Tennessee
Valley
Authority
United States
Environmental Protection
Agency
Division of
Environmental Planning
Chattanooga, Tennessee 37401
Office of Research and
Development
Office of Energy, Minerals
and Industry
Washington, D.C. 20460
E-EP-77-5
EPA-600/7-77-122
November 1977
EXPERIMENTAL AIR
EXCLUSION SYSTEM FOR
FIELD STUDIES OF SO2
EFFECTS ON CROP
PRODUCTIVITY
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
E-EP-77-5
EPA-600/7-77-122
November 1977
EXPERIMENTAL AIR EXCLUSION SYSTEM FOR
FIELD STUDIES OF S02 AFFECTS OR CROP PRODUCTIVITY
by
H. C. Jones, N. L. Lacasse, W. S. Liggett, and Frances Weatherford
Tennessee Valley Authority
Muscle Shoals, Alabama 35660
Interagency Agreement EPA-IAG-D6-721
Project No. E-AP 79 BDL
Program Element No. INE 625B
Project Officer
C. W. Hall
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
Prepared for
OFFICE OF ENERGY, MINERALS, AND INDUSTRY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
-------
DISCLAIMER
This report was prepared by the Tennessee Valley Authority and
has been reviewed by the Office of Energy, Minerals, and Industry,
U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the
views and policies of the Tennessee Valley Authority or the U.S.
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation for
use.
-------
ABSTRACT
The objective of this project is to characterize and quantify the
relationships among sulfur dioxide (S02) exposure, symptomatology of
injury, and yield of crops that are economically important to the south-
eastern United States. Emphasis is placed on the soybean because of its
sensitivity to S02 and its economic importance.
Five experimental sites—four in the vicinity of the Widows Creek
Steam Plant and one at a location that receives exposure only to back-
ground levels of S02—were selected and characterized during the first
year of study. Characterization includes analysis of soil fertility,
pH, soil depth, slope, and content of organic matter. These data and
data for the yield of soybeans were analyzed by regression methods to
identify site factors that might affect yield and mask the effects of
S02 exposure. Results of the soils analyses and regression analysis
were used to adjust soil fertility and pH levels to optimal levels for
soybean production and to eliminate a magnesium-potassium interaction
that was affecting yield.
An air-exclusion system was designed, constructed, and tested
that permits the comparison of plants exposed to S02 with plants at
the same site that were protected from S02 exposure. The system,
to remove or exclude S02-polluted air from 0.004-hectare plots during
ground-level exposure, blows charcoal-filtered air through plastic
tubes laid between rows of soybeans to remove the S02. This system is
activated automatically when S02 concentrations in the vicinity of the
plots equal or exceed 262 |Jg/m3 (0.1 ppm). Tests showed the system to
be as much as 85 percent efficient in excluding or reducing S02 concen-
trations to subthreshold levels during exposure. However, further
testing is needed to improve system efficiency, particularly to protect
plants less than 60 cm high from exposure. The soybean yields on plots
exposed to S02 at one of the sites averaged 12 percent less than those
on plots not exposed to S02. However, regression analyses of the data
indicate that only about one-half of the reduction may be attributed to
exposure to S02, the remainder being attributed to differences in other
factors at that site.
This report was submitted by the Tennessee Valley Authority, Divi-
sion of Environmental Planning, in partial fulfillment of Energy Accom-
plishment Plan 79 BDL, under terms of Interagency Agreement D6-721 with
the Environmental Protection Agency. Work was completed as of June 1977.
111
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgment viii
1. Introduction I
2. Conclusions and Recommendations 6
Characterization of site factors 6
Efficiency of air-exclusion system 8
S02 effects on soybeans 8
3. Materials and Methods 10
Selection and location of study sites 10
Characterization of edaphic factors 11
Soil description 11
Soil pH and fertility 11
Depth to bedrock 11
Air-exclusion system 12
Description 12
Installation and operation 16
Testing efficiency 16
Smoke tests 16
Fog generator tests 16
SC<2 exposure tests 17
Sulfation (Huey) plate tests 17
Foliar sulfur analysis 17
Cultural practices on experimental sites 18
Soil treatment and plot establishment 18
Planting 18
Air-exclusion treatment 18
Weed and pest control 19
Seed harvest and processing 19
Environmental monitoring and observations 19
Plant observations 19
Analysis of data 20
4. Results and Discussion 25
Surface and subsurface characteristics of
experimental sites 25
Efficiency of air-exclusion system 25
Smoke and fog generator tests 25
SC<2 exposure tests 25
Sulfation (Huey) plate tests 30
Foliar sulfur analysis 32
Environmental monitoring 33
Sulfur dioxide 33
Ozone 34
Foliar injury, growth, and yield of soybeans 35
Foliar injury 35
Height growth 36
Yields 37
IV
-------
References 48
Appendixes
A. Plot arrangement at each site and direction from
steam plant 50
B. Soil characteristics, average seasonal precipitation,
and temperature 56
C. Percentage organic matter, pH, and contents of P, K, Mg,
and Ca of the soils of the experimental sites 58
D. List of equipment and vendors 61
E. Planting and harvesting dates for soybeans during 1976 . . 62
F. Depth to bedrock, average relative elevation, and
percent slope of plots at experimental sites 63
G. Sulfur dioxide exposures and estimated percent foliar
chlorosis for the 1976 growing season 64
H. Yields for experimental plots, 1976 growing season,
Widows Creek Steam Plant 65
I. Data set used in regression analysis 66
-------
FIGURES
Number Page
1 Location, of continuous SOz monitors and associated study
sites near Widows Creek Steam Plant 7
2 Filter module containing dust filter and charcoal
filter pads 13
3 Air-exclusion units at monitor 6 showing charcoal filter
box and blower assembly, exhaust duct (arrow), and krene
plastic tubes 14
4 Krene plastic tubes for air-exclusion system 15
5 Comparison of effects of residuals on the size of the
criteria: least-squares and Huber 23
