PNERGY AND
^Resource
(0ONSULTANTS, INC
\
ECONOHIC IMPACTS OF
LOWER CROP YIELDS DUE TO
STRATOSPHERIC OZONE DEPLETION
)
^non.io • ei.oiri.cu.- n,*.(]. I;i:t:: • economics • engincerinc* • ;
• en\ironrm nt * K at!.>::s • ene o\ • environment • t
•uiics • er.^nttrinc ¦ m:»c\ Pi'• economics "• engineering • m*.
environment ¦ telecorn.municatKins • energy * environment • tel*
• engineering • modcMmc- • economics • engineering • modelhno
onment • telecommunications * energy * environment • telecom-
lgineering • modelling • economics • engineering • modelling • t
P.O. Drawer O. Boulder, CO 80306 (303)449-5515
BO LD R COLORADO WASHINGTON, D.C.
-------
ECONOHIC IMPACTS OF
LOWER CROP YIELDS DUE TO
STRATOSPHERIC OZONE DEPLETION
-------
ECONOMIC IMPACTS OF
LOVER CROP YIELDS DUE TO STRATOSPHERIC OZONE
September 30, 1987
By:
Robert D. Rove
Energy and Resource Consultants, Inc.
P.O. Draver 0
Boulder, CO 80306
(303) 449-5515
and
Richard M. Adams
Oregon State University
For:
Steven Anderson
U.S. Environmental Protection Agency
401 M Street, SV
Washington, DC 20460
[FOODS]
The information in this document has been funded by the United States
Enviromental Protection Agency under Contract No. 68-01-7033 Task Order No. 255.
It has not been subject to the Agency's peer and administration reviev. Mention
of trade name or commercial products does not constitute endorsement or
recommendation for use.
-------
TABLE OF CONTENTS
1.0 EXECUTIVE SUMMARY 1-1
1.1 Overview 1-1
1.2 Methods and Assumptions Used in the Analysis 1-1
1.3 Analyses and Results 1-3
1.4 Outline of the Report 1-10
2.0 LITERATURE AND SCENARIOS 2-1
2.1 Climate Scenarios 2-1
2.2 UV-B Crop Loss Studies 2-5
2.3 Tropospheric Ozone Crop Loss Relationships 2-16
3.0 THE ECONOMIC MODEL 3-1
3.1 Introduction 3-1
3.2 Background 3-2
3.3 The Farm Model 3-4
3.4 Generation of Regional State Activities and Crop Mixes 3-6
3.5 The Sector Model 3-8
3.6 Solution Procedure and Data Summary 3-9
3.7 International Trade Component 3-10
4.0 ANALYSIS PLAN AND RESULTS 4-1
4.1 Analysis Plan 4-1
4.2 Results 4-4
5.0 REFERENCES 5-1
APPENDIX: "An Assessment of the Economic Effects of Ozone on U.S. Agriculture."
"The Benefits of Pollution Control: The Case of Ozone and U.S.
Agriculture."
-------
List of Tables and Figures
Chapter 1.0
Page
Table 1-1 Estimates of Annual Economic Damage Due to a 15 Percent
Depletion of Stratospheric Ozone 1-6
Table 1-2 Potential Economic Surplus Changes in the Agricultural
Sector Due to Changes in Stratospheric Ozone — Medium
Scenario 1-9
Chapter 2.0
Table 2-1 Percent Changes in Global Ozone at 40 Degrees North
Under Alternative Control Scenarios 2-3
Table 2-2 Relationship Betveen Ozone Depletion and UV-B 2-4
Table 2-3 Changes in Tropospheric Ozone Concentrations Due to
Changes in Stratospheric Ozone Concentrations 2-6
Table 2-4 Summary of Field Studies Examining the Effects of UV-B
Radiation on Crop Yields 2-10
Table 2-5 Details of Field Study by Teramura (1981-1985) a/ 2-11
Chapter 3.0
Table 3-1 Primary Commodities Included in the Economic Model 3-7
Table 3-2 Secondary Commodities in the Economic Model 3-11
Chapter 4.0
Table 4-1
Table 4-2
Table 4-3
Table 4-4
Table 4-5
Table 4-6
Table 4-7
Table 4-8
Analysis Plan 4-2
Model Prices and Quantities vs. Actual: 1981-83 4-5
Economic Model Response to Tropospheric Ozone Changes
Only 4-7
Estimates of Annual Economic Damage Due to a 15%
Depletion of Stratospheric Ozone 4-9
Economic Model Response to Soybean Yield Changes 4-10
Potential Economic Surplus Changes in the Agricultural
Sector Due to Changes in Stratospheric Ozone — Lowest
Scenario 4-14
Potential Economic Surplus Changes in the Agricultural
Sector Due to Changes in Stratospheric Ozone — Medium
Scenario 4-15
Potential Economic Surplus Changes in the Agricultural
Sector Due to Changes in Stratospheric Ozone — Highest
Scenario 4-16
-------
1.0 DRAFT EXECUTIVE SUMMARY
1.1 OVERVIEW
As part of its stratospheric ozone protection plan, announced in the Federal
Register in January 1987, the EPA is evaluating the potential effects of
stratospheric ozone modification, including the resultant increased ultraviolet
radiation (UV-B), on the environment and human welfare. Researchers have shown
that many commercially important crops appear sensitive to UV-B. In addition to
UV-B effects, decreases in stratospheric ozone may increase levels of
tropospheric ozone. The EPA National Crop Loss Assessment Network. (NCLAN) has
demonstrated that tropospheric ozone has deleterious effects on crops with
important economic consequences.
This report presents analysis of the potential impacts of reduced agricultural
yields in the U.S. due to potential stratospheric ozone depletion. The research
develops generalized relationships between changes in stratospheric ozone and
economic loss. The results are illustrated with sample EPA baseline scenarios
for changes in stratospheric ozone through time.
The results of this study must be considered as suggestive of the potential
damages under the assumptions used. Many factors limit such an analysis. Most
importantly, changes in agricultural practices and prices through time are
unknown and, for the sake of this analysis, we have calculated damages based
upon recent conditions. Other important limitations include limited research
with uncertain conclusions concerning the relationships between changes in UV-B
and crop loss and between changes in stratospheric ozone and tropospheric ozone.
1.2 METHODS AND ASSUMPTIONS USED IN THE ANALYSIS
Key assumptions regarding atmospheric conditions were provided by EPA. These
assumptions included baseline time profiles of changes in stratospheric ozone
that may occur in the future without additional precursor controls (Table 2-1 of
the report). Data were also provided that suggest tropospheric ozone levels
1-1
-------
will increase by just less than one percent for each one percent decrease in
stratospheric ozone (Table 2-3 of the report).
Research suggests increases in UV-B may substantially impact the yield of
important commercial crops, but the evidence is still very limited. For this
analysis the results of Teramura (1987) were used to develop alternative
relationships between UV-B and soybean yields. Over the course of 5 growing
seasons Teramura exposed different soybean cultivars to two levels of UV-B
asserted to correspond to 16 and 23 percent stratospheric ozone depletion.
There are numerous anomalies and limitations with the Teramura data for use in
developing dose response relationships. These include apparent problems with
the experiments in two drought years. A substantial percent of the observations
suggest no statistically significant impact of UV-B on yield, some observations
suggest an initial increase in yields at the 16 percent depletion level, and
there are uncertainties in how the results of the cultivars should be translated
to the mix of cultivars grown in the field now and in the future.
Recognizing these uncertainties in the Teramura results, a "lower bound" of zero
is suggested for the effect of UV-B upon soybean yields. A "point estimate" of
a .3 percent decrease in yield for each one percent depletion in stratospheric
ozone was calculated based on the statistically significant data points in
non-drought years for all cultivars, and an "upper bound" of a .76 percent
decrease in soybean yield for each one percent depletion in stratospheric ozone
was estimated based upon all non-drought year data points (statistically
significant or not) for the most sensitive cultivar (Essex) in the Teramura
study.
Potential UV-B effects on corn and wheat were also considered. In examining
potential yield adjustments for corn and wheat one finds limited and ambiguous
results. For each crop, only one study has found statistically significant
yield depressions due to increased UV-B. Specifically, Teramura reports that
Biggs et al. found a five percent wheat yield reduction at a 16 percent ozone
depletion, while Eisenstart et al. claim a 23 percent corn yield reduction at a
seven percent ozone depletion. Vhile such estimates can perhaps be used with
soybean changes to provide an upper bound impression of direct UV-B yield
effects, these latter values seem extreme due to elements of the experiment
1-2
-------
design or data interpretation and because no other experiments on these two
crops could find any effect of enhanced UV-B radiation. Rather than base a
multi-crop assessment on single data points, ve instead evaluate the sensitivity
of the economic results to hypothetical potential changes in corn and wheat
yields as discussed below. For the base cases we assume no UV-B effect on corn
and wheat.
To evaluate the crop losses due to changes in tropospheric ozone changes, the
most recent NCLAN ozone crop loss functions are employed for soybeans, corn,
wheat, cotton, rice, barley, sorghum and forage.
The economic assessment is conducted with the NCLAN economic model. This model
is a quadratic programming model that accounts for behavior of farmers and the
agricultural markets in response to changes in production conditions. The model
also correctly calculates economic surplus measures for changes in crop yields.
One could obtain economic estimates with a simple damage function approach,
where estimated changes in crop yields are multiplied by current market prices.
However, past analyses have found that simple damage function approaches have,
by ignoring farm and market behavior, overestimated damage by up to 50 percent
in total for all crops, and up to a factor of 7 for individual crops. The use
of the more complex NCLAN model is justified, even though the UV-B crop loss
relationships are only poorly understood, because the tropospheric ozone crop
loss relationships are relatively well understood, the model is readily
assessable, and the model can provide other data of interest including the
relative share of damage incurred by consumers and producers both in the U.S.
and overseas.
1.3 ANALYSES AND RESULTS
The analytical portion of this work provides generalizable functions relating
changes in stratospheric ozone to economic loss that can be applied to future
control scenarios. The functions provide annual economic damage due to crop
injury in the U.S. for 1982, the base year for which the NCLAN model is
calibrated. These functions were derived through the following four analyses
involving 19 different runs of the economic model.
1-3
-------
1. Functions were developed relating changes in tropospheric ozone and economic
damage due to crop losses using the NCLAN crop response functions. No UV-B
damages were assumed in these runs. Based upon 9 runs on the economic model
the following function was estimated:
D1 = -0.0678 * T -0.000195 * T2 R2 = .996 (1)
where:
D1 = annual change in economic surplus, in billions of 1982 dollars, due
to tropospheric ozone only. (-) = damage.
T = percent change in tropospheric ozone where 10.5 percent is
expressed as 10.5. (+) = increase in ozone.
Using the data from EPA (Table 2-3 of the report), a relationship between
stratospheric ozone and tropospheric ozone was estimated as:
T = -.787 * S + .004* S2 if S > -32.6 R2 = .999 (2)
= -30.8 if S < -32.6
where:
T = Percent change in tropospheric ozone (10.5 percent = 10.5)
S = Percent change in stratospheric ozone (10.5 percent = 10.5) (-) ¦
depletion in ozone. Note the function slightly over predicts in
the midrange
Combining the economic function (Dl) with the estimated relationship between
stratospheric ozone and tropospheric ozone results in a "tropospheric ozone
only" loss, for a 15 percent depletion in stratospheric ozone, of $.892
billion per year ($1982), as reported in Table 1-1, where 15 percent was
selected for illustration purposes.
1-4
-------
2. Functions were developed relating changes in stratospheric ozone depletion
and economic damage to reduced yields of soybeans affected by increased
UV-B. Based upon 7 runs of the economic model, a relationship was developed
between soybean yield and economic damage assuming no tropospheric ozone
impacts.
D2 = 0.1068 * SOY -0.00029 * SOY2 R2 = .999 (3)
where:
D2 = annual change in economic surplus, in billions of 1982 dollars,
resulting from changes in soybean yields due to UV-B. (-) =
damage.
SOY = percent change in soybean yields due to UV-B. (-) = decreased
yield.
To convert the soybean damage function to stratospheric ozone depletion, the
following dose response relationships were based upon the Teramura results,
as discussed above, where S again refers to the percent change in
stratospheric ozone.
SOY = 0.30 * S (point estimate)
SOY = 0.76 * S (upper bound) (A)
SOY = 0.00 * S (lower bound)
Combining the soybean economic equation and the above dose response
Functions result in the following:
D2 = 0.03204 * S - 0.000087 * S2 (point estimate)
D2 = 0.08116 * S - 0.000220 * S2 (upper bound) (5)
D2 = 0.0 * S (lower bound)
The UV-B soybean only impact for a 15 percent depletion is estimated to be
$.486 billion per year ($1982) based upon the point estimate of crop loss,
and $1,255 billion per year using the upper bound estimate of crop loss.
1-5
-------
Table 1-1
Estimates of Annual Economic Damage Due To A 15 Percent
Depletion Of Stratospheric Ozone
Case/Assumptions
Economic Surplus
Total Change in
($ billions 1982)
1. Tropospheric ozone effects only - point estimate
$0,892
2. UV-B impacts on soybeans
- lower bound
- point estimate
- upper bound
$0,000
$0,486
$1,255
3. Combined tropospheric ozone
effects plus UV-B impacts on
soybeans
- point estimate
$1,378
THE FOLLOWING SENSITIVITY ANALYSIS ASSUMES HYPOTHETICAL UV-B INDUCED YIELD
LOSSES TO CORN AND WHEAT EQUAL TO THE POINT ESTIMATE DOSE RESPONSE FUNCTION
FOR SOYBEANS.
4. UV-B impacts on soybeans plus
equal percent impacts on corn
and wheat - point estimate $1,673
5. Combined tropospheric ozone
effects plus UV-B induced yield
impacts on soybeans, wheat and
corn - point estimate $2,588
1-6
-------
3. The best estimate of total economic damage due to U.S. agriculture losses
associated with stratospheric ozone depletion combines the UV-B soybean
central case estimate and the tropospheric ozone induced losses. One
approach would be to simply add the two equations; i.e., total damages = D1
+ D2. At the 15 percent depletion level this addition results in estimated
damages of $1,378 billion per year ($1982).
The simultaneous occurrence of the UV-B soybean impact and the tropospheric
ozone effect could result in damage estimates different from the addition of
the two impacts when considered separately. To test this possibility, the
economic model was rerun with both sources of crop loss considered
simultaneously for a 15 percent ozone depletion scenario. The resulting
damage estimate was less than 5 percent different from the corresponding
additive estimates. Therefore, the addition of the two functions serves to
adequately represent total estimated damages.
4. To test the sensitivity of the UV-B analysis to the incorporation of other
crops, it was arbitrarily assumed that the rate of change in yield loss due
to increased UV-B for corn and wheat equaled the point estimate between
percent ozone depletion and percent yield loss for soybeans. Then the
economic model was run for a 15 percent stratospheric depletion case with
the assumed soybean, corn and wheat yield losses. Use of results from this
assessment must be caveated as being based upon hypothetical assumptions for
corn and wheat that are not well grounded in the literature. The estimate
loss of $1,673 billion per year ($1982), about 3.4 times the soybean only
loss.
5. A final run of the economic model considered the hypothetical case of the
combined occurrence of tropospheric ozone losses, UV-B soybean losses using
the point estimate dose response function, and the assumed corn and wheat
UV-B losses with the soybean dose response function. Use of results from
this assessment must be caveated as being based upon hypothetical
assumptions for corn and wheat that are not well grounded in the literature.
The estimated total damages for a 15 percent ozone depletion case are $2,588
billion per year ($1982).
1-7
-------
Table 1-1 illustrates both the potential magnitude of damage and the
significance of the uncertainties in the analysis. First, one notes that
potential tropospheric ozone impacts may be a very significant source of damage,
exceeding the UV-B soybean losses. This suggests more information is needed to
solidify the stratospheric to tropospheric ozone links. Similarly, the
measurement of UV-B induced damage for crops other than soybeans appears to
exceed the current uncertainty in the soybean dose response functions. More
work is needed to solidify the understanding of the soybean response to UV-B and
to improve understanding of the relation for other crops.
To illustrate how the analysis may be used, ve next calculate changes in
economic values due to potential increases in tropospheric ozone and UV-B
associated with stratospheric ozone depletion for alternative baselines provided
by EPA. The "middle" baseline is illustrated in Table 1-2.
Other findings of interest concern the division of damages between producers and
consumers. The economic model calculates that consumer's and producers share
about equally the damage in the "tropospheric ozone only" and "UV-B soybean
only" analyses. However, as multiple impacts occur, the damage shifts to the
producers. In the case where both UV-B soybean impacts and tropospheric impacts
are considered, about two-thirds of the damage is incurred by producers, and
when potential UV-B impacts to other crops are also included about 80 percent of
the damage is incurred by producers. This is because producers are less able to
mitigate damage by putting more and more land in service or to switch crops as
yield losses increase for multiple crops.
Consumer surplus losses are incurred both domestically and internationally.
Foreign markets incur a decreasing share of the consumer surplus losses as more
impacts are considered, with the 79 percent share of the losses for the
tropospheric ozone only scenario, decreasing to 51 percent share in the combined
case of tropospheric ozone impacts plus UV-B impacts to soybeans, corn and
wheat.
Again, it must be stressed that the estimates in this analysis are conditional
upon the assumptions employed. The most important assumptions include the
1-8
-------
Table 1-2
Potential Economic Surplus Changes in the Agricultural Sector
*
Due to Changes in Stratospheric Osone — Radius Scenario
($ Billions 1982)
Economic Impact of UV-B Damage to
Economic Impact of UV-B Damage to Soybeans Plus Tropospheric Osone
Economic Soybeans Impacts
* Change in % Change in Impact of
Stratospheric Tropospheric Tropospheric Central Bigh Central Bigh
Tear Osone Osone Osone Case Case Case Case
198S
0.0
0.000
0.000
0.000
0.000
0.000
0.000
1990
-0.3
0.236
-0.016
-0.009
-0.024
-0.025
-0.040
1995
-0.5
0.395
-0.026
-0.016
-0.040
-0.042
-0.067
2000
-0 5
0.395
-0.026
-0 016
-0.040
-0.042
-0.067
2005
-0.8
0.632
-0.042
-0.025
-0.065
-0.068
-0.108
2010
-1.1
0.871
-0.059
-0.035
-0.089
-0.094
-0.148
2015
-1.4
1.110
-0.075
-0 .044
-0.113
-0.120
-0.189
2020
-1.9
1.510
-0.102
-0.060
-0.154
-0.163
-0.257
2025
-2.5
1.993
-0.135
-0 .080
-0.203
-0.216
-0.339
2030
-3.5
2.804
-0.191
-0.112
-0.286
-0.303
-0.477
2035
-4.2
3.376
-0.231
-0 134
-0.343
-0.365
-0.575
2040
-5.0
4 .035
-0.276
-0 160
-0.410
-0.4 37
-0.686
2045
-6 0
4 .866
-0.334
-0.192
-0.493
-0.527
-0.827
2050
-7.3
5.958
-0.411
-0.234
-0.601
-0.645
-1.012
2055
-8.9
7 321
-0.507
-0.285
-0.735
-0.792
-1.242
2060
-10.9
9 054
-0.630
-0.349
-0.904
-0.979
-1.534
2065
-13 2
11.085
-0.775
-0.423
-1.100
-1.199
-1.876
2070
-18.8
16.209
-1.150
-0.603
-1.585
-1.754
-2.736
2075
-22 2
19.443
-1.392
-0.712
-1.884
-2.105
-3.277
2080
-26.0
23.166
-1.676
-0.834
-2.223
-2.510
-3.899
2085
-30. 3
27.518
-2.014
-0.972
-2.613
-2.987
-4.627
2090
-34.5
30.800
-2.274
-1.108
-2.999
-3.382
-5.274
2095
-38.2
30.800
-2.274
-1.227
-3.345
-3.501
-5.619
2100
-41.3
30.800
-2.274
-1.327
-3.637
-3.601
-5.912
• Values for representative years through ti
values. (-) = damages; (+) = Benefits.
upon scenarios in Table 2-1. Estimates are not converted to present
-------
relationship between changes in tropospheric ozone and stratospheric ozone; the
interpretation of the limited evidence concerning UV-B soybean dose response,
and the assumed stratospheric ozone depletion UV-B relationship assumed in the
Teramura work; the sensitivity assumptions concerning the UV-B impacts on other
crops; and the assumption of unchanging agricultural technology, behavior and
prices into the future.
