 1975). There could be another type
of adjustment (not depending on carbohydrate accumulation) of enzyme activity to keep the plant operating
homeostatically. Porter and Grodzinski (1984) found a drop in activity of two enzymes of carbon fixation,
carbonic anhydrase and ribulose bisphosphate carboxylase (rubisco), in 21-day-old bean plants after 7 days'
exposure to 1200 /iL/L CO,, when they had demonstrabry larger leaf area and dry weight than plants grown in
330 /iL/L CO2. Vu et aL (1983), working with field-grown soybeans, also found extractable carboxylation activity
of rubisco, the chief photosynthetic enzyme, reduced up to 22% early in the season by carbon dioxide enrichment
The substrate for the enzyme (other than CO2X ribulose bisphosphate, built up in these leaves, indicating that
the plant machinery could not handle all the extra carbon dioxide (Vu et aL, 1983).
Among some natural species and cultivated varieties, photosynthesis is maintained at similar rates over a
wide range of environmental conditions (Holtum et al., 1983; Coyne and Bradford, 1984; Korner et aL, 1968;
-------
Rose
Sasek et al., 1985; Sionit et al., 1984; Rogers et at, 1986). Korner et aL (1988) found that plants at high altitudes,
where the carbon dioxide concentration is low, maintained a rate of carbon assimilation relative to their internal
CO2 concentration that was similar to that in low-altitude plants. In other words, the plants adjusted for the
lower CO2 concentration by becoming more efficient at trapping carbon dioxide.
Similarly, Holtum et al. (1983) found that stomatal aperture in the CAM plant Kalanchoe daigremontiana
adjusted to changing CO2 partial pressures to maintain an approximately constant ratio of diffusion resistance
to carboxylation (nonstomatal) resistance in a variety of environmental conditions. Over the entire dark period,
net carbon uptake was similar for plants in 100,330, or 1000 /ibar CO,. Malate accumulated to the same level
in every treatment, and then CO, assimilation shut off. So carbon uptake was not controlled by external CO2
level but by "innate plant control?
Coyne and Bradford (1984) observed nonstomatal resistance to carbon assimilation in Caucasian Bluestem,
a C4 perennial grass, increased greatly as CO2 concentration rose, especially at low temperatures (15°C).
Machler et al. (1986) found active inorganic carbon uptake in Trifolium repens L, a C3 clover, was more efficient
at lower partial pressures of CO2. Plants often seem to have their own optimal photosynthetic rate, and they
will adjust amounts or activities of enzymes to maintain this homeostasis.
Some investigators expect a damping of the effects of increased CO2 concentrations on photosynthesis,
with long-term acclimation and over a large land area, compared to the dramatic photosynthetic increases seen
in short-term chamber experiments (Cough et al., 1981; Azcon-Bieto, 1983; De Lucia et al., 1985; Sionit et aL,
1984; Rogers et al., 1986). Others have found no decline in photosynthetic rates with prolonged exposure to high
levels of CO2 (e.g., soybean, Jones et al., 1985a,b,q cotton, Radin et al., 1987). Until comprehensive experiments
are done accounting for the numerous factors influencing plant growth, it is uncertain whether photosynthesis
will increase greatly, slightly, or not at all among the wide range of crop species.
OTHER CARBON-EXCHANGE REACTIONS
Daik respiration. As discussed above, the apparent rate of net photosynthesis is affected by reactions
where the plant loses carbon dioxide. Dark respiration (the "burning" or oxidation of sugars as fuel to yield
energy) has been reported both to decrease (Reuveni and Gale, 1985) and to increase (Jones et al., 1985c) as
CO2 concentration rises, thus producing an uncertain effect on net carbon fixation.
Giflbrd et al. (1985) found the response differed with species. Respiration in wheat roots or whole plants
grown in high partial pressures of carbon dioxide declined up to 50%, and the contribution of the wasteful
alternate pathway was reduced. However, mung bean root respiration was unaffected by, and sunflower root
cytochrome respiration was increased by, CO, enrichment. Any gain in carbon balance created by a reduction
of respiration could be offset by a speed-up in dark respiration induced by higher temperatures.
Photorespiration. Rising CO, and warmer temperatures also have opposite effects on photorespiration
(the oxygen-dependent inhibition of photosynthesis found only in C3 plants). Increasing the concentration of
CO2 itself caused photorespiration to fall, especially as a percentage of net carbon fixation (Peterson, 1983;
Vines et al., 1983; Valle et al., 1985a; Kobza and Edwards, 1987), as the rubisco enzyme shifted from oxygenase
to carboxylase (Vines et al., 1983). But this effect leveled off at CO, concentrations above 600 pL/L (Peterson,
1983).
On the other hand, higher temperatures, especially those above 40°C (Wynn, 1981; Monson et al., 1982),
enhanced photorespiration (Wynn, 1981; Monson et al., 1982; Azcon-Bieto, 1983; Perry et al., 1983; Espie and
Colman, 1987; Kobza and Edwards, 1987). In cotton, photorespiration used up half the CO2 fixed in net
photosynthesis at 40°C (Perry et al., 1983).
Extractable activity of rubisco decreased as temperatures increased from 15 to 45°C in wheat (Kobza and
Edwards, 1987). The compensation point for CO2 (the concentration at which the rate of photosynthetic CO2
7-7
-------
Rose
uptake equals the release of CO, from photorespiration and dark respiration) rose with temperature, indicating
weaker binding between CO2 and rubisco at hotter temperatures (Monson et aL, 1982; Espie and Colman, 1987).
Kobza and Edwards considered photosynthesis at supraoptimal temeratures was mainly limited by the activation
state of rubisco and by the rise in the oxygen/carbon dioxide solubility ratio.
Photosynthesis decreased greatly from 25 to 45"C, even under the nonphotorespiring conditions of CO2
at 800 nL/L and 2% Oy for wheat (Azcon-Bieto, 1983; Kobza and Edwards, 1987) and for Agropvron smithu.
another C3 grass (Monson et aL, 1982). So while it is not dear which effect will prevail, work done with both
high CO2 concentrations and high temperatures (Monson et aL, 1982; Azcon-Bieto, 1983; Kobza and Edwards,
1987) suggests photorespiration will increase with climate change, causing the rate of carbon fixation to be
somewhat less than expected.
STOMATAL RESPONSES
Effects of enhanced carbon dioxide concentrations on behavior of stomata (the leaf pores) can be reported
in terms of stomatal conductance, a measure of the ease with which water vapor passes out of the leaf, or in
terms of its reciprocal, stomatal resistance, a measure of the difficulty with which water vapor exits the leaf. Both
depend upon degree of stomatal aperture and are generally thought to be affected by evaporative demand, which
increases with higher vapor pressure deficit (VPD, the difference between saturation vapor pressure at that
temperature and actual vapor pressure). Recently Idso (1987) and Idso and Allen (1988) have cast doubt on the
latter theory, contending that stomata are insensitive to VPD and respond only to internal indicators of water
stress. Idso and Allen viewed results which contrasted with this theory as artifacts of porometry.
For C3 plants, it is not clear whether or not there are significant increases in stomatal closure due to
enriched atmospheres of CO2. The response seems to depend upon the species, the dryness of the ah*, and
whether or not the plant is undergoing water deficit stress (see sections on soybeans, wheat, and cotton). In
general, C4 plants seem to respond to higher levels of carbon dioxide by dosing their stomata (Carlson and
Bazzaz, 1980; Morison and Gifford, 1983; 1984; Rogers et al., 1983a,b; King and Greer, 1986).
Morison and Gifford (1983) seem to resolve some confusion (but not the claims of Idso and Allen) by
demostrating that two C3 spedes were as sensitive to increasing carbon dioxide concentrations as two C4 species.
Sensitivity of stomatal conductance to changes in CO- concentration were linearly proportional to the magnitude
of stomatal conductance, as determined by VPD, with the same slope for all four spedes. In other words, at low
concentrations of COg, low VPD and high conductances, sensitivity to changes in the level of CO2 was greatest.
In a later review, Monson (1985) analyzed others' data and determined that with a few exceptions related to the
internal status of the plant (such as hormone concentration), their observations also fit this slope.
Surveying the response of 16 agricultural spedes to carbon dioxide enrichment, Morison and Gifford
(1984) found stomatal conductance decreased by an average of 36%. Several researchers have found the internal
carbon dioxide concentration of the plant, which is affected both by ambient CO2 levels and by CO2 uptake in
photosynthesis, was more important in determining stomatal behavior than was ambient CO2 concentration
(Holtum et aL, 1983; Morison and Gifford, 1983; Mott, 1988). Increasing air and leaf temperatures caused
stomatal closure in a number of spedes (Monson et aL, 1982; Idso et aL, 1987b), which affected transpiration
and water use.
TRANSPIRATION AND WATER USE
Transpiration measures how much water exits the plant. Both transpiration and water use by C3 plants
in enriched atmospheres of CO2 have been seen to increase, to decrease, or to remain the same when compared
to control plants. Stomatal closure, seen in some C3 spedes and in almost all C4 spedes (see Gifford and
Morison, 1985, for an exception), can result in water savings. For spedes using the C4 photosynthetic pathway,
decreases in transpiration due to closure of stomata will probably be the most significant positive effect of
7-8
-------
Rose
elevated amounts of CO2 (Carlson and Bazzaz, 1980; Morison and Gifford, 1983; Rogers et al., 1983a,b; King
and Greer, 1986). However, the predicted global temperature increase may cancel out this effect by increasing
vapor pressure deficit and therefore demand for transpiration, unless the Idso and Allen (1988) hypothesis is
correct.
But even for plants that show increased stomatal resistance upon exposure to high concentrations of CO,*
transpiration per plant or area of ground may not decrease or may even increase. Plants grown in high levels of
carbon dioxide always had higher leaf temperatures, and almost always a greater leaf area than plants grown in
existing atmospheric concentrations of carbon dioxide (Morison and Gifford, 1984). Both these factors increased
transpiration, elevating water use. The outcome in the case of global rise in amounts of carbon dioxide is unclear
and will depend on species, relative humidity or vapor pressure deficit, leaf temperature, and soil moisture.
LEAF TEMPERATURE
Transpiration cools plants by using heat energy to change its water from a liquid state in the leaf to water
vapor. Partial closure of the stomata restricts the transpiration rate and causes leaf temperature to rise. Leaf
temperatures have risen 1 to 3°C or more with exposure to elevated concentrations of carbon dioxide in all
species surveyed thus far (Perry et al., 1983; Morison and Gifford, 1984; Rogers et al., 1984b; Kimball et aL,
1985; Chaudhuri et al., 1986,1987; Idso et al., 1987a,b). These hotter leaf temperatures are the one phenomenon
universally observed when high levels of CO2 are applied.
These results logically follow the stomatal closure seen in some cases with high concentrations of CO,.
But more surprisingly, leaf temperatures have been observed to rise when there was no stomatal closure, by
researchers working with soybeans (Rogers et ah, 1984b), wheat (Chaudhuri et al., 1986, 1987), and cotton
(Kimball et aL, 1985).
When combined with the predicted warmer air temperatures, these hotter leaf temperatures could lead
to biochemically based inhibition of growth and yield as sensitive enzymes react to potentially denaturing
conditions (Hofstra and Hesketh, 1975; Monson et al., 1982; Azcon-Bieto, 1983; Kobza and Edwards, 1987).
As higher temperatures and vapor pressure deficit cause greater evaporation from the soil, more severe drought
stress could combine with these high leaf temperatures to create damage too great to be offset by any CO2
enrichment-induced yield increase.
TEMPERATURE INTERACTIONS
Because rising levels of CO, are predicted to cause a global climate change, it is very important that we
investigate the effects of increased carbon dioxide concentration as it interacts with the concomitant warming.
However, experimental research has not dealt frequently with simultaneous indirect and direct effects on plants,
so it is difficult to assess the relative contributions of the two variables, CO2 enrichment and higher temperatures,
to predicted responses of plants.
High temperatures alone inhibited photosynthesis and other enzymatic reactions in CAM plants (Martin
and Siedow, 1981) as well as in C3 plants, where photorespiration increased with temperature, especially above
40°C (Wynn, 1981; Monson et al., 1982; Perry et al., 1983; Espie and Colman, 1987; Kobza and Edwards, 1987).
As temperatures rose, binding between CO, and rubisco weakened (Monson et al., 1982; Espie and dolman,
1987).
More relevant studies examined the effects of several CO, concentrations at various temperatures. Wheat
photosynthesis was optimal at 25°C and decreased greatly at 3frC and 45°C, even under the nonphotorespiring
conditions of CO2 at 800 pL/L and 2% O, (Azcon-Bieto, 1983; Kobza and Edwards, 1987). Similar inhibitory
results of high temperature on photosynthesis despite elevated carbon dioxide concentrations were found in
Agropvron smithii. another C3 grass (Monson et al., 1982).
7-9
-------
Rose
Temperature sensitivity may reside in photosynthetic enzymes, which also change with exposure to high
concentrations of CO2. At temperatures above 40°C, irreversible enzyme damage and uncoupling of
photophosphorylation occurs in both species mentioned above (Monson et al., 1982; Azcon-Bieto, 1983; Kobza
and Edwards, 1987). Extractable activity of nibisco and fructose 1,6-bisphosphatase decreased as temperatures
increased from 15 to 45°C in wheat (Kobza and Edwards, 1987). Kobza and Edwards considered photosynthesis
at supraoptimal temperatures to be mainly limited by the activation state of rubisco and by the rise in the
oxygen/carbon dioxide solubility ratio.
In several cases, increasing air temperatures were related to intensification of CO2 enrichment effects,
including rising leaf temperatures (Chaudhuri et al., 1987) and enhanced yields (Idso et al., 1987a,b; Kimball
et al., 1986). However, in at least two crops, higher air temperatures combined with high CO, concentrations
did not favorably affect plant development and yield over the whole season in spring wheat (Wall and Baker,
1987) or for several days in soybean (Jones et al., 1985a).
In Trifolium repens L., a C3 clover, CO2 enrichment and increasing temperatures had opposite effects
on CO2 entry into the plant. Active inorganic carbon uptake was more efficient at 30°C than at 10°C, but was
inhibited by increasing concentrations of CO2 (Machler et al., 1986). Machler et al. interpreted the temperature
phenomenon to the greater permeability of membranes at higher temperatures. Jurik et al. (1984) found the
temperature optimum for carbon exchange in bigtooth aspen shifted from 2S°C in 320 /iL/L CO2 to 37°C in
1900 /iL/L CO,. However, the initial slope of the CO, response curve of CER at 20°C leaf temperature did
not differ from that at 30°C (Jurik et al., 1984).
Overall, when plants were given both high CO- concentrations and high temperatures, it seems the
deleterious effects of high temperature prevailed over the beneficial effects of carbon dioxide enrichment.
DROUGHT STRESS
Drought can consist of episodes of water deficit stress which vary in severity, duration, and stage of the
plants' life cycles at which they occur. These latter factors affect the plants' responses to water deficits. Because
of these complexities, it is not usually possible to directly compare research studies. However, a general pattern
of plant behavior that occurs when plants are deprived of water and given enriched atmospheres of CO2 does
emerge.
Depriving plants of adequate water had a very significant influence on their reaction to increased amounts
of carbon dioxide. In general, drought-stressed plants seemed to benefit more than did well-watered plants from
additional amounts of CO2. Drought-stressed plants exposed to high concentrations of carbon dioxide
maintained yields higher than those of nonstressed plants at current levels of CO2 (Gilford, 1979; Jones 1985d;
Kimball et al., 1986). Water deficits also affected stomatal reaction to CO- enrichment of air (Rogers et aL,
1984b).
But for some species, there is evidence that depending on the severity and duration of the water deficit,
the drought-induced inhibition of photosynthesis may not be overcome by photosynthetic increases due to
enhanced CO, (Gifford and Morison, 1985; King and Greer, 1986; Chaudhuri et al., 1987, Schonfeld 1987; Vu
et al., 1987).
Rogers et al. (1986), noting that plants confronted with nutrient deficiencies or high leaf temperatures
may not be able to benefit from extra carbon dioxide, tentatively suggested that drought stress is the only stress
that does not cancel enriched-CO2 enhancement of growth and yield
7-10
-------
Rose
WATER-USE EFFICIENCY
Water-use efficiency (WUE) is a measure of the expenditure of water necessary to produce a certain
amount of yield, which can be expressed as grain yield, gross biomass accumulation, or carbon dioxide molecules
assimilated. In all cases studied, water-use efficiency has improved as CO, concentration has increased. For C3
species, WUE rose because high photosynthesis increased yield. For C4 species, WUE improved because
stomatal closure caused less water to be expended (Morison and Gilford, 1984).
7-11
-------
Rose
CHAPTERS
SOYBEAN
Soybean (Glvcine max L.), the most important oilseed crop in the world, is a C3 legume that originated
in China. Acreage in the VS., which produces half the world's supply, has increased enormously since 1945
(Wilcox and Leffel, 1987). Soybean is grown for a number of products, including oil and its byproducts such as
plastics, fermented bean for tofu and soy sauce, vegetable protein for processed foods, and animal feeds
(Coulman, 1987; Roecklein and Leung, 1987; Wilcox and Leffel, 1987).
The effects of elevated atmospheres of carbon dioxide have been studied more extensively on soybean
than on any other crop. Despite this fact or perhaps because of it, the findings remain contradictory and
confusing. Some of the discrepancies in results may be due to differences in experimental conditions that are
known to influence responses to CO2.
YIELD
Biomass accumulation. In all cases studied, raising the concentration of carbon dioxide causes increases
in the total soybean plant mass at harvest (Carlson and Bazzaz, 1980; Clough et al., 1981; Finn and Brun, 1982;
Rogers et al., 1983a,b, 1984a, 1986; Sionit, 1983; Havelka et al., 1984a; Jones et al., 198Sd; Cure et aL, 1986,
1987; Allen et aL, 1987, 1988). These increases in whole-plant dry weight relative to control were seen
throughout the life of the plant (Allen et al., 1987; Cure et al., 1987).
The first increments of added carbon dioxide (i.e., enriching the concentration of CO2 from 332 to 428
/iL/L rather than from 772 to 910 /iL/L) make the most difference in biomass, but the maximum increase in
total vegetative dry weight was 66% at 910 /iL/L (Rogers et aL, 1984a).
One group reported that the dry weight increase due to enriched atmospheres of carbon dioxide was
proportionate among vegetative plant parts (Rogers et al., 1984a), but many investigators have reported increased
leaf weight (Finn and Brun, 1982), leaf area (Rogers et al., 1983a; Ackerson et al., 1984; Jones et aL, 1984,1985b)
and leaf thickness (Rogers et aL, 1983b; Cure et al., 1986,1987), including an extra layer of photosynthesizing
palisade mesophyU cells (Rogers et aL, 1983b; Campbell et al., 1987).
Yield components and seed yield. Plants grown in atmospheres elevated in carbon dioxide had higher
seed yields (Rogers et al., 1983b, 1984a; Ackerson et al., 1984; Havelka et al., 1984a; Jones et al., 1985d; Cure
et aL, 1986; Allen et aL, 1987; Sionit et al., 1987) by increasing branching (Ackerson et aL, 1984), the number
of seeds per plant (Havelka et aL, 1984a; Rogers et aL, 1984a; Cure et al., 1986; Sionit et aL, 1987) or pods per
plant (Ackerson et al., 1984), and sometimes raised the weight per seed as weU (Cure et aL, 1986).
However, elevated biomass did not always translate into increases in soybean seed yield (Rogers et al.,
1986), because the harvest index (percent of mature plant weight that is seed yield) tended to decline with
carbon dioxide enrichment (Rogers et al., 1984a, 1986; Cure et aL, 1986), although Havelka et aL (1984) found
no change hi harvest index. Allen et aL, (1988) also noted a slight decrease in harvest index, but contended that
the main effect of CO, enrichment seemed to be increasing photoassimilation by soybean canopies while
maintaining consistent allometric relationships (Le, similar relative weight gains by various plant parts). They
found few differences in partitioning between CO2 treatments from 330 to 800 /imol COj/mol air. High
concentrations of carbon dioxide also had deleterious effects on the percent protein of seeds (Rogers et aL,
1984a).
Phenology. Growth response was greatest in the very youngest plants (Cure et aL, 1986; Rogers et aL,
1984a) with the relative growth rate (RGR, dry matter accumulation per unit dry matter) increasing
7-12
-------
Rose
asymptotically with carbon dioxide concentration from S days to 2 weeks after planting. Later, the RGR
remained the same for all CO2 treatments when plants were 2 to 12 weeks old (Rogers et al., 1984a).
The relative carbon dioxide response (percent growth increase due to higher levels of CO2) was cumulative
but the rate progressively decreased while the net assimilation rate (NAR » dry matter accumulation per leaf
area) in the plants exposed to high amounts of CO2 remained above that in the control plants. Later, the dry
matter accumulation for different carbon dioxide treatments was similar, causing the relative response rate to
decline (Cure et al., 1986).
PHOTOSYNTHESIS
In general, researchers have found that exposing soybeans to elevated concentrations of carbon dioxide
causes increases in photosynthesis on a leaf area basis (Gough et al., 1981; Finn and Brun, 1982; Rogers et al.,
1983a,b; Ackerson et al., 1984; Havelka et al., 1984a; Hesketh et al., 1984; Huber et al., 1984; Peet, 1984; Valle
et al., 1985a; Campbell et al., 1987; Sionit et al., 1987) or on a plant canopy basis (Ackerson et al., 1984; Rogers
et al., 1984a; Jones et al., 1985a,b,c,d; Cure et al., 1986; 1987; Allen et al., 1987).
Jones et al. (1985a,b,c) conducted a series of experiments in which plants were grown for about half their
growing seasons in either a high or a low level of CO2 and then switched for a few days to the other
concentration. In each case, canopy photosynthesis (net carbon exchange rate, CER) immediately responded
to the new levels of carbon dioxide. In contrast, Clough et al. (1981) observed that plants grown in ambient
levels of CO2 had higher photosynthetic rates than plants grown in 1000 /iL/L CO2 when measured at either
concentration.
Campbell et al. (1987) studied the photosynthetic increases in plants grown or measured at high partial
pressures of CO2- With either short- or long-term exposure, they found no increases in either the initial or total
in vitro activity of the main enzyme of carbon fixation, ribulose bisphoshate carboxylase (rubisco). Vu et al.
(1983,1987) actually found decreases in these activities with exposure to high levels of carbon dioxide. While
an early study reported an increase in the enzyme's regenerated substrate other than CO,, ribulose bisphosphate
(RuBP) (Vu et al., 1983), a later study by the same group found no change in concentration of this substrate on
exposure to additional carbon dioxide (Campbell et al., 1987). Certainly this does not rule out different in vivo
activities.
Campbell et al. searched for other bases for the photosynthetic differences seen in plants grown in high
and low concentrations of CO^. When carbon dioxide levels were high, the mesophyil cells in the palisade layer
of the leaf proliferated, increasing the internal surface area for gas exchange and the number of chloroplasts for
photosynthetic reactions. They also found more sinks (pods) on the plants grown in high concentrations of COg,
which they felt would enhance assimilate demand, which in turn usually causes a rise in photosynthesis.
Valle et al. (1985a) also found reasons for the enhanced reaction. Soybeans adapted to 660 /unol
COg/mol air and measured at seedfill utilized both light and carbon dioxide more efficiently at elevated CO2
ana through all light levels than leaflets grown at 330 /unol/mol. Binding of CO, to the enzyme rubisco was
tighter in the plants adapted to high amounts of CO2 than in the plants grown in ambient levels of carbon
dioxide, indicating a suppression of the harmful photorespiratory reaction in the former leaflets. On the other
hand, dark respiration was seen to increase in soybeans in elevated atmospheres of CO2 (Jones et al., 1985c).
Experimental conditions. In two careful experiments, Rogers et al. (1986) and Sionit et al. (1984) placed
open-top chambers randomly on preplanted fields. They found that soybean plants grown in open field soil did
not respond at all in photosynthesis (mg CO2 fixed/dm7 h) or growth (net assimilation rate, NAR - dry matter
accumulation per leaf area) to increasing levels of carbon dioxide. Their potted counterparts in the same field
chambers increased photosynthetic rates along with rising levels of COL just as had all other experimental plants
(Sionit et al., 1984), but responded in NAR to elevated carbon dioxide concentrations only when very young
(Rogers et al., 1986).
7-13
-------
Rose
Sionit et al. (1984) felt that there was some harmful effect of confining the roots to a small volume which
altered the plants' response, whereas Rogers et al. (1986) felt some stresses of the field (perhaps low nutrition)
rendered the plants in open-field soil unable to respond to the CO2 enrichment. Whichever explanation is
correct, it is clear that plants in pots cannot serve as models for photosynthetic responses to high concentrations
of carbon dioxide.
More puzzling are the discrepancies between these results and those of many research groups using
soybean planted directly into the soil in closed-top field chambers. The latter have consistently observed
photosynthetic and yield increases in soybeans exposed to rising concentrations of CO,. The closed chambers
seem to have evapotranspiration comparable to that observed in open field situations (Jones et al., 1984,1985a,
b,c,d); but at least in the case of the University of Florida SPAR units, the chambers are placed on reconstituted
~ and therefore not natural — soils, and planting soybean in predetermined locations probably affects
randomization of treatments.
FEEDBACK INHIBITION
Clough et al. (1981) and Peet (1984) found rapid and continuous decreases in photosynthesis per plant
in soybean plants grown in a phytotron in 1000 pL/L CO2. This rate stayed higher than photosynthesis per plant
of plants grown in 350 /*L/L until senescence for Peet (1984) but not for Clough et aL (1981). Jones et al.
(1985c) disagreed with the latter, based on the results of their switch experiments (see above). In contrast to
suggestions that the response to carbon dioxide enrichment could show feedback inhibition, they saw a consistent
positive response of soybean canopy CER to CO, concentration during vegetative growth, regardless of previous
CO, treatment levels. Valle et al. (1985a) and Campbell et aL (1987) found similar results when they measured
photosynthesis in soybean leaflets.
Source-sink relationships. Clough et aL (1981) observed that high-sink plants (those with a high podleaf
ratio) had higher levels of photosynthesis per plant than low-sink plants, whether grown at ambient or high levels
of CO2. However, the dry weight of all plant parts was higher in the low-sink plants. In contrast, Peet (1984)
decreased the pod:leaf ratio from 15:1 to 5:1 and saw no effect on photosynthesis/unit area. When she raised
the COg concentration from 350 to 1000 /JL/L, the total weight and pod weight in 15:1 plants (high-sink or not
sink-limited) increased. But high partial pressures of carbon dioxide had no effect on total weight or pod weight
in 5:1 (sink-limited) plants (Peet, 1984). Ackerson et aL (1984) observed similar results in field-grown soybeans:
plants increased their seed yields only when the elevated atmospheres of CO, were applied at times of high sink
demand, especially at podfill. In other words, plant response to carbon dioxide enrichment during reproductive
growth seemed to depend on the presence of plenty of sinks that could accept newly synthesized carbohydrates.
Carbohydrate accumulation and export. Cure et al. (1987) examined assimilate utilization in soybean
plants grown in elevated atmospheres of carbon dioxide. They felt that the plants mounted a defense against
feedback inhibition of photosynthesis by minimizing the buildup of photoassimilates and adjusting rates of export
from the source leaves according to the strength of both sources and sinks.
Initially (within 2 days of being placed in enriched atmospheres of carbon dioxide) dry weight increases
in 13-day-old vegetative soybean plants exposed to high concentrations of CO2 were associated with
accumulations of high levels of nonstructural carbohydrates in leaves and increases in specific leaf weight (weight
per leaf area). After 3 days, dry weight began to increase rapidly in stems and roots, which was associated with
an early increase in the estimated rate of assimilate utilization in the dark. After this point, specific leaf weight
of plants grown in high partial pressures of CO2 no longer increased relative to that of controls.
Later, assimilate utilization in the light increased as sinks, including stems and roots, began to grow.
Daytime export from source leaves was controlled by sucrose biosynthesis. NAR declined in both carbon dioxide
treatments but remained higher at 700 /iL/L than at 350 /*L/L throughout the experiment. Total leaf area of
plants in carbon dioxide enrichment did not increase significantly over controls until 13 days after the start of
treatment.
7-14
-------
Rose
Cure et al. (1987) contended that dark mobilization of assimilates responded to variations in concentrations
of carbohydrates in source leaves. Export of assimilates in light seemed to adjust more slowly to buildup of
carbohydrates and appeared to be sink-determined, increasing only as these other organs developed throughout
the plant.
The observations of Ackerson et al. (1984), Havelka et aL (1984a), and Huber et al. (1984) on older
soybean plants differ slightly from those of Cure et al. on young vegetative plants (1987). Field-grown soybeans
raised from 22 days after planting in 1232 pL/L CO2 had higher levels of leaf sucrose and glucose than did
control plants until the time of flower bud development, when leaf carbohydrates were depleted (Ackerson et
al., 1984; Havelka et al., 1984a). Plants grown in high levels of carbon dioxide also had large pools of starch and
sucrose in their leaves during seedfill, suggesting that these carbohydrates were not used for seed development
In contrast, control plants were almost devoid of sucrose and starch at plant maturity (Havelka et al., 1984a).
Huber et al. (1984) saw that plants grown in the field at 649 pL/L CO2 had higher photosynthetic rates
but rates of sucrose export and activities of sucrose phosphate synthase that were similar to control plants.
Essentially all the extra carbon in plants grown in elevated partial pressures of CO2 was partitioned into starch.
Sucrose export from the leaves rose with sucrose concentration until it readied 12 mg/dm , after which export
remained constant (Huber et al., 1984).
The average daytime content of leaf starch rose with increasing carbon dioxide concentration (Ackerson
et al., 1984; Havelka et al., 1984a; Huber et al., 1984; Rogers et al. 1984b; Allen et al. 1988), from 85 g/kg dry
weight at 330 /iinol CO2/mol air to 205 gAg dry weight at 800 /imol CO^/mol air (Allen et al., 1988). But
daytime rate of starch accumulation was the same regardless of carbon dioxide concentration (0.64 g/m hr at
48 days after planting; 0.29 g/m2 hr at 69 days after planting) (Allen et al., 1988), indicating some difference in
night export of starch, as was seen in cotton raised in high CO2 levels (Kimball et al., 1985).
STOMATAL RESPONSES
Most investigators have observed that soybean plants in enriched atmospheres of carbon dioxide undergo
either decreases in stomatal conductance (Rogers et al., 1983a,b, 1984b; Havelka et al., 1984a; Sionit et al.,
1984) or increases in stomatal resistance (Huber et al., 1984; Jones et al., 1984,1985b; Valle et at., 1985b). Bulk
canopy resistance (stomatal resistance and boundary layer resistance, a measure of the unstirred layer of air
just outside the leaf surface) at 800 /imol CCL/mol air was 1.6 times greater than the bulk canopy resistance
at 330 /jmol CO2/mol air (Jones, 1984). But Jones et al. (1985b) found total canopy resistance to transpiration
was equal with different carbon dioxide treatments because the increase in stomatal resistance was offset by
increased leaf area.
Rogers et al. (1984b) found a more complex situation, where high concentrations of carbon dioxide had
opposite effects on stomata depending upon the water status of the soybean plants. When the plants were well-
watered, stomatal conductance declined with rising CO^ but in drought-stressed plants stomatal resistance
declined with rising carbon dioxide.
TRANSPIRATION AND WATER USE
Contradictory results have been obtained with respect to transpiration and water use. Several investigators
did not find differences in water use (Jones 1985a) or transpiration (Jones et al., 1985a,b; Valle et al., 1985b) with
CO, treatment, whether plants were grown or measured at different carbon dioxide concentrations. Valle et aL
(1985b) found that transpiration of leaflets acclimated to high levels of CO2 was similar to leaflets grown in
ambient concentrations of CO2. They concluded that this occurred because enriched atmospheres of carbon
dioxide caused both a rise in stomatal resistance, which lowers transpiration, and a partially offsetting rise in leaf
temperature, which in turn increased the vapor pressure gradient from leaf to air, thus raising transpiration.
7-15
-------
Rose
Others found transpiration inversely related to carbon dioxide concentration (Carlson and Bazzaz, 1980;
Huber et aL, 1984; Jones et aL, 1985q ). Jones and coworkers (1985d), measuring transpiration over an entire
season, found transpiration decreased 10% at 660 /imol COg/mol air compared with that at 330 /unol/mol.
Water use was also seen to decline with high carbon dioxide concentration (Jones et al., 1985b; Rogers et al.,
1983a). In well-watered plants, though plants grown in elevated partial pressures of CO2 had higher leaf areas
per plant, water loss per plant decreased because increasing carbon dioxide lowered the stomatal conductance
(Rogers 1984b).
In one study, transpiration responded immediately to a switch in carbon dioxide level from low to high
or vice-versa (Jones et al., 1985c); in another paper, Jones et al. (1985b) stated that while short-term exposure
to different CO2 treatments adequately estimated long-term canopy CER response to carbon dioxide, it didn't
represent the long-term canopy transpiration response. They said this because before the switch, transpiration
rates in the two treatments were essentially the same, but after the switch, plants going from enriched
atmospheres to current levels of CO2 (which plants had higher leaf area) used 36% more water, while those
going the other way (which had lower leaf area) used 18% less water.
LEAF TEMPERATURES
Regardless of carbon dioxide level during adaptation (whether exposure to high concentrations of CO2
was short- or long-term), both stomatal resistance and leaf temperature rose as carbon dioxide concentration
increased. Leaf temperature of plants at double the ambient partial pressures of CO2 was L5°C greater than
plants at ambient levels of the gas (Valle et aL, 198Sb).
TEMPERATURE INTERACTIONS
Soybean plants grown in high levels of CO2 seemed to respond positively to temperature increases within
moderate limits, but there is no evidence that they will do well at temperatures above 26°C. Sionit et aL (1987)
worked with three sets of day/night temperatures (18/12,22/16, and 26/20°C) and three concentrations of CO2
(350,675, and 1000 /iL/L). No seeds were produced at 18/12° at any carbon dioxide concentration. Both the
number of seeds and the number of pods per plant increased with temperature, so that seed yield increase with
high levels of carbon dioxide was greatest at highest temperatures, especially from 350 to 675 /iL/L CO2.
However, the increase in leaf net photosynthesis in response to increasing carbon dioxide concentrations
was best at the intermediate temperatures. At cool temperatures, the plants had extended photosynthetic
capacity (they put off senescence) but this didn't result in allocating more dry matter to the developing pods.
These data might be interpreted as supportive of the Idso et aL (1987a,b) hypothesis that yield increases with
high partial pressues of the gas are enhanced by high temperatures; but the highest temperatures Sionit et al.
used (26/20°C) were very moderate and may be far less than those predicted by global climate change models.
There is no assurance that the yields would continue to increase at daytime temperatures of 30 or 34°C. In
contrast, Jones et al. (1985a) studied the photosynthetic rate (canopy CER) of soybeans grown in 330 and 800
/jmol CO2/mol air with dry-bulb temperature from 28 to 35°C. Photosynthesis was not significantly changed
by variations in temperature over this range.
DROUGHT STRESS
Depriving soybean plants of adequate water seems to be the most significant influence on their reaction
to increased carbon dioxide. (As mentioned above, the term "drought" is descriptive but not quantitative, and
can refer to events varying in severity, duration, and stage of development at which they are applied.) Though
plants grown in enriched carbon dioxide had greater growth, they had lower rates of water use which delayed
and prevented the onset of severe water deficit stress under low moisture availability (Huber et aL, 1984; Rogers
et aL, 1984b).
7-16
-------
Rose
Water-deficit stress imposed at ambient partial pressures of CO, decreased soybean seed yield and final
biomass. But soybean plants exposed to high concentrations of carbon dioxide and drought-stressed during their
reproductive stages maintained yields higher than those of nonstressed plants at current levels of CO2 (Jones
1985d). However, Vu et al. (1987) found the drought-stressed plants' response to extra carbon dioxide was not
large enough to make up for photosynthetic declines caused by severe drought stress.
Rogers et al. (1984b) found high concentrations of CO, had opposite effects on stomata depending upon
water status of the plants. When they were well-watered, stomatal conductance declined with rising carbon
dioxide, but in water-limited plants stomatal resistance declined with rising CO2-
WATER-USE EFFICIENCY
In all cases reported, water-use efficiency (WUE) of soybeans rose with atmospheric enrichment of
carbon dioxide (Carlson and Bazzaz, 1980; Rogers et al., 1983b; Havelka et al., 1984a; Huber et al., 1984; Jones
et al., 1984; Valle et al., 1985b). In a typical example, in leaves adapted to 660 /imol CO2/mol air, WUE
(measured as CER/transpiration) was twice as high as the WUE of leaves adapted to and measured at 330
/imol/mol, chiefly because of a twofold rise in CER at 660 /imol/mol (Valle et al., 1985b). WUE does not
increase with increasing temperatures, because of the temperature-driven increase in transpiration (Jones et al.,
1985a).
NITROGEN FIXATION AND NUTRIENTS
Despite its enhancement of photosynthesis and therefore of available carbohydrates, carbon dioxide
enrichment appeared to have little short-term effect on nitrogen fixation as measured by total or specific root
nodule activity. Dry weight of roots and nodules and total nodule activity increased only after 16 days of
treatment, long after extra carbon had been partitioned to shoot parts. Specific nodule activity was never
enhanced by high levels of COj, indicating that nodule activity responded to extra CO2 only as a consequence
of the general growth respose of the plant. Nodule activity was not limited by photosynthesis at current CO2
concentrations, but by partitioning of photosynthate (Finn and Brun, 1982).
Soybean plants grown at 675 ppm CO, overcame a low nutrient level (1/8-strength Hoagland's solution)
to produce more biomass, more total seed yield, more pods, and larger weight per seed than plants grown at 350
ppm and half-strength Hoagland's solution (Sionit, 1983).
Plants took better advantage of high partial pressures of carbon dioxide (growth response as NAR was
greatest) when nutrients such as nitrogen and phosphorus were at optimal levels. Conversely, at early podfill they
also used nitrogen and phosphorus most efficiently (add g dry weight per mg nitrogen or P) at higher levels of
COy The same levels of phosphorus nutrition were optimal for both plants grown in ambient levels of CO2 (350
ppm) and those grown in a high CO2 concentration (700 ppm); but plants grown at 700 ppm tended to have
lower leaf phosphorus concentrations. Soybeans at 700 ppm CO, required more nitrogen for optimum growth.
Plants grown in enriched atmospheres of carbon dioxide but provided with inadequate concentrations of nitrogen
or phosphorus increased their uptake of those nutrients, but their growth declined with the concentration of these
nutrients (Cure et al., 1986).
INSECT FEEDING
Feeding rates of soybean looper larvae, an insect herbivore, increased 30% on plants grown in 650 jiL/L
CO2 from those grown at 350 /iL/L. These feeding increases were related to declines in leaf content of the
water and nitrogen, indicating that the insects fed more heavily to make up for the poorer food quality (from
their point of view!) of the plants grown in high levels of CO2 (Lincoln et ah, 1986). (The previously noted
7-17
-------
Rose
increases in leaf starch content would create apparent percentage decreases in leaf nitrogen, water, and protein.)
At 800 pmol COu/mol air, Allen et al. (1988) saw a 45% increase in total plant nitrogen in seed and senescent
tissues at season's end These reports do not necessarily contradict each other; for Lincoln et al. (1986) were
measuring photosynthesizing leaf tissue; and if the carbon-based biomass of the soybean rose more than 45%,
there could be a concomitant rise in the C:N ratio, as was seen by Lincoln et al. (1986).
7-18
-------
Rose
CHAPTER 4
WHEAT
Wheat (Triticum aestivum L.) is the most important food crop in the world, planted on a greater area
and feeding even more people than rice (Coalman, 1987; Roecklein and Leung, 1987; Starling, 1987). Originally
from Afghanistan, wheat is grown on every continent. Major production comes from Asia; the Soviet Union,
which raises 16% of the world's crop (about 80 million tons in 1983); the United States, which produced 66
million metric tons in 1983, half of which was exported; and from Canada, Australia, and Argentina (Coulman,
1987; Roecklein and Leung, 1987).
The high gluten content of wheat permits its dough to adhere and increase in volume during baking,
rendering it the most suitable grain for bread. Other wheat products include all types of baked goods, pasta,
breakfast cereals, gravy, alcohol, starch, and straw for animal bedding (Roecklein and Leung, 1987; Starling,
1987).
Wheat grown in North America falls into two categories. Obligate winter wheat is planted in autumn,
must undergo a vernalization period of low temperatures before it will flower and produce seed, and is harvested
in early to midsummer. Winter wheat descended from Ukranian cultivars brought in the 1880's accounts for two-
thirds of the U.S. crop, and grows in the Great Plains from Texas to North Dakota, the Midwest, the Northwest,
and parts of the East Coast. Since these plants frequently encounter desiccating cold, ice, and snow, they have
different tolerance to environmental conditions at various stages than do spring wheats.
Several species and types of spring wheat are planted in spring and grown in summer in the far North
(Montana, the Dakotas, Wyoming, and western Canada), California, and parts of the East, and grown in winter
in the far Southwest and Mexico. The response of wheat to high carbon dioxide concentration depends in part
upon which of these two types it is. However, results for the two groups with respect to yield, phenology, and
photosynthesis are similar, so findings for most of these will be grouped together.
YIELD
Biomass accumulation and vegetative growth. High carbon dioxide concentrations (450, 485, 660, 675,
825, and 1000 pL/L) cause increases in total dry weight both at maturity (Gifford, 1977; Sionit et al., 1980,
1981a,b; Havelka et al., 1984b; Chaudhuri et al., 1987) and incrementally thoughout the season (Sionit et al.,
1981a; Gifford et al., 1985; Schonfeld, 1987). Winter wheat exhibited dry weight increases in leaves, stems, and
tillers (Havelka et al., 1984b; Marc and Gifford, 1984; Chaudhuri et al., 1987). Both types of wheat showed
increased stem height (Sionit et al., 1981b; Chaudhuri et al., 1986,1987).
Leaf area. Both types also developed greater total leaf area in enriched CO2 concentrations (Sionit et
al., 1981b; Marc and Gifford, 1984; Chaudhuri et al., 1987; Wall and Baker, 1987J. Wall and Baker (1987)
surmised that these vegetative changes gave the plants a more favorable canopy structure for light interception.
In spring wheat, this leaf area increase was accomplished by enlarging leaf length and width, but not number of
leaves (Wall and Baker, 1987).
However, in winter wheat both size (Chaudhuri et al., 1986) and number (Chaudhuri et al., 1986; Schonfeld,
1987) of leaves rose with amount of carbon dioxide, so that by 20 days after CO2 application began again in early
spring (March 8,1986), leaf area had doubled; after 40 days, the number of leaves had doubled (Chaudhuri et
al., 1986, 1987).
Roots. Field-rhizotron-grown winter wheat plants responded to increased CO2 with greater root weight
(Chaudhuri et al., 1986) and root length density (Chaudhuri et al., 1987), which usually helps plants enhance
their water absorption.
7-19
-------
Rose
Assimilate partitioning. When plants were well-watered, spring wheat plants grown in a phytotron showed
an increase in their root-shoot ratio (Sionit et aL, 1981b), but in winter wheat the percent increase in shoot
growth was higher than the percent increase in root growth (Chaudhuri et al., 1986).
In spring wheat there was no change in the ratio of leaf area or leaf dry weight to total dry weight of the
plant; or of leaf area to total leaf dry weight (Sionit et al., 1981b) so all vegetative components seemed to
increase in a coordinated fashion. In contrast, in winter wheat specific leaf weight (weight per unit area)
increases with increasing carbon dioxide (Havelka et al., 1984b; Schonfeld, 1987). The specific leaf weight
increase could have been accomplished by an increase in photosynthesizing cells, by an increase in the amount
of carbohydrates, or by both.
Grain yield and yield components. Grain yield increased with elevated carbon dioxide concentration
(Gifford 1977, 1979; Sionit et al., 1980,1981a,b; Havelka et aL, 1984b; Chaudhuri et al., 1986,1987; Wall and
Baker, 1987). Among changes seen in the yield components were increased numbers of tillers (Gifford, 1977;
Sionit et al., 1980, 1981b; Marc and Gifford, 1984; Chaudhuri et al., 1986, 1987; Schonfeld, 1987; Wall and
Baker, 1987), which doubled in at least one case (Chaudhuri et al., 1986); faster rate of tiller emergence (Gifford,
1977; Sionit et al., 19815); and reduced tiller abortion (Havelka et al., 1984b; Wall and Baker, 1987). (The
number of tillers produced determines the maximum number of heads.)
Increases were seen in the number of heads per m2 (Gifford, 1977; Havelka et al., I984b), number of
grains (Gifford, 1979; Havelka et aL, 1984b; Sionit et al., 1980,1981a,b), mass of grains (Sionit et al., 1980,
1981a,b), or all of these yield components (Chaudhuri et aL, 1987, 1986) through effects on phenology,
photosynthesis, and water relations. CO, enrichment caused an increase in the number of tillers which actually
produced heads (Gifford 1977; Sionit et aL, 1981q Havelka et al., 1984b), but not in the number of spikelets per
head (Sionit et aL, 1981c).
Havelka et aL (1984b) found no effect of carbon dioxide concentration on harvest index (ratio of grain
yield to total biomass), but Chaudhuri et al. (1987) saw greater effects on winter wheat kernel weight than on
roots and phytomass. Wheat yields were most affected by high concentrations of CO2 during the period when
the seed number was determined (Havelka et al., 1984b).
Phenology. Gifford (1977) found rate of wheat leaf emergence unchanged by carbon dioxide concentration,
but higher levels of CO, directly affected flowering, including advancing floral initiation (Marc and Gifford,
1984). However, carbon dioxide enrichment was observed to both accelerate development of the immature heads
(Havelka et al., 1984b) and to retard the rate of differentiation of the primordial inflorescence (Marc and
Gifford, 1984).
PHOTOSYNTHESIS
Net assimilation rate rose with increasing carbon dioxide in spring wheat (Wall and Baker, 1987, Sionit
et al., 1981b) and on a per chlorophyll basis in winter wheat (Azcon-Bieto, 1983; Gifford et aL, 1985; Schonfeld,
1987). Gifford (1977) observed an increase in whole-plant net photosynthesis during CO2 enrichment, but not
in the maximum rate of flag leaf photosynthesis; in contrast, Havelka et aL (1984b) saw a 50% increase in
apparent photosynthesis of the flag leaf (the leaf just under the head, whose photosynthesis has great effect on
grain yield). Plants hi a relatively low-light environment responded more to elevated partial pressures of CO,
than did those in higher light (Gifford, 1977).
FEEDBACK INHIBITION
Carbohydrate accumulation. More reduced carbons were available to wheat grown in elevated atmospheres
of CO2 (Sionit et al., 1980,1981d; Azcon-Bieto, 1983; Havelka et aL, 1984b; Chaudhuri et al, 1966,1987; Wall
7-20
-------
Rose
and Baker, 1987), including a doubling of flag leaf sucrose and starch concentrations prior to seed growth
(Havelka et al., 1984b). The carbohydrate accumulation in leaves was associated with decreased net CO2 uptake,
which recovered after a short dark period during which carbohydrates were removed from the leaves, indicating
feedback inhibition (Azcon-Bieto, 1983). Water-soluble carbohydrates in the whole plant, stem, and leaf at
harvest were higher (Havelka et al., 1984b; Chaudhuri et al., 1987), but roots did not accumulate sugars (Gifford
et al., 1985) in plants raised in enriched-CO2 environments.
Source-sink relationships. High flag leaf sucrose and starch concentrations fell rapidly to control levels
with the onset of seed growth, suggesting a rapid mobilization of photosynthate to the head and developing
seeds (Havelka et al., 1984b). Translocation appeared essential for maintenance of a high photosynthetic rate.
CO2 assimilation declined markedly when treatments that lowered the rate of translocation from mature wheat
source leaves (e.g., lower temperatures or chilling the base of the leaf) were applied to plants exposed to 340,
700, or 825 /ibars CO2 (Azcon-Bieto, 1983).
Photosvnthetic regulation. The rate of net CO2 assimilation fell more sharply with time in wheat plants
treated with high levels of carbon dioxide than in control plants, regardless of temperature (Azcon-Bieto, 1983).
Photosynthetic efficiency (assimilation on a leaf area basis as a function of internal carbon dioxide concentration,
c,) was similar for plants grown in ambient and 700 /
-------
Rose
effect on transpiration declined. Another explanation for the rise in transpiration could be the higher canopy
Geaf) temperature at high carbon dioxide concentration (see below).
LEAF TEMPERATURE
Among the most significant findings about high levels of carbon dioxide and winter wheat are studies of
leaf/canopy temperatures by Chaudhuri et al. (1986, 1987). Leaf temperatures of plants grown in enriched
atmospheres of CO2 were higher than those grown at current levels of CO2 (Chaudhuri et al., 1987). Mean
canopy minus air temperatures for plants from high-CO2 treatments averaged 1.0 to L5°C greater than for
plants from ambient atmospheres of CO2.
This difference was more pronounced (plants from high concentrations of CO2 were 2.8°C hotter than
ambient-grown counterparts) when air temperature was "high1' (27.TC, not usually considered a high
temperature). Between 330 and 825 /*L/L COj, the average difference was 3°C. This was true whether the
wheat plants were well-watered or drought-stressed.
The rise in leaf temperatures seen by Chaudhuri et aL (1986, 1987), when combined with wanner air
temperatures, could lead to biochemically-based inhibition of growth and yield as sensitive enzymes react to
potentially denaturing conditions, as seen when Kobza and Edwards (1987) externally manipulated the leaf
temperature as well as the CO2 concentration (see photorespiration, above, and temperature interactions,
below). These high leaf temperatures could create damage too great to be offset by any CO2-induced yield
increase.
TEMPERATURE INTERACTIONS
Of similar importance are the findings of Wall and Baker regarding temperature interactions in spring
wheat (1987). This preliminary report (abstract only) of their study indicates that they grew plants in outdoor
chambers with day/night temperature of 16/8°, 23/15°, or 30/22°C. Plants grew best at the highest CO, levels
and the lowest temperatures. Higher temperatures caused fewer tillers to form and smaller leaves to form on
those tillers.
As mentioned above, photorespiration increased with temperature, even at 800 fiL/L CO,, reducing net
photosynthesis in wheat leaves exposed to temperatures above 25°C. Photosynthetic inhibition at high
temperatures was associated with irreversible damage to enzymes and electron transport, and uncoupling of
photophosphorylation (Azcon-Bieto, 1983; Kobza and Edwards, 1987). As temperatures rose from 15° to 45"C,
extractable activity of rubisco and fructose 1,6-bisphosphatase decreased (the former by 50%), and the thylakoid
membrane showed symptoms of damage (Kobza and Edwards, 1987). Clearly the predicted global warming
could harm wheat plants despite the presence of additional photosynthetic substrate.
Experimental conditions. Chaudhuri et al. (1987) also discussed technical problems of temperature inside
their chambers. They found that the canopy temperature of any treatment was a few degrees hotter than the
open-air temperature (in contrast to the usual finding of well-watered plants being cooler than the surrounding
air), probably due to a 2 to 3°C higher air temperature inside the chambers than in the open field. This latter
difference was greater at night and at cool temperatures — up to 8°C warmer in the chambers at those times.
Although this may render their data inaccurate representatives of current conditions, it means that their
experiments unwittingly represent more closely the high temperatures expected with a high CO2-induced climate
change.
7-22
-------
Rose
DROUGHT STRESS
The most pronounced differences between spring wheats and winter wheats responding to added COj
are found in interactions between drought stress and high carbon dioxide concentration. Again, this may oe
because "drought" and "water deficit" are descriptive terms and are difficult to quantify or compare across
experiments. Well-watered wheat plants of both types make more carbohydrates when Pm m b0i teweb °J
C02 (Sionit et al., 1980, 1981d; Azcon-Bieto, 1983; Havelka et aL, 1984b; Chaudhuri et al., 1986, 1987; Wall and
Baker, 1987), which might enable them to cope better with subsequent drought cycles.
Soring wheat. Water deficits applied to spring and winter wheat yielded different results. Although spring
wheat plants at both high and low CO, levels produced fewer and smaller grains when drought-stressed, plants
at high concentrations of CO, produced more stems and heads than did plants at ambient levels. In fact,
drought-stressed plants in high partial pressures of CO, had grain yields and total dry weight equal to that of
unstressed plants at current CO2 levels (Sionit et al., 1980).
Sionit and coworkers suggested that CO, enrichment increases the yield potential of water-stressed wheat
plants, probably by increasing the concentration of solutes in leaves and thereby enhancing osmotic adjustment,
which helps plants cope with drought (Sionit et al., 1981d, 1980). Since all their work was conducted with plants
grown in pots in controlled indoor environments, the evapotranspiration and thus many gas exchange phenomena,
including photosynthesis and conductance, may not be indicative of those in high-light and windy field conditions
with unrestricted root zones and variable field soils. Thus these results need to be interpreted with caution
before extrapolating to field predictions for spring wheat.
Winter wheat. The results of Sionit and coworkers are also in contrast to findings for drought-stressed
winter wheat. In all cases drought stress lowered winter wheat yield so much that even the highest CO2 levels
(825 /iL/L) did not restore plant parameters to the levels achieved by nonstressed plants at ambient CO2. This
applied to grain yields (Chaudhuri et al., 1987, 1986), photosynthesis (Schonfeld, 1987), and biomass (Schonfeld,
1987, Chaudhuri et al.. 1987, 1986). Grain yield for nonstressed plants grown at 825 /*L/L CO2 was 30% (1986)
or 60% (1985) higher than plants grown at 330 pL/L. Yield increases for drought-stressed plants furnished with
high carbon dioxide concentrations were 63% (1986) or 55% (1985) more than yields of plants grown in current
CO, levels. Winter wheat plants do not seem to be able to use increased levels of carbon dioxide to make up
for the effect of water stress on yield.
From a somewhat different perspective, Gifford (1979) raised wheat in a phytotron at four different levels
of growth-restricting water supply. Plants given less water had a smaller absolute yield response to CO2
enrichment (590 ftL/L) but a greater response relative to control yields. In an extremely arid treatment, plants
in ambient concentrations of CO2 made no grain, but those at a high concentration did make grain.
As mentioned above, Chaudhuri et al. (1986) also found that stomatal response of plants to enriched
atmospheres of carbon dioxide was dependent upon their water status. In well-watered plants, high concentrations
of CO2 increased stomatal resistance, but in drought-stressed plants, they observed no difference in stomatal
resistance based on CO2 treatment.
WATER-USE EFFICIENCY
Raising the carbon dioxide level also affected the WUE of winter wheat plants. Under high partial
pressures of CO^ less water was needed to make a gram of kernel than at ambient CO2 (Chaudhuri et al.,
1987, 1986). Drought-stressed plants given enriched atmospheres of CO, needed similar amounts of water to
nonstressed plants in ambient levels of CO- to produce kernel mass. Nonstressed plants at 825 pL/L CO-
used 14% less water to make one gram ofkernel than at ambient concentrations of CO,; drought stressed
plants used 23% less. z
7-23
-------
Rose
NUTRITION
Sionit and coworkers (1981a) also investigated interaction of CO2 enrichment of the air and mineral
nutrition. In addition to increasing total dry weight, seed weight, and seed number from those seen in control
plants at each nutrient level, high levels of COjenabled plants to use higher concentrations of nutrients without
the inhibition seen at ambient levels of COg elevated carbon dioxide concentration was seen to increase the
grain protein of well-watered winter wheat 35% by Chaudhuri et aL (1986), but no effect of CO, enrichment on
grain protein was observed by Havelka et aL (1984b).
7-24
-------
Rose
CHAPTERS
COTTON
Cotton (Gossvpium hirsutum L.) accounts for 90% of the world's production of fiber and is the second
most important oilseed crop as well (Coulman, 1987; Roecklein and Leung, 1987; Wilcox and Leffel, 1987). ThK
small perennial shrub has been cultivated in both the Old and New World for at least 45 centimes (Lee, 1987;
Roecklein and Leung, 1987). The largest producers are the UJS A., where it grows all across the southern states
from North Carolina to California; the Soviet Union; Turkey; and the People's Republic of China But its high-
temperature adaptation also makes it an important product in many Third World countries such as Egypt, El
Salvador, India, and Mexico (Coulman, 1987; Roecklein and Leung, 1987).
YIELD
Biomass accumulation. Cotton plants exposed to 500 and 650 /iL CO2/L air in open-top field chambers
increased their total dry weight 40 and 51% more than plants at ambient CO2 (Kimball et aL, 1983,1984,1985;
Idso et aL, 1987b); plants grown in a phytotron at 650 and 1000 ftL/L had dry weight 72 and 115% higher than
control plants, respectively (DeLucia et aL, 1985). DeLucia et al. (1985) found dry weight increase was highest
in leaves, as treatment with high levels of CO2 caused leaf area to rise (Kimball et al., 1984).
Economic yield. Cotton yield can be expressed as seed cotton yield (lint and seeds) or lint yield. The
field-grown plants cited above averaged seed cotton yield increases of 70% when exposed to two times the
ambient CO2 concentration over four years of experiments by Kimball et al. (1986, 1985, 1984, 1983). When
plants were nitrogen fertilized, the seed cotton yield increase due to CO, enrichment was 48%; plants with no
added nitrogen showed a 70% increase due to increasing CO, (Kimbalfet al., 1986). In these chambers, lint
yield of plants exposed to 500 and 650 /iL/L increased 31 and 46% (Kimball et al. 1983).
These are among the largest yield responses to high concentrations of carbon dioxide of any species
so far investigated. Some reasons for this may be the following: (1) Cotton has an indeterminant life cycle and
can keep producing flowers and fruits as long as the weather is favorable. (2) Cotton has a much higher
saturation point for CO, than most other agricultural crops (Breen et al., 1986), and therefore can make use
of all the extra carbon oioxide it is given. (3) At least in some cases, cotton does not dose its stomata in
response to carbon dioxide enrichment, enabling it to keep photosynthesizing without a stomatal inhibition
(Radin et al., 1987). (4) Cotton is extremely heat-tolerant. Growing at high temperatures may help it avoid
feedback inhibition, and its emzymes may not suffer the limitations other crops undergo at high temperatures
(Idso et al., 1987a, 1987b). However, Perry et al. (1983) found that at 40°C, cotton photorespiratory losses
equaled 50% of the carbon dioxide fixed in net photosynthesis.
Yield components. Exposure to enriched atmospheres of carbon dioxide appears to affect seed cotton
and lint yield in part by increasing the number of flowers (Kimball et al., 1986,1985,1984,1983), without much
effect on boll retention (Kimball et aL, 1985,1984,1983), seed index, harvest index (Kimball et al., 1985,1984),
percent lint per boll or seed size (Kimball et al., 1984).
Phenology. However, high CO2 concentration affects the life cycle of flowering, boll-loading, and
vegetative growth which cotton plants pass through, with less tune between phases of the life cycle (Kimball et
al., 1986). This renders the response to elevated CO, a function of the phase of the fruiting cycle plants are in
when the season ends and therefore of the length ofthe season.
PHOTOSYNTHESIS
With doubled carbon dioxide concentration, leaf net photosynthesis at midseason was elevated 51 to
77% (Kimball et al., 1986,1985,1984, Radin et al., 1987). This substantial rise in net photosynthesis was still
7-25
-------
Rose
not large enough to account for cotton yield increases seen with elevated carbon dioxide concentration, indicating
an effect of CO2 on phenology, yield components (such as the number of flowers, mentioned above), or some
other influence on growth, such as the leaf area increase seen by Idso et al. (1987b).
Phenology. At ambient levels of carbon dioxide, field-grown cotton plants did not change photosynthetic
rates over the season. But when photosynthesis on these plants was measured at very high CO2 levels, plants
showed a hyperbolic curve in photosynthetic capacity over the season. Shape of the response curve to carbon
dioxide concentration was dependent upon stage of growth (Breen et al., 1986); high concentrations of COL
increased net photosynthesis by 51% at midseason, but only 37% late in the season (Kimball et al., 19861
Photosynthesis at midseason was not saturated until the carbon dioxide concentration reached 3000 cm3/m3
(Breen et al., 1986; Hesketh et al., 1984).
Photosvnthetic capacity. Studies by various investigators have yielded different results concerning
whether or not cotton photosynthetic capacity, also called photosynthetic efficiency (assimilation rate per internal
carbon dioxide concentration, Cj) declines with prolonged exposure to high levels of CO,- Radin et al. (1987) and
Kimball et al. (1985) working in Arizona found photosynthetic efficiency for well-watered field-grown cotton
plants was linear over a wide range of c's, to at least a concentration of 700 pL/L. This phenomenon occurred
in plants grown at 350 or 650 pL/L COy i.e., it was unchanged by elevated carbon dioxide concentration.
Radin et al. (1987) also estimated that cwas at least 85% of atmospheric carbon dioxide concentration
in both ambient and enriched CO2 conditions. The Arizona group found that plants grown in high amounts of
carbon dioxide in their outdoor open-top chambers experienced no decline in photosynthetic capacity, ie., no
inhibition due to acclimation to high partial pressures of CO~ until very late in the season (September) when
temperatures dropped somewhat (Kimball et al., 1985; Radin et al., 1987).
However, the group working at the Duke phytotron (DeLucia et al., 1985; Sasek et al., 1985) found
that cotton plants grown in high carbon dioxide concentrations (675 or 1000 /iL/L) did have a diminished
photosynthetic capacity when measured at 675 and 1000 pL/L CO,, and even at an ambient concentration, 350
pL/L.
FEEDBACK INHIBITION
This drop in photosynthetic capacity after growth in high levels of CO, was correlated with a high
internal (nonstomatal) resistance to carbon assimilation and declines in chlorophyll content and changes in the
chlorophyll a/chlorophyll b ratio (DeLucia et aL, 1985).
Carbohydrate accumulation. Kimball et al. (1985) found that concentrations of leaf-soluble
carbohydrates (glucose, fructose, and sucrose) were unaltered by exposure to 650 /iL COj/L. But they observed
that cotton plants grown in that level of CO2 had starch concentrations two to five times the concentration in
leaves grown in ambient carbon dioxide concentration (Radin et aL, 1987; Kimball et al., 1986, 1985). Early
morning minima of starch concentrations were most strongly affected by CO, enrichment, causing Kimball et
al. (1985) to surmise that starch was not being completely broken down at night. DeLucia et aL (1985) also
observed very high concentrations of carbohydrates, especially starch.
DeLucia et aL (1985) suggested that high carbohydrate accumulation may have limited photosynthesis
by feedback inhibition or even physical damage to chloroplasts, which were crowded with starch grains. Sasek
et aL (1985) reported that both photosynthetic inhibition and starch buildup in plants grown at 675 or 1000 /iL/L
CO2 were reversed upon placing them in an ambient level of CO2 (350 pL/L). In order to rule out effects of
stomatal changes, they calculated photosynthetic efficiency (moles CO, fixed m'2 s~' mole*1 c) and found that
it, like net photosynthesis, recovered 3 to 5 days after being placed in the lower carbon dioxide concentration.