6a Three-dimensional view of surface characteristics at
site 10 (appendix A gives complete plot layout) 26
6b Three-dimensional view of subsurface characteristics
at site 10 (depth to bedrock) 26
7 Smoke from fog generator penetrating plant canopy 27
8 Residual smoke within plant canopy immediately after fog
generator is turned off . . : 28
9 Smoke from fog generator cleared from plant canopy after
fog generator is turned off 29
10 Monthly sulfation rates on three air-exclusion and
two exposed plots at monitor 9 31
11 Yield residuals vs. soil pH adjusted for variables in
model 42
12 Yield residuals vs. log calcium adjusted for variables
in model 42
13 Yield residuals vs. insect-disease index adjusted for
variables in model 43
14 Yield residuals vs. log phosphorus adjusted for other
variables in model 43
15 Yield residuals vs. log potassium adjusted for other
variables in model 44
16 Yield residuals vs. log magnesium adjusted for other
variables in model 44
17 Yield residuals vs. percent organic matter adjusted for
other variables in model 45
18 Residual vs. fitted yields 45
Al Plot arrangement, monitor 6 51
A2 Plot arrangement, monitor 9 52
A3 Plot arrangement, monitor 10 53
A4 Plot arrangement, monitor 19 54
A5 Plot arrangement, monitor 22 55
vi
-------
TABLES
Number Page
1 Sulfur Content of Soybean Foliage on Air-Exclusion and
Exposed Plots at Monitors 6, 9, 10, and 22 30
2 Number of Occurrences of Sulfur Dioxide Exposures Equal to
or Greater Than 1300 M8/J"3 1-h or 3-h Average 31
3 Total Exposure to Sulfur Dioxide Concentration Equal to
or Greater Than 262 (Jg/m3 During Daylight Hours3 for 1976
Growing Season 32
4 Maximum 1-h and 3-h Average Sulfur Dioxide Concentrations
Recorded at Study Sites 6, 9, 10, and 19 During 1976
Soybean Growing Season 32
5 Monthly Average of Hourly Ozone Concentration and Monthly
Average of Daily Maximum Ozone Concentration 33
6 Foliar Chlorosis of Soybean on Study Sites During the
Pod-Filling Stage of Growth 34
7 Average Height Growth of Soybeans by Treatment for
All Sites 35
8 Average Soybean Yield by Treatment for All Sites 36
9 Average Weight of 100 Soybean Seed by Treatment at
Monitor 6 36
10 Least-Squares and Robust Residuals 45
VII
-------
ACKNOWLEDGMENT
This work was conducted as part of the Federal Interagency
Energy/Environment Research and Development Program with funds
administered through the Environmental Protection Agency (EPA
Contract No. EPA-IAG-D6-0721, TVA Contract No. TV-41967A).
The EPA Project Officer for this research project is C. W.
Hall, 401 M Street, SW, Washington, DC. His contribution to the
direction of the research and constructive review of the reported
results are gratefully acknowledged. The TVA Project Director is
Herbert C. Jones, E and D Building, Muscle Shoals, Alabama.
Vlll
-------
SECTION 1
INTRODUCTION
The toxicity of sulfur dioxide (S02) to plants has long been recog-
nized and has been studied more intensively than that of any other
atmospheric pollutant. Most research on the effects of S02 exposure on
vegetation has been concentrated on the relationship between S02 dose
(exposure time x concentration) and foliar injury that manifests itself
as intercostal necrosis or chlorosis. Exposure to concentrations of
S02 greater than 1045 M8/m3 f°r an hour or more can cause visible foliar
injury of sensitive vegetation (1). Some studies indicate that some
species of vegetation can be exposed indefinitely to S02 concentrations
less than 522 |Jg/m3 without causing foliar injury (2,3). Mathematical
equations have been developed that describe the relationship between the
duration of exposure and the concentration of S02 that caused visible
injury to various plant species. However, sensitivity to S02 does not
depend on S02 dose alone; plants exhibit considerable variability in
their sensitivity, depending on species, genetic and physiological
characteristics of the individual plant, and the environment in which
the plant is growing. Therefore, although equations are useful in
predicting injury that might occur under given conditions for a limited
number of species, these equations are not readily adaptable to assess-
ing economic losses sustained by crops that have been exposed to S02 (4,5),
The influence of S02 dosages or foliar injury on plant productivity
is less well documented or understood. Loss in productivity is rela-
tively predictable in those cases in which reduction in foliage consti-
tutes a direct loss in yield, as in the case of foliar S02 injury to
forage crops such as alfalfa (3,6). However, there are few data on the
indirect effects of foliar injury on seed or fruit production or growth.
Linear relationships between total leaf area destroyed by S02 and reduc-
tion in yield have been reported for wheat (7), cotton (8), and soy-
beans (9) after exposure to S02 in closed chambers in the field where
edaphic factors were held more or less constant. Using these data,
regression models (i.e., y = a + bx, where y is percentage of full
yield) were developed for estimating yield losses on the basis of
measurements of a single independent variable such as total leaf area
destroyed (x). Under normal field conditions, it is questionable whether
yield losses resulting from foliar injury can be described by such a
simple model because of the multiplicity of uncontrollable cultural,
edaphic, and climatic factors that also affect yield. Jones et al. (10)
were unable to detect a significant effect of foliar injury on yield
of soybeans exposed to S02 emissions from a coal-fired power plant during
the prebloom stage of growth. However, in a similar study, reductions
in soybean yield as high as 351 kg/ha resulted from exposure to S02
emissions from the same power plant during the pod-filling stage of
growth (11). The reductions in yield were found to be related to the
amount of foliar chlorosis. In general, crop yields are not affected
by exposure to S02 unless visual foliar effects occur, and apparently,
more than five percent of the leaf area must be destroyed to measurably
reduce yield. Experiments with alfalfa indicate that, if leaf destruc-
tion is less than five percent, no residual effect on yield will occur,
even after several exposures (4).