1.4 OUTLINE OF THE REPORT
This Executive Summary serves as Chapter 1 of the report. Chapter 2 reviews the
input data and develops assumptions concerning stratospheric conditions, the
relationship between stratospheric ozone and tropospheric ozone, and the crop
damage literature leading to dose response .functions for UV-B exposure. Chapter
3 discusses the technical nature of the NCLAN economic model. Chapter 4
provides the detailed results and caveats.
1-10
-------
2.0 LITERATURE AND SCENARIOS
2.1 CLIMATE SCENARIOS
Stratospheric Ozone Depletion Scenarios*
EPA has developed six possible baseline scenarios of global ozone depletion
through time. These scenarios are based upon the demand for goods using
chlorofluorocarbons (CFCs), methyl chloroform, carbon tetrachloride, and halons
(e.g., refrigerators, computers, automobile air conditioners, etc.) by examining
the historical relationship between economic activity and the use of these
chemicals. To reflect the large uncertainties inherent in these scenarios,
particularly with respect to technological innovation and the possibility that
industry and consumers will voluntarily limit their future use of these
chemicals due to concern about ozone depletion, a vide range of baseline
scenarios were examined. The scenarios range from a voluntary 80 percent
phase-dovn in the use of CFCs by 2010 to an average annual grovth in use of 5
percent per year from 1985 to 2050. For ozone-modifying gases other than CFCs,
scenarios vere based on recently measured trends, with uncertainties being
evaluated by considering a range of future emissions and concentrations.
Next, atmospheric chemistry models vere used to assess the potential effects of
possible future changes in atmospheric concentrations of these ozone-modifying
gases upon stratospheric ozone. Based upon the models' changes in stratospheric
ozone, concentrations at latitudes of 30, 40, 50 and 60 degrees north vere
projected in five year increments from 1985 to 2100. The methods to calculate
these projections and their reliability are discussed in EPA (1986, Chapter 3).
Of the six scenarios, five address different rates of CFC grovth vithout
legislation and the sixth equals projections under the EPA-proposed CFC
phasedovn.
1. Scenario:- aie those available at the time this research was initiated.
Subsequent refinements are expected.
2-1
-------
Of the five baseline scenarios, each is perceived as equally likely. These five
scenarios are presented in Table 2-1 for AO degrees north latitude, as the
primary crop production regions of the U.S. lie between about 28 and 44 degrees
north. Further, the difference between the projected ozone depletion at 30 and
40 degrees north is always less then 10 percent of either estimate. The
estimates do differ more substantially between 40 and 50 degrees north, but
limited crops are grown above 40N. As a result, the error is minimal in just
using 40N estimates for this analysis. By assumption, the depletion of
stratospheric ozone is limited not to exceed 43.6 percent.
Effects of Stratospheric Ozone Depletion on UV-B
An algorithm designed to estimate the ultraviolet solar flux that reaches the
earth's surface at any location at alternative times of the year has been
developed by Serafino and Frederick. (1986) and is used by EPA to estimate
changes in UV-B for each scenario. The relationship between changes in global
stratospheric ozone and average UV-B in the month of June is illustrated in
Table 2-2. For a crop assessment, different UV-B estimates for
different regions could be utilized, but limited variation across regions and
the limited precision of the available dose-response data suggests that the use
of Region 2 estimates would be sufficient. Two measures of change in
ultra-violet radiation in the United States are reported in Table 2-2; the R-B
meter and human erythema. The latter predicts a rate of change about double the
former and may be the most representative measure for effects on humans, animals
and vegetation. The R-B meter measure is the one most used in the agricultural
impacts assessments to date.
Effects of Stratospheric Ozone Depletion on Tropospheric Ozone
Changes in the stratospheric ozone layer and in global warming may result in
substantial increases in smog and acid rain precursors. This is particularly
important to the agricultural analysis, as the relationship between ambient
ozone, crop loss and economic welfare measures is relatively well established
from work completed for the National Crop Loss Assessment Network (Heck et al.
1984). This work suggests even a 10 percent change in average annual growing
season ozone concentrations may have substantial economic value.
2-2
-------
Table 2-1
Percent Changes in Global Ozone at 40 Degrees North
Under Alternative Control Scenarios*
Scenario of CFC Grovth to 2050 Percent Change in Global Ozone
Lovest (OX)
Low (1.2X)
Medium (2.5Z)
High (3.8Z)
Highest (5Z)
1985
0.0
0.0
0.0
0.0
0.0
1990
-0.2
-0.2
-0.3
-0.3
-0.3
1995
-0.4
-0.4
-0.5
-0.6
-0.7
2000
-0.4
-0.4
-0.8
-1.0
-1.2
2005
-0.5
-0.8
-1.1
-1.5
-1.9
2010
-0.6
-0.9
-1.4
-2.1
-2.9
2015
-0.6
-1.1
-1.9
-3.0
-4.5
2020
-0.6
-1.4
-2.5
-4.3
-6.3
2025
-0.6
-1.6
-3.5
-5.6
-9.0
2030
-0.5
-1.8
-4.2
-7.3
-13.2
2035
-0.5
-2.1
-5.0
-9.7
-21.2
2040
-0.4
-2.3
-6.0
-13.3
-39.6
2045
-0.3
-2.6
-7.3
-19.1
-43.6
2050
-0.2
-3.4
-8.9
-30.3
-43.6
2055
-0.1
-3.7
-10.9
-43.6
-43.6
2060
0.1
-3.9
-13.2
-43.6
-43.6
2065
0.3
-4.2
-18.8
-43.6
-43.6
2070
0.4
-4.4
-22.2
-43.6
-43.6
2075
0.6
-4.7
-22.2
-43.6
-43.6
2080
0.8
-4.9
-26.0
-43.6
-43.6
2085
1.0
-5.1
-30.3
-43.6
-43.6
2090
1.2
-5.2
-34.5
-43.6
-43.6
2095
1.4
-5.4
-38.2
-43.6
-43.6
2100
1.6
-5.5
-41.3
-43.6
-43.6
* Grovth in other trace gases constant across scenarios. CFC use constant
after 2050. See EPA (1986, Chapter 3) for further discussion. (-) indicates
ozone depletion.
2-3
-------
Table 2-2
Relationship Betveen Ozone Depletion and UV-B*
Region 1
Region 2
Region 3
Area
Northern U,
.S.
States
Middle U.S. States
Southern U.S.
States
Average Latitude 43.3N
39.IN
31.8N
Percent Ozone
Depletion
Percent Change in UV-
-B
Human
Human
Human
R-B Meter
Erythema R-B Meter
Erythema R-B Meter
Erythema
-10.0
- 7.4
-14.4
- 7.3
-14.3
- 7.0
-14.0
- 5.0
- 3.8
- 7.6
- 3.7
- 7.5
- 3.6
- 7.4
- 2.0
- 1.5
- 3.1
- 1.5
- 3.1
- 1.5
- 3.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.0
1.6
3.3
1.6
3.3
1.6
3.4
5.0
4.1
8.6
4.0
8.6
3.9
8.6
10.0
8.5
18.5
8.4
18.4
8.1.
17.9
20.0
18.4
42.1
18.1
41.9
17.5
41.1
30.0
30.1
74.1
29.7
43.3
28.4
70.7
* Measured
in
percent
change in
annual average
for the
month of June
based upon
Serafino
and
Frederick
(1986)
and used in EPA
(1986).
-------
Work in progress by Styles et al. has been used to estimate how changes in the
stratospheric ozone layer relate to changes in the average tropospheric ozone
concentrations at different cities for the EPA scenarios. The results for the
"mid-case" location of Philadelphia are summarized in Table 2-3.
The results in Table 2-3 suggest that each one percent change in ozone in the
stratospheric layer will have an effect on annual average tropospheric ozone
concentrations of about .7 to .9 percent, when averaged across several
locations. Preliminary model results suggest the rate of tropospheric ozone
change may be greater in northern U.S. cities and smaller in southern U.S.
cities. By assumption, the percent change in tropospheric ozone is limited to
30.8 percent, the value predicted in 2085 under the medium CPC growth scenario.
Additional analysis of the Table 2-2 data are found in Section A.2.
2.2 UV-B CROP LOSS STUDIES
This review focuses on three crops: soybeans, wheat and corn. While research
has, to some degree, considered other crops, these three vere selected because:
o They are the top 3 crops in terms of commercial value of production in
the U.S., totaling over 54 percent of total U.S. production value;
o One or more studies suggest UV-B effects yield for each of the crops;
and
o There have been multiple studies performed for each of these crops, with
replications over time to at least allow one to make reasoned
speculation regarding possible yield impacts.
This review finds there is limited evidence to construct dose-response
relationships over a range of UV-B values for any crop. In fact, most studies
have only examined UV-B for 16 and 23 percent stratospheric ozone depletion
2-5
-------
Table 2-3
Changes in Tropospheric Ozone Concentrations
Due to Changes in Stratospheric Ozone Concentrations*
Medium Growth In CPC Scenario at 50 Degrees North Latitude
X Change in Tropospheric
X Change in Ozone (Annual Average Across
Tear Stratospheric Ozone 3 U.S. Cities)
1985
0.0
0.0
1990
- 0.4
0.3
1995
- 0.7
0.6
2000
- 1.2
1.0
2005
- 1.6
1.3
2010
- 2.1
1.8
2015
- 2.8
2.3
2020
- 3.6
3.0
2025
- 4.8
4.0
2030
- 5.5
4.6
2035
- 6.4
5.4
2040
- 7.6
6.3
2045
- 9.0
7.5
2050
-10.8
9.0
2055
-13.0
10.8
2060
-15.6
13.0
2065
-18.5
15.8
2070
-21.8
19.1
2075
-25.7
23.0
2080
-30.0
27.4
2085
-34.7
30.8
2090
-39.4
30.8
2095
-43.6
30.8
2100
-47.1
30.8
* Summarized from results provided by ICF, Inc., based upon results and models
in Styles et al. (1986).
2-6
-------
scenarios. One of the significant limitations of studies to date has been the
failure to examine crop impacts at multiple UV-B levels.
Background
The primary objective of this review was to identify a series of dose-response
relationships that could be used in combination with alternative stratospheric
ozone-UV-B scenarios to measure changes in crop yields. Such relationships are
available for gaseous pollutants and acid deposition (e.g., Heck et al.) and
have been the basis for numerous environmental assessments of these pollutants
(e.g., Adams and McCarl, Kopp et al.). However, as the review progressed, it
became clear that the extant literature (data) on UV-B stress is insufficient to
allow the construction of a set of representative response functions for these
three crops for use in major agricultural areas of the United States (Teramura,
personal communication). Instead, what emerges from this review is a body of
plant science knowledge that has evolved from exploratory greenhouse or growth
chamber investigations of potential plant sensitivity (screenings) to a series
of ongoing field experiments aimed at quantifying the responses of specific crop
cultivars, mostly of soybean, to changes in UV-B radiation. Field studies are
most relevant for use in economic assessments, as field studies are believed to
be the best indicators of how commercially grown crops will respond to stress
(Heck et al.). Teramura's work (1983) with UV-B soybean response indicates
significant differences between field and greenhouse response, reinforcing the
need to use field-based response information in the assessment. Unfortunately,
few field studies use an experimental design amenable to estimation of a
response function. As a result, while the level of understanding has advanced
sufficiently to allow some generalization of UV-B effects (Teramura, 1986,
1987), the evolution has not advanced to the point where representative response
functions exist for these crop species.
Despite the lack of available response functions for use in economic
assessments, some existing studies do provide quantitative evidence of changes
in yields of specific crops associated with certain depletions of the ozone
column. For example, field-based experiments with soybeans have demonstrated a
statistically significant negative (yield depressing) effect associated with
2-7
-------
increases in UV-B radiation for some cultivars on some years (Teramura, 1986).
Effects on the other two crops of interest, corn and wheat, however, are less
secure in that alternative signs have been noted across different experiments
(Bartholic, Halsey and Garrard, Eisenstark et al.). However, if one focuses
only on those few UV-B yield experiments that demonstrated statistically
significant effects, it is possible to develop a set of preliminary yield
adjustments for use in this assessment. It should be stressed that the
potentially large uncertainties in these responses places greater importance on
the sensitivity analyses to accompany the assessment.
General Findings
Most of the published research on the effects of UV-B radiation on crops stems
from two major research programs. The first is the Climate Impacts Assessment
Program (CIAP) funded by the Department of Transportation during the mid-70s.
Most of the initial research into the effects of enhanced UV-B radiation appear
to have been motivated by the CIAP program. Since the elimination of this
program, the second major funding source for this research has been the
Environmental Protection Agency's Photobiology Research Program. In fact,
ongoing UV-B research on crop productivity appears to be funded exclusively by
the EPA (Vorrest, personal communication). Thus, the literature reveals two
major thresholds, or stages, in the understanding of photobiology effects on
crops; the first is defined by what was learned in the CIAP (mostly from
greenhouse and growth chamber studies), and the second represents advancements
in the CIAP findings brought about by the EPA photobiology program.
Another notable feature of UV-B research is the relatively small number of
individuals who have made major contributions to this area. Much of the
published literature on vegetative effects traces back, to Caldwell and
associates work at Utah State University (see, for example, Caldwell 1971),
Biggs and associates at the University of Florida, and more recently, some of
their former students, such as Alan Teramura and colleagues at the University of
Maryland. Professor Teramura's recent reviews of vegetative effects for the EPA
and U.S. Senate (1986, 1987), as well as personal communication with Teramura;
Dr. Joe Sullivan, University of Maryland; and Dr. Robert Vorrest, EPA,
Corvallis, form the basis for much of the review and evaluation contained here.
2-8
-------
Teramura's reviev of the effects of UV-B on plants indicates that of the 200 or
so plant species and cultivars that have been screened for UV-B sensitivity,
about tvo out of three show some degree of sensitivity. It should be noted that
sensitivity does not necessarily imply an adverse consequence (yield reduction),
as some species have demonstrated enhanced productivity under UV-B stress.
These positive responses may be an artifact of experimental design, although
Teramura notes that plants do have some natural adaptive mechanisms to deal with
ambient levels of UV-B radiation. The nature of the vegetative response to date
has been derived largely from species and cultivars that are identified as
agriculturally important. Hence, little is known about large groupings of plant
types, such as forests.
With respect to agricultural crops, Teramura has noted a general trend in the
crop experiments for reductions in yield as a result of increasing UV-B
radiation. Specifically, he estimates that approximately 70 percent of crops
species tested were sensitive to UV-B stress. The remaining 30 percent were not
affected. His review also indicates that, as with many other types of
environmental change, UV-B stress is just one factor among many that has the
potential to alter crop yields. It is likely that there is interaction among
these many effects and that it is these interactions that are of consequence.
For example, it is suspected that increased UV-B stress may increase the
incidence of disease or insect attacks on crop species. Reinforcing this
possibility is the high degree of temporal variability in response of a given
crop cultivar across time. Teramura's soybean research displays examples of
cultivar responses being negative in some years yet positive in others.
However, there are few studies that explicitly include interactions, although
Teramura has ongoing interaction studies (between CO2 and UV-B). Furthermore,
UV-B stress may alter the competitive balance among species, although at least
one study indicated that the competitive advantage may go to the agricultural
crop and not to the competing or weed species (Caldwell).
Specific Effects on Crop Yields
Nine field-based studies have looked at the effects of UV-B on crop yields for
22 crops. Seven studies have included the three crops of interest here. Tables
2-9
-------
Table 2-4
Summary of Field Studies Examining the Effects of
UV-B Radiation on Crop Tields
Values Represent Percent Changes From Controls
Ambler Bartholic Biggs & Biggs Eisenstark Hart
et al. et al. Kossuth et al. et al. et al.
(1978)a (1975)b <1978)c (1984)c <1985)c (1975)a
Wheat
(Triticum
Aestivum)
Corn 0 +29 - +39
(Zea mays)
Soybean 0
(Glycine
max)
-5d
0 0 -23 - -32d 0
0
Source: Teramura (1987), Table ll-4c.
a. Unfiltered Vestinghouse BZS-CLG and FS-40 sunlamps.
b. Ambient UV filtered with Mylar Type S or polyethylene.
c. Vestinghouse's FS-40 sunlamps filtered with cellulose acetate or Mylar.
d. Statistically significant effects detected in one or more years.
2-10
-------
Table 2-5
Details of Field Studies by Teramura <1981-1985) a/
Cultivar b/
Year:
1981
1982
1983
10% yield reduction) in the greenhouse, these six
were chosen for field experimentation. The six represent the full range of. UV sensitivity found in the greenhouse, including
very sensitive and very tolerant cultivars. Beginning in 1983 only Essex (very sensitive) and Williams (very tolerant) were
planted in the field to increaso the experimental sample size to 200. As shown in the table, yields for Essex were generally
reduced at the higher UV level (otone change of -25%), but were mixed (relative to controls) at the other UV level. Yields for
Williams were generally enhanced by increased UV. Of note is that Essex is currently replacing other older cultivars
(including Williams) and is becoming one of the most widely planted soybeans in the U.S. In a UV enriched environment, Essex
will be deleteriously affected. Therefore, superior cultivars being developed today by crop breeders may not be suitable for
the future should the UV environment change,
c/ Drought year. Low yields for both controls and experimental (i.e., dosed) plants,
d/ HE a Not evaluated.
e/ Significantly different at paO.OS level.