Experimental conditions. This apparent discrepancy in results concerning feedback inhibition of
photosynthesis may be due to several differences in methodology between the two groups. Since the Duke
7-26
-------
Rose
group (DeLucia et al., 1985, and Sasek et aL, 1985) did their work in a phytotron, it might be suspected that
the light intensity would be lower than that obtained by the Phoenix researchers (Kimball et aL, 1983, 1984,
1985,1986; Idso et al., 1987a,b; Radin et al., 1987) growing plants outside in sunny Arizona. However, the light
level measured by Sasek et al. (1985) is almost identical to that measured by Radin et aL (1987). Alternatively,
the phytotron plants were grown in pots, which method is known to affect root growth and may have affected
(inhibited) translocation of carbohydrates.
TEMPERATURE INTERACTIONS
The cause of the different results concerning feedback inhibition may be the differences hi growth
temperatures used by the two research groups. DeLucia et aL (1985) and Sasek et al. (1985) used day/night
temperatures of 26/20°C (in a thermo- and photoperiod of 12h/12h, a short day for a summer crop). The
Arizona group's average maximum daytime temperatures were 455°C in July, 43.8*0 in August, and 3&5°C in
September, the latter being the one time feedback inhibition was observed in Phoenix. Radin et aL (1987)
ascribe this inhibition to the drop in temperature.
The Phoenix group has a theory based on observations of five different species, including cotton at
various natural, uncontrolled (but generally very hot) temperatures, that plant response to elevated atmospheres
of CO, is dependent upon temperature. At temperatures < 12°C, plants experienced a negative yield response;
at 19° C, plants yielded about the same in high carbon dioxide concentration as at ambient (but cotton was not
grown at either of those lower temperatures). At 34°C, plants averaged 130% greater yields when amounts of
carbon dioxide were doubled, and cotton seemed to respond positively to even higher temperatures (Kimball et
al., 1985, 1986; Idso et al., 1987a,b; Radin et aL, 1987).
Possibly cotton had such a good yield response to CO2 enrichment because higher rates of enzyme
activity at high temperatures speeded up metabolism, including respiration and translocation, enough to eliminate
any dysfunctional carbohydrates clogging up the system and inhibiting photosynthesis. But at the much lower
temperatures used by the Duke researchers, photosynthesis could have created many more new carbohydrates
than the other plant processes could use. Sasek et al. (1985) suggested that their reversal experiments indicate
if strong sinks (such as more flowers, fruits, stems, or roots) appeared, starch could decrease and photosynthesis
recover during periods of high photosynthate demand, even under continuous high concentrations of CO,.
However, in the Arizona field experiments, cotton temperatures were not controlled, and the plants
were aging as the season progressed and the temperatures dropped. Considering the changes in photosynthetic
response to high partial pressures of CO, according to stage of development mentioned above (Radin et aL,
1987; Breen et al., 1986; Kimball et al., 1986,1985) one cannot be sure that this decline in photosynthesis with
time is due to temperature, senescence, or yet another factor that is also changing with the season. Experiments
done at externally controlled temperatures intermediate between the two groups (e.g. 35/25°C) and with
thermoperiod and photoperiod typical of southern temperate summers would help sort out these influences.
STOMATAL RESPONSES
There is contradiction in the published experimental results on the effect of enhanced carbon dioxide
concentrations on cotton stomata. One report (Kimball et al., 1985) even found opposite results with the same
measuring device: stomatal resistance dropped 47% in well-watered and 69% in draughted plants exposed to high
amounts of CO,, but that stomatal conductance is unaffected (not reduced) by the same conditions. Since the
two measures, stomatal resistance and stomatal conductance, are reciprocals of each other, if one rises the other
should fall. Other groups found that added CO, treatments increased stomatal resistance (Kimball et aL, 1983);
or did not affect stomatal resistance when plants were well watered (Kimball et al., 1984, DeLucia et al., 1985),
but increased stomatal resistance if plants were drought stressed (Kimball et al., 1984). Carbon dioxide
enrichment has also been reported as not affecting the high stomatal conductance seen at midseason (Radin et
al., 1987); or decreasing stomatal conductance slightly (11%) at midseason but more (25%) later in the season
7-27
-------
Rose
(Kimball et al., 1986). In this latter case, a sharp drop in stomatal conductance late in the season was larger than
any treatment differences. The decline was attributed to falling temperatures, which would cause less water to
evaporate from the stomata; or to lower relative humidity, which would increase evaporative demand and might
therefore induce stomatal closure.
The recent work of Idso (1987) and Idso and Allen (1988) might explain these discrepancies as artifacts
of porometry. Or perhaps the answer is found in Sasek et al. (1985), who observed that stomatal function seems
to adjust to high concentration of carbon dioxide over time. Plants in ambient CO2 started out with higher
conductance, which gradually fell (stomatal resistance increased) over time; plants in 1000 /*L/L CO2 started
out with lower conductance, which gradually rose (stomatal resistance decreased) over time. Therefore, the level
of conductance in the'two treatments approached each other as the season progressed, one of many instances
of plants reacting to environmental change by homeostatically tending towards core metabolic levels. If plants
were switched from very high levels of CO, to ambient levels, conductance rose more quickly to previous levels
(Sasek et al., 1985).
Another explanation is that stomatal apertures are controlled by a number of plant factors, such as
leaf temperature, presence of absrisic acid in the extracellular fluid, water status of the leaf, and concentration
of solutes; and environmental factors, including humidity, air temperature, carbon dioxide concentration, light,
and available water in the soil. Obviously carbon dioxide concentration is just one of many influences on
stomata, and in many instances, it may not be the most important or controlling factor.
The Arizona research group, who generally did not see declines in stomatal conductance in plants
exposed to high carbon dioxide concentrations (Radin et aL, 1987, Kimball et al., 1986, 1985), felt that this
maintenance of open stomata was perhaps the reason why cotton showed a greater positive yield response to
increased CO^ with a correlation between stomatal conductance and photosynthetic capacity. They showed
that plants grown in high amounts of CO, (Radin et al., 1987, Kimball et al., 1986) were also more reponsive
to injections of ABA, which causes stomatal closure. Finally, they observed a 10% increase in the number of
stomata/leaf area with high CO2 levels.
TRANSPIRATION, WATER USE, AND WATER-USE EFFICIENCY
Since plants grown in high partial pressures of carbon dioxide had greater leaf area, in some cases their
transpiration and water use per unit land area were higher, which caused the soil to be more depleted (Kimball
et al., 1985). Higher leaf temperature of those plants also raised transpiration/ground area. This was
counteracted by the increasing stomatal resistance (decreasing conductance), seen in at least some cases, which
decreased transpiration per leaf area. In well-watered cotton plants exposed to double the current concentration
of COy transpiration declined 73% (Kimball et aL, 1985; Idso et al., 1987a,b). Relative water content of the
leaves decreased with increasing levels of CO2 five or six days after irrigation (Kimball et al., 1985).
Results concerning the effect of high partial pressures of CO2 on water use are inconclusive, whether
measured with lysimeter (a device to weigh both soil and plants) or neutron probe (a device that bounces
neutrons off soil water molecules) (Kimball et aL, 1984,1983). However, the large yield increases with a small
change in water use translated into large increases in water-use efficiency.
LEAF TEMPERATURE
Despite the differences in experimental results with stomata, there is agreement that increases in leaf
temperature (which are generally secondary effects of stomatal closure) occur when high concentrations of CO,
are administered to cotton plants (and all other species studied). Kimball et aL (1984) found that well-watered
plants exposed to high levels of CO2 under humid conditions (vapor pressure deficit « 1 kPa) had leaf
temperatures within 1°C of leaf temperatures of ambient-grown plants. As air became drier (vapor pressure
deficit » 4 kPa), a doubling of carbon dioxide concentration caused leaf temperatures to increase by about 2°C.
7-28
-------
Rose
They point out that leaf temperature data were somewhat in conflict with stomatal resistance data," but since
there were many more temperature data, the researchers concluded the latter data were probably more reliable.
Leaf temperature increases of about 1° with carbon dioxide concentrations in the 600-650 /iL/L range
were also found in other years (Kimball et aL, 1983,1985,1986; Idso et al., 1987a,b). Idso et al. (1987a,b) felt
this leaf temperature rise, which caused plant transpiration to fall, was the key to cotton yield increases and
developed a formula for prediction of the latter from the former.
DROUGHT STRESS
Drought stress and yield. CO2-enrichment effects on water-limited cotton are dramatic. Cotton plants
exposed to high levels of COgachieved yield increases that overwhelmed any yield reductions due to drought
stress (Kimball et al., 1986). Raising the concentration of CO, from ambient to 650 ppm increased the seed
cotton yield of water-limited plants from 476 to 717 g/m2 (51%; when no nitrogen was added, and from 449 to
763 (70%) when nitrogen fertilized. This lifted the yields even above those of the well-watered plants growing
in 330 ppm CO^ which were 557 (N-) and 658 (N+ ) (Kimball et al., 1986). Kimball et al. (1985) found that seed
yield was increased even more, to 104% of ambient levels, by placing plants in enriched atmospheres of COy,
and that total dry weight rose 71%.
However, the water-deficit stress in these experiments was applied by adding two-thirds of the normal
weekly amount of irrigation water. The results achieved in the very dry Arizona climate, with its high vapor
pressure deficit, may not be representative of plants in more humid areas that may not experience such severe
stress even though the percentage drop in amount of water from well-watered to drought ed maybe much greater.
Alternatively, plants that live through much more extended drying cycles may experience more severe stress than
these mild and constant stresses.
Drought stress and metabolism. Carbon assimilation rates of drought-stressed plants grown under
CO, concentration of 650 /iL/L exceeded that of plants in current atmospheres of CO2 by 39% (Kimball et
al., 1984) or 52% (Radin et al., 1987). But photosynthesis in plants deprived of adequate water showed a
saturation response to c,, unlike well-watered plants that kept increasing assimilation with c, up to 700 /iL/L
(Radin et al., 1987). Again, even the substantial rise in photosynthesis does not account for yield increases seen
with high levels of CO2- Kimball et aL (1986) found no effect of drought stress alone on photosynthesis, though
well-watered plants showed a greater positive effect (70% increase) of CO, enrichment on photosynthesis (Radin
et al., 1987).
While Kimball et al. (1986) found that drought stress caused leaf temperatures to rise, they observed
no concomitant, significant drop in stomatal conductance with drought stress. As noted above, according to
new calculations of Idso and Allen (1988) this could be an artifact of steady-state porometer methodology.
However, Radin et aL (1987) saw increased stomatal closure in drought-stressed plants responding to elevated
amounts of CO2.
7-29
-------
Rose
CHAPTER 6
CORN
Corn or maize (Zea mays L.) is a very high-yielding C4 grass of Central American origin that has been
bred by humans for grain production for 25,000 years. Maize is the third largest crop in the world, with
production highest in North America and acreage highest in Asia (Coalman, 1987; Roecklein and Leung, 1987;
Starling, 1987). It is the number one U.S. crop, whether counted by acreage with 20.9 million ha in 1983-84, or
by production with 106.8 million metric tons in the same years (Coulman, 1987; Roecklein and Leung, 1987;
Starling, 1987). In this country corn is used extensively for animal feed and for processed foods, often as com
flakes, high-fructose corn syrup, com oil, and com starch. In other countries, maize is still a primary part of the
human diet, as flour for tortillas and bread or as a vegetable. It has many industrial uses, including alcohol,
adhesives, paper, linoleum, paint, soap, and textile sizing (Roecklein and Leung, 1987; Starling, 1987).
PHOTOSYNTHESIS
Despite its economic importance to American agriculture, very little research has been done recently
concerning the effects of elevated carbon dioxide concentration on this crop. The lack of study on corn probably
stems in part from the fact that plants like com which use the C4 pathway of photosynthesis already trap ambient
CO, quite efficiently. Therefore, com showed no photosynthetic increase (in CO2 fixation) with increases in
carbon dioxide concentration (Rogers et al. 1983a,b; Surano and Shinn, 1984). However, Usuda (1987) recently
found that increasing internal CO, level of maize increased the concentrations of intermediates of the pentose
phosphate pathway, which in turn increased the level of total C4 cycle intermediates, indicating a possible effect
on photosynthetic carbon metabolism.
YIELD
Although few effects on photosynthesis have been noted, biomass of these plants does increase with
moderate increases in carbon dioxide concentration. At 600 ppm CO,, total plant weight increased 24% (Carlson
and Bazzaz, 1980) or 7% across several water treatments (King ana Greer, 1986); at 850 ppm King and Greer
(1986) found no further increase, but at 910 ppm Rogers et aL (1983b) observed a 60% rise in dry matter over
ambient concentration of carbon dioxide. At 1200 ppm COj, Carlson and Bazzaz (1980) found a 7% decline in
biomass production.
Rogers' research group (1983a,b) also found enhanced growth of plant parts, including leaf area and
seed yield They suggested that the increases in leaf area and biomass occurred because the plants grown hi
high partial pressures of CO2 did not wilt on hot summer afternoons. The wilting observed at ambient levels
of CO2 inhibits leaf expansion. Surano and Shinn (1984) found that carbon dioxide enrichment increased root
growth in well-watered plants and stem height in drought-stressed plants, but had no effect on yield in either
case. In contrast to effects seen in species using the C3 pathway of photosynthesis, high carbon dioxide
concentration had no effect on leaf thickness, node number, number of reproductive parts (Rogers et aL 1983b),
or partitioning of dry matter (King and Greer, 1986).
PHENOLOGY
Leaf emergence occurred more rapidly in com plants subjected to 850 pL/L carbon dioxide than in
plants grown in 350 /*L/L. The former also consistently had significantly (P * 0.001) more fully senesced,
nongreen leaves than the latter, indicating an acceleration of phenology hi the plants exposed to high partial
pressures of carbon dioxide (King and Greer, 1986). But Surano and Shinn (1984) found no effect on phenology
of com in CO2 enrichment
7-30
-------
Rose
STOMATAL RESPONSE, TRANSPIRATION, AND WATER-USE EFFICIENCY
Stomatal conductance in corn decreased from 1.8 to 0.8 cm/sec as carbon dioxide concentration rose
from 340 ppm to 910 ppm (Rogers et al. 1983a,b). Morison and Gifford (1983) found the stomatal responses
of corn and of Paspalum plicatun). another C4 grass, were similar to that of rice and another C3 grass. All four
species showed greater stomatal sensitivity when stomatal conductance was largest, that is when carbon dioxide
concentration was low and VPD was high.
Transpiration was reduced whether maize plants were grown for 35 days in a glasshouse (Carlson and
Bazzaz 1980), in controlled environments, (Morison and Gifford, 1983; King and Greer, 1986), or for 3 months
in open-top field chambers (Rogers et al. 1983a,b). Water-use efficiency improved in all cases, up to 55% in the
controlled environment (King and Greer, 1986).
LEAF TEMPERATURE
Corn leaf temperature, like that of all other species measured, rose in response to carbon dioxide
enrichment (Morison and Gifford, 1983,1984). This is not surprising, since the drop in stomatal conductance
in corn exposed to high CO2 concentrations was the largest of any crop in the literature. Corn is very tolerant
of high temperatures; like all C4 plants it is not subject to high temperature-induced yield losses due to
photorespiration. But combined with rising temperatures predicted by climate change, corn leaf temperatures
may increase enough to actually inhibit some phases of plant metabolism (Hesketh and Hofstra, 1975).
DROUGHT STRESS
Surano and Shinn (1984) observed no effect of high concentrations of carbon dioxide on yield, biomass,
leaf area, or number and weight of ears in com plants that were deprived of adequate water, which was similar
to the findings of Gifford and Morison (1985) on Paspalum plicatum. another C4 grass. In contrast, King and
Greer (1986) observed that draughted corn had a greater percentage response to CO2 enrichment of total plant
dry weight and of grain yield than did control plants (as long as there was enough water to produce grain at all).
However, additional amounts of carbon dioxide did not make up for the very large yield reductions caused by
drought (King and Greer, 1986).
7-31
-------
Rose
REFERENCES
Ackerson, R.C., Havelka, U.D. and Boyle, M.G., 1984. CO,-enrichment effects on soybean physiology. D.
Effects of stage-specific CO2 exposure. Crop ScL, 24:1150-1154.
Acock, B., and Allen, L.H., Jr., 1985. Crop responses to elevated CO, concentrations. In: B.R. Strain and J.D.
Cure, (editors), Direct Effects of Increasing CO, on Vegetation. VS. Dept. of Energy DOE/ER-0238, N.T J.S.,
Springfield, VA, pp 55-97.
Allen, L.H., Jr., Boote, KJ., Jones, J.W., Jones, P.H., Valle, R.R., Acock, B., Rogers, H.H. and Dahlman, R.C.,
1987. Response of vegetation to rising carbon dioxide: Photosynthesis, biomass, and seed yield. Global
Biogeochemical Cycles, 1:1-14.
Allen, L.H., Jr., Vu, J.C.V., Valle, R.R., Boote, KJ. and Jones, PH., 1988. Nonstructural carbohydrates and
nitrogen of soybeans grown under CO2 enrichment. Crop ScL, 28:84-94.
Azcon-Bieto, J., 1983. Inhibition of photosynthesis by carbohydrates in wheat leaves. Plant PhysioL, 73:681-
686.
Begonia, G.B., Hesketh, JJX, Wooley, J.T., and Peters, D.B., 1987. Variability in leaf photosynthetic CO2
exchange rates near saturating irradiance and CO2- Photosynthetica, 21:584-587.
Breen, PJ., Hesketh, JJ>., and Peters, D.B., 1986. Field measurements of leaf photosynthesis of C3 and C4
species under high irradiance and enriched CO2. Photosynthetica, 20:281-285.
Campbell, W J., Allen, L.H., Jr., and Bowes, G., 1987. Effects of short-term and long-term exposures to varying
CO, concetrations on soybean photosynthesis. In: Biggins, J., (editor) Progress in Photosynthesis Research, Vol.
IV. Martinus Nijhoff Publishers, Dordrecht, the Netherlands.
Carlson, R.W., and Bazzaz, FA., 1980. Effects of elevated CO2 concentrations on growth, photosynthesis,
transpiration, and water use efficiency of plants. In: JJ. Singh and A. Deepak (editors), Environmental and
climatic impact of coal utilization. Academic Press, New York.
Chaudhuri, U.N., Burnett, R.B. Kanemasu, E.T. and Kirkham, M.B., 1986. Effect of elevated levels of CO,
on winter wheat under two moisture regimes. 029 in series: Response of Vegetation to Carbon Dioxide. USi
Depts. of Energy and Agriculture.
Chaudhuri, U.N., Burnett, R.B. Kanemasu, E.T. and Kirkham, M.B., 1987. Effect of elevated levels of CO,
on winter wheat under two moisture regimes. 040 in series: Response of Vegetation to Carbon Dioxide. UJ&
Depts. of Energy and Agriculture.
Clough, J.M., Peet, M.M. and Kramer, PJ., 1981. Effects of high atmospheric CO2 and sink size on rates of
photosynthesis of a soybean cultivar. Plant PhysioL, 67:1007-1010
Coulman, B.E., 1987. World crop production and distribution. In B.R. Christie, (ed.) CRC Handbook of Plant
Sciences in Agriculture, VoL n, CRC Press, Boca Raton, FT, pp. 77-87.
Coyne, P J. and Bradford, J A., 1984. Leaf gas exchange in 'Caucasian' bluestem in relation to light, temperature,
humidity and CO2- Agron. J., 76:107-113.
Cure, J.D., Galloway, L.R, Israel, D.W. and Rufty, T.W., 1986. Influence of nutrition on vegetation response to
carbon dioxide. I. Interactions of nitrogen and phosphorus supply on soybean growth and nutritional parameters.
033 in series: Response of Vegetation to Carbon Dioxide. ILS. Depts. of Energy and Agriculture.
7-32
-------
Rose
Cure, J.D., Rufty, T.W. and Israel, D.W., 1987. Assimilate utilization in the leaf canopy and whole plant growth
of soybeans during acclimation to elevated CO2. Bot. Gat, 148:67-72.
DeLucia, E.H., Sasek, T.W., and Strain, B.R., 1985. Photosynthetic inhibition after long-term exposure to
elevated levels of atmospheric carbon dioxide. Photosyn. Res., 7:175-184.
Espie, G.G. and Coulman, B., 1987. The effect of pH, Oy and temperature on the CO2 compensation point of
isolated Asparagus mesophyll cells. Plant PhysioL, 83:113-117.
Finn, G A. and Brun, W A., 1982. Effect of atmospheric CO2 enrichment on growth, nonstructural carbohydrate
content, and root nodule activity in soybean. Plant PhysioL, 69:327-331.
Gent, M.P.N., 1984. Carbohydrate level and the growth of tomato plants. Plant PhysioL, 76:694-699.
Gifford, R.M., 1977. Growth pattern, carbon dioxide exchange, and dry weight distribution in wheat growing
under different photosynthetic environments. AustraL J. Plant PhysioL, 4:99-110.
Gifford, R.M., 1979. Growth and yield of CCyenriched wheat under water-limited conditions. Austral. J. Plant
PhysioL, 6:367-378.
Gifford, R.M. and Morison, J.I.L., 1985. Photosynthesis, water use and growth of a C4 grass stand at high CO2
concentration. Photosynthesis Res., 7:77-90.
Gifford, R.M., Lambers, H. and Morison, J.I.L., 1985. Respiration of crop species under CO, enrichment.
PhysioL Plant., 63:351-356.
Havelka, U.D., Ackerson, R.C., Boyle, M.G. and Wittenbach, VA., 1984a. CCyenrichment effects on soybean
physiology. I. Effects of long-term CO2 exposure. Crop Sci., 24:1146-1150.
Havelka, U.D., Wittenbach, VA. and Boyle, M.G., 19845. CCyenrichment effects on wheat yield and
physiology. Crop Sci., 24:1163-1168.
Hesketh, J.D., Woolley, J.T., and Peters, D.B., 1984. Leaf photosynthetic CO2 exchange rates in light and CO2
enriched environments. Photosynthetica, 18: 536-540.
Hokum, JA.H., O'Leary, M.H. and Osmond, C.B., 1983. Effect of varying CO2 partial pressure on
photosynthesis and on carbon isotope composition of carbon-4 of malate from the Crassulacean Acid Metabolism
plant Kalanchoe daigremontiana Hamet et Perr., Plant PhysioL, 71:602-609.
Huber, S.C., 1983. Role of sucrose-phosphate synthesis in partitioning of carbon in leaves. Plant PhysioL, 71:818-
821.
Huber, S.C., Rogers, H.H., and Mowry, F.L., 1984. Effects of water stress on photosynthesis and carbon
partitioning in soybean (Gfycine max [L.J Merr.) plants grown in the field at different CO2 levels. Plant PhysioL,
76:244-249.
Idso, S.B., 1987. An apparent discrepancy between porometry and infrared thermometry relative to the
dependence of plant stomatal conductance on air vapor pressure deficit. Agric. For. MeteoroL, 40:105-106.
Idso, S.B. and Allen, S.G., 1988. Problems with porometry: measuring stomatal conductances of potentially
transpiring plants. Agric. For. MeteoroL, 43:49-58.
7-33
-------
Rose
Idso, S.B., Kimball, BA. and Mauney, J.R., 1987a. Atmospheric carbon dioxide enrichment effects on cotton
midday foliage temperature: Implication for plant water use and crop yield Agroa J., 79:667-672.
Idso, S.B., Kimball, BA., Anderson, M.G. and Mauney, J.R., 1987b. Effects of atmospheric CO2 enrichment
on plant growth: the interactive role of air temperature. Agric. Ecosystems and Envir., 20:1-10.
Jones, PH., Allen, L.H., Jr., Jones, J.W., Boote, KJ., and Campbell, WJ., 1984. Soybean canopy growth,
photosynthesis, and transpiration responses to whole-season carbon dioxide enrichment. Agroa J., 76:633-637.
Jones, P.H., Allen, L.H., Jr., and Jones, J.W., 1985a. Responses of soybean canopy photosynthesis and
transpiration to whole-day temperature changes in different CO2 environments. Agron. J., 77:242-249.
Jones, P H., Allen, L.H., Jr., Jones, J.W., and Valle, R.R., 198Sb. Photosynthesis and transpiration responses of
soybean canopies to short- and long-term CO2 treatments. Agron. J., 77:119-126.
Jones, P.H., Jones, J.W., and Allen, L.H., Jr., 1985c. Carbon dioxide effects on photosynthesis and transpiration
during vegetative growth in soybeans. Soil and Crop Sci. Soc. of Florida Proc., 44:129-134.
Jones, P.H., Jones, J.W., and Allen, L.H., Jr., 1985d. Seasonal carbon and water balances of soybeans grown
under stress treatments in sunlit chambers. Trans. Amer. Soc. Agric. Engineers, 28:2021-2028.
Jurik, T.W., Weber, JA., and Gates, D.M., 1984. Short-term effects of CO, on gas exchange of leaves of
bigtooth aspen (Populus grandidentata) in the field. Plant PhysioL, 75:1022-1026.
Kimball, B A., 1983. Carbon dioxide and agricultural yield: an assemblage and analysis of 430 prior observations.
Agroa J., 75:779-788.
Kimball, BA., Mauney, J.R., Guinn, G., Nakayama, F.S., Pinter, PJ., Jr., Clawson, K.L., Reginato, R J., Idso,
S.B., 1983. Effects of increasing atmospheric CO, on the yield and water use of crops. 021 in series: Response
of Vegetation to Carbon Dioxide. U.S. Depts. of Energy and Agriculture.
Kimball, BA,, Mauney, J.R., Guinn, G., Nakayama, F.S., Pinter, PJ., Jr., Clawson, K.L., Idso, S.B., Butler, G.D.,
Radin, J.W., 1984. Effects of increasing atmospheric CO2 on the yield and water use of crops. 023 in series:
Response of Vegetation to Carbon Dioxide. U.S. Depts. of Energy and Agriculture.
Kimball, BA., Mauney, J.R., Guinn, G., Nakayama, FJS., Idso, S.B., Radin, J.W., Hendrix, D.L., Butler, G.D.,
Zarembinsld, T.L, Nixon, P.E. n., 1985. Effects of increasing atmospheric CO, on the yield and water use of
crops. 027 in series: Response of Vegetation to Carbon Dioxide. U.S. Depts. of Energy and Agriculture.
Kimball, BA, Mauney, Radin, J.W., Nakayama, F.S., Idso, S.B^ Hendrix, D.L., Akey, D.H., Allen, S.G.,
Anderson, M.G., Hartung, W., 1986. Effects of increasing atmospheric COupn the yield and water use of crops.
039 in series: Response of Vegetation to Carbon Dioxide. UJS. Depts. ofJEnergy and Agriculture.
King, KM. and Greer, D.H., 1986. Effects of carbon dioxide enrichment and soil water on maize. Agroa J.,
78:515-521.
Kobza, J. and Edwards, GIL, 1987. Influence of leaf temperature on photosynthetic carbon metabolism in wheat.
Plant Physiol., 83:69-74.
Korner, G, Farquhar, G.D., and Roksandic, Z., 1988. A global survey of carbon isotope discrimination in plants
from high altitude. Oecologia, 74:6223-632,
Lee, C-S. and RawUngs, J.O., 1982. Design of experiments in growth chambers-uniformity trials in the North
Carolina State University Phytotroa Crop Sd, 22:551-558.
7-34
-------
Rose
Lee, J A., 1987. Plant fibers. In B.R. Christie, (ed.) CRC Handbook of Plant Sciences in Agriculture, VoL n,
CRC Press, Boca Raton, FT, pp. 173-182.
Lincoln, D.E., Couvet, D., and Sionit, N., 1986. Response of an insect herbivore to host plants grown in carbon
dioxide enriched atmospheres. Oecologia, 69:556-560.
Machler, F., Lehnherr, B., and Nosberger, J., 1986. Is transfer of CO, to C3 plants purely by diffusion? Plant
Physiol., 82:1167-1168.
Marc, J. and Gifford, R.M., 1984. Floral initiation in wheat, sunflower and sorghum under carbon dioxide
enrichment. Can. J. Bot., 62:9-14.
Martin, C.E. and Siedow, J.N., 1981. Crassulacean acid metabolism in the epiphyte TUlandsia usneoides L.
(Spanish moss). Plant Physiol. 68:335-339.
Monson, R.K., Stidham, MA., Williams, GJ. ED, Edwards, G.E. and Uribe, E.G., 1982. Temperature
dependence of photosynthesis in Agropyron smithii Rydb. I. Factors affecting net CO, uptake in intact leaves
and contributions from ribulose-l,5-bisphosphate carboxylase measured in vivo and in vitro. Plant Physiol.,
69:921-928.
Morison, J.I.L., 1985. Sensitivity of stomata and water use efficiency to high CO,. Plant, Cell and Envir., 8:467-
474.
Morison, J.I.L. and Gifford, R.M., 1983. Stomatal sensitivity to carbon dioxide and humidity. A comparison of
two C3 and two C4 grass species. Plant Physiol., 71:789-7%.
Morison, J.I.L. and Gifford, R.M., 1984a. Plant growth and water use with limited water supply in high CO2
concentrations. Austral. J. Plant Physiol., 11:361-374.
Mott, KA,, 1988. Do stomata respond to CO, concentrations other than intercellular? Plant Physiol., 86:0200-
0203.
Peet, M.M., 1984. CO2 enrichment of soybeans. Effect of leaf/ pod ratio. Physiol. Plant., 60:38-42.
Perry, S.W., Krieg, D.R. and Hutmacher, R.B., 1983. Photosynthetic rate control in cotton. Photorespiration.
Plant Physiol., 73:662-665.
Peterson, R.B., 1983. Estimation of photorespiration based on the initial rate of postillumination CO2 release.
Plant Physiol., 73:983-988.
Porter, MA. and Grodzinski, B., 1984. Acclimation to high CO2 in bean. Carbonic anhydrase and ribulose
bisphosphate carboxylase. Plant Physiol., 74:413-416.
Potvin, C., Goeschl, J.D. and Strain B.R., 1984. Efects of temperature and CO2 enrichment on carbon
translocation of plants of the c4 grass species Echinochloa crus-galli (L.) Beauv. from cool and warm
environments. Plant Physiol., 75:1054-1057.
Radin, J.W., Kimball, BA. Kimball, Hendrix, D.L., and Mauney, J.R., 1987. Photosynthesis of cotton plants
exposed to elevated levels of carbon dioxide in the field. Photosyn. Res., 12:191-203.
Reuveni, J., and Gale, J., 1985. The effect of high levels of carbon dioxide on dark respiration and growth of
plants. Plant Cell Envir., 8:623-628.
7-35
-------
Rose
Roecklein, J.C., and Leung, P.-S., 1987. A Profile of Economic Plants. Transaction Books, New Brunswick, New
Jersey 09803, pp. 36-37; 45; 131; 352-353.
Rogers, H.H., Bingham, G.E., Cure, JJD., Smith, J.M. , and Surano, R A., 1983a. Responses of selected plant
species to elevated carbon dioxide in the field. J. Envir. Qua!., 12:569-574.
Rogers, H.H., Thomas, J.F., and Bingham, G.E., 1983b. Response of agronomic and forest species to elevated
atmospheric carbon dioxide. Science, 220:428-429.
Rogers, H.H., Cure, J.D., Thomas, J.F., and Smith, JJVI., 1984a. Influence of elevated CO9 on growth of soybean
plants. Crop Sci, 24:361-366.
Rogers, H.H., Sionit, R, Cure, J.D., Smith, J.M., and Bingham, G.E., 1984b. Influence of elevated carbon
dioxide on water relations of soybeans. Plant Physiol., 74:233-238.
Rogers, H.H., Cure, J.D., and Smith, JJvI., 1986. Soybean growth and yield response to elevated carbon dioxide.
Agric. Ecosystems and Envir., 16:113-128.
Rufty, T.W. and Huber, S.C., 1983. Changes in starch formation and activities of sucrose-phosphate synthase
and cytoplasmic fructose -1,6-bisphosphatase in response to source-sink alterations. Plant Physiol., 72:474-480
Sage, R.F. and Sharkey, TJD., 1987. The effect of temperature on the occurrence of O2 and CO2 insensitive
photosynthesis in field-grown plants. Plant Physiol., 84:658-664
Sasek, T.W., DeLucia, E.H., and Strain, B JR., 1985. Reversibility of photosynthetic inhibition in cotton after long-
term exposure to elevated CO2 concentration. Plant PhysioL, 78:619-621
Schonfeld, M., 1987. Drought sensitivity in winter wheat populations under field conditions and elevated
atmospheric CO2. PhJD. Thesis, Oklahoma State University. 76pp.
Sionit, N., 1983. Response of soybean to two levels of mineral nutrition in CCy enriched atmosphere. Crop ScL,
23:329-333.
Sionit, N., Hellmers, H., and Strain, B JR.., 1980. Growth and yield of wheat under CO2 enrichment and water
stress. Crop ScL, 20:687-690.
Sionit, N., Mortenson, DA., Strain, B.R. and Hellmers, H., 1981a. Growth response of wheat to CO2 enrichment
and different levels of mineral nutrition. Agron. J., 73:1023-1027.
Sionit, N., Strain, B JR. Hellmers, H., and Kramer, PJ., 1981b. Effects of atmospheric CO2 concentration and
water stress on die water relations of wheat. Bot. Ga&, 142:191-196.
Sionit, N., Strain, B JR., and Beckford, HAu, 1981c. Environmental controls on the growth and yield of okra.
I. Effects of temperature and of CO2 enrichment at cool temperature. Crop ScL, 21:885-888.
Sionit, Nn Strain, B.R., and Hellmers, H., 1981d Effects of different concentrations of atmospheric CO2 on
growth and yield components of wheat. J. Agric. ScL, Camb., 79:335-339.
Sionit, N., Rogers, H JH., Bingham, G JB., and Strain, B JR, 1984. Photosynthesis and stomatal conductance with
CO, enrichment of container and filed-grown soybeans. Agron. J., 76:447-451.
Sionit, N., Strain, B.R., and Flint, E.P., 1987. Interaction of temperature and CO2 enrichment on soybean:
photosynthesis and seed yield Can. J. Plant ScL, 67:629-636.
7-36
-------
Rose
Starling, T.M., 1987. Cereals. In B.R. Christie, (ed.) CRC Handbook of Plant Sciences in Agriculture, Vol. D,
CRC Press, Boca Raton, Fl, pp. 117-124.
Stitt, M., Herzog, B. and Heldt, H.W., 1984a. Control of photosynthetic sucrose synthesis by fructose 2,6-
bisphosphate. I. Coordination of CO2 fixation and sucrose synthesis. Plant PhysioL, 75:548-553.
Stitt, M., KurzeL B. and Heldt, H.W., 1984b. Control of photosynthetic sucrose synthesis by fructose 2,6-
bisphosphate. H. Partitioning between sucrose and starch. Plant PhysioL, 75:554-560.
Strain, B.R., 1985. Background on the response of vegetation to atmospheric CO2 enrichment. In B.R. Strain
and J.D. Cure, (editors), Direct Effects of Increasing CO, on Vegetation. US. Dept. of Energy DOE/ER-0238,
N.T.I.S., Springfield, VA.
Strain, B.R., and Cure, J.D., (Editors), 1985. Direct Effects of Increasing CO, on Vegetation. UJS. Dept. of
Energy DOE/ER-0238, N.T.I.S., Springfield, VA.
Surano, KA. and Shinn, J.H., 1984. CO2 and water stress effects on yield, water-use efficiency and photosynthate
partitioning in field grown corn. Lawrenc Livermore National Laboratory, UCRL-90771, 30 pp.
Szarek, S.R., Holthe, PA., and Ting, I.P., 1987. Minor physiological response to elevated CO2 by the CAM plant
Agave vUmomiana. Plant PhysioL, 83:938-940.
Usuda, H., 1987. Changes in levels of intermediates of the C4 cycle and reductive pentose phosphate pathway
under various concentrations of CO2 in maize leaves. Plant PhysioL, 83:29-32.
Valle, R.R., Mishoe, J.W., Campbell, WJ., Jones, J.W., and Allen, L.H., Jr., 1985a. Photosynthetic responses
of 'Bragg* soybean leaves adapted to different CO2 environments. Crop Sci., 25:333-339.
Valle, R.R., Mishoe, J.W., Jones, J.W., and Allen, LJL, Jr., 1985b. Transpiration rate and water use efficiency
of soybean leaves adapted to different CO2 environments. Crop Sci., 25:477-482.
Vines, H.M., Tu, Z.-P., Armitage, A.M., Chen, S.-S., and Black, C.C., Jr., 1983. Environmental responses of the
post-lower illumination CO2 burst as related to leaf photorespiration. Plant PhysioL, 73:25-30.
Vu, J.C.V., Allen, L.H., Jr., and Bowes, G., 1983. Effects of light and elevated atmospheric CO2 on the ribulose
bisphosphate carboxylase activity and ribulose bisphosphate level of soybean leaves. Plant PhysioL, 73:729-734.
Vu, J.C.V., Allen, L.H., Jr., and Bowes, G., 1987. Drought stress and elevated CO^effects on soybean ribulose
bisphosphate carboxylase activity and canopy photosynthetic rates. Plant PhysioL, o3:573-578.
Wall, G.W. and Baker, D.N., 1987. Effects of temperature and elevated carbon dioxide concentrations on the
phenology of spring wheat. Agron. Abstr. 1987, p.18.
Wilcox, J.R., and Leffel, R.C., 1987. Oilseed crops. In B.R. Christie, (ed.) CRC Handbook of Plant Sciences
in Agriculture, Vol. H, CRC Press, Boca Raton, FT, pp. 145-166.
Wynn, T., 1981. Effects of O2 tension and temperature during light uptake of 14-CO2 on the dark release of 14-
CO2 by excised C3 and C4 leaves. Plant PhysioL, 68:1253-1256.
Wyse, R.E. and Saftner, R A^ 1982. Reduction in sink-mobilizing ability following periods of high carbon flux.
Plant PhysioL, 69:226-228.
7-37
-------
POTENTIAL EFFECTS OF CLIMATE CHANGE ON PLANT-PEST INTERACTIONS
by
Benjamin R. Stinner
Robin A. J. Taylor
Ronald B. Hammond
Foster F. Purrington
David A. McCartney
Department of Entomology
Ohio Agricultural Research and
Development Center
Wooster, OH 44691
and
Nick Rodenhouse
Gary W. Barrett
Department of Zoology
Miami University
Oxford, OH 45056
EIN No. 316401599A1
-------
CONTENTS
FINDINGS 8-1
CHAPTER 1: APPROACH AND OBJECTIVES 8-2
CHAPTER 2: ECOLOGICAL MECHANISMS AND LITERATURE REVIEW 8-4
ABIOTIC STRESS 8-4
INTERSPECIFIC INTERACTIONS 8-6
POPULATION GROWTH RATES 8-8
OVERWINTERING 8-8
GEOGRAPHIC EXPANSION 8-9
SEASON LENGTH 8-10
PLANT-PEST SYNCHRONY 8-11
SUMMARY OF TRENDS FOR CLIMATE CHANGE SCENARIOS INFLUENCING
ECOLOGICAL MECHANISMS FOR CROP/PEST INTERACTIONS 8-13
SIMULATION STUDIES OF AN INSECT-PLANT
INTERACTION 8-14
The Simulation Model 8-14
Plant Submodel 8-16
Insect Submodel 8-16
Simulations 8-17
Results 8-18
Caveats 8-18
ESTIMATES OF LAND USE AND PEST IMPACT CHANGE AS INFLUENCED BY
TEMPERATURE AND PRECIPITATION CHANGES 8-21
INTERPRETATION OF RESULTS, RESEARCH NEEDS, AND POLICY IMPLICATIONS 8-23
REFERENCES 8-24
11
-------
S tinner
FINDINGS1
The literature survey indicated that temperature increases associated with the climate change scenarios
could result in the following trends: (1) Overwintering survival will increase for migratory and nonmigratory
insect pest species. (2) Northern range will extend for current pests in the higher latitudes and result in southern
species migrating into upper Grain Belt regions. (3) Pest species with multigeneration per year life cycles will
increase in upper Grain Belt areas. (4) Pest populations will establish earlier in the growing season, which would
result in pest populations being particularly abundant during the more susceptible growth stages of crops.
The results of modeling earworm (Heliothis zeaVsoybean interaction showed that this pest, under warming
climate conditions, will increase dramatically in severity in Grain Belt areas. It is predicted that the extent of
damage by this species will result in significant economic losses by grain farmers in the Midwest. This result was
particularly marked with the GFDL scenario. Both GFDL and GISS models predicted northward range
extensions in overwintering capability for four insect pest species: potato leafhopper, black cutworm, sunflower
moth, and green doverworm. This pattern indicates that there is the potential for at least these pests to move
northward and invade cropping systems earlier in the growing season.
Regression analysis incorporating both temperature and precipitation variables indicated significant
cropping system-pest interactions. In general, corn acreage decreases with increasing temperature scenarios.
This change will be greater in southern and western areas than in the more eastern areas of the Grain Belt.
Insect severity increased with temperature rise in the Lake State region and decreased in severity in the southern
region of the Grain or Corn Belt. Insect pest severity on soybean increased with temperature rise in all regions.
The literature survey and modeling suggest that the result of climatic warming could result in a substantial rise
in pesticide use with the accompanying environmental hazards. This report recommends increased emphasis on
integrated pest management targeted toward the predicted climate changes.
'Although the information contained in this report has been funded wholly or partially by the U.S.
Environmental Protection Agency under BIN No. 316401599A1, it does not necessarily reflect the Agency's views,
and no official endorsement should be inferred from it.
8-1
-------
Stinner
CHAPTER 1
APPROACH AND OBJECTIVES
Compared to the existing information on the potential climate change effects on crop physiology and
agronomic production parameters, there has been relatively little effort directed toward assessing the influence
of climate change on plant-pest interactions. Atmospheric increases in temperature and COg, and changes in
moisture regimes, all have the potential to affect interactions between pest organisms and crops either directly
or indirectly.
In this report, we will identify insect pest interactions of major agronomic significance that are sensitive
to climate change. Major emphasis is placed on temperature increase effects, particularly in relation to the
Great Lakes and Midwestern Grain Belt regions.
The objectives of the report are as follows:
1. Formulate key ecological concepts and compile a literature survey relating projected temperature
change scenarios to the major classes of mechanisms by which pest-crop interactions could be
influenced
2. Model the effects of climate change scenarios on a major insect pest-crop relationship.
3. Suggest needs for future research and policy development.
In preparing this report, we have had to make a number of assumptions that we feel are somewhat
unrealistic, though necessary for a first attempt in formulating a response to the risk of undesirable change to
the agroecosystems due to climate change. While we consider it highly probable that there will be major changes
to the agricultural system as we know it over the next decade and beyond, we have been forced to make the
assumption that there will be no fundamental changes to the way agriculture is conducted, except as a function
of climate change. Thus we have based our analyses on agricultural practices as they currently exist, although
the authors are cognizant of projections for agricultural practices to change in the future. In particular, there is
considerable interest in a growing movement toward the adoption of lower chemical and energy inputs to farming
systems in order to sustain and stabilize productivity and to minimize negative environmental inputs. Cropping
system design is expected to change accordingly. Should this trend become a major trajectory for agricultural
change, we believe that the analyses and conclusions concerning climate effects on insect populations will remain
valid. However, the impact of the climate/insect interactions on agricultural productivity and economics would
vary with specific agricultural practices and the nature of future cropping systems.
We have strictly limited our discussion to basic biological processes that we know to be temperature driven,
and attempted to construct likely scenarios of population changes in response to increased heat availability. Our
end point is an assessment of what this may mean to agricultural production and costs. Essentially we have taken
a population biology approach and measured the outcome on crop damage resulting from shifts in spatial and
temporal distribution combined with changed, usually increased, longevity and fecundity of pest insects. We have
interpreted crop damage in terms of costs in control measures as well as yield reduction: both these factors are
central issues for the farmer attempting to make a living from the land.
Because of our own expertise, we have limited our discussion to two crops: corn and soybeans. They are,
in the Midwest, the two most important crops, and so constitute a major agricultural investment. We foresee
that there may be shifts in the economics of these crops due to both macro-economic trends and to climate
warming. We have made the assumption that these crops will remain important for the next two or three
decades. Our arguments, however, apply to other crops that we have not considered: rangeland, forests, cotton,
as well as small grains and vegetables will be influenced by shifts in pest status. For some of these crops,
8-2
-------
Stiimer
management and predictive models exist that might be used to investigate in more detail possible changes to
the distribution and economics of these commodities.
Throughout this report we have assumed we know which pests will be dominant. This is probably the least
supportable assumptioa We know so little about the millions of insects in the world that it is virtually impossible
to predict which ones will be pests: there are innumerable instances of closely related species living in the same
crop, only one of which is ever a serious pest. We know far less about the nonpest species, so that the impact
of climate change on pest status is impossible to estimate. We believe that it is entirely possible that a warming
trend could reduce the pest status of one species, only to increase the status of a relative. To be able to guess
which potential pests will actually become pests under certain circumstances requires long-term monitoring of
pest and nonpest species' populations. This essential work has never been done in North America
In spite of these caveats, we unhesitatingly predict that there will be changes to insect-plant relationships, and
we outline what kind of changes are possible, and furthermore we predict that these plant-pest interactions will
contribute to changes in the economics of agriculture.
8-3
-------
S tinner
CHAPTER 2
ECOLOGICAL MECHANISMS AND LITERATURE REVIEW
Within this section, we develop concepts and review the literature on some of the major mechanisms
through which climate change may impact upon pest-plant interactions. Table 1 summarizes the major crop
pests by region for the reader's convenience.
ABIOTIC STRESS
Adverse abiotic conditions can stress both arthropods (Scriber and Slansky, 1981) and plants (Boyer, 1982).
"Stressed" indicates that arthropod or plant physiology has been deflected away from optimum by an external
factor (e.g., heat, moisture), and such stress results in reduced survival, growth, or reproduction (Odum et al.,
1979). Drought stress will likely increase in frequency in the midwestern and southern United States as the
climate of these regions becomes warmer and drier. Drought is often associated with pest outbreaks
(McPherson and Allen, 1981; Wallner, 1987). Drought, however, can both directly impede the growth of some
pest populations by creating suboptimal conditions for reproduction and survival (e.g., Stinner et at, 1982), and
indirectly benefit population growth of others by altering plant characteristics (e.g., food quality) (Mattson and
Haack, 1987a,b). Also, the effects of drought on plants and pest populations interact in complex ways with
xenobiotics including air pollutants, for example, ozone (Heagle et al., 1987; Flagler et al., 1987; King and Nelson,
1987; King, 1987) and sulfur dioxide (Black, 1982; Hughes et al., 1982; Hughes et al., 1983; Griffith and
Campbell, 1987), as well as pesticides (Wrensch, 1985; Kogan and Turnipseed, 1987). Despite these complexities,
most current evidence indicates that pest populations will be enhanced by warmer, drier conditions, particularly
in the northern portions of their ranges and in spring when cool temperatures and rainfall limit population
growth.
Direct effects of stressful climate conditions on arthropod reproduction, survival, and movement are
documented for many species. Reproduction, fecundity, and hatching success of foliage-dwelling arthropods can
be markedly reduced by temperature and humidity that are too high or low (e.g., Sprenkel and Rabb, 1981;
Moscardi et al., 1981b). Fecundity and larval survival of Mexican bean beetles (Epilachna varivestis). a pest of
soybeans and many other leguminous crops, can be severely reduced by moderately high temperatures and low
humidity (Kitayama et al., 1979; Lockwood et al., 1979; Sprenkel and Rabb, 1981; Wilson et al., 1982; Mellors
and Bassow, 1983). However, fecundity of spider mites (Tetranvchus spp.X pests of many crops including
soybeans, apples, maize, sorghum, wheat, and tomatoes, is favored by warmer temperatures and reduced at high
humidity (Wrensch, 1979,1985). Indeed, during the summer of 1988 when the entire U.S. Corn Belt experienced
its worst drought in the past 100 years along with higher than normal temperatures, twospotted spider mites
(Tetranvchus urticae) became a severe problem on soybeans throughout the whole region. In Ohio alone,
acreage sprayed was estimated to be over 1 million acres, with cost-to-control and loss estimates in that single
state placed at $15-20 million (unpublished data). In other Corn Belt states where soybean acreage is much
higher, estimates of acreage sprayed and losses were even greater. Estimates indicate that up to 8 million acres
were sprayed with insecticides for spider mite control
Parasitoids released to control pests have highest emergence rates within relatively narrow ranges of
humidity and temperature (e.g., Powell et al., 1981; Orr et al., 1985). For pest species with soil-dwelling stages,
suboptimal soil moisture (e.g., Regniere et al., 1981) and temperature (e.g., Marrone and Stinner, 1983) can
significantly reduce egg and/or larval survivorship in the soil Marrone and Stinner (1983, 1984) found that
survival of bean leaf beetle (Cerotoma trifurcata) eggs was greatest at intermediate levels of soil moisture and
temperature, and that survival of larvae and pupae was lower in dry soils. Survivorship of larval and adult pests
and predators will likely be enhanced by the predicted climate changes. Survivorship of pests can be dramatically
lowered by ephemeral, adverse weather events such as heavy rainfall (Jones, 1979; Puttaswamy and
Channabasavana, 1981; Wellington and Trimble, 1984), but summer populations may reach high numbers when
8-4
-------
Stinner
Table 1. Major Insect Pests and Pathogens by Region for Corn and Soybeans
REGION INSECTS
PATHOGENS
Northern Plains
Lake States
Corn Belt
Delta States
Southeastern
States
Northern Plains
Lake States
Corn Belt
Delta States
Southeastern
States
CORN
Corn rootworms, European corn
borer, cutworms, mites,
grasshoppers
Cora rootworms, European corn
borer, armywonn, cutworms,
stalk borers
Cora rootworms, European corn
borer, cutworms, stalk borers,
wireworms
European corn borer, cutworms,
armywonn, fall armywonn, white
grubs
Billbugs, European corn borer,
cutworms, wireworms, southern
com rootworm
Stalk rots, charcoal
rot, ear rots, goss'
leaf blight, seed rots
Stalk rots, eyespot,
ear rots, smut,
northern leaf spot
Stalk rots, northern
corn leaf blight, ear
rots, southern corn
leaf blight, nematodes
Charcoal rot, stalk
rots, ear rots, viruses,
helminthosporium
Aspergillus, nematodes,
stalk rots, viruses,
seed rots
SOYBEANS
Green doverworm, Mexican bean
beetle, bean leaf beetle, corn
earworms, cutworms
Green doverworm, cutworms,
spider mites, potato leaf
hopper, bean leaf beetle
Mexican bean beetle, bean
leaf beetle, Japanese
beetle, seed corn maggot,
cornearworm
Cora earworm, soybean
loopers, stinkbugs,
velvetbean caterpillar,
green doverworm
Cora earworm, stinkbugs,
soybean loopers, velvetbean
caterpillar, bean leaf
beetle
Phytopthora, seed rots,
charcoal rot, pod and
stem blight, stem canker
Phytopthora, seed rots,
brown stem, pod and stem
blight, brown spot
Phytopthora, cyst
nematodes, pod and
stem blight, brown stem,
brown spot
Anthracnose, cyst
nematodes, pod and stem
blight, charcoal rot,
seed rots
Cyst nematodes, root knot
nematodes, stem canker,
brown spot, and stem
blight
REGIONS: Northern Plains « Colorado, Kansas, Nebraska, N. Dakota, S. Dakota
Lake States - Michigan, Minnesota, Wisconsin
Corn Belt - Illinois, Indiana, Iowa, Missouri, Ohio
Delta States - Arkansas, Louisiana, Kentucky, Mississippi,
Tennessee, Oklahoma, Texas
Southeastern States « Alabama, Georgia, N. Carolina, S. Carolina
8-5
-------
S tinner
conditions are drier and populations are not checked by such intermittent events. Survivorship of predators may
also be enhanced by moderately drier conditions as, for example, survivorship of juvenile spiders is reduced by
heavy rainfall (LeSar and Unzicker, 1978).
Climate conditions may differentially affect the survivorship of males and females skewing sex ratios (Walker,
1984), and hence potentially influencing population growth rates (see below). Movement of arthropods, both
local (e.g., Blau and Stumer, 1983) and migratory (Rankin and Rankin, 1979), will likely be increased by the
predicted climate changes. Temperature, humidity, and wind seem particularly important in determining the
proportion of arthropods moving (Wolfenbarger, 1975) and the spatial extent of their movements (see Rabb and
Kennedy, 1979; Pedgley, 1982; MacKenzie et al., 1984). Warmer temperatures are often associated with greater
movement. For example, Ives (1981) found that local movement rates of coccinellids (a major predator of
aphids) significantly increased with temperature. Also, increased rates of movement of many species are often
associated with stressful environmental conditions. Kennedy and Smitley (1985) report that movement of spider
mites is stimulated by declining host plant quality, high population density, and particularly, low relative humidity.
Aerial transport of pathogens is also strongly affected by temperature, humidity, and wind (Pedigo et al., 1983;
Davis, 1987). The indirect effects of abiotic stress on pest arthropods through changes in host plant quality may
have greater impact than the direct effects described above. Host plant quality is determined by the balance
between numerous morphological and biochemical factors including the host plants' content of nutrients
(particularly nitrogen-containing compounds) and defensive compounds (Rhoades, 1979; Strong et aL, 1984;
Brodbeck and Strong, 1987), and is an important factor affecting pest populations because it is a major
determinant of arthropod reproduction (Wermelinger et al., 1985; WeUings and Dixon, 1987), survival (e.g.,
Rufener et al., 1986), and movement (Rankin and Singer, 1984). Host plant quality for herbivores is often
enhanced by drought-induced plant stress (White, 1984; Mattson and Haack, 1987a,b), but maybe lowered when
plants are stressed by conditions including foliar damage (Carroll and Hoffman, 1980; Reynolds and Smith, 1985;
Kraemer et al., 1987), soil nutrient deficiency (Muller et al., 1987), and numerous other environmental factors
(see Kogan and Paxton, 1983).
Reduced host plant quality due to plant stress has been been associated with increased concentrations of
plant defensive compounds (Gershenzon, 1984), but can be caused by changes in foliar nutrient or water content
(Scriber and Slansky, 1981), or increases in the C/N (carbon/nitrogen) ratio of foliage (see Lincoln et al., 1984).
As atmospheric concentrations of CO, rise, C/N ratio will increase, and higher C/N ratio of foliage will
stimulate greater feeding by some herbivores (Lincoln et al., 1984). Major crop types will be differentially
affected by such increased herbivory because C/N ratio will increase more for plants with C, metabolism (e.g.,
soybeans, wheat, rice) than for C4 plants (e.g., corn, sorghum, sugarcane) (Kimball, 1983). Because both
increases in host plant quality due to plant stress and rising C/N ratio will stimulate feeding by pests, greater
crop damage is likely for most climate change scenarios.
INTERSPECIFIC INTERACTIONS
Interspecific interactions among pests, pathogens, parasites, and predators are of major importance in
agroecosystems because there is strong evidence that natural enemies, in many cases, significantly reduce pest
populations (Price, 1987). Interspecific interactions in agroecosystems can be highly complex (Duffey and Bloem,
1986), and although agroecosystems often have low species richness (Odum, 1984), they usually possess all major
structural components and types of species interactions found in natural systems (Risser, 1986). Current
understanding of both natural and agricultural ecosystems indicates that a minimum of three trophic levels ~
plant, pest, natural enemy - interact to determine the dynamics of pest populations (Lawton and McNeilL 1979;
Price et aL, 1980; Duffey and Bloem, 1986).
Furthermore, a strong argument can be made that interspecific interactions are best viewed as part of the
functioning of whole agroecosystems (House and Stinner, 1983; Wiens, 1984; Tilman, 1986; O'Neill et al., 1986).
For example, spatial variability in soil moisture, nutrient availability, or plant competition can affect host plant
quality for herbivorous arthropods (Harborne, 1982; Coley et al., 1985, and see above). Herbivores tend to
aggregate in patches with higher host plant quality (Hassell and Southwood, 1978; StiUng et al., 1982), and
natural enemies (predators, parasites and pathogens) cause greatest mortality where prey are concentrated
8-6
-------
Stinner
(Hassell and Southwood, 1978; Yeargan et al., 1983; Reichert and Lockley, 1984; Ewald, 1987). The linkage
between all trophic levels, from soil nutrients to parasitoids, is clear and permits a broad range of complex
dynamics throughout agroecosystem food webs.
Climate changes may have strong effects on interspecific interactions within agroecosystems. Pest-parasite
and pest-predator interactions should be particularly sensitive to such changes because temperature often
differentially affects interacting species (Hughes et al., 1984; Strong et al., 1984; Mattson and Haack, 1987a),
potentially "releasing" pests from control by natural enemies (Risen, 1987). For example, aphid populations
(aphids are pests of major crops including corn, cotton, alfalfa, peas, hops, and potatoes) are able to maintain
positive growth at lower temperatures than their parasites and predators and may escape control by natural
enemies during periods of cool weather (Wettings and Dixon, 1987). However, warm temperatures enhance the
activity of predators and parasites (e.g., searching rates, attack rates, and handling times) and make them capable
of causing the local extinction of aphid populations (Van Emden, 1966). Although local extinction of a pest
species may seem beneficial for crop production (see Murdoch et al., 1985), such a phenomenon may also result
in the decimation or extinction of natural enemies due to starvation, thus increasing the probability of pest
outbreaks at a later time (Ferro, 1987). Pests usually have greater dispersal abilities than their natural enemies
(Price, 1976; Mayse and Price, 1978; Rey and McCoy, 1979). The models of May (1986) suggest that
pest-pathogen dynamics are most stable when both pest and pathogen are maintained together in the
environment at low abundances, and Hassell (1981) reports that biological control of pests is more successful
in perennial than annual cropping systems, probably because perennial crops permit a continuous interaction
between predator and prey.
A second major way in which climate changes may affect interspecific interactions involving pests and natural
enemies within agroecosystems is through effects on crop plants. Greater crop growth associated with CO,
increase (Rodgers et al., 1983) probably will not affect pest populations because few species are currently limited
by the amount of food available (White, 1984); however, increases in food quality due to plant stress (see above,
Abiotic Stress) may permit dramatic increases in the growth rate of pest populations (White, 1984; Brodbeck
and Strong, 1987), obviating control by natural enemies that have greatest impact on slow-growing pest
populations (Schultz, 1983; Duffey and Bloem, 1987). Sap-feeding arthropods (e.g., planthoppers that are major
pests of rice) seem most strongly favored by stress-induced changes in host plant quality (McNeill and Prestige,
1982; Buckley, 1987). Decreases in host plant quality that slow the growth of pest populations may enhance the
effectiveness of natural enemies in controlling pest populations (Schultz, 1983; Duffey and Bloem, 1986).
A third mechanism in which climate change may alter interspecific interactions involving pests is by alteration
of the structure of pest-natural enemy communities within agroecosystems (Liss et al., 1986). Geographical range
extensions and phenological shifts in species occurrences and relative abundances will create new interactions
between species and probably alter the dynamics of existing interactions. These changes may be particularly
dramatic where generalist predators and parasites (Le., those that prey on numerous insect species) predominate,
such as in soybean agroecosystems (Turnipseed and Kogan, 1983). Switching of a generalist predator between
prey types may stabilize simple predator-prey systems (Luck, 1984), but in more species-rich systems, generalist
predators or parasites may compete for prey (Liss et aL, 1986), and "apparent" competition (Holt, 1977) may
occur between prey species. Also, beneficial predators and parasites may victimize each other, potentially
reducing their abilities to control pests (Messenger, 1975). Evidence that shifts in community interactions may
destabilize existing communities and result in outbreaks of pests comes from studies of (1) species introductions
(e.g., Messenger, 1975), and (2) disruption of food webs by insecticides (see Metcalf, 1986; Ferro, 1987), and
modeling studies of species removals (Redfearn and Pimm, 1987). Whether the climate changes associated with
a doubling of CO, will stabilize or destabilize interspecific interactions involving arthropod pests will depend on
the following: (1) the type of crop and stage of development; (2) the type, magnitude and duration of plant stress;
(3) the types and developmental stages of the pests and natural enemies involved; and (4) the differential impact
of climate change on pests and natural enemies. We currently lack much of the ecological information needed
to make predictions about the outcomes of such complex interactions for even the best known cropping systems.
However, the differential effects of weather on pests and natural enemies (see aboveX and the association of pest
outbreaks in agroecosystems with periods of warm, dry weather (Risch, 1987), strongly suggest to us that
destabilization of arthropod communities and a consequent increase in the frequency of pest outbreaks is likely.
8-7
-------
S tinner
POPULATION GROWTH RATES
Whether a potential pest causes economic damage in a particular year is often determined by the
favorableness of the environment for population growth. Conditions of climate, food quantity and quality, and
the abundance and activity of natural enemies determine environmental favorableness. Climate is, perhaps, the
most important factor because it also strongly affects food quality (see Abiotic Stress), the functional and
numerical response of parasites and predators (Hassell and Southwood, 1978X and the virulence and spread of
pathogens (e.g., Pedigo et a]., 1983; Boucias et al., 1984). Climate directly influences six variables that
significantly affect the growth of pest populations: (1) initial population size (i.e., the number of overwintering
survivors); (2) developmental rate (and hence time to reproductive maturity); (3) per capita lifetime fecundity;
(4) survivorship of immature and adult stages; (5) sex ratio of the population (particularly for species with
parthenogenic capability); and (6) movement (i.e., dispersal) rates. Effects of climate change on overwintering
survival, fecundity, survivorship, and movement are each dealt with in other sections of this section; hence, only
developmental rate and sex ratio are elaborated here. The rate of arthropod physiological development
(developmental rate) can strongly influence the growth of pest arthropod populations (Snell, 1978; Hughes et al.,
1984) and is largely determined by temperature, humidity, and food quality and quantity (Gordon, 1984).
Developmental rate is maximized at optimal values of temperature and humidity (e.g., Wrensch, 1985), and
distinct arthropod biotypes (populations adapted to local environmental conditions) have often been identified
(e.g., Buckley, 1987). Shifts in rainfall or temperature regimes, therefore, might result in increased or decreased
developmental rates, depending on whether prevailing environmental conditions are below or above optimum.
Spring temperatures, and those at the northern edge of a species range, probably are lower than optimum; thus,
temperature increases would result in shorter development times and more rapid increase of pest populations.
Also, the developmental rates of many pest species will be indirectly favored by warmer, drier conditions because
of increases in food quality (see Abiotic Stress). Where the growth rates of arthropod populations are altered
by climate changes, smaller arthropods, e.g., aphids and mites, will likely be more affected than larger ones.