-------
-2-
Most documented cases of S02 exposures that occurred under
field conditions and caused permanent adverse effects on the growth
or yield of vegetation have been associated with emissions from
smelters. The effects were caused by relatively high frequency
of exposure to excessive concentrations of SC>2. This type of expo-
sure does not occur normally in the vicinity of modern, well-sited
coal-fired power plants that are equipped with tall stacks (5). No
evidence has been presented to show that emissions from such plants
have caused reductions in plant productivity. The only report, for
the United States, of power plant emissions affecting the productivity
of a seed crop is that of Jones et al. (11) for the Shawnee Steam
Plant in western Kentucky, which at the time of the exposure was
equipped with short stacks.
Foliar injury can occur in crop and forest species when they are
exposed to S02 dosages that do not exceed the national secondary ambient
air quality standard of 1306 M8/ffl3 f°r 3 h (1). However, there are
insufficient data to predict (1) the amount of foliar injury that may
occur for exposure at or below the standard or (2) whether such expo-
sures affect productivity (12). Furthermore, the SC>2 may interact
with other pollutants, such as fluorides and oxides of nitrogen, that
are emitted from coal-fired power plants or with oxidants in the ambient
air to produce effects on vegetation.
Development of a capability for assessing the direct impacts of
atmospheric emissions of 862 from coal-fired power plants will require
characterization of dose-response relationships among S02 dosages
(alone and in mixtures of other pollutants), occurrence of foliar
injury, and effects on productivity under controlled exposure condi-
tions. Emphasis should be placed on (1) dosages that are near and
within the secondary standard for S02 and (2) species such as soy-
beans that are especially sensitive to SC>2 and, at the same time,
economically or ecologically important. Of equal importance are
(1) documentation of the occurrence of both short- and long-term
effects of ambient S02 levels on the productivity of vegetation and
(2) validation of dose-response relationships obtained under controlled
exposure in the field. The principal obstacle to characterization and
quantification of S02 effects in the ambient environment is the selec-
tion of adequate controls or baselines for making valid comparisons of
productivity between exposed and unexposed vegetation; that is, the
difficulty is to isolate an effect caused by S02 exposure from one
caused by other factors in the environment that are difficult to con-
trol or identify. In general, three types of approaches were used to
isolate S02 effects in the field:
1. Effects on plants grown in experimental plots were com-
pared at increasing distances from an S(>2 source, or
effects at plots near the source were compared with those
at a location remote from sources of the type being investi-
gated. Plots close to the source experience a higher fre-
quency of exposure and a higher dosage than plots at increasing
distance from the source. To make valid comparisons, all
other factors affecting growth must be constant or equalized,
a task which is difficult and sometimes impossible to achieve.
-------
-3-
2. Effects on plants grown in containers were compared at
increasing distances from the emission source. In this
approach, most edaphic factors are held constant, but cli-
matic factors cannot be controlled. This approach is
impractical if the plants are to be grown to maturity
because restricted root growth may affect normal growth
and development.
3. Effects on plants grown in field chambers that are venti-
lated with charcoal-filtered air are compared with effects
on plants grown in chambers ventilated with ambient air.
This approach is the most promising because exposed and
unexposed plants can be compared at the same location,
minimizing the effects of differing site factors. However,
the use of field chambers has resulted in special problems.
Mandl et al. (13) summarized the development and use of field expo-
sure chambers since 1896. All field chambers used until 1971 were of a
completely closed design. In a closed system, filtered and unfiltered
ambient air is circulated through paired chambers and vented under posi-
tive pressure to the atmosphere. Because of the closed design and the
characteristics of the covering material, radiant heat is trapped,
necessitating constant air circulation. This air circulation causes
abnormal air movement around foliar surfaces and changes the water
economy of plants. Exposure to natural precipitation and plant pests,
which causes problems in field plots, is precluded.
As a partial solution to these problems, Mandl et al. (13) designed
and tested a field chamber with an open top. This chamber design essen-
tially eliminates heat buildup within the chamber, admits sunlight and
natural precipitation, and prevents infestation with insects and disease
that occurs in field plots. However, continuous air circulation is still
required, and a pair of chambers (one with filtered air circulation and
one with ambient air circulation) is needed for each replication.
Howell (14) compared yields of soybeans grown in paired, open-top
chambers with yields in field plots. Significant reduction in soybean
yields occurred in the plants grown in chambers ventilated with ambient
air as compared with those grown in chambers with filtered air or those
grown in the field plots. There was no difference in yields between
the filtered air chambers and the field plots. Howell concluded that,
under the conditions of his experiments, field chambers in some way
increased the sensitivity of soybeans to pollution.
TVA biologists used all three approaches to determine the impact
of stack emissions from coal-fired power plants' on the appearance,
growth, and yield of soybeans. Soybeans were selected for study
because of their high sensitivity to S02 and their economic importance
to the Tennessee Valley region. Three specific types of experiments
were conducted:
-------
-4-
1. Soybeans were grown in portable greenhouses that were parti-
tioned into two equal sections, with one section receiving
filtered air and the other receiving ambient air. The green-
houses were located about 3 km from the steam plant in an
area receiving a relatively high frequency of exposure to
high S02 concentrations.