-------
2-4 and 2-5 summarize the range of findings for these three crops. Despite the
great degree of uncertainty evident from these tables, there is evidence of a
negative or yield-depressing effect of enhanced UV-B stress in some of the years
for some of the crops. Teramura's work over the period 1981-1985 has
demonstrated a range of sensitivities to UV-B radiation across numerous soybean
cultivars and the specific magnitude of the yield changes for several key
cultivars. One cultivar, Essex, displayed a statistically significant 25
percent yield reduction in one experiment under a simulated 25 percent
stratospheric ozone column depletion, and a negative response to this exposure
three out of four years. Other cultivars (e.g., Bay, Forrest) were less
sensitive or displayed no sensitivity to enhanced UV-B radiation. One cultivar,
Villiams, generally shoved a positive response to UV-B enhancement.
Nonetheless, the importance of Essex in the total mix of commercial soybean
cultivars and the intermediate sensitivity .displayed by other important
cultivars, such as York and James, suggest that under levels of ozone depletion
being discussed in current policy deliberations, some reduction in aggregate
soybean yields in some years appears to be likely.
Information with respect to corn and wheat, the other two crops of interest in
this assessment, is even less clear than for soybeans. One multi-year study by
Biggs et al. (1984), at the University of Florida, demonstrates a statistically
significant yield depression of spring wheat associated with a UV-B Increase in
one year. Specifically, a 25 percent increase in UV-B resulted in a 5 percent
reduction in wheat yields. Other studies have been unable to establish a
statistically significant effect of UV-B on wheat yields (Amber et al.). Yield
response for field corn under ozone depletion scenarios is mixed. Eisenstark et
al. report yield depressions of up to 32 percent associated with 21 percent
depletion in the ozone column. Conversely, other studies have shown no effect
of UV-B on corn yields (Biggs et al., 1984).
In the comparisons of wheat and corn response reported here, it is important to
note that cultivars varied across each experiment. In view of the cultivar
response differences reported by Teramura for soybeans, some genotypic
differences may also exist for corn and wheat. An implication is that the
contradictory results among each crop's experiments do not necessarily rule out
adverse consequences of UV-B radiation on some subset of commercial cultivars.
2-12
-------
In his review, Teramura concludes that, despite the conflicting results, "there
are still more instances of significant reductions in yields than reports of no
effects" (Teramura, 1987). Vhile that statement reflects the state of knowledge
concerning crop response, it does not address the positive responses for some
cultivars nor the temporal variability across experiments. Ultimately,
sensitivity screenings of commercially-grown wheat and corn cultivars, along
with interactive studies of specific environment variables, are needed to test
these issues.
Effects of UV-B on Crop Quality
Like yield, the quality of crop production has economic implications. Evidence
to date suggests that at sufficiently high levels, UV-B stress can affect the
quality of crops. This phenomenon appears to be particularly important to
certain types of vegetable crops, where the reported higher incidence of cracks,
sun scald, and other factors would render the crop less valuable (Biggs and
Kossuth). Vith respect to the crops of interest here, namely, soybeans, corn
and wheat, Teramura (1982, 1985) notes that seed protein within some soybean
cultivars declined slightly in two of the four years of the crop experiments.
Further, seed oil concentrations were also slightly reduced in one of the
experiments. As was the case with yields, information on corn and wheat with
respect to crop quality is ill-defined at present. A related issue concerns the
likelihood of increases of disease and pest attacks associated with UV-B stress.
Again, Teramura (1982) reports some changes in insect and disease frequency
associated with UV-B stress. In some situations, the incidence of insect or
pathogen infestation appears to have been reduced (Owens and Krizek, Essex), in
others the opposite situation (increase in infestation) occurred. Thus, while
potentially an issue, there are insufficient data to include crop quality as a
component of an economic assessment.
UV-B Dose Response Relationships Selected for the Empirical Analysis
Soybean Central Case Estimates. Teramura's studies at Maryland are relied upon
for this assessment, as they are the only study that represent field conditions,
and the study spans the longest time period at any site (five years of reported
data). However, limitations in the available research indicate that caution is
2-13
-------
required in using these data to construct a "central case" or representative
analysis for soybeans over the entire U.S.
The Teramura data can provide a number of distinct interpretations concerning
the potential effects of UV-B radiation on soybean yields. Specifically, over
the multiple cultivars and years, both increases and decreases on yields are
observed under two levels of ozone depletion, 16 and 25 percent. In fact, the
data suggests a potential trend of increasing then decreasing yields as UV-B
increases. This might be the result of experimental design problems, UV-B
affecting insects more severely than plants at low increases, or other unknown
reasons. Further, the effects vary within a cultivar and year across the two
treatment levels. For example, in at least five experiments, the effect of
enhanced UV-B on yield reversed between the simulated 16 percent ozone depletion
scenario and the 25 percent depletion level.
Another complication in interpreting the Teramura results is that drought and
experimental conditions during the 1984 and 1985 crop years had a dramatic
(positive) effect on yield responses, particularly at the 16 percent depletion
treatment. The author has indicated that drought conditions may have affected
the conduct of the experiment in ways other than direct effects on yields and
the results may be less reliable. However, extensive NCLAN and other literature
does suggest that drought conditions reduce agricultural plant response to
increased levels of ozone, from which one might infer the drought year results
are plausible.
To calculate a representative case analyses for soybeans, we first limit the
acceptable data to those reported by Teramura as statistically significant
(.05). These observed yield changes are then added across each cultivar and
year and the simple average yield change for each treatment is recorded. In
doing this calculation, the "drought" years experiments of 1983 and 1984 are
excluded from the data. Using this highly restricted set of observations
results in a response of approximately -.3 percent yield change for each one
percent depletion in stratospheric ozone. As a point of comparison, this
soybean yield response is fairly close to the weighted average response of
soybeans to tropospheric ozone as reported by NCLAN researchers.
2-14
-------
Note that this simple calculation procedure gives equal weight to each
observation in the data set. Thus, a yield change for the Williams cultivar in
1981 is assumed to be equally valid as for the Essex cultivar in 1985. The
problem with such an assumption is that cultivars are not equally represented
over time in the aggregate production of soybeans in the U.S. For example,
Villiams accounted for almost ten percent of total soybean production in 1982,
while Forrest, York, and Essex were much less. However, cultivar mixes change
over time due to genetic improvements. Because of the evolution in genetic
materials, anticipating the national composition of cultivar mixes more than
five years in the future is a difficult task. The assumption of equal weight
implies that the set of cultivars in the Teramura data will be representative of
cultivar mixes over the time span of the stratospheric ozone depletion analyses.
This assumption is more tenuous, given that the cultivars in the Teramura data
set are mid-latitude, intermediate length growing season cultivars. Soybean
cultivars from more northern or southern growing regions are thus not well
represented in these data. Fortunately, the Essex and Villiams cultivars used
most extensively in the Teramura work do represent the responsive and
unresponsive range of cultivars tested, and the majority of cultivars tested
were at least somewhat responsive.
Soybean Upper Range Estimate. The upper range dose response function was
calculated using all observations for non-drought years for the Essex cultivar
only. An average response of .76 percent change in yield for each one percent
stratospheric ozone depletion was estimated. This is perceived as the upper
bound based upon Essex as the most consistently sensitive cultivar in the
Teramura study.
Soybean Lower Range Estimate. The lover range estimate is taken to be zero, or
no response to UV-Bs based upon the significant uncertainty in the experimental
results to date. Some cultivars may even increase in yields, but there is no
assurance these are, all things considered, the highest yield cultivars. The
early Teramura results suggest that the majority of cultivars do show some
adverse yield response to UV-B.
2-15
-------
In an alternative calculation of yield changes for soybeans, the drought years
of 1984 and 1985 were included in the data set. This results in a "mixed" yield
response under the two treatment levels. Specifically, at a 16 percent ozone
depletion assumption, soybean yields vould actually increase by 4.8 percent (a
+.3 response per one percent stratospheric ozone depletion). However, at a 25
percent ozone depletion, yields would decrease by 2.2 percent (a -.1 yield
response per one percent depletion in stratospheric ozone). These mixed
results, with the other limitations noted, again suggest zero as a reasonable
lower bound.
Wheat and Corn UV-B Dose Response Assumptions. In examining potential yield
adjustments for corn and wheat there are only two statistically significant
results reported in the UV-B literature, one for each crop. Specifically,
Teramura reports that Biggs et al. found a five percent wheat yield reduction at
a 16 percent ozone depletion, while Eisenstark et al. claim a 23 percent corn
yield reduction at a seven percent stratospheric ozone depletion. Vhile such
estimates can perhaps be used with the soybean changes to provide an upper bound
impression of direct UV-B yield effects, these latter values seem extreme in
that no other experiments on these two crops could find any effect of enhanced
UV-B radiation.
Rather than base a multi-crop assessment on single data points, we instead
evaluate the sensitivity of the economic model results to the potential for
changes in corn and wheat yields. Strictly for the purpose of a "what if"
sensitivity analysis, we assume a yield loss response relationship for corn and
wheat equal to the soybeans central case response rate.
2.3 TROPOSPHERIC OZONE CROP LOSS RELATIONSHIPS
Research by the National Crop Loss Assessment Network (NCLAN) and others have
identified a number of commercially important agricultural crops as being
sensitive to ambient levels of tropospheric ozone. NCLAN research has estimated
dose response functions for many major crops. Appendix A includes a summary of
the NCLAN functions as of 1985. Continued NCLAN work to update the functions
are included in the model run for this report, but represent relatively minor
2-16
-------
adjustments compared to other assumptions required in this analysis. It should
be noted that potentially sensitive crops such as fruits and nuts are omitted
and the economic estimates are therefore likely to be understated.
2-17
-------
3.0 THE ECONOMIC MODEL
3.1 INTRODUCTION
This section discusses the structure and assumptions of the economic model used
in the analysis. The economic model has recently been used in a series of
regulatory analyses of tropospheric ozone performed for EPA's National Crop Loss
Assessment Network (NCLAN). The features of this model and empirical results of
the ozone analysis are described in detail in Adams, Hamilton and McCarl (1984).
Much of the following discussion has been drawn from that report.
The primary differences between the model reported in Adams, Hamilton and McCarl
and the current version of the model, as noted subsequently, include:
1) updating of model to reflect economic, agronomic and environmental
conditions through 1983;
2) slightly greater spatial resolution (from 55 to 63 contiguous
production regions) in the U.S.;
3) generalization of the model to allow characterization of individual or
multi-year "base" periods (e.g., 1983 or 1980-1983);
4) reestimation and expansion of livestock components of the model; and,
5) adjustment of demand and supply elasticities to reflect changes in
national and world markets.
The model is based upon detailed analysis of the behavior of typical farms
throughout the country and of the agricultural sector as a whole. Changes in
theoretically correct economic surplus measures are calculated for changes in
ambient ozone depletion scenarios and incorporate individual farm and market
response behaviors such as crop substitution and price induced changes in
demand.
3-1
-------
One could more simply obtain economic estimates of damage due to ozone depletion
with a damage function approach where estimated changes in crop yields are
multiplied by current market prices. However, past analyses have found that
simple damage function approaches have overestimated damages by at least 50
percent or total for all crops and up to a factor of 7 for individual crops by
ignoring mitigating behavior (Rowe and Chestnut 1985). The use of the more
complex model is justified even through the UV-B crop loss relationships are
only poorly understood. This is the case because the tropospheric ozone crop
loss relationships are relatively well understood and comprehensive, the model
is readily modified for this application, the model can address export
implication, and by using correct methods the model adds limited further
inaccuracy to the already uncertain analysis.
The remainder of this section summarizes the technical, conceptual and
application details of the model, which the non-economist may wish to skip over.
3.2 BACKGROUND
Economists have devoted considerable effort to assessing the consequences of
policy or technically induced change on participants within the agricultural
sector. Many types of changes have been examined. For example, Freeman (1979)
reviews approaches examining the consequences of environmental change; Feder,
Just and Zilberman (1983) examine the consequences of technical changes on
agriculture, and Adams, Hamilton and HcCarl review the effects of alternative
air pollution levels or policies on agriculture. The assessments of such
changes have been performed at different levels of generality within the
agricultural sector. In particular, assessments of the benefits of change have
been performed at both the farm and sectoral levels. Analysis at each of these
levels potentially leads to quite different results. For example, it is not
difficult to imagine a case in which the analysis on a single Corn Belt farm
would lead to the conclusion that a ban on pesticides without effective
substitutes would lead to Corn Belt farmers receiving less income. But
conversely, when considering a similar ban where the pesticide is widely used
throughout the Corn belt on commodities subject to inelastic demand, it is
possible that an increase in farm sector gross revenue would occur while
3-2
-------
consumers would lose. Ultimately, changes in producer welfare will be a
function of the characteristics of demand and supply. This suggests the need to
consider broader implications in evaluation of induced change rather than simply
isolating the farm level effects as economists have traditionally done (Heady
and Srivistava, 1975; McCarl and Spreen, 1980; and Norton and Schiefer, 1980
review such larger efforts).
Focusing on broader effects typically involves research tradeoffs. The
evaluation of changes at the sectoral level often requires one to sacrifice
microeconomic detail to keep the problem tractable. This can have serious
consequences. For example, when doing appraisals of induced change with
aggregate programming models (i.e., models such as those used in Heady and
Srivistava, 1975; Baumes, 1978; or Burton, 1982) one often finds extreme
specialization in production. That is, on^ gets solutions where whole regions
are devoted to a single crop. This situation usually leads to the imposition of
inflexible "flexibility" constraints (Sahi and Craddock, 1974). A way to avoid
this specialization in production and thereby generate more plausible results
within a sectoral analysis is to link microeconomic considerations with the
sector model through a Dantzig-Volfe decomposition scheme using heuristic
procedures (McCarl, 1982a). Implementing such an approach requires both
detailed farm-level models and a macro or sector model.
The methodology to be used here contains such a linking of micro (farm level)
and sector models. One component of the methodology is the agricultural sector
model, a price-endogenous mathematical programming model of the agricultural
sector, i.e., an activity analysis spatial equilibrium model (Takayama and
Judge, 1971). Such sector models have been used extensively by agricultural eco
policies or technological change (Heady and Srivistava, 1975; Duloy and Norton,
1973). Among the various analytical techniques available to formulate policy
models, mathematical programming has proven to be a particularly useful tool
given its ability to predict potential consequence of as yet unrealized
policies.
The second component of the methodology consists of a series of farm models
based on linear programming (LP) models utilizing historical data.
3-3
-------
Specifically, historical data were used to generate whole farm plans. Crop
records by state for 1970-1983 were used both to develop representative state
level crop mixes and to derive econometric estimates of crop yield changes
associated with crop mix changes. In turn, the whole farm plans were used to
generate activities for the sector model documented utilizing the USDA (FEDS)
budgets. Such detailed specification and estimation of crop-mix activities was
used to overcome the aggregation problems identified in McCarl (1982a).
The following sections provide a detailed discussion of the farm model, the
sector model and the linkage of these models to assess the economic impacts of
environmental change on agriculture.
3.3 THE FARM MODEL
As noted above, the mathematical programming model applied here consists of both
a micro or producer level component and a sector component. The mathematical
programming sector model contains activities that represent production and
consumption of various commodities. The production and consumption sectors are
assumed to be made up of a large number of individuals, each of whom maximizes
some objective function under competitive market conditions. Specifically, each
producer is expected to maximize profits subject to a set of technical and
behavioral constraints. That is:
n m
Profits - I ( tph 2hk - I r, «,k)
k=l h=l i=l
where:
p^ is the market price per unit of the h1^ output (h=l,..., m); and
is the yield of the h**1 output from the k**1 production process
(h=l, . . . , m; k=l,..., n);
3-4
-------
rj is the market price per unit of the i1*1 purchased factor
s).
is the use of the i**1 purchased factor in the k**1 production process
(i=l, . .. , s; k=l,..., n).
This constrained profit maximization problem is then specified as a mathematical
(linear) programming problem (McCarl and Spreen, 1980). Necessary and
sufficient conditions for solution of a constrained maximum problem are defined
by the Kuhn-Tucker conditions, conditions easily traced to standard
microeconomic marginal conditions for profit maximization.
In this analysis, producer level behavior for each production region is
portrayed by a series of representative farms based on FEDS budgets, which
define input and output relationships to represent crop production. These farm
models are assumed to mimic the input-output or technical and economic
environment of the producers in a given region. The farms are modeled with
considerable detail on cropping activities (e.g., twelve annual crops plus hay
and beef, pork, and milk production), input use, and environmental and other
fixed constraints. Multiple activities are defined for a given crop in a given
region to cover the wide range of production technologies available to
producers. Such detail is needed to model adequately potential producer
mitigative behavior in the face of environmental changes. As noted by Crocker
(1982), failure to account for these adaptive opportunities tends to overstate
potential damages and understate benefits.
The multiple activities (within crops) generate the set of primary agricultural
commodities that interface with the sector models. These primary commodities
are listed in Table 3-1. In addition to crops, the analysis also includes
livestock products. An endogenous livestock component is needed, given that
livestock are the "consumers" of most of the primary crop commodities, as well
as accounting for approximately 50 percent of overall economic activity in the
agricultural sector.
3-5
-------
RFP ROUTING
RFP #: DACW72-88-R-0001
RFP Title:US ARMY: Analytical Support Services for Water Resources Policy
Programs & Projects
Date Proposal Due 01/05/88
Bid If Bid, Code Name
No Bid
Mike Carol Bruce Bob Richard 4-&J-
Dan Bob/D.C. r-\ Laurie Don Chris Jim
Please return this form to Colleen after examining the summary attached. It is
very important that you do this in order to help keep track of the RFPs
circulating.
Attachment: RFP Title Page and Statement of Work
NOTE: The attached is a synopsis of the proposed work. If the work is of
potential interest, please request entire RFP from Colleen.
RFP f: 2886-1
RFP Title:EPRI: Feedwater
Date Proposal Due 02/12/88
Bid
No Bid
RFP ROUTING
Heater Life Cycle Advisor
If Bid, Code Name
Mike Carol Bruce Bob Richard
Dan Bob/D.C. Laurie Don Chris Jim
Please return this form to Colleen after examining the summary attached. It is
very important that you do this in order to help keep track of the RFPs
circulating.
Attachment: RFP Title Page and Statement of Work
NOTE: The attached is a synopsis of the proposed work. If the work is of
potential interest, please request entire RFP from Colleen.
-------
RFP ROUTING
RFP #: PSC#048802
RFP TitleiMD PUBLIC SERVICE COMM: Management Audit of Choptank Electric
Cooperative, Inc.
Date Proposal Due 01/08/88
Bid If Bid, Code Name
No Bid
Mike Carol y Bruce Bob Richard
Dan v- Bob/D.C. Laurie Don Chris Jim
Please return this form to Colleen after examining the summary attached. It is
very Important that you do this in order to help keep track of the RFPs
circulating.
Attachment: RFP Title Page and Statement of Work
NOTE: The attached is a synopsis of the proposed work. If the vork is of
potential interest, please request entire RFP from Colleen.