Although smaller arthropods are more sensitive to adverse climate conditions (e.g., desiccation, heavy rainfall)
than larger arthropods, smaller species usually have greater reproductive potential (e.g., greater fecundity, shorter
development times, more generations per year), and hence are better able to track climatic fluctuations (Ito,
1980). Population growth rates may also be enhanced by a shift in the sex ratio (M/F) toward more females
(Hughes et al., 1984). Rapidly growing populations are characterized by a female-biased sex ratio (e.g., mites
[Sabelis, 1985], aphids [Wellings and Dixon, 1987], pentatomids [Schumann and Todd, 1982]). Various external
factors can influence sex ratio, including population density (Wrensch, 1985), food quality (Moscardi et al.,
1981a), and differential mortality (Walker, 1984). The principal climate variable influencing sex ratio seems to
be temperature. For example, Shepard and Gale (1977) found that higher temperatures favored the production
of male Pediobius foveolatus (an important parasitoid of Mexican bean beetles), and hence would decrease
population growth of this beneficial species. Sex ratio may be a particularly important variable driving the
growth rates of populations containing all-female strains (Mitter and Schneider, 1987) or that are cyclically
parthenogenetic (e.g., aphids and mites). Species with all-female strains occur in the orders Lepidoptera,
Homoptera, and Coleoptera (Mitter and Schneider, 1987). The patterns in arthropod growth rates described
above indicate that a warmer, drier climate will most enhance the population growth rates of small, sap-feeding
arthropods with highly variable sex ratio or parthenogenic strains (e.g., mites and homopterans). Control of this
group of pest species will be most difficult because pesticide resistance is prevalent and develops rapidly among
these species (Metcalf, 1986). Population growth rates of all pest arthropod species will benefit from increases
in temperature through its effect on development rates, particularly in spring and in the northern regions.
OVERWINTERING
Warmer temperatures will impact insect biology during those months in which crops are not growing or are
dormant. The most critical time is during the cold months when pests endemic to a region must overwinter for
survival. Higher winter temperatures have the potential to increase the survival rate of many pest species. Most
insects falling into this group usually survive very well during the winter temperature extremes that occur, e.g.,
European corn borer and corn rootworms on corn, Mexican bean beetle and bean leaf beetle on soybeans.
8-8
-------
S tinner
However, extremes in low temperatures can play a significant role as a mortality factor that can serve to
drastically reduce the following season's insect population. During the winter of 1983-84, a severe cold spell in
the Midwest (two consecutive days of temperatures down to -24 C with little snow cover to act as insulation)
was a partial reason given for the nearly total population reduction of both Mexican bean beetles and bean leaf
beetles on soybeans (the other condition adding to their mortality was a drought during the summer of 1983).
A similar decline in overwintering survival brought about by a short period of extreme cold winter temperatures
was observed in Louisiana for bean leaf beetle in 1982 (Payah and Boethel, 1985). As average winter
temperatures rise, the extreme low temperatures associated with overwintering mortality might not occur as
readily, Le., diurnal variability would remain relative to increased average temperatures. Thus, insects currently
overwintering in the North that are affected by such extremes should successively overwinter more frequently,
with higher survivorship.
Numerous pests of corn and soybeans, unable to overwinter in northern regions owing to cold winter
temperatures, migrate from wanner climes prior to becoming economic problems. Black cutworms on corn in
northern midwestern states (Von Raster and Showers, 1982; Beck, 1987) and green cloverworm on soybeans in
all the Midwest (Wolf et al., 1987) are two important examples of such crop pests. Although warmer average
temperatures during the winter might allow these insects to overwinter farther north, it is still uncertain whether
they will become greater problems in these newer overwintering areas. Green cloverworm, for example, is not
a significant problem on soybeans where the insect currently overwinters. Examining the insect's damage
potential in areas of their current overwintering is helpful in determining their future damage potential
An additional consideration should be given to those insects that are currently not pests in the Midwest
because they do not migrate (1) in sufficient numbers or (2) early enough to cause economic losses. As an
example, corn earworm passes through numerous generations in more southern climes, with the second
generation damaging corn and the third and fourth generations causing damage to soybeans in late summer
(Pedigo et al., 1981). The com earworm migrates to more northern locations in midsummer and causes serious
damage to sweet corn and, less frequently, to field corn. Populations do not have sufficient time to develop on
soybeans, and thus, are not considered a problem on soybeans in the Midwest. If the ability of the com earworm
to overwinter in northern regions becomes greater because of wanner winters, population levels in the North
might develop that could cause greater damage to field corn and, later in the summer, to soybeans (see Modeling
section).
In summary, warmer winter conditions will probably improve the overwintering ability of certain insects while
allowing others, currently unable to overwinter in the Midwest, to overwinter farther north. Whether this will
increase an insect's damage potential is unclear; however, insects currently affected by severe winters should be
more often present and in sufficient numbers in the spring.
GEOGRAPHIC EXPANSION
Related to overwintering ability is the geographical expansion that might occur with many corn and soybean
pests. For insects such as the black cutworm and green cloverworm (Wolf et al., 1987) that migrate from
warmer dimes, any ability to overwinter farther north will increase their chances to extend their range northward.
However, both these insects already reach the more northern limits of crop production. An insect that does not
reach the more northern crop limits, or does so in insufficient numbers to be of importance, is the corn earworm.
As discussed earlier, if this insect were able to overwinter farther north, we would see greater populations in the
Midwest, sufficient to accumulate numbers on field corn and then move over to soybeans.
Where geographical expansion by an insect is of greater importance is in the predicted movement of crop
areas into Canada. The question is whether insect pests in the Midwest will move concomitantly with the crop
into these more northern regions. Because many of the insects are able to overwinter and develop into damaging
populations under current conditions found in the Midwest, it is likely that they will expand into the more
northern regions along with the crop. Indeed, Kogan (1981) has suggested for soybean pests that most of the
oligophagous species associated with the crop adapted rather quickly throughout much of the range of soybean
cultivation. He argued that the insects' expansion of their range into the most northern areas was probably
8-9
-------
S tinner
limited by climatic factors as much as the lack or presence of the crop, and is one possible reason why those
northern soybean growing areas do not have major pest problems. Assuming that other environmental
conditions change appreciably along with temperature changes in the areas into which crops expand (e.g., soil
type and moisture [rootworms and bean leaf beetles] presence of overwintering sites [Mexican bean beetles:
Bernhardt and Shepard, 1978]), the pests will probably move with the crop. Whether there will be a time lag
between when a crop expands into an area and when the pest insect arrives is unclear. However, if we can
assume that temperature changes will occur slowly, allowing for a gradual migration northward of the crop rather
than an abrupt movement, there should be an expansion of both crop and pest into these more northern areas.
(Of special interest regarding geographical expansion is the situation with the European corn borer, for which
we find varying generations - from multiple in the South, two in the Midwest, to one in the far North. This
difference in voltinism, resulting from the effect of environmental conditions on diapause induction, is discussed
in the section on seasonal length).
Finally, certain southern insect pests that do not occur in the Midwest may expand farther northward with
the advent of wanner temperatures. Indeed, we expect this expansion to occur because many of these insects
might gain the capability to exist in the Midwest through the ability to overwinter farther north or by increasing
their migrating range. However, we must not automatically assume all southern species will expand northward.
An example serves to illustrate this point. The velvetbean caterpillar, a pest of soybeans in southern states, is
not able to overwinter in most gulf coast states; overwintering occurs in the United States only in southern
Florida, with most of the migrants arriving from Central America (Newsom et al., 1980; Buschman et al., 1981).
Because future predictions are that the southern states will not see an increase in higher temperatures to the
extent of the northern states, we expect the ability of the velvetbean caterpillar to overwinter in other regions
of the United States to remain unchanged. It is unlikely that this insect species will be able to overwinter in the
southern United States as it still needs the tropical conditions of Central America. If predictions are validated
and southern states do become warmer and more tropical, the velvetbean caterpillar might indeed gain the ability
to overwinter throughout the southern United States. In this latter scenario, the insect could become a much
greater problem in the South, and also begin migrating to more northern regions.
SEASON LENGTH
If it is assumed that the predicted increase in average temperature occurs over the entire year, rather than
being limited to the summer season, there should be a lengthened growing season for both corn and soybeans.
Whether this addition to the cropping season is measured in weeks or months, and whether the extension is at
the beginning of the season in the spring or the end in the fall (or both), would be speculative. Of primary
importance, prior to examining the effect of a lengthened season on insect pests, is whether or not the crops will
be modified by breeding to exploit the longer season. Soybeans (and com to a lesser amount) are strongly
affected by day length, which will not change as temperatures rise. Crops will first need to be developed for the
longer season that will be experienced, knowing that the day lengths will remain the same. This varietal
development will in all likelihood occur.
The effect of a lengthened season on insects will be variable depending on the insect's biology. For those
insects that have two or more generations per year, a lengthened season might allow for additional generations
and, thus, potentially higher population buildups. On soybeans, both the Mexican bean beetle and bean leaf
beetle can have multiple generations, partially depending on the season length. Both insects have two
generations in the Midwest and have three generations farther south in the Carolinas. In Minnesota, the bean
leaf beetle has only one generation (Loughran and Ragsdale, 1986). Obviously, given a longer season for
population development, both insects might develop a third generation in the Midwest and a second in more
northern regions. Knowing that higher daily temperatures will accompany an increased season length will
enhance the possibility of additional generations; thus, the damage potential will increase as well.
A longer season length, however, will not automatically increase the potential for additional generations.
Other contributing factors to an insect's voltinism have to be taken into consideration. The European com borer
has three ecorypes in North America- a northern population, which is univoltine in Minnesota and Quebec -
a bivoltine ecotype appeared in Quebec in 1976, but has remained at low levels (Boivin et al., 1986); a central
8-10
-------
Stinner
population in the Midwest (Iowa, Nebraska, and Ohio), which is bivoltine; and a southern ecotype (Alabama,
Georgia, and southern Missouri) with three to four generations per year (Showers et al., 1975). This pest
species has been successful in North America partially because of its adaptive plasticity, which allows it to modify
voltinism to suit local conditions, voltinism being controlled by the diapause response (Showers, 1981). Lee and
Spence (1986) believe that variation in diapause increases the insect's fitness in uncertain environments and
suggest that at the northwestern edge of its range, Alberta, Canada, univoltinism is its safest strategy because a
second-generation population would be unlikely to mature before the onset of winter in most years. Voltinism
in the European corn borer is under control of thermoperiod - Beck (1982) suggested a thermoperiodic
response threshold of 175 C - and photoperiod. The current understanding is that there is a significant
interaction between the two; Beck (1987) states that the photoperiodic determination of larval diapause is
strongly influenced by superimposed thermoperiods. As temperatures in more northern areas warm, this
interaction between thermoperiod and photoperiod will become complex. While the former will be affected, the
latter will not. The question is, with a warmer thermoperiod but similar photoperiod, will induction of diapause
in the European com borer be delayed so as to permit larvae from the first generation to pupate, leading to a
second generation. Assuming that this occurs, a lengthened season (combined with more rapid developmental
rates) should allow for the completion of a second generation. The bivoltine ecotype might begin to occur in
locations farther north where it currently does not exist and might also become the dominant ecotype in areas
where both ecotypes currently are present (e.g., Quebec).
For univoltine insect species, or those insects where phenology of the insect and plant is a more important
criterion for causing damage, a lengthened season will probably not be important. Rootworms, a univoltine
species, will not be affected by an increased season unless the biology of the insect changes to increase the
number of generations (an unlikely scenario). For numerous insects, damage is limited to feeding during certain
growth stages of the crop. Black cutworm is a species whose damage potential is directly related to crop
phenology; it is a problem on seedling corn. The species does not cause damage to corn during other plant
growth periods. A lengthened season will not cause the insect to damage other plant parts. Of more importance
with rootworms and the black cutworm are their phonological relationships with the crop, i.e., whether the insects
are present during the time when the damage potential is greatest (see section on phenology).
PLANT-PEST SYNCHRONY
The importance of phenological synchrony in affecting the severity of pest damage to crops has been
demonstrated for a number of major pest species and their host plants ~ European corn borer on corn, potato
leafhopper on alfalfa, black cutworm on corn, and green stink bug on soybean, to name several specific
interactions. It is not atypical to observe rigid requirements placed upon insects by phenological development
of their host plants. Insect development can be severely hampered if synchronization with the host plant lags
even several weeks out of phase (Strong et al., 1984). The key aspects of synchrony impact upon crop damage
center on either (1) pest development exceeding crop growth (larger growth stages of insects can consume more
crop biomass per unit time), or (2) pests being abundant when a crop is particularly vulnerable. Both of these
mechanisms can be impacted very markedly by temperature. Our working hypothesis is that, in general, a
wanning trend would favor more rapid insect development over crop growth. During spring months in the Grain
Belt, several of the major insect pests could accumulate degree days more rapidly, leading to earlier
establishment of pest populations while crops are at young, relatively susceptible growth stages. This pattern
would result in a lessened ability of crops to outgrow pest damage.
Waldbauer and Kogan (1976) presented data that indicated that a 3°C temperature increase during April
and May would cause bean leaf beetle (Cerotoma trifurcata) to developmental time to decrease from 90 to 60
days. This temperature change would result in soybeans being attacked at a growth stage when the plants are
quite vulnerable. Moreover, this temperature shift and more rapid development by the bean beetle would cause
second-generation populations to peak before soybean senescence and consequently lead to increased damage
on maturing seed pods.
Growth stages of crops vary in their susceptibility to insect feeding, Moscardi et al. (1981) reported that
feeding by the velvetbean caterpillar declined as soybean phenology progressed. Additionally, 6th instar larvae
8-11
-------
S tinner
of this pest consumed almost five times as much plant biomass as did 5th instar individuals. Growth rates in this
insect, as with many species, are very temperature dependent, so that when insect growth proceeds at a more
rapid rate because of increased temperature regimes, larger herbivorous insects will not only be present early
in the growing season but they will also consume more biomass from the earlier, more vulnerable growth stages
of crop plants.
In another experiment, Margolies and Kennedy (1984) found that the population of the spider mites
(Tetranvchus urticae) increased at a significantly greater rate on crops, both corn and legumes, when the plants
were in the early vegetative stages versus later growth stages. Warmer growing temperatures would cause spider
mite populations, a major pest of corn and soybean in the Midwest, to develop more rapidly, coinciding with
earlier growth stages of these crops. This shift would presumably result in more damage by this pest species.
In fact, this phenomenon was illustrated during the severe twospotted spider mite problem in the summer of 1988
referred to earlier in this report. In previous drought situations, mites have increased and damaged midwestern
field crops in late summer (late August through early September) usually after the grain has matured. During
1988 when the drought began early in the spring and continued throughout most of the summer, the mite
population explosion was also much earlier (mid-late JulyX and the damage to soybeans occurred during the
critical flowering, pod development, and pod fill growth stages. The damage that occurred during these stages
resulted in yield losses much greater than had ever been experienced.
A caveat needs to be added here that wanner climatic scenarios should also have the effect of shifting crop
planting dates to a time earlier in the season. This shift could result in crops developing earlier and in turn
influence synchrony between pests and crop net effect of this planting date. Pest development shifting could
either be positive or negative in terms of crop growth.
Moisture conditions and drought incidence can also significantly influence crop-pest synchrony and
subsequent crop damage. Menezes et al. (1985) found that green stink bug (Nezara viridula) populations and
their impact on soybean yields were appreciably influenced by precipitation patterns impacting upon insect-plant
phenological synchronies. Similarly, spider mite populations are higher and damage more severe under drought
conditions. This pattern results both from increased mortality on the mites and a lessened ability of the crops
to outgrow arthropod damage under water stress conditions (Simpson and Council, 1973).
8-12
-------
Stinner
SUMMARY OF TRENDS FOR CLIMATE CHANGE SCENARIOS INFLUENCING
ECOLOGICAL MECHANISMS FOR CROP/PEST INTERACTIONS
Based on the preceding literature review and our knowledge of plant/pest interactions, it is predicted that
the climate change scenarios will:
1. Increase drought stress, leading to more insect outbreaks.
2. Increase C/N ratios in crop biomass thus stimulating feeding by some pest species.
3. Change temperature scenarios such that they have different impacts on pest bisects and their natural
enemies - predator and parasites.
4. Three-level trophic interactions - crop, pest, and natural enemies - are key in predicting the impact
of climate change on pest severity.
5. Create new interactions between pest and natural enemies as geographical range extensions and
phenological shifts in species occur.
6. Favor developmental rates of many pest species under warmer climate scenarios.
7. Increase overwintering survival potential for both migratory and nonmigratory insect pest species.
8. Result in northward range extensions of current pests in the higher latitudes and result in southern
species migrating into upper Grain Belt regions.
9. Increase the number of generations of pest species with multigeneration per year life cycles in upper
Grain Belt areas.
10. Result in pest populations being particularly abundant during the more susceptible growth stages of
crops if pest populations establish themselves earlier in the growing season.
8-13
-------
S tinner
SIMULATION STUDIES OF AN INSECT-PLANT INTERACTION
Expansion of population or overwintering range as a result of an increase in winter temperatures is likely to
have profound effects on the population dynamics of many temperature-limited insect species. This will be
especially true for many migrant species that rely on overwintering survival in milder regions for their future
summer success in cooler northern areas.
The present and potential overwintering ranges of four migrant species, which are major northern pests, are
represented in Figures 1-4. The potato leafhopper, Empoasca fabae. is a serious pest on a wide variety of crops,
including alfalfa and soybeans. At present, the species overwinters only in a narrow band along the coast of the
Gulf of Mexico (Figure 1). Increases in winter temperature predicted by both the GFDL and GISS models
suggest a doubling or tripling of the overwintering range, respectively. Increases in winter population density of
this size would increase the invasion populations in the northern states by similar factors. The invasions would
also be earlier; both features are likely to lead to increased insect load and damage.
The green cloverworm, sunflower moth, and black cutworm have more extensive ranges (Figures 2-4), so
the increase in overwintering range as a result of climate warming will not be proportionally as great, but even
these small increases are likely to have an impact on the population dynamics of these species. These are but
a few of the insect pests that would be likely candidates to increase their damage potential as a result of climate
warming.
We know enough about the population dynamics of all four of these species to speculate concerning what
might happen but, unfortunately, not enough to be able to investigate in detail the likely consequences of climate
change. In this section we will describe and use an insect-plant simulation model to investigate the potential for
climate-induced changes in agricultural practices and profits in soybeans. The model, called SICM (for Soybean
Integrated Crop Management), was made available to us by researchers at North Carolina State University. In
particular, we would like to thank Dr. Gail Wilkerson who sent us the FORTRAN code in a form suitable for
these investigations.
SICM was developed to allow researchers to study the effects on yield or profit of desired changes in crop
management or changes in the biological structure of the crop, or crop-pest interactions (Jones et al., 1986). It
comprises several component submodels: the crop, soil and water, economics, and three pests (velvet bean
caterpillar, southern green stink bug, and corn earworm). All three species are pests of soybeans in North
Carolina, but only the corn earworm, Heliothis zea. will be considered here. It is a cosmopolitan species, highly
polyphagous, and highly mobile. At present it is a serious and regular pest of soybeans only in the southern
states, but it is an ideal model insect for our purposes because it is a marginal pest on soybeans throughout the
Corn Belt. The purpose of this investigation is to determine if climate change could trigger a change in the
damage distribution of this species and, if so, to estimate the costs of increased corn earworm numbers on
soybeans in the Corn Belt.
The Simulation Model
A variation of SICM, called STRATEGY, allows the simulation of many conditions and weather years to
evaluate various production changes. STRATEGY uses historical weather data and expected market information
to develop strategies to manage future cropping seasons. The major differences between SICM and STRATEGY
are the interactive interface to select inputs to the model, and the level of reporting by the model. SICM allows
detailed interactive selection of inputs and interactive graphic reporting of results
for users who may not be familiar with computers. It is intended ultimately to be a practical management tool.
STRATEGY, on the other hand is intended specifically for research work. It is the latter version we used to
8-14
-------
S tinner
Patatl laafhaapar
Enaaasca fabaa
GISS
CFDL
Present
Figure 1. Potato leafhopper overwintering range.
8-15
-------
Stinner
Grtta Glowworm
Plathypoia scabra
rosoit
Figure 2. Green cloverworm overwintering range.
8-16
-------
Stinner
Sunflower moth
Homoeosoma electellum
Figure 3. Sunflower moth overwintering range.
8-17
-------
S tinner
Black cutworm
Agrotis ipsilon
Figure 4. Black cutworm overwintering range.
8-18
-------
S tinner
investigate the potential consequences of atmospheric warming. What follows is a brief description of the plant
and insect submodels extracted from the Jones et al. (1986) description of SICM. The economic and soil models
are not described: soil and economic conditions were assumed to be constant.
Plant Submodel
For many crop plants, the partitioning of dry matter to growth in the different plant parts depends on the
stage of development. In order to predict accurately the growth and yield of these crops, one must be able to
predict the timing and duration of the various growth stages. Soybeans have several distinct stages of
development between planting and final maturity. The stages of development for soybean, as described by Fehr
et al. (1971) were used as a basis for the plant model, although additional stages are now defined Six of the
eleven stages depend on both night length (scotophase) and on temperature; others depend only on temperature.
The reproductive stages of soybean depend on temperature and scotophase: soybeans flower sooner in long
nights than in short ones. Some cultivars are highly sensitive to night length, whereas others are not. Soybean
sensitivity to night length is the major source of soybean adaptation to various latitudes. A phenological model
was developed to permit the prediction of timing and duration of stages under a range of latitudes and planting
dates (Mishoe et al., 1985). For the temperature-sensitive phases, a physiological time approach was adopted
This approach is equivalent to degree-days for the range of suitable temperatures, but is more flexible in that
it allows for decreased rates of development at higher temperatures. Plant transpiration is based on root length
and soil water distribution and on the potential evapotranspiration rate, determined by weather conditions. The
potential evapotranspiration rate is calculated from temperature and radiation. The actual plant transpiration,
and thus the actual water extraction by roots, is the lessor of the potential rate and the water supplying capability
of the soil-root system. The soil-root system capability is calculated from the rate of root water absorption per
unit length of root (using the steady-state solution of the radial flow equation) multiplied by root length density.
Since soil hydraulic conductivity varies with water content, water absorption may be limited by root length or by
water content. Water stress occurs when the water supply to the root system is less than demand
Photosynthesis is reduced in proportion to the ratio of supply to demand Leaf and stem growth rates are
reduced when the supply to roots is less than 1.5 times the potential evapotranspiration rate. Thus as soil dries
or as roots are damaged, water stress occurs, first causing reductions in leaf and stem expansion and increased
partitioning of CH2O to roots, followed by decreased photosynthesis and transpiration.
Insect Submodel
The corn earworm (Heliothis zea) is extremely polyphagous, attacking a number of crops including corn and
soybeans, upon which very large populations can develop, causing extreme defoliation (Neunzig, 1969). Because
corn earworm populations can vary markedly from field to field and from year to year, population models
developed for this species must be year- and site-specific.
The corn earworm submodel is derived from the population dynamics model of Stinner et aL (1974). In it,
populations are divided into six developmental stages: eggs, small larvae, medium larvae, large larvae, pupae, and
adults. Larvae are separated by size rather than instar because field identification of instars is often difficult, and
separating samples on the basis of size is much easier. The small size class corresponds roughly to larval instars
I and II; medium to instars n and IV; and large to the later instars.
In the population model, the age structure (a) is maintained within each developmental stage (i). The change
in population density (N, - N,(a,t)) in each stage can be expressed as:
dN, dN,
„. + ._» .m N - TN, + E,
•. * !• II
dt da
8-19
-------
S tinner
where: m, is the stage-specific mortality, T, is a developmental function, and Ei is the number of insects that
enter a stage, either immigrating into the field or developing from the previous stage. For stages small
larva to adult (i = 2..6), E, is the number of insects that make the transition from the previous stage.
For the adult stage, E, includes the immigrating adults also. For eggs, E, is the total number of eggs
ovipositing on day t. E,, m,, and T, are functions of a and t.
The effect of varying temperature is incorporated into the model by calculating physiological days. One
physiological day is defined as the proportion of development completed in one day at a reference temperature,
taken to be 27°C. For example, at 18°C an average com earworm requires twice as long to go from egg to adult
as it does at 27° C. Thus, one day at 18°C equals OJ5 physiological days. A lower development threshold of 13°C
was used (Mangat and Apple, 1966a).
The variability of individual developmental rates within the population was accounted for in the model by
using a cumulative distribution function, F|(a-), taken from Stinner et al. (1975), where a_ represents
physiological age.
The proportion of those insects present at time t of age a, that complete stage i at time t, T,(a,t), is obtained
by differentiating the cumulative distribution function
T, (a,t) = --«- --------- [1 - m,]
1 -
The term [1 - nij] is included because the cumulative distribution function is based on the number that survive
a stage and develop into the next.
The total number of eggs laid by a female corn earworm in her lifetime is a function of temperature (Isely,
1935). The proportion of those eggs laid on any particular day is a function of temperature and adult stage
(Isely, 1935). Each day, the number of eggs each female lays is calculated from that day's temperature. The
eggs are distributed on the plant according to data of Johnson et al. (1975).
Mortalities inflicted by predators, mjp, insecticides, mjT, and food shortages, m(F, are incorporated into the
model explicitly. Functionally, mortality is represented as
m. = 1 - (1 - m,B)(l - m|F)(l - m|p)(l - m|T)
where m,B represents background mortality due to factors other than predators, insecticides, or food shortages.
For all stages except pupae, background mortality (miB) is considered to be a constant proportion per day
throughout the season: 0.10, 0.05, 0.05, 0.02, 0.05 for eggs, small, medium and large larvae, and adults
respectively. Background mortality for pupae is a function of crop leaf area index.
Mortality due to food (mjF) is equal to zero unless leaf area has been zero for at least 2 days. This factor
increases linearly thereafter until it reaches 1.0 ten days after leaf area has been zero.
Corn earworm is a migratory insect that overwinters across most of its range as diapausing pupae (Mangat
and Apple, 1966a, 1966b). Pupae breaking diapause develop into adults that usually lay the first brood on corn.
Later generations invade other crops, including soybeans. The first adults appear in light traps in June or July;
larvae are not usually found in soybean fields until early August.
Simulations
SICM was run for Castana, Iowa, for which weather and soil condition data were known for 1979. This site
was chosen over several other sites for which data were available, because it is well north of the area where com
earworm is a major pest on soybeans. This Iowa site is also representative of the major grain producing areas
8-20
-------
S tinner
in the Midwest. No insecticide was used in the simulations; this defines a baseline for comparison. Insecticide
application adds significantly to costs, so the expected return must be enough to cover those costs. The insect
population level at which insecticide application becomes desirable is called the economic threshold. Of course
this parameter varies with the current price of the commodity, so that the same insect population density on two
different occasions may have different economic thresholds. For this reason, the economic environment was also
kept constant in all simulations.
Three sets of simulations were performed: the first using actual weather data for Castana, Iowa, in 1979; the
second used GISS predictions for a doubling of the CO2 level; and the third used the GFDL predictions. To
obtain predictions for Castana, the weather model outputs were interpolated in both space and time to give
expected daily temperature and precipitation changes (Figure 5). The expected changes were then added to the
1979 weather data to create modified weather data files, which were fed into the simulation. The outputs were
compared to those from the model using unmodified weather data. The output variables recorded are the date
at which the corn earworm population peaked, the peak population, the yield in bushels per hectare, and the
profit per hectare.
Results
Table 2 gives a summary of the output from SICM. Using the 1979 weather data (the first block in the
table), the corn earworm moths left corn in early August, too late to be a serious pest on soybeans. Even fairly
large influxes of corn earworm into soybean have no effect on yield or profit if they are not early, but an early
influx does significantly reduce yield and profit. This contrasts with the modified weather scenarios. Not only
do the higher summer temperatures and reduced precipitation reduce the yield of this particular cultivar, but they
have a synergistic effect on the plant-insect interaction. The major impact of the higher temperatures on insect
population is more rapid growth, earlier in the season. This results in a larger, earlier emigration from corn, and
a corresponding increase in damage to the soybeans. Interestingly, the two climate scenarios produce
quantitatively similar results. Qualitatively, the GFDL scenario is much worse, probably due to its lower
predicted summer precipitation. Under no circumstances can a profit be made under the GFDL scenario
without insecticide application: the combination of early influx and larger than average numbers completely
destroys an already moisture stressed crop. With the GISS scenario, normal and larger than normal influxes
have no effect on profitability. But, again, an early influx is devastating.
iveats
As mentioned in the description of the model, the latitude and temperature regime dictates the soybean
cultivar that can be used. In these simulations, cultivar "Wayne" was used throughout. It is an ideal choice for
central Iowa for today's conditions, but the simulations show that if the climate does warm, and especially if
summer precipitation declines, "Wayne" would be an inappropriate choice. Also, the absence of any insecticide
is an unusual situation. Other insects feed on soybeans, against which a fanner would use an insecticide. Thus,
the results are not as bleak as they at first appear. They do suggest that an expanded use of insecticide would
be inevitable, a point well illustrated by the spider mite outbreak during 1988, during which insecticide use on
soybeans increased over tenfold.
8-21
-------
S tinner
(A)
8
7
6
n
*
O
3
4-*
CO
0)
a
1=
00
20
40-
60-
80-
00-
20-
40-
60-
80-
00
+ GISS Prediction
x GFDL Prediction
567
Month
8 10 11 12
(B)
1.50
+ GISS Prediction
x GFDL Prediction
-1.50
567
Month
Interpolation by cubic spline
Figure 5. (A) Temperature change and (B) precipitation change in Iowa.
8-22
-------
S tinner
Table 2. Summary of Effects of Climate Change on Soybean-Hellothls Interactions
Influx of Hellothls zea:
Climate
Model
NONE
GISS
GFDL
Output
Variable None
Day of peak n/a
Larval density n/a
Yield (bu/ha) 24
Profit ($/ha) 37
Day of peak n/a
Larval density n/a
Yield (bu/ha) 20
Profit ($/ha) 11
Day of peak n/a
Larval density n/a
Yield (bu/ha) 10
Profit ($/ha) -62
Normal
223
0.45
24
37
220
0.25
20
11
223
0.35
10
-62
Early
198
58
22
27
192
81
17
-1
200
100
8
-72
Large
223
1.70
24
37
220
1
20
11
223
1.40
10
-62
Early &
Large
198
230
21
15
192
330
8
-72
200
430
0
-130
8-23
-------
Stinner
ESTIMATES OF LAND USE AND PEST IMPACT
CHANGE AS INFLUENCED BY TEMPERATURE AND PRECIPITATION CHANGES
In Table 3, we have summarized the results of regression analysis incorporating temperature and precipitation
variables. The objective was to develop a summary of climate change effects on pest severity based upon mean annual
temperature changes of 2°C and 4°C.
The major patterns that emerged from this exercise are as follows:
1. Corn acreage decreases with temperature increase.
2. Changes in corn acreage are greater in the southern and western regions than in the northern, more humid
areas.
3. Pest severity on corn increased with temperature rise only in the Lake State region, remained unchanged in
the Northern Plains and Corn Belt, and decreased in severity in the southern regions.
4. No trend in pathogen severity was observed.
5. Soybean acreage was less affected than corn by temperature change.
6. Pest severity on soybean increased with temperature rise in all regions, but was most marked in the northern
areas.
7. Pathogen severity on soybean increased with temperature and most notably in the northern areas.
8-24
-------
S tinner
Table 3. Summary of Current and Projected Scenarios For Land Usage and Percent
Yield Loss Due to Pest and Pathogen Damage
SCENARIO NORTHERN LAKE CORN DELTA SOUTHEAST SOUTHERN SUMMARY
TEMP.,PREC. PLAINS STATES BELT STATES STATES PLAINS
MILLIONS OF ACRES -
CURRENT
+2, +0
+4, +0
+2 , +10
+4 , +20
PERCENT YIELD
CURRENT
+2, +0
+4, +0
+2 , +10
+4, +20
PERCENT YIELD
CURRENT
+2, +0
+4, +0
+2 , +10
+4, +20
13.1
8.9
3.9
10.6
6.8
LOSS
27.0
27.8
27.8
27.5
27.2
LOSS
13.1
13.1
13.1
12.8
12.5
MILLIONS OF ACRES -
CURRENT
+2, +0
+4, +0
+2 , +10
+4, +20
PERCENT YIELD
CURRENT
+2, +0
+4, +0
+2, +10
+4, +20
PERCENT YIELD
CURRENT
+2, +0
+4, +0
+2, +10
+4, +20
5.0
3.4
.9
6.3
6.3
LOSS
.0
3.7
7.4
5.4
11.5
LOSS
7.2
10.2
13.3
10.3
13.4
CORN
13.
11.
7.
12.
10.
WITHOUT
9.
11.
13.
11.
13.
WITHOUT
7.
7.
7.
7.
7.
9
6
6
8
0
37.2
31.3
26.6
33.1
29.2
PESTICIDES
8
7
6
4
0
27.8
28.2
27.0
27.9
26.4
PESTICIDES
6
6
6
3
0
8.3
8.3
8.3
8.0
7.7
2.3
.6
.0
1.0
.2
FROM
15.
13.
11.
13.
10.
FROM
7.
7.
7.
7.
6.
INSECT
3
4
5
1
9
3.2
.6
.0
.9
.1
PESTS
25.1
23.2
21.3
22.9
20.7
1.2
.0
.0
.0
.0
- CORN
15.
13.
11.
13.
10.
3
4
5
1
9
70.
53.
38.
58.
46.
23.
24.
24.
23.
23.
9
0
1
4
3
4
3
4
9
6
PATHOGENS - CORN
5
5
5
2
9
10.9
10.9
10.9
10.6
10.3
7.
7.
7.
7.
6.
5
5
5
2
9
9.
9.
8.
8.
8.
1
0
7
7
3
SOYBEANS
5.
5.
4.
6.
6.
WITHOUT
t
4.
8.
6.
12.
WITHOUT
2.
5.
8.
5.
9
1
0
6
6
30.3
28.9
25.5
30.3
30.3
PESTICIDES
9
0
6
0
7
1.8
6.4
11.1
8.5
15.2
PESTICIDES
3
3
4
4
8.5
6.3
9.3
12.4
9.4
12.5
14.
12.
10.
14.
14.
FROM
18.
23.
27.
25.
31.
FROM
30.
33.
37.
3
1
0
3
3
INSECT
5
1
8
2
9
8.4
6.7
4.9
8.4
8.4
PESTS
61.2
65.8
70.5
67.9
74.6
PATHOGENS -
9
9
0
34.0
37.1
12.1
15.1
18.2
15.2
18.3
1.
.
•
1.
1.
1
0
0
1
1
65.
0
56.2
45.3
67.0
67.0
- SOYBEANS
18.5
23.1
27.8
25.2
31.9
SOYBEANS
30.9
33.9
37
34
37
.0
.0
.1
13.2
16.7
20.9
19.2
25.9
12
15
18
15
18
.6
.0
.1
.5
.6
REGIONS: NORTHERN PLAINS - COLORADO, KANSAS, NEBRASKA, NORTH DAKOTA, SOUTH
DAKOTA. LAKE STATES - MICHIGAN, MINNESOTA, WISCONSIN. CORN BELT - ILLINOIS,
INDIANA, IOWA, MISSOURI, OHIO. DELTA STATES - ARKANSAS, LOUISIANA, KENTUCKY,
MISSISSIPPI, TENNESSEE. SOUTHEAST STATES - ALABAMA, GEORGIA, NORTH CAROLINA,
SOUTH CAROLINA. SOUTHERN PLAINS - OKLAHOMA, TEXAS. TEMP., PREC. - DEVIATION
FROM MEAN ANNUAL TEMPERATURE IN DEGREES C AND FROM TOTAL ANNUAL PRECIPITATION
IN CENTIMETERS.
8-25
-------
S tinner
INTERPRETATION OF RESULTS, RESEARCH NEEDS, AND POLICY IMPLICATIONS
For both the literature search on ecological mechanisms and the modeling efforts, both temperature and
precipitation patterns were key variables affecting crop-pest interactions. The overall, emerging pattern is that
pest severity will increase with the predicted wanning patterns and that this effect will be more marked in the
higher latitudes. Interactive effects with precipitation patterns are more difficult to discern and depend quite
heavily on crop stress phenomena.
If pest damage to crops does increase as predicted, then the consequent economic and environmental
ramifications would be substantial. As our modeling indicated, the economic losses due to only one species could
be sufficient to remove what is an already small profit margin on field crops. In all probability, pesticides would
be effective in at least reducing damage severity. Yet the economic and environmental costs of remedial
treatment can also be substantial. Pesticide application would cost up to $20 per acre per treatment. The
increased use of pesticides would create an additional threat to ecosystem integrity through soil and water
contamination and through the increased risks to public health. The predicted increases in pest severity resulting
from climate change would not only cause more intense problems in areas where the pest-crop interactions
already occur, but would also lead to economic damage in new areas particularly in the northern farming regions.
If the farming community is not to rely increasingly on potentially environmentally damaging chemicals, then
there surely will be an increased need for alternative pest management strategies. In terms of policy implications,
perhaps the primary take home message is that under the anticipated climate changes, more than ever, there will
be a need for intensive support of alternative pest management strategies that utilize a range of tactics, such as
biological control, genetic resistance, and innovative cropping systems, instead of increased reliance on chemical
toxicants.
Research needs addressing these concerns can be summarized as follows:
1. Carry out targeted studies on the effects of climate change variables on specific pest-crop interactions
in controlled and semicontrolled environments.
2. Carry out long-term studies to develop cropping systems that use natural ecological processes to
buffer the negative impacts of climate change.
3. Perform modeling studies that examine a range of site- and species-specific interactions.
4. Develop of a better understanding of thermoperiodic response in more pest and potential pest
species, so that modeling studies may be performed for more species.
5. Perform integrative modeling studies that incorporate crop, soil, pest and biological control
phenomena.
6. Conduct research programs directed toward designing integrative pest management strategies relevant
to the climate change scenarios.
8-26
-------
S tinner
REFERENCES
Beck, S. D. 1982. Thennoperiodic induction of larval diapause in the European corn borer, Ostrinia nubilalis.
J. Insect. PhysioL 28: 273-277.
Beck, S. D. 1986. Effects of photoperiod and thermoperiod on growth of Agrotis ipsilon (Lepidoptera:
Noctuidae). Ann. Entomol. Soc. Am. 79: 821-828.
Bernhardt, J. L. and M. Shepard. 1978. Overwintered Mexican bean beetles: emergence from overwintering
sites, fecundity, fertility, and longevity. Ann. Entomol. Soc. Am. 71: 724-727.
Black, V. J. 1982. Effects of sulphur dioxide on physiological processes in plants. Pages 67-91 in M. H.
Unsworth and D. P. Ormrod, editors. Effects of gaseous air pollution in agriculture and horticulture.
Butterworth Scientific, London, UK.
Blau, W. S. and R. E. Stinner. 1983. Temporal flight patterns in the Mexican bean beetle (Coleoptere
Coccinellidae) and their relation to weather. Environmental Entomology 12:1047-1054.
Boivin, G., C. Ritchot, P. Mattel, and D. G. R. McLeod. 1986. Evidence for regional differences in the
emergence pattern of the European com borer (Lepidoptera: Pyralidae) in Quebec. J. Econ. Entomol.
79:978-980.
Boucias, D. L. 1984. Susceptibility of the velvetbean caterpillar and soybean looper (Lepidoptera: Noctuidae)
to Nomuraea rileyi: effects of pathotype, dosage, temperature, and host age. Journal of Economic
Entomology 77:247-253.
Boyer, J. S. 1982. Plant productivity and environment. Science 218:443-448.
Brodbeck, B. and D. Strong. 1987. Amino acid nutrition of herbivorous insects and stress to host plants. Pages
347-364 in P. Barbosa and J. C. Schultz, editors. Insect Outbreaks. Academic Press, San Diego,
California, USA,
Buckley, R. C. 1987. Interactions involving plants, homoptera, and ants. Annual Review of Ecology and
Systematics 18:111-135.
Buschman, L. L., H. N. Pitre, C. H. Hovermale, and N. C. Richards, Jr. 1981. Occurrence of the velvetbean
caterpillar in Mississippi: winter survival or immigration. Environ. Entomol. 10: 45-52.
Carroll, C. R. and C. A. Hoffman. 1980. Chemical feeding deterrent mobilized in response to insect herbivory
and counter adaptation by Epilachna trederimnotata Science 209:414-416.
Coley, P. D., J. P. Bryant, and F. S. Chapin, ffl. 1985. Resource availability and plant antiherbivore defense.
Science 230:895-899.
Davis, J. M. 1987. Modeling the long-range transport of plant pathogens in the atmosphere. Annual Review
of Phytopathology 25:169-188.
Duffey, S. S. and K. A. Bloem. 1986. Plant defensive-herbivore-parasite interactions and biological control.
Pages 135-183 in M. Kogan, editor. Ecological theory and integrated pest management practice. John Wiley
& Sons, New York, New York, USA.
Ewald, P. W. 1987. Pathogen-induced cycling of outbreak insect populations. Pages 269-286 in P. Barbosa
and J. C. Schultz, editors. Insect Outbreaks. Academic Press, San Diego, California, USA.
8-27
-------
Stinner
Fehr, W.R., C. E. Caviness, D. T. Burmood, and J. S. Pennington. 1971. Stage of development descriptions for
soybeans, Glycine max (L.) Merrill. Crop. Sti. 11:929-931.
Ferro, D. N. 1987. Insect pest outbreaks in agroecosystems. Pages 195-215 in P. Barbosa and J. C. Schultz,
editors. Insect Outbreaks. Academic Press, San Diego, California, USA.
Flagler, R. B., R. P. Patterson, A. S. Heagle, and W. W. Heck. 1987. Ozone and soil moisture deficit effects
on nitrogen metabolism of soybean. Crop Science 27:1177-1184.
Gershenzon, J. 1984. Plant secondary metabolite production under stress. Recent advances in Phytochemistry
18:273-320.
Gordon, H. T. 1984. Growth and development of insects. Pages 53-77 in C. B. Huffaker and R. L. Rabb,
editors. Ecological Entomology. John Wiley & Sons, New York, New York, USA.
Griffith, S. M., and W. F. Campbell. 1987. Effects of sulfur dioxide on nitrogen fixation, carbon partitioning,
and yield components in snapbean. Journal of Environmental Quality 16:77-8.
Harborne, J. B. 1982. Introduction to ecological biochemistry, 2nd Edition. Academic Press, New York, New
York, USA.
Hassell, M. P. 1981. Arthropod predator-prey systems. Pages 105-131 in R.M. May, editor. Theoretical
ecology, 2nd edition. Sinaur Associates, Sunderland, Massachusetts, USA.
Hassell, M. P., and T. R. E. Southwood. 1978. Foraging strategies of insects. Annual Review of Ecological
Systems 9:75-98.
Heagle, A. S., R. B. Flagler, R. P. Patterson, V. M. Lesser, S. R. Shafer, and W. W. Heck. 1987. Injury and
yield response of soybean to chronic doses of ozone and soil moisture deficit. Crop Science 27:1016-1024.
Holt, R. D. 1977. Predation, apparent competition, and the structure of prey communiites. Theoretical
Population Biology 12:197-229.
House, G. J. and B. R. Stinner. 1983. Arthropods in no-tillage soybean agroecosystems: community composition
and ecosystem interactions. Environmental Management 7:23-28.
Hughes, P. R., A. I. Dickie, and M. A. Penton. 1983. Increased success of the Mexican bean beetle on
field-grown soybeans exposed to sulfur dioxide. Journal of Environmental Quality 12:565-568.
Hughes, P. R., J. E. Potter, and L. H. Weinstein. 1982. Effects of air pollution on plant-insect interactions:
increased susceptibility of greenhouse-grown soybeans to the Mexican bean beetle after plant exposure to SO2.
Environmental Entomology 11:173-176.
Hughes, R. D., R. E. Jones, and A. P. Gutierrez. 1984. Short-term patterns of population change: the life
system. Pages 309-357 in C. B. Huffaker and R. L. Rabb, editors. Ecological Entomology. John Wiley &
Sons, New York, New York, USA.
Isely, D. 1935. Relation of hosts to abundance of cotton bollworm. Ark. Agr. Exp. Stn. Bull. 320:1-30.
Ito, Y. 1980. Comparative ecology. Cambridge University Press, New York, New York, USA.
Ives, P. M. 1981. Estimation of coccinellid numbers and movement in the field. Canadian Entomologist
113:981-997. ^
8-28
-------
Stinner
Johnson, M. W., R. E. Stinner, and R. L. Rabb. 1975. Ovipositional response of Heliothis zea (Boddie) to its
major hosts in North Carolina. Environ. Entomol. 4:291-297.
Johnson, K. B., C. G. Watrin, E. B. Radcliffe. 1988. Vertical distributions of potato leafhopper nymphs
(Homoptera- Cicadellidae) on potatoes relative to leaf position, plant age and population size. J. Econ. EntomoL
81: 304-309.
Jones, M. G. 1979. Abundance of aphids o cereals from before 1973 to 1977. Journal of Applied Ecology
16:1-22.
Jones, J. W., J. W. Mishoe, G. Wilkerson, J. L. Stimac, and W. G. Boggess. 1986. Integration of soybean crop
and pest models. In Integrated Pest Management on Major agricultural Systems. Eds. R. E. Frisbie and P. E.
Adkisson. Texas Ag. Exp. Stn. Misc. Pub. MP-1616.
Kennedy, G. G. and Dr. R. Smitley. 1985. Dispersal. Pages 233-242 in W. Helle and M. W. Sabelis, editors.
Spider mites: their biology, natural enemies and control, Volume 1A. Elsevier, New York, New York, USA.
Kimball, B. A. 1983. Carbon dioxide and agricultural yield: an assemblage and analysis of 430 prior
observations. Agronomy Journal 75:779-788.
Kimball, B. A., J. R. Mauney, G. Guinis, F. S. Nakayama, S. B. Idso, J. W. Radin, D. L. Hendrix, G. D. Butler,
Jr., T. I. Zarembinski and P. E. Nixon, in. 1985. Effects of increasing atmospheric CO2 on the yield
and biomass of crops. Greenlock Series No. 027, VS. Dept. of Ecology, Wash., D. C., 75 pp.
King, D. A. 1987. A model for predicting the influence of moisture stress on crop losses caused by ozone.
Ecological Modelling 35:29-44.
King, D. A. and W. L. Nelson. 1987. Assessing the impacts of soil moisture stress on regional soybean yield
and its sensitivity to ozone. Agriculture, Ecosystems, and Environment 20:23-35.
Kitayama, K., R. E. Stinner, and R. L. Rabb. 1979. Effects of temperature, humidity, and soybean maturity
on longevity and fecundity of the adult Mexican bean beetle, Epilachna varivestis. Environmental Entomology
8:458-464.
Kogan, M. 1981. Dynamics of insect adaptations to soybeans: impact of integrated pest management. Environ.
Entomol. 10:363-371.
Kogan, M.and Paxton. 1983. Natural inducers of plant resistance to insects. Pages 153-171 in P. A. Hedin,
editor. Plant resistance to insects. American Chemical Society, Washington, D. C., USA..
Kogan, M. and S. G. Turnipseed. 1987. Ecology and management of soybean arthropods. Annual Review of
Entomology 32:507-538.
Kraemer, M. E., M. Rangappa, W. Gade, and P. S. Benepal. 1987. Induction of Trypsin inhibitors in soybean
leaves by Mexican bean beetle (Coleoptere Coctinellidae) defoliation. Journal of Economic Entomology
80:237-241.
Lawton, J. H. and S. McNeill. 1979. Between the devil and the deep blue sea: on the problem of being a
herbivore. Pages 223-244 in R. M. Anderson, B. D. Turner, and L R. Taylor, editors. Population dynamics.
The 20th Symposium of the British Ecological Society. Blackwell Scientific
Publications, London, UK.
Lee, D. A. and J. R. Spence. 1986. Phenology of the European com borer (Ostrinia nubilalis) in Alberta.
Can. EntomoL 118:943-956.
8-29
-------
S tinner
LeSar.CD.andJ.D.Unzicker. 1978. Soybean spiders: species composition, population densities, and vertical
distribution. Illinois Natural History Survey; Biological Notes No. 107:14 pages.
Lincoln, D. E., N. Sionit, and B. R. Strain. 1984. Growth and feeding response of Pseudoplusia indudens
(Lepidotpera: Noctuidae) to host plants grown in controlled carbon dioxide atmospheres. Environmental
Entomology 13:1527-1530.
Liss, W. J., L. J. Gut, P. H. Westigard, and C. E. Warren. 1986. Perspectives on arthropod community
structure, organization, and development in agriculural crops. Annual Review of Entomology
31:455-478.
Lockwood, D. F., R. L. Rabb, R. E Stinner, and R. K. Sprenkel. 1979. Effects of two hosts and host phenology
on three population parameters of adult Mexican bean beetle in North Carolina. Journal of the Georgia
Entomological Society 14:220-229.
Loughran, J. C. and D. W. Ragsdale. 1986. Life cycle of the bean leaf beetle, Cerotoma trifurcata (Coloeptera
Chrysomellidae), in southern Minnesota. Ann. Entomol. Soc. Am. 79: 34-38.
Luck, R. F. 1984. Principles of arthropod predation. Pages 498-529 in C. B. Huffaker and R. L. Rabb, editors.
Ecological Entomology. John Wiley & Sons, New York, New York, USA.
MacKenzie, D. R., C. S. Barfield, G. G. Kennedy, R. D. Berger, and D. J. Taranto, editors. 1984. The
movement and dispersal of agriculturally important biotic agents. Claitor's Publishing Division, Baton Rouge,
Louisiana, USA
Mangat, B. S. and J. W. Apple. 1966a. Termination of diapause in the com earworm. Proc. N. C. Br. Entomol.
Soc. Am. 21:123-125.
Mangat, B. S. and J. W. Apple. 1966b. Corn earworm development in relation to temperature. J. Econ. Entomol.
59:1005-1006.
Margolies, D. C. and G. C. Kennedy. 1984. Population response of the twospotted spider mite, Tetranychus
urticae, to host phenology in corn and peanut. Entomol. Exp. Appl. 36: 193-1%
Marrone, P. G. and R. E. Stinner. 1983. Effects of soil moisture and texture on the bean leaf beetle, Cerotoma
trifurcata (Forster) (Coleoptere Chrysomelidae). Environmental Entomology 12:136-138.
Mattson, W. J. and R. A. Haack. 1987a. The role of drought in outbreaks of plant-eating insects. BioScience
37:110-118.
Mattson, W. J. and R. A. Haack. 1987b. The role of drought stress hi provoking outbreaks of phytophagous
insects. Pages 365-407 in P. Barbosa and J. C. Schultz, editors. Insect Outbreaks. Academic Press,
San Diego, California, USA.
May, R. M. 1986. Ecology of insect-pathogen interactions and some possible applications. Pages 185-201 in
M. Kogan, editor. Ecological theory and integrated pest management practice. John Wiley & Sons, New York,
New York, USA.
Mayse, M. A. and P. W. Price. 1978. Seasonal development of soybean arthropod communities in east central
Illinois. Agro-Ecosystems 4:387-405.
McNeil], S. and R. A. Prestidge. 1982. Plant nutritional strategies and insect herbivore community dynamics.
Pages 225-235 in J. H. Visser and A. K. Minks, editors. Proceedings of the 5th international symposium on
insect-plant relationships. Pudoc, Wageningen, Netherlands.
8-30
-------
Stinner
McPherson, R. M. and W. A. Allen. 1981. Drought and insects: bad news for beans. Agrichemical Age
April:16-42.
Menezes, E. B., D. C. Herzog, I. D. Teare, and R. K. SprenkeL 1985. Phenological events affecting Southern
green stink bug on soybean. Soil and Crop Science Society of Florida 44: 227-231.
Metcalf, R. L. 1986. The ecology of insecticides and the chemical control of insects. Pages 251-297 in M.
Kogan, editor. Ecological theory and integrated pest management practice. John Wiley & Sons, New York,
New York, USA.
Mellors, W. K. and F. E. Bassow. 1983. Temperature-dependent developmant of Mexican bean beetle
(Coleoptere Coccinellidae) immatures on snap bean and soybean foliage. Annals of the Entomological Society
of America 76:692-698.
Messenger, P. S. 1975. Parasites, predators, and population dynamics. Pages 201-223 in D. Pimentel, editor.
Insects, science, & society. Academic Press, New York, New York, USA.
Mishoe, J. W., J. D. Hesketh, K. J. Boote, D. Herzog, and Y. J. TsaL 1985. Phenology of soybean: development
of a model. Environ. Entomol. 12:432-439.
Mitter, C. and J. C. Schneider. 1987. Genetic change and insect outbreaks. Pages 505-532 in P. Barbosa and
J. C. Schultz, editors. Insect Outbreaks. Academic Press, San Diego, California, USA.
Moscardi, F., C. S. Barfield and G. E. Allen. 1981a. Inpact of soybean phenology on velvetbean caterpillar
(Lepidoptere Noctuidae): oviposition, egg hatch, and adult longevity. Can. Entomol. 113:113-119.
Moscardi, F., C. S. Barfield, G. E. Allen. 1981b. Effects of temperature on adult velvetbean caterpillar
oviposition, egg hatch, and longevity. Annals of the Entomological Society of America 74:167-171.
Muller, R. N., P. J. Kalisz, and T. W. Kimmerer. 1987. Intraspecific variation in production of astringent
phenolics over a vegetation-resource availability gradient. Oecologia 72:211-215.
Murdoch, W. W., J. Chesson, and P. L. Chessoa 1985. Biological control in theory and practice. American
Naturalist 125344-366.
Newsom, L. D., M. Kogan, F. D. Miner, R. L. Rabb, S. G. Tumipseed, and W. H. Whitcomb. 1980. General
accomplishments toward better pest control in soybean. In New Technology of Pest Control, C. B. Huffacker,
ed. WUey-Interscience Pub. 500 pp.
Neunzig, H. H. 1969. The biology of the tobacco hornworm and the corn earworm in North Carolina. N. C.
Agr. Exp. Stn. Tech. Bull 196:1-76.
Odum, E.P. 1984. Properties of agroecosystems. Pages 5-11 in R. Lowrance, B. R. Stinner, and G. J. House,
editors. Agricultural ecosystems: unifying concepts. John Wiley & Sons, New York, New York, USA.
Odum, E. P., J. T. Finn, and E. H. Franz. 1979. Perturbation theory and the subsidy-stress gradient. BioScience
29:349-352.
O'Neill, R. V., D. L. DeAngelis, J. B. Waide, and T. F. H. Allen. 1986. A hierarchical concept of ecosystems.
Monographs in Population Biology 23. Princeton University Press, Princeton, New Jersey, USA,
Orr, D. B., D. J. Boethel, and W. A. Jones. 1985. Development and emergence of Telenomus chloropus and
Trissolcus basalis (Hymenoptera: Scelionidae) at various temperatures and relative humidities. Annals of the
Entomological Society of America 78:615-619.
8-31
-------
Stinner
Payah, W. S. and D. J. BoetheL 1985. Overwintering bean leaf beetles, Cerotoma trifurcata (Forster)
(Coleoptera: Chrysomelidae): density of populations, temporal patterns of movement, and survival. Environ.
Entomol. 14: 65-73.
Pedgley, D. E. 1982. Windborne pests and diseases. Ellis Horwood, Chichester, UK.
Pedigo, L. P., R. A. Higgins, R. B. Hammond, and E. J. BechinskL 1981. Soybean pest management. In CRC
Handbook of Pest Management, Vol m. D. Pimental, ed. CRC Press, Inc. 656pp.
Pedigo, L. P., H. N. Pitre, W. H. Whitcomb, and S. Y. Young. 1983. Assessment of the role of natural enemies
in regulation of soybean pest populations. Pages 39-48 in H. N. Pitre, editor. Natural enemies of arthropod
pests in soybean. Southern Cooperative Series Bulletin 285. South Carolina Agricultural Experiment Station,
Clemson, South Carolina, USA.
Powell, J. E., M. Shepard, and M. J. Sullivan. 1981. Use of heating degree day and physiological day equations
for predicting development of the parasitoid Trissolcus basalis. Environmental Entomology 10:1008-1011.
Price, P. W. 1976. Colonization of crops by arthropods: nonequilibrium communities in soybean fields.
Environmental Entomology 5:605-611.
Price, P. W. 1987. The role of natural enemies in insect populations. Pages 287-312 in P. Barbosa and J. C.
Schultz, editors. Insect Outbreaks. Academic Press, San Diego, California, USA.
Price, P. W., C. E. Bouton, P. Gross, B. A. McPheron, J. N. Thompson, and A. E. Weis. 1980. Interactions
among three trophic level: influence of plants on interactions between insect herbivores and natural enemies.
Annual Review of Ecology and Systematics 11:41-65.
Puttaswamy and G. P. Channabasavanna. 1980. Influence of weather factors and predators on the populations
of the spidermite Tetranychus ludeni (Acari: Tetranychidae). Indian Journal of Acarology 5:69-79.
Rabb, R. L. and G. G. Kennedy, editors. 1979. Movement of highly mobile insects: concepts and metholodology
in research. University Graphics, North Carolina State University, Raleigh, North Carolina, USA.
Rankin, M. A. and S. M. Rankia 1979. Physiological aspects of insect migratory behavior. Pages 35-63 in R.
L. Rabb and G. G. Kennedy, editors. Movement of highly mobile insects: concepts and metholodology
in research. University Graphics, North Carolina State University, Raleigh, North Carolina, USA.
Rankin, M. A. and M. C. Singer. 1984. Insect movement: mechanisms and effects. Pages 185-216 in C. B.
Huffaker and R. L. Rabb, editors. Ecological Entomology. John Wiley & Sons, New York, New York, USA.
Redfearn, A. and S. L. Pimm. 1987. Insect outbreaks and community structure. Pages 99-133 in P. Barbosa
and J. C. Schultz, editors. Insect Outbreaks. Academic Press, San Diego, California, USA.
Regniere, J., R. L. Rabb, and R. E. Stinner. 1981. Popillia japonica: the effect of soil moisture and texture
on survival and development of eggs and Ist-instar grubs. Environmental Entomology 10:654-660.
Regiere, J., 1984. A method for describing and using variability in development rates for the simulation of
insect phenology. Can. Ent. 116:1367-1376.
Rey, J. R. and E. D. McCoy. 1979. Application of island biogeographic theory to pests of cultivated crops.
Environmental Entomology 8:577-582.
Reynolds, G. W. and C. M. Smith. 1985. Effects of leaf position, leaf wounding, and plant age of two soybean
genotypes on soybean looper (Lepidoptera Noctuidae) growth. Environmental Entomology 14:475-478.
8-32
-------
Stinner
Rhoades, D. F. 1979. Evolution of plant defense against herbivores. Pages 1-54 in G. A. Rosenthal and D.
H. Janzen, editors. Herbivores: their interaction with secondary plant metabolites. Academic Press, New York,
New York, USA.
Riechert, S. E. and T. Lockley. 1984. Spiders as biological control agents. Annual Review of Entomology
29:299-320.
Risen, S. J. 1987. Agricultural ecology and insect outbreaks. Pages 217-238 In: P. Barbosa and J. C. Schultz,
editors. Insect Outbreaks. Academic Press, San Diego, California, USA.
Risser, P. G. 1986. Agroecosystems-strucrure, analysis, and modeling. Pages 321-343 In: M. Kogan, editor.
Ecological theory and integrated pest management practice. John Wiley & Sons, New York, New York, USA.
Rodgers, H. H., J. F. Thomas, and G. E. Bingham. 1983. Response of agronomic and forest species to elevated
atmospheric carbon dioxide. Science 220:428-429.
Rufener, G. K., III., R. B. Hammond, R. L. Cooper, and S. K. Martin. 1986. Mexican bean beetle (Coleoptera:
Coccinellidae) developmemt on resistant and susceptible soybean lines in the laboratory and relationship to
field selection. Journal of Economic Entomology. 79:1354-1358.
Sabelis, M. W. 1985. Reproductive strategies. Pages 265-278 in W. Helle and M. W. Sabelis, editors. Spider
mites: their biology, natural enemies and control, Volume 1A. Elsevier, New York, New York, USA.
Schultz, J. C. 1983. Impact of variable plant defensive chemistry on susceptibility of insects to natural enemies.
Pages 37-54 in P. A. Hedin, editor. Plant resistance to insects. American Chemical Society,
Washington, D. C, USA.
Schumann, G. W. and J. W. Todd. 1982. Population dynamics of the southern green stink bug (Heteroptera:
Pentatomidae) in relation to soybean phenology. Environmental Entomology 75:748-753.
Scriber, J. M. and F. Slansky Jr. 1981. The nutritional ecology of immature insects. Annual Review of
Entomology 26:183-211.
Shepard, M., and G. T. Gale. 1977. Superparasitism of Epilachna varivestis (Col: Coccinellidae) by Pediobus
foveolatus (Hym: Eulophidae): influence of termperature and parasitoid-host ratio. Entomophaga 22:315-321.
Showers, W. B., H. C. Chaing, A. J. Keaster, R. E. Hill, G. L. Reed, A. N. Sparks, and G. J. Musick. 1975.
Ecotypes of the European com borer in North America. Enviroa Entomol. 4: 753- 760.
Showers, W. B., M. B. De Rozari, G. L. Reed and R. H. Shaw. 1978. Temperature-reelated climatic effects
on survivorship of the European corn borer. Environ. Entomol. 7:717-723.
Showers, W. B. 1981. Geographic variation of the diapause response in the European corn borer. In Insect
Life History Patterns: Habitat and Geographic Variation, R. F. Denno and H Dingle, eds.
Springer-Verlag, N.Y. pp. 97-111.
Simpson, K. W., and W. A. Connell. 1973. Mites on soybeans: moisture and temperature relations. Environ.
Entomol. 2: 319-323.
Snell, T. W. 1978. Fecundity, development time, and population growth rate. Oecologia 32:119-125.
Sprenkel, R. K. and R. L. Rabb. 1981. Effects of micrometeorological conditions on survival and fecundity of
the Mexican bean beetle in soybean fields. Environmental Entomology 10:219-221.
8-33
-------
S tinner
Stiling, P. D., B. V. Brodbeck, and D. R. Strong. 1982. Foliar nitrogen and larval parasitism as determinants
of leafminer distribution patterns on Spartina alterniflora. Ecological Entomology 7:447-452.
Stinner, R. E., G. D. Butler, J. S. Bacheler, and C. Turtle. 1975. Simulation of temperature-dependent
development in population dynamics models. Can. EntomoL 107:1167-1174.
Stinner, R. E., R. L. Rabb, and J. R. Bradley. 1974. Population dynamics of Heliothis zea and H. virescens in
North Carolina: a simulation model Environ. EntomoL 3:163-168.
Stinner, R. E., J. Regniere, and K. Wilson. 1981 Differential effects of agroecosystem structure on dynamics
of three soybean herbivores. Environmental Entomology 11:538-543.
Strong, D. R., J. H. Lawton, and R. Southwood. 1984. Insects on plants: community patterns and mechanisms.
Harvard University Press, Cambridge, Massachusetts, USA.
Strong, D. R., J. H. Lawton and T. R. E. Southwood 1984. Insects on Plants. Hammond Univ. Press,
Cambridge, Mass. 373 pp.
Tilman, D. 1986. A consumer resource approach to community structure. American Zoologist 26:5-22.
Turnipseed, S. and M. Kogan. 1983. Soybean pests and indigenous natural enemies. Pages 1-6 in H. N. Pitre,
editor. Natural enemies of arthropod pests in soybean. Southern Cooperative Series Bulletin 285. South
Carolina Agricultural Experiment Station, Clemson, South Carolina, USA.
van Emden, H. F. 1966. The effectiveness of aphidophagous insects in reducing aphid populations. Pages
227-235 in I. Hodek, editor. Ecology of Aphidophagous insects. Junk, The Hague, Netherlands.