2. Soybeans were grown in six field plots located 3.5 to 56.5 km
from the steam plant. All plots were located in the same
type of soil and were fertilized and limed to the same levels.
3. Soybeans were grown in 19 small (3- by 3-m) field plots
located 3.2 to 43.8 km from the steam plant. To minimize
edaphic differences among sites, plants were grown at each
location in four 30-quart styrofoam chests containing a
standard soil mixture. Rainfall was supplemented with
irrigation so that all plots received essentially the
same weekly ration of water.
During the five years of operation of the filtered and unfiltered
greenhouses, the soybean yields have averaged 24 percent less in the
unfiltered portion of the greenhouse than in the filtered portion.
Because the plants grown in the filtered section are protected from
contaminants from all existing sources, TVA biologists cannot conclude
that the steam plant emissions were the sole cause of the reductions
in yield in the unfiltered sections. Also, because continuous air
circulation is required to prevent excessive temperatures, plants must
be irrigated to offset evapotranspiration losses. The irrigation tends
to produce luxuriant growth, and yields have been considerably above
those of field-grown plants. Comparisons between yields of crops grown
under optimum growing conditions and yields of those grown under field
conditions must, therefore, be interpreted with caution. No significant
differences were detected between the growth and yields of soybeans
grown near the steam plant in areas receiving exposure to S02 and the
growth and yields of soybeans grown at remote locations. However,
there was considerable variability, possibly because of edaphic and
climatic factors, in yields within sites and between sites. In general,
plots near the steam plant tended to produce higher yields.
To overcome the problems associated with field chambers and with
field plots in different locations, a prototype air-exclusion system was
designed and installed in 1975. The system (see Materials and Methods)
eliminates the need for an enclosure and, therefore, the need for con-
tinuous air circulation. Filtered air is circulated through the plant
canopy only when the pollutant concentration reaches a predetermined
concentration. Because the system operates for only a small part of
the time, a parallel system for control is unnecessary. Results of
this preliminary test showed that the yield of soybeans (c. v. Forrest)
was 28.6 percent lower in the plot without the air-exclusion system.
The yield from the air-exclusion plot was comparable to yields normally
expected under field conditions, but was less than half the yields
obtained in the greenhouses. However, the percentage reduction in
yield between greenhouse treatments was the same as that observed
-------
-5-
for the air-exclusion study. As a result of the successful test with
the 1975 prototype, plans were developed to expand the study. It was
proposed that 30 air-exclusion units be constructed, installed, and
evaluated on five sites in 1976.
This report describes the air-exclusion system, its operation,
and evaluation of its efficiency at excluding pollutants from the crop;
the establishment and characterization of the study sites; and the
results obtained for the 1976 growing season.
-------
-6-
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
Experimental plots (Figure 1) were established at four sites on
Sand Mountain near the Widows Creek Steam Plant in northeast Alabama
and at a remote location at the Auburn University's Sand Mountain
Experiment Station to study the effects of S02 emissions from the
steam plant on crop productivity. A system was designed and con-
structed to blow air that has been filtered by charcoal to remove
S02 through plastic tubes to exclude S02-polluted air from experi-
mental plots during ground-level exposures. The air-exclusion sys-
tem is activated by a continuous S02 monitor when the monitor registers
S02 concentrations of 262 |Jg/m3 or more. Three air-exclusion systems
were installed on soybean plots at each site to compare the yields
of plots protected from S02 exposure by the systems with those of
plots that were not protected. The primary objectives of the first
year of investigation were (1) to identify the sources and magnitude
of variation within sites and between sites attributable to factors,
primarily edaphic, that might mask S02 effects so that measures could
be taken to minimize them and (2) to characterize the efficiency of
the air-exclusion system in protecting vegetation from exposure to
S02. The secondary objective was to measure the effect of S02 expo-
sure in terms of foliar injury and productivity of soybeans. Soybean
plants are extremely sensitive to S02 and are a crop of great economic
importance in the Southeast. Several years might be required to obtain
sufficient data for characterization of the effects of S02 on the pro-
ductivity of soybeans grown under ambient conditions.
CHARACTERIZATION OF SITE FACTORS
Analyses of the surface and subsurface characteristics of the
sites showed that soil depth was extremely variable at site 19.
Because the growth and yield of plots at site 19 reflected the
shallowness of soils in some areas, site 19 was abandoned at the
end of the growing season and a new site was established. Some
shallow spots that were identified at site 6 will not be used for
plot locations in the future. No effect of soil depth on yield
was detected by regression analyses for sites 6, 10, 9, and 22.
Measurements of slope identified areas requiring drainage to mini-
mize excessive soil moisture at some sites during periods of high
rainfall.
Soil analyses showed that the phosphorus (P) and magnesium (Mg)
contents and pH of the soil varied within plots at site 6 and that
Mg and pH varied at site 10. Furthermore, levels of P, Mg, and
calcium (Ca), percent organic matter, and pH varied considerably
among sites. Regression analyses were used to identify soil factors
that influenced yields among the experimental plots and that might
mask the effect of S02 exposure on yields; most significant was a
magnesium-potassium (Mg-K) interaction. Increased yields were
associated with increasing ratios of Mg to K. Levels of Mg were
-------
-7-
WIDOWS CREEK
STEAM PLANT
SCALE
0 I 2 3km
Figure 1. Location of continuous SC>2 monitors and associated study
sites near Widows Creek Steam Plant.
-------
-8-
optimized at all sites by the addition of dolomitic lime or magnesium
sulfate, depending on whether there was also a need to optimize pH
levels. Levels of P at site 9 tended to be excessive, and regression
analyses indicated that yields might be reduced at the higher levels.