RFP #: A8-0338
RFP Title:AZ DEPT OF REVENUE:
Date Proposal Due 12/23/87
Bid
No Bid
RFP ROUTING
Capitalization Rate Study
If Bid, Code Name
Mike Carol Bruce Bob Richard
Dan Bob/D.C. Laurie Don Chris Jim
Please return this form to Colleen after examining the summary attached. It is
very important that you do this in order to help keep track of the RFPs
circulating.
Attachment: RFP Title Page and Statement of Work
NOTE: The attached is a synopsis of the proposed vork. If the vork is of
potential interest, please request entire RFP from Colleen.
ft. rwuj. c/ul^
-------
3.4 GENERATION OP REGIONAL ACTIVITIES AND CROP MIXES
Yield-acreage response relationships observed over time in each region are used
to develop a mix of crop activities from the FEDS budgets for these regions.
This procedure is used to insure that the representative farms provide an
economically and technically realistic portrayal of producer's behavioral
response. These relationships are derived econometrically from historical
patterns on relative crop acreage (percentage of a unit land area devoted to
each crop). Each crop mix implies a given cost structure and corresponding
yield. Since yields are expected to change with crop mix, these yield
adjustments are predicted by the yield-acreage response functions.
Specifically, the yield of a crop can vary, not only in response to the acreage
planted of that crop, but also the acreage planted of other crops.
The rationale and procedure used to derive econometrically the yield-acreage
relationships is described in Adams, Hamilton and HcCarl. In general, however,
the procedure involves estimating a system of yield equations expressed as a
function of crop acreages ("own" and "other") and other variables based on
historical data from 1970 to 1983. These relationships were then "normalized"
to a 1983 base value. This gave rise to a series of annual crop mix-yield
adjustment ratios. The ratios are used in combination with the FEDS crop
budgets for each crop to generate a set of farm level activities that portray
various cropping alternatives for each state representative farm.
The alternative to explicitly accounting for these relationships would have been
to simply use the FEDS budgets and a given annual crop mix to define crop
alternatives (activities) for each representative farm. Vhile analytically
simpler than trying to account for crop mix-yield changes, this was not done
because it ignores potentially important yield response information.
Specifically, since expected profit is assumed to be a primary factor in the
producers planting decision (acreage and crop mix) and since these acreage and
crop mix decisions can affect yields, it is important to account for these
responses in the modeling of farm level production. If the yield response
relationship is ignored, the acreage response arising from changes in expected
profit would likely be overstated. This would then bias subsequent analyses
3-6
-------
Table 3-1
Primary Commodities Included in the Economic Model
Commodity
Units
Field Crop Commodities
Cotton
Corn
Soybeans
Wheat
Sorghum
Oats
Barley
Rice
Sugar cane
Sugar beets
Silage
Hay
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
bales
bushels
bushels
bushels
bushels
bushels
bushels
cvt
tons
tons
tons
tons
Livestock Commodities
Milk 1000 cvt
Culled dairy cows 1000 head
Culled dairy calves 1000 head
Culled beef covs 1000 head
Live heifers 1000 cvt
Live calves 1000 cvt
Non-fed beef available for slaughter 1000 cvt
Fed beef available for slaughter 1000 cvt
Calves available for slaughter 1000 cvt
Feeder pigs 1000 cvt
Hogs available for slaughter 1000 cvt
3-7
-------
where changes in relative crop yields are used to portray the effects of
alternative environmental levels, such as stratospheric ozone.
3.5 THE SECTOR MODEL
The producer level responses generated through the farm level models are
interfaced with the macro component of the sector model to obtain a measure of
social benefits. Consistent aggregation is achieved by building on the above
micro conditions, which allows the aggregate and micro processes to be linked.
The macro model features constant elasticity of demand relationships for the
outputs (commodities) of the micro models. The elasticities vary with end use
and across domestic and export markets. Given the long-run nature of the model,
export demand assumes major importance. Export demand elasticities for corn,
sorghum, wheat, soybeans, soybean oil and soybean meal are derived from the USDA
Forecast Support Group. The export elasticities vary, from -.18 for cotton, to
-.82 and -.80 for soybeans (whole) and sorghum, respectively. Assuming supply
and demand functions which are integrable and independent of sector activity,
first order conditions are then achieved in macro model specification. The
objective function of this specification is:
Maximize n = I gj(Z.) - I e.(X.) - L CmYm
i11 j J mmm
where it is the sum of ordinary consumers' and producers' surplus and the
integrals are evaluated from zero to Z^*, the amount of the i1'1 commodity
produced and sold to consumers; and from zero to Xj*> the amount of the j1*1
factor used. The parameters are as follows:
g^(Z^) is the area under the demand function for the i^ product;
e^(Xj) is the area under the supply function for the j^ factor;
C is the miscellaneous cost of production,
m
3-8
-------
subject to a set of technical and behavioral constraints. Given the micro and
macro structure of a model, the sector model solution then simulates a long-run,
perfectly competitive equilibrium. The full empirical detail of the objective
function is provided in Appendix A paper by Adams, Hamilton and McCarl.
Following Samuelson (1952), the objective function II may be interpreted as a
measure of ordinary consumers' and producers' surplus (quasi-rents) or net
social benefit. Analytically, this is defined as the area between the demand
and supply curves to the left of their intersection. The demand functions are
specified at the national level, as are aggregate production responses. Thus,
the solution of the sectoral model provides objective function values at the
national level.
The linking of the detailed producer behavioral model with a macro model
measured in (changes in) consumers' and producers' surplus provides a useful
policy model. Justification for the use of economic surplus in policy analysis
is well documented in the literature (Willig, 1976; Just et al., 1981) and is
particularly relevant to agricultural uses where aggregate distributional
consequences are of concern. By imposing alternative environmental assumptions
(e.g., reduced crop yields due to enhanced UV-B radiation flux) on farm level
behavior, as manifested in yields predicted by available response data, changes
in production and consumption, and ultimately economic surplus may be measured.
Comparisons of changes in consumer and producer surplus between the alternative
environmental states and current ambient concentration indicates the benefits
for these alternative ozone levels.
3.6 SOLUTION PROCEDURE AND DATA SUMMARY
The sector model is solved under a mix of demand curves (constant elasticity,
stepped demand, displaying a range of elasticities) using the MINOS software
package (Murtaugh and Saunders). A schematic of the model structure (in a two
region example) is presented in Figure 3-1 of Adams, Hamilton and McCarl. The
basic data utilized in implementing the model are United States Department of
Agriculture Firm Enterprise Data System (FEDS) budgets. The United States is
disaggregated into ten regions consisting of the 48 states but with the Corn
3-9
-------
Belt (Iowa, Indiana, Ohio, Illinois, Missouri), California and Texas
disaggregated into 22 subregions. This results in a total of 63 production
regions. The primary crop coverage (commodities) were presented in Table 3-1.
Table 3-2 summarizes the secondary commodities arising from these primary items.
In its current version, the FEDS budgets used in the analysis were updated to
1983, using 1983 yields, acreage, and prices. The modified budgets include
transportation costs, chemicals, machinery, fuel and repairs and interest.
Labor and land availability depends on endogenous prices. The miscellaneous
production costs were altered following the procedures outlined in Fajardo,
McCarl, and Thompson (1981). (These procedures involve calculating
miscellaneous costs so that the miscellaneous costs exactly equaled the
difference between the value of production and the cost of the endogenously
priced inputs within the model.)
Finally, demand levels for products and the supply prices are drawn from 1983
agricultural statistics. As discussed earlier, elasticities are from USDA, and
vary according to end use, for both domestic and export markets.
3.7 INTERNATIONAL TRADE COMPONENT
One feature of the sector model of interest here is the export component.
Specifically, since a large fraction of many U.S. commodities enter world trade,
any economic model of U.S. agriculture must contain a representation of this
world demand for these commodities. Eight primary commodities are assumed to
enter into world trade: cotton, corn, soybeans, wheat, sorghum, rice, barley and
oats. To reflect their respective demand situations, constant elasticity of
demand functions are assumed. As a part of the model solution, equilibrium
prices and quantities for these eight commodities are derived. This allows some
demonstration of the effects of imposed environmental change (in the U.S.) on
not only domestic producers and consumers but also foreign consumers (and
indirectly, foreign producers). Specifically, consumers' surplus changes
arising from foreign consumption is measured in the model solution. As a
result, the transboundary effects of environmental policies in one country (the
U.S.) can be observed in an aggregate sense by noting corresponding changes in
foreign consumption and welfare.
3-10
-------
Table 3-2
Secondary Commodities in the Economic Model
Commodity Units
Soybean meal
1000
lbs.
Soybean oil
1000
cvt.
Poultry feed
1000
lbs.
Feed grains
1000
lbs.
Protein supplement dairy feed
1000
lbs.
High protein svine feed
1000
lbs.
Lov protein svine feed
1000
lbs.
Veal
1000
cvt.
Non-fed beef
1000
cvt.
Fed beef
1000
cvt.
Pork
1000
cvt.
3-11
-------
It should be noted that the model assumes no change in conditions abroad. If
increased UV-B reduce yields worldwide, the U.S. reductions could have even more
serious implications than modeled herein. It should also be noted that
individual, country by country, welfare changes are not measured with this
economic model. To break down the aggregate changes in foreign consumption of
U.S. commodities would require demand relationships (elasticities) for each
importing country. In addition, specific import quantities for each country are
needed. Such data are difficult to obtain for many countries. Thus, the model
analysis of economic effects of stratospheric modifications on importing
countries will be limited to aggregate effects.
3-12
-------
4.0 ANALYSIS PLAN AND RESULTS
4.1 ANALYSIS PLAN
The analysis plan was developed to:
1. provide generalizable functions relating changes in stratospheric ozone
depletion and economic loss,
2. examine the importance of the central and upper estimates on the UV-B
soybean dose response relationship,
3. examine the importance of the "whqt if" assumption concerning UV-B
impacts on corn and wheat, and
4. illustrate how the results can be applied to alternative scenarios of
changes in stratospheric ozone depletion through time.
Five analyses undertaken to meet the first 3 goals are discussed immediately
below and summarized in Table 4-1. Application of the results to meet goal 4 is
presented in Section 4.2.
Analysis 1: Tropospheric Ozone Effects Only - Economic Response Function. To
account for changes in agricultural yields, and therefore economic values, due
to changes in tropospheric ozone arising from depletion of the stratospheric
ozone column, a set of nine tropospheric ozone situations are evaluated, from a
25 percent reduction to a 40 percent increase. This range of tropospheric ozone
increases would correspond to depletions in stratospheric ozone of up to 47
percent. For each percent change in tropospheric ozone, changes in crop yields
were calculated with the NCLAN equations and changes in economic surplus
generated with a run of the economic model. The pairs of economic surplus-
tropospheric ozone points generated from these model runs provide an economic
"response function" relating changes in economic value to changes in
tropospheric ozone. As illustrated in Section 4.2, this response function can
4-1
-------
Table 4-1
Analysis Plan
Analysis/Title
Assumption
Yield Change
Analysis 1;
Tropospheric ozone
effects only —
economic response
function
Analysis 2:
UV-B soybean effects
only-economic response
functions
Percentage changes in
tropospheric ozone and
corresponding changes
in yields of eight
crops based on NCLAN
data. Analyses measure
sensitivity of economic
model objective func-
tion to tropospheric
ozone changes.
Percentage changes in
soybean yields to
measure sensitivity of
economic model
objective function
(economic surplus)
value.
Nine model runs with
tropospheric ozone
adjustments from -25 to
+40 percent from 1983
actual ambient levels.
Seven model runs
ranging from +10 to -25
percent change in
yields from 1983 actual
levels.
ANALYSIS 3:
Combined tropospheric
ozone and UV-B soybean
effects at 152
depletion
ANALYSIS 4
UV-B soybean central
case dose response plus
NCLAN dose response
functions combined.
4.52 reduction for
soybean (.3 x 152)
plus soybean and other
yield reductions due to
tropospheric ozone.
UV-B soybean plus corn
and wheat effects at
152 depletion
ANALYSIS 5:
UV-B soybean central
case dose response plus
equal percent input for
corn and wheat.
4.5% reduction for
soybean, corn and wheat
(.3 x 152), plus
soybean and other yield
reductions due to
tropospheric ozone.
Combined tropospheric
ozone plus UV-B soy-
bean, corn and wheat
effects at 152
deplet ion
Analysis 4 plus NCLAN
tropospheric ozone dose
response function.
4.52 reduction for
soybean, corn & wheat
due to UV-B combined
with NCLAN losses, plus
soybean and other yield
reductions due to
troposphere ozone.
4-2
-------
be tied to changes in ozone with a relationship betveen stratospheric ozone and
tropospheric ozone.
Analysis 2: UV-B Soybean Effects - Only Economic Response Function. This
analysis focuses specifically on changes in economic surplus due to changes in
soybean yields, holding all other crop yields constant (at 1983 levels). Six
runs, ranging from +10 percent to -25 percent changes in soybean yields are
evaluated. Like the tropospheric ozone analyses, this provides a "response
function" for the economic model estimates of economic surplus to changes in
soybean yields. As illustrated in Section 4.2, this economic response function
can be tied to changes in stratospheric ozone through the "lower," "central" and
"upper" bound estimates relating changes in ozone depletion to soybean yields
from Chapter 2.
Analysis 3: Combined Tropospheric Ozone and UV-B Effects at a 15% Depletion
Level. The simultaneous occurrence of the UV-B soybean impact and the
tropospheric ozone effects could result in damage estimates different from the
addition of the two impacts when considered separately. However, if the results
of both impacts considered simultaneously in the model are not substantially
different from the additive results of Analysis 1 and Analysis 2, then resources
could be conserved in this and future applications. To test this possibility,
the model was rerun with both sources of crop loss considered simultaneously for
a 15 percent ozone depletion scenario, which could readily be compared with the
additive results. For this comparison, the "central" UV-B soybean yield
relationship is used, which for a 15 percent depletion results in a 4.5 percent
yield loss.
Analysis A: UV-B Soybean plus Corn and Wheat Effects at a 15% Depletion Level.
To test the sensitivity of the total economic estimates to assumptions on other
crops, a 15 percent depletion analysis was run assuming the corn and wheat
percent yield losses equal to soybean percent yield losses. Due to the
hypothetical nature of the dose response function assumption, resources were not
spent to develop a complete economic response function over multiple ozone
depletion levels. At 15 percent depletion, then, all crops are assumed to
experience a A.5 percent yield decline.
4-3
-------
Analysis 5: Tropospheric Ozone plus UV-B Soybean, Corn and Wheat Effects at a
15% Depletion Level. As in Analysis 3, this analysis considers the simultaneous
effects, rather than additive, of all impacts at the 15% depletion level. As in
analyses 3 and 4, only one model run is executed. For comparison sake, a 15
percent stratospheric ozone depletion is associated with a 12.5 percent increase
in tropospheric ozone.
4.2 RESULTS
Base Case
This section first reports the results of the base model — the model without
any impacts due to stratospheric ozone depletion — and compares these outputs
with 1981-83 values. The model generates prices and quantities in its optimal
solution. Since changes in crop yields due to UV-B radiation, tropospheric
ozone or other environmental changes will alter crop production and prices, it
is important that the base model provide an accurate portrayal of actual
commodity quantities and prices. Further, the validity of the economic
estimates rests on these endogeneously determined prices and quantities. That
is, changes in the objective function value between the base and the various
stratospheric ozone analyses are the estimates of the economic consequences of
changes in those ozone levels. For these reasons, the model price, quantity and
objective function values assume major importance in establishing the
credibility of the assessment.
The actual and model prices and quantities in 1981-83 for 12 primary commodities
are presented in Table 4-2. The model predicts prices within five percent of
the actual prices observed during this period. For most crops (8 of the 12) the
model price levels are equal to or slightly higher than actual. In terms of
quantities, the model results are within ten percent of actual (the exceptions
are cotton, 15 percent; rice, 20 percent; and barley, 14 percent).
A comparison of livestock products, while not reported here, displays similar
correspondence. The inclusion of a livestock component is an important feature
4-4
-------
Table 4-2
Model Prices and Quantities vs. Actual: 1981-83
Commodity
Prices
a/
Quantities b/
Model
Actual
Model
Actual
($ per
unit)
(millions)
Cotton
284.90
202.10
10.
24
11.79
Corn
2.69
2.68
6,519
6,839
Soybeans
5.66
5.65
1,875
1,915
Wheat
3.56
3.50
2,285
2,419
Sorghum
2.52
2.50
670
730
Rice
8.09
8.01
120
145
Barley
2.23
2.20
438
498
Oats
1.55
1.69
518
526
Silage
21.58
NA
47
NA
Hay
62.56
68.95
74
82
a/ Prices for all commodities are $ per bushel, except for cotton ($ per 480
pound bale), rice ($ per hundred weight), and silage and hay ($ per ton),
b/ Units all same as in footnote a/.
-------
since most of the grain and oil seed consumption in the U.S. is by livestock,
not directly by consumers. As with crops, model livestock prices fall within
three percent of actual, while quantities are generally within ten percent of
actual levels observed in 1981-83. Overall, the model prices and quantities for
these commodities capture the relative magnitudes of equilibrium prices and
quantities observed in recent years.
Tropospheric Ozone Impacts
Based upon 9 runs on the economic model with results reported in Table 4-3, the
following function was estimated for the U.S. in $1982 (with t-ratios in
parentheses).
D1 = -0.0678 * T - 0.000195 * T2 (Quadratic) R2 = .999 (1)
(-153) (-13.7)
D1 = -.07232 * T (linear) R2 = .997
(-45.9)
where:
D1 = annual change in economic surplus, in billions of 1982 dollars, due
to tropospheric ozone only. (-) = damage.
T = percent change in tropospheric ozone where 10.5 percent is
expressed as 10.5. (+) = increase in ozone.
Using the data from EPA (Table 2-3 of the report), a relationship between
stratospheric ozone and tropospheric ozone was estimated as (with t-ratios in
parentheses):
T = - .7875 * S + .004 * S2 if S > -32.6 (quadratic) R2 = .999 (2)
(-116) (14)
= -30.8 if S < -32.6 (quadratic)
4-6
-------
Table 4-3
Economic Model Response to Tropospheric Ozone Changes Only
Ozone
Analysis
Economic Surplus
Change in
Economic Surplus
($ billion)
($ billion)
Base
158.785
-25%
160.377
+1.434
+ 5%
158.447
- .309
+ 10%
158.102
- .624
+ 15%
157.744
- .949
+20%
157.370
-1.289
+25%
156.984
-1.644
+30%
156.581
-2.014
+40%
155.733
-2.797
4-7
-------
T = - .87747 * S
(-122)
= -30.8
if S > -32.6 (linear) R2 = .998
if S < -32.6 (linear)
where:
T = Percent change in tropospheric ozone (10.5 percent = 10.5)
S = Percent change in stratospheric ozone (10.5 percent = 10.5) (-) =
depletion in ozone. Note the quadratic function slightly over
predicts in the midrange, and the linear function significantly
over-predicts in the midrange.
Combining the economic function (Dl) with the estimated relationship between
stratospheric ozone and tropospheric ozone results in a "tropospheric ozone
only" loss, for a 15 percent depletion in stratospheric ozone, of $.892 billion
per year ($1982), as reported in Table 4-4.