Von Kaster, L. 1981 Evidence of spring immigration and autumn reproductive diapause of the adult black
cutworm in Iowa. Environ. EntomoL 11: 306-311
Waldbauer, G. P. and M. Kogan. 1976. Bean leaf beetle: Phenological relationship with soybean in Illinois.
Environ. EntomoL 5: 35-44.
Walker, T. J. 1984. Do populations self-regulate? Pages 531-558 in C. B. Huffaker and R. L. Rabb, editors.
Ecological Entomology. John Wiley & Sons, New York, New York, USA.
Wallner, W. E. 1987. Factors affecting insect population dynamics: differences between outbreak and
non-outbreak species. Annual Review of Entomology 31317-340.
Wellings, P. W. and A. F. G. Dixon. 1987. The role of weather and natural enemies in determining aphid
outbreaks. Pages 314-346 in P. Barbosa and J. C. Schultz, editors. Insect Outbreaks. Academic Press, San
Diego, California, USA.
Wellington, W. G. and R. M. Trimble. 1984. Weather. Pages 399-425 in C. B. Huffaker and R. L. Rabb,
editors. Ecological Entomology. John Wiley & Sons, New York, New York, USA.
Wermelinger, B., J. J. Oertli, and V. Delucchl 1985. Effect of host plant nitrogen fertilization on the biology
of the two-spotted spider mite, Tetranychus urticae. Entomologia Experimentalis et Applicata 38:23-28.
White, T. R. C. 1984. The abundance of invertebrate herbivores in relation to the availability of nitrogen in
stressed food plants. Oecologia 63:90-105.
Wiens, J. A. 1984. Resource systems, populations, and communities. Pages 398-436 in P. W. Price, C. N.
Sloboddrikoff, and W. S. Gaud, editors. A new ecology: novel approaches to interactive systems. John Wiley
& Sons, New York, New York, USA.
8-34
-------
Stinner
Williams, J. 1978. Carbon Dioxide, Climate and Society. Proc. IIASA Workshop Pergamon Press, Oxford.
167 pp.
Wilson, K. G., R. E. Stinner, and R. L. Rabb. 1982. Effects of temperature, relative humidity, and host plant
on larval survival of the Mexican bean beetle, Epilachna varivestis Mulsant. Environmental Entomology
11:121-126.
Wolf, R. A., L. P. Pedigo, R. H. Shaw, and L. D. Newsome. 1987. Migration/transport of the green cloverworm,
Plathypena scabra (F.) (Lepidoptera: Noctuidae), into Iowa as determined by synoptic-scale
weather patterns. Environ. Entomol. 16:1169-1174.
Wolfenbarger, D. O. 1975. Factors affecting dispersal distances of small organisms. Exposition Press, Hicksville,
New York, USA.
Wrensch, D. L. 1979. Components of reproductive success in spider mites. Pages 155-164 in J. G. Rodriguez,
editor. Recent advances in acarology, Volume I. Academic Press, New York, New York, USA.
Wrensch, D. L. 1985. Reproductive parameters. Pages 165-170 in W. Helle and M. W. Sabelis, editors.
Spider mites: their biology, natural enemies and control, Volume LA. Elsevier, New York, New York, USA.
Yeargan, K. V., J. W. Todd, A. J. Mueller, and W. C. Yearian. 1983. Dynamics of natural enemy populations.
Pages 32-38 in H. N. Pitre, editor. Natural enemies of arthropod pests in soybean. Southern Cooperative
Series Bulletin 285. South Carolina Agricultural Experiment Station, Clemson, South Carolina, USA.
8-35
-------
IMPACTS OF CLIMATE CHANGE ON THE TRANSPORT OF AGRICULTURAL
CHEMICALS ACROSS THE USA GREAT PLAINS AND CENTRAL PRAIRIE
by
Howard L. Johnson
Ellen J. Cooler
Robert J. Sladewski
Oklahoma Climatological Survey
University of Oklahoma
Norman, OK 73019
Contract No. CR-814591-01-0
-------
CONTENTS
FINDINGS 9-1
CHAPTER 1: INTRODUCTION 9-3
THE PROBLEM: WATER QUALITY IN THE U.S. GREAT PLAINS
AND CENTRAL PRAIRIE 9-3
LITERATURE REVIEW 9-4
PROJECT OVERVIEW 9-4
THE STUDY REGION 9-8
CHAPTER 2: METHODOLOGY 9-9
THE EFFECTS MODEL 9-9
Development of the Modeling System 9-9
Further Limitations of the Model 9-12
THE CLIMATE CHANGE SCENARIOS 9-12
Scenarios Used 9-12
Issues Resulting From the Climate Scenario 9-13
RESULTS 9-13
REGIONAL RESULTS AND INTERPRETATION 9-13
CHAPTER 3: IMPLICATIONS OF RESULTS 9-18
SURFACE WATER IMPLICATIONS 9-22
GROUNDWATER IMPLICATIONS 9-22
REFERENCES 9-25
APPENDIX A: THE STATISTICAL NATURE OF PRZM ESTIMATED MODEL
OUTPUT AND IMPLICATIONS FOR CLIMATE CHANGE
SCENARIO INTERPRETATION 9-29
APPENDIX B: REGIONAL CLIMATE CHANGE RESULTS AND
INTERPRETATION 9-30
-------
Johnson
FINDINGS1
Pesticide pollutants are ranked as a high priority non-point source pollution problem in at least one-half of
the states within the VS. Great Plains and Central Prairie. Potentially toxic agricultural chemicals can be
removed from application sites through degradation, surface runoff, sediment transport, and downward
percolation. An understanding of the importance of these processes under the present environment, as well as
under alternative environmental futures, is essential to the thorough evaluation of potential threats to drinking
water quality.
The present study focused on interactions among agricultural soil/tillage/management systems, pesticide
transport, and the implications of CO, doubling on subsequent water quality in the U.S. Great Plains and Central
Prairie. A model which mathematically simulates the vertical and horizontal surface movement of water and its
contents across a field and through the crop root zone was applied. The critical parameters supplied to the
model included soil profiles, crop phenology, tillage/management practices, and pesticide use. Weather data
included mean daily temperature and precipitation.
These parameters were supplied at 26 locations, spanning four cropping regions; spring wheat, winter wheat,
cotton, and corn. Each site was characterized by three different soils (sand, silt, clay), three tillage/management
systems (conservation, conventional and conventional with crop residues left on the field after harvest), and three
classes of pesticides: those which are highly soluble and persistent, relatively insoluble and persistent, and highly
soluble but not persistent in the soil. The study is limited to those preplan! pesticides applied to the soil surface
or applied to the soil surface and shallowly incorporated.
Primary limitations of the model's use in this application include: 1) the sensitivity of the model to the
selection of cropping system parameters; 2) no modeling of direct impacts of increased CO2 on vegetative cover;
3) no modeling of macro-pore or vadose zone transport; 4) the sensitivity of the results to changes in
precipitation, whose processes are not well modeled by present GCM's (General Circulation Models); and 5)
the care with which regional results must be prepared, interpreted, and applied when derived from a point
estimate model
Statistically significant chemical loading results for each crop region include:
1) Northern Plains (Spring Wheat): Under GFDL-estimated temperature and precipitation conditions,
runoff and erosion of highly soluble short-lived pesticides could increase while erosion of less soluble long-lived
pesticides could decrease. Leaching of all pesticides could decrease.
2) Central Plains (Winter Wheat): Under GISS-estimated temperature and precipitation conditions, runoff
and erosion of all pesticide types studied, on all soils studied, could increase.
3) Southern Plains (Cotton): Under GISS-estimated temperature and precipitation conditions, regional
pesticide runoff could increase. Surface erosion of some pesticides could also increase. Under GFDL-estimated
temperature and precipitation conditions, surface runoff losses of some pesticides could increase.
4) Central Prairie (Corn): Under GFDL-estimated temperature and precipitation conditions, model-
generated regional surface runoff and subsurface leaching of all the chemicals studied declined.
'Although the information in this report has been funded wholly or partly by the U.S. Environmental
Protection Agency under contract no. CR-814591-01-0 to the Oklahoma Climatological Survey, it does not
necessarily reflect the Agency's views, and no official endorsement should be inferred from it.
9-1
-------
Johnson
These results are not forecasts, but represent a range of alternative futures under numerically estimated
weather conditions. Another set of model-generated weather outcomes would likely result in a different set of
pesticide transport results.
9-2
-------
Johnson
CHAPTER 1
INTRODUCTION
THE PROBLEM: WATER QUALITY IN THE US. GREAT PLAINS AND CENTRAL PRAIRIE
Even a brief overview of the status and future of water quality in the Great Plains and Central Prairie can
quickly evolve into an extensive treatise. Such matters are important considerations under current climatic
conditions and are not expected to disappear.
At least two major documents summarizing nonpoint source pollution information on a state-by-state basis
have appeared in recent years: the STEP Report (ASIWPCA, 1984), and the SNAP Report (ASIWPCA, 1985).
Based on these ASIWPCA documents, eight of the twelve states in the Great Plains and Central Prairie region
consider agricultural nonpoint source pollution to be a major source of their water quality problems (ASIWPCA,
1985). Three other states identified agricultural nonpoint source pollution as a significant problem. The SNAP
report indicates that 43% of all river miles assessed and 39% of all lake surface acres assessed sustain at least
moderate impacts from agricultural nonpoint source pollution (ASIWPCA, 1985).
Pesticide pollutants, a specific category of agricultural nonpoint pollution, ranked as a high-priority problem
in six of the twelve study area states: Illinois, Iowa, Kansas, Oklahoma, South Dakota, and Texas (EPA, 1984).
For the US. Great Plains and Central Prairie region, the SNAP report indicates that water use of 60,812 surface
acres and 21,246 river miles is at least threatened by pesticide pollutants. In this same area, there are more
than 200 public and private drinking water wells and more than 45,000 mi2 of aquifers that are known or
suspected to be polluted by pesticides (ASIWPCA, 1985).
Delivery of pesticides to water bodies varies according to crop adsorption characteristics, whether the
chemicals are transported as solutes in water or attached to sediment particles, rainfall, slope and soil type, and
the proximity of the land to a waterway. Newer pesticides have fewer long-term impacts, but are more likely to
be water soluble. Thus, water (rather than sediment) is the primary vehicle by which these chemicals enter
water bodies. In most cases, sediment control measures also control runoff water, but concern remains as to
whether they provide adequate protection. Water quality impairments resulting from pesticides are related to
the toxic effects of pesticides and are due more to concentrations than to total loadings (Maas, et al., 1985).
If pesticide application rates are held constant, techniques such as conservation tillage, terraces, and contouring
assume that pesticide surface loadings will be reduced by reducing surface runoff volume. However, as a result
of decreased dilution these practices may not significantly reduce edge-of-field pesticide concentrations.
In addition to these concerns, toxic water-soluble chemicals in pesticides may be more biologically available
when freely waterborne than they are when bound to sediment (EPA, 1984). Thus, they may cause acute
short-term surface water impacts and eventually have serious effects on ground water resources through
percolation. Two specific case studies in Kansas and Ohio are presented to illustrate the extent of recent
pesticide pollution problems in the study region.
Butler and Arruda (1985) report that since 1977, a total of 21 different pesticides have been found above
their detection limits in Kansas streams and rivers. Over the period of record (1977-1984), atrazine, alachlor,
metolochlor, 2,4-D, and metribuzin have accounted for 77 percent of the total detections. The frequency of
detection of these pesticides has increased over time. Water from some of the lakes and certain stream segments
with detected pesticides is used for domestic consumption after treatment, and pesticides were found in the
treated drinking water.
Baker et al. (1985) studied the effects of intensive agricultural land use on regional water quality in
northwestern Ohio. Simazine, atrazine, alachlor, metolachlor, linuron, and cyanazine were found in tap water
at Tiffin, Fremont, and Bowling Green, Ohio. Seasonal occurrences of herbicides in stream systems also result
9-3
-------
Johnson
in seasonal pesticide exposures in drinking water derived from river systems. Most soluble pesticides pass
through conventional water treatment plants with very little attenuation (Baker et al., 1985).
Although long-range policy decisions concerning non-point water pollution begin by considering today's
problems, possible future scenarios must also be considered. One set of "possible futures is derived from the
expected impacts of increasing concentrations of CO, and other trace gases (Titus, 1986). The task at hand is
to assess the potential impacts of numerically derived CO0-related climate changes on pesticide fate in the Great
; potential impacts of numerically de:
Plains and Central Prairie.
LITERATURE REVIEW
The impacts of CO2-doubling climate change on hydrologic aspects of agriculture in the Great Plains have
been studied by a number of researchers. Robertson et al. (1987) consider the impacts of climate change on
yields and soil erosion for selected crops in the Southern U.S., Central Prairie and Northern Plains using the
•f * _ .. f^ «M^V • • ^m^^ • . • •• • * *
Imhoff et al., 1983; Knisel, 1980; Aller et al., 1985; and Donigian and Carsel, 1985). These studies focus on the
development and verification of field or watershed models under a variety of geographic, climatological and
agricultural management conditions. Although chemical transport research which explicitly addresses the impacts
of CO2-induced climate change on chemical transport is virtually non-existent, the range of model sensitivity
analyses and verification and their use in "what if type studies provides a logical springboard to climate change
applications. An example of such a preliminary study of these impacts across the southern US. is Cooler
(1987).
PROJECT OVERVIEW
The present study focuses on interactions among agricultural soil/tillage/management systems, pesticide
movement, and the potential impacts of CO2-doubling on subsequent hydrologic quality in the VS. Great Plains.
In this case, the Great Plains is defined as including a northern spring wheat production region, a central winter
wheat production region, a southern cotton production region, and a Central Prairie corn production region.
The purpose of the research to be presented is neither to defend nor support specific general circulation model
results, but to apply the best currently available information to issues of national interest. Although agricultural
water quality issues will be addressed, this document does not represent a toxics study. Statements concerning
toxicity and chemical concentrations are desirable goals, but the case will be presented that the present
state-of-the-art models may be insufficient to support definitive location-specific policy statements.
Weather data from twenty-six National Weather Service (NWS) First Order and cooperative observation
stations were used to represent the climates of the study areas. Stations representative of each region were
selected on the basis of completeness of record for the period 1951-1980. The selected stations and their
proximity to cropping regions are shown in Figure 1. Three generic soil types (all loams) were analyzed at each
weather site; sand, silt, and clay. Specific soils were selected to represent each of these three general types (see
Table 1). Three general tillages were selected to represent agricultural management in each region: conservation,
conventional, and conventional with residue remaining. The conservation tillage management system was taken
to be one in which a chisel plow was used. Descriptions of the complete management systems are provided in
the methodology portion of this report and are summarized in Table 2.
Given these regional production conditions, a process type chemical transport computer model was run for
each station, soil, and tillage combination contained in Figure 1 and Tables 1 and 2. This range of production
scenarios facilitates the examination of pesticide movement in terms of the quantity of chemical that could be
eroded or washed off the soil into surface water supplies, or that could be leached out of the bottom of the soil
profile and thus become available for further vertical or horizontal movement into drinking water supplies.
9-4
-------
Johnson
SPRING WHEAT
COTTON
and
rownsville
Figure 1. Weather stations and cropping areas utilized in the climate change analysis.
9-5
-------
Johir.on
Table 1. Sand, Silt, and Clay Loam Soils Selected for Each Site Analysis
STATION
Watseka, IL
Windsor, IL
Wheatfield, IN
Webster City, IA
Dodge City, KS
Goodland, KS
Wichita, KS
Glasgow, MT
Scottsbluff, NE
Bismarck, ND
Fargo, ND
Minot, ND
Cleveland, OH
Toledo, OH
Oklahoma City, OK
Huron, SD
Pierre, SD
Abilene, TX
Amarillo, TX
Brownsville, TX
Corpus Christi, TX
Midland, TX
San Antonio, TX
Waco, TX
La Crosse, WI
Madison, WI
SANDY
LOAM
Plainfield
Oakville
Granby
Plainfield
Pratt
Otero
Holdrege
Amor
Otero
Amor
Emrick
Hecla
Red Hook
Rocky Ford
Pratt
Embden
Embden
Waldeck
Amarillo
Delfina
Segno
Amarillo
Crockett
Crockett
Plainfield
Coloma
SILTY
LOAM
Flanagan
Flanagan
Flanagan
Clarion
Harney
Keith
Hastings
Williams
Weld
Williams
Fargo
Williams
Spinks
Tappan
Grant
Clarno
Glenham
Miles
St. Paul
Hidalgo
Norwood
Acuff
Venus
Bosque
Fayette
Piano
CLAY
LOAM
Drummer
Drummer
Drummer
Webster
Nib son
Nibson
Garvin
Bowdine
Uly
Parnell
Nutley
Parnell
Sebring
Ellsworth
Garvin
Parnell
Promise
Foard
Pullman
Harlingen
Beaumont
Pullman
Heiden
Houston-Blac
Clyde
Reddick
9-6
-------
Table 2. Continuous cropping tillage//management systems for each crop and
region used as input to PRZM. All chemicals applied at planting.
(T. = surface soil temperature.)
CROP
TILLAGE METHOD
TILLAGE
OPERATION
RESIDUE OR
NO RESIDUE
PLANTING
DATE
PLANTING
CRITERIA
HARVEST
(DAYS FROM PLANT)
Spring Wheat Conservation Chisel Plow No Residue April 1- T > 21°C
May 30
Conventional Noldboard Residue
Conventional holdboard No Residue
135
Winter Wheat Conservation
Conventional
Conventional
Chisel Plow No Residue Sept 15-
Oct 15
Moldboard
Moldboard
Residue
No Residue
Previous 7-
day precipita-
tion > 2.54 cm
Soil moisture
< field
capacity
249 (Texas and
Oklahoma)
267 (Kansas)
293 (Nebraska)
Cotton
Conservation
Chisel Plow No Residue March 1-
June 25
T > 28.3°C
s8il moisture
< field
capacity
165
Conventional
Conventional
Moldboard No Residue
Moldboard Residue
Corn
Conservation
Chisel Plow No Residue April 1-
June 15
Conventional
Conventional
Fall
Moldboard
Spring
Moldboard
No Residue
Residue
T > 21°C
s8il moisture
< field
capacity
165
* This management scenario is not practiced in the cotton study area at present.
scenario was included to examine possible implications of its future use.
The
-------
Johnson
Major research results are presented in terms of direction and magnitude of regional changes in levels of
pesticide losses to surface runoff, erosion, and subsurface leaching. Secondary discussions of within region and
inter-scenario variability are presented to reinforce discussions of model limitations and result interpretation
(Appendices A and B). The remainder of this report describes the modeling methodology applied and its
limitations, presents and interprets the results, and addresses the environmental and sodoeconomic implications
of these results.
THE STUDY REGION
Norum et al. (1957) propose that the most compelling fact of agriculture in the cool, temperate Northern
Great Plains is the irregular and generally deficient rainfall. They conclude that the climate, topography, soil
and native vegetation of the area comprising North Dakota, South Dakota, and the sand hills region of Nebraska
encourage a cropping agriculture broadly devoted to the production of spring wheat. The spring wheat
production area extends from the Red River Valley westward and southward to the range country of Montana
and South Dakota and includes mainly loamy and sandy soils developed on varied glacial deposits. An area of
several counties in north central Montana has soil and topography favorable for production of spring wheat,
although rangeland separates it from the major spring wheat area With precipitation low relative to evaporation
and transpiration, unevenly distributed seasonally, and possessing great interannual variability, small negative
deviations in precipitation are sufficient to initiate a condition of drought. Negative deviations, small and large,
are frequent in these regions (Rosenberg, 1980).
South of the Northern High Plains is a region in which winter wheat production predominates. This area
comprises Kansas, southwestern Nebraska, and the northwestern one-third of Oklahoma (Hobbs, 1957). There
is considerable variation in temperature and precipitation across this region. Annual precipitation ranges from
38 inches in the southeast to 14 inches along the western edge. The frost-free growing season ranges from more
than 210 days in the southeast to fewer than 140 days in the northwest. Many marginal soils exist but they are
generally productive and respond well to good management (Hobbs, 1957). The main problems are the
irregularity of precipitation, the need for supplemental fertilization, and water and wind erosion. Water erosion
may result in sizable soil losses from sloping fields in this area Areas of sandy soils are extensive in
southwestern Kansas, and the Panhandles of Oklahoma and Texas. Many have a low content of clay in the
surface layers and are particularly erosive when they are cultivated.
The southernmost region to be considered comprises the Southern High Plains and Blacklands areas of
Texas. Cotton is a major cash crop, and growing it is a highly developed, mechanized enterprise (Johnston,
1957; USDC, 1985). The topography varies from nearly level to strongly rolling. Surface drainage is well
developed, except in the part of the region that is on the Southern High Plains. The soils range from slightly
acid to calcareous and from deep sands and days to thin soils. These productive soils absorb moisture quickly,
have good water holding ability, and are easily worked. The sandier sections are subject to severe wind erosion.
Extreme variations in seasonal and yearly precipitation make year-to-year production uncertain. Such variations
in precipitation are more pronounced in the western part of the region than in the east. Extreme variations in
temperature sometimes do great damage to crops.
The final study area (Central Prairie) comprises substantial portions of the Corn Belt including northern
Illinois, eastern and north central Iowa, and south central Minnesota Cora is the crop of greatest overall
economic importance in this region. Most of the land is level to gently rolling (Pierre and Riecken, 1957). The
soils are generally medium to fine in texture. They have good structure and hold moisture well. More than 25
percent of the soils of this Central Prairie subregion are of level to nearly level topography and usually require
artificial drainage. Associated with the level soils in many places are soils of moderate slope and good natural
drainage. These soils are apt to erode because of their topography when they are cropped intensively. Excessive
tillage, especially for corn, has been partly responsible for a decline in soil organic matter, a deterioration of tilth,
and an increase in soil erosion (Pierre and Riecken, 1957).
9-8
-------
Johnson
CHAPTER 2
METHODOLOGY
THE EFFECTS MODEL
Development of the Modeling System
The model requirements for the present study include wide geographic applicability, ease of calibration so
that results at many locations and under many initial conditions can be examined rapidly, use of environmental
parameters that are part of the climate change model output, availability of non-climate input parameters, and
inclusion of farm management options so that the role of on-farm response to climate change can be included.
Haan et al. (1982) provide an annotated bibliography of 80 watershed models developed prior to 1978. Of
these, only one explicitly addresses pesticide-soil interactions (Agricultural Runoff Management Model (ARM),
Crawford and Donigian, 1973). ARM simulation performance is compared with the Cornell Pesticide Model
(CPM) in Steenhuis (1979). Other recent watershed models that include chemical transport but were not
summarized by Haan are CREAMS (Knisel, 1980), CADIL (Emerson et al., 1984) and HSPF (Johanson et aL,
1984).
Dean et aL (1984) review a survey of 50 chemical transport models. They consider in detail SEGOL (SegoL,
1976), SUMATRA-1 (van Genchten, 1978X TARGET (Sharma, 1979), and FEMWATER/FEMWASTE (Yen
and Ward, 1980,1981). In addition, the authors consider the PESTAN (Enfield et aL, 1982) and PRZM (Carsel
et al., 1984) models. PRZM was selected over the others for use in the regional Leaching Evaluation of
Agricultural Chemicals (LEACH) study because of its link to the SCS curve number approach for surface
hydrology to supply infiltration volumes, and its ability to be operated quite inexpensively for long simulation
time periods (Dean et al., 1984).
PRZM has also been selected for the present study because of its flexibility, accessibility, and ease of use.
It meets fully the stated selection criteria, including validation under a wide variety of farm management
situations. It is a dynamic, compartmental model which mathematically simulates the movement of water and
its dissolved constituents (pesticides) through the upper soil profile, which includes the active root zone of
growing crops. The model considers parameterized pesticide-soil and pesticide-water interactions. Rainfall is
partitioned into interception (by plants), runoff, and infiltration components based on the Soil Conservation
Service curve number technique. Erosion, crop growth, evapotranspiration, and chemical interactions are all
simulated or parameterized. Figure 2 provides a schematic view of the simulated processes. The model is
described in detail in Carsel et aL (1984).
Field level validation of PRZM has been performed and is documented in Donigian et al. (1986). The
application of results from a field-level model to a region surrounding a weather site assumes that the same
processes that take place on a 1-na field take place on many fields throughout the region. No spatial aggregation
of PRZM results is performed in this research. All PRZM values are expressed as quantity lost/unit area/year.
The careful selection of representative regional input parameters is essential. The critical parameters
supplied to the model for the present application include weather, soil profiles, crop phenology,
tillage/management practices, and pesticide characteristics. Many of these parameters are used by PRZM to
modify rainfall partitioning to runoff and infiltration and, consequently, the distribution and transport of pesticide
within and out of the soil profile.
Climatological input values include mean daily temperature, 24-hour precipitation, and storm duration
interval. PRZM assumes that all precipitation reported for a day is delivered by a single storm. As a result,
9-9
-------
Johnson
Figure 2. Schematic of pesticide root zone model (from Carsel et al., 1984).
9-W
-------
Johnson
large 24-hour precipitation reports result in the simulation of more intense rainfall than smaller daily
observations. The same storm duration is assumed for all storm events. It is further assumed that the majority
of surface losses will be associated with storms which occur near planting, and it is the characteristics of these
storms that are used to define storm intensity.
The selection of specific sand, silt, and clay soils assigned to each weather site was based on the greater soil
groups contained in "USDA/SCS Land Resource Regions and Major Land Resource Areas of the UJS." (1981),
and the short verbal description of each soil contained within the Erosion Productivity Impact Calculator (EPIC)
soils data base (Grassland, Soil, and Water Research Laboratory, Temple, TX). Once the soil is selected, soil
parameters required as input to PRZM for each layer include: depth (cm), bulk density (g/cm3), wilting point
moisture content (cm3 of water/cm3 of soil), field capacity (cm3 of water/cnrof soil), percent organic carbon
content, and the universal soil loss equation K value for the profile. Values for these parameters are provided
by the EPIC soils data base.
Crop phenology is modeled in PRZM by a fixed second-order polynomial function. Because of this,
within-season PRZM phenology estimates are not always representative. When PRZM is linked to process-type
yield models, the estimated phenology is usually replaced by that predicted by the crop model ( Johnson, 1986;
Cooler, 1986). In its original form, PRZM requires specified dates of emergence, maturity, and harvest. PRZM
uses these values to adjust parameters such as the daily runoff curve number (bare soil or full cover), crop
interception, and root depth.
Planting date can affect the entire pesticide system if it is assumed that pesticides are applied at that time.
As soon as the pesticide is applied it can be affected by the environment. If a fixed or unrealistic planting date
is provided to PRZM, the model could produce unrepresentative distributions of pesticide wash-off, runoff, or
leaching. For the present study, simple conditional algorithms are applied within fixed planting windows to
estimate dates of planting. Emergence was set to occur 15 days after the estimated planting date for all crops.
Harvest dates were modified to allow for geographic variability across crop regions with large latitudinal ranges
(Table 2) (Nuttonson, 19SS).
Tillage practices are listed by cropping region in Table 2. Each element of the tillage/management practice
implies some modification to the soil water runoff curve, CN. The PRZM users manual provides appropriate
parameter estimates for each tillage management combination.
The last group of PRZM input parameters are those associated with the pesticides to be monitored. Rather
than attempt to deal with individual pesticides, three classes of pesticide are considered: 1) highly
soluble/long-lived (ex. metolochlor, atrazine); 2) highly soluble/short-lived (ex. metribuzin); and 3) slightly
soluble/long-lived (ex. trifluralin). The organic carbon distribution coefficient K^ is an inverse indicator of
solubility. In this study, high solubility is characterized by K values of less ton 100. Low solubility is
characterized by K values greater than 500. Long residence is Defined as a chemical half-life in soil of about
140 days. Short residence is defined as a chemical half-life in soil of about 14 days. These combinations of
characteristics are most likely to result in undesirable chemicals reaching surface and groundwater systems. For
example, in the case of a highly soluble, long-residence pesticide, high solubility permits the chemical to mix
readily with surrounding water, which would facilitate its further transport beyond the application site in surface
runoff. A long residence time implies a slow rate of chemical degradation. Once the chemical enters the
hydrologic system, it will tend to persist for an extended period of time. This increases the likelihood that
undesirable chemicals could remain intact long enough to reach major surface or groundwater bodies.
PRZM permits a wide range of pesticide application rates and methods. The model pesticides in this study
are applied in field furrows once a year, at planting, at a rate of 1.12 kg/ha (1 Ib/ac), or are surface banded or
banded and incorporated into the soil to a depth of 5 cm. Wauchope (1978) indicates this is a reasonable
application rate, but more than one application may be made per year. This application scenario is most
appropriate for spray or granular preplan! herbicides or soil insecticides used under all but no-till conditions.
Additional simulations are necessary to estimate foliar applied pesticide transport.
9-11
-------
Johnson
Further Limitations of the Model
There are several limitations to the use of the PRZM model in climate change applications. For example,
the PRZM ground cover and interception estimates do not take into account the direct effects of increased
atmospheric CO2 on the efficiency of crop biomass production (Baker et al., 1982). Neither does PRZM
consider the climate change effect on length of growing season (Newman, 1982). These impacts are important
because ground cover (crop/no crop) and leaf area index are used by PRZM to modify the partitioning of rainfall
into runoff and infiltration.
Second, PRZM is a one-dimensional(z) flow model No subsurface lateral flow is computed Although this
can result in an overstatement of recharge and subsequent leaching, long-term trends and qualitative comparisons
should remain consistent with field observations. PRZM does not model macro-pore transport. Transport below
the root zone can be estimated but has not been verified and so was not used in the present study.
Although the PRZM model has been validated under many environmental conditions and geographic
locations, there are certain limitations to its hydrologic budget, particularly in the estimation of recharge. The
model performs best in well-drained soil conditions typical of most large-scale UJS. commercial agriculture. It
does not perform well under poorly drained or ponding conditions such as those favorable to rice or cranberry
production.
Finally, the rate of pesticide decay is directly proportional to temperature. The higher the temperature, the
more rapid the decay. PRZM assumes a constant rate of chemical decay. Under a climate change scenario
which includes atmospheric warming and, consequently, increased soil temperatures, this could result in
overestimates of chemical persistence.
In summary, the most serious limitations to the use of PRZM in the present study are: 1) the need for
careful selection of cropping system parameters and the sensitivity of the model to these parameters; 2) the lack
of CO- doubling impacts on biomass for ground cover; 3) the lack of subsurface horizontal or vadose zone
transport; 4) the lack of chemical decay rate response to temperature; and 5) the care with which regional results
are prepared, interpreted, and applied when based on a point estimate model.
THE CLIMATE CHANGE SCENARIOS
Scenarios Used
A 30-year time-series of weather data was supplied for each study site in Figure 1. Daily values of historical
temperature and precipitation for the period 1951-1980 are used to represent the "base" weather scenario
(McDonald et al., 1983). These data are 24-hour summaries of daily values for stations contained in the NWS
First Order and Cooperative Meteorological Observation Networks. Although generally complete, these station
data often contain missing observations. A statistical interpolation in space and time using nearby stations has
been performed to replace missing observations with statistically representative values (Eddy, 1985). Daily data
were then modified according to estimated monthly temperature and precipitation changes based on 2xCO,
results from each of two global climate models (GCM), GISS (Hanson et al., 1986) and GFDL (Manabe and
Weatherald, 1980). This was done by multiplying daily precipitation by a monthly precipitation factor, or by
adding a monthly daily temperature change to daily mean temperature values. GCM model output represents
the impacts of CO, doubling on weather variables at the center of 8° lat x 10° Ion grid boxes for GISS and 45°
lat x 7.5° Ion grid boxes for GFDL. No spatial interpolation of the grid box centroid values was performed.
Model results are presented and discussed in terms of the base (historical weather) scenario and 2xCO2-induced
changes from this base.
9-12
-------
Johnson
Issues Resulting from *he Climate Scenarjn
One issue not easily resolved is the use of results generated by models of differing scale. In the case of the
GCM's, point manifestations (weather sites) are inferred from a regional (grid box) generalization. In the case
of PRZM, regional trends are inferred from a point model In the present study, it is assumed that the GCMs
are addressing large-scale changes in weather parameters. Smaller scale (within region) variability is assumed
to remain unchanged from the historical record. Although temperature change limitations are of some concern
to PRZM, limitations related to the representation of precipitation by the GCM's are far greater.
Sufficient GCM data are not available to ascertain 2xCO2-induced changes in storm frequency or duration
resulting from changes in the rainfall delivery system from convective showers to warm frontal rain, or vice versa
Since runoff and erosion are single-storm event processes, this is a serious shortcoming of the induced climate
change scenario and limits the time scale appropriate for analysis of results to one or more 24-hour periods
rather than to a rainfall event.
The present GCM output requires that the time sequences of the occurrence of precipitation are left
unchanged from those of the base scenario data set. Under these conditions, estimates of total annual erosion
could easily underestimate the magnitude of the problem in any given year. On the other hand, since a single
percentage adjustment is applied to all 24-hour precipitation events within a month, the magnitude of extreme
rainfall events is exaggerated. This exaggeration can result in changes in the correlation of temperature and
precipitation Gelds and could result in an overestimation of surface and subsurface pesticide losses in an
individual year. Hence, the presentation and interpretation of model-generated results must be pursued with
caution.
RESULTS
REGIONAL RESULTS AND INTERPRETATION
Major research findings are presented over a broad geographic scale in qualitative terms. In the absence
of an objective means of assessing the validity of GCM environmental projections, consensus impact results may
suggest those physical processes or geographic areas that are relatively insensitive to the precise manner of
temperature or precipitation estimation. For instance, if both GCM model estimates result in declines of
regional pesticide leaching, the implication is that the "signal" of a leaching decrease is stronger than the "noise"
created by the application of weather changes estimated under differing GCM assumptions and methods of
computation. Consensus directional results are, then, statistically robust. Table 3 summarizes the consensus
of direction of change in median values over the base between GISS and GFDL climate change scenarios.
Blanks in Table 3 indicate no consensus.
The model output contained in Table 3 indicates that the consensus is that surface pesticide erosion and
subsurface ^rhing could decrease in the northern spring-wheat region of the Great Plains. Central Plains
surface changes are scattered, but model output indicates that short-lived pesticide leaching could increase and
long-lived pesticide leaching could decrease. Surface runoff and erosion of all pesticide types on all soil types
could increase in the southern Great Plains. Short-lived pesticide leaching on silt soils is modeled as increasing
and longer lived pesticide leaching on day soils is modeled as decreasing in this region. All pesticide transport
(surface and subsurface) could be decreased in the Central Prairie com region.
When further regional studies of the scenarios listed in Table 3 were performed, it was found that
distributions of certain model predicted changes are significantly different (statistically) from the base case
distributions as well as possessing medians for which there is directional consensus across the models (Table 4).
In several cases, although there is consensus as to direction of change, the results do not appear to be statistically
significant. This is largely the result of the statistical nature of the model results. Model-estimated values of
pesticide transport are highly variable and their frequency distributions are skewed towards zero; that is, there
are many more small values than large ones. A discussion of these characteristics is provided in Appendix A,
9-13
-------
Johnson
Table 3. Summary of GCM Model Consensus of PRZM Pesticide Transport by
Cropping Region, Pesticide, and Soil (+ indicates Median Values
Increase Under Climate Change; - Indicates Median Values Decrease
Under Climate Change, Blank Indicates No Consensus Among Median
Values)
SURFACE SURFACE
PESTICIDE PESTICIDE PESTICIDE
RUNOFF LOSSES EROSION LOSSES LEACHING
SPRING WHEAT:
Region(all soils, pesticide,
tillage scenarios) +
*Highly Soluble/Short-Lived + +
Highly Soluble/Long-Lived +
Slightly Soluble/Long-Lived
**Clay +
Sand
Silt
WINTER WHEAT:
Region(all soil, pesticide,
tillage scenarios)
Highly Soluble/Short-Lived + +
Highly Soluble/Long-Lived +
Slightly Soluble/Long-Lived
Clay +
Sand +
Silt
COTTON:
Region(all soil, pesticide,
tillage scenarios)
Highly Soluble/Short-Lived + +
Highly Soluble/Long Lived + +
Slightly Soluble/Long-Lived + +
Clay + +
Sand +
Silt + + +
CORN:
Region(all soil, pesticide,
tillage scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
* ex. Median value of all tillage, soil, weather site, scenarios for
highly soluble/short-lived pesticides in the spring wheat crop area.
** ex. Median value of all tillage, pesticide, weather site scenarios
for clay soils in the spring wheat crop area.
9-14
-------
Johnson
Table 4. Statistically Significant* Climate Change Impacts
on Pesticide Transport Across the USA Great
Plains and Central Prairie
Spring Wheat:
Winter Wheat:
Cotton:
Corn:
Under GFDL estimated temperature and precipitation
conditions, runoff and leaching of relatively
insoluble long-lived pesticides could decrease
as could leaching of highly soluble long-lived
pesticides. There are no statistically
significant changes under GISS climate change.
Under GISS estimated temperature and precipitation
conditions, runoff and erosion losses of all
pesticide types studied on all soils could
increase. There are no statistically significant
changes under GFDL climate change.
Under GISS estimated temperature and precipitation
conditions, regional runoff could increase.
Runoff of highly soluble short-lived, and
relatively insoluble long-lived pesticides
could increase. Surface erosion of highly
soluble short-lived pesticides could increase.
Under GFDL estimated temperature and precipi-
tation conditions, surface runoff of relatively
insoluble long-lived pesticides could increase.
Under GFDL estimated temperature and precipitation
conditions, regional surface runoff and subsurface
leaching of the chemicals studied declined. Sur-
face runoff of highly soluble long-lived chemicals
declines as does leaching of highly soluble short-
lived chemicals. There are no statistically
significant changes under GISS climate change.
* The null hypothesis is that there is no difference between the
population of pesticide transport results under historical or
climate change conditions. A result is declared statistically
significant if there is less than a 10% chance that a signifi-
cant difference has been detected when there really is no
significant difference.
9-15
-------
Johnson
but the important implications are: 1) statistics such as mean and standard deviation are not good descriptive
measures for pesticide transport; 2) any analysis for policy implications should consider extremes and spread of
possible outcomes; and 3) careful stratification of results and use of non-parametric tests may prove helpful in
identifying statistically significant double CO2 impact "signals" from background "noise" generated by different
fields, soils, tillages, pesticides, and GCM constraints.
A final result stems from a more detailed regional analysis provided in Appendix B and summarized by
Figure 3. Each transport variable analyzed for this study is presented as a percentage of the base or historical
case (base = 100%) and is a function of GCM weather scenario and region. For pesticide lost through runoff
transport, the most striking impact is the 100% increase over base conditions under GISS climate change
conditions in the winter wheat growing season. For pesticide erosion which is, in general, at least one order of
magnitude less than pesticide lost through runoff, large increases are noted under the GISS scenario in the
winter wheat and cotton growing regions. For pesticide leaching, very large declines are noted under GISS in
the spring wheat area, and under GFDL in the corn area. Moderate declines are noted for winter wheat and
cotton areas under the GFDL.
A detailed analysis of the temperature and precipitation sources of the changes noted in Tables 3 and 4 and
Figure 3 are discussed in some detail in Appendix B. These results generally stem from higher daily
temperatures and changes in the amount and timing of precipitation falling on the simulated field. Both GCM's
indicate significant temperature increases under double CO^, but disagree on the pattern of precipitation
delivery. This, in large part, accounts for the blank entries indicating no pesticide transport consensus in Table
3. Increased annual pesticide lost in runoff and erosion stem from precipitation increases at planting time,
especially in the winter wheat area under the GISS scenario conditions, or increased precipitation during fallow
periods.
Areas in which runoff and erosion decline suggest that precipitation changes (increases) were not adequate
to overcome increased moisture demands from wanner temperatures or that the large monthly precipitation
increases occur at times less favorable for significant pesticide losses. Leaching declines can result from either
temperature or precipitation changes. Given constant frequencies, precipitation increases over base conditions
can result in higher PRZM estimated soil and adhered pesticide losses via the Modified Universal Soil Loss
Equation (MUSLE). If higher than base case soil moistures result from modeled precipitation increases, PRZM
modifies the runoff curve number (CN) to partition a larger percent of subsequent rainfall to runoff. This
generates greater subsequent losses of pesticides carried by surface runoff. These two processes result in less
pesticide remaining to be leached downward through the soil profile. A second contributing factor is higher
than base case daytime temperatures which result in greater evaporative demand from the soil. Less moisture
is then available in the profile to transport pesticides downward. This can result in less pesticide being leached
below the root zone and a larger proportion of the pesticide being stranded in the profile to eventually decay
into inactive constituents. Even greater leaching declines are anticipated if the effect of higher soil temperatures
on rate of chemical decay had been considered. In most cases, decay rates increase as temperatures rise. Since
the chemical decay rate is a constant in PRZM, the model may overestimate the quantity of leachable pesticide
in the profile under CO2 warming conditions. It is most likely that the regional impacts summarized in Figure
3 result from a combination of temperature increases and precipitation intensity changes. Further sensitivity
analyses could establish the point at which one factor (temperature or precipitation) overwhelms the other.
9-16
-------
°flZM Modeled Impacts of 2XC02 on
Pesticide Runoff
X of Base toe.
Johnson
crop region
PRZM Modeled Inacts of 2XC02 on
Pesticide SeoiBent Load
COIN
crop region
PRZM Modeled Inoacts of 2XC02 on
pesticide Leaching
crop region
Figure 3. Regional summary of surface and subsurface pesticide loss as a percentage of the base
weather scenario losses (base = 100%, SW = Spring Wheat, WW • Winter Wheat, COT
Cotton, CORN » com).
9-17
-------
Johnson
CHAPTERS
IMPLICATIONS OF RESULTS
Figures 4 and 5 represent summaries of PRZM results under observed, historical weather for soils which have
the greatest chemical erosion potential and the greatest leaching potential. Soil erodability is based on Soil
Conservation Service (SCS) class/sub-class designation. Under this system, a designation of 2E indicates a
slightly erodable soil. 6E indicates a severely erodable soil. The soils selected for this study had SCS
designations of 4E, 5E, or 6E. Only scenarios in Table 2 that include conservation tillage practices were selected
for the maximum erosion and leaching studies. This assumes that the farmer realizes there may be erosion
problems on his property and is attempting to minimize his soil losses. The specific sandy type EPIC soil
selected at a particular analysis site was assumed to possess the greatest leaching potential.
Figures 4a-c contain the areas of greatest model-indicated 30-year median pesticide surface losses under the
base weather scenario. Figure 4a highlights those areas in which surface losses of highly soluble short-lived
pesticides are the greatest. This area comprises the Texas cotton croppuig areas and Oklahoma. Median annual
chemical losses (loadings) at sites within this area range from 2.0 x l(r to 9.0 x 10 mg of pesticide lost/ha/yr.
Maximum storm event concentrations can be calculated from PRZM output parameters. The distribution of
maximum storm event concentrations for each of the 30 base scenario years was computed at Corpus Christi,
Texas. These concentration values are compared to sample experimental site observations found in Wauchope
(1978) and are presented in Table 5. Although the experimental site values contained in Table 5 fall within the
range of values obtained from PRZM simulations for Corpus Christi, the simulated concentrations tend to be
high. Wauchope (1978) comments, "Pesticide concentrations in runoff may vaiy by an order of magnitude or
more during a single runoff event, and even event averages for a given chemical are almost unique for each
situation, depending on rate of application, storm intensity and timing, field site, etc" Given these comments,
the PRZM limitations in terms of precipitation event modeling on a time scale less than 1 day, variations between
specific chemicals within the generic group of more than an order of magnitude, and uncertainty in estimates of
pesticide properties such as half-life, the PRZM model outcome is remarkably good.
Figure 4b highlights those areas in which surface losses of highly soluble/long-lived pesticides were the
greatest. This area comprises the Texas cotton areas, Ohio, western Kansas, South Dakota, and southeastern
North Dakota. Median annual chemical losses at sites within this area range from 3.0 x 104 to 2.0 x 105 mg of
pesticide lost/ha/yr. Representative PRZM- estimated pesticide runoff concentration and percent pesticide losses
at Cleveland, Ohio, were compared to experimental data (Table 5).
Figure 4c highlights those areas in which surface losses of slightly soluble/long-lived pesticides were greatest.
The area comprises Texas, Oklahoma and Ohio. Median loading values at the sites within this area range from
2.0 x 104 to 1.8 x 10° mg of pesticide lost/ha/year. Corpus Christi, Texas was again used as a comparison site
for experimental runoff concentration and percentage loss results. The Table 5 comparison of this case was the
poorest of the three generic groups addressed. Future studies might include a more careful examination of the
chemical parameters chosed to represent this generic group and consideration of a range of experimental results
rather than a single case.
Figures 5a-b illustrate the areas of PRZM-estimated maximum median annual pesticide leaching under base
scenario climatologkal conditions. Figure 5a summarizes areas of maximum leaching of highly
soluble/short-lived as well as slightly soluble/long-lived pesticides. Although areas of maximum leaching are the
same for these pesticide types, the quantity of slightly soluble/long-lived pesticide leached is approximately twice
that of highly soluble/short-lived pesticide at some locations. Maximum median quantities of highly soluble
short-lived pesticide range from 3.7 x 103 mg/ha/yr to 3.0 x 10* mg/ha/year. Maximum median quantities of
slightly soluble/long-lived pesticides range from 335 mg/ha/yr to 4.4 x 104 mg/ha/year.
9-18
-------
Figure Aa.
PRZM modeled areas of
maximum potential 30-year
median annual pesticide
runoff and erosion for
highly soluble/short-lived
pesticides under historical
(base) weather conditions
(runoff + erosion >
20,000 mg/ha/yr).
Figure 4b. PRZM modeled areas of
maximum potential 30-year
median annual pesticide
runoff and erosion for
highly soluble/long-lived
pesticides under historical
(base) weather conditions
(runoff+erosion >
30,000 mg/ha/yr).
Figure 4c. PRZM modeled areas of
maximum potential 30-year
median annual pesticide
runoff and erosion for
slightly soluble/long-lived
pesticides under historical
(base) weather conditions
(runoff + erosion >
20,000 mg/ha/yr). v,
-------
Figure 5a. PRZM modeled areas of maximum potential 30-year
median annual subsurface leaching for highly
soluble/short-lived and slightly soluble/
long-lived pesticides under historical (base)
2 weather conditions (leaching > 100 mg/ha/yr).
Figure 5b. PRZM modeled areas of maximum potential !
30-year median annual subsurface leaching
for highly soluble/long-lived pesticides 1
under historical (base) weather conditions!
(leaching > 20,000 mg/ha/yr).
-------
Table 5. Comparison of experimental and PRZM simulated surface pesticide losses.
s
Experiment
PRZM - BASE
- GISS
- CFDL
Experiment
PRZH - BASE
- GISS
- CFDL
Experiment
PRZH - BASE
- GISS
- CFDL
Chemical
Hetrlbuzln
Highly Soluble
Short -Lived
Atrazlne
Highly Soluble
Long -Lived
Trlfluralln
Slightly
Soluble
Long -Lived
Location
Stonevllle,
MS
Corpus Christ 1,
TX
Coshocton,
OH
Cleveland,
OH
Levlsberg,
NC
Corpus Christ!.
TX
Soil
Texture
Sandy
loam
Sandy
loam
HA
Sandy
loam
Sandy
loam
Sandy
loam
Cover
Crop
Soybeans
Cotton
Corn
Corn
Cotton
Cotton
Length of
Experiment
1-5 months
30 years
30 years
30 yeara
3 months
30 years
30 years
30 years
6-7 months
30 years
30 years
30 years
Application Concentration
Rate Estimated
kg/ha (pg/llter) Reference
.56 10-53 Wauchope, et al.
(1977)
1.12 2-1529
32-1436
2-1485
1.12 180 Edwards (1972)
1.12 221-973
355-1017
317-1018
1.12 7-24 Sheets, et al.
(1972)
1.12 158-488
299-475
191-474
-------
Johnson
The mass of slightly soluble/long-lived or highly soluble/short-lived pesticide leached is compared to that
which was initially applied in Table 6. Table 6 indicates that the median percent of applied highly
soluble/short-lived pesticide lost per year under base conditions ranges from .02% to 2.0%. Under the GISS
climate scenario, the range is unchanged. Under the GFDL scenario, the range decreases dramatically. The
median percent of less soluble/long-lived pesticide lost per year under base conditions ranges from .03% to 3.9%.
Neither climate change scenario results in significant change in the mass of this pesticide lost per year. The
values in Table 6 represent the mass of pesticide moving out of the bottom of the crop root zone. No downward
transport beyond the root zone is included in the present study, so no statement can be made concerning when
(rate of movement) or if (degradation rate) these leached pesticides will actually reach an aquifer. Areas which
the model indicates experience the greatest leaching of highly soluble/short-lived and slightly soluble/long-lived
pesticides on sandy soils under base conditions include the Texas Panhandle, Oklahoma, southwestern Kansas,
Iowa, Illinois, and Wisconsin.
Figure 5b illustrates maximum leaching areas for highly soluble/long-lived pesticides. The areas of maximum
model estimated leaching are the same as for highly soluble/short-lived and slightly soluble/long-lived pesticides,
but also include portions of the Texas Blacklands and Ohio. Mean annual losses of highly soluble/long-lived
pesticide under base weather conditions range from 2.0 x 104 mg/ha/yr to 3.0 x 108 mg/ha/yr. Table 6 indicates
that this represents an annual loss of 332% to 27% of the initial pesticide application. Both GISS and GFDL
climate change scenarios lead to a downward shift of the range of mass losses, but the significance of the shift
is undetermined. The results presented in Figures 5a-b and Table 6 represent worst case scenarios and do not
necessarily reflect the mixture of erodable and non-erodable soils used in current agricultural production.
SURFACE WATER IMPLICATIONS
If the consensus results of Table 3 are applied to Figures 4a-c, the implications are that those areas the
model indicates to be most likely to experience surface pesticide load problems under current weather conditions
in the spring wheat and corn regions could improve (decrease) under the CO2 doubling. Although chemical
load may decline, it does not necessarily follow that concentrations will decline. Table 5 summarizes base and
climate change maximum concentration events at selected locations. The same procedure used to determine the
statistical significance of changes in chemical loadings was applied to the concentration distributions. Results
indicate that at the two specific locations presented in Table 5, there is no significant change in the distribution
of maximum concentration events. Thus, even though the chemical runoff load increases significantly under the
GFDL scenario across the corn region, the concentration of the chemical entering an adjacent water body from
a point within the region may not necessarily increase as well.
If a pesticide is removed from a field primarily in the sediment phase of runoff, then its loss might be
controlled by standard soil conservation practices. Unfortunately, many of today's pesticides are lost mainly in
the water phase (Wauchope, 1978; EPA, 1984; Maas et al., 1985). Erosion control practices, unless they control
water as well as sediment losses, can be expected to have little effect on runoff losses of pesticides, except those
having solubilities of 1 ppm or less (Wauchope, 1978; Maas et al., 1985). This implies that soil conservation
technology that emphasizes sediment control could help to offset the surface loss increases indicated by PRZM
for slightly soluble/long-lived pesticides (solubility = .05 ppm). Such measures are not expected to help surface
losses of highly soluble/short-lived pesticides (solubility = 1,220 ppm) or highly soluble/long-lived pesticides
(solubility = 33 ppm), unless runoff control measures are also implemented.
GROUNDWATER IMPLICATIONS
The consensus results of Table 3 may be applied to Figures 5a and b to assess the implications to
groundwater. Leaching of pesticides below the crop root zone should decrease everywhere except perhaps on
silty soils in the cotton region. This overall decline, as discussed earlier, most likely is a result of higher
evaporative demands and increased runoff volumes, resulting in less moisture available for infiltration and deep
9-22
-------
Johnson
Table 6. Median Annual Leaching Losses Through Sandy Loam Soils in the Maximum Loss Areas of
Figures 5a and b Expressed as a Percent of Annual Pesticide Application (1.12 kg/ha)
SCENARIO
PESTICIDE BASE 6ISS GFDL
Highly Soluble
Short-Lived .02% to 2.00% .01% to 2.9% .01% to .77%
Highly Soluble
Long-Lived 3.32% to 27.00% .02% to 21.60% .02% to 23.90%
Less Soluble
Long-Lived .03% to 3.90% .02% to 3.10% .10% to 3.00%
9-23
-------
Johnson
percolation. However, as in the case of surface losses, decreased loads do not necessarily imply decreased
concentrations of pesticide transported beyond the root zone.
From a water quality perspective, decreased pesticide leaching may be advantageous. From a water quantity
perspective, these results could be cause for concern. Less leaching can imply less water moving through soil
profiles and less water available for aquifer recharge. If water demands were to increase at the same time there
was a decrease in the recharge rate, competition for scarce water resources would increase dramatically.
A second implication, particularly for rapidly recharging aquifers, arises from the management response to
increased surface pesticide losses suggested earlier. Runoff reducing conservation practices will generally reduce
surface pesticide loadings more than surface concentrations. While such practices may increase infiltration, more
pesticide also could remain to be transported downward and pesticide concentrations in infiltration water could
increase (Maas et al., 1985). Situations could arise in which decisions concerning which resource to protect,
surface or groundwater would be required. Given the PRZM chemical load estimates and recent estimates of
county-scale groundwater vulnerability (Alexander et al., 1986), those study areas most likely to be facing such
decisions are the Cotton and Winter Wheat cropping regions.
9-24
-------
Johnson
REFERENCES
Alexander, W. J., S. K. Liddle, R. E. Mason and W. B. Yeager. "Ground-Water Vulnerability Assessment in
Support of the first stage of the National Pesticide Survey." EPA Contract No. 68-01-6646. Research Triangle
Institute, 1986. 22 pp.
Aller, L., T. Bennett, J. H. Lehr, and R. J. Petty. "DRASTIC: A Standardized System for Evaluating Ground
Water Pollution Potential Using Hydrogeologjc Settings." Robert S. Kerr, USEPA Report EPA/600/2-85/018,
US. Environmental Protection Agency, Washington, D.C., 1985. 163 pp.
Association of State and Interstate Water Pollution Control Administrators. "America's Clean Water: The
States' Nonpoint Source Assessment 1985, Appendix." ASIWPCA, Washington, D.C., 1985.
Association of State and Interstate Water Pollution Control Administrators. "America's Clean Water: The
States' Evaluation of Progress 1972-1982, Appendix." ASIWPCA, Washington, D.C., 1984.
Baker, D. Nn L. H. Allen, Jr., and S. R. Lambert. "Effects of Increased CO2 on Photosynthesis and Agricultural
Productivity." Environmental and Societal Consequences of a Possible CO2 Induced Climate Change,
DOE/EV/1009-6, Volume 2 of 2, NTIS, Springfield, Virginia, 1981 104 pp.
Baker, D. B., K. A. Krieger, R. P. Richards, and J. W. Dramer. "Effects of Intensive Agricultural Land Use on
Regional Water Quality in Northwestern Ohio." In: Perspectives on Nonpoint Source Pollution. Proceedings of
a National Conference. Kansas City, Missouri, May 19-22, 1985, U.S Environmental Protection Agency,
Washington, D.C., 1985. pp. 201-207.
Butler, M.K., and J. A. Arruda. "Pesticide Monitoring in Kansas Surface Waters: 1973-1984." In: Perspectives
on Nonpoint Source Pollution. Proceedings of a National Conference. Kansas City, Missouri, May 19-22,1985,
U.S. Environmental Protection Agency, Washington, D.C.,1985. pp. 196-200.
Carsel, R. F., C. N. Smith, L A. Mulkey, J. D. Dean, and P. Jowise. "Users Manual for the Pesticide Root Zone
Model: PRZM." USEPA/ERL Report, EPA-600-3-84-109, Washington, D.C., 1984. 216pp.
Cooler, E. J. The Use of Plant Process Models to Assess Potential Impacts of Regional Pesticide Policies on
Corn and Soybean Production." Agricultural Pesticide Policy Analysis for the Upper Mississippi Basin, Report
4, Oklahoma Climatological Survey, Norman, Oklahoma, 1986. 150 pp.
Cooler, E. J. "The Impact of Climate Change on Cora Production in the Southern United States." In:
Proceedings of the Symposium on Climate Change in the Southern United States: Future Impacts and Present
Policy Issues. M. Meo, Ed., University of Oklahoma, Science and Public Policy Program, Norman, Oklahoma,
1987. pp. 89-134.
Crawford, N. H., and A. S. Donigian, Jr. "Pesticide Transport and Runoff Model for Agricultural Lands." Office
of Research and Development, USEPA, EPA-660/2-74-013, Washington, D.C., 1973. 211 pp.
Dean, J. D., P. P. Jowise, and A. S. Donigian, Jr. "Leaching Evaluation of Agricultural Chemicals (LEACH)
Handbook."EPA-600/3-84-068, U.S. Environmental Protection Agency,
Athens, Georgia, 1984. 407 pp.
Donigian, A. S., and R. F. Carsel. "Impact of Conservation Tillage on Pesticide Concentrations in Ground Water,
Surface Water and Air". Temple, Barker, and Sloan, Inc., Final Report to USEPA on Contract No. 68-01-7002,
Work Assignment No. 600, VS. Environmental Protection Agency, Washington, D.C., 1985. 79pp.
9-25
-------
Johnson
Donigian, A. S., D. W. Meier, and P. O. Jowise. "Stream Transport and Agricultural Runoff of Pesticides for
Exposure Assessment: A Methodology." USEPA/ERL Report EPA/600/3- 86/Olla, VS. Environmental
Protection Agency, Washington, D.C., 1986. 109pp.
Eddy, G A. "Interpolation, Quality Assurance and Archiving of Climate Data from Selected Foreign Countries."
Final Report (Supplementary) on NOAA Contract NA 8344-H-00019,1985. pp 9.
Eddy, G. A., and R. J. Sladewski. "Some Biophysical Implications of Changes in the Water Delivery System of
the Atmosphere." IWRA World Congress on Water Resources, Vol. II Ottawa, Ontario, Canada, 1988. 13pp.
Edwards, W. M. "Agricultural Chemical Pollution as Affected by Reduced Tillage Systems." In: Proc.
No-Tillage Systems Symposium. Ohio State University, Columbus, Ohio, 1972. pp. 30-40
Emerson, C. J., B. Thomas, Jr., and R. J. Luxmoore. "CADIL: Model Documentation for Chemical Absorption
and Degradation in Land." Environmental Sciences Division, Pub. No. 2246, ORNL/TM-8972, Oak Ridge
National Laboratory, Oak Ridge, Tennessee, 1984. 127 pp.
Enfield, G. G., R. F. Carsel, S. E. Cohen, T. Phan and M. F. Walters. "Approximating Pollutant Transport to
Ground Waters." Ground Water 20(6): 711,1982.
Environmental Protection Agency. "Report to Congress: Nonpoint Source Pollution in the UJS." VS.
Environmental Protection Agency, Washington, D.C., 1984. 82 pp.
Haan,C. T., H. P. Johnson, and D. L. Brakensiek. "Hydrologic Modeling of Small Watersheds". In ASAE
Monograph. No. 5. American Society of Agricultural Engineers, St. Joseph, Michigan, 1982. pp. 504-522.
Hansen, J., A. Lacis, D. Rind, G. Russel, I. Fung, P. Ashcraft, S. Lebedeff, R. Ruedy, and P. Stone. The
Greenhouse Effects of Changes in Stratospheric Ozone and Global Climate." In: Effects of Changes in
Stratospheric Ozone and Global Climate. Volume I: Overview. USEP/UNEP, Washington, D.C., 1986. pp
199-218.
Hobbs, J. A. "The Winter Wheat Grazing Region". In: Soils: The 1957 Yearbook of Agriculture. GPO,
Washington, D.C., 1957. pp. 505-515.
Imhoff, J. C., B. R. Bicknell, and A. S. Donigian. "Preliminary Application of HSPF to the Iowa River Basin to
Model Water Quality and the Effects of Agricultural Best Management Practices." Anderson-Nichols & Co.,
Report to USEPA/ERL on Contract No. 68-03-2895, VS. Environmental Protection Agency, Washington, D.C.,
1983. 108pp.
Johanson, R. G, J. C. Imhoff, J. L. Kittle, Jr., and A. S. Donigian, Jr. "Hydrological Simulation
Program-FORTRAN (HSPF): Users Manual for Release 8.0." EPA-600/3-84-066, VS. Evironmental
Protection Agency, Athens, Georgia, 1984. 767 pp.
Johnson, H. L. The Surface Water Module." Agricultural Pesticide Policy Analysis for the Upper Mississippi
Basin, Report No. 6, Oklahoma Climatological Survey, Norman, Oklahoma, 1986. 73 pp.
Johnston, J. R. The Southern Plains". In: Soils: The 1957 Yearbook of Agriculture. GPO. Washington. D.C..
1957. pp. 516-523.
Knisel, W. G., ed "CREAMS: A Field-Scale Model for Chemicals, Runoff, and Erosion from Agricultural
Management Systems." USDA, Conservation Research Report No. 26, Washington, D.C., 1980. 640 pp.
9-26
-------
Johnson
Maas, R. P., M. D. Smolen, S. A. Dressing, C A. Janueson, and J. Spooner. "Practical Guidelines for Selecting
Critical Areas for Controlling Nonpoint Source Pesticide Contamination of Aquatic Systems.* In: Perspectives
on Nonooint Source Pollution. Proceedings of a National Conference. Kansas City, Missouri, May 19-22, 1985,
VS. Environmental Protection Agency, Washington, D.C, 1985. pp. 363-367.
McDonald, P. A., E. J. Cooter, and A. Eddy. "Develop a Daily Climatological Surface Data Base for Multiple
Users." Oklahoma Climatological Survey, Norman, Oklahoma, 1983. 50pp.
Manabe, S., and R.T. Wetherald. "On the Distribution of Climate Change Resulting from an Increase in CO2
Content of the Atmosphere." Journal of the Atmospheric Sciences. 37: 99-118,1980.
Mather, J. R., and J. Feddema. "Hydrologjc Consequences of Increases in Trace Gases and CO2 in the
Atmosphere." In: Effects of Changes in Stratospheric Ozone and Global Climate. Volume 3: Climate Change.
S. G. Titus, ed., VS. Environmental Protection Agency, Washington, D.C., 1986. pp. 251-274.
Newman, J. E. "Impacts of Rising Atmospheric Carbon Dioxide Levels on Agricultural Water Use Efficiencies."
Environmental and Societal Consequences of a Possible CO2- Induced Climate Change, Volume 11, Part 8,
DOE/EV/10019-8, Volume 2 of 2, NTIS, Springfield, Virginia, 1982. 29 pp.
Norum, E. B., B. A. Krantz, and H. J. Haas. The Northern Great Plains". In: Soils: The 1957 Yearbook of
Agriculture. Washington, D.C., 1957. pp. 494-504.
Nuttonson, M. Y. Wheat-Climate Relationships and the Use of Phenology in Ascertaining the Thermal and
Photo-Thermal Requirements of Wheat. American Institute of Crop Ecology, Washington, D.C., 1955. 388 pp.
Pierre, W. H., and F. F. Riecken. "The Midland Feed Region". In: Soils: The 1957 Yearbook of Agriculture.
Washington, D.C., 1957. pp. 535-546.