However, levels of P were uniform among the plots on the site and,
therefore, should not affect comparisons of treatments at the site.
Content of organic matter in the soil was an important variable
in explaining variation in yields among the sites. However, percent
organic matter is relatively uniform among plots within sites and should
not affect comparisons of treatments. Regression analyses also identi-
fied strong interactions among percent organic matter, pH, and Ca, with
greater amounts of Ca being required to produce the same pH in soils
with high organic content as in those with lower organic content.
EFFICIENCY OF AIR-EXCLUSION SYSTEM
Smoke, fog, and S02 exposure tests indicated that the air-exclusion
system might be as much as 85 percent efficient in excluding or removing
S02 from the soybean plots after the plants are about 2 feet tall. The
system was least efficient during the seedling stage of growth. Any
S02 that was not removed from the plots was reduced to concentrations
below the 1000- to 1300-|Jg/m3 level generally required to cause threshold
injury to most sensitive species of vegetation. Comparisons of observa-
tions of chlorosis and measurements of foliar sulfur between plots with
air-exclusion systems and those without indicated that the exclusion
systems effectively reduced both chlorosis caused by S02 and accumula-
tion of sulfur on foliage. After an S02 exposure at site 6, exposed
plots averaged 6.4 percent chlorosis whereas the air-exclusion plots
averaged 0.9 percent chlorosis. The sulfur content of the foliage at
the exposed plots was 34 percent higher than that at the air-exclusion
plots.
Results of the tests suggest that further testing is needed to
complete characterization of the efficiency of the air-exclusion
system and that the efficiency of the system could be improved by
preventing air loss from the periphery of the plots below the plant
canopy. Studies are planned for 1977 in which artificial S02 expo-
sures will be used to evaluate the effect of placing air barriers
around the periphery of the plots to increase the efficiency of the
system over the life cycle of soybeans. In addition, S02 will be
monitored continuously within the plots over the growing season to
evaluate the system under ambient conditions.
S02 EFFECTS ON SOYBEANS
Exposure to concentrations of about 1500 M8/m3 (0.6 ppm) S02
maximum 1-h average and 1100 |Jg/m3 (0.4 ppm) S02 maximum 3-h average
was required for visible threshold injury to the foliage of soybeans.
Exposure to concentrations of about 2350 pg/m3 (0.9 ppm) S02 maximum
-------
-9-
1-h average and 1500 |Jg/m3 (0.6 ppm) S02 maximum 3-h average was
required for significant injury (>5 percent leaf area affected). The
response of soybeans to these concentrations would be expected to vary
somewhat, depending on environmental conditions before, during, and
after an exposure and the variety and physiological condition of the
individual plants. However, other data for the relationship between
SC>2 concentrations and the occurrence of foliar effects on soybeans
in the field generally agree with these determinations.
Of the three sites on Sand Mountain, site 6 exhibited the most
evidence of yield reduction due to SC>2. Site 6 experienced the greatest
frequency of exposure to high concentrations of S02 at ground level
(X1300 (Jg/m3) and was the only site at which exposed plots exhibited
significant foliar injury. At site 6 the exposed plots had signifi-
cantly greater amounts of foliar sulfur. The difference in yields
between protected and exposed plots at this site averaged about
402 kg/ha. However, a local regression model attributed 222 kg/ha
of the difference to SQ% exposure and the rest to differences in
other factors at site 6. Thus, although not conclusive, the data
indicate that a reduction in yield of about 6 percent was attributable
to SC>2 exposure at site 6. Air-exclusion plots at all the other sites
produced somewhat greater yields than their comparable controls.
The residuals (unexplained variation) for yields generated by the
local model suggest that strong positive or negative bias might exist
for plots within treatments at some of the sites as the result of
unidentified site factors. Therefore, the residuals were used to
relocate those plots that will be used next year within treatments at
each site to reduce the chance that these factors might bias treatment
in a single direction and result in a misinterpretation of the effect
of SC>2 exposure on yields.
-------
-10-
SECTION 3
MATERIALS AND METHODS
SELECTION AND LOCATION OF STUDY SITES
The study area is located on Sand Mountain in Jackson County in
northeast Alabama, about 48 km southwest of Chattanooga, Tennessee,
in the proximity of the Widows Creek Steam Plant. The steam plant is
the third largest coal-fired electric-generating facility in the TVA
system and has a rated generating capacity of about 1971 megawatts (MW).
It consists of six 141-MW units, one 575-MW unit, and one 550-MW unit.
Stack heights are 48 ra for the six small units and 152 m for the two
larger units. The steam plant is located on the Valley floor, about
183 m above sea level. The study area on Sand Mountain is about 3 km
east of the plant at an elevation of 457 m above sea level. A control
site for the project was installed at Auburn University's Sand Mountain
Experiment Station at Crossville, DeKalb County, Alabama, about 69 km
southwest of the steam plant.
The Widows Creek area was selected for this study for the following
reasons:
1. The relatively high frequency of occurrence of ground-level
S02 exposures in the area. This phenomenon results from a
combination of the short stacks, topography, and prevailing
meteorological conditions. Dispersion of pollutants is
restricted by the Sand Mountain Plateau, located 1.6 km
east of the steam plant, and the Cumberland Plateau, located
8 km northwest of the steam plant. Adverse meteorological
conditions are associated with a midmorning shift in airflow
from a down-valley to an up-valley direction, a situation
that frequently carries the plume across the Sand Mountain
Plateau. Because of the proximity of the Sand Mountain
Plateau to the steam plant, persistent ground-level expo-
sures may occur before adequate dispersion has taken place.
2. The availability of existing continuous S02 monitors and
meteorological stations in the area.
3. The availability of agricultural land for establishment
of experimental sites adjacent to S02 monitors.