UV-B Soybean Effects Only
Based upon 7 runs of the economic model with results in Table 4-5, a
relationship was developed between soybean yield and economic damage assuming no
tropospheric ozone impacts (t-ratios in parentheses).
D2 = 0.1068 * SOY - 0.00029 * SOY2 (quadratic) R2 = .999
(3)
(170)
(-27)
D2 = 0.1121 * SOY
(53)
(linear)
R2 = .998
where:
D2 = annual change in economic surplus, in billions of 1982 dollars,
resulting from changes in soybean yields due to UV-B. (-) =
damage.
4-8
-------
Table 4-4
Estimates of Annual Economic Damage Due to a 15X
Depletion of Stratospheric Ozone
Total Change in Percent of Loss in
Economic Surplus Producer Consumer
Case/Assumptions Estimate ($ billions 1982) Surplus Surplus
1. Trospospheric
ozone effects
only
point estimate
2. UV-B impacts
on soybeans
lower bound
point estimate
upper bound
$0,892
52
48
$0,000
$0,486
$1,255
47
53
3. Combined
tropospheric
ozone effects
plus UV-B impacts
on soybeans point estimate
$1,378 37 63
THE FOLLOWING SENSITIVITY ANALYSES ASSUME HYPOTHETICAL UV-B INDUCED YIELD
LOSSES TO CORN AND VHEAT EQUAL TO THE POINT ESTIMATE DOSE RESPONSE FUNCTION
FOR SOYBEANS.
4. UV-B impacts on
soybeans plus
equal percent
impacts on corn
and wheat point estimate $1,673 35 65
5. Combined
tropospheri c
ozone effects
plus UV-B induced
yield impacts on
soybeans, wheat
and corn. point estimate $2,588 18 82
4-9
-------
Table 4-5
Economic Model Response to Soybean Yield Changes
Soybean
Analysis
Economic Surplus
Change in
Economic Surplus
($ billion)
($ billion)
Base
158.785
+ 10
159.801
+1.016
- 5
158.243
- .542
-10
157.683
-1.102
-15
157.112
-1.673
-20
156.544
-2.241
-25
155.923
-2.862
4-10
-------
SOY = percent change in soybean yields due to UV-B. (-) = decreased
yield.
To convert the soybean damage function to stratospheric ozone depletion, the
following dose response relationships were based upon the Teramura results, as
discussed above, where S again refers to the percent change in stratospheric
ozone.
SOY = 0.30 * S (point estimate)
SOY = 0.76 * S (upper bound) (4)
SOY = 0.00 * S (lower bound)
Combining the soybean economic equation and the above dose response functions
results in the following quadratic functions:
D2 = 0.03204 * S - 0.000087 * S2 (point estimate)
D2 = 0.08116 * S - 0.000220 * S2 (upper bound) (5)
D2 = 0.0 * S (lower bound)
The UV-B soybean only impact for a 15 percent depletion is estimated to be $.486
billion per year ($1982) based upon the point estimate of crop loss, and $1,255
billion per year using the upper bound estimate of crop loss.
Analyses 3, A, and 5
The simultaneous occurrence of the UV-B soybean impact and the tropospheric
ozone effect could result in damage estimates different from the addition of the
two impacts when considered separately. At the 15 percent depletion level, the
addition of results from analyses D1 + D2 yields an estimated damages of $1,378
billion per year ($1982). The economic model rerun with both sources of crop
loss considered simultaneously for a 15 percent ozone depletion scenario results
in a damage estimate less than 5 percent different from the corresponding
4-11
-------
additive estimates. Therefore, the addition of the two function serves to
adequately represent total estimated damages.
To test the sensitivity of the UV-B analysis to the incorporation of other
crops, the economic model was run for a 15 percent stratospheric depletion case
with the assumed soybean, corn and wheat yield losses. Use of results from this
assessment must be caveated as being based upon hypothetical assumptions for
corn and wheat that are not well established in the literature. The estimated
loss of $1,673 billion per year ($1982) is about 3.4 times the soybean only
loss.
A final run of the economic model considered the hypothetical case of the
combined occurrence of tropospheric ozone losses, UV-B soybean losses using the
point estimate dose response function, and the assumed corn and wheat UV-B
losses with the soybean dose response function. Use of results from this
assessment must be caveated as being based upon hypothetical assumptions for
corn and wheat that are not well established in the literature. The estimated
total damages for the 15 percent ozone depletion case are $2,588 billion per
year ($1982).
Table 4-4 illustrates both the potential magnitude of damage and the
significance of the uncertainties in the analysis. First one notes that
potential tropospheric ozone impacts may be a very significant source of damage,
exceeding the UV-B soybean losses. This suggests more information is needed to
solidify the stratospheric to tropospheric ozone links. Similarly, the measurem
of UV-B induced damage for crops other than soybeans appears to exceed the
current uncertainty in the soybean dose response functions. More work is needed
to solidify the understanding of the soybean response to UV-B and to improve
understanding of the relation for other crops.
Other findings of interest concern the division of damages between producers and
consumers as reported in Table 4-4. Both producers and consumers share in the
losses from stratospheric ozone assumed in these analyses. The bulk of the
costs are borne by consumers as a result of commodity price increases. The
existence of producer losses from reduced yields in the face of generally
inelastic demands for model commodities is attributable to the presence of
4-12
-------
secondary commodities in the analysis. Producer losses occur through the
livestock sector, where reductions in feed grains due to tropospheric ozone
increase the costs of livestock production. Finally, while not reported in
Table 4-4 the consumer losses are felt primarily in the export markets, where
losses in foreign markets as a percentage of total consumer losses range from a
minimum of 51 percent for Case 5 up to 79 percent for Analysis 2.
Application to Stratospheric Ozone Depletion Scenarios
To illustrate how the results may be used, the quadratic versions of Equations
1, 2, and 5 are applied to the "lowest," "medium" and "highest" baselines
scenarios from Table 2-1. The central and high cases refer to the use of the
central and high UV-B effects on soybean yields. The results are reported in
Tables 4-6 through 4-8.
Again, it must be stressed the estimates in this analysis are conditional upon
the assumptions employed. The most important assumptions include the
relationship between changes in tropospheric ozone and stratospheric ozone; the
interpretation of the limited evidence concerning UV-B soybean dose response,
and the assumed stratospheric ozone depletion UV-B relationship assumed in the
Teramura work; the sensitivity assumptions concerning the UV-B impacts on other
crops; and the assumption of unchanging agricultural technology, behavior and
prices into the future.
4-13
-------
Table 4-6
Potential Bconoaic Surplus Changes in the Agricultural Sector
*
Due to Changes in Stratospheric Osone — Lowest Scenario
($ Billions 1982)
Tear
% Change in
Stratospheric
Osone
% Change in
Tropospheric
Osone
Economic
Iapact of
Tropospheric
Osone
Econoaic Iapact of UV-B
Soybeans
Daaage to
Bconoaic Iapact of UV-B Daaage to
Soybeans Plus Tropospheric Osone
lapacts
Central
Case
Bigh
Case
Central
Case
High
Case
1985
0.0
0.000
0.000
0.000
0.000
0.000
0.000
1990
-0.2
0.158
-0.010
-0.006
-0.016
-0.017
-0.026
1995
-0.4
0.316
-0.021
-0.012
-0.032
-0.034
-0.053
2000
-0.5
0.395
-0.026
-0.016
-0.040
-0.042
-0.067
2005
-0.5
0.395
-0.026
-0.016
-0.040
-0.042
-0.067
2010
-0.6
0.474
-0.032
-0.019
-0.048
-0.051
-0.080
2015
-0.6
0.474
-0.032
-0.019
-0.048
-0.051
-0.080
2020
-0.6
0.474
-0.032
-0.019
-0.048
-0.051
-0.080
2025
-0 6
0.474
-0.032
-0 019
-0.048
-0.051
-0.080
2030
-0 5
0.395
-0.026
-0.016
-0.040
-0.042
-0.067
2035
-0.5
0.395
-0.026
-0.016
-0.040
-0.042
-0.067
2040
-0.4
0.316
-0.021
-0.012
-0.032
-0.034
-0.053
2045
-0.3
0.237
-0.016
-0.009
-0.024
-0.025
-0.040
2050
-0.2
0.158
-0.010
-0.006
-0.016
-0.017
-0.026
205S
-0.1
0.079
-0.005
-0.003
-0.008
-0.008
-0.013
2060
0.1
-0.078
0.005
0.003
0.008
0.009
0.013
2065
0.3
-0.235
0.016
0.010
0.024
0.026
0 040
2070
0.4
-0.314
0.021
0.013
0.032
0.034
0.054
2075
0.6
-0.471
0.032
0.019
0.049
0.051
0.081
2080
0.8
-0.627
0.042
0.026
0.065
0.068
0.108
2085
1.0
-0.783
0.053
0.032
0.081
0.085
0.134
2090
1.2
-0.939
0.064
0.038
0.098
0.102
0.161
2095
i . 4
-1.094
0.074
0.045
0.114
0.119
0.188
2100
1.6
-1.249
0.084
0.051
0.130
0.136
0.215
' Values (or representative years through tine are based upon scenarios in Table 2-1. Estimates are not converted to present
values. (-1 » damages; (+) = Benefits. Central Case « Central estinate of UV-B effect on soybean yields. High case • High
estimate of UV-B effects on soybean yields.
-------
TECHNICAL PAPERS
JAPCA S& 938-943 (1985)
An Assessment of the Economic Effects of Ozone on
U.S. Agriculture
RJ1 Adams and SJL Hamflton
Department of Agricultural and Resource Economics
Oregon State University
Corvaltis, Oregon
B. A. McCart
Department of Agriarftural Economics
Texas AAM University
College Station, Texas
Past attempts to measure the economic consequences of ozone on agriculture have been
based on limited plant science Information. This paper reports on an economic aisenmentof
ozone on U.S. agriculture using recent crop response data from the National Clrop Loss
Assessment Network (NCLAN). The results are derived from a US. •gricuftural sector model
that Includes major crop and livestock production as wefl as domestic consumption, Ivestock
feetfng and export uses. The economic effects of four hypothetical ambient ozone levels are
Investigated. The analysis Indicates that the benefits to society of moderate (25%) ozone
reductions are approximately $1.7 bODon. A 25% Increase In ozone poButfon results In costs
(negative benefits) of $2.1 ballon. These estimates do not reflect compHance costs of
achieving the ozone changes and hence are not net benefits.
The harmful effects of ozone and other
air pollutants on vegetation have been
documented for at least 35 years.1'* At-
tempts to assess the monetary impact
of these effects soon followed the recog-
nition of an ozone problem.4 Until re-
cently however, assessments of the
benefits of ozone control to agriculture
have been l>ased on only sparse or eveq
contradictory biological-information. „
>The Rational Crop Loss Assessment
Network (fyCLAN) has-improved the
state of knowledge in the area of crop
response to air pollutants, specifically
ozone.5 The data from NCLAN are in-
tended to form the basis for economic
assessments of the national conse-
quences of ozone on agriculture. In
turn, such economic assessments are
intended to help perform cost-benefit
analyses of various ozone control regu-
lations.
This paper reports on the prelimi-
nary NCLAN national economic as-
sessment of ozone effects on agricul-
ture. The analysis and underlying-data
and results represent the collective bio-
logical, meteorological and economic
knowledge gained from the NCLAN
program through 1983. The results are
derived from an. economic analysis
¦based on a U.S. agricultural sector
model adapted from Chattin et al.6
The direct effects of ozone on six major
Copynffet 188$-Air Pollution Cootrol Aoocatioo
crops (com, soybeans, wheat, cotton,
grain sorghum, and barley) are evaluat-
ed using NCLAN data. In addition, the
model develops estimates of the indi-
rect effects on livestock production and
other crops. The analysis also tests the
sensitivity of the benefit estimates to
varying yield forecasts reflecting
sources of uncertainty within the
NCLAN response data and the as-
sumptions used in the economic analy-
sis. Hie latter results can be used to
place bounds on the expected economic
effects as well as to provide insight on
the importance of improved response
information in subsequent benefit as-
sessments of ozone effects on agricul-
ture.
The Model
Hie analysis and methodology are
conceptually similar to the numerous
induced change analyses found in the
agricultural economics literature. Spe-
cifically, a spatial equilibrium model
formulated as a mathematical pro-
gramming problem is used.7 This gen-
eral methodology has been recently ap-
plied in regional air pollution assess-
ments.*"11 The current analysis
/Table L Commodities included in the NCLAN economic model.
Primary commodities
Secondary commodities .
Held crop eommoditiei
Soybean meal
Soybean oil
Feedgrains
Protein supplement deity feed
High protein livestock feed
High protein swine feed
Low protein swine feed
Cotton
Corn
Soybeans
Wheat
Sorghum
Oats
Barley
Rice
Silage
Hay
Livestock commodities
Poultry
Veal
Nonfedbeef
Fed beef
Pork
Milk
Culled dairy cows
Culled dairy calves
Culled beef cows
Live heifers
Live calves
Nonfed beef available for slaughter
Fed beef available for slaughter
Calves available for slaughter
Feeder pigs
Hogs available for slaughter
-------
• MW• ••««••• >MWW%< kJki
Source: Heck et al. (Tables I and 111)", and Heck et aL (Table 2).17 Response functions
incorporating 1983 data (the pooled soybean models) are from W. W. Cure (personal
communication).
Crop/cultivar
Parameters
Degrees
of
freedom
Residual
mean
squares
Individual/pooled cultivar**
Barley
1.988
0.205
4.278
21
0.0248
(Paco cv.)
(0.051)® (0.669)
(17.15)
Com
0.158
3£30
(2 pooled cultivars)4
Cotton
3686
0.112
2.577
21
52171
(Stonev31e-213 cv.)*
(140)
(0.004)
(0.416)
Cotton
0.226
L445
(2 pooled cultivars)'
Grain Sorghum
8137
0.296
2^17
17
195875
(Dekalb-29 cv.)
(218)
(0.019)
(1.229)
Soybeans (all data)
0.165
1.303
(16 pooled data sets)*
Wheat, spring
4.480
0.186
3.200
(4 pooled cuhivars)h
(0.200)
(0.040)
(1.860)
Wheat, winter
0.146
2.235
(2 pooled cultivars)1
Non-homogeneous adtiuarti
Com
12533
0.155
3.091
17
774833
(Pioneer-3780 cv.)
(323)
(0.004)
(0.461)
Cotton
5351
0.092
2.530
21
422162
(Acala SJ-2 cv.)
(310)
(0.005)
(0.731)
Soybeans
5593
0.128
0.872
43
91452
(Davis cv.-1981)
(863)
(0.019)
(0.284)
Wheat, winter
5479
0.113
1.633
17
91726
(Roland cv.)
(312)
(0.005)
(0.288)
* The general form of the Weibull model is:
Y=aexp[(XM«l + «
where Y is yield and X is the ozone dose in seasonal seven-hour per day mean concentra-
tion in ppm. The parameters are a, the hypothetical maximum yield at zero ozone
concentration; c, the ozone concentration when yield is 0.37; c, a dimenaionless shape
parameter (e-g., c = 1 is the exponential loss function and larger c values imply less
response).
"This set of crop/cultivar responses is used to generate the yield adjustments in the
principal ozone analysis. The pooled responses in this set use all statistically homoge-
neous cultivars.
' Values in parentheses are standard errors.
1 Includes Pioneer 3780 and PAG 397 cultivars. -
• StonevQle-213 is 'a southern, nonirrigated cottop cultivar used in the axlalyses to
¦portray cotton response in the south and southeast production areas. ¦
Includes Acala SJ-2, -irrigated and dry cultivar experiments. Used to portray cotton
response in the southwest
| Includes Amsoy, Corsoy 1980, 1983; Davis 1982, 1983; Essex, Forrest 1982, 1983,
irrigated and dry; Hodgson and Williams 1981, 1982, irrigated and dry and 1983,
irrigated and dry cultivar experiments.
1 Includes Abe and Arthur 71 cultivars.
1 Includes Blueboy II; Coker 47-27; Holly and Oasis cultivars.
I These "extreme" cultivars are the basis for the sensitivity analysis. They project
greater yield adjustments than the pooled response for the same crop.
features a model structured at the na-
tional level, calibrated to represent the
U.S. agriculture sector in 1980.
The model represents production
and consumption of numerous agricul-
tural commodities. Processing of agri-
cultural products into secondary com-
modities is also included in the model,
rhe set of included primary and sec-
ondary commodities is listed in Table I.
The production and consumption sec-
tors are assumed to be made up of a
large number of individuals, each of
whom operates under competitive mar-
ket conditions. This leads to a model
which nwrnmirffl the area under the
demand curves less the areas under the
supply curves. The assumptions and
analytics of this methodology are dis-
cussed in more detail in Adams et al.li
and McCarl and Spree n.13
The model actually consists of two
components, a set of micro or farm lev-
el models integrated with a national
(sector) level model. Producer level be-
havior is estimated for a series of repre-
physical and economic environment of
agricultural producers in each of 55
production regions in the model. The
farm level supply response generated
from the farm models are linked to na-
tional demand through the sector mod-
el objective function which features de-
mand relationships for various market
outlets for the included commodities.
The model simulates a long-run, per-
fectly competitive equilibrium, as re-
flected in 1980 economic and environ-
mental parameters. Following Samuel-
son,14 the value of the objective
function, expressed in 1980 dollars,
may be interpreted as a measure of eco-
nomic surplus* (sum of ordinary consu-
mers' and producers' surplus) or net
social benefit.
Model Use
The ozone assessment involves ex-
amination of the alterations induced
by changes in ozone from a base solu-
tion of the economic modeL Once the
base-case solution is established, the
model is rerun using alternative yield
adjustments (modified from actual
1980 yields) to simulate different hypo-
thetical ambient ozone levels in each
production area. The yield adjust-
ments are derived from the NCLAN
ozone-yield response functions. In this
analysis four alternative ozone levels
are evaluated: a 10%, 25%, and 40% re-
duction in ambient ozone concentra-
tions and a 25% increase in ozone. This
range of levels is intended to bound
possible changes arising from alterna-
tive secondary standards.
Comparing the base model solution
with model solutions arising from dif-
ferent ozone assumptions then indi-
cates the effects of ozone, as yield
changes generated by alternative ozone
concentrations trigger changes in pro-
duction and consumption and ulti-
mately economic surplus. Differences
in economic surplus between the cur-
rent (base) situation and the alterna-
tive ozone concentrations are the esti-
mates of benefits for these alternative
levels. Differences in economic surplus
arising from varying response data can
also form the basis of a sensitivity anal-
ysis of underlying response relation-
ships.
The Data: Sources and Assumptions
Implementing the economic model
requires biological and meterological
information. This section discusses the
sources of these data and assumptions
inherent in their use.
Tb» um of tcooociK surplus fti a measure of >ocu) welfare
to evaluatiaf policy changes ts well documented id the
economic literature1**w and a particularly relevant to agri*
cultural analyse* where dalnbutKwal consequences are of
coocern.