Robertson, T., V. W. Benson, J. R. Williams, J. R. Kiniry, and C. A. Jones. "Impacts of Climate Change on
Yields and Erosion for Selected Crops in the Southern UJS." In: Proceedings of the Symposium on Climate
Change in the Southern United States: Future Impacts and Present Policy Issues. M. Meo, ed., Science and
Public Policy Program, University of Oklahoma, Norman, Oklahoma, 1987. pp. 73-88.
Rosenberg, N. J, ed. Drought in the Great Plains: Research on Impacts and Strategies. Water Resources
Publications, Littleton, Colorado, 1980. pp. 1-6.
SAS Institute, Inc. User's Guide: Statistics. SAS Institute, Inc., Cory, North Carolina, 1982. pp. 479-484.
Segol, G. "A Three-Dimensional Galerkin Finite-Element Model for the Analysis of Contaminant Transport in
Variably Saturated Porous Media, User's Guide." Dept. of Earth Sciences, University of Waterloo, Waterloo,
Ontario, Canada. 1976.
Sharma, D. "A Comprehensive Mathematical Model Capable of Predicting Chemical-Species Transport in
Porous Media." Rep. TN-DN-42, Dames and Moore, Advanced Technology Group, 1979.
Sheets, T. J., J. R. Bradley, Jr., and M. D. Jackson. "Contamination of Surface and Ground Water with
Pesticides Applied to Cotton." Report No. 60, University of North Carolina Water Resource Res. lost., Chapel
Hill, NC, 1972. 63 pp.
Steenhuis, T. S. "Simulation of the Action of Soil and Water Conservation Practices in Controlling Pesticides."
In: Effectiveness of Soil and Water Conservation Practices for Pollution Control. D. A. Haith and R. C. Loehr,
eds. EPA- 600/3-79-106, U.S. Environmental Protection Agency, Athens, Georgia, 1979. pp. 106-146.
9-27
-------
Johnson
Titus, J. G., ed. Effects of Changes in Stratospheric Ozone and Global Climate. Volume I: Overview.
USEPA/UNEP, Washington, D.C., 1986. 379 pp.
USDA/SCS. Land Resource Regions and Major Land Resource Areas of the United States. Agricultural
Handbook 296, GPO, Washington, D.C., 1981. 156 pp.
USDC. 1982 Census of Agriculture. Volume 2. Subject Series. Part 3. Ranking of States and Counties. United
States Department of Commerce, 1985. pp. 62-63.
van Genuchten, M. Th. "Mass Transport in Saturated Unsaturated Media; One-Dimensional Solutions."
Research Report 78-WR-11, Water Resources Program, Princeton University, Princeton, New Jersey, 1978.
Wauchope, R. D. "The Pesticide Content of Surface Water Draining from Agricultural Fields - A Review." J_.
Environ. Oual.. Vol. 7, No. 4, 1978. pp. 459-472.
Wauchope, R. D., L. E. Asmussen, and E. W. Hauser. "MSMA losses in runoff from small watersheds in the
Mississippi Delta and Georgia coastal plain." Proc. South. Weed Sci. Soc. 30: 395,1977.
Yen, G. T. and D. S. Ward. "FEMWASTE: A Finite-Element Model of Waste Transport Through Porous
Saturated-Unsaturated Media" Environmental Sciences Division Pub. No. 1462, ORNL-5601, Oak Ridge
National Laboratory, Oak Ridge, Tennessee, 1981.133 pp.
Yeh, G. T. and D. S. Ward. "FEMWATER: A Finite-Element Model of Water Flow Through
Saturated-Unsaturated Porus Media." Environmental Sciences Division Pub. No. 1462, ORNL-5567, Oak Ridge
National Laboratory, Oak Ridge, Tennessee, 1980. 149 pp.
9-28
-------
Johnson
APPENDIX A
THE STATISTICAL NATURE OF PRZM ESTIMATED MODEL OUTPUT
AND IMPLICATIONS FOR CLIMATE CHANGE SCENARIO
INTERPRETATION
The skewed nature of the chemical transport results presents some problems for even a simple
descriptive statistical summary. In response, the central tendency of a sample is presented in terms of the
median rather than mean. Climate change impacts are presented in terms of percentage change from the
base case median to the GISS or GFDL medians.
Since variability in the model output is about 2 to 3 orders of magnitude greater than the median, it
is prudent to consider the statistical significance of the model indicated climate change impacts. The extreme
non-normality of the sample does not lend itself to standard parametric tests which often assume Gaussian
distributions. A two-sample t-test applied to the sample ranks is equivalent to a Mann-Whitney-U using the
t approximation for the significance level (SAS, 1982). The two sided null hypothesis is, H.,: the populations
have identical distributions and H,: the populations differ with respect to location. Impact results contained
in the regional analysis that are significant at the .90 level are denoted by * in the tables of this report. Those
results significant at the .95 level are denoted by **.
For an example of further statistical considerations connected with PRZM model output, consider the
climate change impact results contained in Table 8, Appendix B (Regional Analysis). The absolute values
of the changes contained in Table 8 vary greatly. Note also the rarity of statistically significant results (none
for the GISS scenario and five cases for GFDL). Both of these observations arise from the large variability
in the model results. Even though some regional model changes are very large, sample variability is much
larger, resulting in an inability to reject the null hypothesis of identical distributions.
As sample size increases among homogeneous observations, variance usually decreases and the sample
distribution approaches the normal distribution (central limit theorem). In the case of regional climate change
results, this does not occur. The variance of the distribution of annual PRZM estimated chemical transport
does not decrease as more sites and scenarios are added to a region. This suggests that there is significant
variability among locations and scenarios within the region. That is, climatologically and physically the stations
are not homogeneous.
Another factor which influences the variability of the results is the use of multiple climate change grid
boxes within a single crop region. The direction and magnitude of modeled climate change is rarely consistent
across grid box or climate scenario. This serves to alter the existing time-space autocorrelation patterns
among stations within the cropping region. There is every indication that natural spatial variability across
climate, soil, tillage, and pesticide within a region, can be greater than the climate change whose impact is to
be assessed (Eddy and Sladewski, 1988). Large percentage changes and consensus results may be obtained,
but the changes infrequently attain statistical significance. Some of these regional "signal" versus "noise" biases
might be decreased by appropriate data stratification prior to significance testing.
9-29
-------
Johnson
APPENDIX B
REGIONAL CLIMATE CHANGE RESULTS AND INTERPRETATION
The Northern Great Plains (Spring Wheat)
Table 7 contains mean monthly GCM estimated temperature and precipitation modifications across
the spring wheat region. Table 8 contains a summary of resultant pesticide transport changes for this cropping
area. The values in Table 8 were obtained by determining the 30 year mean parameter value for across 81
scenarios at each weather site within the region. For spring wheat this is a sample of 486 30-year means. The
regional transport values in Table 8 represent comparisons of the median of the base 486 scenario means with
the median of each set of 486 GCM scenario means. The 3 sets of means were then stratified by soil or
pesticide and the same comparison of medians performed, this time with samples of 162 scenario means.
Table 7 can be used to suggest possible sources of the changes noted in Table 8. For instance, when all
management scenarios are considered together, there is an overall decrease in pesticide runoff transport under
the GFDL scenario. The GFDL scenario contains an increase in planting time precipitation, which could then
generate an increase in runoff and erosion of highly soluble/short-lived pesticides. Decreased runoff and
erosion of longer lived chemicals reflects GFDL precipitation decreases for the remainder of the growing
season. Leaching of active chemical below the root zone is decreased throughout. This is the result of a
smaller volume of water available to move chemicals through the profile or less pesticide remaining on the
surface for downward transport.
The Central Plains (Winter Wheat)
Table 9 contains a summary of regional climate change impacts for the Central Plains. GISS runoff
results and regional erosion show high statistical significance. That is, the alternative hypothesis of a shift
in the distribution of pesticide transport cannot be rejected. The smallest change under the GISS scenario
is noted on clay soil with slightly soluble/long-lived pesticides. The largest increases under the GISS scenario
are found when the entire population of scenarios in the region are considered with highly soluble/short-lived
and slightly soluble/long-lived pesticide subgroups. GFDL runoff results are mixed. Greatest chemical runoff
increases are noted among the highly soluble/long-lived pesticides. Greatest chemical runoff declines are
noted on sandy soils.
Table 10 contains the mean monthly GISS and GFDL model-produced regional precipitation and
temperature changes across the winter wheat cropping region. Under both climate scenarios, relatively dry
plant and preplan! conditions exist, but the remainder of the growing season (October through May) shows
increased precipitation. Monthly precipitation changes across the two climate scenarios are consistent except
during August. There is no crop in the field at this time, and surface losses of remaining pesticides will be
at their greatest Table 10 suggests that GISS planting conditions are more moist. If so, there is less soil
infiltration because the profile is already full and more pesticide is available to be washed-off when the rains
return. Dry August conditions under the GFDL climate change scenario encourage greater soil infiltration
when water is available leaving less chemical on the surface for later removal. Subsurface leaching of pesticide
is less under GFDL than GISS. This is the result of greater GFDL summertime temperature increases. As
in the Northern Plains, less moisture is left for later downward transport of chemicals.
9-30
-------
Table 7.
30-Year Mean Monthly Precipitation and Temperature
Changes For the Spring Wheat Cropping Region Under
GISS and GFDL Climate Change Scenarios
Johnson
MONTH
PRECIPITATION
(% CHANGE FROM BASE)
Jan
Feb
Mar
Apr (P)*
May
Jun
Jly
Aug (H)*
Sep
Oct
Nov
Dec
GISS
-18
+30
-12
-8
-3
-12
-17
+35
+104
-2
-26
+5
GFDL
+19
+30
+27
+20
-13
-28
-16
-17
-10
-10
+31
+1
TEMPERATURE
(°C CHANGE FROM BASE)
GISS
+6.16
+5.71
+5.51
+4.85
+3.50
+3.59
+3.38
+3.88
+5.43
+4.49
+3.82
+5.47
GFDL
+4.47
+5.97
+6.03
+5.38
+4.01
+8.73
+8.62
+7.95
+6.54
+5.57
+6.25
+6.19
* (P) - Crop planted
(H) - Crop harvested
Table 8. GISS and GFDL Scenario Impacts on Median Values of Pesticide
Transport in the Spring Wheat Cropping Region
GISS IMPACT
(% CHANGE FROM BASE)
Surface Runoff:
Region (all scenarios) +6
Highly Soluble/Short-Lived +2
Highly Soluble/Long-Lived +4
Slightly Soluble/Long-Lived +5
Clay +8
Sand +4
Silt +9
Surface Erosion:
Region (all scenarios) +17
Highly Soluble/Short-Lived +41
Highly Soluble/Long-Lived +1
Slightly Soluble/Long-Llved +10
Clay +10
Sand +25
Silt +19
Subsurface Leaching:
Region (all scenarios) -23
Highly Soluble/Short-Lived -41
Highly Soluble/Long-Lived -22
Slightly Soluble/Long-Lived -46
Clay -22
Sand +71
Silt -10
* Result significant at the .90 level
** Result significant at the .95 level
GFDL IMPACT
(% CHANGE FROM BASE)
-11
+45**
-21
-25**
-2
-7
0
+17
+126**
+5
-9
0
+50
+7
-63
-53
-69**
-84**
-41
-19
-11
9-31
-------
Johnson
Table 9. GISS and GFDL Scenario Impacts on Median Values of Pesticide
Transport in the Winter Wheat Cropping Region
Surface Runoff:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
Surface Erosion:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
Subsurface Leaching:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
* Result significant at the
** Result significant at the
GISS IMPACT
(% CHANGE FROM BASE)
+100**
+74
+96**
+69**
+168**
+104*
+73**
+49*
+42
+65
+47
+63
+34
+43
-6
+10
+10
+29
+127
-28
-32
.90 level
.95 level
GFDL IMPACT
(% CHANGE FROM BASE)
-5
+1
-16
-8
-1
+19
-5
-13
-16
+12
-19
+15
-3
-7
-36
-43
-32
-15
+7
-43
-65
Table 10.
30-Year Mean Monthly Precipitation and Temperature Changes
For the Winter Wheat Cropping Region Under GISS and GFDL
Climate Change Scenarios.
MONTH
PRECIPITATION
(% CHANGE FROM BASE)
GISS GFDL
TEMPERATURE
(°C CHANGE FROM BASE)
GISS GFDL
Jan
Feb
Mar
Apr
May
Jun (H)*
Jly
Aug
Sep
Oct (P)*
Nov
Dec
* (P) - Crop
(H) - Crop
+12.0
+12.3
+32.4
+8.7
+13.4
-.4
-.8
+18.2
-14.5
-15.2
+20.0
+10.5
planted
harvested
+38.4
+28.2
+10.1
+37.0
+7.3
-44.4
-40.0
-32.0
-17.0
-18.9
+49.1
+18.6
+5.54
+4.44
+5.53
+4.85
+3.11
+3.39
+4.20
+4.65
+4.19
+5.66
+5.52
+5.67
+4.13
+5.30
+4.61
+4.81
+5.84
+5.87
+6.23
+4.57
+4.79
+4.89
+4.65
+5.39
9-32
-------
Johnson
The Southern Hieh Plains and Blackla
Table 11 contains the regional double CO, impacts computed from PRZM model results for the
Southeraigh Plains and Blacklands areas. Statistically significant GISS impacts include: a regional chemical
runoff increase; an increase in highly soluble/short-lived and slightly soluble/long-lived pesticide runoff; and
an increase in highly soluble/short-lived surface pesticide erosion. The only statistically significant GFDL
result is an increase in slightly soluble/long-lived pesticide surface erosion. In general, GISS and GFDL
impact results suggest increased surface pesticide runoff and erosion across all pesticides and all soils.
Subsurface leaching decreased across most pesticides and soils.
Table 12 which contains monthly GISS and GFDL precipitation and temperature changes, provides a
source for such impacts. As in the case for winter wheat, both GISS and GFDL scenarios indicate
precipitation declines during the time of year when the fields have the least ground cover (October-February).
In February, there is a GISS precipitation increase. Some increase of GISS surface erosion and runoff could
be noted here. Planting conditions would be more moist and, since the soil profile is filled, more surface
runoff would result than in the drier GFDL case. Leaching in both scenarios generally decreases because of
increased surface losses and decreased excess moisture capable of carrying pesticides out the bottom of the
root zone.
Central Prairie (Corn)
Table 13 contains regional 2XCO2 induced pesticide transport changes across the corn cropping area
No GISS change scenario results were statistically significant. Significant GFDL results include: a decrease
in regional surface runoff; a decrease in highly soluble/long-lived surface runoff; and a decrease in regional
subsurface leaching and highly soluble/short-lived leaching. Most GISS and GFDL results for surface and
subsurface transport appear to agree that the GCM estimated climate change would result in decreases in
chemical losses in the Central Prairie.
Precipitation increases under the climate change scenarios occur primarily during December, January, and
February when sub-freezing temperatures would likely result in the moisture being stored as snowpack (Table
14). If the snowpack melts slowly when warm spring temperatures arrive, the impact of increased wintertime
precipitation on chemical runoff and erosion is minimized. The majority of the moisture would serve to
recharge soil moisture. If the snow melts rapidly over a frozen subsoil, erosion would be substantially higher.
Summertime temperatures increase for both climate change scenarios, most dramatically in the GFDL
scenario, increasing soil moisture demands. Increases in plant and evaporative demand resulting from these
much higher temperatures exceed modest early season precipitation gains. Late season precipitation declines
exacerbate the situation. The result is that more soil water is needed to support plant and evaporative
demands and less is available for pesticide transport out of the root zone. More pesticide is stranded within
the profile to degrade into non-reactive chemical constituents.
9-33
-------
Johnson
Table 11. GISS and GFDL Scenario Impacts on Median Values of
Pesticide Transport in the Cotton Cropping Region
Surface Runoff:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
Surface Erosion:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
Subsurface Leaching:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
* Result significant at the
** Result significant at the
GISS IMPACT
(% CHANGE FROM BASE)
+23**
+50**
+14
+13*
+18
+26
+19
+35
+37*
+22
+17
+25
+53
+31
0
+97
-40
-31
-33
-34
+5
.90 level
.95 level
GFDL IMPACT
(% CHANGE FROM BASE)
+17
+11
+36
+46**
+16
+11
0
+8
-21
+1
+18
+12
-12
+15
-35
+29
-51
-52
-66
+150
+14
Table 12.
30-Year Mean Monthly Precipitation and Temperature
Changes For the Cotton Cropping Region Under GISS
and GFDL Climate Change Scenarios
MONTH
PRECIPITATION
(% CHANGE FROM BASE)
GISS GFDL
TEMPERATURE
(°C CHANGE FROM BASE)
GISS GFDL
Jan
Feb
Mar
Apr (P)*
May
Jun
Jly
Aug
Sep
Oct (H)*
Nov
Dec
* (P) - Crop
(H) - Crop
-33.3
+3.8
-27.2
-1.2
+22.2
-19.8
-6.1
+12.1
+53.7
-16.2
-39.9
-4.8
planted
harvested
-12.5
-1.6
+35.9
-14.9
-29.1
+103.7
-2.3
+82.4
-25.4
+5.1
+17.8
-12.3
+4.69
+3.87
+4.82
+4.69
+2.87
+4.23
+4.26
+4.73
+3.29
+5.29
+5.27
+5.91
+3.65
+4.85
+4.46
+3.87
+5.90
+3.37
+4.29
+3.55
+4.35
+4.72
+4.38
+4.85
9-34
-------
Johnson
Table 13. GISS and GFDL Scenario Impacts on Median Values of
Pesticide Transport in the Corn Cropping Region
Surface Runoff:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
Surface Erosion:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
Subsurface Leaching:
Region (all scenarios)
Highly Soluble/Short-Lived
Highly Soluble/Long-Lived
Slightly Soluble/Long-Lived
Clay
Sand
Silt
* Result significant at the
** Result significant at the
GISS IMPACT
(% CHANGE FROM BASE)
+1
-12
0
0
+10
0
+3
-4
-12
-5
-5
0
+3
-4
-30
-23
-28
-66
-61
-15
-21
GFDL IMPACT
(% CHANGE FROM BASE)
-28**
-15
-31**
-32
-24
-30
-27
-12
-7
-13
-20
-5
-8
-2
-87*
-82*
-24
-57
-69
-54
-88
.90
.95
level
level
Table 14.
30-Year Mean Monthly Precipitation and Temperature
Changes For the Corn Cropping Region Under GISS and
GFDL Climate Change Scenarios
MONTH
Jan
Feb
Mar
Apr
May (P)*
Jun
Jly
Aug
Sep
Oct (H)*
Nov
Dec
* (P) - Crop
(H) - Crop
+11.9
+6.3
+17.7
+4.8
+8.9
+5.0
+12.7
-2.1
-33.9
-12.5
-14.1
+16.4
planted
harvested
PRECIPITATION
(% CHANGE FROM BASE)
GISS GFDL
+35.3
+30.1
+10.0
+1.9
+8.9
-25.2
-29.0
+8.8
-29.0
+6.5
+14.5
+36.6
TEMPERATURE
(°C CHANGE FROM BASE)
GISS GFDL
+6.25
+5.17
+5.11
+5.10
+2.92
+3.67
+2.71
+3.92
+6.55
+3.87
+5.71
+6.05
+5.35
+4.43
+5.15
+4.87
+3.79
+7.51
+7.78
+5.07
+6.58
+5.55
+5.51
+6.34
9-35
-------
FARM-LEVEL ADJUSTMENTS BY ILLINOIS CORN PRODUCERS
TO CLIMATIC CHANGE
by
William E. Easterling
Climate Resources Program
Resources for the Future
Washington, DC 20036
Contract No. CR-814653-01-0
-------
CONTENTS
FINDINGS 10-1
Specific Adjustment Possibilities 10-2
Specific Adaptation Possibilities 10-2
CHAPTER 1: INTRODUCTION 10-4
Background 10-4
Research Objectives 10-4
CHAPTER H: METHODOLOGY • 10-5
Focus on Farm-level Decision Making .'.... 10-6
Knowledge Engineering Procedure 10-8
Study Assumptions 10-12
CHAPTER 3: CLIMATE CHANGE SCENARIOS 10-14
CHAPTER 4: RESULTS 10-18
Analysis of the GISS 2xCO2 Scenario 10-18
Analysis of the GFDL 2xCO, Scenario 10-23
Summarizing Adjustments: Composite Portfolios 10-27
CHAPTER 5: INTERPRETATION OF RESULTS 10-30
General Conclusions 10-30
Specific Adjustment Possibilities 10-31
Specific Adaptation Possibilities 10-32
CHAPTER 6: POLICY IMPLICATIONS FOR EPA 10-33
Acknowledgments 10-34
REFERENCES 10-35
-------
Easterling
FINDINGS1
Adjustments to existing production practices are the first line of defense against climate change.
Production practices currently in use by corn producers can be altered with little cost to the farmer in order to
cope with some types of climatic fluctuation. Examples of such adjustments include altering planting dates, crop
mixes, tillage practices, and fertilizer applications (these are elaborated on below). This analysis suggests the
possibility that, in some scenarios of climate change, such adjustments could totally eliminate negative climate
impacts or even create a positive growth situation. The net effect of climate change on corn production after
adjustments are made by farmers will be significantly different from the effect of climate change without
adjustments.
More severe climatic changes are likely to make major adaptations necessary. If climatic conditions were
to become as hot and dry during the summer growing season in Illinois as is projected by the GFDL model, then
farmers who are used to raising full-season corn will be forced to adopt production practices that are quite
different from those in use today. Such adaptations would be necessary to deal with the potentially high heat
and moisture stress conditions that are likely to accompany such climatic changes. Examples of adaptations that
are likely to be made include installation and use of irrigation systems; changes in crop mix to include some
combination of short-season corn, soybeans and grain sorghum; and removal of some marginal land from
production on a given farm.
Climate changes in nongrowing season parts of the year will be important, and nongrowing season
adjustments are likely to follow. Climate change will occur in all seasons. Indeed, in the Northern Hemisphere
mid-latitudes, changes may be most dramatic in winter. Climate changes in the nongrowing season (fall
postharvest and winter) will affect some adjustment decisions. For example, a warmer and wetter winter would
increase nitrogen leaching and denitrification thus leading to a likely reduction of pregrowing season fall-applied
fertilizer.
Changes in regional comparative advantage will determine the extent of adjustment versus adaptation.
The effect of climate on agricultural regions competing with Illinois will directly influence how adjustments and
adaptations are made by Illinois producers. Economics apart, neither the GFDL nor the GISS 2xCO2 climate
scenarios would preclude corn from being raised in Illinois if appropriate adjustments (e.g., earlier planting,
altered tillage) and adaptations (e.g., irrigation, short season varieties) are made. However, when economics
come into play, many of these adjustments may raise production costs high enough to make corn production
unprofitable relative to other regions or other crops. On the one hand, if other production regions are similarly
adversely affected by the climate change and world demand for food and fiber cannot be met, then market prices
could rise. Such price increases would encourage more extensive efforts to adjust and adapt production to the
climate change. On the other hand, if comparative advantages begin to weigh more heavily in favor of other
regions as climate changes, prices do not go up and necessary adjustments by Illinois farmers become too costly,
the only strategy left for Illinois farmers is to leave farming altogether or, more likely, to choose a new crop to
grow.
If prices for crops uniformly rise, as predicted by Adams (this volume), and input prices do not rise
significantly, then adjustments are likely to be extensive in Illinois. I consider what sorts of specific adjustments
and adaptations that might be in store in the following two sections.
'Although the information in this report has been funded wholly or partly by the U.S. Environmental
Protection Agency under Contract No. CR-814653-01-0, it does not necessarily reflect the Agency's vieews, and
no official endorsement should be inferred from it.
10-1
-------
Easterling
Specific Adjustment Possibilities
A number of adjustments to existing production practices could be made by Illinois com producers in
response to changes in climate:
1. Changes in tillage practices. This research points to increased fall conservation tillage in Illinois. This would
offset the negative effects of a smaller possible number of field days in the following spring planting season and
would take advantage of dry conditions in the fall season. Though this response may not, by itself, have a large
effect on yields, it will contribute in a significant way to how farmers may respond
2. Less fall-applied nitrogen. Questions aside about the continued production of com in Illinois under climate
change, less nitrogen is likely to be applied in fall If nitrogen-fixing legumes such as soybeans replace com in
Illinois, then less nitrogen use needs no explanation. However, if com and other crops which benefit from
nitrogen application continue to be grown then the explanation for less fall-applied nitrogen is the desire to avoid
leaching and denitrification during the warmer (and possibly wetter) winter.
3. Planting earlier in the spring. It will benefit farmers to plant corn earlier in the spring than at present.
Earlier planting will result in the occurrence of anthesis (a period of high temperature and moisture stress
sensitivity) earlier in the summer when temperatures and soil moisture levels are less likely to be at stress levels.
Earlier planting could be enhanced by warmer spring temperatures. However, the warmer spring temperatures
could be offset somewhat by the greater heat storage capacity of a wetter soil (wetter from higher than current
precipitation in winter), i.e., as soil moisture increases, more energy is needed to raise the temperature of the
soil.
4. Changes in corn varieties. Climatic change will encourage farmers to alter the corn varieties planted either
to maximize potentially beneficial climate changes or to minimize potentially negative climate changes. If planted
earlier than present, there is a possibility that high yielding, full season varieties would do well under the GISS
2xCO2 summer climate. Also, if planted earlier than now, low yielding short-season corn varieties could be
grown under the relatively harsh GFDL 2xCO2 summer climate.
5. Lower planting densities. Potentially drier soils during summer will likely lead to changes in planting
densities. Lowering the number of plants per unit area would conserve soil moisture. Given that planting
densities have been increasing steadily in recent years, it is likely that densities will be higher than present but
not as high as they would have been without climate change.
6. More attention to pests. Warmer temperatures are likely to stimulate greater pest problems (this is handled
more extensively by Skinner et at., this volume). It would be reasonable to expect farmers to respond with some
increase in pesticide use but perhaps accompanied by greater emphasis on integrated pest management practices.
7. Earlier harvest. If planting does occur earlier in the spring than now, it foUows that harvest will begin earlier
than now. This is especially true if short-season varieties were planted in spring. Also, changes toward warmer
and drier late summers and early falls are likely to favor earlier corn maturation and to accelerate drydown. Not
only would farmers be able to harvest earlier than now, but there is likely to be less need for artificial drying.
Specific Adaptation Possibilities
1. Altered crop mixes. Depending upon the exact nature of climatic change, farmers are likely to change not
only the varieties of corn to be raised, but also to alter the types of crops grown. Crop mixes could become more
diverse with likely combinations (either intercropped or annually rotated) of corn, soybeans, and grain sorghum.
If com is grown under the GFDL 2xCO2 scenario, then almost certainly it will be a short-season variety.
2. Increased irrigation. According to the experts, current climatic and economic conditions would support use
of irrigation in Illinois approximately 3 years out of 10 years. Climatic changes, especially changes which leave
10-2
-------
Easterling
soils drier during summer than now, will likely lead to increased use of irrigation in Illinois. Increased use of
irrigation is supported by projected price increases for all crops likely to be grown in Illinois. However, not all
farmers can be expected to turn to irrigation as climate changes. Insufficient water supply and nonirrigable soils
will limit where irrigation may come into practice. Moreover, even with severe climatic changes, some crops may
be profitably grown without irrigation (e.g., soybeans, grain sorghum).
3. Marginal land taken out of production. If climatic changes cause production costs to rise in Illinois, there
is likely to be a corresponding decrease in total acreage planted on a given farm. Poorer quality land (tess
fertile, highly erosive) on any farm is likely to be taken out of production. Such action would tend to help hold
the line on reductions of productivity per unit of land.
10-3
-------
Easterling
CHAPTER 1
INTRODUCTION
Background
Scientific consensus is emerging that climatic change will impact production agriculture (Rosenberg,
1986; Ciborowski Mid Abntamson, 1966). The extent of impacts is not dear nor is it known, on balance,
whether impacts wffl be positive or negative. This uncertainty stems partly from incomplete knowledge of how
climate will change and pvtry Cram lack of understanding of how agricultural systems will adjust to impacts of
changing climate.
The purpose of this paper is to provide insight on how existing farm level production technologies might
be adjusted in response to dynamic climate. The focus will be on corn production in Illinois as representative
of production agriculture in the core of the midwestern Grain Belt.
Why corn production and why Illinois? Corn was second only to soybeans over the period 1978-1982 in
share of total U.S. agricultural output held by major grains (corn, wheat, soybeans, grain sorghum). It is a
major agricultural export commodity, again second only to soybeans among grains and it is, thus, a major
international trade commodity in the overall VS. export economy.
. Illinois (17%) is second only to Iowa (21%) in percentage of U.S. corn produced. Approximately 81%
of the total land area of Illinois comes under the plow. Much of this 81% is cultivated with corn. Moreover,
in 1985, Illinois led all Corn Belt states in value per acre of land and buildings.
Corn, regardless of where it is grown, is sensitive to vagaries of climate and weather. Perhaps more than
any other major grain crop, corn yields can be substantially reduced (or enhanced) by anomalous weather
conditions. Recent Illinois droughts (e.g., 1976,1980,1983,1988) have resulted in major documented corn yield
disruptions across the state. Indeed, Richman and Easterling (in press) have shown that anomalous weather
conditions strung out over several growing seasons can be detected in corn yield time series, account being taken
for technology enhancements of yields.
Research Objectives
Specific objectives of the research reported here are the following:
1. To identify the climate sensitivities of current farm-level com production technologies in Illinois, especially
those which may be subject to adjustments under climate changes specified in NASA/GISS and GFDL GCM
runs with doubled ambient carbon dioxide levels (henceforth 2xCO2).
2. To determine, under a set of constraining assumptions, decision options for adjusting production technologies
to climate change.
3. To identify situations (e.g., applications of policy, economic conditions) that might facilitate or hinder certain
adjustment decisions.
4. To develop a set of hypothetical scenarios in which farmers adjust to climate change.
5. To identify a set of environmental policy issues relating to farm level adjustments to climate change.
10-4
-------
Easterling
CHAPTER H
METHODOLOGY
Despite several decades of rapid technological advance in production agriculture, climate continues to
exert strong influences on productivity (Warrick, 1984). Sensitivity and resiliency of agricultural systems have
been suggested in recent empirical analyses. These studies not only document climate impacts on production,
but also the capacity of agriculture to adjust and adapt to climate. For example, Parry et al. (1986) have
coordinated a series of studies showing that crop production at various growing margins (e.g., cold, warm/dry)
around the world are responsive to small shifts in climate. Rosenberg (1982) demonstrated that hard red winter
wheat production has expanded geographically in the North American Great Plains across spatial climatic
gradients that exceed predicted climatic changes from the greenhouse effect. Adaptability of wheat culture was
inferred from this expansion.
While useful insight is gained from empirical hindsight studies such as those cited above, it is not entirely
possible to base predictions of future agricultural response to climatic change on such analysis. New production
technologies and occurrence of economic and social conditions unlike any of today will likely make agriculture
quite different in the future. Yet, existing predictive studies have their own problems. Biasing and Solomon
(1982) examined climatic constraints on com production in light of a scenario of future climatic change leading
to changes in temperature and precipitation. They concluded that the Corn Belt would shift on an axis from
southwest to northeast in the United States (and into S. Canada). Moreover, they concluded that much of the
Corn Belt that is now in dryland farming would be forced to rely on irrigation, and many currently irrigated Com
Belt locations would be forced out of corn production.
Rosenzweig (1985) used a similar approach to Biasing and Solomon. She used a general circulation model
(GCM)-based projection of climatic change, under CO2 doubling, to examine wheat production in the United
States. In general, potential wheat producing areas were found to expand under climatic change. Some cultivars
would likely be stressed by high temperatures along the current southern boundary of wheat production.
Though studies such as these provide useful first approximations of how climate might affect production,
they do have deficiencies. First, they tend to rely on a single scenario of climatic change. Rosenberg (1987) has
demonstrated major differences between sophisticated GCM predictions of climate change for important
agricultural regions such as the U.S. Midwest. He concludes it is best to consider a range of possible climate
change scenarios in any attempt to estimate its impacts.
Second, these models do not explicitly consider potential incremental adjustments by fanners and related
agribusiness. Yet, adjustments to production practices in response to climate change will play a major role in
determining actual impacts on agricultural productivity. These adjustments must be considered before an
accurate assessment of impacts can be made. Although Decker et al. (1986) cataloged a variety of agricultural
impacts from CO2-induced climatic change and possible adjustment strategies to climatic stress, little analysis
was provided of conditions under which these adjustments might be made.
Third, in these analyses, agricultural impacts of climate change are examined in isolation from other less
directly climate-linked trends in production agriculture. In essence, the future economic context of agriculture
is ignored. Therefore, these studies, like empirical hindsight studies, must be viewed with caution in making
projections about the future. Some foreseeable trends such as more restrictive use of pesticides on the farm
may conflict with logical strategies for coping with pest outbreaks brought about in part by changing climate.
These trends make it important that agricultural impact studies be cast within a broader context of climate and
nondimate influences on production.
10-5
-------
Easterling
Focus on Farm-level Decision Making
In this study I focus on farm level decision making as the key to the understanding of how farmers might
adjust existing technologies in response to climate change. I presume that decisions by a farmer on how to
manage a crop are influenced by factors which affect productivity and hence profits. These factors include his
knowledge of existing conditions and his rational expectations of near future conditions (Sonka and Lamb, 1987).
Figure 1 shows knowledge and expectations of current and near future economic conditions such as, for example,
market prices for inputs and outputs; environmental conditions such as, for example, soils, water quality and
climate; and social values such as, for example, maintaining cultural traditions. Some of these conditions weigh
more heavily than others in influencing a farmer's decision strategy. But they all must be examined if one is to
try to understand how a fanner might behave in a particular way in managing a crop.
In this study my focus is on those production decisions by farmers which are influenced directly or
indirectly by climate conditions. A number of techniques have been suggested for gaining insight into how
production decisions might be influenced by climate. Lovell and Smith (1985) identify several microeconomic
modeling approaches as being appropriate for this task including microsimulation, econometric, systems analysis,
and input-output modeling. Decision analysis (Ausubel, 1980; Sonka and Lamb, 1987), in which some index of
farmer well being is examined in light of various combinations of production decisions, is another possible
approach. However, these approaches are somewhat time-consuming to develop and data-intensive and thus
inappropriate for the task at hand.
The approach I chose for this study was to marshal expert judgment as to how farmers might adjust
existing technologies to climate change. Specifically, I used what social scientists refer to as a case-scenario
approach, which was modified with the inclusion of principles of the development of expert systems.
Case-scenario approaches involve the presentation of scenarios to individuals and then asking them how
they might respond to any of the scenarios if they occurred. This is an accepted methodology for climate impact
assessment (Wilhite and Easterling, 1987). The objective of this is to allow individuals to draw on their
accumulated expertise in resolving a hypothetical problem. Glantz (1977) used this approach to ask decision
makers what they might have done differently in response to a drought had they had available a perfect forecast
of the drought beforehand. In the current study, experts were provided different scenarios of climate change and
were asked how farmers-decision makers, given all of the climate and nonclimate factors that influence
production decisions, might adjust existing technologies to cope with any problems the climate change might
present. The justification for the use of the case-scenario approach and the manner in which the approach was
carried out for this study will be amplified further below.
The procedure followed in this study to present climate scenarios to experts and to retrieve their
assessments involved the use of expert system development principles.2 I stress that this was not intended to lead
up to the development of an expert system (at least not for this study) but rather was to provide a systematic
framework for the gathering and synthesis of expert judgment. Put another way, the expert systems development
principles are simply a means of adding rigor to the case-scenario approach.
Expert systems are developed using a process known as "knowledge engineering." Knowledge engineering
is a structured interviewing process wherein modeler(s) and expert(s) interact to build a progression of expert
logic beginning with basic assumptions, facts, and heuristics which lead to some solution or derision. In this
research, the intent is to use knowledge engineering to assess factors influencing changes in farm production
practices under climate change.
Expert systems are formally a branch of artificial intelligence. They are defined as computerized
representations of human problem solving. All facts and heuristics ("niles-of-thumb") that an expert uses in a
particular domain of expertise can be represented symbolically on a computer in such a form as to be used
directly in resolving practical problems (Hayes-Roth, et al., 1983).
10-6
-------
Easterfing
(Current Year's Planning)
Expectations o
market conditions
for inputs and output
\
Expectations of
climatic events, j— —
pests, and diseases/
Knowledge of production
requirements based on
current technology
Knowledge of
actual conditions
xpected
profit
xpected
yields
Financial, societal and
physical constraints
Non-economic goals
(Actual Events)
Production
decisions
Incidence of
pests and diseases
Climatic
events
Resulting crop
(Market Effects)
Demand
conditions
Supply conditions in
non-affected areas
Market
price
(Following Year's Planning)
/ Expectations of
[market conditions
\for inputs & outputs
/Expectations of
I climatic events
\pests and diseasesj
Figure 1. Schematic of decision making at the farm level (after Sonka and Lamb, 1987).
10-7
-------
Easterling
Why use expert judgment and knowledge engineering for this study? Both offer several desirable
characteristics in meeting the stated objectives of this research. Clark (1986) and Brewer (1986) argue that
expert judgment is a valid means of developing insight into how complex systems behave. They imply that expert
judgment is probably no better and no worse than large numerical models in predicting complex alternative
futures, given data and methodology limitations of the large models. The use of expert judgment and the use
of numerical models are not mutually exclusive. Indeed, the systematic combination of expert judgment and
numerical modeling is a particularly strong form of analysis (Coulson, et al.t 1987). However, expert judgment
need not necessarily be linked to numerical models to be useful This is particularly true when numerical models
are limited by data scarcity, poor understanding of the processes being modeled, or a combination of these two.
Much practical expertise can be used to further the understanding of impacts of future climate change.
There is considerable knowledge of how crops respond to weather and how fanners can adapt to
weather-induced crop stress. This is supported by an extensive agroclimatic modeling and basic agronomic
research literature and the accumulated expertise of county extension agents and crop consultants. But as
pointed out above and elsewhere, such research is yet to be compiled and synthesized to useful models or other
forms of information (Easterling and Sonka, submitted). It is proposed here that knowledge engineering
provides a means of integrating undisputed factual information with less well documented heuristic information
requiring expert judgment (e.g., interpretation of model output, use of nonscientific but valid knowledge in
decisionmaking, assessment of reasonableness of future scenarios).
Knowledge Engineering Procedure
Two crop consultants (Dr. Harold Reetz, of the Potash and Phosphate Institute, and Mr. Ronald Olsen,
of Top Soil, Inc.) with considerable knowledge of how climate affects Illinois corn production participated in the
knowledge engineering (henceforth used interchangeably with knowledge acquisition). I intentionally avoided
the inclusion of active farmers as experts. At this time most farmers are generally unaware of climate change
implications for their own operations. However, I recognize that this is changing and future assessments should
include farmers. The goal of the knowledge acquisition was to elicit from these experts ideas on likely responses
by farmers to two alternative scenarios of climate change (discussed below).
Knowledge aquisition was structured to follow a simple conceptual model of climate and society
interaction. Such models are discussed at length in Kates et al. (1985). Figure 2 shows that the process of
climate and society interaction begins with climate changes. These are represented by GCM-based scenarios
which are discussed further in the next section.
These climate changes result in physical impacts such as altered soil moisture, runoff, soil erosion, and
so on. Climate changes directly cause physical impacts, which result in immediate agronomic impacts and likely
prompt adjustments.
The knowledge acquisition procedure is based on the above schematic. It is repeated for individual
seasons of the year (as these seasons are defined below). In addition to insight into how impacts and
adjustments might occur within individual seasons, knowledge was sought as to how climate-altered conditions
within one season might relate to agronomic conditions (including climate) in a previous or uncoming season.
In many instances, a farmer's knowledge of climate-altered conditions in a past season or expectations of
climate-altered conditions in an upcoming season may represent necessary contingencies for adjustment decisions
in the current season. For example, if the farmer expects the upcoming winter to be warmer and wetter than
now, then his fall (current) decision might be to apply less nitrogen (to avoid winter leaching) even though the
now drier fall favors nitrogen application activities.
Another point raised in Figure 2 relates to the general economic, social, and political environment within
which corn will be produced as climate changes. Nonclimate factors will undoubtedly influence how farmers will
choose to respond to changing climatic conditions. Trends hi technology, production costs, government
programs, and markets, to name a few, will have some influence on how farmers will adjust to climate change.
10-8
-------
Easterling
Some of these nonclimate trends were identified in the knowledge aquisition procedure, and qualitative insight
was sought as to their likely effect on climate adjustment decisions by farmers. Though I was unable to approach
these nonclimate factors as systematically as I would have liked, any consideration of them at all breaks new
ground in the climate impacts literature.
The experts were prepared for their interviews as follows:
1. They received a written statement of the research problem and the research objectives, including
dear explanation of what was expected of them in knowledge engineering sessions.
2. They were shown scenarios of climate change produced by the GFDL ZxCOj, and the GISS
2xCO2 model runs (these are described below more fully).
3. They were shown projected yields for com and soybeans based on the climate change scenarios
(from Richie, this volume). Such yield projections were given for Illinois and surrounding states.
(These are discussed below.)
4. They were shown projected prices (discussed below) for corn and soybeans based on projected
yields (from Adams, this volume).
5. They were briefed on the conceptual model discussed above, which was used to organize retrieval
of expert knowledge. Experts were asked to respond specifically to this model. In responding,
experts were required to provide all assumptions, facts, and heuristics brought to bear in
identifying potential farmer adjustments to climate change. These are to be represented as a
knowledge base of farm-level adjustments.
After results from the knowledge engineering sessions were synthesized they were portrayed in
Diagrammatic form. In this form, they might be thought of as a simple systems model. As such, they were
presented to the experts for evaluation and resolution of conflicts of opinion between experts. When I identified
a conflict of opinion, it was pointed out to the two experts. Fuller explanations of the item of conflict were
sought from each expert. Often this additional clarification was sufficient to resolve the conflict. For example,
there was a conflicting opinion on whether planting densities would rise or fall in response to a particular climate
scenario. After receiving clarification on this issue, I determined that both experts agreed that planting densities
will rise over the long term but at a slower rate than if climate had not changed. If a conflict of opinion
remained after the clarification exercise, a third expert would be consulted to evaluate the issue and either break
the tie or recommend that the conflict be maintained and represented as an uncertainty in the analysis.
I presented estimated yield and price responses to climate change to the experts when they become
available from other EPA researchers. The methods used to develop and analyze these projections appear
elsewhere in this volume (see Richie for more on yields and Adams for more on prices).
Projected yields (both for dryland and irrigated production) from Illinois and a proximate region to the
south (Missouri) were presented to experts (Table 1). Inclusion of yields from Missouri is to allow experts to
see the climate effect on a competing nearby area Projected price changes will of course be uniform across all
regions. Projected prices are shown in Table 2.
The experts were provided with these projections at the same time they were provided with the results.
They were asked to evaluate the systems model in light of the yield and price projections. Thus the results
appearing in this report incorporate the experts' consideration of projected yields and prices and their effects on
adjustment/adaptation strategies.
It is useful to make a distinction in this study between changes in farm production which I represent as
adjustments versus changes which I represent as adaptations. Adjustments are meant as the fine tuning of
10-9
-------
Easterling
Seasonal Climate
Change Scenario
(GISS & GFDL)
1st Order Impacts
(Biophysical)
2nd Order Impacts
(immediate agro-
nomic significance)
Contingency
for decision
response
Decision
Response
Goal
Other influences
on decision
response
(competing,
obscuring)
I
Figure 2. Farm level adjustments knowledge acquisition framework.
10-10
-------
Table 1. Projected Crop Yields (from Ritchie, this Volume)
Easterling
Location
and Crop
Baseline Yield
(Kg/Ha)
GISS 2xC09 Yield
(Kg/fla)
GFDL 2xC09 Yield
(Kg/Har
Peoria dryland corn
Peoria irrigated corn
St. Louis dryland soybeans
*St. Louis irrigated soybeans
Springfield, MO dryland corn
Springfied, MO irrigated corn
Springfield, MO dryland
soybeans
Springfield, Mo irrigated
soybeans
7,429
12,709
2,090
3,407
6,069
11,632
1,772
3,244
6,517
10,517
2,058
3,167
6,674
8,398
1,921
3,024
2,427
6,196
1,107
2,630
4,805
6,923
1,165
2,596
*St. Louis soybean yields were used since Peoria soybean yields were
unavailable at this writing.
Table 2, Projected Crop Prices (from Adams, this Volume)
Crop
Baseline Price
GISS 2xC02
Price % of baseline
GFDL 2xC02
Price % of baseline
Corn
Soybeans
Sorghum
2.68
5.78
2.58
3.10
6.19
3.74
116
107
144
5.95
14.21
5.00
222
245
193
10-11
-------
Easterling
on-farm production processes already in place. Examples include changes in tillage practices, planting dates,
and varieties. These are not likely to result in major investments in new machinery or radically different farming
practices. Adaptations, however, are more fundamental and long term. They represent changes to production
operations which were not previously in place on the farm. Examples include changes in crop species,
introduction of irrigation, and even farm abandonment. The distinction between adjustment and adaptation is
useful in assessing the feasibility of continuing to grow corn in Illinois if the climate changes. Adjustments imply
that corn could remain feasible under climate change from an agronomic standpoint. Adaptations suggest that
corn would be only marginally feasible or not be feasible at all.
Study Assumptions
A number of constraining assumptions were made in developing the analysis. These assumptions were
partly to simplify what is an obviously complex task and partly to keep this analysis consistent with
complementary initiatives of the the U.S. EPA's agricultural effects report. Many assumptions were derived in
committee during earlier group planning exercises. Key assumptions were the following:
1. GCM climate change scenarios are prescribed by EPA. EPA principal investigators were provided
three GCM scenario sets. These included the NASA/GISS and Princeton/GFDL model output for atmospheric
CO2 doubled over preindustrial levels. An additional GISS run provided 10-year transients beginning today and
running to CO, doubling. The transient run was not included in this research. The technique for adapting these
scenarios to Illinois was prescribed by EPA study organizers. This is discussed below for the 2xCO2 scenarios.
2. Technology is held constant. As agreed in earlier planning meetings, technology is to be kept at
current levels. In this study, this was interpreted to mean that new technology not currently available (e.g., crops
genetically engineered for CO2 increase and climate change) might not be included, but that current
technological trends not yet fully adopted on all or most farms (e.g., increasing on-farm mechanization) might
be included in the analysis.
3. Corn as initial crop of choice. Corn (or corn-soybean rotation) is assumed to be the first choice of
producers prior to imposition of a climate change. However, shifting from corn to some other crop or shifting
out of production altogether are considered valid responses to climate change.
4. Direct effects of CO2 not considered. The potential direct impact of increasing atmospheric CO,
levels on plants, such as increased plant water-use efficiency, though important, is not explicitly considered. This
constraint had to be imposed because none of the CO2 direct effect results were available from the crop
modelers involved in the study at the time of this writing.
5. Market prices increase but movement in overall costs of production are unknown. There are
conflicting views as to what climate change will do to commodity prices. In a paper for a forthcoming Resources
for the Future book, Easterling, Parry and Crosson posit that overproduction in the western world will continue
to plague grain production (even with climate change) and that some countries may benefit agriculturally from
climate change. How prices will respond is unclear, though a fall is not out of the question when global effects
of climatic change are considered. Given the uncertainty surrounding prices, the current study was conducted
under an initial assumption of no relative change in market prices. This assumption is relaxed later and
projected price increases derived by Adams for this report are introduced.
I assume production costs to be relatively unchanging. It is not easy to justify this assumption within the
framework of this study since I assume constant technology. In reality, however, if productivity is permitted to
continue to increase on average in the future as it has in the past - and we think it will (Easterling, Parry and
Crosson, 1989) -- and provided environmental costs are equitably managed, there is little reason to expect
production costs to rise, climate change apart. I do stress, however, that climate change may put pressure on
some costs to rise and thus affect how certain decisions are made.
10-12
-------
Easterling
6. Government programs continue. This may be the weakest of our assumptions. Some speculate that
the protectionism currently practiced in the US. will give way to freer market-driven trade. Moreover, growing
emphasis on environmental protection and resource conservation are likely to have increasing influence on farm
level production practices. However, formal policies have been demonstrated to be slow to change (e.g^ the
General Agricultural Trade and Tariff Agreement, GATT, was implemented in the late 40's and remains the
basis for many of today's farm support policies). To project new policies not even under consideration today
would prove to be an almost intractable problem. Thus, for this analysis, we assume no change in legislation
or executive agency practice, but that positive attitudes toward resource conservation and environmental
protection will continue to influence on-farm production practices.
7. All biophysical impacts except yields are subjectively determined. This is the assumption most easily
relaxed in this analysis. In the initial conduct of this study experts were asked to suggest how the climate change
scenarios might affect growing conditions. The basis for this request rests on the experts' perceived skill in
anticipating crop problems based on a limited amount of climate information. Dearly this is not the best
available method of assessing the kinds of impacts which farmers may face and be forced to adjust to. In fact,
as crop simulation model yield projections under climate change became available from other EPA investigators,
such projections were used in this study.
Perhaps the greatest limitation of this analysis lies in the subjective and qualitative nature of the research
design. Though intensive efforts were made to create an objective knowledge acquisition protocol, including
informing experts of the bounds of known facts about climate change and accompanying impacts, the fact
remains that there is much expert opinion in the results. These results are a useful first approximation of steps
com producers can be expected to take, under the aforementioned constraining assumptions, in responding to
the various impacts of climatic change.
Future research will seek to provide more quantitative understanding of the likelihood of certain
adjustments being made. Expert knowledge gained from the above exercise will be coded into an expert system
computer program. Efforts are under way to integrate the farm level decision analysis with mechanistic crop
models and outputs of market level economic models. This can all be incorporated into a single integrated
expert system (as discussed by Coulson et al., 1987). The expert system could then be used with actual decision
makers to record how they would actually respond to simulated farm conditions.
10-13
-------
Easterling
CHAPTERS
CLIMATE CHANGE SCENARIOS
Climate change scenarios used in the case scenario analysis were constructed from the NASA/GISS and
GFDL models. Specifically, monthly mean temperature averages and monthly precipitation totals for a 2xCO,
climate change were derived for a representative point in Illinois. This was done in a two-step process. First
the ratio of monthly climate under a icCOjrun to monthly climate in a control run (simulation of present
climate) is calculated from the GISS and GFDL models. Second, these ratios were multiplied by the actual
average monthly temperatures and precipitation totals to calculate the marginal amount of temperature and
precipitation change. These monthly temperature and precipitation changes were added to actual current monthly
normals to produce scenarios of 2xCO2 climate in Illinois. Peoria was chosen as representative of the most
productive com growing regions in Illinois. Thus the actual (observed) climate data is from Peoria
Figure 3a-c shows the monthly plots of 2xCO2 temperature and precipitation for the GISS and GFDL
model runs. The current monthly normals are plotted on each figure for comparison.
Temperatures are consistently elevated throughout the year in the 2xCO2 GISS scenario, with a small peak
occurring in September. Precipitation is somewhat below current normals during the late summer and fall but
above current normals from winter to mid-summer.
Temperature changes in the GFDL 2xCO2 scenario were similar to those of the GISS scenario except for
one important difference: temperatures during the agronomically critical mid-summer are considerably higher
in the GFDL scenario. GFDL precipitation changes were markedly different from GISS changes in both sign
and magnitude. According to the GFDL model run, precipitation during the mid-summer would be reduced
considerably. This pattern would remain in effect until late fall. In late fall precipitation increases to slightly
above the current normal until early summer.
The scenarios given to the experts were limited to these monthly changes rather than weekly or daily
changes in order to avoid information overload. Moreover, the experts were to suggest how, in view of their
experience with the Illinois climate, these monthly changes might be reflected during shorter periods within
months. Were further time and resources available to do so, these scenarios could be constructed to reflect
climatic conditions during periods when crops are particularly sensitive such as during silking and tasseling.
Nonetheless, it was felt that much useful information could be derived with monthly scenarios.
One serious concern deserving of comment here exists in the analytical structure. Some of the more
extreme climate changes (especially in the GFDL scenario) might exceed the range of experience of the experts.
That is, the average climate changes in the 2xCO2 scenario could be greater than even the extremes that are
occasionally experienced in Illinois's current climate. Useful information about adjustments should be extractable
under these conditions as well. Experts pointed out that corn is currently grown outside Illinois under climatic
conditions that are similar to the Illinois GFDL 2xCO2 climate (e.g., southwest Texas). Moreover, both experts
were knowledgeable of corn production in such regions. Even if experts were not experienced in climates of
other regions, it can be argued that when experts are asked to consider conditions outside their own experience
with Illinois climate, they will reason from the most extreme events that they have experienced. We cannot
exclude the possibility that, sometime between now and JxCOy extremes matching the experts' outer range of
experience will occur. Thus, their response may represent an outer range of experience, admittedly less extreme
than 2xCO2, but a scenario that will likely occur some time in the future.
10-14
-------
Easterling
GCM PREDICTIONS UNDER 2X CO2 SCENARIO
TEMPERATURE - PEORIA, ILLINOIS
u_
o
-------
Easterling
GCM PREDICTIONS UNDER 2X COS SCENARIO
PRECIPITATION PEORIA, ILLINOIS
4.5 -
4 -
3.5 -
a
'o
CD
2.5 -
1.5 -
0 CONTROL
+ GISS Model
1 2 3 4 5 6 7 8 9 10 11 12
Month
Figure 3. GCM climate scenarios for Peoria, Illinois - (b) precipitation - GISS.
10-16
-------
Easterling
GCM PREDICTIONS UNDER 3X CO2 SCENARIO
PRECIPITATION - PEORIA, ILLINOIS
n CONTROL
o GFDL Model
10 11 12
Figure 3. GCM climate scenarios for Peoria, Illinois - (c) precipitation - GFDL.
10-17
-------
Easterling
CHAPTER 4
RESULTS
Results of the knowledge aquisition sessions were represented diagrammatically in a series of flow models
of potential climate impacts and farmer adjustments. It is important that these results be interpreted as
representing how an Illinois corn producer might adjust production practices and what related issues he might
face if the climate were to shift suddenly to a new equilibrium state. An instantaneous response of climate is
implied by doubling CO, levels in the GISS and GFDL GCMs. Though assessments of adjustments are best
discussed in terms of gradual warming over several decades, analysis of response to quick change is a useful
starting place.
A separate analysis was done for certain agronomicaUy meaningful periods within the year. Linkages of
decisions between periods were considered. These periods are (a) fall post-harvest; (b) spring and early summer
preplan!, plant; (c) summer anthesis and grainfill; and (d) fall drydown and harvest.
Analysis of the GISS 2xCO2 Scenario
Fall Post-Harvest. Upon review of the GISS 2xCO2 year, the experts felt that corn production could
remain feasible with only adjustments to existing technologies already in place on most farms. Assessment begins
with potential adjustment decision paths during the fall post-harvest period (Figure 4). Tillage operations and
fertilizer applications, each being sensitive to climate change, are normally under way at this time. The
immediate consequence of wanner and drier post-harvest conditions is a drier soil. The immediate agronomic
impact is an increased probability of more field days -- days in which farmers can move heavy equipment into
the fields.
Ii farmers have grown accustomed to expecting wetter conditions (fewer field days) during the upcoming
spring preplanting and planting period (as GISS predicts), then the greater number of post-harvest field days will
encourage more fall tillage operations and fall-applications of potash and phosphate fertilizers where such is
possible in Illinois. In effect, more fall field days provide the opportunity to compensate for the fewer spring
field days expected. But this strategy is somewhat dependant on soil type. Soils in southern Illinois are not and
would not be worked in fall.
If farmers grow to expect warm wet winters (as GISS predicts), the incentives to apply nitrate-based
fertilizers in fall will decrease. In such conditions denitrification and nitrogen leaching increases and the usual
effectiveness of fall fertilizer application (i.e., allowing nitrogen to diffuse into the soil) is diminished.
Nonclimatic influences on these decisions derive largely from the growing concerns for environmental
management at the farm level. Any increase in fall tillage will conflict with soil conservation policies such as
T-by-20003 and with general conservation concepts implicit in the Food Security Act of 1985. These conflicting
interests might lead to a general increase in fall conservation tillage.
Environmental concerns, especially pertaining to agricultural chemicals in runoff, would likely reinforce
a reduction of fall application of nitrate-based fertilizers. However, phosphate and potash-based fertilizers are
less mobile in the soil than nitrates and, thus, might engender less concern.
3T-by-2000 is an expression describing attempts by states to encourage farmers to follow practices which
reduce soil erosion to rates which fall within certain tolerances as described by a model known as the
Universal Soil Equation.
10-18
-------
Fall Post-Harvest
(October-December)
Easterling
Seasonal climate
scenario
GISS climate:
warmer,
dryer
Biophysical
impacts
soils remain
dry
Immediate agronomic
significance
more possible
field days
Contingency for
decision response
Decision response
(contingency met)
increased tillage,
increased P, K
application
reduction or
or elimination
of N application
Decision response
goal
increase field
preparation
efficiency
reduce negative
effects of N
application
Other influences
(t = supports decision
response, may override
or obscure)
(- = conflicts with
decision response,
may override or offset)
conservation
policies e.g.
T by 2000
restrictive use of
certain chemicals
Figure 4. GISS fall post-harvest adjustments.
10-19
-------
Easterling
Spring Preplant and Planting. Farmers face a number of production decisions daring spring which are
directly or indirectly climate-dependent (Figure 5). These decisions focus on fertilization, field preparation,
planting, and pesticide applications.
According to the GISS scenario spring will be warmer and wetter than now. Biophysical impacts will likely
include soils more often at field capacity early in the season. Soils may warm earlier in the spring although
wetter soil would likely offset such warming to an extent. The immediate agronomic significance of such impacts
will be fewer field day opportunities in spring although warmer conditions may allow field activities to begin
earlier.
If nitrogen-based fertilizers cannot be applied so readily in fall, more nitrogen will have to be applied in
preplant or side dressed in early summer than at present. Additionally, there may be agronomic incentives to
increase total nitrogen application in order to take advantage of a shift toward generally more favorable climatic
conditions and higher yields. The experts emphasized, however, that increasing concern over water contamination
from agricultural runoff may limit use of nitrate-based fertilizers in the future.
If the previous winter were warm (as predicted by GISS) and warm conditions are expected to occur over
the upcoming summer (predicted by GISS) then pest populations are likely to increase (see Skinner et al., this
volume). The decision response would favor increased pesticide applications, even if fanners embrace "apply
only when pests are present" practices. Growing restrictions on the uses of agricultural chemicals (as discussed
above) conflict with this outcome. Intensified development of integrated pest management practices (innovative
biological and nonchemical management technologies) provides an alternative.
If tillage is done the previous fall, only a single pass over the field with a cultivator in early March will be
needed as preparation for planting. Not only will field preparations begin about two weeks earlier, but a single
equipment pass will speed up field preparation compensating for the aforementioned loss of spring field days.
If high temperatures occur at anthesis (as predicted by GISS) then decisions as to choice of variety,
planting date, and planting density must be made in the spring. High temperatures during anthesis, especially
if plant water requirements cannot be met, would stress plants and increase the likelihood of poor pollination.
This would likely be due to lack of coordination between pollen shed and silk emergence. Losses in yield would
follow. In anticipation of such conditions, planting could begin earlier taking advantage of the earlier spring.
Thus anthesis would occur prior to the highest summer temperatures. Earlier warming in spring mentioned
above could facilitate this strategy provided that soil temperature thresholds for germination also occur earlier.
Because of the potentially longer growing season, farmers could resort more to the use of full season varieties
in their crop mix. It is likely that variety diversity will be maintained on any given farm leading to a mix of full
season and heat tolerant varieties. Cultivars now planted in the southern United States could become more
common in Illinois, for example.
Planting densities are already on the rise on most farms and could reach a maximum density of 30,000
seeds per acre. This trend will likely continue with or without climate change. The effect of higher densities may
be to make the crop more vulnerable to the dry conditions in mid-summer. If farmers alter planting density
practices at all it would likely be to decrease the rate of increase in planting densities. Lower planting densities
will decrease the competition of plants for available soil moisture.
Summer Anthesis. Once spring and early summer are past, there is little a farmer can do to affect
production until fall. However, if previous climate-dependent decisions leave the farmer's soils with less than
optimum nitrogen levels, sidedressing prior to and up to anthesis becomes an option (Figure 6). This is highly
contingent on a farmer having access to specialized equipment for sidedressing the growing crop (in some cases
up to the time of anthesis). There are indications that such technology is becoming more available and more
widely used.
10-20
-------
Spring Preplanting
and Planting
(March mid-April)
GISS climate:
varraer,
wetter
Easterling
Biophysical impact
Soils at field
capacity
Immediate agronomic
significance
Contingency for
decision response
> f
fever field days
days but
earlier planting
opportunities
Decision
response
(contingency
met)
N
application
not done
in fall
>
f
single pass
with field
cultivator in
early Harch
Planting begins
mid-March with
full season &
heat resistent
varieties
greater rate of
pesticide appli-
cation; more in-
tegrated pest
managemen t
Decision
response goal
take advantage
of possible
climate-enhanced
yields
1
I
rate of increase
in planting den-
sities decreased
/
\
V
r
Time anthesis
before high temps;
water conserve
Other
influences
on decision
response
Fluctuating N
prices; restric-
tive use trends
t
more
efficient
mechanization
/
t
(t = support decision response; may override or obscure)
(- = conflicts with decision response may override/offset)
Figure 5. GISS spring preplant and planting adjustments.
10-21
-------
Easterling
Summer
Anthesis
(July - August)
Climate
Change
Scenario
Biophysical
Impact
Immediate
Agronomic
Significance
GISS Climate:
Warmer, wetter
\u
Not
clear
^
Not
clear
Contingency
For Decision
Response
Decision
Response
Decision
Response
Goal
N
sidedressing "^>
capability, -'
•-- exists--
vi/
N sidedressing
Maximize N efficiency
if N is not fall
applied
Figure 6. GISS summer anthesis adjustments.
10-22
-------
Easterling
Fall Harvest. The GISS model projects a warmer, drier climate during the early fall harvest period. The
most significant biophysical consequence of this is increased evaporation (Figure 7). Harvest in Illinois would
likely occur earlier in the summer. This would complement the effects of earlier planting by allowing normal
or even more rapid and thorough drydown to take place. More rapid corn drydown would encourage earlier
harvesting. This would result in higher quality yields. More thorough drydown would also be likely to reduce
the need for expensive artificial drying procedures. Moreover, there are no readily identifiable factors that might
impede adoption of these practices. The incentive to reduce costs of fuel for artificial drying would actually
encourage adoption.
Analysis of the GFDL 2xCO2 Scenario
The relatively more severe climate changes suggested by the GFDL model require a slightly different
approach to the discussion of adjustment. The experts agreed that the hot and dry conditions of mid-summer
predicted in the GFDL 2xCO2 scenario would force corn producers to make major adaptations. In the logic of
this analysis, these decisions would occur in fall immediately after corn harvest.