4. The importance of agriculture and forestry to the area.
The air monitoring network at the Widows Creek Steam Plant consists
of 14 stations at which S02 is monitored. These stations were located
in this area because of the probability of frequent exposures by S02
based on dispersion models and field observations of vegetation affected
by S02. Sites at monitoring stations 6, 9, 10, and 19 (Figure 1) were
selected as experimental study areas because research plots had pre-
viously been located there and additional land was available for lease.
-------
- -11-
A Philips S02 monitor was installed at the control site at Auburn
University's Sand Mountain Experiment Station to verify that only
background-level S02 exposures occur at the site.
Parcels of land adjacent to the monitors, each about 0.4 hectare (ha)
in size, were leased from their owners. The sites differ in configura-
tion because of terrain limitations or inability to lease better sites.
Sites are referred to by monitor numbers. Each site was subdivided
into twenty-four 0.004-ha plots—12 for air-exclusion plots and 12 for
exposure to ambient air. The arrangement of plots at each site and the
direction from the steam plant are shown in the figures in appendix A.
Air-exclusion and exposed plots cropped in 1976 are also shown in these
figures.
CHARACTERIZATION OF EDAPHIC FACTORS
Soil Description
The soils at all monitors except monitor 19 are of the Hartsells
series. Monitor 19 is of the Hanceville series. A complete description
of the soil characteristics, productivity, average seasonal precipitation
levels, and average temperature is given in appendix B.
Soil pH and Fertility
Soils were sampled before planting and were analyzed for P, K, Ca,
Mg, and pH by the Soils Testing Laboratory at Auburn University, Auburn,
Alabama. The sulfur content was determined by the Leco sulfur analysis
method (15).
An attempt was made to optimize the pH and fertility of the plots
by adding 4.5 metric tons of dolomitic lime and 1.1 metric tons of
4-24-24 NPK fertilizer per hectare in the spring of 1972 as soon as
soil conditions permitted.
The soils were sampled again after harvest in 1976 and were analyzed
for content of organic matter and the variables listed above. These
data were used to adjust soil fertility and pH levels for 1977 studies
and in regression analyses to account for variations in soybean yields
among the experimental plots. (For results of the soil analyses, see
appendix C.)
Depth to Bedrock
The soils on Sand Mountain, although used extensively for truck
farming, are relatively shallow and, therefore, subject to drought
during periods of low rainfall. Depth to bedrock at individual plots
was measured to identify subsurface characteristics that may affect
plant growth and development.
-------
-12-
Three measurements per plot were made with a 90-cm soil auger. One
measurement was made at the center of the plots, and the other two measure-
ments were made at the front and rear of the plots. (The front of the
plot is identified by the location of the plot number stake.) Points
between plots were interpolated to obtain a complete rectangular array
of data for each site. An interactive three-dimensional graphics
computer routine was used to depict the subsurface characteristics.
The terrain at each site was characterized to determine whether
slope or low-lying areas where surface runoff could accumulate affected
growth, development, and symptom expression of plants. Characterization
of the surface was accomplished by measuring with a transit and rod the
relative elevation of each plot on a 6- by 6-m grid. A temporary bench-
mark was established at each site to determine the height of the instru-
ment for future reference. An interactive three-dimensional graphics
computer routine was used to depict surface characteristics. The slope
of the plots was determined by difference in elevation from front to
rear of the plots.
AIR-EXCLUSION SYSTEM
Description
The air-exclusion system is a type of field exposure facility that
consists of three modules: air filtration, air circulation, and air
distribution.
The air filtration module consists of a dust filter and six panels,
each containing about 4.3 kg of charcoal in a galvanized metal box. The
box is airtight, except for the inlet blower, and is attached to the air
circulating module. The charcoal is of the activated CH type.1 Access
to the filters is through a hinged lid that, in the closed position, is
sealed to the box with a gasket.
The filter module containing the dust filter and charcoal filter
panels is shown in Figure 2.
The air circulating module consists of a wheel blower with a rated
air delivery of 122 m3/min. The blower is powered by a 1.5-hp motor
energized by 220 V a.c. The blower and the filter box are mounted on
an angle iron frame.
The air distribution module consists of two components: an exhaust
duct and four plastic tubes. The exhaust duct is constructed of galvan-
ized sheet metal and contains fins and deflectors to distribute the air-
flow evenly into four exhaust ports (Figure 3). Krene plastic tubes,
attached to the posts, extend between the rows of the crops. The dimen-
sion of the tubes, hole size, and distribution are shown in Figure 4.
1. All equipment and vendors are listed in appendix D.
-------
Figure 2. Filter module containing dust filter
and charcoal filter panels.
-------
Figure 3. Air-exclusion units at monitor 6 showing charcoal
filter box and blower assembly, exhaust duct (arrow),
and krene plastic tubes.
-------
CROSS SECTION OF TUBES AS PLACED IN FIELD
(Projections show orientation of holes)
A A
Left Right
A. Twenty-four holes, 2.5 cm diameter each, spaced 30.5 cm apart along length
B. Forty-eight holes, 2.5 cm diameter each, spaced 30.5 cm apart along length
en
I
SIDE VIEW OF AIR-EXCLUSION TUBE (LEFT)
/
30.5
cm
K
7.6 m
Figure 4. Krene plastic tubes for air-exclusion system.
-------
-16-
Installation and Operation
The air-exclusion units were installed on patio blocks and wired
through underground service lines to a central breaker panel. A relay
in the S02 recorder activates the air-exclusion units when the S02 con-
centration reaches 262.0 |Jg/m3. The units were installed during the
first three weeks in May 1976 and put into operation when the seedlings
emerged from the ground during the first week in June.