Sentpmhpr V/Altimo Nn o
Q40
-------
iable ill. Actual ana muuei isou prices turn qimuuues. iur i»uuar> aops mm
commodities.
1980 prices* 1980 quantities
($/unit)» (million units)
Commodity Model Actual ~ Model Actual
Cotton
366.72
358.00
17.45
15.65
Corn
3.25
3.11
7,339.85
6,645.84
Soybeans
7.74
7.57
1,778.07
1.792.06
Wheat
3.71
3J1
2^74.31
Sorghum
3.00
2.94
700.88
579.20
Rice
12.79
12.80
164.78
146.15
Barley
2.91
245
335.50
360.96
Oats
1.93
1.79
472.91
458J26
Silage
19.46
NA
91.24
110.97
Ray
70.90
71.00
141.58
131.03
Soybean meal
0.11
0.11
46,180.80
50.624.00
Soybean oil
0.24
0.23
10.755.81
11,270.00
• Units are as follows: 500 pound bales for cotton; bushels for com, soybeans, wheat,
barley, oats, and sorghum; hundredweight for rice; tons for hay and silage; pounds for
soybean meal and ofl.
Plant Response
The NCLAN program has developed
a base of plant science information on the
response of crops to ozone as reported
in this journal and elsewhere.17-19 The
date used here cover four yean of crop
experiments (1980-83), conducted on
multiple crops and cultivan at five
sites in the U.S. The summary publica-
tion by Heck et aL19 is the source of the
individual crop response functions
used in this economic analysis. All
functions are reported in the Weibull
form. The Weibull is a defensible char-
acterization of the relationship be-
tween yield and dose, based upon sta-
tistical and plant .science consider-
ations.17* 20 For some crops, response
functions are estimated from both
pooled and individual cultivar data.
These response functions are summa-
rized in Table ILB
Note that separate response func-
tions are available for various eultivars
of the same crop in addition to multiple
sets of pooled responses. Thus, some
judgments must be made as to which to
include in the economic analysis. We
have used common or pooled responses
for soybeans, corn, cotton and wheat, in
combinations with the single cultivar
dose-yield functions available for bar-
ley and sorghum. This set of response
function* is presented in Table II un-
der the first subheading. For crops for
which NCLAN response data are lack-
ing (oats, rice, sugar cpne, and sugar
beets) the respective yields are held
constant across the ozone evaluations.0
Ozone Data
Two types of air quality data are
needed to use the NCLAN response
functions. First, some measures of ac-
tual ambient ozone concentrations in
rural areas are necessary. These ozone
data must be presented in a dose mea-
sure consistent with that used in the
NCLAN experiments (seasonal seven-
hour concentration) at a geographical
scale sufficient to be compatible with
the 55 productioh regions in the sec-
toral modeL Ideally, actual monitored
data for all agricultural production ar-
eas in the economic model should be
used. Unfortunately, such complete
data do not exist. However, a surrogate
'Additional pooled ooobcnatioae of cakfam are abo ev*3-
able.Tbeae vert ua*d in etnaitivityanalywe not reported
btn. The full beting of thru crop* and wtenngifand
So Adam* ft ot (Ttbli
•An «u«mpt I* mod* to modd powiblt bay yWd cfaangea,
given that hay fa an fanportani Input far &v«todipndi»e-
boo. By otiai a wmpti yield tmynrm itpnntuled by
the ai&ple average of reeponacaaf the as NC1AN crepe, a
aenahHty analyate of the amnrwn ir eertmataa to hay yWd
changea b performed. Ahernitm moca of the available
lopotttt data are alio ueed to teat the aenstmty of the
among ic ftwnatm to the mponic data, reported id Ad-
am* rt olu SpoaftcaDy, the cuhirare Uetod cinder the eee-
ood eobbeadisg lo Table II era taed m conbimtwc with
the individual crop cuhnran in the Tint category to an
alternative (and more rcspoohrc) »*t of jrMd adjustments.
data set has been estimated by J. Rea-
gan, U.S. EPA, Research Triangle
Park. County-level ambient ozone con-
centrations are derived by a spatial in-
terpolation of the EPA Storage and Re-
trieval of Aerometric Data (SAROAD)
system data. The procedures, results
and limitations associated with this
data set are discussed elsewhere0."
The second piece of ozone informa-
tion needed to develop yield adjust-
ments are the hypothetical ozone levels
to be examined in the economic analy-
sis. For the purposes of this assessment
effort, these are 10,25, and 40% reduc-
tions in ozone from ambient and a 25%
increase in ambient ozone. Each hypo-
thetical ambient ozone level is then
used in conjunction with the Weibull
response functions for each crop to pre-
dict the yield changes associated with
these changes in ambient ozone con-
centrations. The 10 and 25% adjust-
ments in ambient ozone are considered
plausible chfcnges, in that changes of
this magnitude are encompassed in the
temporal .variability displayed in the
five-year AEone data set "
The tesponse functions, in combina-
tion with the respective ambient ozone
assumptions, result in a projected yield
adjustment (from current yields) for
each crop in each region. For a given
ozone level, up to 55 yield adjustments
may be required for each crop (one for
each of the 55 regions in the model).
While not all crops are grown in all
regions, there are still a large number of
individual yield adjustments to be cal-
culated to implement the assessment
The entire set of adjustments are re-
ported in Adams et al.13 but some aver-
age values can serve to define the mag-
Interpolation tHaean monitoring ajtaaa baaed on a"Kn-
gtag* procedure that impttdUy free greater weight to
ntawt aooitoring cites, ft ehouid be ootod that *00*
atatea have as few aa three nonitonng eitea. Even where
ritee are nor* mrmrroua (over 100 in California, for etas-
pi*), noat tend to be In urban or euburban area*. Tbua. this
gengnphral inurpotalioo of the national SAROAD data
•et n wubjeet to uncertainties.
nitude of yield adjustments for each
ozone case. Specifically, for the 10, 25
and 40% changes in ozone, average
yields for the six crops change by 1.25,
2.9, and 4.4% respectively. Within
these averages, rather dramatic differ-
ences occur across crops, given differ-
ential sensitivity. The mmimum yield
changes observed within the above
ozone changes are 3.8,12.6 and 17.9%.
Results and Implications
This section first reports the results
of the base model and compares these
outputs with 1980 values. The results
and implications of the major ozone
yield evaluations (as measured against
the base situation) are then presented.
Additional information, such as region-
al effects and price changes by com-
modity, is contained in Adams et ol.11
Base Model Results
The model generates prices and
quantities in its optimal solution. Since
changes in crop yields due to ozone will
alter crop production and prices it is
important that the base model provide
an accurate portrayal of actual com-
modity quantities and prices. Further,
the validity of the economic estimates
rests on these endogenously deter-
mined prices and quantities. (Changes
in the objective function value between
the base and the ozone analyses are the
estimates of the economic conse-
quences of changes in ozone levels.) For
these reasons, the model price, quanti-
ty and objective function values as-
sume major importance in establishing
the credibility of the assessment
The actual21 and model prices and
quantities in 1980 for 12 crop commod-
ities are presented in Table III. The
model predicts prices within five per-
cent of the actual prices observed in
1980. For most crops (11 of the 12) the
model price levels are equal to or slight-
-------
A butc A » • AWVUCU OJUU UlUUCi AWW WIU aw* win* J «*mw
Source: USDA, Statistical Reporting Service. Statistical Bulletin 552.
Product
1980 prices
($/unit)»
Model Actual
1980 quantities
(mil
ieT
Milk
1235
13.00
1,28244
1,28640
Pork
139.00
139.50
141.68
165.77
Fed beef
237-50
237.60
13840
159.36
Veal
310.30
309.50
3.66
4.11
Nonfed beef
15020
149.76
64.40
7342
* Units are hundredweight. Meat prices are average retail prices for finished meat
products.
ly higher than actual. In terms of quan-
tities, the model results are within 10%
of actual (the exceptions are sorghum,
17%, and silage, 20%). For Beven of the
12 commodities, model production ex-
ceeds actual, lite oversupply of sor-
ghum and silage depicted in the base
model is due to the presence of a
drought in the Com Belt and Southern
Plains regions in 1980 that lowered
yields which in turn led to a shift of
corn planted for gTain into silage use.
However, when the silage and sorghum
quantifies are compared with 1979 and
1981 actual levels, the model supply
projections are within ±3%.
A similar comparison of livestock
products is provided in Table IV. The
inclusion of a livestock component is an
important feature, since most of the
.grain and oO seed consumption in the
U.S. is by livestock, not directly by con-
sumers. As with crops, model livestock
prices fall within 3 percent of actual,
while quantities are again within 10
percent of actual levels observed in
1980. Overall, the model prices and
quantities for both crop and livestock
commodities appear to capture the rel-
ative jnagnitudes of equilibrium prices
and quantities observed in recent
years.
Ozone Analyses
The principal set of ozone evalua-
tions used pooled response functions
based on all statistically homogeneous
experiments (across cultivars and
years) for corn, soybean, and wheat
(spring and winter). In addition, the
single cultivar responses for barley,
sorghum, and cotton (irrigated and
dry land) are used. These cultivar re-
sponse functions, together with the
four changes in ambient ozone levels
(measured by the 1980 seasonal seven-
hour average in each of the 55 produc-
tion regions), are used to calculate the
yield adjustment coefficients with
which to modify the base model yields.
The resulting changes in objective
function values (from the base case)
then define the economic benefits or
costs to society for each ozone adjust-
ment.
The model output of greatest inter-
est is the economic surplus estimate
generated under each model solution.
These values are presented in Table V.
Also presented are the changes in bene-
fits from the base case, as well as pro-
ducer and consumer surplus compo-
nents of each total. The changes in eco-
nomic surplus represent the benefits to
society from alternative ozone levels.
The analysis indicates that improve-
ments in air quality (reductions in
ozone from 1980 ambient levels) result
in benefits to society. Specifically, the
benefits of a 10% reduction in rural
ozone are $669 million. For a 25% re-
duction in ozone, benefits are $1.71 bil-
lion, and for more extreme (40%) ozone
reduction benefits are $2.52 billion. Fi-
nally, an increase in ozone levels (air
quality degradation) by 25% produces a
cost (negative benefit) to society of ap-
proximately $2.10 billion.8
Tbi aoU t—44—-*— displ*; wyiaf wi»lU»ltj to Ba-
tumi* 1cm uomilnt
-------
ent ozone levels are plausible changes,
in that temporal variations of this mag-
nitude are observed in the seasonal av-
erages over the periods 1978-1982. The
40% reduction in ambient levels is ap-
proaching what some researchers sug-
gest to be background or natural lev-
els.12 While the attainment of such a
change may not be feasible, the bene-
fits value imputed to this level of ozone
reduction can suggest the agricultural
consequences of all anthroprogenic
ozone.
particularly with respect to the region-
al adjustment process to ozone-in-
duced supply shifts, the similarity of
th«u> estimates derived from theoreti-
cally defensible economic approaches
suggests that benefits of reduced ozone
to agriculture are of this general magni-
tude. To use such estimates in a regula-
tory setting, however, additional Infor-
mation is needed, including meteoro-
logical models to link secondary
standards (measured as one-hour max-
imums) to rural seasonal concentra-
tial sources of uncertainty that may
bias the economic estimates. These in-
clude the issue of exposure dynamics
(seven-hour seasonal mean as used in
NCLAN periods va. alternative expo-
sures), the lack of environmental inter-
actions, such as moisture stress, in the
response experiments, omission of
some crops and the extrapolation from
a limited number of cultivars of each
crop to the entire production response
for that crop. Each of these response
issues is being addressed in current
Table VL Effect of crone on regional producers' wrplui (I billion).
10% reduction
26% reduction
40% reduction
25% increase
Region
Base
Surplus
% change
Surplus
.%change
Surplus
% change
Surplus
% change
Northeast
0551
0.669
337
OS75
4JX
0.601
9.07
¦ 0.516
-6.35
Lakes
3.743
3.768
L20
3.782
L04
Z27
3£S1
-1.66
Cora Belt
7.261
7.404
L97
7.492
3.18
7.633
5.12
7.106
-2.13
Northern Plains
3-576
&670
a
&£78
a
3£99
060
3£59
-0.50
Appalachia
Z238
2*131
L44
2J64
2£T
2.402
4£3
ZJ22A
-3.22
Southeast
1.377
L409
232
1.440
4*8
L475
7.12
L274
-7.48
Delta
L34S
1.395
3.49
1.448
7.42
1.495
1031
1.224
-9.20
Southern Plains
2J540
2£49
045
2£73
L30
2J503
2.48
2.485
—2.17
Mountain
2-508
££06
a
2-570
2.47
.Z576
2.71
2.439
-2.75
Pacific
0l813
0.820
086
0.984
2L03
L059
30.26
0.614
-24.48
Total
26.015
26.337
26.807
27.271
25.122
' Less than 0.10% change.
The 25 and 40% ozone reduction ana-
lyses can also serve as a useful compari-
son with some recent national level es-
timates derived by other researchers,
given similarities to the (none levels
used in those studies. Specifically, a
study by Shriner et aL,a using 1981
NCLAN data for corn, soybeans, wheat
and peanuts estimated the 1978 costs
of ozone to be approximately $3.0 bil-
lion. This cost was evaluated against a
background or "clean air" situation de-
fined as ambient ozone concentrations
of 25 ppb. This is «milar to the ozone
levels simulated in the 40% change
evaluated in the present study. When
adjusted to 1980 dollars, the Shriner
study value is over $1 billion higher
than in this analysis. However, Shriner
et al. do not use a model of economic
behavior (rather, they multiplied com-
modity prices times predicted quantity
change) nor did they have the benefit
of the more recent NCLAN data.
A second national benefits analysis
by Kopp et at.94 uses a set of NCLAN
data to derive yield adjustments for use
in a comprehensive economic modeL In
the Kopp et al. analysis, improvement
in air quality from current ambient lev-
els to levels similar to the 25% reduc-
tion result in benefits to society of ap-
proximately $1-2 billion measured in
1978 dollars. When the Kopp etal. esti-
mates are adjusted to reflect 1980 dol-
lars and the same set of crops, the esti-
mates are within 20%. While the assess-
tions as well as information on benefits
that may accrue to other Bectors and
compliance costs of reducing ozone lev-
els.
Concluding Comments
Overall, the results of this assess-
ment indicate that the benefits of mod-
erate ozone reductions are substantial
in absolute terms (approximately $2
billion) but are a relatively small per-
centage of total agricultural value (ap-
proximately 3% of gross crop value).
The benefits of ozone reductions ac-
crue to both producers and consumers,
with about 60% of the consumer bene-
fits accruing to foreign consumers. Do-
mestic consumers benefit from slightly
lower prices of livestock products, due
to increased supplies of feedgrains and
oil seed. Producers of commodities that
enter foreign markets benefit if those
markets absorb the increases in supply.
Regionally, the major beneficiaries are
those areas with high relative levels of
ozone and ozone-sensitive crops. It
should be noted that the benefits re-
ported here do not include compliance
costs of achieving these ozone changes
and hence are not net benefits.
A number of limitations or caveats
need to be attached to these estimates.
While they are derived from a concep-
tually sound economic model and use
the best available supporting biological
and meteorological data in implement-
NCLAN research. In addition, sensi-
tivity analysis was done on these ques-
tions to assess their importance in fu-
ture economic assessments (Adams et
al.13). Also, the ozone data are based on
a limited set of EPA's SAROAD moni-
toring sites, mainly in urban and sub-
urban areas. While the spatial interpo-
lation process (Kriging) results in a
fairly dose correspondence between
predicted and actual ozone levels at a
few validation points, (e.g., the
NCLAN sites) there is a need for more
monitoring sites in rural areas. Spatial
dispersion models linking rural ambi-
ent ozone levels and federal secondary
standards are also needed in the devel-
opment of ozone scenarios for regula-
tory analyses.
The economic model, with its many
variables, parameters and the underly-
ing data used to derive these values, is a
potential source of error. For example,
the model does not explicitly consider
changes in the operation and structure
of federal farm programs that may
arise from ozone-induced supply
changes. Also, while using 1980 output
prices and input costs, the model as-
sumes a 1974 production technology.
To the extent that these relationships
affect or are affected by ozone changes,
biases will exist in these economic esti-
mates. Finally, the model is a long-run
equilibrium model and assumes that
excess supply does not exist The mod-
el was validated against historical val-
.fotrrnal of th« Air Pollution Control Association
-------
greatest when addressing changes
bounded by historical levels, rather
than quantum adjustments. The mod-
est adjustments portrayed in the ozone
analyses fall within these historical lev-
els.
Acknowledgments
Although the research described in
this article has been funded by the
United States Environmental Protec-
tion Agency through a cooperative
agreement (CR-819297-01) with Ore-
gon State University, it has not been
subjected to the Agency's required peer
and policy review and therefore does
not necessarily reflect the views of the
Agency and no official endorsement
should be inferred.
The authors gratefully acknowledge
the support of the NCLAN Research
Management Committee and scien-
tists. The paper has benefited from
comments by Don Davis, Don Holt,
Earl Swanson, Bob Taylor, Norm
Whittlesey and three anonymous re-
viewers on an earlier draft Technical
Paper No. 7589 of the Oregon State
Agricultural Experiment Station.
References
1. J.T. Middleton, J.B. Kendricks, Jr.,
H.W. Schiwalm, "Injury to herbaceous
plants by smog or air pollution," Plant
Disease Reporter 34:245 (1950).
2. HJL Brialey, W.W. Jones, "Sulfur diox-
ide fumigation of wheat with special ref-
erence to effect on yield," Plant Phy-
siol 25:666 (1950).
3. DJCA.BleasedaIe," Atmospheric pollu-
tion and plant growth,'' Nature 1(9:376
(1952).
4. 1LM. Benedict, C J. Miller, J.S. Smith,
(if Bfflnowiif ImiMft «f Air
Pollutants on Vegetation in the United
States: 1969-1971/* EPA-650/5-78-002,
Stanford Research Institute, Menlo
Park, CA, 1971.
U. Bingham, J. Miller, E. Preston, L.
Weinstein, "Assessment of crop loss
from ozone," 32:353 (1982).
6. B. Chattin, B-A. McCarl, H. Baumes,
Jr., User* Guide and Documentation
for a Partial Equilibrium Sector Model
of US. Agriculture, Agriculture Experi-
ment Station Bulletin No. 313, Purdue
University, Lafayette, IN, 1983.
7. T. Takayama, G. Judge, Spatial and
Temporal Price and Allocation Models,
North Holland Publishing Company,
Amsterdam, 1971.
& RM. Adams, TJ>. Crocker, "Dose-re-
sponse information and environmental
damage assessments: an economic per-
spective," 32:1062 (1982).
9. RM. Adams, B.A. McCarl, "Assessing
the benefits of alternative oxidant stan-
dards on agriculture; the role of re-
sponse information," J. Environ. Econ.
Management, in press, (1985).
10. R-E. Howitt, TJs. Gossard, RM. Ad-
ams, "Effects of alternative ozone levels
and response data on economic assess-
ments: the case of California crops,"
JAPCA 34:1122 (1984).