Figure 8 is a schematic of fall post-harvest decision paths available to farmers under the GFDL 2xCO2
climate change scenario. Like the GISS scenario, the analysis begins with warm, dry conditions in fall likely to
lead to more fall field days. At this point, however, the GFDL-driven decision scenario departs significantly from
the GISS-driven scenario. Expectations in fall of the next season's hot, dry mid-to-late summer would likely
force fanners to choose among three initial alternatives. Two of the alternatives would involve abandoning corn
production totally. These alternatives represent major adaptations.
(1) Changing crops or land use. If climate changes were to shift comparative advantage to heat and low-
moisture hardy crops such as grain sorghum to Illinois farmers, then Illinois farmers could conceivably switch
to such crops. If productivity in these crops in nearby competing regions is unaffected or affected beneficially
by climate change, Illinois farmers could be forced either to leave agricultural production altogether or to
produce their corn for even smaller profits than now. In the former case, competing land uses, some
nonagricultural (e.g., urban encroachment), could force lands, especially marginal lands, out of production. If
corn production is to continue, however, farmers would almost certainly be forced to grow shorter season, lower
yield, varieties. These varieties could probably be intercropped with a heat and moisture stress resistant crop
such as grain sorghum.
(2) Irrigation. Irrigation is a potential adaptive decision. Irrigation may not be widery applicable in Illinois
since groundwater availability and soil moisture holding characteristics may not be suitable in many parts of the
state. For example, some soils in southern Illinois contain too much day for effective irrigation (water simply
runs off). Also, at current production costs and crop prices the operation of irrigation systems may not be
economical in Illinois, even under climate change conditions. Adams (this volume) suggests sizable price
increases for corn under the GFDL scenario. If he is correct, experts felt it would be reasonable to expect some
increase in irrigation in Illinois under the GFDL scenario. But not all or even most farmers would be likely to
acquire and operate irrigation systems.
(3) Continuing with corn. If a farmer decides to continue raising corn, especially if he uses short-season
varieties possibly intercropped with grain sorghum, then the most important decision in fall would be the amount
of acreage to be tilled in preparation for spring planting. The experts agreed that under the GFDL scenario,
more marginally productive land would likely be taken out of crops-a logical response to anticipated higher
costs and reduced productivity on these lands. Since the GFDL and GISS 2xCO2 winter and spring temperature
and precipitation scenarios are generally similar, other fall post-harvest decisions would likely be as described
under the GISS results (refer to discussion in the previous section).
In connection with their implications for continued production of corn, the GISS and GFDL 2xCO2
scenarios differ most in the summer. On the one hand, the GISS model projects higher temperatures and higher
rainfall than present during the anthesis and grainfill periods (on the dates that such periods presently occur).
10-23
-------
Easterling
Fall Harvest
(Late August -
Hid October)
Climate
Change
Scenario
Biophysical
Impact
Immediate
Agronomic
Significance
Contingency
For Decision
Response
Decision
Response
Decision
Response
Goal
GISS Climate:
Warmer, drier
High evaporative loss
from plants & soil
earlier, faster
dry-down
Planting
as earl
Early harvest, less
artificial drying
Avoid field losses,
minimize fuel expenses
for drying
Figure 7. GISS fall harvest adjustments.
10-24
-------
Fall Corn Post-Harvest (GPDL)
Climate characteristics
of current season
warmer, dryer
post-harvest
fall
immediate agronomic
significance
I
Easterling
more field
days are
likely
contingencies
relevant to
current decision
mid to late summer
expected to be too
hot and dry for
current corn varieties
I
\
winter to mid
expected to
warmer and vt
than now
f
spring
be
Jtter
adaptation
decision
(interannual)
continue to
grow corn?
NO
YES
implement ir-
rigation where
possible
adjustments
(intrannual)
/I
switch to inter-
cropped soybeans
grain sorghum
rotation
fill market
opportunity for
crops being
forces out of
of production
elsewhere
decision incentive
1
off-farm influences
on decisions
short season
corn varieties
mixed with soy-
bean and sorghum
marginal
acreage
taken out
of production
4
more intense con-
servation tillage
but less absolute
acreage tilled &
less fall-applied
nitrogen
speed up spring
planting and
minimize winter
nitrogen loss
climate effects in
other production
regions as related
to prices
far less
tillage
far less
nitrogen
resource conser-
vation policies
& environmental
protection policies
Figure 8. GFDL fall post-harvest adaptations and adjustments.
10-25
-------
Easterling
On the other hand, GFDL projects the anthesis and grainfill periods to be much higher with lower rainfall than
present. Even taking into account differences between these models, farmers' spring decisions will aim to
minimize the negative impacts of a wet spring and avoid the high temperatures of the anthesis-grainfill period.
Most GISS strategies for dealing with the high temperatures in mid to late summer (e.g., earlier planting, lower
population densities, more varietal and crop diversity) would also apply to the GFDL scenario of warmer and
significantly drier summer conditions.
One major production adjustment that could be considered under the GFDL scenario would be to hold
nitrogen application levels constant or even to reduce nitrogen applications in spring and summer as well as fall
The incentive to do this is twofold: (1) reduced summer rainfall would result in less nitrogen leaching, thus less
total nitrogen required; and (2) lower yield expectations would encourage cost minimization. A reduction in
nitrogen usage might help reduce costs.
With earlier planting and increased reliance on shorter season varieties, harvest could occur even earlier
than under the GISS scenario. Aside from the probability of harvests as early as July (for short-season varieties)
under the GFDL scenario, the same harvest adjustments would apply for both climate change scenarios (i.e., less
artificial drying).
(4) Farm abandonment. Returning to the fall post-harvest adaptive decisions -- if the corn grower decides
to quit agricultural production altogether, then the analysis ends at that point. This is obviously an extreme
action, one that is not likely to occur often. However, if many make this decision the impacts will be felt far
beyond the farmgate. It is beyond the scope of this study to consider these impacts. Research is under way at
Resources for the Future to consider climate change impact linkages between the farm and market levels.
Changes in cultural practices that might accompany crop changes. If the decision made in fall is to switch
to a different cropping pattern, then the farmer will be faced with a new set of production decisions. Double
cropping, defined as the following of one crop with another crop within a growing season, is practiced in
southern Illinois. For example, wheat is followed by soybeans on many southern Illinois farms. Double cropping
is seen as maximizing the economic potential of the land, given agronomic constraints such as soils, climate, input
costs, market prices, and so on. Double cropping is not widespread in Illinois because only a limited part of the
state experiences a long enough growing season to support two crops. The areas in Illinois where
wheat-soybeans double cropping is practiced are often characterized by poor soils and a longer growing season.
Wheat-soybeans double cropping is practiced in other proximate regions south of the major corn
production area of Illinois. To an extent, these proximate regions currently experience temperatures approaching
the GFDL 2xCO2 temperatures for central Illinois. Thus it is tempting to suggest that wheat-soybeans double
cropping would supplant corn in Illinois as temperatures warm. However, such double cropping will likely not
be widespread in Illinois under the GFDL scenario. Dry and hot summer conditions would deplete soil moisture
under the wheat crop to the extent to be insufficient to support the soybean crop.
The cropping pattern that experts felt to be most likely to replace corn would be straight soybeans or
possibly soybeans intercropped (grown at the same time) with grain sorghum. Both of these crops have higher
moisture and heat stress tolerances than corn. Soybeans are routinely grown in rotation with com by many
Illinois farmers. Moreover, grain sorghum production practices are similar to corn production practices. Thus
the switching to some combination of soybeans and grain sorghum would require less new learning and
reequipping on the farm than would switching to another crop such as wheat.
Experts reasoned that with projected increases in prices for soybeans under the 2xCO- GFDL climate
(predicted to rise by more than 200% over today's prices), farmers would find soybeans to oe profitable in
Illinois even given major predicted yield decreases. Such was also the case for grain sorghum.
10-26
-------
Easterling
SUMMARIZING ADJUSTMENTS: COMPOSITE PORTFOLIOS
It is useful to compile from the previous analysis a comprehensive set of on-farm adaptations and
production adjustments (call them - a portfolio). Such a portfolio would reflect all adjustments that a farmer
might make to optimize his operations in response to climate change. Influences on adjustment decisions other
than climate (as discussed above)
were not considered in building the portfolio. This illustrates what steps could be taken on the farm, all things
being equal, to adjust production to climate change.
Table 3 shows adjustments by seasons based on the GISS 2xCO2 climate change scenario. The bulk of
adjustments will occur during the fall post-harvest and the spring pre-plant and planting periods. Only one
adjustment option is available to farmers after planting and before harvest, i.e., nitrogen sidedressing.
types
Table 4 shows adjustments and adaptation that may follow the GFDL 2xCO2 warming. Note that the
.,..__ of adjustments to farming operations will follow somewhat from choice of adaptation. For example, if
farmers choose to continue with growing corn (but they adapt with short-season varieties) then nitrogen will still
be applied although such application might be eliminated in fall (adjustment) and reduced in spring and summer
(adjustment).
These scenarios, though qualitative, could provide information for crop simulation models on how to
manipulate certain variables (e.g., varieties, planting dates, planting densities). This would allow alternative
on-farm adjustments and adaptations to be reflected in the crop models. The result would be crop model
estimates of productivity, modified by management response, as an alternative to productivity estimates generated
by simply varying climate and some other less systematically chosen model components. Moreover, as the
adjustments analysis is refined, the reliability of estimates of crop yield response to climatic change would
improve.
10-27
-------
Easterling
Table 3. Composite Adjustments to Optimize For Climate Only: GISS Scenario
Season Adjustment
1. Fall Post harvest
(October-December)
increased tillage
(mostly conservation)
increased application of
phosphates and potash
reduction or elimination
of fall-applied nitrogen
Winter
(December-March)
none
Spring Pre-plant and
plant (March-April)
reduction of field preparation
to single pass with field
cultivator (early March)
multiple varieties but more
full-season and heat resistant
(planting begins mid-March)
slower rate of increase in
planting densities
none nitrate fertilizer applied
more pesticides applied
(some increase in integrated
pest management)
Late Spring-Early Summer
(Mary-June).
none
Summer Anthesis
(late June-July)
nitrate fertilizer sidedress
(when possible)
6. Late Summer
none
7. Fall Harvest
(September-October)
earlier harvest
less artificial drying
10-28
-------
Easterling
Table 4. Composite Adjustments to Optimize For Climate Only: GFDL Scenario
Adaptations
Alter crops planted to a mix of
short season corn varieties and
soybeans/grain/sorghum which are
better adapted to heat and moisture
stress than full season corn
Plant only on best land (reducing
total acreage planted) to -minimize
climate stresses on productivity*
Establish or expand irrigation when
water supplies exist and soils are
amenable to irrigation in order to
moisture and heat stress in
mid-summer
Adjustments
Increased fall conservation tillage
to shorten field preparation activ-
ities in spring
Decreased fall applied nitrogen to
reduce winter leaching and
denitrification
Earlier planting to avoid mid-
summer high temperatures and
aridity
Lowered rate of increase in offset
planting densities to decrease
competition for moisture
Increased pesticide applications
and integrated pest management
practices) to control warming-
induced insect and weed population
expansions
Less nitrogen applied in spring
partly to reflect lowered nitrogen
leaching in a drier summer and
partly to reflect need to minimize
input costs
Earlier harvest and less artificial
drying in order to benefit from
warm and dry late summer and early
If input costs (e.g. fertilizer, pesticides, seeds) rise, then more land,
not less, might be brought into production as a substitute for more expensive
inputs (Crosson, 1986).
10-29
-------
Easterling
CHAPTERS
INTERPRETATION OF RESULTS
From the foregoing discussion, it is clear that climatic changes occurring within the bounds of the GISS
and GFDL 2xCO-model runs will affect Illinois corn production. Though no attempt is made here to predict
the economic feasibility of corn production in Illinois after climate change, I argue that farmers will try to adapt
and adjust to climate as it changes. Moreover, knowledge of how such adjustments and adaptations might take
place is necessary to calculate the net effect of climate on farmers and their crops. In the following paragraphs,
I provide a set of conclusions on farmer responses to climatic change which are based on the above analysis.
Such responses are divided into categories of general conclusions, specific adjustment possibilities, and specific
adaptation possibilities.
General Conclusions
1. Adjustments to existing production practices are the first line of defense against climate change.
Production practices currently in use by corn producers can be altered with little cost to the fanner to cope with
some types of climatic fluctuation.
Examples of such adjustments include altering planting dates, crop mixes, tillage practices, and fertilizer
applications (these are elaborated below). This analysis suggests the possibility that, in some scenarios of climate
change, such adjustments could totally eliminate negative climate impacts or even create a positive growth
situation. The net effect of climate change on corn production after adjustment(s) are made by farmers will be
significantly different from the effect of climate change without adjustments.
2. More severe climatic changes are likely to make major adaptations necessary. If climatic conditions
were to become as hot and dry during the summer growing season in Illinois as projected by the GFDL model,
then farmers who are used to raising full-season corn will be forced to adopt production practices quite different
from those in use today. Such adaptations would be necessary to deal with the potentially high heat and
moisture stress conditions likely to accompany such climatic changes. Examples of adaptations that are likely
to be made include installation and use of irrigation systems; changes in crop mix to include some combination
of short- season corn, soybeans, and grain sorghum; and removal of some marginal land from production on a
given farm.
3. Climate changes in nongrowing season parts of the year will be important, and nongrowing season
adjustments are likely to follow. Climate change will occur in all seasons. Indeed, in the Northern Hemisphere
mid-latitudes, changes may be most dramatic in winter. Climate changes in the nongrowing season (fall
post-harvest and winter) will affect some adjustment decisions. For example, a wanner and wetter winter would
increase nitrogen leaching and denitrification, thus leading to a likely reduction of pre-growing season fall-applied
fertilizer.
4. Changes in regional comparative advantage will determine the extent of adjustment versus adaptatioa
The effect of climate on agricultural regions competing with Illinois will directly influence how adjustments and
adaptations are made by Illinois producers. Economics apart, neither the GFDL nor the GISS 2xCO2 climate
scenarios would preclude corn from being raised in Illinois if appropriate adjustments (e.g., earlier planting,
altered tillage) and adaptations (e.g., irrigation, short-season varieties) are made. However, when economics
come into play, many of these adjustments may raise production costs high enough to make corn production
unprofitable relative to other regions or other crops. On the one hand, if other production regions are similarly
adversely affected by the climate change and world demand for food and fiber cannot be met, then market prices
could rise. Such price increases would encourage more extensive efforts to adjust and adapt production to the
climate change. On the other hand, if comparative advantages begin to weigh more heavily in favor of other
regions as climate changes, prices do not go up, and necessary adjustments by Illinois farmers become too costly,
the only strategy left for Illinois farmers is to leave farming altogether or, more likely, to choose a new crop to
10-30
-------
Easterling
grow.
If prices for crops uniformly rise, as predicted by Adams (this volume), and input prices do not rise
significantly, then adjustments are likely to be extensive in Illinois. I consider the sorts of specific adjustments
and adaptations that might be in store in the following two sections.
Specific Adjustment Possibilities
A number of adjustments to existing production practices could be made by Illinois corn producers in
response to changes in climate:
1. Changes in tillage practices. This research points to increased fall conservation tillage in Illinois. This
would offset the negative effects of a smaller possible number of field days in the following spring planting
season and would take advantage of dry conditions in the fall season. Though this response may not, by itself,
have a large effect on yields, it will contribute in a significant way to how farmers may respond.
2. Less fall-applied nitrogen. Questions aside about the continued production of corn in Illinois under
climate change, less nitrogen is likely to be applied in fall. If nitrogen-fixing legumes such as soybeans replace
corn in Illinois, then less nitrogen use needs no explanation. However, if corn and other crops which benefit
from nitrogen application continue to be grown, then the explanation for less fall-applied nitrogen is the desire
to avoid leaching and denitrification during the warmer (and possibly wetter) winter.
3. Planting earlier in the spring. It will benefit fanners to plant corn earlier in the spring than at present.
Earlier planting will result in the occurrence of anthesis (a period of high temperature and moisture stress
sensitivity) earlier in the summer when temperatures and soil moisture levels are less likely to be at stress levels.
Earlier planting could be enhanced by warmer spring temperatures. However, the warmer spring temperatures
could be offset somewhat by the greater heat storage capacity of a wetter soil (wetter from higher than current
precipitation in winter), i.e., as soil moisture increases, more energy is needed to raise the temperature of the
soil.
4. Changes in corn varieties. Climatic change will encourage farmers to alter the com varieties planted
either to maximize potentially beneficial climate changes or to minimize potentially negative climate changes.
If planted earlier than at present, there is a possibility that high yielding, full-season varieties would do well
under the GISS 2xCO2 summer climate. Also, if planted earlier than now, low yielding short-season corn
varieties could be grown under the relatively harsh GFDL 2xCO2 summer climate.
5. Lower planting densities. Potentially drier soils during summer will likely lead to changes in planting
densities. Lowering the number of plants per unit area would conserve soil moisture. Given that planting
densities have been increasing steadily in recent years, it is likely that densities will be higher than at present but
not as high as they would have been without climate change.
6. More attention to pests. Warmer temperatures are likely to stimulate greater pest problems (this is
handled more extensively by Skinner et aL, this volume). It would be reasonable to expect farmers to respond
with some increase in pesticide use but perhaps accompanied by greater emphasis on integrated pest
management practices.
7. Earlier harvest. If planting does occur earlier in the spring than now, it follows that harvest will begin
earlier than now. This is especially true if short-season varieties were planted in spring. Also, changes toward
wanner and drier late summers and early falls are likely to favor earlier corn maturation and to accelerate
drydown. Not only would farmers be able to harvest earlier than now, but there is likely to be less need for
artificial drying.
10-31
-------
Easterling
Specific Adaptation Possibilities
1. Altered crop mixes. Depending upon the exact nature of climatic change, fanners are likely to change
not only the varieties of corn to be raised, but also to alter the types of crops grown. Crop mixes could become
more diverse with likely combinations (either intercropped or annually rotated) of corn, soybeans, and grain
sorghum. If corn is grown under the GFDL 2xCO2 scenario, then almost certainly it will be a short-season
variety.
2. Increased irrigation. According to the experts, current climatic and economic conditions would support
use of irrigation in Illinois approximately 3 years out of 10 years. Climatic changes, especially changes which
leave soils drier during summer than now, will likely lead to increased use of irrigation in Illinois. Increased use
of irrigation is supported by projected price increases for all crops likely to be grown in Illinois. However, not
all farmers can be expected to turn to irrigation as climate changes. Insufficient water supply and nonirrigable
soils will limit where irrigation may come into practice. Moreover, even with severe climatic changes, some crops
may be profitably grown without irrigation (e.g., soybeans, grain sorghum).
3. Marginal land taken out of production. If climatic changes cause production costs to rise in Illinois,
there is likely to be a corresponding decrease in total acreage planted on a given farm. Poorer quality land (less
fertile, highly erosive) on any farm is likely to be taken out of production. Such action would tend to help hold
the line on reductions of productivity per unit of land.
10-32
-------
Easterling
CHAPTER 6
POLICY IMPLICATIONS FOR EPA
I take as given that EPA is most concerned about implications of climatic change for managing
environmental resources. Adjustments and adaptations to climatic change by Illinois fanners may raise some
environmental concerns. Perhaps one way to target potential topics for further policy analysis is to identify those
adjustments and adaptations which interact directly (either in a negative sense or a positive sense) with processes
resulting in changes hi environmental quality. Two such topics covered below are soil erosion and water quality.
Soil erosion. Soil erosion leads to two major types of environmental degradation. First, topsoils are
transported off the farm leaving behind thinner soils. Second, soils that are transported off the farm can be
deposited in places which cause environmental problems such as silting in of reservoirs and shipping channels
and altered groundwater characteristics.
There are two climate change adjustment/adaptation strategies that may directly affect soil erosion.
Increased tillage in fall, even if restricted to conservation tillage practices, will leave soils more susceptible to soil
erosion during whiter and early spring. This problem may be exacerbated somewhat by a warmer, wetter whiter
which leaves soils unfrozen and more exposed to water transport from runoff.
Reduction of planted acreage will result in less productive marginal lands being taken out of production.
Marginal lands seem to be most susceptible to soil erosion. Thus, the fallowing of marginal kinds could act to
slow down rates of erosion. Moreover, such fallowing may serve to offset the negative effect of increased fall
tillage on soil erosion. It is not clear what the net soil erosion outcome will be from increased tillage. However,
there should be some inclusion of how climate change might affect tillage practices hi the development of
follow-on policies to soil conservation concerns such as T-by-2000.
Water quality. Adequate water quality is at risk from current agricultural practices in Illinois. Certain
agricultural adjustment/adaptation strategies for coping with future climatic change, if such strategies are widely
practiced, could worsen the situation. Such strategies deemed to be important hi the possible degradation of
water quality include increased irrigation and increased pesticide use. Another strategy, changes hi nitrogen
fertilizer application practices, is likely to affect water quality, but it is not certain whether effects will be positive
or negative (more on this below).
If climate warming and drying is severe enough to warrant increased irrigation, then it can be concluded
that surface water and groundwater supplies will be reduced to some extent. Some of this reduction would come
from increased evaporation together with decreased precipitation. Withdrawal of water for irrigation from
surface and groundwater sources would further reduce supplies.
Reduced water supply will certainly decrease water quality. With less total water supply, water which is
left has less capacity to assimilate toxic substances. Effluents of any land, agricultural or otherwise, will tend to
be hi higher concentrations hi water bodies, all things being equal, if the amount of water hi those bodies is
reduced. Moreover, reduced water supply would probably lead to other problems like disrupted navigation of
rivers, drying up of shallow wells, and reductions of water use to safe yield ratios for major urban water supply
systems.
Pesticide residues hi farm runoff lower water quality. If pesticide application rates were to increase hi
response to climatically induced pest population increases, then more pesticide residues will enter groundwater
and surface watersheds. Once hi a water system, pesticide residues can accumulate to levels that are toxic to
aquatic life and to other animals that consume aquatic life, including humans.
10-33
-------
Easterling
Nitrogen-based fertilizers pose health risks to humans when they enter public water supply systems. In
agricultural regions where nitrogen is applied to crops in large amounts, toxic nitrogen compounds have shown
up in public water supply systems within those regions. At times, concentrations of nitrogen compounds have
become great enough to force temporary closings of municipal water supply systems.
Climatic changes which bode well for corn production could provide incentive to increase fertilizer
applications. The rationale behind such an increase has been discussed above. If more nitrogen fertilizers are
applied to crops, there is likely to be more risk of contaminated public water supply systems in the regions where
the nitrogen application is
taking place.
Climatic changes leading to hotter and drier summer conditions will bode poorly for corn production.
However, drier conditions than now would reduce nitrogen leaching from soils and into groundwater. Reduction
of nitrogen loss in soils could lead to less nitrogen reapplication. A reduction of reapplication rates might also
be helped by generally poorer yield expectations by farmers (discussed above). The net effect of a reduction of
nitrogen application would be to lessen the amounts of toxic nitrogen compounds that enter public water supply
systems to nontoxic levels.
The above issues of water quality changes as the result of climatically induced adjustments/adaptations
on the farm are not well understood. Indeed, none of the potential environmental problems from adjusting and
adapting to climate change that I have raised above may materialize. Nevertheless, the possibility of occurrence
of these problems is sufficient to raise concerns. Perhaps the best way to manage the possible environmental
effects of future climate adjustment/adaptation strategies at the farm level is to gather more facts about how
fanners will make decisions in the face of climatic change. More sophisticated assessments are needed of the
integrated climate-farm-market system. We cannot assess the potential environmental damages of future farming
practices until we know how these practices will differ from today with or without climate change.
Acknowledgments
i wish to extend my appreciation to the Illinois State Water Survey where this effort was initiated I also
wish to thank Dennis Smith for his able assistance in the compilation and analysis of the data, Angela Blake for
her help in word processing, and Norman J. Rosenberg for his review of this research.
10-34
-------
Easterling
REFERENCES
Ausubel, J. H., 1983. Can we assess the impacts of climatic changes? Climatic Change, 5, 7-14.
Ausubel, J. H. 1980. "Economics in the Air," in J. H. Ausubel and A. K. Biswas, eds., Climatic Constraints and
Human Activities, IIASA, Oxford, Pergamon Press.
Biasing, T. J. and A. Solomon. 1981 Response of the North American Corn Belt to Climate Warming.
Prepared for the US. Dept. of Energy, Springfield, Virginia, NTIS.
Brewer, G. D. 1986. "Methods for synthesis: policy exercises", chap. 17 in W. C. Clark and R. E. Munn (eds.)
Sustainable Development of the Atmosphere, Cambridge: Cambridge University Press, pp. 455-474.
Ciborowski, P. and D. Abrahamson, 1986. The Granary and the Greenhouse Problem," in C.F. Runge, ed., The
Future of the North American Granary: Politics, Economics, and Resource Constraints in North American
Agriculture, Iowa State Univ. Press.
Clark, W. C., 1986. Chapter 1: "Sustainable Development of the Biosphere: Themes for a Research Program,
in W. C. Clark and R. E. Munn, eds., Sustainable Development of the Biosphere, Cambridge: Cambridge
University Press, 491 pp.
Clark, W.C., and R. E. Munn. 1986. Sustainable Development of the Biosphere. Cambridge: Cambridge
University Press, 491 pp.
Coulson, R., L. Folse, and D. Loh. 1987. "Artificial Intelligence and Natural Resource Management," Science
vol. 237, 262-267.
Decker, W., V. Jones, R. Achutuni. 1986. The Impact of Climate Change from Increased Atmospheric Carbon
Dioxide on American Agriculture", US. Dept. of Energy, Springfield, Virginia, NTIS.
Easterling, W. E. and J. Mjelde. 1987. The Importance of Seasonal Climate Prediction Lead Tone in
Agricultural Decision Making" Agricultural and Forest Meteorology.
Easterling, W.E., J. Mjelde, and S.T. Sonka. submitted. "Broadening Possibilities for Artificial Intelligence in
Geographic Research: An Expert Systems-based Analysis of Agricultural Sensitivity to Climatic Variability.
Submitted to Annals of Assoc. of American Geographers.
Glantz, M. J. 1988. The Value of a Long-Range Weather Forecast for the West African Sahel," Bulletin of the
American Meteorological Society vol. 58, pp. 150-158.
Hayes-Roth, F., D. Waterman, D. Lenat. 1983. Building expert systems, London: Addison-Wesley.
Kates, R., J. Ausubel, M. Berberian. 1985. Climate impact assessment: Studies of the interaciton of climate
and society, New York, Wiley and Sons.
Lamb, P. J. 1987. On the development of regional climatic scenarios for policy-oriented climate-impact
assessments. Bull, of American Met. Soc. 68(a), 1116-1123.
Lovell, AA^ and V.C. Smith. 1985. "Chapter 12: Microeconomic Analysis." In: Kates, R.W., J.H. Ausubel, and
M. Berderian, eds. Climate Impact Assessment: Studies of the Interaction of Climate and Society, Scope 27,
New York: John Wiley & Sons, pp 293-322.
10-35
-------
Easterling
Mjelde, J. W. 1985. "Dynamic Programming Model of the Corn Production Decision Process with Stochastic
Climate Forecasts" (unpublished Ph.D. dissertation, University of Illinois).
Parry, M. L. 1986. "Some Implications of Climatic Change for Human Development," in W. Clark and R.
Munn, eds., Ibid.
Parry, M. L., T. Carter, N. Konigen (eds.) 1988. Assessment of Climatic Variations on Agriculture, Volume 1
Assessment in Cool Temperature and Cold Regions (The Netherlands, Reidel).
Easterling, W. E., M. L. Parry, and P. Crosson. 1989. "Adapting agriculture to climate change" in N. J.
Rosenberg and W. E. Easterling (eds.) Greenhouse Wanning: Abatement and Adaptation, Washington, D.G,
Resources for the Future.
Rosenberg, N. J. 1982. The increasing CO2 concentraion in the atmosphere and its implication on agricultural
productivity. II. Effects through CCyinduced climatic change. Clim. Change 4, 239-254.
Rosenberg, NJ. 1986. Climate Change: A Primer, Resources for the Future, Washington, D.C.
Rosenberg, NJ. 1987. "Drought and Climate Change: For Better or Worse?" in D. A. Wilhite and W. E.
Easterling, eds., Planning for Drought: Toward a Reduction of Societal Vulnerability (Boulder, Colorado,
Westview Press).
Rosenzweig, C. 1985. "Potential CCyinduced climate effects on North American wheat-producing regions,"
Clim. Change 7(4), 367-390.
Sonka, S.T. and P. Lamb. 1987. "On climate change and economic analysis," Clim. Change, 11(3), 291-312.
Warrick, R. 1984. "The possible impacts on wheat production of a recurrence of the 1930's drought in the US.
Great Plains," Clim. Change 6(1), 5-26.
Wilhite, D. A., and W. E. Easterling. 1986. Planning for Drought: Toward a Reduction of Societal
Vulnerability (Boulder, Westview Press).
Winkler, R., A. Murphy, R. Katz. 1983. The value of climatic information: A decision analytic approach," J.
of Climatology.
10-36
-------
CHANGING ANIMAL DISEASE PATTERNS INDUCED BY THE
GREENHOUSE EFFECT
by
Edgar Stem
Gregory A. Mertz
J. Dirck Stryker
Monika Huppi
Tufts University
North Grafton, MA 01536
Purchase Order No. 8W-3651-NTEX
-------
CONTENTS
FINDINGS 11-1
Objectives, Methods, and Report Presentation Format 11-3
Economic Methodology. 11-4
Methodology for Determining Predictions of Future Disease and Disease
Vector Distributions 11-5
General Changes Likely To Be Seen as a Result of Greenhouse Gas-Induced
Global Climate Changes 11-5
CHAPTER 1: BLUETONGUE 11-10
Introduction 11-10
Clinical Signs, Epidemiology, and Vector Biology of the Disease 11-10
Economic Importance of Bluetongue 11-11
Control of Bluetongue 11-11
Possible Effects of Global Climate Change on Bluetongue 11-13
Economic Model of the Effects of Global Climate Changes on Bluetongue 11-14
CHAPTER 2: ANAPLASMOSIS 11-16
Introduction 11-16
Clinical Signs, Epidemiology, and Vector Biology of Anaplasmosis 11-16
Economic Importance of Anaplasmosis 11-17
Control of Anaplasmosis 11-17
Possible Effects of Global Climate Change on Anaplasmosis 11-18
CHAPTER 3: RIFT VALLEY FEVER 11-19
Introduction 11-19
Clinical Signs, Epidemiology, and Vector Biology of Rift Valley Fever 11-19
Economic Importance of Rift Valley Fever 11-21
Control of Rift Valley Fever 11-21
Possible Effects of Global Climate Change on Rift Valley Fever 11-21
Economic Model of the Effects of Global Climate Changes on Rift
Valley Fever 11-25
CHAPTER 4: AFRICAN SWINE FEVER 11-26
Introduction 11-26
Clinical Signs, Epidemiology, and Vector Biology of the Disease 11-26
Economic Importance of African Swine Fever 11-27
Control of African Swine Fever 11-27
Possible Effects of Global Climate Change on African Swine Fever 11-27
Economic Model of the Effects of Global Climate Changes on African
Swine Fever 11-29
Suggested Topics For an In-depth Study on the Effects of Greenhouse-Induced
Climate Changes on Pattern of Animal Health 11-29
REFERENCES 11-32
ANNEX A: ECONOMIC TABLES 11-38
-------
Stem
FINDINGS1
The United States today produces about 20 million metric tons of red meat annually. This is almost twice
the output of its nearest competitor, the Soviet Union (McCoy and Sarhan, 1988). In addition, dairy products,
skins and hides, wool, and other animal products contribute to the value of production in the livestock sector.
About two-thirds of the agricultural land in the United States is devoted to grazing. Much of this land
has no alternative food production use. In addition, livestock utilize enormous quantities of roughage and crop
residues that would otherwise be wasted. The number of people employed in livestock production, processing,
and marketing may total 10 million (McCoy and Sarhan, 1988).
Animal production tends to be in the western United States, close to the native forage of the plains and
mountain states and to the land under cultivation in the mid and far west. Hog production is more concentrated,
mostly in the north central states, than is that of cattle and sheep. More than three-quarters of the sheep are
raised in the west, where plentiful feed sources are available.
Increased incidence of animal disease as a result of global climate wanning due to the greenhouse effect
has serious implications for the U.S. and world populations as a whole, but especially those directly involved in
animal agriculture. Many animal diseases are kept in check largely because of the restrictions climate places
on vectors, environmental habitats, and the disease-causing agents themselves (Blood et al., 1983; Gillespie and
Timony, 1981; and Fraser, 1986).
Changes in animal health patterns can have important effects upon the quality of human life:
economically, owing to increased price and decreased availability; nutritionally, through decreased consumption
of animal products; physically, through changing zoonotic disease patterns; and socially, through the effects of
disease states of companion and sport animals.
A detailed study is necessary to comprehend and elaborate the range of important changes that could
result from climate changes induced by an accumulation of greenhouse gases.
Animal diseases that are likely to become increasingly important as the earth warms, due to the
accumulation of greenhouse gases, fall into several categories: vector-borne diseases, zoonotic diseases, and
foreign animal diseases.
The vector-borne diseases such as anaplasmosis, bluetongue, babesiosis, and Lyme disease may become
increasingly important as larger areas of the United States are likely to become suitable habitats for the vectors
of these and other diseases.
Zoonotic diseases such as Eastern equine encephalitis, intestinal nematodes such as canine hookworm and
roundworms, Q fever, toxoplasmosis, and aflatoxicosis will probably also become more important as the
environmental changes permit longer seasonality of these diseases and as their animal or environmental reservoir
grows. Poultry diseases such as salmonellosis, coctidiosis, influenza, and New Castle disease may also be
affected.
Foreign animal diseases, which even today pose an enormous threat to our animal industry, could become
more important if movements of people and animals increase, taxing the control mechanisms that have been
'Although the information in this report has been funded wholly by the US. Environmental Protection
Agency under Purchase Order No. 8W-3651-NTEX, it does not necessarily reflect the Agency's views, and no
official endorsement should be inferred from it.
11-1
-------
Stem
put inio place by the USDA's Animal and Plant Health Inspection Services (APHIS). Diseases such as foot and
mouth disease can spread through the action of wind and move across a region, infecting farm animals Math
lightning speed.
Vector-borne foreign animal diseases including African swine fever and Rift Valley fever could have an
even greater chance of becoming established, once introduced, because the habitats and sizes of vector
populations might increase owing to the climate changes induced by the greenhouse effect. With greater
numbers of vectors dispersed through larger geographic areas, the chances of introduction of a foreign animal
disease into the American vector reservoir will increase proportionately.
Control of these and other arthropod-borne diseases will be increasingly difficult; as environmental
temperatures increase, the generation time of these populations decreases, which has the effect of increasing the
rate of arthropod evolution. Resistance to pesticides and possibly other control methods will evolve more
rapidly, frustrating control efforts and making these diseases even more important economically.
Based on the epidemiology of the disease and the disease agent and vector biology, we chose four vector-
borne diseases as examples of diseases whose distributions are likely to change owing to the greenhouse effect.
Two of these diseases, anaplasmosis and bluetongue, are important diseases with wide distribution in the United
States today. The remaining two, African swine fever and Rift Valley fever, although presently exotic to the
United States, are likely to become more of a threat to American animal agriculture as climate changes induce
an expansion of American vector habitats for these diseases and as movements of animals and people increase,
precipitated in part by the predicted environmental and climate changes. As discussed on the preceding page,
many other diseases will likely be affected. Detailed discussion of all of these diseases is beyond the scope of
this preliminary study.
We draw the conclusions outlined below as a result of the literature search for the diseases that we have
examined in depth and our Firsthand experience with these diseases in the United States and throughout the
world. Detailed discussions are presented in the specific sections for each disease.
Bluetongue As vector overwintering and year-around vector reproduction become possible in more states
and the average lifespan of adult midges increases, we expect to see the geographical distribution of this disease
expand northward and eastward. These predictions were made based on the known vector biology, nature of
the bluetongue virus, and predicted climate changes as described by the GISS and GDFL models. These factors
are described in detail in the bluetongue section of this report. The states in which bluetongue occurs in
sporadic epizootics today might become endemic states for this disease as the climate warms. States that are
now free from the disease could become areas that will experience occasional outbreaks (Figures 2 and 3). It
is not very likely that the predicted climate changes will result in a significant vector habitat loss for the
bluetongue vector. Bluetongue and its vector Culicoides currently are highly prevalent in some of the world's
warmest and driest regions, including the African Sahel.
Anaplasmosis Likewise, as the temperature of northern states increases and, in particular, winters become
milder than at present, and based on the known vector biology as presented in today's scientific literature, we
predict that anaplasmosis might take on increasing importance as a disease of serious economic implications to
U-S. animal agriculture (it is presently ranked second in animal diseases of economic importance) (Goodger et
al., 1979).
Although the ticks that transmit anaplasmosis have feeding and reproductive habits largely based upon
changes in day lengths, the tick numbers may well increase and their range may expand in northern states owing
to less severe winter climates. In the south central states, where decreased precipitation and relative humidity
levels are anticipated, it is possible that this disease could decrease in importance — except where rangelands are,
or become, a mixture of bushland and shrubs. The micro-environments surrounding bushes and small trees
provide sufficient levels of relative humidity to keep these ticks from dying because of desiccation.
11-2
-------
Stem
Furthermore, in the Midwest, if the climate warms and precipitation decreases, it is possible that crops
such as com will become less productive, resulting in the transformation of some of these marginal cropping
areas to rangeland. This could make a disease such as anaplasmosis of greater importance in these areas.
Anaplasmosis is a disease that, once established in a herd, is clinically worse in adult animals. Stress,
including that precipitated by high environmental temperatures, increases associated morbidity and mortality
rates. Thus, the frequency of outbreaks of anaplasmosis and the resulting mortality and production losses will
likely increase in the southern states, particularly during summer months as temperatures increase.
African Swine Fever As the temperature increases in the north-central and southeastern states, the tick
carrier or vector for African swine fever could potentially extend its geographical habitat (Figures 6 and 7). If
this disease is introduced into the United States, this increase in vector distribution and numbers would increase
the risk of establishment of the disease. This is particularly important for African swine fever because it is
thought that once this disease has become established in the soft tick vector population, it may be impossible to
eradicate (Butler and Gibbs, 1984).
The chances that such a disease might be introduced into the United States will certainly be increased if
geographical re-location of both human and animal populations occurs as a result of environmental changes
that may be precipitated by the accumulation of greenhouse gases. This will necessitate increased efforts by
the U.S.D A.'s Animal and Plant Health Inspection Service (APHIS) to prevent the entry of African swine fever
into the south.
Rift Vallev Fever As with African swine fever, the greenhouse effect will have an impact on the possibility
of introduction of this disease in two ways. First, if environmental change stimulates an increase in the
movement and relocation of people and animals, there will be an increased risk of introduction of disease. This
is particularly true for Rift Valley fever, because as a disease which affects humans, it is quite possible that
human carriers of the virus could act as sources for introduction of the disease into our animal population.
Secondly, this disease is transmitted principally by mosquitoes, and at least three mosquito species widely
distributed in the United States today are capable of doing so. Warmer temperatures, less severe winters, and
precipitation increases all will contribute to increased mosquito populations and to an increase in the length of
mosquito season in these areas. This increases the likelihood that once introduced, Rift Valley fever will become
established in our animal populations. Although mosquito vectors capable of transmitting Rift Valley fever exist
throughout the country, it is likely that this disease can become established only in the areas of the country where
winter temperatures will be able to support active mosquito populations (Figures 4 and 5).
Objectives. Methods, and Report Presentation Format
The objectives of this preliminary study are as follows:
Examine current literature on the greenhouse effect in light of the implications for changes in the patterns
of animal disease in the United States;
Undertake a literature review of four animal diseases that are either important or potentially significant
in that their distribution and occurrence might undergo alterations due to climatic and environmental
changes induced by the greenhouse effect;
Examine the potential economic effects of these changes;
Summarize the findings in a report presented to the Environmental Protection Agency; and
Provide recommendations for further study.
11-3
-------
Stem
This study was performed as a preliminary effort with the intention that it is likely to raise more questions
than it answers. It is expected that this study will result in the allocation of sufficient resources to permit a
detailed study of the effects of global warming on animal agriculture not only in the United States, but
throughout the world. In the livestock industry as in other economic sectors, the effective links with world
markets and the interdependency that is present today render the findings of a detailed study within the confines
of a single country questionable. A thorough study should emphasize the United States but cannot ignore the
changes likely to occur in livestock industries throughout the world.
The predictive findings presented in this report are based on an understanding of the diseases obtained
through the literature search and personal experience with the diseases and are made according to the 2xCO2
regional values obtained with the GISS and GFDL models (Rind and Goldberg, 1988; Jenne, 1987).
This report begins with a summary section entitled Findings. Here important observations and conclusions
are discussed. Additionally, the need for a more definitive study is pointed out.
The general changes which may be seen are presented and include a brief discussion of some of the
potential changes that may be precipitated in the poultry industry. For the sake of brevity, we have concentrated
this discussion on the confined chicken industry.
Next, four diseases are discussed that we feel represent the types of changing disease patterns that may
be induced by the climate changes. We performed a detailed review of the recent literature on these diseases
and have presented the important aspects of these diseases as they relate to climatic and environmental changes
induced by the greenhouse effect. A discussion detailing the clinical signs, vector biology, epidemiology,
economic importance, and means of control is included for each disease in addition to an economic analysis on
the possible effects that changes in global climate may have on these diseases and their relative importance.
Two of these diseases, bluetongue and anaplasmosis, presently occur in the United States and are good
examples of diseases that will probably become more important with the predicted climate changes. The other
two diseases, Rift Valley fever and African swine fever, are foreign animal diseases that pose an everpresent
risk of introduction into the Unites States. We believe that as the climate and environment change, not only
will the risk of introduction of these diseases be increased because of shifts in human and animal populations,
but the risk of these diseases becoming established and entering the vector reservoir will increase as vector
populations and geographical distributions grow.
Economic Methodology
The economic methodology used for this study consists of estimating the impact of the spread of each
animal disease on (1) decreased annual production resulting from the mortality associated with the disease, (2)
decreased animal production resulting from the morbidity of the survivors of the disease, and (3) the decline in
the value of the capital stock invested in animal herds because of the mortality associated with the disease.
Herd-specific mortality and morbidity rates were estimated separately for cattle and sheep, and for whether
or not the animals are mature. These rates apply when the disease is first introduced on a large scale and would
be reduced considerably once the disease becomes endemic. The rates used were based on reports in the
literature and vary according to the disease and the its rate of spread when introduced into a new environment.
Normally, it would take several years for these diseases to become established to these levels, and consequently
the analysis is calculated to reflect the effects over a several-year period. Morbidity rates were adjusted to
reflect the losses in production associated with infection. Both mortality and morbidity rates were applied on
a state by state basis assuming the following percentage of herds to be infected within each state once the disease
is endemic: Bluetongue 50%, Rift Valley Fever 20%.
To the losses from mortality and morbidity, as estimated on the basis of the cumulative decline in annual
production as the disease spreads over a several-year period, is added the capital loss associated with the herd
as a reproductive and growing out or fattening unit. The impact of the spread of the disease is estimated by
11-4
-------
Stem
aggregating the losses associated with that spread across the states in which the disease has become endemic.
No losses are estimated for states in which there are only occasional outbreaks. Nor is any account taken of
efforts to control the disease or of the costs of this control The numbers, therefore, represent a worst-case
scenario of what might happen if no control or containment measures were exercised.
The detailed tables involved in these calculations are presented in Annex A. Notes to those tables list
sources and describe details concerning the calculations.
Methodology for Determining Predictions of Future Disease and Disease Vector Distributions
The distributions of future disease and disease vectors summarized in Figures 3,5, and 7 were made based
on a combination of current knowledge of the diseases and the vectors combined with the GISS and GFDL
climate change scenarios as predicted for 2x levels of CO,.
Current knowledge as presented in today's scientific literature is variable depending upon the disease and
the vectors involved. Reasonably good data are available on the present distribution and vector biology of
bluetongue, which permitted us to make projections that take into consideration the climate models, the vector
biology, and the natural environmental resistance of the bluetongue virus. The literature on anaplasmosis, on
the other hand, lacks details and specificity in terms of current distribution of both the vector and the disease
in the United States. Consequently, no attempt was made to display either the present distribution of
anaplasmosis or its vectors, and no attempt made to predict their potential future distribution. For Rift Valley
fever and African swine fever, sufficient data in the literature on both the current distribution of the disease and
the vector distribution and biology enable us to make general predictions on areas of the United States where
these diseases could become established if introduced to America in the future.
Estimates of temperature and precipitation at 2xCO2 levels from the GISS and GFDL models were used
in these disease projections. Monthly average mean temperatures from the GFDL model were used by region
(southern Great Plains, Southeast, West Coast, and Great Lakes) for the months of January and July. The
diurnal variations for these months and regions were then added and subtracted from these means to arrive at
the high and low temperatures for the period. These values were then used to define the geographic limits of
the vectors and/or disease agents according to the biology of the specific vector or disease-causing agent. The
net change in precipitation predicted by these models was also taken into consideration.
The objective of the accompanying maps is to display to the reader the trends that we see possible. It is
important to note that the inclusion or exclusion of a state in a group on any of the maps is at this time only
speculative. There may be yet undescribed factors that are determinate in the distribution of a vector or disease
in any of these states. Furthermore, the specificity of the climate data is not detailed enough to be able to
determine, with any degree of certainty, changes between adjacent states. Finally, variation of disease
distribution within states was not taken into consideration in this study. The inclusion of a vector or disease, in
any given state, in the current or the projected maps is not meant to infer that the entire state is, or will be,
affected.
i Likely To Be Seen as a Result of Greenhouse Gas-Induced Global Climate Changes
As a result of global warming conditions, the meterological and environmental effects of the greenhouse
effect are likely to induce changes on animal health patterns in three general ways. Changes can be anticipated
in the habitats of bisects that transmit animal disease, in floral environments, and in the patterns of animal and
human movement. Each of these potentially could profoundly change the patterns of animal diseases.
Direct changes in environmental habitats will be precipitated as a result of the temperature and humidity
changes induced by the accumulation of greenhouse gases. As the coastal regions experience increases in mean
temperatures accompanied by precipitation increases, the habitats that can be exploited by various disease
vectors will increase. Southern summers may become too hot for optimum tick activity in some areas; however,
it is likely that activity will increase for ticks in these regions during spring and fall months. Many animal
11-5
-------
Stem
diseases are limited strictly by vector habitats (Fraser, 1986; El Shoura, 1987). As the environment changes,
permitting an expansion of the vector habitat, the geographic distribution of a disease carried or spread by that
vector is also likely to expand.
Vectors are living agents that carry disease organisms from one host to another. They are often
arthropods that carry viruses, bacteria, or protozoans. There are two types of vectors: infected and mechanical.
Infected vectors are ones in which the disease organism invades tissue and eventually multiplies itself within the
vector's body. Infected vectors can often transmit the disease organism to their offspring. As a result, disease
organisms can be maintained in vector populations for a number of generations. These vector populations can
become reservoirs of infectious diseases. Mechanical vectors are ones that simply, by virtue of what sticks to
a bug's feet or mouthparts, transport pathogens from one host to another.
Biological vectors become infected by the disease organism, but do so without showing any pathological
reaction. In most cases, ingestion of the disease organism happens at feeding. The agent is carried to the gut
along with the ingested blood, but instead of being digested it invades the cells lining the gut. Here, replication
of the agent generally takes place before it spreads throughout the vector's body. Later the disease agent is
disseminated to a new host, either through saliva injected when feeding or through waste or coxal fluids passed
during or soon after feeding on the host.
For diseases occurring in a given region, incidence is frequently limited by the seasonal activity of vectors
(Blood et al., 1983; Fraser, 1986; El Shoura, 1987). The vectors in the central and northern states do not
undergo year-round activity owing to the temperature effects of winter, which halt or severely limit the
reproductive cycles of the vectors, vector population, and animal-vector exposure, thereby reducing disease spread
in these areas.
Based on an understanding of a disease, its epidemiology, vector biology, and current distribution, and on
environmental changes that may occur according to the predictions produced at 2xCO2 values of the GISS and
GFDL climate models, one can predict the future distribution of that disease. It appears that for some of the
vector-borne diseases there may be an expansion of the present-day distribution northward, resulting in an
increase in overall disease distribution. Unfortunately, the southern ranges of the diseases are likely to remain
the same and, in fact, in many areas the diseases may become even more important. Despite excessively hot
summers for some disease vectors, there may be an increase in fall, winter, and spring season vector activity,
permitting the vector to undergo an increase in the number of reproductive cycles and a decrease in the
generation time for each cycle, resulting in increased numbers and concentrations of vectors. This could result
in an increase in the disease transmission rates and number of disease particles or units introduced into
individual animals, provoking an increase in severity of these diseases.
The GISS and the GFDL models predict that the central areas of the great continental land masses, such
as North America, will experience increases in temperatures and a decrease in relative humidity, causing a
corresponding decrease in soil moisture. This will likely precipitate important environmental changes. In states
where cereal and grain crops are marginally productive today, the importance of these crops to the state's
agricultural industry is likely to decrease because of these decreases in precipitation and humidity. Nonirrigated
land may dry to the point that it may be unable to sustain cereal cropping activities. These areas will take on
increased importance as rangelands, permitting the expansion of extensive animal agriculture. The geographical
distribution of the diseases may expand northward as the system of extensive range animal agriculture extends
from regions of Texas, Oklahoma, New Mexico, and Arizona to the traditional breadbasket regions of the
Midwest. Animal agriculture in the form of extensive rangelands is likely to remain important to the economies
of the present-day southern range states, and the disease incidence in these areas could be expected to remain
about the same. The seasonality of these diseases may change, with incidence increasing in the spring and fall,
and decreasing during summer months owing to the effects of high temperatures and low humidity on the
efficiency of vector reproductive and activity cycles. This increase in rangelands may also be enlarged as the
American taste preference for beef continues to shift from the fat-laden meat of grain-fed feedlot animals to
the leaner red meat produced by range animals reared under extensive range conditions.
11-6
-------
Stem
Furthermore, in the Midwest as the environment dries and the predominant form of agriculture begins
to shift away from crop-based systems to extensive animal agriculture, the ecosystems may undergo a change
resulting in an increase in the distribution of bushes and shrubs. These are ideal habitats for ticks, as the
humidity in the micro-environment surrounding the bushes and shrubs is considerably higher than the overall
environmental average permitting tick survival and transmission of diseases such as anaplasmosis and babesiosis.
Also in the Midwest as subterranean water supplies permit, there may be an increase in irrigation for plant
agriculture. If the importance of animal agriculture increases around these regions, the level of vector activity
of such water-breeding insects as mosquitoes and midges may increase owing to the presence of standing water
resulting from irrigation practices. This could be significant for diseases such as bluetongue, Eastern, Western,
and Venezuelan encephalitis, and Rift Valley fever, should the latter disease be introduced to the United States.
Finally, as the effects of the climate changes become apparent on a worldwide basis, there could be large
population shifts in both animal and human populations. This will result in the introduction of diseases and
disease strains from one state to another, as the relative distributions of animal and human populations change.
More important, however, will be the increased threat from animal diseases that are currently exotic to the
United States. Among this group of diseases are several of the most important diseases known, including foot
and mouth disease, rinderpest, hog cholera, and African swine fever. It is this group of diseases that poses the
greatest threat to American agriculture. At least two of these diseases, foot and mouth disease and African
swine fever, are capable of wind-borne spread, which can potentially make a small outbreak a major multi-state
epidemic in a matter of a few days.
Compounding this risk is the speed and distances over which American livestock are presently transported.
In a study of the animal shipment patterns from three major livestock markets of the Midwest during a seven-
day period in the month of July, animals were transported to 42 different states (Figure 1) (Saulmon, 1969).
If a foreign animal disease is introduced to one or more of the important livestock markets, it can spread to
many of the contiguous 48 states in the space of a few days. It is important to keep in mind that the incubation
period of rinderpest and foot and mouth can be long enough that a shipment of animals incubating the disease
might possibly go unnoticed by all but the most alert animal inspectors at these markets. If introduced and
spread in this fashion, a disease such as rinderpest or foot and mouth truly could be a disaster for the livestock
industry, the effects of which would be felt throughout the American economy. In fact, foot and mouth disease
has been introduced into the United States on eight occasions (McCauley et al., 1977). The most serious
outbreak occurred in 1914, where the epidemic began in Michigan and spread to 22 states by 1915 following
introduction into the Chicago stockyards.
Control of these diseases must be swift and extremely aggressive if the disease is to be eradicated from
American soil, once introduced. Unless the disease spreads to the point of being extensively distributed, the
control method of choice is total slaughter of all animals that potentially have been in contact with the disease.
This, of course, is a very expensive undertaking, although far less expensive than permitting the disease to
become established.
During the eradication program of the foot and mouth disease epidemic of 1914 -1915,172,000 cattle,
sheep, goats, and swine were slaughtered (McCauley et al., 1977). If a worse case scenario is envisioned, and
foot and mouth disease is introduced into the United States to the extent that to eradicate the disease 1% of
the livestock population must be slaughtered, the costs of such a control program would be estimated at $539
million (1977 dollars). The net effect on consumer prices of meat could be up to $1.02 billion (McCauley et aL,
1977).
The effect of the increased temperatures resulting from the greenhouse effect will likely have important
economic ramifications for the poultry industry. Although today's chicken industry is a highly intensive form of
animal agriculture with considerable attention given to environmental control in closed units, the effect of
temperature increases on the industry could be considerable. Major investments may be necessary to cope with
the increased temperatures that are expected in the central and southern states where much of our poultry
11-7
-------
Stem
LIVESTOCK SHIPMENTS FROM
THREE MAJOR STOCKYARDS
Figure 1. Cattle movements during one week in July from Chicago, Omaha, and Kansas City (Saulmon 1969).
11-8
-------
Stem
industry is concentrated. In many cases, cooling units and air quality control installations will have to be
completely re-done in order to prevent heat stress and increased respiratory disease and related problems.
. ... Salmondtoris, a stress-related disease, is likely to become more important. This carries significant public
health implications in light of recent evidence of transmission of Salmonella enteritidis to chicken eggs before
the egg shell is formed (St. Louis et al, 1988). These researchers estimate that egg-borne Salmonellosis in
humans has been responsible for up to 70% of the cases of food poisoning in the northeastern United States in
a two-year period from January 1985 to March 1987. The public health implications are clear if the prevalence
of Salmonella enteritidis in chicking ii»-roMf 5 r
The recent extended heat wave of 1986 may be used as a good illustration of this problem. A drought
°??/£S m South®a?t'wmch tegan »n spring of that year. By July, average daily temperatures reached highs
ol 1WF or more. This put severe stress on the climate control systems of chicken houses from southern
Pennsylvania through Georgia. At these temperatures, birds become reluctant to move even to reach nearby
watering troughs resulting in dehydration, heat stress, and for many, death. By early July, some farmers in
North Carolina were losing as many as 500 birds a day (Murphy, 1986). By the end of July when the heat wave
broke, losses to the poultry industry in North Carolina were estimated to be $48.5 million, whereas losses of
poultry in Georgia were estimated to be $5 million (United Press International, 1986). The temperatures
experienced during this period are likely to be much more common when we have reached 2xCO0 levels,
according to GISS and GFDL models.
These changes could result in a slight northward expansion of the poultry industry, and depending upon
the economics and availability of energy at the time, could result in a contraction of the southern margins of the
poultry industry - when and if the environment becomes too hot to permit the continuation of efficient poultry
raising.
11-9
-------
Stem
CHAPTER 1
BLUETONGUE
Introduction
Bluetongue is a viral disease that causes debilitating lesions of the mouth in a variety of wild and domestic
ruminants. Several species of insects, in particular the biting midges (punkies and no-see-ums), act as insect
vectors by transmitting the disease. Sheep manifest the worst clinical signs, but the disease is also significant in
cattle and rarely so, in goats. Goats can, however, act as a reservoir. Clinical signs include erosions,
hemorrhage, and edema of the mouth, tongue, and oral cavity. Frequently, there is a viscous mucopurulent nasal
discharge that crusts on the muzzle; this is sometimes accompanied by diarrhea. Erosions are also common on
the coronet band of the hoof causing acute lameness. Pregnant ewes and cows abort from the disease, and in
some cases sheep will abort from modified live vaccines administered during gestation (Fraser, 1986). An
infected fetus will show hydrencephaly (Gillespie and Timony, 1981).
In this country, the principal vector is Culicoides variipennis. one of the punkies that lives in and around
farm wastewater. Bluetongue morbidity is highest in late summer and early fall and disappears with the onset
of frost (Gillespie and Timony, 1981). This correlates to the biting activity of the vector. Increased incidence of
the disease is seen when the warm season is wet (Gillespie and Timony, 1981).
Clinical Signs. Epidemiology, and Vector Biology of the Disease
Bluetongue was originally described as a clinical entity in South Africa in the early part of 20th century
(Blood et al., 1983). It was first described in the United States in 1952 (Gillespie and Timony, 1981). In recent
years, evidence of the disease has been demonstrated in a variety of geographic areas, and it is probable that the
virus is moving into new regions (Metcalf et al., 1981).
Bluetongue may be approaching a cosmopolitan distribution. Incidence of the virus has been verified in
most areas of Africa and extends along the Mediterranean coast of Asia and eastward through Pakistan and
India. There have been unsubstantiated reports in Peru and recent confirmation of the virus in Australia (Della-
Porta et al., 1981). In North America, only the United States officially reports outbreaks of the disease,
although it is highly probable that it occurs in Mexico and Canada (Walton et al., 1984).
There are 20 immunological serotypes recognized worldwide (Gillespie and Timony, 1981), only a few of
which occur in the United States (Gillespie and Timony, 1981). Virulence of the disease varies regionally and
from outbreak to outbreak. In Africa clinical signs appear worse than those seen in the United States. These
differences probably reflect the geographic and temporal distribution of particular serotypes (Gillespie and
Timony, 1981).
Signs of the disease are similar in sheep and cattle, but severity and mortality are much greater in sheep
(Gillespie and Timony, 1981). Mortality in sheep can approach 30% in this country (Fraser, 1986). Goats,
although they can host the virus, rarely show clinical signs (Gillespie and Timony, 1981).
The most significant clinical signs are evidenced as erosions, hemorrhage, and edema of the mouth, tongue,
and oral cavity. Often there is a viscous nasal discharge that crusts on the muzzle. Early in the course of the
disease, affected animals show a high (105° - 107°F) temperature. There is commonly inflammation of the
coronet band, and sometimes diarrhea Pregnant ewes and cows frequently abort. An infected fetus will show
hydrencephaly (Gillespie and Timony, 1981).
Morbidity is greatest in lambs that have recently lost colostral antibody protection or have been unexposed
to the bluetongue virus during the period of passive antibody protection (Gillespie and Timony, 1981). In herds
11-10
-------
Stem
where the disease is endemic, older sheep are less severely and more infrequently affected. However, the
resulting reproductive losses can be disastrous for sheep fanners.
Anderson (1970) shows serological evidence of bluetongue in 39% of dairy cattle and 48% of beef cattle
(Anderson, 1970; Sellers, 1981). He also isolated virus from Culicoides spp. in Minnesota and concludes that
bluetongue is more prevalent than previously thought and that, because the symptoms in cattle resemble bovine
viral diarrhea, the disease is greatly underdiagnosed (Anderson, 1970; Sellers, 1981).
Clinical signs of bluetongue are variable and similarities exist with diseases such as foot and mouth disease,
mycotic stomatitis, infectious bovine rhmotracheitis, bovine virus diarrhea, rinderpest, akabani disease, and
vesicular stomatitis, making it difficult to readily identify the etiological cause in a particular outbreak (Gillespie
and Timony, 1981). As a result, in areas where the disease has been previously unknown or infrequent, definitive
diagnosis depends on extensive serological testing, viral isolation, or transmission studies. In areas where
bluetongue is common, diagnosis of the disease is based upon clinical observation (Gillespie and Timony, 1981;
Fraser, 1986). Elk, antelope, big horn sheep, Barbary sheep, moose, and several species of deer have all shown
serological evidence of infection (Gillespie and Timony, 1981). This means that these wildlife species probably
serve as reservoirs for the virus. Clinical signs and death have been induced experimentally in white-tailed deer
(Callis, 1985).
Experiments show that the virus is only transmitted by insect vectors and that simple direct contact does
not result in disease spread (Gillespie and Timony, 1981). As a result, distribution of the virus follows that of
the vector, although there have been reports of outbreaks in areas where no known vectors occur.
Culicoides variipennis is most important vector for bluetongue in the U.S. (Wirth, 1979). It is a proven
transmitter in the laboratory and there have been frequent isolations of the virus from midges of this species in
nature (Wirth, 1979). While the distribution of the bluetongue virus is reliant on the Culicoides vector, the
existence of reservoir populations is just as essential. Without reservoirs to maintain the virus between outbreaks
of the disease, there would be no source for the Culicoides vector to acquire the infection. In the case of
bluetongue, a definitive host is still unproven, but it is likely that one or several species of wild ruminants,
domestic cattle, or goats host the virus from outbreak to outbreak without evidencing severe clinical symptoms.
Its occurrence in sheep is probably accidental.
The prevalence of bluetongue in the United States was studied in the winter of 1977-78, when 19,758 blood
samples from American cattle were examined for serological evidence of exposure to this disease. Overall,
18.9% of the samples were positive, with the highest rates of prevalence of the disease in the southwestern states
and the lowest rates in the northern states. The highest rates of infection were in Puerto Rico and the Virgin
Islands. High rates in the Southwest probably reflected a recent outbreak that occurred throughout that region
(Metcalf et al., 1981). Generally, bluetongue epizootics have occurred in the southern regions of the United
States where the climate is warmer and more favorable for the culicoides vectors (Figure 2).
The insect carrier has been found to be an efficient transmitter of bluetongue. The organisms rapidly
multiply in the midge. It has been estimated that by 20 days following a bite of a bluetongue-infected animal,
one midge has the potential to infect three animals with sufficient numbers of the bluetongue virus to result in
clinical disease (Sellers, 1981).
Economic Importance of Bluetongue
Bluetongue is of significant economic importance today in the United States, although accurate figures are
very difficult to come by.
Control of Bluetongue
Control of the disease, whether in epizootic proportions or not, is probably best accomplished by a
combination of vaccination programs and by minimizing contact between the vector and ruminant host
11-11
-------
Stem
ENDEMIC
EPIZOOTIC
FTiEE
BLUETONGUE: PRESENT DISTRIBUTION
Figure 2. Present distribution of bluetongue (from Walton et aL, 1984).
11-12
-------
Stem
The vaccine against bluetongue is less than ideal, as pregnant animals should not be vaccinated because
of a risk of abortion or teratogenitity. There have been reports of abortion following administration of modified
live bluetongue vaccines to pregnant sheep (Fraser, 1986). Vaccination of sheep should take place after shearing
because there is some evidence that the vaccine can cause wool breakage in animals showing severe reactions
to the vaccine (Gillespie and Timony, 1981).
Minimizing contact between biting midges and animals includes decreasing vector populations in standing
pools of water, increasing water runoff, and closeting susceptible animals during evening and nighttime hours
when the vectors are active. However, the breeding habitats of Culicoides variipennis are discrete and not easily
identified, but are highly productive, making control measures aimed at this vector difficult (Barnard and Jones,
1980).