A digital timer was wired into the S02 recorder relay that activates
the blowers to record the number of hours of operation. The timers were
read and reset each week. An average time of operation was compiled each
week, and the blowers at monitor 22 were operated an equal period during
each following week. The latter operation was controlled through a timer
switch. The blowers at Crossville were operated for an amount of time
equal to the average operating time of blowers at the Widows Creek site
to determine whether blowing air through the crop affects growth and
yield of the crop.
The efficiency of the charcoal filters was determined by passing
S02 from a permeation tube through a known quantity (1 g) of charcoal
inserted in the sampling train of an S02 monitor. The amount of S02
adsorbed by the charcoal was calculated by measuring the flow through
the monitor. Estimating the average S02 concentration in the Widows
Creek area would allow calculation of the amount of time that the air-
exclusion units could operate efficiently to remove S02. As a check to
the above method, the actual S02 exposures that were recorded by the
monitors at the experimental site during the 1976 growing season were
calculated to determine whether the S02 burden exceeded capacity for
S02 adsorption of the charcoal charge of the air-exclusion units.
Testing Efficiency
Extensive tests were conducted during the summer of 1976 to evalu-
ate the efficiency of the air-exclusion system for excluding air pollut-
ants. Five types of tests were used: smoke grenades, fog generation,
artificial S02 exposure, sulfation plates, and foliar sulfur analysis.
Smoke Tests-
Tests were performed with 3-min smoke grenades immediately after
the system was installed and before the seedlings emerged from the
ground. The first test was conducted on June 2 between 5:45 a.m. and
6:20 a.m. Wind speed during the test was 0.5 to 1.5 m/s. Observations
were recorded with a still camera.
Fog Generator Tests—
The second test was conducted with a fog generator July 21 between
7:30 a.m. and 10:30 a.m. The soybeans averaged 66 cm in height. Wind
speed during the tests averaged 2 m/s. Observations were recorded with
a still camera and a movie camera.
-------
-17-
S02 Exposure Tests--
The third test was conducted by artificial S02 exposure on
September 11 between 10 a.m. and 2 p.m. The average height of the
soybeans was 102 cm. During this test, the plot was enclosed with
a 60-cm-high fiberglass air barrier. The air barrier around the
periphery of the plot was placed to reduce lateral air loss from
the side of the plot and increase vertical air movement. Two portable
SC>2 monitors were placed inside the plot. A third monitor equipped
with a 4.5-m sampling probe was placed outside the plot, and samples
were drawn from various locations within the plot. Wind speed during
the test averaged 1.5 m/s. Anhydrous S02 was injected into a 5-cm
plastic pipe that had been perforated along one side to simulate a
line source of S02. The S02 was diluted by forcing air into the
pipe with a squirrel cage blower.
The fourth test also consisted of an artificial S02 exposure
conducted October 8 between 9 a.m. and 12 p.m. Wind speed during
the test averaged 0.6 m/s. This test was performed on a cotton plot
because the soybeans were defoliated. The plot was enclosed with a
30-cm-high fiberglass barrier between the inner and outer buffer rows.2
One portable S02 monitor was placed in the center of the plot, and the
monitor with the 4.5-m probe was used again as in the previous test.
This test was performed in two parts. The first part of the test was
conducted to determine whether reducing the plot size by placing the
barrier between the buffer rows would increase vertical airflow. The
second part of the test consisted of reducing by half the length of
plastic tubes and reducing the plot size by half its original length,
thereby doubling the airflow through the plot.
Sulfation (Huey) Plate Tests—
Sulfation plates were used to quantify the efficiency of the
air-exclusion system in excluding S02 from the plots (16). Plates
were installed at monitor 9 on three air-exclusion plots and on two
exposed plots. Plates were set in the center of each plot at 30 and
60 cm above ground on July 15 and at 120 cm above ground on August 5.
Plates were changed monthly and analyzed by the turbidimetric barium
sulfate method (17). All plates were removed October 8 when the
air-exclusion system was deactivated.
Foliar Sulfur Analysis—
Soybean foliage was analyzed for total sulfur content at the end
of the growing season in another attempt to quantify the efficiency of
the air-exclusion system. Samples of soybean foliage were collected
Buffer rows and buffer plants are rows of plants or additional
plants around the periphery of an experimental area to reduce the
edge effect on the test plants. Plants on the periphery of a
plot are exposed to more sunlight than plants grown in the interior
of the plot, thus resulting in an edge effect.
-------
-18-
from all sites on September 28 and 29 according to the following
procedure: Five soybean plants from the inner buffer rows were
harvested, placed in paper bags, and transported to Muscle Shoals.
The following day the five topmost trifoliates from each plant were
cut, excluding the petioles, and washed in a mild soap solution,
rinsed with running tap water, and rerinsed with distilled water.
The leaves were dried at 70°C for 24 h in a forced-draft oven. The
dried leaves were ground in a Wiley mill and analyzed by the Leco
sulfur analysis method (15).
CULTURAL PRACTICES ON EXPERIMENTAL SITES
Soil Treatment and Plot Establishment
The experimental sites were plowed and disked in the spring of
1976 as soon as field conditions permitted. The sites were then
limed, fertilized, and disked again. Plots were staked (appendix A),
and electrical service lines were installed underground.
Two weeks before planting, Trifluralin was applied at the rate
of 0.126 cm3/m2 and tilled to a depth of 7 to 10 cm immediately
following application to control weed growth.