11. RP. Rowe, L.G. Chestnut, C. Miller,
. RM Adams, M Threshow, H.O. Ma-
son, RE. Howitt, J. Trijonis, Economic
Assessment of the Effect of Air Pollu-
tion in the San Joaauin Valley, draft
report to the Research Divison, Califor- '
nia Air Resources Board, Energy and
Resource Consultants, Inc., Boulder,
CO 1984.
12. R.M. Adams, S.A. Hamilton, B.A.
McCarl, The Economic Effects of
Ozone on Agriculture, U.S. Environ-
mental Protection Agency, EPA-600/3-
84-090, September. 1984.
13. BJV. McCarl, T. Spreen, Trice endoge-
nous mathematical programming as a
tool for sector analysis. Am. J. Agric.
Econ. 62:87 (1980).
14. P.A. Samuelson, "Spatial price equilib-
rium and linear programming," Am.
Econ. Review 42:283-303.
15. RD. Willig, "Consumers' surplus with-
out apology," Am. Econ. Review 66:589
(1976).
16. RE. Just, D.L. Hueth, A. Schmitz, Ap-
plied Welfare Economics and Piiblic
Policy, Prentice-Hall, New York, 1982.
17. W.W. Heck, O.C. Taylor, R.M. Adams,
J.E. Miller, L. Weinstein, National
Crop Loss Assessment Network
{NCLAN) 1982 Annual Report, report
Research Laboratory, June 1983.
18. W.W. Heck, RM Adams, W.W. Cure,
AJ5. Heagle, H.E. H^gestadt, RJ. Ko-
hut, L.W. Kress, J.OT Rawlings, O.C.
Taylor, "A reassessment of crop loss
from ozone," Environ. ScL TechnoL 17:
573A (1983).
19. W.W. Heck, W.W. Cure, J.O. Rawlings,
LJ. Zaragoza, AS. Heagle, HJS. Heg-
gestad, RJ. Kohut, lZwTKress, PJ.
Temple, "Assessing impacts of ozone on
agricultural oops: IL Crop yield func-
tions and alternative exposure statis-
tics." JAPCA 34 810 (1984).
20. JJ). Rawlings, W.W. Core, "The Wei-
bull function as a dose-response model
for studying air pollution effects," Crop
Sci- in press, (1985).
21. Agricultural Statistics, 1981, US. De-
partment of Agriculture, U.S. Govern-
ment Printing Office, Washington, DC,
1982.
22. W.W. Heck, W.W. Cure, J.O. Rawlings,
LJ. Zaragoza, AJS. Heagle, HJ5. Heg-
Rested, RJ. Kohut, L.W. Kress, PJ.
Temple, "Assessing impacts of ozone on
agricultural crops: I. Overview,"
JAPCA 34:729 (1984).
R M Adams is an associate profes-
sor of Agricultural and Resource Eco-
nomics, Oregon State University,
Corvallis, OR 97331; S. A. Hamilton
is an economist with Northwest Eco-
nomic Associates, Vancouver, WA;
and B. A. McCarl is a professor of
Agricultural Economics, Texas A&M
University, College Station, TX. This
tjwtiniwl paper was submitted for
peer review July 3,1984; the revised
manuscript was received June 27,
1985.
Seotember 1985 Volume 35 No 9
943
-------
The Benefits of Pollution Control:
The Case of Ozone and U.S. Agriculture
R. M. Adams, S. A. Hamilton, and B. A. McCarl
Tbe advene, effect* of ozooe and other kir poUutams oo crop yields are well
documented. This paper reports on an assessment of tbe benefits to agriculture arising
from reductions u »ozone pofltrtkw. Estimites are derived using recent plant
science data as input Cor a ¦p*'*' equilibrium model of U.S. agriculture. Sensitivity
of benefit estimates to biological and eeoaoaiic sources of uncertainly b also
Investigated. Results suggest that the benefits of a 25 percent reduction in ambient
ozooe ate substantial, amounting to $1.7 billion, Tbe robustness of these estimates
varies across alternative assumptions concerning response data and export markets.
Key words: agriculture, air pollution, benefits of control, regulatory efficiency, sector
morirfiitg.
The adverse effects of air pollution on vegeta-
tion, including crops, are well documented
(U.S. Environmental Protection Agency
[USEPA] 1984, Heck et aL). Ozone (Oj) is one
of the major air pollutants, accounting for over
90% of vegetation damage. While man-made
sources (primarily automobiles) contribute
about 40%-50% of total ambient ozone, the
long-range transport of ozone results in ele-
vated levels in rural areas. As a result, am-
bient pollution concentrations in important ag-
ricultural production areas are sufficiently
hig])10 reduce crop yields (USEPA 1984).
Major reductions in agricultural productiv-
ity from ozone or other air pollutants could
K. H. Atoms b an associate professor. Department of Apicul-
tunl and Resource Economics. Orcfoa State University; S. A.
HtmPtoa "a senior Nattmt Economic AukUics;
B. A. McCirt is » professor, DqwUKSt of Ajrieuhar*] Econom-
ics. Texas AAM University. Seniority of authorship is shared.
Technical Piper 7672 of tbe Oregon Ajricultmil Experiment
Srnira.
Although tbe research described in this article has been funded
by Ibe U.S. Earironaacutal Protection Agency through a Coopera-
tive Agreement (CR-8 WWT-01) with Oregon Stste Ueiwuty, it
has oat bees subjected to tbe agency's rojuind peer Mai poficy
review sad therefore does not necessarily reflect tbe views of (be
agency, and no official endorsement should be isfcrod.
The authors Ihnt Doe Dsvis. Don Hob. £u{ Swisson, Bob
Taylor, and Horn Whittlesey lor detailed reviews. Ride Freeman
and Bill Martin aho provided mnnwrms oo a (dated manuscript.
In addition, two AJAE n-yiewm provided useful suggestions on
an earlier manuscript. Tbe authors' thanks also to tbe NCLAN
research team and Eric Acsun, USEPA preyed officer for din.
guidance, and iiutraction. Any remaining etrors are the respoasi-
biOty of the auibcrs.
Review «u coordinated by Alan Randall, editor.
have a substantia] effect on societal welfare.
Such possible adverse effects were one of the
motivations behind the establishment of Sec-
ondary National Ambient Air Quality Stan-
dards (SNAAQS) as mandated by the Clean
Air Act and its amendments.1 Assessments of
the economic benefits of present or proposed
secondary air quality standards are needed to
formulate efficient regulatory actions. In as-
sessing agricultural benefits, information is re-
quired linking ambient ozone to crop yields
realized by producers. The National Crop
Loss Assessment Network (NCLAN) was es-
tablished by USEPA to provide such data for
use in economic assessments. One phase of
the NCLAN program involves multiycar-
multisite field experiments on crop-yield sen-
sitivity to ozone. Hie second phase involves
use of these data to assess the economic
benefits and costs arising from ambient ozone
changes. The results reported here are the out-
growth of that effort.
The overall purpose of this paper is to re-
port estimates of the societal benefits of alter-
native ambient ozone levels arising from the
agricultural sector. Effects on nonagriculturaJ
commodities and compliance costs of reduc-
ing ozone concentrations are not evaluated
' Tbe dean Air Act also mandates the USEPA to set air qualu Y
standards to maintain human health values. Specifically. USEPA
sets Pnmary National Ambient Air Quality Standards based on
measures of human health (morbidity and mortality).
Coovrieht 1986 American Aencultural Economics Association
-------
Adams, Hamilton, and McCcri
here; hence, these estimates are not net eco-
nomic effects. However, an attempt is made
to provide a perspective on the magnitude of
these estimates vis-d-vis compliance costs and
benefits accruing to other sectors. The specific
objectives of the research underlying this pa-
per are (a) to estimate benefits arising from the
agricultural sector under alterations in am-
bient ozone levels; and (b) to provide a per-
spective on the economic costs and benefits of
ozone control considering other sectors and
compliance costs.
In addition, the analysis investigates the
sensitivity of the benefit estimates to biolog-
ical and economic sources of uncertainty.
Methodological Considerations
Assessment of the societal costs and benefits
of ambient ozone changes arising through the
U.S. agricultural sector involves estimation of
both the physical and economic effects associ-
ated with alternative ozone concentrations.
However, relevant market situations cannot
be. directly observed since ozone concentra-
tions have not occurred over the range of in-
terest in a regulatory analysis. Consequently,
the effects of alternative ozone levels must be
simulated. Agricultural sector models are fre-
quently used to construct such marketplace
simulations, with mathematical programming
providing the simulation method (Heady
and Srivastava, McCari and Spreen). This
methodology has been applied in regional air
pollution studies (e.g., Adams, Crocker, and
Thanavibulchai; Adams and McCari).
_> A mathematical programming-based assess-
ment of ozone is performed by adjusting the
mathematical programming model to account
for ozone's input and output effects. Data per-
tinent to such adjustments are available from
the NCLAN program. However, the NCLAN
plant science and aerometric data are not ex-
haustive (not all crops and regions are cov-
ered) nor without variation. For example,
yield response estimates differ based on crop
variety and environmental stress. Thus, the
assessment results are investigated as to their
sensitivity to the range of effects revealed by
the NCLAN data. This is an extension of ear-
lier work by Adams and McCari that identified
critical biological uncertainties in bioeco-
nomic modeling of air pollution effects.
The economic parameters of the assessment
are also subject to uncertainty. Perhaps the
Pollution Control 887
most critical of these is the agricultural export
sector. Specifically, exchange rates have
undergone a major realignment, substantially
altering the export demand for U.S. products.
Agricultural exports are an important factor in
terms of the welfare derived within the ag-
ricultural sector. Consequently, an explor-
atory analysis of the effect of shifts in export
demand is also performed.
The Sector Model
The mathematical programming model is of
the long-run equilibrium type encompassing
production, processing, imports, domestic
consumption, and export of thirty mqjor crop
and livestock commodities, including feed and
feed concentrates.3 The model contains a ten-
region disaggregation of the United States
containing fifty-five subregjons. These subrc-
gions include twelve subregions in the Com
Belt (as developed by Brown and Pheasant)
and forty-three non-Corn Belt states. The base
cropping information is derived from the 1977
Federal Enterprise Data System (FEDS)
budgets for crops and livestock as reformatted
by Burton. These were tonight to a 1980
basis, using USDA state-level yields, acre-
ages, and prices. The model objective function
maximizes the area under the demand curves
less the area under the supply curves plus the
revenue under infinitely elastic demand curves
less the cost under infinitely elastic supply as
in a quadratic, spatial equilibrium model
(McCari and Spreen). However, constant
elasticity rather than constant slope curves are
used.
A complete description of the model is be-
yond the scope of this paper. Readers inter-
ested in more detail should refer to Chattin,
McCari, and Baumes; House; Baumes; or
Burton. A detailed, summation notation de-
scription of the model is also available upon
request. In addition, related conceptual mate-
rial is given in McCari and Spreen and Heady
and Srivastava. Nevertheless, a description of
how the microeconomic detail was incorpo-
rated in the sector model is in order because
we used crop mixes for activities rather than
individual crops (McCari 1982a).
The Corn Belt and the rest of the United
' The particular model used here was developed by Baumes,
improved in data specification by Buitoo, documented in Chattin,
McCari, aad EUumct. and was updated ud revived for this study.
-------
888 November 1986
States were treated differently in defining the
crop mix activities. The Com Belt model was
developed using representative farm models.
Linear programming fann-level models were
set up for twelve Corn Belt subregions by
Brown and Pheasant, using the Purdue
REPFARM system (McCarl 1982b). These
models were assumed to mimic the technical
and economic environment of producers in
each subregion. The models were solved
with five corn prices and twelve soybean-to-
corn price ratios as well as five wheat prices, a
total of300 different price ratios. The resultant
solutions were summarized to yield a series of
unique crop mixes (i.e., land use patterns)
and accompanying yields for each of the rep-
resentative farms. Specifically, when a model
was solved under a particular price ratio, its
solution contained a particular combination of
crops; for example, 45% com, 35% soybeans,
and 20% wheat, with associated yields. These
were used to generate the activities in the sec-
tor model. The relevant FEDS budgets were
multiplied by the crop mix percentages and
summed. Thus, each LP activity represents a
whole-farm, multiple-crop plan rather than a
single-crop activity.
Our use of Brown and Pheasant's results
suggested an important consideration not cov-
ered in McCarl (1982a). Crop yields were
found to be sensitive to crop mix. This is not
surprising, as an increase in one crop's acre-
age would be expected to be accompanied by
a change in the yield of other crops, as rota-
tion effects, less favorable land, and/or differ-
ent planting/harvesting conditions are encoun-
tered. To reflect such influences, yield
changes were incorporated in the cropping ac-
tivities. The FEDS yields were adjusted to ac-
count for yield difference due to crop mix.
For areas outside the Corn Belt, historical
data procedures were utilized to develop
whole farm plans. Yield adjustments reflecting
crop mix changes were also included. State-
level crop records for 1970-81 were used both
to develop representative state-level crop
mixes and to derive econometric estimates
of yield response to crop mix changes.
Specifically, yield estimates were derived us-
ing historical data on relative crop acreages
(percentage of a unit of land area devoted to
each crop) as the principal independent vari-
ables; i.e., the yield of a crop was estimated as
a function of its planted acreage, the planted
acreage of other crops, and other variables.
Activities were then formed reflecting a multi-
Amer. J. Agr. Ecoi
ple-crop budget with the FEDS budgets ap
gregated according to the crop mixes observe-
in the 1970-81 years including estimates c
yield change given the crop mix change.
The sector model provides a basis for a>
sessment of the economic effects of changes i>
rural ozone concentrations. However, befor
the model was used for such experimentation
an attempt was made at verifying the modi
base solution. This was done by comparin
model price, quantity, and acreage result
generated using 1980 production data wii
1980 actual values (table 1). The model pn
diets equilibrium crop prices within 5% or le*
of the actual 1980 prices, while livestoc
prices are within 3% of actual. The mod*
quantity results presented are generally withi-
10% of actual.
Distributional effects in regional welfare ai
another dimension of the regulatory policy is
sue. The model exhibited close correspor
dence between predicted and actual region;
acreages. (Adams, Hamilton, and McCarl pre
vide more details.) The model results ar
slightly below actual regional acreage for a'
crops because of the exclusion of minor sp<
cialty crops. Subregional crop acreage also e:
hibits dose correspondence. The accurac
with which the 1980 base model solulic
simulated the observed 1980 results was coi
sidered acceptable, and so the model wt
deemed suitable for use in the subsequei
analysis.
Data on Air Pollution Effects
The NCLAN program has generated fi\
years of plant science information on crop r<
sponse to ozone and other stresses. Heck el a
summarize NCLAN-generated dose res pom
functions for major crops. The response dai
underlying these functions were drawn froi
experiments on com, soybeans, wheat, co
ton, grain sorghum, and bariey. Aerometr
data also are required on ambient ozone coi
centrations for U.S. agricultural productic
areas. Such information is needed for u:
in the response functions to predict yie
changes simulating alternate ozone levels. Ui
fortunately, a complete set of aerometric da
does not exist. However, an interpolated da
set is available through NCLAN.
These data were incorporated in the mod
assuming that ozone imposes a neutral techn
logical change. Thus, when ozone concentr
-------
Adam*. Hamilton, and McCaH
^SV)
&°T
Pollution Control 889
Commodity/Product
1980 Prices
1980 Quantities
Model
Actual
Model
Actual
Crops
(units)
(S/unit)
(million units)
Cotton
(500 lb. bales)
366.72
358.00
17.45
15.65
Corn
(bushels)
3.25
3.11
7,339.85
6.645.84
Soybeans
(bushels)
7.74
7.57
1,778.07
1.792.06
Wheat
(bushels)
3.71
3.91
2.633.94
2.374.31
Sorghum
(bushels)
3.00
2.94
700.88
579.20
Rice
(cwt)
12.79
12.80
164.78
146.15
Bailey
(bushels)
2.91
2.85
335.50
360.96
Oats
(bushels)
1.93
1.79
472.91
458.26
SQage
(tons)
19.46
NA
91.24
110.97
Hay
(tons)
70.90
71.00
141.58
131.03
Soybean meal
(pounds)
0.11
0.11
46.180.80
50.624.00
Soybean oQ
(pounds)
0.24
0.23
10,755.81
11.270.00
livestock
Mfflc
(cwt)
12.95
13.00
1.282.24
1.286.20
Pork
(cwO
139.00
139 JO
141.68
165.77
Fed beef
(cwt)
237JO
237.60
138.20
159.36
Veal
(cwt)
310 JO
309Sd
3.66
4.11
Noo-fedbeef
(cwt)
150 M
149.76
64.40
73.22
Source: Mode) rraifti tod USDiA. 1982a.
tions are altered, the yield of each production
activity is adjusted by the predicted yield
change generated from the NCLAN response
functions. The mechanics of this adjustment
within the model framework are discussed in a
technical appendix available upon request.
The Oxone Assessment
(SNAAQS of 0.10 and 0.08 parts per million
not to be exceeded more than once per year).
The 40% ozone reduction is an extreme analy-
sis reflecting ambient concentrations equal to
or below what is generally thought to be back-
ground or natural ozone levels. These results
can suggest the maximum benefits of ozone
control, assuming control of most or all man-
made sources of ozone.
Multiple analyses were performed investigat-
ing the effect of alternative ambient ozone
concentrations, as wen as sensitivity to crop
response assumptions and export demand
conditions. Each analysis considers four am-
bient ozone concentration scenarios: 10%,
25%, and 40% reductions and a 25% increase.
These changes are measured as departures
from the 1980 actual ambient ozone levels and
are assumed to occur solely within the United
States.3
The 10% and 25% adjustments around am-
bient levels are considered plausible in a regu-
latory context. Changes of this magnitude are
found in the recent temporal variability dis-
played by ambient ozone levels. The 10% and
25% improvements in ozone levels are also of
policy importance in that these adjustments
parallel proposed alternative regulations
' Ozooe cooccaualioos sic iimmr<1 constinl ia the rat of the
wodd. Such an itmmptioc b (cncnUy continent with the
meteorology of ozooe foraatioo aad transport.
Overall Assessment
The NCLAN data contain information on the
ozone response of several corn, soybean,
wheat, and cotton cultivars (varieties). Differ-
ences in cultivar response are not statistically
significant within most of the soybean, wheat,
and corn data (Heck et al.); thus, pooled re-
sponse functions are used. For cotton, indi-
vidual response functions for irrigated (west-
ern U.S.) and nonirrigated (southeast) cotton
cultivars are used. Only single cultivar re-
sponse functions are available for grain sor-
ghum and barley. Using this mix of response
functions, the four ozone alternatives then are
translated into corresponding yield changes
for use in the benefit evaluations.4
4 The rente otozooe-crop sensitivities captured ia the NCLAN
data vary from eottoo aad soybean (owst sensitive) to bailey (least
sensitive). For (be 25% ozooe reductioa alternative, the respective
average yield adjustments (increases) are: cotton. 9.0%: soybean.