Possible Effects of Global Climate Change on Bluetongue
Each generation of adult punkies is short lived, and it is this stage that is capable of transmitting the
bluetongue virus to ruminants. There are a variety of immature stages (instars) that live in fresh/brackish water
in the bottom sediments. Temperature in part determines the speed with which these instars pass from one
stage to another. A shorter generation time potentially increases the number of reproductive cycles in any given
season. Additionally, as described below, climate changes will also affect adult survivorship and the length of
the active season for Culicoides.
The impact of a changing climate is manifested in altered patterns of water distribution. Here it is
probable that the diseases would become more important in areas where there is an increase of standing pools
of water, e.g. a partially flooded meadow down slope from herded cattle or sheep. Such an environment would
be ideal for the increased propagation of Culicoides variipennis.
Introduction of bluetongue into a new geographic area has been blamed on strong surface winds blowing
the midges on the wind. This was reported for the 1977 outbreak in Cyprus. The source of midges was
interpreted to be-Syria and eastern Turkey (Sellers et at, 1979a; Sellers et aL, 1979b). Wind-borne introduction
of bluetongue and Culicoides is potential!) possible in states or regions that currently do not have the disease
but that may develop into acceptable habitats as their environment changes.
Barnard and Jones (1980) report that seven generations of £. variipennis occur in Colorado each year.
Generations happen between June and September and one overwintering generation develops slowly during
favorable weather between October and March (Barnard and Jones, 1980). The predicted climate changes
probably represent an opportunity for survival of greater numbers of adult and larval midges and more rapid
development if the winters become warmer. Longer warm season periods would increase the season of biting
activity for midges. A longer season of activity for midges would increase the number of animals exposed to
midges by an increase in the length of the exposure period and by an increase in the vector population from
an increased number of midge generations (Mullens and Rutz, 1984). Adult survivorship of Culicoides species
is very important in the maintenance of the disease. There is evidence that bluetongue virus numbers peak in
midges 10-14 days after ingestion of virus from an infected host (Foster and Jones, 1979). A wanner climate
would therefore favor adult survivorship and increased disease transmission. Figure 2 divides the Unites States
into areas where bluetongue is currently enzootic (occurring with a high degree of frequency and well established
in the environment), epizootic (occurring infrequently in outbreak form when conditions for its propagation are
ideal), and bluetongue-free areas. The principal factor preventing the establishment of the disease in the
bluetongue-free states and the states where it occurs sporadically in outbreaks appears to be climate related
(Walton et al., 1984).
Apparently there are two important factors that limit the distribution of this disease (Walton et aL, 1984).
First, bluetongue is established in areas where there is sufficient precipitation, standing water, and warm enough
winters to permit over-wintering of the midge vectors to occur. Secondly, bluetongue virus reproduction is
11-13
-------
Stem
arrested in midges whose metabolic rate has been reduced below a critical level by environmental temperatures
(Walton et al., 1984). Summertime temperatures of present-day Canada and the northeastern United States
have temperatures that appear to be either too low or of too short duration to permit bhietongue virus
replication (Walton et al., 1984; Metcalf and Luedke, 1980). As the winter temperatures warm in the northern
Great Lake states, permitting establishment, maintenance, and proliferation of the midge populations and
sufficient replication of the bluetongue virus within the midges, the states that are presently bluetongue free may
become areas where outbreaks occasionally occur. The states that today experience occasional outbreaks could
become enzootic areas for the disease as it becomes permanently established (Figure 3).
It is not likely that the predicted climate changes will result in a significant vector habitat loss for the
bluetongue vector. Bluetongue and its vector Culicoides currently are highly prevalent in some of the world's
warmest and driest regions. For example, Niger has an annual temperature mean greater than 80° and mean
annual precipitation less than 15 inches, but has a bluetongue prevalence of more then 80% (Mariner et al., in
press).
The changes predicted by the GISS and the GFDL climate models appear to be sufficient to permit these
changes in bluetongue distribution to occur. These changes are likely to occur gradually and may not be
apparent as cases of bluetongue are misdiagnosed as other similar-appearing diseases. This indicates a strong
need for epidemiologjc monitoring in these areas so that proper control and prevention measures may be
initiated.
Economic Model of the Effects of Global Climate Changes on Bluetongue
Bluetongue disease is assumed in this analysis to become endemic in the states in which there are now
only occasional outbreaks: Alabama, Arkansas, Georgia, Kentucky, Louisiana, Mississippi, Missouri, Montana,
North Carolina, Oregon, South Carolina, South Dakota, Tennessee, and Wyoming.
Herd-specific mortality and morbidity rates were estimated separately for cattle and sheep and for whether
or not the animals are mature. These rates apply when the disease is first introduced on a large scale and would
be reduced considerably once the disease becomes endemic. Normally it would take several years for these
diseases to become established to these levels, and consequently the analysis is calculated to reflect the effects
over a several-year period. To the losses from mortality and morbidity, as estimated on the basis of the
cumulative decline in annual production, is added the capital loss associated with the herd as a reproductive and
growing out or fattening unit The impact of the spread of the disease is estimated by aggregating the losses
associated with that spread across the states in which the disease has become endemic. No losses are estimated
for states in which there are only occasional outbreaks. Nor is any estimate made of efforts to control the
disease or of the costs of this control The numbers, therefore, represent a worst case scenario of what might
happen if no control or containment measures were exercised.
Left uncontrolled, this would result in annual production losses associated with the disease equal to $634
million for cattle and $21 million for sheep, plus a decline in the value of the herd equal to $1287 million for
cattle and $12 million for sheep. This totals 8% and 6% of the actual value of production in 1987 of cattle and
sheep, respectively (models 1 and 2, Annex A).
11-14
-------
Stem
ENDEMIC 1666663
EPIZOOTIC i;:;:;:;:i:
BLUETONGUE: PROJECTED DISTRIBUTION
Figure 3. Projected future distribution (2xCO2 levels) of bluetongue: projections are based on diurnal
variations of mean temperatures for the months of January and Jury and mean precipitation
predictions of GISS and GDFL climate models, present day disease distribution (Walton et aL,
1984), and known vector and agent biology as presented in current literature. Projected presence
in any given state does not imply statewide distribution due to localized environmental variability.
11-15
-------
Stem
CHAPTER 2
ANAPLASMOSIS
Introduction
Anaplasmosis is a vector-borne rickettsial infection of cattle, sheep, goats, and wild ruminants including
deer, big horn sheep, and elk. It is the second most important disease of cattle in the United States, with
estimates of annual economic losses ranging from $36 to $100 million a year (Blood et al., 1983; Rebhun, 1984).
The anaplasma organism is a parasite of red blood cells, and a corresponding major clinical feature of the
disease is anemia. Losses are due to chronic debility, decreased growth and production, and death.
The distribution of anaplasmosis is nearly cosmopolitan, occurring in Africa, Asia, Australia, South
America, and North America (Blood et al., 1983). Distribution of the causative rickettsial agent generally
parallels the distribution of a variety of tick species, although its incidence does occur beyond the boundaries of
known tick regions. Infection can be mechanically transmitted through biting flies, contaminated needles, and
syringes, surgical equipment, and any blood to blood contact (Gillespie and Timony, 1981). Control and
eradication of the disease is made more difficult by the common occurrence of a sub-clinical carrier state in
infected adult animals. Precise distribution of anaplasmosis in the United States today is not known. This is a
concern to the U.S.D.A., which is currently undertaking efforts to determine the distribution and severity of this
disease. For this reason, we were not able to produce detailed current and projected distribution maps and
economic projections for anaplasmosis.
Anaplasmosis is a significant disease not only because of its current economic impact but also because of
the incipient nature of this disease. With the expansion of vector habitats, the disease has the potential for rapid
spread and wide dissemination to new regions.
Clinical Signs. Epidemiology, and Vector Biology of Anaplasmosis
Clinical signs of Anaplasma marginale infection are difficult to discern. Infected animals frequently show
no clinical signs for long periods of time, whereas others may die from chronic debility or even acute anemic
crisis. Fever is the first clinical sign, but in range-fed or feedlot animals this often goes undetected.
Furthermore, fever in anaplasmosis usually disappears before other symptoms become apparent (Gillespie and
Timony, 1981). In fact, body temperature may be subnormal in the later stages of the disease (Gillespie and
Timony, 1981; Blood et al., 1983).
Other clinical signs of anaplasmosis include depression, dullness, firm feces, icterus, peripheral lymph
node enlargement, occasional edema, and less commonly, catarrhal enteritis as evidenced by mucus-tinged feces
(Gillespie and Timony, 1981; Blood et al., 1983). As a disease, anaplasmosis is most important for adult animals
under stress associated with high environmental temperatures, transport, change of feed, or calving (Jubb and
Kennedy, 1970).
Diagnosis of anaplasmosis is best made by examining stained blood smears which show the rickettsia
within the red blood cell. Usually, but by no means always, the anaplasma organism, which is contained within
its own membrane, is seen pushed against the cell wall. Babesiosis, a similar tickborne disease, and other red
blood cell parasites are also diagnosed in the same way by the specific morphology and placement of these
inclusions.
Clinical anaplasmosis is primarily a disease of older animals. Young animals can be infected, and will
remain infected (therefore are potential carriers if never treated) but may never show disease manifestations any
time during their lives (Jubb and Kennedy, 1970). Instead, their bodies accept the infection and maintain a low
level of circulating parasites that can seed rickettsia to the vector organisms. These vectors, usually ticks, can
then pass the infection onto other susceptible animals. As a result, animals infected when they are young may
11-16
-------
Stem
not suffer clinical disease, but may serve as a source of infection for life as carriers. Additionally, these animals
often fall into the category of "chronic poor doers" showing reduced production parameters and representing a
significant economic drain.
If previously unexposed cattle are infected at the age of three years, they often become clinically ill,
suffering from either a peracute or a chronic form of anaplasmosis. Those animals suffering from the peracute
form may die within 2 to 3 days of the first clinical symptoms. Those affected chronically linger with weakness,
icterus, and emaciation. In animals previously affected, clinical signs of the disease are often triggered in stress
which can be demonstrated in the form of intercurrent disease, heat stress, or any physiological strain (Jubb and
Kennedy, 1970). As previously mentioned, transport, feed changes, seasonal changes in climate, or calving are
common stress-inducing factors that precipitate clinical anaplasmosis.
Losses from production factors and mortality due to anaplasmosis are worse during hot weather (Gillespie
and Timony, 1981). Mortality may reach more than 50% depending upon the concurrent stress and if the disease
is introduced into a herd or a group of susceptible individuals for the first time (Blood et al., 1983). Animals
that survive will become carriers for an extended period of time (Gillespie and Timony, 1981). Additionally,
these animals, unless a biological cure is effected, will detract from farm profits due to resulting production
losses (Blood et al., 1983).
At least 20 species of ticks have been shown to experimentally transmit anaplasmosis (Blood et al., 1983).
In addition, mechanical transmission has been documented by horseflies and mosquitoes (Kocan, 1986; Gillespie
and Timony, 1981). Fortunately, only a few of these species of ticks are able to support a multiplication of the
rickettsial organism within their bodies before passing it vertically to their offspring, which can then be the source
of transmission to other animals. In the United States, the most important of these biological vectors are the
ticks Boophilus aimulatus. Dermacentor occidentalis. Dermacentor andersoni (Gillespie and Timony, 1981). The
significance of these species is that the anaplasma organism can persist by reproducing within the tick-wildlife
population indefinitely without contact with the domestic farm animals. This makes the eradication of
anaplasmosis very difficult, if not impossible, once is has become established within a given ecosystem.
Development of the rickettsial organism within the tick starts with infection of the tick gut cells after
feeding on an infected animal. Several developmental stages follow and to some extent parallel developmental
stages of the tick. Eventually, the tick hemocyte (equivalent to a mammalian white blood cell) becomes infected.
Finally, the tick's salivary glands become infected and the organism is then transmitted to the tick's host during
the next blood meal. Surprisingly, anaplasma is not transmitted to the host rapidly, as tick attachment to the host
must approach 6 days before the rickettsial infection is transmitted (Kocan, 1986). As discussed below, this has
important implications for the control of the disease in domestic animals.
Economic Importance of Anaplasmosis
Anaplasmosis is said to be the second most important disease of cattle in the United States (Goodger et
aL, 1979). Estimates of losses vary from $35 to 100 million per year (Rebhun, 1984; Blood et aL, 1983).
Nevertheless, there is a scarcity of information on the geographic distribution, prevalence, and economic effects
of this disease.
Control of Anaplasmosis
The primary problem in control of anaplasmosis is keeping adult uninfected animals out of endemic areas,
and keeping infected carrier animals out of nonendemic areas. The extensive nature and rapidity of livestock
transport in present-day America makes this difficult (Jubb and Kennedy, 1970; Saulmon, 1969). Likewise, if
cattle from a distant place move into an endemic area where the disease is subclinicaL these older cattle can
become rapidly infected.
Fortunately, there are effective treatments for anaplasmosis. The carrier state can be effectively eliminated
from animals through the application of one of the following antibiotic protocols: (1) injection of 5 mg/Ib
11-17
-------
Stem
oxytetracycline, daily for 12-14 days; (2) injection of 15 mg/lb aureomyrin, daily for 16 days; or (3) injection of
5 mg/lb tetracycline for 10 days (Fraser, 1986; Howard, 1986; Blood et al., 1983).
Clinical anaplasmosis may also be treated by tetracyclines. Imidocarb, a newer drug, causes rapid
clearance of the anaplasma bodies from red blood cells of clinically ill animals at a dosage rate of 5 mg/kg given
every 14 days for two to three cycles (Howard, 1986; Blood et aL, 1983; Fraser, 1986).
Since it is thought that a tick must feed on an animal for up to 6 days before seeding the host animal
with enough anaplasma bodies to be infective, a good tick control program can be an effective means of
prevention (Kocan, 1986).
All of the above-described control systems are economically expensive (Gillespie and Timony, 1981).
Vaccinations against anaplasmosis have been successfully produced, but they are not used extensively in
the United States owing to the presence of the vaccine-induced antibodies, which cannot be detected from
antibodies produced as a result of natural disease. Thus, there is no way to differentiate between vaccinated
animals and those naturally infected. This becomes economically significant for interstate or international animal
transport. Few purchasers or regulatory bodies will risk the potentially disastrous results by importing vaccinated
cattle.
Both inactivated and attenuated vaccines have been developed. They appear to be effective which makes
'global" or wide area vaccination enticing. However, such extensive vaccination of all susceptible animals for the
disease would be a very expensive undertaking, and one would be unable to eliminate the disease entirely owing
to the wild animal reservoir. Furthermore, it is highly unlikely that an effective vaccination program could be
mounted in enough countries to permit international trade to proceed unhampered, and if this were not the case,
most countries would not accept the risk of vaccination of their animals, which would render them
indistinguishable from those that are clinically infected.
Possible Effects of Global Climate Change on Anaplasmosis
There are several striking features about the disease that make it a concern in a warming climate. These
factors combine to make an already extremely important cattle disease in the United States even more serious.
First, the day to day activity of the vector insects will probably increase, causing a rise in vector-host contact and
an increase in the transmission of the disease. The exception to this might be in the southern central states
where, according to the GISS and GFDL climate models, summertime temperatures will increase and relative
humidity will decrease. This may result in a decrease in the vector activities during these periods. Depending
upon the photoperiod response characteristics of individual tick species, however, there may be an increase in
the vector activities in either the spring or fall in this region. Second, the distribution of the habitat conducive
to the vector may expand. This will be due largely to the increased warming in the northern states, in particular
the Great Lake states.
The added stress of increased temperatures is likely to precipitate clinical anaplasmosis. Furthermore,
losses from anaplasmosis are more severe during hot weather and in warmer climates (Gillespie and Timony,
1981). Changing vector distribution will cause anaplasmosis to creep into areas where there are populations of
susceptible adults and juvenile cattle.
In cattle-producing systems, carrier animals, which can be impossible to identify without the collection of
blood samples followed by laboratory testing for antibodies, are a reservoir for introducing the anaplasma
infection to previously unexposed animals. With rapid and frequent transport of animals from one part of the
country to another, the problem becomes worse because if infected animals move into new territory where the
infection has not previously been encountered, the carrier animal becomes a source of infection through
arthropod vectors to other animals. The resulting losses from outbreaks of anaplasmosis can be disastrous. This
is especially true as the clinical signs of anaplasmosis are not particularly specific for this disease, complicating
and often delaying diagnosis.
11-18
-------
Stem
CHAPTERS
RUT VALLEY FEVER
Introduction
Rift Valley fever is a vectorborne viral disease of sheep and cattle characterized by explosive outbreaks,
symptoms of severe liver disease, abortion, and frequent rapid death.
Although its worst impact is on sheep and cattle, it also affects humans, camels, monkeys, mice, rats,
ferrets, hamsters, and goats (Blood et aL, 1983). Previously limited to Africa, Rift Valley fever has recently
spread to the Asian continent in the Sinai peninsula (Hoogstraal et aL, 1979; Blood et aL, 1983; Gillespie and
Timony, 1981). Spread to other areas of the world is a prominent and potential problem.
Whenever outbreaks are seen in animals, there is frequently concurrent morbidity in humans (Gillespie
and Timony, 1981). As a result, incidence of the disease has important public health implications (Gillespie and
Timony, 1981). In humans, symptoms resemble influenza with occasional cases complicated by long-term eye
problems and more rarely death (Blood et aL, 1983).
riliniral Signs. Epidemiology, and Vector Biology of Rift Vallev Fever
In domestic animals, sheep are most seriously affected, cattle are less affected, and goats rarely exhibit
serious signs. Morbidity and mortality percentages are generally higher for young animals than for adults.
Young lambs suffer 95-100% mortality (Gillespie and Timony, 1981); and young calves can show 70%
mortality (Blood et aL, 1983; Gillespie and Timony, 1981). Death in these animals is often seen within 72 hours
of the start of incubation. The rapid evolution of the disease process affords little time for treatment.
Pregnant ewes with Rift Valley fever usually abort, and mortality in sheep is often 20-30%. Mortality may
be characterized by peracute death with no clinical signs, or death may follow a display of a variety of signs
including vomiting, purulent nasal discharge, or hemorrhagic diarrhea (Gillespie and Timony, 1981). Signs in
adult cattle mimic those of sheep except that death ensues in only about 10% of the cases. Losses in any case,
whether death or debility, are severe.
In humans, the disease manifests with signs similar to influenza (headache, muscular pain, fever), although
there can be short- and long-term degeneration of the retina with resulting loss of vision (Schrine, 1951; Freed,
1951; Ayoub et aL, 1978).
Rift Valley fever is caused by a single stranded RNA virus in the Bunyaviridae family (Gillespie and
Timony, 1981). The virus is classified technically as an arbovirus since its principal mode of transmission is
through infection of insect vectors (Blood et aL, 1983). Although the disease was first described in Kenya in
1931, it is probable that this disease had previously occurred for years (Gillespie and Timony, 1981). The recent
history of the disease shows a virus with variation that may be becoming more virulent and biologically broader
in its adaptability to new ecological settings (Walsh, 1988).
Rift Valley fever had been a stable endemic disease of the Rift Valley in eastern Africa for a number of
years (Kaschula, 1956). In the last twenty-five years, the virus appears to be actively moving into new geographic
areas, new ecosystems, and new animal populations (Kaschula, 1956).
A large epizootic occurred in 1951 in South Africa when 100,000 sheep and cattle died. About 20,000
people were infected but no deaths were reported (Simpson, 1978). There was also another epizootic in South
Africa in 1974-75, in which human fatalities were reported (Simpson, 1978). Another epizootic broke in Egypt
11-19
-------
Stem
Li 1977 just northwest of Cairo. Over 200,000 people were infected, with approximately 600 deaths occurring
(Miller, 1980; Walsh, 1988). The fact that this epizootic continued throughout two years suggests an over-
wintering mechanism for the virus. It is possible that the virus maintains itself in the vector populations either
in individual insects that survive the winter or by transovarial passage to insect offspring (Hoogstaal et al., 1979).
One study reports that the virus is not maintained for any length of time in mosquito populations. It is also quite
possible that the virus survives in an unknown vertebrate host, which reinfects vector populations when they take
a blood meal in the spring or when the next epizootic begins (Mclntosh et aL, 1980).
As a clinical entity, Rift Valley fever, to this point in time, occurs only on the continent of Africa, although
recent serological investigation shows evidence of the virus in man and animals in Asia on the Sinai Peninsula
just east of the Nile River. This in itself is unremarkable because the distance traveled from the 1977-78
epizootic is small (Miller, 1980). However, the event along with the 1977-78 outbreak signifies movement of the
virus into a new ecological setting and technically onto a new continent (Hoogstraal et al., 1979).
More recently, in 1987-88 there has been an outbreak of Rift Valley fever in West Africa involving the
countries of Mauritania, Senegal, Gambia, and Mali (Walsh, 1988). Like other recent outbreaks, this one has
affected man and domestic animals causing significant mortalities (Walsh, 1988). The Pasteur Institute of Dakar
estimated the total number of human cases of Rift Valley fever in Mauritania at 1264 with 224 known deaths
(Walsh, 1988). Serological evidence for Rift Valley fever has been reported in several other west African
countries including Cameroon, Nigeria, and more recently Niger (Mariner et al., in press; Walsh, 1988; Ezeifeka,
1982; Fagbami, 1973). In the pastoral zone of Niger, a 1988 survey revealed over a 40% prevalence rate among
camels in the region (Mariner, 1988).
Rift Valley fever seems to have outbreaks on a 5-7 year cycle, some more severe than others. Incursion
into domestic animals seems accidental and appears only when conditions are right; that is, when there are high
numbers of both vectors and susceptible animals, infection of domestic animals results when the virus particles
are transferred to the blood The carrier of the virus can be arthropod or mechanical; reused needles or surgical
equipment can serve as mechanical carriers. Likewise, bloodsucking insects can also be mechanical carriers, but
in addition some insect species can become infected and can serve as biological vectors. Spread of the disease
by vector is more efficient than mechanical spread since the vector organism can serve as a persistent source of
virus particles, whereas a mechanical carrier exhausts its supply of virus particles with time and activity.
In animal populations, insect vectors are the principal means of introducing Rift Valley fever virus,
mosquitoes being the worst offenders. In humans, close contact with infected animals frequently induces the
disease. It seems that in the latter case, contact with blood products, especially in persons with skin abrasions,
may be the infective route (Blood et al., 1983), although inhalation of blood products in abattoirs has also been
suspect (Miller, 1980).
As many as 26 species of arthropods may act as vectors for the Rift Valley fever virus on the African
continent (Miller, 1980). Most of these species are mechanical carriers, although there are several very efficient
biological vectors in the group. Such biological latitude ensures successful transmission and survival of the virus
in a variety of ecological settings, which poses significant problems for disease control.
Several species of mosquitoes indigenous to Africa that carry Rift Valley fever are also found in the
United States. It is also possible that the virus could also be successfully carried by arthropods native to only
North America.
One study in Africa attributes the mosquitoes Culex theileri. Culex univattatus. Aedes juppi. and Aedes
lineotopennis as the most efficient vectors (Mclntosh et al., 1980). Culex theileri is thought to be biggest
offender, especially on the plateau areas of central Africa (Hoogstraal et al., 1979). It feeds on man, cattle,
sheep, and other animals. It is also long-lived, from early summer to mid-winter, and occurs in great numbers.
Other investigators have attributed the worst offender status to Culex pipiens citing that the 1977-78 epizootic
in Egypt was probably maintained by Culex pipiens (Hoogstraal et al., 1979; Miller, 1980)
11-20
-------
Stem
Of the vectors shown to transmit Rift Valley fever, three species of mosquitoes appear in the United
States: Culex pipiens. Aedes aegvpti. and Aedes ^pjerhynchus. Their existence in the United States would
provide a viable mechanism for disease spread if the virus were to become established here.
Aedes taenierhvnchus has a widespread distribution ranging from California to Peru and from
Massachusetts to Brazil along coastal areas and inland salt waterways (Stone et aL, 1983).
Aedes aegypti in the subgenus Stegomyia is a pantropical species with the highest incidence reported to
be in Egypt. Two subspecies have been reported in Georgia- Aedes aegypti taeniatus. and Aedes aegypti
excitans (Stone et aL, 1983).
Aedes aegypti is reported to transmit Rift Valley fever in experimental trials (Turrell and Bailey, 1987).
However, it has not yet been found to cany Rift Valley fever in nature (Turrell and Bailey, 1987). It is also
shown to be an inefficient vector (Turrell and Bailey, 1987).
Distribution of Culex pipiens is worldwide within 60°N latitude and 40°S latitude (HorsfalL 1955). It can,
in places, extend northward to the Arctic Circle as in Norway (HorsfalL 1955).
It has also been found that with decreasing feeding efficiency there is an increase in virus dissemination
on the part of the vectors. In other words, the more contacts a vector insect has with a mammalian host, the
greater the incidence of disease transmission (Turrell and Bailey, 1987). This has important ramifications for
the spread of the disease during an outbreak, as one vector can potentially infect many host animals or people.
Economic Importance of Rift Valley Fever
The economic importance of Rift Valley fever has to date not received attention in the literature. In the
recent outbreaks in Egypt and Mauritania, Senegal, Mali, and Gambia, it is clear that the economic effects have
been significant in terms of animal mortality, animal production losses, human mortality, and human morbidity.
Control of Rift Valley Fever
A human vaccine against Rift Valley fever has been developed and is currently being tested by the UJS.
Army. An animal vaccine under development soon will be tested in west Africa.
Currently, most control methods are aimed at mosquito control. In a study done on time of mosquito
activity, it was found that Culex pipiens had greatest flight activity after 12 p.m. (midnight). This time would then
be the best time for insecticide administration (Mitchell, 1982). Most control efforts to date have been largely
directed toward control of mosquito breeding grounds.
Possible Effects of Q|nhal Climate Change on Rift Vallev Fever
The potential for spread of Rift Valley fever from Africa to other regions is great It has already made
great progress across the African Continent with some evidence of its occurrence along the Mediterranean Sinai.
One theory on the route of spread from East Africa to Egypt, western North Africa including the Maghreb, and
down to West Africa involves the trans-Saharan camel caravan trade routes. Human travel is perhaps the
greatest threat in the spread of Rift Valley fever worldwide. People appear to be quite capable of acting as
carriers and that, in conjunction with the diseases's apparent ability to "infect" a wide variety of arthropod vectors,
make this disease a conceivable threat worldwide. The impact of a changing climate can only exacerbate the
problem. Mild winters, longer summers, and increased precipitation in coastal areas will contribute to increased
vector populations and seasons, for much of the United States. This will serve to increase the risk of this disease
becoming established in the environment if introduced.
The shorter the generation time for any population, the more effective the population usually is as a vector,
transmitting pathogens to domestic and wild animals. The short generation time of the mosquito permits rapid
11-21
-------
Stem
colonization and facilitates expansion into new areas. Optimum temperature for development is around 25°C
(Horsfall, 1955). At approximately 35°C development halts (Horsfall, 1955).
Several unanswered questions require further study. Rift Valley fever seems to be a disease on the move.
In the last half century, the geographic range of the disease has extended across the African continent even into
such geographically isolated ecological territories as South Africa, the Rift Valley in East Africa, and northern
Egypt. The virus has been able to successfully infect a variety of vectors in these geographic areas. Presently,
there is a real concern for eventual worldwide distribution.
Some concern exists that humans may be a suitable carrier host for the virus into other parts of the world.
It is still not clear if humans, camels, or other yet unknown hosts were responsible for the spread of Rift Valley
fever into North Africa from East Africa. Apparently, humans become viremic (carriers of virus particles in their
blood); this has important ramifications concerning the spread of this disease (Miller, 1980). A possible scenario
is that if someone in an endemic area at the outset of an epizootic was bitten, and traveled to another country
while in the incubation period, he could introduce the virus in local indigenous vector arthropods and eventually
to domestic and wild animals, and to other humans. There is, therefore, real concern that humans could act
as carriers to biting arthropods (Miller, 1980).
The impact of a changing climate, especially a wanning one, would be that if the virus was in the vector
population during a time when there was a great increase in vector numbers, as could be expected along the
costal areas that are expected to warm considerably and remain humid, the spread of the disease would be very
rapid. Culex pipiens. and Aedes aegypti and Aedes taenierhvnchus all are more successful in areas of increased
rainfall and stagnant, and polluted waters. Aedes taenierhvnchus likes to live along coastal areas where there
may be a significant rise in sea level (Horsfall, 1955).
Extended cold weather, 2°C or less is detrimental to larval development and growth. Incubation optimum
temperature lies between 21°C and 30-5°C (Horsfall, 1955). Development throughout the mosquito lifespan is
best in the 25°-30°C temperature span. Extreme cold, 3°C or less, is detrimental to the development of Culex
pipiens (Horsfall, 1955). Therefore, the expected global warming would not only open up new vector habitats
but also increase the rate of vector development and extend the season of vector activity. At temperatures above
25°C, mosquitoes tend to seek out cooler areas, and at temperatures above 30°C, their activity is generally
diminished.
Over-wintering of Culex adults is generally in protected areas, although some adults even can survive
temperatures well below freezing (Horsfall, 1955). It is clear that as the climate warms, over-wintering will
become an important factor in increasing vector population sizes.
Although the complete epidemiologjcal picture of Rift Valley fever is not yet clear, and capable vectors
are known to exist throughout the United States, the disease is more likely to become established in regions that
are sufficiently warm and moist to sustain reproductive and feeding activity on a year-round basis. States where
mosquito activity may be sufficient to permit the maintenance of Rift Valley fever in the host-vector cycles are
shown in Figure 4. It is important to emphasize that Rift Valley fever has never been introduced into the
United States. These states were grouped according to the known vector biology of the mosquitoes that are
proven vectors for this disease. It is possible that other, yet unidentified factors may well prevent the
establishment of the disease in these states. Furthermore, environmental variability within any one state may
preclude establishment of the disease in certain areas. Figure 5 demonstrates the hypothetical expansion of areas
that may be able to support sufficient mosquito vector activity to permit establishment of a reservoir of Rift
Valley fever, should it be introduced into any of these states. This figure incorporates the environmental changes
at 2xCO2 levels as predicted by the GISS and GFDL climate models. The exact distribution of mosquito
populations able to support this disease is beyond both the present knowledge as presented in current literature
and the scope of this study. The important point is that there are areas of the United States that could
potentially support this disease, and these are likely to expand with the environmental changes predicted by these
climate models.
11-22
-------
Stem
CURRENT VECTOR ACTIVITY SUFFICIENT;
FOR DISEASE MAINTENANCE ^
RFT VALLEY FEVER
Figure 4. States where present day mosquito activity may be sufficient to permit the establishment and
maintenance of Rift Valley fever assumptions are based on present day diurnal variations of
mean temperatures for the months of January and July and mean precipitation as described in the
GISS and GDFL climate models, and known vector and agent biology as presented in current
literature. Rift Valley fever does not currently occur in the United States. If introduced through
entry of infected people or animals, the shaded states are those where this disease is more likely
to become established owing to the current vector activity.
11-23
-------
Stem
PROJECTED VECTOR ACTIVITY SUFFICIENT
FOR DISEASE MAINTENANCE
V///7/\
MCflTH Q*!OT>, 1MNNE901&
)
MSOQNSN
''' '////¥/}&//>
/ x /•'
RIFT VALLEY FEVER
Figure 5. States where future (2xCO, levels) mosquito activity may be sufficient to permit the establishment
and maintenance of Rift valley fever: projections are based on diurnal variations of mean
temperatures for the months of January and July and mean precipitation predictions of GISS and
GDFL climate models, and known vector and agent biology as presented in current literature.
Rift Valley fever does not currently occur in the United States. If introduced through entry of
infected people or animals, the shaded states are those where this disease is more likely to become
established owing to the potential vector activity.
11-24
-------
Stem
Culex pipiens is a nonmigratory species, although they have been recorded to travel on wing a distance
of 5 kilometers (Horsfall, 1955).
Wind velocities above 16 km/hr curtail mosquito flights. Flight activity increases as temperatures rise to
25°. Above this, flight activity is correlated to humidity, not to temperature (Horsfall, 1955).
The distribution and success of any populations of mosquitoes, is inextricably interlocked with climate and
the dynamics of day to day variations in weather and environment. For instance, predation on the larvae of
Aedes taenierhynchus is limited by frequent rainfall. Flooding into new flat water pools is not deep enough
for Gambusia puncticulata. Lima cavmanensis. and Rivulus marmoratus minnows. These pools are sufficiently
deep for Aedes taenierhvnchus to live and develop in, and development of the larval forms is done early in the
flooding of new areas. Drier wet seasons are sometimes more favorable for the mosquitoes, for there are not
great increases in the Gambusia populations concurrent with the increases in the mosquito population. Deeper
water disperses Gambusia into such wide areas they are unable to feed effectively on the developing mosquitoes
at the beginning of the wet season, when there are few Gambusia about. This is the time when most mosquito
larvae are developing (Todd and Giglioli, 1983).
Economic Model of the Effects of Global Climate Changes on Rift Vallev Fever
Rift Valley fever is not currently found in the United States. It could, however, be introduced and become
established in several states. The analysis was undertaken for two of these: California and Florida. Based upon
mortality rates of 2% for cattle and 14% for calves and morbidity rates of 8% herd-specific mortality and
morbidity rates were estimated separately for calves and cattle (Blood et al., 1981; Howard, 1986; Gillespie and
Timony, 1981) These rates apply when the disease is first introduced on a large scale and would be reduced
considerably once the disease becomes endemic. Normally it would take several years for Rift Valley fever to
become established to these levels and consequently the analysis is calculated to reflect the effects over a several-
year period. To the losses from mortality and morbidity, as estimated on the basis of the cumulative decline
hi annual production, is added the capital loss associated with the herd as a reproductive and growing out or
fattening unit. The impact of the spread of the disease is estimated by aggregating the losses associated with that
spread across the states in which the disease has become endemic. No estimate is made of efforts to control
the disease nor of the costs of this control. The numbers, therefore, represent a worst case scenario of what
might happen if no control or containment measures were exercised.
The analysis suggests annual production losses associated with the disease infecting cattle equal to $56
million in California and $28 million in Florida. Capital losses would equal $76 million in California and $31
million in Florida. Together these losses would total 12% of the value of production in California and 20% of
its value in Florida (model 3, Annex A).
11-25
-------
Stem
CHAPTER 4
AFRICAN SWINE FEVER
Introduction
African swine fever is a virus that affects wild and domestic swine (Blood et al., 1983; Gillespie and
Timony, 1981). Most feral swine do not show clinical signs of disease. It is however a serious threat to
domesticated swine. At the present time, it is a foreign animal disease and was recently eradicated from the
island countries of the Dominican Republic, Haiti, and Cuba in the Caribbean. The virus belongs to the family
of viruses known as Iridoviridae, although the classification of this virus is currently under review and may be
reclassified as a pox virus (Hess, 1981).
Transmission of the virus can follow several routes. First, and the one historically blamed for spread of
the disease from one region of the world to another, is the feeding of raw pork products and garbage to swine
(Jubb and Kennedy, 1970). Once a focus of infection is established transmission can occur by several routes.
Direct contact between pigs is the most likely route but arthropod vectors, especially soft-shelled ticks, act as
carriers of the disease (Blood et al., 1983).
The impact of the greenhouse effect upon the distribution and activity of tick populations could be
significant enough to cause outbreaks of the disease in a variety of areas. Ornithodoros coriaceous, a tick native
to the United States, has experimentally infected domestic swine (Blood et al., 1983). Haematopinus suis. the
hog louse, is also a probable vector of African swine fever (Blood et al., 1983).
One can envision a scenario where African swine fever is introduced into the UJS. and arthropod vectors
carry the organism into wild pig populations. Once a reservoir such as this is established, direct contact or
arthropod vectors could then sporadically and repeatedly carry the disease back into domestic swine populations
where it would rapidly disseminate into commercial piggeries through direct contact (Degner et al., 1983).
Furthermore, once African swine fever is established in the soft tick population, it may be impossible to eradicate
(Butler and Gibbs, 1984). Apparently, this situation has recently occurred in Spain where the disease has spread
from Africa.
Clinical Signs. Epidemiology, and Vector Biology of the Disease
The disease is an acute generalized viremia that dramatically demonstrates damage to the vascular
endothelium. As a result blood vessels break down, showing hemorrhage and edema as major clinical signs
(Blood et al., 1983). Macrophages and lymphocytes are also invaded by the virus with a resulting lymphopenia
(Blood et al., 1983). In chronically infected animals immune complexes may form, as evidenced by disseminated
intravascular coagualopathies (Blood et al., 1983).
Signs of the disease are quite similar to hog cholera. In fact, it is difficult to distinguish the two in the
field. Laboratory diagnosis in the form of virus isolation is often needed for diagnosis. In African swine fever
as well as hog cholera, there are pyrexia, depression, vomiting, and diarrhea, but swine affected with hog cholera
will not eat (Blood et al., 1983). In African swine fever animals continue to eat. There is no evidence that
African swine fever affects man.
One important aspect of African swine fever, like many diseases that manifest extremely volatile outbreaks,
is that initial epizootics show acute and peracute signs with high mortality. With the passage of time the
outbreaks become less severe, mortality drops, and more diseased animals show chronic signs with a greater
increase in the carrier animal state. In these cases the damage, economically and biologically, is less direct but
no less debilitating.
11-26
-------
Stem
Economic Importance of African Swine Fever
If African swine fever were to be introduced into the United States, the effects on the swine industry
could be truly disastrous. Recently this disease was introduced into the island of Hispaniola. In order to
eradicate the disease from the island before it was established in the soft tick population, plans were made and
executed to depopulate the entire island of swine. Following this, sentinel swine were introduced for a period
of time to be sure that the disease had been eradicated Following this, the island was repopulated with swine
from the United States. The level of donor assistance and loans that has been necessary to eradicate African
swine fever from Haiti alone is estimated to be more than $36 million (Mulhern, 1984).
Florida has more feral swine than any other state in the Union (Degner et al., 1983). Census shows
approximately 500,000 head occurring in 66 of 67 counties (Degner et al., 1983). These swine are domestic swine
that have gone wild from herds imported by Spanish explorers (Degner et al., 1983). It has been projected that
an annual loss of $6 million would result if the feral swine population in Florida were to be lost owing to African
swine fever (Degner et al., 1983).
Control of African Swine Fever
There is no effective vaccine or treatment for African swine fever (Meyer, 1971). It has the potential for
decimating swine populations where it is introduced, and once established in the tick vectors, it would be
extremely difficult, if not impossible, to eradicate (Butler and Gibbs, 1984). Fortunately in Haiti and the
Dominican Republic, the disease had not established itself in the Omithodoros vectors to the extent that could
prevent its eradication. Direct costs for eradication of the disease in Haiti have been documented to be more
than $36 million (Mulhern, 1984).
Possible Effects of Global Climate Change on African Swine Fever
Climate has the greatest effect on the vector organism. Warm-blooded vertebrate hosts moderate and
maintain temperature for the virus organism. Too low a temperature induces hibernation with no transmission.
Cold-blooded (poikilothermic) arthropod v ,ctors are unable to regulate temperature and as a result climate plays
an important role in the activity of vectors and the transmission of virus (Sellers, 1981).
If African swine fever is introduced into the United States and becomes established in the Omithodoros
tick reservoir, eradication of this disease may well become impossible. It has been shown that Omithodoros ticks
can live up to 11 years, and infected ticks have been able to transmit the virus to swine up to 8 years following
infection (Butler and Gibbs, 1984). Furthermore, once infected, ticks can spread the virus to their offspring and
through sexual contact, permitting maintenance of the disease in the tick reservoir indefinitely (Butler and Gibbs,
1984; Fraser, 1986). Because this vector is so important in the control and eradication of this disease, any
expansion of its habitat proportionately increases the risk of establishment in the tick population if this disease
is introduced and would be justification for increased animal disease surveillance efforts at our borders.
Changing climates, especially those projected by the GISS climate model, would be conducive to expanding
habitats and populations of the tick vectors and possibly the wild or semi-wild pig populations.
The current distribution of the Qnithodorus species of soft ticks that could support African swine fever
can be seen in Figure 6. Environmental factors seem to be very important limiting the distribution of this tick
(El Shoura, 1987; Butler and Gibbs, 1984). Experimental evidence and the present day natural distribution of
these ticks show that they are able to survive hot weather quite well (El Shoura, 1987; Butler and Gibbs, 1984).
Increases in temperatures enhance digestion, nutrient absorption, and egg maturation, demonstrating the
important effect of temperature on the physiologic activity and reproduction of this tick (El Shoura, 1987).
However, the activity of the Ornithodoros tick is affected by cool weather, as temperatures at or below 22°C
significantly reduce tick activity, larval development, and reproduction (El Shoura, 1987; Cunliffe and Nuttall,
1921). Present day distribution of this tick appears to be limited to areas where temperatures are warm enough
for a sufficient period of time to permit the tick to complete its rather complex life cycle.
11-27
-------
Stem
CURRENT ORNITHODOROS DISTRIBUTKEZ3
AFRICAN SWNE FEVER
Figure 6. Present day distribution of African swine fever vectors: assumptions are based on present day
diurnal variations of mean temperatures for the months of January and July and mean
precipitation as described in the GISS and GDFL climate models, current vector distribution
(Butler and Gibbs, 1984), and known vector and agent biology as presented in current literature.
African swine fever does not currently occur in the United States. If introduced, through the entry
of infected animals, uncooked garbage, wild swine, or other means, the shaded states are those
where this disease is more likely to become established owing to present vector distribution.
11-28
-------
Stem
The environmental changes predicted at 2xCO2 levels of the GISS and GFDL models would permit an
expansion of the geographical distribution of this vector species of African swine fever. Generally, the
distribution could be expected to expand in a northern direction, possibly extending to the northern limits of the
southern states, where, if the climate model predictions are realized, the tick should be able to complete its life
cycle (Figure 7). It is important to note that these projections are based on what is currently known about the
biology of this vector and the predictions made by the GISS and GDFL climate models. It is possible that as
yet undefined environmental or other limitations may restrict the distribution of this tick species, as further work
is necessary to fully elaborate the factors limiting its distributioa The tick is likely to expand into areas where
over-wintering becomes possible owing to the occurrence of milder winters in these areas. If the Onithodorus
tick habitat expands northward, the implications are serious for the control and eradication of African swine fever
if it is introduced into the United States.
Economic Model of the Effects of Global Climate Changes on African Swine Fever
The economic losses associated with the introduction of African swine fever into the United States, but
under the assumption that various control measures are used to limit its spread, have been estimated by
McCauley and Sundquist (1979). These estimates could be used to measure the increase in costs that might be
incurred as a result of the projected changes in climate.
The economic consequences of African swine fever and its control in the United States were estimated
under three alternative scenarios:
(1) A small outbreak of the disease is controlled and the disease is successfully eradicated through
slaughter and eradication.
(2) Widespread outbreaks result in the quarantine of a larger area (e.g., a state) for a longer period of
time (e.g., three years).
(3) The disease becomes endemic, and a program is mounted to reduce its incidence and eventually to
eradicate it over a period of several years.
The first of these scenarios is estimated to cost about $73 million, of which $3 million (1977 market
prices) would be used to indemnify owners of slaughtered pigs. The cost for the second scenario would be about
$152 million, of which $63 million would be for indemnification. Under the third scenario, the program would
cost $290 million ($211 in indemnities) over five years and $559 million ($362 in indemnities) if it had to be
carried out over ten years. In addition, there would be additional losses to consumers under the third scenario
equal to about $23 billion.
Suggested Topics For an In-depth Study on the Effects of Greenhouse-Induced Climate Changes on Pattern of
Animal Health
1. Consideration should be given not only to animal agriculture in the United States but also to livestock
production in the world to be able to arrive at clear implications for the American livestock industry.
2. The effects of these changes on the poultry industry should be addressed in detail.
3. The effects on companion animals and animals kept for sport should be addressed.
4. Control costs should be projected for diseases examined in detail as well as the implications of the
introduction, establishment, or expansion of a disease on consumers of livestock products.
5. The potential for utilizing emerging technologies such as biotechnology and genetic engineering should be
examined in light of developing disease and heat stress-resistant animals.
11-29
-------
Stem
PROJECTED ORNITHODOROS DISTRIBUTION
XX\WW¥VWWWW^^
\XxW\vNf/<^\fVItf^^4«W"5^'Ofx xXXxx'
x x V Y SfV\' W x yV^^L
^S66rt^WWWSK^
AFRICAN SWNE FEVER
Figure 7. Projected future (2x00, levels) distribution of African swine fever vectors: projections are based
on diurnal variations 67 mean temperatures for the months of January and July and mean
precipitation predictions of GISS and GDFL climate models, current vector distribution (Butler
and Gibbs, 1984), and known vector and agent biology as presented in current literature. African
Swine Fever does not currently occur in the United States. If introduced through the entry of
infected animals, uncooked garbage, wild swine, or other means, the shaded states are those where
this disease is more likely to become established owing to the potential vector distribution.
11-30
-------
Stem
6. Livestock management techniques should be reviewed so that recommendations may be made on reducing
production losses and risks of disease introduction to farms.
7. One limitation of the analysis presented here is that it does not take into account many of the dynamic effects
of the spread of disease. This would require the construction of herd projection models that would simulate the
impact of changing mortality and morbidity over time. It would also involve discounting the costs associated with
the disease so as to take into account the time value of money. Finally, future research should focus on the
effects of disease control measures and their costs in relation to the costs associated with production losses and
the decline in the capital invested in animal herds.
11-31
-------
Stem
REFERENCES
Allen, Jon, C. and Daniel Gonzalez. The Effects of Temperature and Host Refuge on Insect Host-Parasite
Models." Environmental Entomology. Vol. 4, No. 1: 57-65, 1975.
Anderson, J.F. "Epizooliology and Experimental Studies of Bluetongue Virus in Cattle." Master's Thesis,
University of Minnesota, 1970.
Arthur , Don, R. Ticks: A Monograph of the Ixodidae. Cambridge At The University Press, 1960. Part V,
pp. 38-63.
Arthur, Don, R. Ticks: A Monograph of the Ixodidae. Cambridge At The University Press, 1960. Part V,
pp. 205-225.
Ayoub, M., Barhoma, G., and Zaghlol, I. "The Fundus Oculi and the Retina Ocular Manifestations of Rift
Valley Fever." Bull. 71(75) Qphthalmol. Soc.. Egypt, 1978. pp. 125-133.
Barnard, Donald, R. "Seasonal Changes in Feeding on Cattle and Reproduction by Amblvomma americanum
(Acari: Ixodidae) Under Held Conditions." J. Med. Entomol. Vol. 25, No.l: 20-25, 1988.
Barnard, D.R., and Jones, R.H. "Culicoides Variipennis: Seasonal Abundance, Overwintering, and Voltinism in
Northeastern Colorado." Environmental Entomology. 9(5): 709-712, 1980.
Belozerov, V.N. "Diapause and Biological Rhythms in Ticks," In: Physiology of Ticks. Frederick D. Obenchain,
and Rachel Galun, eds. Pergamon Press, Oxford, 1981 pp. 469-500.
Blood, D. C., O. M. Radostits, J. A. Henderson. Veterinary Medicine. Bailliere Tindall, London. 6th ed. 1983.
1310 pp.
Boorman, J. "Some Aspects of the Biology of Major Vectors of Bluetongue Disease in Regions Other Than the
USA," In: Round Table Tech Discussion Bluetongue. College Park, MD, 1979. pp. 1-21.
Bowman, Kenneth, P. "Global Trends in Total Ozone." Science. Vol. 239: 48-50. 1988.
Butler, J.F. and E.P. J. Gibbs. "Distribution of Potential Soft Tick Vectors of African Swine Fever in the
Caribbean Region." Prev. Vet. Med. 2:63-70. 1984.
Callis, J. "Bluetongue in the United States." in: Bluetongue and Related Orbiviruses. Alan R. Liss, Inc. 1985.
pp. 37-42.
Campbell, A., and Harris, D.L. "Reproduction of the American Dog Tick, Dermacentor variabilis1, under
Laboratory and Field Conditions2." Entomological Society of America. 8(4): 734-739, 1979.
Clifford, Carleton M., Anastos, George, and Elbl, Alena, "The Larval Ixodid Ticks of the Eastern United States
(Acarina-Ixodidae)." Miscellaneous Publication of the Entomological Society of AP1prir^i 5: 213-237, 1961.
Cunliffe, N. and G.H.F Nuttall. "Some Observations on the Biology and Structure of Ornithodoros Moubata.
Murray." Parasitology 13:327-344.1921.
Degner, R.L., Rodan, L.W., Mathis, W.K., and Gibbs, E.PJ. "The Recreational and Commercial Importance
of Feral Swine in Florida: Relevance to the Possible Introduction of African Swine Fever into the U-SA.,"
Preventive Veterinary Medicine. 1: 371-381, 1983.
11-32
-------
Stem
Della-Porta, AJ., McPhee, DA., Wark, M.C., St. George, T.D., and Cybinski, D.H. "Serological Studies of Two
Additional Australian Bluetongue Virus Isolates CSIRO154 and CSIRO156." Vet. MicrobioL 6(3): 23-245,
1981.
Detwiler, R.P., and Charles A. S. Hall. Tropical Forests and the Global Carbon Cycle." Science. VoL 239:42-
47.1988.
El Shoura, S. M. "Effect of Temperature and Relative Humidity on the Live Cycle of Ornithodoros
(Pavlovskvella) Erraticus (Ixodoidea: Argasidae)". J. Parisitologv. 73:1102-1108.1987.
Ezeifeka GO, Umoh JU, Belino ED, Ezeokoli CD. "A Serological Survey for Rift Valley Fever Antibody in Food
Animals in Kaduna and Sokoto States of Nigeria". Int J Zoonosis. 9:147-151, 1982.
Fagbami AH, Tomori O, Kemp GE. "A Survey of Nigerian Domestic and Wild Animals for Serum Neutralizing
Antibody to Indigenous Rift Valley Fever Virus". Nigerian Vet J. :2:45-48, 1973.
Feldman - Muhsam, B. "Copulatory Behavior and Fecunity of Male Ornithodoros Ticks." Recent Advances in
Acarologv. 2:159-166, 1978.
Foster, N.M. and R.H. Jones. "Multiplication Rate of Bluetongue Virus in the Vector Culicoides variipennis
(Diptera: Ceratopogonidae) Infected Orally". J. of Med. Entomol. 15(3):302-303. 1979.
Fraser.Clarence M., ed. Merck Veterinary Manual. Merck and Co. Rahway. 6th ed. 1986. 1677 pp.
Freed, I. "Rift Valley Fever in Man Complicated by Retinal Changes and Loss of Vision." South Afr. Med. J..
25:930-932, 1951.
Gillespie, J.H., and J. F. Timony. Hagen and Bruner's Infectious Diseases of Domestic Animals. Cornell Univ.
Press. Ithaca. 1981.
Gillett,J.D. Mosauitos. Weidenfeld and Nicolson, London, UK, 1971. pp 208-240.
Goodger, W. J., Tim Carpenter, Hans Riemann. "Estimation of Economic Loss Associated with Anaplasmosis
in California Beef Cattle" J. Am. Vet. Med. Assoc. 174:1333-1336. 1979.
Hess, William R. "African Swine Fever A Reassessment." in Advances in Veterinary Science and Comparative
Medicine. 25:35-69.1981.
Hansen, J., G. Russell, A. Laos, I. Fung, D. Rind., and P. Stone. "Climate Response Times: Dependence on
Climate Sensitivity and Ocean Mixing," Science. VoL 229: 857-859, 1985.
Hoogstraal, H., Meegan, J.M., and Khalil, G.M., and Adham, FJC "The Rift Valley Fever Epizootic in Egypt
1977-1978 Ecological and Entomological Studies," Trans. Roval Society Medical Hygiene. 73 (6): 624-629, 1979.
Horsfall, Wm. R. Mosquitoes - Their Bionomics and Relation to Disease. The Ronald Press Co., N.Y., 1955.
pp. 476-589.
Howard, J. L. Current Veterinary Therapy, food animal practice. W. B. Saunders Co. Philadelphia. 2nd ed.
1986. 1008pp.
Hussein, Hussein, S., and Mustafa, Bedawi E. "Temperature and Humidity Effects on the Life Cycle of
Haemaphysalis Spinulosa and Rhipicephalus Simus (Acait Ixodidae)." J. Med. Entomol.. Vol. 24: No. 1: 77-
81,1987.
11-33
-------
Stem
Jaranback, Hugo. The Culicoides of New York State (Diptera: Ceratopogonidae)." Bull No. 399. New York
State Museum and Science Service, New York, New York, 1965. pp. 110-113.
Jarvis, Level. "Contributions of Animal Agriculture to National Economies" at Animal Agriculture Symposium:
Development Priorities Toward the Year 2000. June 1-3 1988. Chantilly VA.
Jenne, Roy. "Print of U.S. Data from GISS Model." NCAR/EPA Project Document 2. October 1987.
Jones, R.H. "Biting Flies Collected from Recumbent Bluetongue - Infected Sheep in Idaho." Mosauito News.
41(2): 183, 1981.
Jubb, K. V. F. and Peter C. Kennedy. Pathology of Domestic Animals. Academic Press. New York. 2nd ed.
1970. 2vols. 1290pp.
Kaschula, V.R. "Rift Valley Fever." Proc. 1956 Reg. Meet. Foreign Anim. Pis.. 1956. pp. 113-118.
Kasting, James F., Thomas P. Ackerman. "Climatic Consequences of Very High Carbon Dioxide Levels in the
Earth's Early Atmosphere." Science. Vol. 234: 1383-1385, 1986.
Kemp, D.H., Stone, B.F., and Binnington, K.C. Tick Attachment and Feeding: Role of the Mouthparts, Feeding
Appartus, Salivary Gland Secretions and the Host Response." In: Physiology of Ticks. Frederick D. Obenchain,
and Rachel Galun, eds. Pergamon Press, Oxford, 1982. pp. 119-168.
Kerr, Richard, P. "Is the Greenhouse Here?" Science. VoL 239: 559-561, 1988.
Kerr, Richard, A. " Carbon Dioxide and the Control of Ice Ages." Science. Vol. 223:1053-1054, 1984.
Kerr, Richard, P. The Deepest Antarctic Ozone Hole Ever Seen." Science. VoL 239, 46 pp. 1988.
Kerr, Richard, A. "How Much Drying from a Greenhouse Warming," Research News. 435 pp. 1987.
Kerr, Richard, A. "Carbon Dioxide and a Changing Climate." Research News. 1983.
Kirk, B. "Potential Impacts of African Swine Fever." Agricultural Canada Market Economic Branch, Canada,
1980, 43 pp.
Knulle, W^and Rudolph, D. "Humidity Relationships and Water Balance of Ticks." In: Physiology of Ticks.
Obenchain, Frederick, D., and Rachel Galun. eds. Pergamon Press, Oxford, 1982. pp. 43-70.
Kocan, Katherine, M. "Development of Anaplasma Mareinale Theiler in Ixodid Ticks: Coordinated
Development of a Rickettsial Organism and its Tick Host." In: Morphology Physiology & Behavioral Biology
of Ticks. Sauer, John R., and Hain, J. Alexander, eds. John Wiley & Sons, 1986.
Mariner, J. "Epidemiological Research on Rift Valley Fever" in Final Report of Activities of the Vaccine
Development Specialist. Niger Integrated Livestock Production Project, Tufts University, Republic of Niger,
U.SA.I.D. N. Grafton, MA. 1988. pp. 27-29.
Mariner, J., E. (Chip) Stem, S. Salifou, C. Mamini, K. Baare, R. Yedloutschrig, C. Mebus, A. Soiled.
"Serosurvey of Sahelian Cattle for Evidence of Epizootic Disease". Prev. Vet. Med. In press.
McCauley, E.H. and W.B. Sundquist, "Potential Economic Consequences of African Swine Fever and Its Control
in the United States," Staff Paper 79-11, Department of Agricultural and Applied Economics, University of
Minnesota, April 1979.
11-34
-------
Stem
McCauley, E. Hunt and W. B. Sunquist. A Study of the Potential Economic Impact of Foot and Mouth Disease
in the United States. Superintendent of Documents, U.S. Government Printing Office. Washington, D. C. 1977.
241pp.
McCosker, P J. "Global Aspects of the Management and Control of Ticks of Veterinary Importance," Recent
Advances in Acarologv. Vol. II: 45-54, 1978.
McCoy, John H. and M.E. Sarhan, Livestock and Meat Marketing. New York: Van Nostrand Reinhold, 1988.
Mclntosh, B.M., P. G. Jupp, I. Dos Santos, and B. J. Barnard. "Vector Studies on Rift Valley Fever Virus in
South Africa." S. Afr. Med. J.. 5§ (3): 127-132, 1980.
Mclntosh, BJVI. "Mosquitoes as Vectors of Viruses in Southern Africa." Entomology Memoir Department of
Agricultural Technical Services Republic of South Africa. 43:1-19. 1975.
Metcalf, H.E and AJ. Leudke. "Epizootiology of Bluetongue in the United States of America." Bull. Off. Int.
Epiz. 92:507-513.1980.
Metcalf, H.E., Pearson, J.E., and Klingsporn, A.L. "Bluetongue in Cattle: A Serologjc Survey of Slaughter Cattle
in the United States." Am. J. Vet. Res.. 42(6): 1057-1061, 1981.
Meyer, NX. "Paper No. 11-African Swine Fever." Proc. Indus. Advis. Group FMD, Washington, D.C., 1971,
pp. 45-47.
Miller, Laury, Robert E. Cheney, and Bruce C. Douglas. "GEOSAT Altimeter Observations of Kelvin Waves
and the 1986-87 El Nino." Science. Vol. 239: 52-54, 1988.
Miller, JA. "African Epidemic Where is it Heading." Science News. Vol. 117 (11): 170-171, 1980.
Mitchell, Lee. "Time-Segregated Mosquito Collections with a CDC Miniature Light Trap." Mosquito News.
VoL 12, No. 1, 1982.
Monastersky, R. "Rising sea levels: Predictions and plans." Science News, p. 326. 1987
MorriL, J. C, G. B. Jennings, H. Caplen, MJ. Turell, AJ. Johnson, and CJ. Peters. "Pathogenicity and
immunogenicity of a mutagen-attenuated Rift Valley fever virus immunogen in pregnant ewes." Am J. Vet. Res.
48(7)1042-1047,1987.
Mulhern, F. "Update on African Swine Fever in Haiti." In 1984 Official Proceedings of the Livestock
Conservation Institute. Crofton, MD. 1985. pp. 235-237.
Mullens, BA, and DA. Rutz. "Age Structure and Survivorship of Culicoides Variipennis
(DipteraCeratopogonidae) in Central New York State, U.SA." J. Med. Entomol. 21(2):194-203. 1984.
Murphy, Robert, J. press report: "Scorching temperatures grip Dixie." United Press International. July 9,1986.
Owen, Robert, M., and David K. Rea. "Sea-Floor Hydrothermal Activity Links Climate to Tectonics: The
Eocene Carbon Dioxide Greenhouse." Science. VoL 227:166-169, 1983.
Peters, Robert, 1*, and Joan D.S. Darling. "The Greenhouse Effect and Nature Reserves." BioScience. Vol.
35: No. 11., 707-717. 1985.
Raloff, J. "Pacific's CO2 levels: Cause for concern?" Science News. VoL 129: p 198. 1988.
11-35
-------
Stem
Rayah, EL, Groun, G., and NA. Abu. "Effect of Temperature on Hatching Eggs and Embryonic Survival in the
Mosquito Culex Quinquefasdatus." Entomol. EXP. Aool.. 23(3): 349-351, 1983.
Rebhun, William C. "Other Infectious Diseases" in: 1984 Yearbook of Agriculture: Animal Health - Livestock
and Pets. UJS. Government Printing Office. 1984. p. 172.
Rind, D. and R. Goldberg. "Change in Climate Variability in the 21st Century." Goddard Space Flight Center,
New York, 1988.
St. Louis, M.E., D.L. Morse, M.E. Potter, T.M. DeMelfi, JJ. Guzewich, R.V. Tauxe, and PA. Blake. The
emergence of grade A eggs as a major source of Salmonella enteritidis infections. New Implications for the
Control of Salmonellosis. J. Am. Med. Assoc. 259(14):2103-2107, 1988.
Saulmon, E. E. "Control of Foot and Mouth Disease in the United States." J. Am. Vet. Med. Assoc. 154(1):40-
41, 1969.
Schrine, L. "Muscular Changes in Rift Valley Fever." South Afr. Med. J.. 25: 926-930, 1951.
Sellers, R.F. "Bluetongue and Related Diseases." In: Virus Diseases of Food Animals. E.PJ. Gibbs, ed.
Academic Press, London, UK, 1981. pp. 567-584.
Sellers, R.F., Gibbs, E.PJ., and Herniman, KAJ. "Possible Origin of the Bluetongue Epidemic in Cyprus
August, 1977." In: Round Table Tech Discussion Bluetongue. College Park, MD, 1979a, pp. 1-21.
Sellers, R.F., Gibbs, E.PJ., Herniman, KAJ., Pedgle, D.E., and Tucker, M.R. "Possible Origin of the
Bluetongue Epidemic in Cyprus August 1977." J. Hvg.. 83: 547-555, 1979b.
Simpson, D.I.H. "Viral Haemorrhagic Fevers in Man." Bull. 56 (6): World Health Organization, 1978. pp.
819-832.
Singh, E.L., Eaglesome, M.D., Thomas, F.C., Papp-Vid, G., and Hare, W.C.D. "Embryo Transfer as a Means
of Controlling the Transmission of Viral Infections. I. The in Vitro Exposal of Preimplantation Bovine Embryos
to Akabare, Bluetongue, and Bovine Viral Diarrhea Viruses." Theriogenology. 17(4): 437-444, 1982.
Stone, Alan, Curtis W. Sabrosky, Willis W. Wirth, Richard H. Foote, and Jack R. Coulson, "A Catalog of the
Dipteraof America North of Mexico." Smithsonian Institution Press, Washington, D. C., 1983. pp. 104-142.
Todd, G.R., and Giglioli, M.E.C. The Failure of Gambusia Puncticulatea and Other Minnows to Control Aedes
Taenierhvnchus in a Mangrove Swamp on Grand Cayman, W.I." Mosauito News. 43(4): 419-425, 1983.
Turrell, Michael, J., and Charles L. Bailey. Transmission Studies in Mosquitoes (Diptera Culicidae) With
Disseminated Rift Valley Fever Virus Infections1." J. Med. Entomol.. Vol. 24, No. 1: 11-18, 1987.
Turrell, Michael, J., and Charles L. Bailey, press report: Agricultural losses attributed to the 1986 drought and
heat wave in the Southeastern U.S. United Press International. July 28, 1986.
Turrell, Michael, J., and Charles L. Bailey. "USDA Surveys Cattle for Bluetongue Disease." USDA News, pp.
124-81(1-2), 1981.
Waladde, S.M., and Rice, M J. The Sensory Basis of Tick Feeding Behavior." Ire Physiology of Ticks.
Frederick D. Obenchain, and Rachel Galun, eds. Pergamon Press, Oxford, 1982. pp. 71-118.
Walsh, John. "Rift Valley Fever Rears its Head." Science 240:1397-1399. 1988.
11-36
-------
Stem
Walton, T.E., T.L. Barber, R. H. Jones and A. J. Luedke. "Epizootiology of Bluetongue Virus" Transmission
Cycle Vectors, and Serotypic Distribution in the Americas." Prev. Vet. Med. 2:379-388. 1984.
Walton, GA. "A Taxonomic Review of the Ornithodoros Moubata Species in Africa." Recent Advances in
Acarology. % 49-500, 1978.
Watts, Douglas M., Gary G. Clark, C. L. Crabbs, Cynthia A. Rossi, Timothy R. Olin, and Charles L. Bailey.
"Ecological Evidence Against Vertical Transmission of Eastern Equine Encephalitis Virus by Mosquitoes
(Diptera: Culicidae) on the Delmarva Peninsula, USA1." J. Med. EntomoL Vol. 24: No. 1: 91-98, 1987.
Weisburd, S. "More grounding for global warming." Science News, p.87 1986.
Weisburd, Stefi. "Waiting for the Warming: The Catch-22 of CO~" Science News. Vol. 128: pp. 170 & 174.
1987.
Wikel, Stephen, K., and Allen, John, R. "Immunological Basis of Host Resistance to Ticks." In: Physiology of
Ticks. Frederick D. Obenchain, and Rachel Galun, eds. Pergamon Press, Oxford, 1982. pp. 169-196.
Wirth, W.W. "Biology of Major Bluetongue Vectors in
the United States." In: Round Table Tech. Discussion Bluetongue. College Park, MD, 1979. pp. 1-12.
Zimmerman, R.H., Barker, S J., and Turner, Jr., E.C. "Swarming and Mating Behavior of a Natural Population
of Culicoides Variipennis (Diptera: Ceratopogonidae)1." J. Med. EntomoL. Vol. 19, No. 2:151-156, 1982.
11-37
-------
Stem
ANNEX A: ECONOMIC TABLES
NOTES TO TABLES
(a) USDA, Meat Animals Production, Deposition and Income and USDA, Cattle, Sheep.
(b) Actual deaths plus mortality rates multiplied times inventory.
(c) Derived from the same sources cited in Note (a).
(d) Actual production minus average production loss due to mortality and morbidity over the course
of a year.
(e) See note (a).
(0 Actual value of production from the same sources cited in Note (a). Projected value of production
obtained by multiplying projected production times average price.
(g) Actual minus projected value of production.
(h) Projected minus actual deaths multiplied times average weight and price.
11-38
-------
llr FirMlltri:
InlM norlilllr rilr
cillli fil»ei
I LSI IS.Ot
I
Hwftldltr rili
Clttlf cilvn ntlfhli: |n 100 Ibil
11.01 IS.9I cittli 11.5
cilvii MS
emu MO «iv»
Still
til
III!
Irli
Irl
tllll
Clll
Cm
Oil
ru
Cl
Kl»
Ml*
III
Inl
lovi
Iini
IT
U
Kline
NO
RIM
Rich
linn
Kin
id
Nwil
libr
lit
in
11
1 Nil
II
K
IOU
OMl
Okll
9rti
H
II
SC
! (il
liin
In
tttih
It
vj
•lift
« tl
III
U.I
Imintirr
Cittli
1.000 NO
1.151
1.7S1
711
1.17?
I. ISO
1.105
11
IS
1.115
I.IH
ISO
1.115
1.710
1.010
1.110
1.515
MSI
tit
11
HI
It
.015
.111
.015
.110
.111
.170
111
10
71
1.017
I.S15
ill
1.511
I.MO
1.111
1.971
l.!»t
5.7
171
1.IJ1
.755
10.500
115
115
I.ISO
I.O'O
111
1.115
I.OH
7m. 1117 III
Cil.ci
1.090 MO
110
1.70
111
111
1.100
its
II
1
IIS
111
15
115
510
ItO
1.110
1.110
Sit
HI
17
II
II
100
177
111
1.110
111
1.110
151
11
II
111
111
HI
1S7
110
1.145
111
Ml
1
111
IK
IIS
1.190
US
150
110
no
in
17S
HI
Mull
Clltll
1.000 NO
H
0
11
IS
10
SS
1
1
IS
11
1
II
IS
H
10
10
47
17
1
5
1
11
SO
I]
IS
It
10
7
1
1
11
10
10
II
IS
10
19
I]
0
II
!5
.'$
110
IS
II
21
IS
11
SO
30
Oiilhi 111
CllVfl
1.000 NO
II
1
10
11
111
70
1
1
10
17
1
70
SO
It
110
IIS
11
11
1
11
1
70
110
IS
110
ts
IIS
IS
1
1
11
17
11
SS
45
no
50
79
0
11
110
10
110
11
17
70
15
IS
315
11
Frijicttd Dttlhi Itl Mull rroductlM Icl
Clltll Cllm Clltlt Cilvii
1.000 HO
171
0
H
111
70
1)
1
|
11
110
i
II
IS
11
10
10
III
11
1
S
1
11
50
110
111
111
10
1
1
1
II
10
71
II
IS
10
111
J]
0
SI
155
1'.)
170
IS
II
;j
15
IS
!0
IS
1,000 HO
101
1
10
111
111
70
1
1
10
III
1
70
SO
It
IM
115
111
IS
1
11
4
70
110
U
111
It
IIS
IS
1
1
11
17
71
11
1!
110
9»
?c
0
u
151
117
159
11
57
70
IS
IS
115
ri
1.000 111
417.111
1.510
510.715
570.111
1. 111. 101
1.417.110
11.150
7,111
US. 110
111.110
11.111
101.111
711.111
151. lit
1.111. ISO
1.015.7)5
541.774
111. 101
11.117
11.510
11.110
Ill.lll
1. 101.171
111.171
1.051.144
100.141
I. 110. SIO
111.117
11.010
1,040
SIS. 701
157.011
711.131
457.141
515.110
3.014.311
151.515
591. JOS
1175
SOS. 3)1
1. IH.II1
411.010
1.147.511
151.101
45.717
111.711
537. ))»
15.541
1. 151. 115
111.147
1.000 US
57.111
IIS
0
17.711
71, Itl
1
1.100
HI
111,110
111.151
l.ltl
I). 701
1.111
41.014
0
0
110. ill
134.441
Mil
11.415
1.100
11. Ill
11.117
110. Ill
Ill.lll
17.111
0
11.111
1.100
10.111
14.117
15.111
11.771
11,111
IS, 100
35,151
14.779
41,157
7!l
37 4)7
10.111
1S.7I5
77.
-------
ltdll I 1: IM«I l"< lllti Illut Icniuil
Id rViKtin:
Itllwi "orlillli nit
Itorbldlti rill
|h»»i liibi •tliMi: In 100 IMl
11.01 I! 0» iKftD I )0
tub O.JO
mo unil
Mill
Illl
• Ml
(Illl
(III
COM
liiM
III
M
IMI
Ilnl
II
U
Ailni
BO
Kill
fllch
mm
n»
nenl
lebr
• it
-------
Jodtl I 1- CattU '«lft Valler F-verl
Ker Paraieltri:
<««ion rortjlltr rat«
ctttlt cilvei
llorbiditi rut
cittle calvei xeijhti: in 100 Ibil
1 2.01 II. Ot
2
3
•talc Invcntorr Jan. 19!7 lit
Cattle Calvci
1.000 M l.OOO HO
Calif 3.450 1.100
Ma 1.72) Hi
!.0t
Act'Jal
Cattle
1.000 HO
40
35
l.flt
CAtlie AND
Oiathi lal
Calvei
1.000 HD
1(2
(0
cittle
calvei
CALVES
12. i
2.:5
Projected Deathi (Dl Actual Production Id
Cattle Calvei Cattle Calvei
1.000 HO
163
70
1.000 HO 1.000 IBS 1.000 l!S
316 1.131. 101 7l.it?
9} 125.120 29V750
Projected Production Idl Average Prlct
oer 100 oound i lei
Cattle Calvei Cattle Calv*a
1.000 IBS 1,000 IK USI USt
1,741,67! 61.0JJ 56.30 71.60
118, 961 261.156 (7.60 79.20
total Value of Production If) Value of
oroiected
Actual Projected production
leu lo)
1.0001 l.COO! 1,0001
1.0)2.117 1.026.112 56,01)
296.9)4 269.211 27.71)
Value of
herd
capital
lott Ihl
1.0001
76.IS3
!0,!81
-------
EFFECT OF CLIMATIC WARMING ON POPULATIONS OF THE
HORN FLY, WITH ASSOCIATED IMPACT ON WEIGHT GAIN AND
MILK PRODUCTION IN CATTLE
by
E. T. Schmidtmann
Livestock Insects Laboratory
and
J. A. Miller
US. Livestock Insects Laboratory
Agricultural Research Service, USDA
Beltsville, MD 20705 and
Kerrville, TX 78028
Contract No. W12932972-01-0
-------
CONTENTS
FINDINGS 12-1
CHAPTER I: INTRODUCTION 12-2
DESCRIPTION OF THE ECOLOGICAL SYSTEM 12-2
Horn flies, Cattle, and Economic Losses 12-2
Horn Flies and Temperature 12-2
Climate and Insect Population Models 12-2
CHAPTER 2: METHODOLOGY 12-3
THE EFFECTS MODEL 12-3
Description of the Model 12-3
Limitations Resulting from the Model 12-3
THE SCENARIOS 12-3
Scenario Used 12-3
Estimation of Economic Loss 12-3
CHAPTER 3: RESULTS 12-6
HORN FLY PREDICTIONS: CHANGES, CAUSES, AND EFFECTS 12-6
Horn Fly Density and Seasonal Activity 12-6
Cause of Changes in Horn Fly Populations 12-6
Effect on Weight Gain in Beef Cattle 12-6
Summer Suppression of Horn Flies 12-6
Horn Flies and Dairy Cattle 12-7
CHAPTER 4: IMPLICATIONS OF RESULTS 12-11
ECONOMIC IMPLICATIONS 12-11
Beef Cattle 12-11
Dairy Cattle 12-11
SOCIOECONOMIC IMPLICATIONS 12-11
REFERENCES 12-12
-------
Schmidtmann
FINDINGS1
Climatic wanning predicted under a doubling of atmospheric carbon dioxide (GFDL, 2xCO2) will have
a pronounced effect on populations of the horn fly, a widespread pest of pastured cattle. Throughout most of
the United States, annual adult horn fly activity will commence 4 to 6 weeks earlier in the spring, average 100
to 400 flies per head higher through the summer, and strongly persist 4 to 6 weeks longer in the fall. This
condition will (1) cause substantial reductions in the seasonal average weight gain of growing beef cattle (range
2 to 12 kg/head), and (2) represent a potentially important new stress on dairy cattle in northern dairy regions.
In contrast, horn fly populations from Mississippi west to Texas and southern California will be suppressed by
maximum summer temperatures; this will result in weight-gain losses in beef cattle that are less than those
currently experienced, especially in Texas.
'Although the information in this report has been funded wholly or partly by the U.S. Environmental
Protection Agency under Contract No. W12932972-01-0, it does not necessarily reflect the Agency's views, and
no official endorsement should be inferred from it.
12-1
-------
Schmidtmann
CHAPTER I
INTRODUCTION
DESCRIPTION OF THE ECOLOGICAL SYSTEM
This assessment of the impact of 2xCO2-induced climatic wanning on animal health focuses on the
relationship between climate, an ectoparasite (the horn fly), and growth and milk production in domestic cattle.
Horn Flies. Cattle, and Economic Losses
The horn fly, Haematobia irritans L., is a ubiquitous pest of pastured cattle throughout the United
States, commonly attaining levels of 300 to 500 flies per head on untreated stock. Adult horn flies live on cattle
and derive nutrition by blood feeding from their host ca. 20-30 times per day (Harris et al., 1974). This activity
causes annoyance and irritation that is reflected behaviorally as head throwing, skin rippling, and tail switching
(Harvey and Launchbaugh, 1982). Economic loss attributable to the horn fly is expressed as a reduction in
weight gain and milk production, as well as less efficent feed conversion (Steelman, 1974). Estimates of horn
fly-induced reductions in average daily weight gain (ADG) in beef cattle range from negligible to 0.2 kg per head
per day (Drummond, 1987). Losses in milk production due to horn flies have been estimated at 80 million
dollars annually (Anonymous, 1976), or one percent of total milk production (Drummond et al., 1981). Recently,
the collective loss in beef and dairy cattle production due to horn flies was estimated at 7303 million dollars
annually, irrespective of control costs (Drummond, 1987).
Horn Flies and Temperature
Like many insects, the phenology and density of hom fly populations are regulated largely by
temperature (Bruce, 1964). Annual adult horn fly activity commences with wanning ambient temperature in
spring (March-MayX and thereafter successive overlapping generations are produced until decreasing
temperature and daylength in the fall trigger an overwintering diapause. In general, the climatic conditions
projected for a doubling of carbon dioxide in the atmosphere can be anticipated to (1) extend the seasonal
duration of horn fly activity during both spring and fall months, and (2) accelerate horn fly population
growth/activity, with the result that greater numbers of flies will occur on cattle during favorable periods; also,
extreme summer temperatures may be lethal.
Climate and Insect Population Models
The use of deterministic, climate-driven models to predict the seasonal abundance of insect populations
is an accepted procedure widely used to estimate population growth in insects (Haynes and Tummala, 1976).
Such models are typically based on life-stage-specific, temperature-dependent growth rates, and can be run with
real or substituted temperature values. This report describes the application of a deterministic population
model (Miller, 1986) to estimate the effect of 2xCO2-induced climatic warming on populations of the horn fly;
the results are used to estimate reductions in weight gain for beef cattle and milk production in dairy cattle.
12-2
-------
Schmidtmann
CHAPTER 2
METHODOLOGY
THE EFFECTS MODEL
Description of the Model
The abundance of horn flies under 2XCO2-induced climatic warming (GFDL output) was predicted on
a monthly basis using the horn fly population model of Miller (1986). The model is based on life-stage-specific
growth and survival characteristics of the horn fly, as determined by temperature and rainfall. Monthly horn fly
population levels were estimated on the basis of climatic conditions predicted for nine states: California, Idaho,
Texas, Nebraska, Minnesota, Mississippi, Kentucky, Florida, and Maryland.
{.imitations Resulting from the Model
Horn fly population levels predicted for average temperatures >40°C (104°F) cannot be compared
with current horn fly densities, since these temperatures now occur only in Southwestern deserts where cattle
are confined in feedlots during the summer months and therefore are not exposed to horn flies. Model
predictions of horn fly phenology/density are based on 30-year climatic averages, whereas the current horn fly
data are based on fly counts taken at various times of the season under variable weather and animal management
conditions and by various individuals. Nevertheless, existing horn fly densities generally correlate closely with
those predicted by the model with concurrent ambient temperature and rainfall (Miller, 1986). Though not a
limitation of the model, it should be noted that our interpretation of economic loss associated with predicted
horn fly levels is based on existing cattle management practices, breeds, and distribution; possible changes in
these factors and their impact on cattle production are beyond the scope of this study. We know of no reason
to suspect that a doubling in atmospheric carbon dioxide alone will account for the changes in horn fly
population density or phenology predicted by 2xCO2-induced climatic warming.
THE SCENARIOS
Used
The climatic scenario used to estimate horn fly abundance was the GFDL^CO^ using values projected
for temperature in degrees Celsius and precipitation as percent departure from current averages.
Estimation of Economic Loss
The effect of projected horn fly population levels on weight gain in beef cattle was estimated by the
following procedure. First, the relationship between horn fly population levels and rate of gain in beef cattle was
examined by computing various regressions for horn fly level (independent variable) and average daily gain
(ADG) for growing beef cattle (calves and feeder/stocker cattle) (dependent variable). Twenty-five data points
representing available economic loss values (see review by Drummond, 1987) were used; three outlier values that
did not represent the majority of data were not used. The best estimate for the relationship between horn fly
population level and reduction in ADG was the linear regression Y = 0.14 + 0.0002X (Figure 1). The correlation
coefficient, r2 » 0.71, is statistically significant (p = <0.05).
12-3
-------
K
g
0.25-
0.23-
0.21-
0.19-
0.17-
0.15-
0.13-
0.11-
0.09-
0.07-
0.05
0.03
0.01
Relationship between numbers of horn flies and
reduction in average daily gain (ADG) in kg.
T
100 200 300 400 500 600
Flies
1 I ' ' ' ' 1 ' ' ' ' I ' ' ' ' | ' ' ' i |
700 BOO 900 1000 1100
vw
§
Linear regression with 95X confidence limits on mean predicted values.
-------
Schmidtmann
The effect of projected horn fly levels on weight gain was estimated by using model-predicted horn fly
levels in the horn fly - ADG regression. Reductions in ADG corresponding to monthly horn fly levels that
equaled or exceeded 100 flies/head were calculated for each state. These values were then multiplied by 30 to
give the monthly reduction in weight gain attributable to hom flies. Monthly weight gain reductions were linked
with the growth of calves and feeder /stocker cattle (steers and heifers) by summing monthly reductions across
the periods that calves and feeder/stocker cattle are exposed to horn flies in respective states. We assumed a
7-month horn fly-exposure period (from birth to weaning) for calves, based on spring calving in northern and
central states, fall calving in California, and "year arouncT calving in Florida and Texas. For feeder/stocker
cattle, a 4-month Gate spring into summer) period was assumed for exposure to horn flies.
Weight gain values calculated under climatic warming were then compared with reductions in weight
gain based on current horn fly levels. This value was then multiplied by the numbers of beef calves and
stocker/feeder animals by state (Anonymous, 1986). The numbers of stocker/feeder cattle used in calculations
represent the numbers of steers (« stackers) plus 75 percent of heifer calves (- feeders from the calf crop of
the preceding year); the balance of heifers remain with herds as replacement cows. The resulting weight gain
values, either increases or decreases relative to current horn fly-induced losses, represent the change in weight
gain that will be experienced by calves and feeder/stocker cattle due to horn flies under climatic warming.
12-5
-------
Schmidtmann
CHAPTERS
RESULTS
HORN FLY PREDICTIONS: CHANGES, CAUSES, AND EFFECTS
Horn Fly Density and Seasonal Activity
When compared with existing population levels (Figure 2), the numbers of horn flies that will occur
under climatic wanning differ dramatically; both increases and decreases will occur. Conditions favoring
increases in horn fly levels will occur in all areas of the U.S. except the southern states from Mississippi west to
central California; in this area, horn fly populations will be suppressed during the summer months. As illustrated
by projected monthly population levels (Figure 2), adult horn fly activity throughout most of the United States
will increase rapidly during the spring months under climatic warming (see Florida, Idaho, Maryland, Minnesota,
and Nebraska), and summer population levels will be higher than current horn fly levels. Also, high-level horn
fly activity will persist during the fall months, with populations that exceed 500 flies per head experienced
throughout central and northern areas of the United States; current horn fly levels average ca, 100 to 300 flies
per head in these areas during this period. Thus, climatic warming will (1) expand the length of time that cattle
are exposed to economically important levels of horn flies, and (2) increase the density (number of flies per
animal) of many populations, especially during spring and fall months.
Cause of Changes in Horn Fly Populations
The results indicate that ambient temperature is the primary factor underlying projected changes in horn
fly population levels. The marked increase in duration and intensity of horn fly activity projected for Northwest
and Midwest regions, where increases in temperature will occur but precipitation will remain stable or decrease
slightly, support this interpretation. Also, ambient temperature exceeding 34° C for two hours (« 44° C in cattle
dung pats) is lethal to immature horn flies (Palmer and Bay, 1984); as the model predicts, a summer depression
of horn fly density will be pronounced, particularly in southern states.
Effect on Weight Gain in Beef Cattle
The exposure of beef cattle to greater numbers of horn flies for longer periods will increase reductions
in weight gain; the weight of cattle will be reduced by from ca. 2 to 12 kg/head across most of the U.S. (Table
1). The greatest weight reduction per animal will occur in the mid-Atlantic region, but the Northwest, Southeast,
and the Midwest regions will also experience considerable losses. In Nebraska, for example, projected increases
of 225-250 horn flies/head will result in greater weight reductions of 10.2 and 5.1 kg/head for calves and
feeder/stocker cattle, respectively. For the 17,728,000 calves and 1,851,000 feeder/stocker animals in the state,
total reductions in weight will total 27,168,100 kg greater than currently experienced.
Summer Suppression of Horn Flies
During the summer months, horn fly populations from Mississippi across to southern California will be
suppressed under climatic wanning (Figure 2). Predicted spring and fall climatic conditions will be favorable
for horn flies, but summer temperatures will limit horn fly larval (maggot) survival in cattle dung pats and
populations will decline. The net effect will be fewer horn flies over fewer days, hence less impact on cattle.
The summer suppression of horn fly activity will be strongly expressed in Texas, and horn fly-induced reductions
in weight gain wUl be 10.5 and 9.6 kg/head less than currently experienced by calves and feeder/stocker cattle,
respectively.
12-6
-------
Schmidtmann
Horn Flies and Dairv Cattle
The levels of horn flies projected for the upper Midwest, Northeast, and mid-Atlantic regions, which
exceed current levels by several hundred flies/head from May through October, will be associated largely with
dairy cattle. Economic losses in milk production due to increased horn fly activity under climatic warming cannot
be reliably predicted with available data However, it is reasonable to assume that losses in response to
increased horn fly-induced annoyance and irritation will exceed either the estimated 80 million dollars in reduced
milk production, or the 1 percent loss in milk production that are estimated to occur with existing levels of horn
fly activity.
12-7
-------
CURRENT AND PROJECTED (GFDL)
HORNFLY POPULATIONS
Figure 2. Current hornfly populations and projected hornfly populations (GFDL).
-------
Table 1. Summary of current and projected horn fly densities and associated changes In weight gain for beef cattle.
State
California
calves 3
feeders/stockers
Florida
calves
f coder s/stocker s
Idaho
calves
feeders/stockers
Kentucky
calves
feeders/stockers
Maryland
calves
feeders/stockers
Minnesota
calves
feeders/stockers
Mississippi
calves
feeders/stockers
Nebraska
calves
feeders/stockers
Texas
calves
feeders/stockers
Average
per
Current
139
400
105
105
104
100
200
200
48
46
82
138
304
294
154
175
445
675
number of
month per
Projected
182
200
530
530
393
475
243
275
414
400
243
350
164
250
412
400
135
138
horn flies
head
Difference
+ 43
-200
+425
+425
+289
+375
+ 43
+ 75
+366
+161
+212
-140
- 44
+258
+225
-310
-537
Average reduction
gain per head
Current
5.4
9.0
6.3
3.6
3.9
2.4
8.5
4.9
0.9
0
3.3
3.3
11.4
6.2
6.3
4.2
17.5
14.1
Projected
7.1
4.5
18.9
10.8
15.9
10.8
10.2
6.6
16.2
9.0
10.2
8.1
7.2
6.3
16.5
9.3
7.0
4.5
in weight No. cattle (xlOOO)1
(kg)
Difference2 Calves Feeders /Stockers
+ 1.7 950
- 6.3 1381
+12.6 1135
+ 7.2 222
+12.0 526
+ 8.4 447
+ 1.7 976
+1.7 482
+15.3 71
+ 9.0 52
+ 6.9 396
+ 4.8 821
- 4.2 686
+0.1 207
+10.2 1738
+ 5.1 1851
-10.5 5178
- 9.6 5548
Change in weight gain (kg)
xlOOO
+ 1650
- 8700
+14301
+ 1598
+ 6312
+ 3758
+ 1659
+ 819
+ 1086
+ 468
+ 2732
+ 3941
- 2881
+ 21
+17728
+ 9440
-54329
-34060
Total
- 7,085,300
+15,899,400
+10,066,800
+ 2,478,400
+ 1,554,000
+ 6,672,800
- 2.860,300
+27,168,100
-88,389,900
1 + values represent weight gain losses greater than currently experienced; - values represent losses less than currently experienced.
2 Data taken from Agricultural Statistics, 1986.
3 Values represent 75Z of heifer calves (• feeders) plus steers (- scockers).
-------
Schmidtmann
CHAPTER 4
IMPLICATIONS OF RESULTS
ECONOMIC IMPLICATIONS
The economic implications of projected increases and decreases in horn fly populations are several-fold.
Most notably, the exposure of beef and dairy cattle across most of the United States to both longer periods and
higher levels of horn fly activity will (1) clearly result in increased reductions in weight gain for beef cattle, and
(2) probably suppress milk production and certainly increase control costs for dairy cattle.
Beef Cattle
Assuming an average market sale price of $55/100 Ibs. (Anonymous, 1986), losses due to increased horn
fly activity (based on the nine states investigated) will range from approximately 2 to 30 million dollars per state
in reduced beef sales annually. Only states from Mississippi west to southern California will be spared reductions
in beef production.
The ca. eight million plus head of beef cattle (carves and feeder/stocker cattle) affected by horn flies
in Texas represent ca. 15 percent of cattle in these age classes in the United States. Under climatic warming,
Texas cattle will sustain economic losses greater than those currently experienced during spring and fall months,
but this condition will have little effect on the strong overall economic benefit that will accrue from a summer
suppression of horn fly activity. As compared with current horn fly-induced reductions in weight gain, losses
under climatic warming will be as much as 10.5 and 9.6 kg/head less in calves and feeder/stocker cattle,
respectively. This will amount to increased beef production of ca. 86 million kg. annually; on a monetary basis,
this represents ca. 104 million dollars in beef sales.
Dairy Cattle
The economic impact of increased horn fly activity in the Great Lakes and Northeast dairy regions is
less readily interpreted. Few data are available concerning the effect of horn flies on milk production, and those
that have been conducted are confounded by the presence of other fly species, e^, stable flies (Morgan and
Bailie, 1980) and face flies (Burton et aL, 1984). It is, however, reasonable to infer that higher levels and longer
periods of horn fly activity will be expressed as losses that exceed either the 1 percent reduction in milk sales or
the 80 million dollars in reduced production that are estimated for current horn fly populations (Drummond et
aL, 1984; Anonymous, 1976). However, the trend toward confinement housing of cattle in free-stall barns in the
Northeast, which precludes horn fly attack, may offset any effect of increased horn fly activity. The decreased
horn fly levels projected for the Southwest will have little impact on dairy cattle, as horn flies are uncommon in
the drylot-type holding areas of the large dairy operations common to this region.
SOCIOECONOMIC IMPLICATIONS
It is reasonable to assume that higher levels and longer periods of horn fly activity, along with increased
economic losses, will intensify the need for methods to suppress horn fly populations. Current horn fly control
strategies are limited to the application of insecticides, either directly to cattle (primarily as
insecticide-impregnated ear tags), or as feed additives fed to animals that render the manure toxic to immature
stages. A dependency upon chemical control has in recent years resulted in a rapid selection for genetic
resistance (Schmidt et aL, 1987). Since horn flies hi many states are now resistant to chemicals registered for
their control, alternative horn fly control methods, including practices safe for lactating cows, will be necessary
under climatic warming to offset increased losses in animal production.
12-10
-------
Schmidtmann
REFERENCES
Anonymous. "Agricultural Statistics", U. S. Government Printing Office, Washington D. C, 1986, 436 pp.
Anonymous. "ARS National Research Program, NRP No. 20480, Control of Insects Affecting Livestock", Agric.
Res. Serv., 1976, 99pp.
Bruce, W. G. "The history and biology of horn fly, Haematobia irritans (Linnaeus); With comments on control."
Tech. Bull No. 157., North Carolina Agric. Expt. Sta and EntomoL Res. Div., UJS.DA.. 1964, 32 pp.
Burton, J. H., I. McMillan, and G. Surgeoner. "An evaluation of fenvalerate (Bovaid) ear tags as an improvement
in fly control and methods from increased milk production in dairy herds." Caflfflf i Ji Anim fin. 64:113-117,1984.
Drummond, R. (X, G. Lambert, H. E. Smalleyjr., and C. E. TerrilL "Estimated losses of stock to pests." In:
CRC Handbook of Pest Management in Agriculture. VoL 1, D. Pimentel, ed. CRC Press, 1981, 597 pp.
Drummond, R. O. "Economic aspects of ectoparasites of cattle in North America. In: Proc. MSD AGVET
Symp., The economic impact of parasitisms in cattle. W. H. D. Leaning and J. Guerroe, eds. Montreal, Quebec.
1987.106 pp.
Harris, R. I*, J. A. Miller. "Horn flies and stable flies: Feeding activity." Ann. Entomol. Soc. Amer. 67:891-894,
1974;
Harvey, T. L., and J. L. Launchbaugh. "Effect of horn flies on behavior of cattle." J. of Econ. Entomol. 75:25-27.
1982.
Haynes, D. N, and R. L. Tummala. "Development and use of prediction models in the management of cereal
leaf beetle populations." In: Modeling for pest management: Concepts, techniques and applications. East
Lansing, Michigan St. Univ. Press, 1976. pp. 53-76.
Miller, J. R. "Computer modeling populations of the hom fly." In: Modeling and simulation: Tools for
management of Veterinary pests. J. A. Miller ed, USDA, ARS-46,1986, pp. 33-40.
Morgan, D. W. T., and H. D. Bailie. "A field trial to determine the effect of fly control using permethrin on milk
yields in dairy cattle in the U.K." Vet. Rec. 106:121-123,1980.
Palmer, W. A., D. E. Bay, and D. J. H. Sharpe. "Influence of temperature on the development and survival of
the immature stages of the horn fly, Haematobia irritans irritans (L.). Prot. EcoL3:299-309.1981.
Schmidt, C. D., S. E. King, H. D. Petersen, and J. L. Robertson. "Resistance of horn flies (DipteraMuscidae)
to permethrin and fenvalerate." J. Econ. Entomol. 78:401-406,1985.
Steelman, C. D. "Effects of external and internal arethropod parasites on domestic livestock production." Annu.
Rev. EntomoL 21:155-178.1976.
12-11
-------
AGRICULTURAL POLICIES FOR CLIMATE CHANGES
INDUCED BY GREENHOUSE GASES
by
G. Edward Schuh
Humphrey Institute of Public Affairs
University of Minnesota
Minneapolis, MN 55455
Contract No. 58-3AEL-9-80030
-------
CONTENTS
FINDINGS 13-1
CHAPTER 1: IMPORTANCE AND COMPLEXITY OF THE PROBLEM 13-2
CHAPTER* REDUCING AGRICULTURAL SOURCES OF
GREENHOUSE EMISSIONS 13-3
Nitrogen Fertilizers 13-3
Rice Cultivation 13-3
Ruminant Animals 13-4
CHAPTER 3: SCIENCE AND TECHNOLOGY POLICY 13-6
Dealing with Externalities 13-6
CHAPTER 4: POLICIES NEEDED TO HELP AGRICULTURE ADJUST
TO THE GREENHOUSE EFFECT 13-8
Commodity Policies 13-8
Adjustment Policies 13-10
Biocultural Restoration 13-11
Irrigation Policies 13-11
Research Policies 13-12
Other Issues 13-12
CHAPTER 5: CONCLUSIONS 13-14
11
-------
Schuh
FINDINGS1
Climate changes caused by ever-increasing quantities of greenhouse gases released by human activities
are expected to have a significant effect on VS. agriculture and rural America. Among other things, these
effects are expected to include: (1) significant shifts in the location of agricultural production within the United
States; (2) major changes in land and water use; (3) significant geographic shifts in population and rural
industries; (4) important environmental effects from changes in fertilizer and pesticide runoff; (5) impacts on
biodiversity and availability of germ plasm; (6) significant shifts in international comparative advantage; and (7)
increased risk and uncertainty for agriculture.
This paper addresses the policy issues associated with these prospective effects. There are at least two
dimensions to this problem. First, agriculture itself is an important source of greenhouse emissions. This raises
the issue of what policies or policy changes are appropriate and necessary to reduce these emissions. Second,
climate changes will have significant changes on the configuration of U.S. agriculture and rural America. This
raises the issue of what policies are needed to facilitate the inevitable adjustments, and what policies are needed
to assure that the agricultural and rural sectors contribute in an efficient and equitable way to the further
development of the U.S. economy.
'Although the information in this report has been funded wholly by the UJS. Environmental Protection
Agency under Contract No. 58-3AEL-9-80030, it does not necessarily reflect the Agency's views, and no official
endorsement should be inferred from it.
13-1
-------
Schuh
CHAPTER 1
IMPORTANCE AND COMPLEXITY OF THE PROBLEM
A number of factors contribute to the importance and complexity of these problems. First, much of
North America's agricultural production takes place under dryland conditions using annual crop species under
technologically advanced cropping systems that are primarily monocultural in nature. The robustness associated
with general, multiple-enterprise farms has, long since passed from the scene on any economically significant
scale. Thus we are left with an agricultural system constituted mostly of specialized farms producing
commodities in geographically specialized patterns. This system is vulnerable to changes in climate, and
adjustment costs can be quite high.
Second, U.S. agriculture is an integral part of a global, international agricultural sector. Consequently,
agricultural policies cannot be devised for US. agriculture in isolation from the rest of the world. The optimal
configuration of agriculture will thus depend very much on how changes in climate affect global agriculture. In
addition, it will depend very much on the policy responses of other countries to the changes in climate.
Third, the knowledge base on which sound agricultural and economic policies can be conceived and
articulated is extremely weak. The precise effects of climate changes on technical agriculture are unknown even
in the United States, to say nothing of the rest of the world, and especially in developing countries. In the same
way, we know little in a systematic or parametric way about the response of farmers and people in the rural
sector to the changes in climate.
Because of these problems, the best that one can do is focus on general principles that provide
guidelines for the development of policy. In that regard, the basic principles that have guided our own policy
prescriptions are to let markets be the basis for facilitating the multifarious adjustments needed, and to use
government interventions only when there are clear cases of externalities in which private costs and benefits are
significantly different from public costs and benefits, respectively.
2We do know, however, that after the adverse weather of the mid-1930's and mid-1950's, there were
substantial changes in cultural practices in the Great Plains to reduce erosion and to conserve or more
effectively use moisture. These changes included, for example, the replacement of the moldboard plow and
an increase in the use of summer fallowing.
13-2
-------
Schuh
CHAPTER 2
REDUCING AGRICULTURAL SOURCES OF GREENHOUSE EMISSIONS
There are three sources of greenhouse emissions from agriculture: (1) the use of nitrogen fertilizers;
(2) the cultivation of rice; and (3) enteric fermentation in domestic animals, with cattle being the major culprit.
Policy measures should be directed to reducing these emissions so that agriculture's contribution to the
greenhouse effect is reduced to a minimum.
Nitrogen Fertilizers
It happens that current agricultural policies3 contribute to augmenting greenhouse emissions in each of
the above cases. For example, target prices and loan levels set above what would otherwise be market-clearing
levels increase the demand for fertilizer. Land set-asides to bring production into balance attenuate somewhat
the effect of higher commodity prices on the demand for fertilizer. However, land set-asides also cause farmers
to search for land substitutes, and fertilizers are an excellent land substitute. Eliminating these commodity
programs and shifting to a more market-oriented policy would reduce the demand for fertilizer and in turn the
supply of emissions to the atmosphere.
Rice Cultivation
Rice is not one of the major crops in the United States, although regionally it is relatively important in
places such as Arkansas, California and Louisiana. Moreover, this nation produces more rice than it should
according to efficiency criteria by virtue of commodity programs that set the price of rice above what it would
be in the absence of such programs. Similar to the case of fertilizer, commodity programs that were more
market oriented, with income transfers to farmers decoupled from the price of the commodity, would lead to a
decline in the production of rice and thus to a reduction in greenhouse emissions.
Globally, the expectation is that the consumption of rice will decline in relative terms as economic
development proceeds, and perhaps eventually in absolute terms. The income elasticity of demand for rice is
relatively low so as per capita incomes rise worldwide, and especially in South and Southeast Asia, consumers
will increase their consumption of fruits, vegetables, and livestock products. Equally as important, consumption
of rice is time-intensive in the household since, compared to wheat bread, for example, it takes more time to
prepare. Bread is "manufactured" in bakeries and becomes a convenience food in the household, while rice has
to be cooked, with all that that entails.
As national economies expand and develop and the opportunity costs of time to the household rise, there
will be added incentive to shift away from the consumption of rice to the consumption of other cereals.
Opportunity costs of time in the household rise as wage rates rise and as women participate increasingly in the
nation's work force.
Globally, this shift in consumption patterns away from rice suggests the importance of diversification
policies to help shift resources out of the production of rice. Given the importance of irrigated rice in Asia and
the lack of mobility of resources in the production of irrigated rice, adjustment of resources out of this activity
can be expected to be relatively slow. A full panoply of policies, ranging from expanded and diversified
agricultural research efforts, changes in irrigation infrastructure, changes in marketing infrastructure, and proper
price incentives, will be needed to facilitate and accelerate the rate at which resources are adjusted out of the
sector.
an overview of these policies, see B. Gardner's, The Governing of Agriculture. Lawrence, Kansas:
Regent's Press, 1981.
13-3
-------
Schuh
Economic policies in some Asian countries also encourage the excess production of rice. Japan is a
classic case in which price policies encourage the production of more rice than the domestic economy can
effectively absorb, and consumption policies keep the price of wheat and other commodities high relative to the
price of rice, thereby causing the consumption of rice to be higher than it otherwise would be. These policies
should be the subject of international negotiations designed to change the policies to reflect comparative
advantage principles.
Ruminant Animals
U.S. commodity programs also encourage a cattle herd in this country that is much larger than it
otherwise would be. Dairy policies in particular have for over a decade contributed to the excess production of
milk by a number of means. First, the price of milk has been set above market-clearing levels, thus encouraging
the production of excess supplies.4 The cattle herd would be significantly lower if it weren't for these policies,
even granting that a decline in production per cow would attenuate this reduction somewhat.
The marketing orders5 which are an integral part of dairy policy discourage the adoption of available
technology in the processing sector. This also causes the production capacity in the sector to be larger than it
would otherwise be.
Notable in this regard is the failure to take full advantage of technology which makes it possible to
reconstitute milk components in such a way that the reconstituted product cannot be distinguished from fresh
milk either by consumer taste panels or by laboratory tests.6 This technology makes it possible to extract the
water from milk and convert it into powdered form. The powder can then be transported long distances at lower
cost, thus making it possible to produce the milk in low-cost regions. In addition, the powdered product can be
transported through time by storing. In this way, full advantage can be taken of the flush season of production
and the product carried into the slack production time of the year. Present policies, which encourage herd sizes
to meet needs in the slack season when costs of production are highest, could thus be abandoned and a smaller
cow herd would be needed.
Whether the milk is transported geographically or through time, reconstituting it is fairly simple. The
need is to add water, and a modest proportion of fresh milk to provide taste.7
Milk marketing regulations embedded in marketing orders reduce the incentive to use this technology.
They do this by requiring that the reconstituted milk be priced at the same price as the higher cost fresh milk.
There is thus no incentive to invest in the equipment and operating costs needed to use the technology.
Elimination of this regulation would reduce the size of the dairy herd, while at the same time using resources
more efficiently and making the final product available to consumers at a lower cost.
Over the last decade, these excess supplies are estimated to have ranged from 6 to 12 percent at prevailing
price levels.
Marketing orders are a rather precise set of rules which prescribe how milk is to be marketed. They
prescribe relationships among prices of various classes of milk, the geographic pattern of milk prices, and rules
for how milk is to be marketed. For more detail, see Brooks, Karen, "Dairying and Dairy Policy in the 1980's,"
AEI Occasional Papers, American Enterprise Institute, Washington, D.C., 1984.
Hammond, Jerome W., Boyd Buxton, and Cameron Thraen, Potential Impact of Reconstituted Milk
on Regional Prices. Utilization, and Production. Agricultural Experiment Station, University of Minnesota, Station
Bulletin 529, 1979.
7Ibid.
13-4
-------
Schuh
VS. grain policies also cause the dairy herd to be larger than it would otherwise be. These policies
operate by means of deficiency payments. In effect, grain prices on the basis of which income transfers are paid
to farmers are set above market-clearing levels. This causes the supply of grain to be larger than it otherwise
would be. The larger supply comes on the market and in effect drives price down to the loan level, which is set
lower than what would be market-clearing levels. This lower price is in effect a subsidy to the dairy industry,
causing the cost of production for milk to be lower than it otherwise would be.
In recent years land has been pulled out of production as a means to reduce stocks in government hands
and to raise the price of the grains. Whether prices have been increased to what would otherwise be their
market-clearing levels is open to question, however. Policy should be to reduce this implicit subsidy and thus
the incentives for a larger dairy herd than would otherwise be the case.
This same implicit subsidy affects the beef sector, and causes it, too, to be larger than it otherwise would
be. The beef sector is in addition protected by a tariff, which in recent years has been converted in effect to a
variable levy, and by the periodic negotiation of voluntary export agreements which limit imports. The sector
is thus subsidized from two directions, both of which cause the beef herd to be larger than it otherwise would
be. The protection of the beef sector is modest, but should be eliminated. Elimination of both kinds of subsidies
would reduce the size of the herd and in turn the generation of greenhouse gases.
13-5
-------
Schuh
CHAPTERS
SCIENCE AND TECHNOLOGY POLICY
Science and technology can also be used to reduce greenhouse emissions. For example, the biological
fixation of nitrogen as a substitute for the use of commercial fertilizers would reduce the production of methane
from these materials. Research to expand the biological fixation of nitrogen should thus be a part of policies
designed to reduce the generation of greenhouse gases.
It may also be possible to develop fertilizer materials which release nitrogen more slowly. The available
nitrogen would be absorbed more efficiently, thus reducing the amount released to the atmosphere.
More generally, research efforts directed to reducing methane emission from all of its agricultural
sources should be expanded. This includes the emissions from fertilizers, from rice, and from the cattle herd.
An alternative approach is to develop new techniques for capturing and using the methane. At the
height of the energy crisis in the 1970's, research on methane was augmented for this purpose. But resources
for such research declined as energy prices declined in the 1980's. It is time these efforts were revitalized, but
with the different goal of reduction of methane emissions.
Development of sustainable agricultural systems may be another way of reducing greenhouse emissions.
Such systems require that nutrients be produced within the farm for on-farm commodities. The problem with
this approach is that it tends to imply a slower growth in agricultural output than would be the case with plant
nutrients purchased from the nonfarm sector. The costs of such a production system in terms of higher food
prices could be quite high.
Current research on growth hormones for dairy cows offers another means to reducing cattle numbers.
These hormones promise to increase production per cow by 25 percent or more with little additional feed intake.
The potential for reducing cow numbers is thus great, even if the increase in productivity is reflected in lower
milk prices, since the price elasticity of demand for milk is relatively low.
An important issue with this technological breakthrough is whether dairy fanners will permit the use of
these hormones, or whether its use will be regulated or precluded entirely. Adjustment policies may be needed
to help dairy producers shift to other income and employment opportunities. This points to the more general
research and development issue of developing production alternatives that facilitate the shift of resources out of
commodities that produce greenhouse gases.
Dealing with Externalities
Finally, methane emissions constitute a negative externality in that they impose social costs on others.
Users of nitrogen fertilizers, producers of rice, and owners of cattle earn profits from those activities, but they
don't pay all the costs incurred by society in their use and production since the greenhouse effect is "external"
to their specific economic activities.
This is a classic problem in public policy. The solution to the problem is to impose a tax on the activity
so the entrepreneur, whoever he or she may be, bears the full cost. Hence, the solution would be to impose a
tax on the use of nitrogen fertilizer, and on the production of rice and cattle. This tax would reduce these
activities to a socially optimal level.
The size of this tax should reflect the amount of external costs actually imposed by the activity.
However, given the present state of knowledge, calculating the amount of the tax would be difficult.
13-6
-------
Schuh
To conclude, a number of policy actions are required to reduce greenhouse emissions from agriculture.
The first is to reduce the subsidies that stimulate those activities which generate the emissions. The second is
to increase R & D expenditures on those activities designed to reduce methane emissions and on activities
designed to find ways to capture and make use of these emissions. The third is to tax those activities that give
rise to methane emissions so as to reduce them to socially optimal levels.
Aside from political difficulties, the first two actions are presently feasible. The third is not feasible
owing to the lack of an adequate knowledge base.
13-7
-------
Schuh
CHAPTER 4
POLICIES NEEDED TO HELP AGRICULTURE ADJUST
TO THE GREENHOUSE EFFECT
If efficient use is to be made of this nation's agricultural resources, policies designed to facilitate the
adjustment to greenhouse effects must take into account that these effects will be global in nature and that U.S.
agriculture is part of a global economy.
The implication of the first point is that comparative advantage on a global scale will change significantly
as a consequence of these effects. Regions of the world, including those in the United States, that were once
efficient producers will no longer be so, while regions that previously were not efficient producers will become
so. Similarly, greenhouse effects eventually can be expected to have an impact on where populations and
economic activities are located. This will change the location of the foci of demand, the other important
dimension of comparative advantage.
The implication of the second point is that border-price pricing principles9 should be followed so that
the use of national resources is gradually adjusted to the changing conditions of comparative advantage. Subject
to the caveats in the previous section and in sub-sections which are to follow, such principles will allow markets
to bring about the needed adjustments. This is a proper policy to follow in light of the inadequate knowledge
base on which policies will inevitably need to be developed.
Subject to the above constraints, a number of kinds of policies can be pursued to facilitate the needed
adjustment. Each kind of policy and the principles which should guide it are discussed in the following sections.
Commodity Policies
The role of commodity programs in contributing to increased greenhouse emissions was discussed in the
previous section. In that context the objective of policy was viewed as reducing the subsidies to activities which
produce greenhouse emissions. In the present context the policy objective is similar, but the goal is different.
The goal is now to have price and program policies which will permit agriculture to adjust to the changing
conditions of demand and supply that evolve from the greenhouse effect.
The means to do that is to have policies which let domestic prices reflect border price opportunity costs.
For export commodities such as wheat, corn, and soybeans, that will be domestic prices consistent with FOB
export prices. For commodities that compete with imports (beef, for example), it will be domestic prices
For an assessment of the greenhouse effects on Great Plains agriculture, see Ciborowski, Peter, and Dean
Abrahamson, The Granary and the Greenhouse Problem"; McQuigg, James D., "Climate Change and
Agricultural Response: The Short Run"; and Rosenberg, Norman J., "Climate, Technology, Climate Changes,
and Policy: The Long Run." All of these are in The Future of the North American Granary, edited by C. Ford
Runge. Ames: Iowa State University, 1986.
9 Border-price pricing principles assert that the domestic price of tradeable goods and services should be
consistant with prices in international markets on the grounds that the latter prices reflect opportunity cost uses
of resources. This is an efficiency criterion for the allocation of domestic resources.
10 FOB prices are the prices of commodities at the port at which they leave the country.
13-8
-------
Schuh
consistent with GIF prices of imported commodities.11 For non-tradeable commodities (the best example being
a service, haircuts), conditions of domestic demand and supply, as conditioned by the demand and supply for
tradeable commodities, should be allowed to work themselves out.
The implication of such principles is that prices in the domestic economy should reflect conditions of
demand and supply, and not be used as the means to transfer income to the agricultural sector. Similarly,
commodity programs should not be used for price stabilization means. Neither of these propositions is to imply
that equity or stabilization objectives should be ignored in the design of policy. What it means is that the price
and market system should be permitted to do what it does best - efficiently allocate scarce resources, while other
measures are used to attain other policy goals. In light of the inevitable lack of adequate information to centrally
allocate resources in the context of the expected changes in demand and supply induced by changes in climate,
greater reliance on free markets is the only feasible alternative to pursue.
Present commodity programs also deter regional shifts in production of affected crops by means of
acreage bases assigned to fanners as the basis of program management. The sector should be deregulated and
these acreage bases phased out.
Income policies for agriculture should be made consistent with other welfare policies in the general
economy. There is no inherent reason why income or welfare policies for agriculture should be any different
than for the rest of the economy. The criterion of transparency requires that income transfers be direct and in
the form of cash payments, and not in the form of implicit income transfers by means of distortions in prices.
Similar comments apply to price stabilization measures. There is little evidence to support a case for
government intervention in the market for stocks as a stabilization measure.12 As long as an adequate market
information system is available and capital markets are reasonably efficient, the private sector can be expected
to carry an adequate level of stocks, and to carry them in an efficient manner, so that markets will have the
optimal level of stability. If it should be deemed necessary to carry a higher level of stocks than market forces
will bring forth, modest storage subsidies paid on a per unit basis can increase the level of stocks carried with
a minimal distortion to market forces.
Two factors argue for depending on market forces to reallocate resources in response to greenhouse
effects. The first is that these effects are expected to be spread out over time. Resource reallocations are
expected to be gradual, not abrupt and once-for-alL Markets are the most efficient means of pooling the
enormous amount of data that will be generated as greenhouse effects emerge. The only caveat is that UJS.
agriculture should be part of a global system of free markets so that border prices truly reflect global conditions
of demand and resource scarcity. That is already a goal of UJS. policy. The potential seriousness of the
greenhouse effect may provide added incentive to establish such policies globally.
Second, the consequences of errors in establishing prices by central means, or hi imposing quantitative
restrictions such as the amount of nitrogen per hectare, can be large, and errors can be expected to be frequent.
Agriculture involves substantial fixed investments - investments that have a long gestation period and that are
relatively immobile. Mistakes in such investments can result in significant economic wastage.
Errors result from the centralized fixing of prices. No individual, or small group of individuals, can be
expected to absorb and understand the large amount of information needed to fix prices, especially when climatic
conditions are influencing those markets in unknown and uncertain ways. The only practical alternative is to
depend on market forces.
GIF prices are the prices of commodities laid down at the port of entry.
12For a comprehensive treatment of the storage or stockpiling issue, see B. Gardner's, Optimal Stockpiling
of Grain. Lexington, Mass.: Lexington Books, 1979.
13-9
-------
Schuh
Adjustment Policies
Greenhouse effects are expected to lead to significant geographic shifts in the location of economic
activities, both within agriculture and within the nonfarm sector. Helping to facilitate the relocation of economic
activities and the associated labor force is a legitimate role for government. Given that agriculture in the United
States is largely a part-time activity, with some 72 percent of the income of farm families coming from nonfarm
activities, adjustment policies for agriculture cannot be designed in isolation of activities in the nonfarm sector.
Consider first the issue of labor adjustment. Two kinds of policies are important. The first is retraining
for new economic activities. This may be for new agricultural activities in a new location. It may be for non-
agricultural activities in other regions. It may be for non-agricultural activities in the regions where labor is
presently located. Or it may be to train the nonfarm labor force for agricultural activities in the regions where
agriculture is expected to be more important as a consequence of climatic changes.
Members of the labor force can be expected to recoup some of the benefits of such retraining. Hence,
it is approriate that individual workers pay some portion of these costs. However, a case can be made for
government subsidies on the grounds of obtaining a more efficient allocation of resources and at a faster rate.
Equity issues also enter since in the absence of subsidies a select group of individuals would have to bear the
costs of adjustment that are a national issue, and not due to their own individual actions.
Government subsidies for retraining can take a variety of forms. One form is to provide the retraining
at less than its full cost. Another form is to provide the trainees with fellowships and scholarships. Still another
form the subsidy could take is to pay the trainee for attending the training program. This form of subsidy would
address the fact that foregone income is frequently the most important private cost of participation in such
programs.
Similarly, a case can be made for the government subsidization of relocation costs. In this case the
subsidies might be in the form of moving costs, and for subsistence while individuals search for alternative
employment. An important implicit subsidy for relocation is to maintain adequate information and technical
assistance services for the labor market. These services should provide information to the labor force on where
jobs are emerging and where they are declining. It should provide similar information to private and public
businesses. And it should provide technical assistance to both employees and employers about how they can
better take advantage of market opportunities.
Labor is not the only asset for which mobility will be an issue. Farmers in particular tend to have other
assets, many of which are highly immobile. These include barns and other buildings, fences, drainage systems,
and other site-bound assets. Assets such as tractors and equipment will be more mobile in a physical sense, but
they may be obsolescent due to a change in economic activities associated with relocation. If a significant
geographic area finds itself no longer vital in a particular economic activity, the widespread dumping of such
assets can impose large losses on individual asset owners. These declines in asset values in regions losing
comparative advantage can exacerbate the adjustment problem because they cause the declining area to be more
competitive with newly expanding areas where the costs of production are expected to be higher on the margin.
Society faces a real quandary and challenge in dealing with these asset markets. To compensate all asset
owners for the losses they incur due to changes in climate could be both difficult and costly. In the first place,
it will be difficult to know what share of decline in asset values should be attributed to the greenhouse effects.
The possibilities of moral hazard in which individuals insure only when the chances of loss tend to be highest,
thus disturbing the actuarial base on which the insurance is issued, are great. So are the possibilities of errors
in valuation, since there will be no markets to establish values. In addition, the costs of compensating for such
losses can be very large.
An important issue will be whether alternative uses for the resources that might be used for
compensation might have a higher payoff. For example, in a resource constrained situation, these same
resources might have a higher social payoff in building the infrastructure in areas that are benefited by the
13-10
-------
Schuh
change in climate, or in agricultural research designed to understand the impact of the change in climate and
to develop varieties of crops better adapted to the changed climatic conditions.
On the assumption that greenhouse effects can be expected to be gradual and spread out over time, the
best policy may be to not compensate for losses in asset values. The costs of the assets can be depreciated over
their expected life, and the net income of the individuals will be compensated by means of income tax schedules.
This system will require that sound information be made available on what the precise effects of the climate
change will be. and their geographic locations.
If there should be acute, localized consequences of the greenhouse effect, compensation might be
considered. This would be for such things as substantial increases in localized flooding or the sudden
disappearance of water supplies.
The land market is another sector in which government intervention may be needed. Some areas may
require the organization of land into larger farms so that the farm unit can provide an adequate income for the
farm family. Other areas may require that farm size be smaller due to the emergence of agricultural activities
that are more intensive. In the latter case, the market for land can probably accommodate the adjustments with
relative ease. In the former case, a secondary market13 may be needed to reorganize the resources efficiently.
Whether this will be needed will depend on the expected speed of adjustment
Finally, there is no a priori reason why the labor force should be forced to bear all the adjustment to
changing economic conditions. For areas which experience a significant decline in agricultural activities, a surplus
labor force can be expected to appear. Policy can promote the development of new economic activities in these
regions as well as promote the relocation of labor. Retraining the labor force for such new activities will be an
important form of subsidy for the relocation of non-farm activities. The provision of an adequate infrastructure
such as roads, other transportation installations, and communication networks, is another. It is not dear that
subsidies should go much beyond that, as long as an efficient local capital market is available.
Biocultural Restoration
Present analyses suggest that a significant greenhouse effect may lead to the virtual abandonment of
agriculture in some parts of the United States. If such forces do indeed play themselves out, one can imagine
a significant drainage of population from such areas and the virtual abandonment of some regions. Policies to
reestablish some system of plant life on these abandoned lands will be important. In the first place, it may be
the basis for a new set of economic activities. In the second place, it will be needed to prevent erosion and loss
of topsoil. And finally, plants are consumers of carbon dioxide. Reestablishing as vital a plant life as is possible
will be one means of minimising the greenhouse effect itself.
To the extent that a new plant growth such as trees and shrubs should provide the basis of new
economic activities, the private sector can be expected to handle it as long as they know what the economic
opportunities are and capital markets are efficient. If that is not the case, then the government will need to take
the lead in making investments and establishing vegetative growth which will control wind and water erosion and
provide for plant cover. Such activities may be costly, but they are critical for conserving the basic resources for
the future.
Irrigation Policies
An important aspect of the greenhouse effect is expected to be droughts and potential effects on water
supplies. Measures to respond to these effects include our expansion in the area that can be irrigated and, if
13 A secondary market is one in which "third" persons acquire the land on the expectation of reselling it later
at a gain, and with no intention of actually farming it. In effect they provide a floor under the market, while at
the same time providing an efficient means of reorganizing the land into efficient sized farms.
13-11
-------
Schuh
necessary, the capability to respond to a decline in available water supplies in a given area if the need for such
a decline should occur. These actions will be fairly location specific, and will need to be implemented as more
information becomes available on the regional and commodity specificity of the greenhouse effects.
More generally, measures which improve the efficiency with which irrigation water is used should be
promoted. This should include pricing policies which price water at its true value. This is in stark contrast to
present policies which price water at substantially less than its economic value and thus lead to substantial
wastage. In addition, research and physical investments should be directed to reducing water losses, for example,
trickle irrigation and instruments that help determine soil moisture.
Research Policies
The demand for new knowledge about how to adjust to changing conditions will be very great.
Investments in the research to generate that knowledge may have as high a payoff as any investment the
government can make.14 Research should concentrate on generating new knowledge and information that is
useful to both the private and public sectors.
The kinds of new knowledge needed are legion. The greenhouse effect itself needs to be better
understood. Understanding its expected consequences should also be given high priority, especially on global
comparative advantage. Such information is critical to enable the private sector to adjust efficiently to the
changed economic conditions.
Research designed to produce new technology for a changed agriculture is also critical. Agricultural
technology tends to be highly location-specific. In regions where agriculture becomes more vital as a
consequence of the changes in climate, new technologies will need to be developed. In many cases this will
require the development of a new capacity for agricultural research. In regions where agriculture declines,
research will be needed to identify and develop the knowledge base for whatever vegetative and animal base will
survive. In some cases, such new knowledge may be critical to establishing vegetative coverage of any kind.
Research should also focus on developing a more flexible agriculture. To the extent the changes in
climate create more risk and uncertainty for farmers, they need farming systems with more flexibility, and which
involve informal insurance schemes such as crop mixes in which the crops respond differently to climatic
developments each year. A world of more risk and uncertainty may shift U.S. agriculture towards a combination
of crops and livestocks on individual farms, and away from past trends towards specializatioa Research on the
inter-relationships among these enterprises will be needed.
Finally, there is a need for socio-economic research on a broad basis. This research should provide the
knowledge base for devising sound socio-economic policies, and for facilitating the adjustment of the private
sector to changed economic conditions. It should also focus on identifying changes in comparative advantage,
both within the United States and on a global scale.
Other Issues
Three additional issues are important in helping society adjust to and adapt to prospective changes in
climate from greenhouse effects. The first is the need for a sound information policy. The second is the need
14 The social rate of return to investments in agricultural research generally tends to be high, ranging from
35 percent to as high as 600 to 700 percent. See Ruttan, Vernon W., Agricultural Research Policy. Minneapolis,
MN, University of Minnesota Press, 1982. With dramatically changed ecological conditions for agriculture and
the need to do a great deal of adaptive research, the social rate of return to additional investments in agriculture
could be expected to be quite high.
13-12
-------
Schuh
for a capacity for strategic planning to help establish policy as technical and economic conditions change. The
third is the need for investments in infrastructure to accommodate a changed configuration of the economy.
Some remarks on each of these issues follow.
We proposed above that maximum reliance be placed on markets in bringing about the reallocation of
resources associated with expected greenhouse effects. Markets can be expected to do this in an efficient way
only if they have an adequate information base on which to respond. The provision of such information is a
proper role of government. Hence, in addition to expanded research programs designed to produce new
information and knowledge, it will be necessary to have an information system which makes this information fully
available to members of society.
Such information is needed not only for men and women in their economic roles. It is also needed so
that citizens can vote and act in an informed way on the issues. Hence, the information needs to be disseminated
as widely as possible.
To accommodate the significant changes expected to come about as a consequence of climatic change,
society needs a continuing capacity for strategic planning. This planning should be of a long-term nature, and
designed to identify what the expected changes in the economy and society will be, and what resources will be
required. This capacity need not be a large centralized organization. Some centralized means of coordinating
the strategic analysis will be needed, however.
This nation is endowed with a remarkable set of land grant universities, most of which have Colleges
of Agriculture. These Colleges, which tend to focus on problems of the state in which they are located, should
be encouraged to take a longer term, strategic perspective to these problems, with particular emphasis on the
potential consequences of climatic change. The VS. Department of Agriculture has a unique analytical capacity
in its Economic Research Service. This Service should also be encouraged to take a longer-term perspective
focusing on climatic change. The Economic Research Service is also in a unique position to coordinate the
research and analysis which goes on at the state level.
Finally, significant changes in the economy can be expected to bring about significant changes in
infrastructure. Changed and new transportation and communication systems may be needed. New irrigation
systems may be required, together with new drainage systems. And new systems in the form of formal
adjustment policies and the institutions to implement them are required to accommodate the relocation of
economic activities. It is a proper role of government to provide such facilities.
13-13
-------
Schuh
CHAPTERS
CONCLUSIONS
Agricultural policies for climate changes induced by greenhouse gases need to focus on two broad classes
of problems: (1) reducing greenhouse emissions from agricultural sources; and (2) facilitating the adjustment to
the changes in climate in an efficient and equitable way.
Policies to deal with the first set of problems, in their order of importance, include: (1) deregulation of
US. agriculture by the dismantling of present commodity programs and the shift to a more market-oriented
agriculture, with closer integration of ILS. agriculture with the global agricultural sector; (2) shift of agricultural
science and technology policy to give more attention to reducing greenhouse emissions; (3) promote the
diversification out of rice production in the global economy; and (4) as appropriate, and as knowledge becomes
available, use taxes and subsidies to internalize the external costs and benefits of activities which generate
greenhouse emissions.
Policies to help agriculture adjust to changes in climate, in their order of importance, include: (1)
commodity policies which facilitate the adjustment of US. agriculture to changing conditions of demand and
supply in the global economy and in the United States; (2) explicit adjustment policies which facilitate the
adjustment of labor to changed employment opportunities and which support other agricultural input markets;
(3) expanded and redirected agricultural research programs to generate new production technologies for
agriculture, to better understand the aggregate adjustment problem agriculture will face, and to facilitate
adjustment at the individual farm level; (4) an expansion in irrigable area and measures that make for a more
efficient use of available water supplies, including increased investments in irrigation and water use and
management; (5) more strategic planning to make for a more efficient and equitable longer-term adjustment of
agriculture; (6) a strengthened information system on the consequences of climatic change so individual citizens
can vote more intelligently on the issues and can adjust their own economic activities in their own best interests;
(7) construction of new infrastructure in regions that become newly adapted to agricultural activities; and (8)
biocultural restoration in areas in which agriculture declines as to conserve resources for the future.
Timing is an important issue in taking action to reduce agricultural sources of greenhouse emissions and
in helping agriculture to adjust to possible greenhouse effects. There are lags in the build up of possible effects,
and similar lags in the response to any measures taken. Moreover, some of the measures proposed, such as
increased investments in agricultural research, have long gestation periods. In the case of agricultural research
in particular, the lag between increased expenditures on research and effects on farmers' fields is from six to ten
years.
Measures to reduce agricultural sources of emissions and to help agriculture adjust to possible
greenhouse emissions should thus begin immediately. An important feature of the measures that have been
outlined above is that most of them are in the best interest of the nation in their own right, and that they
improve the efficiency with which the nation's agricultural resources are used. The efficiency gains provide the
basis for a more rapid rate of economic growth, and thus provide the resources for dealing with greenhouse
effects if and when they come.
Finally, the extent to which a possible greenhouse effect is a global problem means that steps toward
international cooperation in dealing with the problem should begin immediately. A sense of urgency and
immediacy is especially important in addressing the problems associated with the agricultural sector.
13-14
------- |