U%; wheat (winter). 3.4%; wheat (sprint). 1.3%; corn, 1.2%.
pain sorghum. 1.0%: and bailey, J2%.
-------
890 November 1986
Table 2. Annual Benefits of Alternative Ozone Levels In 1980
Aimer. J. Agr. Econ.
Ozooe
Assumption
Economic Surplus
Changes in
Economic Suiphis
Producers*
Surplus
Consumers'
Surplus
Tool
Surplus
Producers'
Surplus
Consumers'
Surplus
Tout
Surplus
Base
10% reduction
2596 reduction
4096 reduction
2596 increase
26.015
2&2S0
26.567
26.788
25.413
114.957
115.390
116.116
116.701
113.462
140.971
141.640
142.683
143.489
138.875
(S billion)
0-235
0J52
0.773
-0.607
0.433
1.159
1.744
-1.495
0.669
1.712
2J18
-2496
The simulated economic effects of ozone
changes are portrayed in table 2. Estimates of
the annual benefits from reduced ozone in-
crease as ozone pollution is reduced. Spe-
cifically, the benefit estimates of 10%, 25%,
and 40% reductions in ambient ozone are ap-
proximately $0.7, $1.7, and $2.5 billion, re-
spectively. The cost of a 25% increase in
ozone is $2.1 billion. These economic esti-
- mates amount to percentage changes of the
objective function of approximately 0.5%,
1.4%, 2.0%, and —1.7%, respectively. These
changes are triggered by ozone-induced aver-
age crop yield changes (across the six crops)
of 1.1%, 2.5%, 3.8%, and -3.0%. Market-
place substitutions permit a partial mitigation
of the yield reductions.
The distributional consequences of regula-
tory policies are also relevant. In this analysis,
both producers and consumers gained. In ab-
solute terms, consumers benefit substantially
more than producers. The benefit to domestic
consumers arises from falling prices for sev-
eral of the commodities. The increase in pro-
ducers* surplus under conditions of increased
supply arises because of the complex interac-
tion of the demand and supply relationships
within the model. Specifically, domestic and
foreign demand are characterized by varying
elasticities (including elastic assumptions for
some exports). Further, shifts in supply within
the model are not always characterized by
parallel shifts. In addition, primary (e.g., feed
grains) and intermediate (e.g., livestock) com-
modities are included in the model, with cor-
responding derived demand implications
vis-d-vis producers' effects. Under these con-
ditions, increases in producers* surplus with
increased supply are realized.
Another distributional result involves com-
parative welfare of domestic and foreign con-
sumers. Most of the consumers* surplus arises
from domestic consumption. However, in per-
centage terms, the foreign consumers exhibit a
larger relative change in surplus. This is not
surprising given the greater relative elasticities
of the export demand curves.
A final distributional aspect concerns re-
gional effects. Regions display different am-
bient ozone levels and hence different yield
responses/ Almost all regions benefit from re-
duced ozone and suffer losses from increased
ozone. The greatest absolute benefits geoer- _
ally accrue to regions with the greatest value *
of included crops, Le., the Corn Belt How-
ever, in relative terms, the distribution of
gains and losses is somewhat different. The
regions exhibiting the most sensitivity in per*
centage terms from reduced ozone have fairly
high ambient levels and a crop mix dominated
by sensitive crops, i.e., soybeans and cotton.
These regions are the Pacific (including
California), Delta, Northeast, and Southeast.
Regions like the Northern Plains received al-
most no benefits (or losses) from adjustments
in ozone due to relatively low ambient ozone
levels and a lack of ozone-sensitive crops.
Overall, the results indicate that the benefits
of moderate ozone reductions are substantial
in absolute terms but a relatively small per-
centage of total agricultural value (approxi-
mately 3% of gross crop value). The benefits
of ozone reductions accrue to both producers
and consumers, with about 60% of the con-
* These results also have a methodological fanpticaiioo, in that
the implementation of the McCarl (1982a) proposal within the sec-
tor awdd yielded satisfactory results. Specifically, the sector
model generated plausible regional acreages and crop mixes for
the various ozooe analyses using finear-programming-cepre-
seatative farm models to generate Cora Belt crop activities. The
linear programming procedure provided a more satisfactory range
of potential crop activities than did the use of historical cropping
paneres (used for the remaining U.S. subrcgions). However, data
and computational costs were correspoodingly higher for the
hnear programming procedure. Also, the procedure did need to be
modified so that yield changes associated with crop mix changes
are incorporated. A more extensive evaluation of this procedure b
prevented in Hamilton. McCarl, and Adams.
-------
Adams, Hamilton, and McCari
Pollution Control 891
sumer benefits accruing to foreign consumers.
Regionally, the major beneficiaries are those
areas with high relative levels of ozone and
ozone-sensitive crops.
These results can be compared to two re-
cent studies of agricultural benefits of ozone
control. Specifically, Shriner et al. estimate
the benefits of eliminating all man-made
sources of pollution to be from $2 to $5 billion.
These estimates are derived from a damage
function analysis (i.e., change in crop yield
times an invariant market price) for each com-
modity and hence are not true economic esti-
mates. In addition, the response data repre-
sent only one year of the NCLAN program
(1980). A recent study by Kopp, Vaughn, and
Hnrflia employs a more rigorous and defensi-
ble economic framework in estimating agricul-
tural benefits of ozone control. In that study,
the benefits (measured in economic surplus) of
a more stringent ozone standard of .08 ppm
are estimated to be $1.2 billion On 1978
dollars). In terms of an ozone assumption,
the Kopp, Vaughn, and Ha7.illa analysis is
somewhat comparable to the 25% reduction
scenario reported here. Accounting for the dif-
ference in time period, crop coverage, and re-
sponse data, the results are quite similar, sug-
gesting that the agricultural benefits of ozone
are indeed of this general magnitude.
Sensitivity Analysis
A number of factors that may affect the benefit
estimates are missing from or are incompletely
treated in the NCLAN data. Specifically, we
investigated the effects of including hay, po-
tential moisture stress-ozone interactions and
export market conditions on the benefit es-
timates.6
Hay. One important commodity for which
NCLAN data are not currently available is
hay. Alfalfa or grass-legume hay is an impor-
tant crop in several regions and in the overall
feed-livestock balance. While NCLAN data
are not available currently to measure the ef-
fects of ozone on hay yields, a study of alfalfa
hay response indicates moderate sensitivity of
hay yield to ozone (Oshima et al.). To gauge
the potential effects of hay on overall benefits,
a hay response adjustment, assumed to be the
4 Sensitivity analyses were also performed on a frw cultivars of
coca, soybeans, and wheat that display responses unlike the
pooled cultivar respoases used in generating the preceding esti-
mates. Using the most and least sensitive cultivars in place of the
footed response resulted in benefit estimates approximately 50%
greater and lower than from the above analysis
average yield response of the other crops, was
included. With the inclusion of hay, the
benefit values increase to $0.75, $1.9, $2.9,
and —$2.4 billion, an approximate 13% in-
crease in benefit estimates. This points out the
importance of the feed-livestock linkage in the
model and demonstrates the need for data
on the ozone sensitivity of hay, pasture and
range.
Moisture stress. Preliminary evidence sug-
gest that moisture-stress and other environ-
mental covariates alter the effect of ozone on
crop yield (Tingey et al.). This type of interac-
tion has important implications for nonir-
rigated crops. Ozone concentrations tend to
increase during hot, clear weather. Conse-
quently, high ozone levels would often be ac-
companied by low water availability or plant
moisture-stress. While most NCLAN data are
generated under adequate moisture condi-
tions, limited data are available on ozone
moisture-stress interactions from three
NCLAN experiments (one on cotton and two
on soybeans). In addition, simulation results
on moisture-stress-ozone interactions are
available (King and Snow). King and Snow's
data allow for a preliminary analysis of ozone
effects in the presence of moisture-stress us-
ing the assessment model. The results indicate
that the interaction of moisture stress lowers
the benefit estimates about 25% for each
ozone assumption. For example, the 25%
ozone reduction now results in benefits of ap-
proximately $1.56 versus $1.94 billion. These
results point to a need for more comprehen-
sive experimental designs in dealing with envi-
ronmental stress relationships.
Export market conditions. A large amount
of U.S. agricultural production is exported. In
turn (and as recent history has shown) agricul-
tural exports are sensitive to the relative ex-
change rates of U.S. and foreign currencies.
To test the effect of exchange rate changes on
the benefit estimates reported here, an analy-
sis was performed to portray an approximate
10% increase in the strength of the dollars vis-
a-vis major importers of U.S. commodities (a
commodity-weighted exchange rate adjust-
ment against the 1980 value of the U.S. dol-
lar). This adjustment was manifest in a lower-
ing of the export demand relationships.7 The
effect of such an adjustment was an approxi-
mate 14% reduction in the total benefits of
* Such an exchange rate change would affect numerous other
factors including factor supply and domestic demand. The analy-
sis here is done only 10 be suggestive of overall effects.
-------
892 November 1986
Amer. J. Agr. Econ.
ozone control to agriculture. This indicates
the importance of export assumptions and
suggests the need for assessments using data
under alternative exchange rate conditions.
The Relative Importance
of Agricultural Benefits
The above results indicate substantial eco-
nomic consequences of ozone within the
agricultural sector. However, perspective is
needed on the magnitude of these estimates
before they can be useful in policy formation
and evaluation. WhOe there are insufficient
data to perform a complete comparison of
these benefits and costs, some USEPA evi-
dence, in combination with the recent studies
cited above, can provide needed perspective.
Agricultural benefits are only one category
of potential benefits accruing to ozone control.
In addition, effects of air pollution on health
and welfare (materials, other vegetation, aes-
thetic values) are also used in promulgat-
ing pollution control standards. A recent
USEPA report (1985) suggests that among the
primary and secondary effects categories eval-
uated, agricultural benefits contributed ap-
proximately half of total economic benefits of
ozone control.
In a rank ordering, agricultural benefits
were followed by human health (as measured
by lost work days, etc.), materials, and other
vegetation. While not covering all effects
(e.g., insufficient data exist to evaluate effects
on forests), the ranking can be used to place
these benefits in perspective.
The costs of reducing ambient ozone levels
primarily involves controls on automobiles
and truck emissions, as well as stationary
sources, such as power plants, to reduce
ozone precursors. Costs of compliance reflect-
ing various control technologies and ozone
standards are reported by the USEPA (1979).
Because of the nature of the control technol-
ogy and other assumptions, a fairly wide range
of cost estimates for each standard is pro-
vided. The costs of controlling ozone from the
present standard (.12 ppm hourly maximum
not to be exceeded more than once per year)
to a more stringent standard (.08 ppm) are
from approximately $2 to about S3 billion, de-
pending on the control technology. When
compared with 23% ozone reductions benefits
reported here for agriculture ($1.7 billion), and
USEPA *s estimated ordering of other benefit
categories, it is possible that the benefits
across all sectors may equal or exceed such
costs. It should be noted, however, that these
compliance cost estimates reflect 1977 tech-
nology and dollars and are sensitive to the
control technology assumptions employed.
Current cost estimates are needed to compare
fully the benefits and costs of ozone control.
Concluding Comments
This study addresses the potential welfare ef-
fects of alternative ozone pollution levels on
the U.S. agricultural sector. Overall, the re-
sults indicate that there are substantial abso-
lute, but small relative (approximately 3% of
the value of agricultural production), benefits
of moderate ozone reductions. The 10%, 25%,
and 40% reductions in ozone result in annual
benefits of $.756, $1,937, and $2,859 billion.
A 25% increase in ozone results in annual
damages of $2,363 billion. These benefits are
not spread uniformly among participants in
the agricultural sector. About 60% of the an-
nual consumer benefits accrue to foreign con-
sumers. Domestic consumers' benefit from
slightly lower prices of processed grain and
livestock products because of increased sup-
plies of feed grains and oQseed. Producer
benefits accrue from export and factor market
adjustments. Regionally, the major benefi-
ciaries are those areas with high relative levels
of ozone and ozone-sensitive crops.
More extreme crop (variety) response infor-
mation derived from the NCLAN data sets in-
creases the range of the benefit estimates.
Conversely, the inclusion of interactions be-
tween ozone pollution effects and moisture-
stress reduced the expected annual economic
benefits by approximately 25%. Sufficient
prior biological and statistical information is
available to indicate that the extreme crop re-
sponses are indeed extreme responses. How-
ever, there is also sufficient evidence concern-
ing moisture stress interactions to suggest that
such effects are plausible and indeed likely.
Thus, future assessments of environmental
stress on agriculture need to include such in-
teractive effects, given their likely influence
on benefit calculations. The export sensitivity
experiment also indicates that future policy
assessments should involve a careful evalua-
tion of this factor.
The benefits assessment above not only ac-
counts for biological responses but also cap-
tures the economic responses as portrayed by
microlevel producer behavior and the market
-------
Adams, Hamilton, and McCarl
Pollution Control 893
structure in the sector model. The resultant
benefit estimates of alternative ozone pollu-
tion levels obtained from this bioeconomic
analysis provide one measure of the efficiency
of air pollution control strategies. It should be
noted, however, that the compliance costs of
achieving such ozone changes are not in-
cluded directly in these benefit estimates.
Thus, the net benefits to society from ozone
changes are not evaluated. While a compari-
son of USEPA cost estimates with the benefits
calculated in this analysis indicates that the
total benefits of increased ozone control may
approach or exceed costs, additional research
on benefits to other categories (health, materi-
als, etc.) as well as current compliance cost
estimates are required.
[Received September 1984; final revision
received November 1985.]
References
Adams, R. M.. and B. A. McCarl. "Assessing the
Benefits of Alternative Ozone Standards: The Role of
Response Information." J. Environ. Econ. and Man-
age. 12(1985)^64-76.
Adams, R. M., T. D. Crocker, and N. ThanavibulcfaaL
"An Economic Assessment of Air Pollution Damages
to Selected Annual Crops in Southern California." J.
Environ. Econ. and Manage. 9(1982):42-58.
Adams, R. M., S. A. Hamilton, and B. A. McCarl. The
Economic Effect* of Ozone on Agriculture. EPA-600/
3-84-090. Corvallis OR. Sep. 1985.
Baumes, H. "A Partial Equilibrium Sector Model of U.S.
Agriculture Open to Trade: A Domestic Agricultural
and Agricultural Trade Policy Analysis." Ph.D.
thesis, Purdue University, 1978.
Brown, D., and J. Pheasant. A Linear Programming As-
sessment of Economic Damages to Midwest Agri-
culture Due to Ozone. Final Report to Corvallis
Environmental Research Laboratory, U.S. En-
vironmental Protection Agency. Corvallis OR. 1983.
Burton, R. "Reduced Herbicide Availability: An Analysis
of the Economic Impacts on Agriculture." Ph.D.
thesis. Purdue University, 1982.
Chat tin, B., B. A. McCarl, and H. Baumes. Jr. Users'
Guide and Documentation for a Partial Equilibrium
Sector Model of US. Agriculture. Purdue University
Agr. Exp. Sta. Bull. No. 313. 1983.
Duloy, J. H.. and R. D. Norton. "CHAC: A Programming
Model of Mexican Agriculture." Multilevel Planning:
Case Studies in Mexico, ed. L. Goreux and A.
Manne, pp. 291-312. Amsterdam: North-Holland
Publishing Co.. 1973.
Hamilton, S. A., B. A. McCarl. and R. M. Adams. "Im-
proving the Performance of Programming-Based En-
vironmental Analyses of Agriculture." Amer. J. Agr.
Econ. 67(1983):407-I3.
Heady, E. O., and U. K. Sri vast*va. Spatial Sector Pro-
gramming Models in Agriculture. Ames: Iowa State
University Press, 1975.
Heck. W. W.. W. W. Core, J. O. Rawfings. L. J.
Zaragoza, A. S. Heagle, H. E. Heggestad, R. J.
Kohut, L. W. Kress, and P. J. Temple. "Assessing
Impacts of Ozooe on Agricultural Grape Crop Yield
Functions and Alternative Exposure Statistics." J.
Air Pollution Control Assoc. 34(1984)^10-17.
House, R. M. USMP: A Mathematical Programming
Model for Agriculture Sector Policy Analysis. Wash-
ington DC U.S. Department of Agriculture, Ecoo.
Res. Serv. Draft Rep., 1983.
King, D., and M. Snow. "Current Pwgrtu in Modeling
Drought by Ozooe Interactions." Paper presented at
NCLAN Annual Mrrting, Corvallis OR. 28 Jan.
1984.
Kopp, R. J., W. T. Vaughn, and M. HazQla. Agricultural
Benefits Analysis: Alternative Ozone and Photo-
chemical Oxidant Standards. Final Report to Eco-
nomic Analysis Brandt, U.S. Environmental Protec-
tion Ageacy, Research Triangle Park NC, Juoe 1984.
McCarl, B. A. "Cropping Activities in Agricultural Sector
Models: A Methodological Proposal." Amer. J. Agr.
Econ. 64(1982a):768-72.
. "REPFARM: Design, Calculation and Interpreta-
tion of the 1 incar Programming ModeL" Purdue Uni-
versity Agr. Exp. Sta. BuIL No. 38S, 1982b.
McCarl. B. A., and T. H. Spreen. "Price Endogenous
Mathematical Programming as a Tool for Sector
Analysis." Amer. J. Agr. Econ. 62(1980)^7-102.
Oshima, R. J., M. P. Foe. P. K. Braegelmann, D. W.
Baldwin, and V. VanWay. "Ozooe Dosage-Crop
Loss Function for Alfalfa: A Standardized Method
for Assessing Crop Losses from Air Pollutants." J.
Air Pollution Control Assoc. 26(1976):861-65.
Shriner, D. S.. W. W. Cure. A. S. Heagle. W. W. Heck.
5. W. Johnson, R. J. Olson, and J. M. Kelly. An
Analysis of Potential Agriculture and Forestry Im-
pacts of Long-Range Transport Air Pollutants.
ORNL-590, Oak Ridge National Laboratory, Oak
Ridge TN, 1982.
Tingey, D. T., G. L. Thuft, M. L. Gumpertz, and W. E.
Hogsett. "Plant Water Status Influences Ozooe
Sensitivity of Bean Plants/' Agr. and Environ.
7(1982)243-54.
U.S. Department of Agriculture. Agricultural Statistics,
1981. Washington DC, 1982a.
. Statist. Rptg. Svc. Statist. Bull. No. 552. Wash-
ington DC. 1982b.
U.S. Environmental Protection Agency. Air Quality Cri-
teria for Ozone and Other Photochemical Oxidants,
vol. 3. EPA-600/8-84-020A, Washington DC. June
1984.
. Costs and Benefits of Reducing Lead in Gasoline:
Final Regulatory Impact Analysis. EPA-230-05-85-
006. Washington DC. 1985.
. Costs and Economic Impact Assessment for Al-
ternative Levels of the National Ambient Air Quality
Standards for Ozone. EPA-405/5-79-002. Washington
DC. 1979.
------- |