Planting
Soybean [Glycine max (L.) Merr. c.v. Essex] was selected as a
test crop because of its growing importance in the agricultural
economy of the Widows Creek area, the Tennessee Valley, and the
Nation. The Essex variety was chosen because of its maturity group-
ing and nonbranching habit. Seed were obtained from the Auburn
University Sand Mountain Experiment Station, treated with a com-
mercial preparation of Captan and molybdenum, and inoculated with
Rhizobium bacterium. Seed were planted by hand in furrows at the
rate of 1 to 2 seed every 5 to 7 cm. Seven rows, 7.3 m long, were
planted 75 cm apart on each plot. The two outer rows on each side
of the plots served as buffer rows, and the plants in a 0.61-m
length at each end of the three inner sampling rows served as
buffer plants for the row. The sampling rows were divided into
10 sections 0.61 m in length, staked, and numbered to provide
permanent identification of the section. Planting dates are
presented in appendix E.
Air-Exclusion Treatment
The locations of the soybean plots at each monitor are shown
in appendix A. Each site had three air-exclusion plots (test plots)
and three exposed plots. The exposed plots had plastic strips between
the sampling rows to simulate the mulching effect of the plastic tubes
of the air-exclusion system. Three additional soybean plots were
installed without the plastic strips at monitors 6, 9, and 10 to deter-
mine whether plastic mulching had an effect on growth and yield.
-------
- -19-
Weed and Pest Control
The plots were kept free of weeds throughout the growing season
by hand weeding and hoeing. Pests were controlled as necessary. Soy-
beans were sprayed biweekly with either Malathion or a broad-spectrum
insecticide containing Carbaryl, Meta-Systox, and Kelthane to control
the Mexican bean beetles, stink bugs, and grasshoppers. Sprays were
applied with either a backpack sprayer or a high-pressure orchard
sprayer, according to recommendations of the manufacturers.
Seed Harvest and Processing
Soybeans were threshed by row (including the two inner buffer
rows) in the field with a plot thresher. Considerable differences
existed between and within sites in reaching harvest maturity.
Harvesting dates are presented in appendix E.
Beans were cleaned by hand to remove soil particles and plant
debris. The bean samples were subsequently dried in a forced-draft
oven at 70°C to constant weight, placed in a desiccator to cool, and
then weighed.
Bean size was determined by weighing 10 replicate samples of 100
beans from each of the sampling rows.
ENVIRONMENTAL MONITORING AND OBSERVATIONS
S02 was measured continuously at each site with a Philips S02
monitor. 03 was measured at monitor 9 with a chemiluminescent Og
monitor from June 22 to September 21 to determine whether its con-
centration was of sufficient intensity to cause injury or syner-
gistic effects. Temperature and relative humidity were measured
with hygrothermographs at monitors 6, 9, and 19 and at the TVA
mountain meteorology station in the vicinity of monitor 9. Solar
radiation was measured at the TVA meteorology station located about
1.6 km south of the Widows Creek Steam Plant. Temperature, relative
humidity, and solar radiation were measured because of their influence
on the sensitivity of plants to S02. Precipitation gauges were
installed at all sites except site 22 to record precipitation;
precipitation measurements from the Sand Mountain Experiment Sub-
station were used for site 22. These measurements were made to
determine differences in moisture levels among sites and as an
indicator of moisture stress.
PLANT OBSERVATIONS
Measurements of plant growth, development, symptom expression,
and incidence of insect activity and plant diseases were made each
week beginning the first week in July and ending the second week in
September. Weekly observations were made to compare visible S02
injury with various stages of plant development, subsequent reductions
in yield, if any occurred, and S02 exposures.
-------
-20-
Observations were taken according to the following procedure.
Two sections, 0.61 m long, from each of the three inner rows in each
soybean plot were used as observation points every week. One section
was randomly selected at the beginning of the growing season and was
constant for the remainder of the growing season. The second section
was randomly selected each week. Observations were made by trained
individuals on a rotation basis. A team of two individuals made
observations each week, with one member of each team having been a
member of the team of the previous week. This procedure provided
continuity in observation from week to week.
ANALYSIS OF DATA
Data for this study were analyzed by one-way analysis of variance
(ANOVA); the stepwise regression procedure; scatter plots of measured
and estimated variables; correlation coefficients; and multiple regres-
sion analysis, including least-squares and "robust" regression. The
analyses were performed with two interactive computer packages of
statistical procedures, the Statistical Analysis System (SAS76) (18)
and MIT-SNAP (19).
Differences in factors that might confirm the presence of S02 in
the plots (percentage foliar chlorosis, percentage total foliar sulfur)
and factors that ultimately might be the mechanism for yield reduction
caused by S02 (final plant height, bean, weight) were compared using
ANOVA. Yield differences among treatments also were tested in this
manner. To determine the effect of S02 on yield using ANOVA, all
other variables affecting yield must be controlled. Since this was
not the case, further analyses were necessary to isolate the effect
of S02 on yield. To accomplish this, regression analysis was used to
obtain a model expressing the size of the effect of any limiting
variables, including S02, on yield. Because of the number of factors
that might influence the outcome of a large field experiment of this
type, it was necessary to identify which factors were most important
among all the factors for which data were available in describing the
outcome and use of these variables in regression.
Several techniques were used to identify these variables. The
stepwise regression procedure from SAS76, which is capable of selecting
a model containing a combination of variables from a large group of
independent variables, was used to provide a basis for selecting
variables to be used in multiple regression. Scatter plots were
generated by MIT-SNAP to permit examination of the relationship between
the measured and estimated variables. These plots were useful in
identifying variables influencing yield; variables that should be
transformed to provide a better fit in the regression analysis; and
interactions between independent variables. Correlation coefficients
obtained from SAS76 provided additional information on which variables
were most highly associated with yield and on the degree of interaction
of any of the independent variables. The variables selected from the
techniques and the form of expressing the variables were subsequently
used in the multiple regression analyses (least-squares and "robust").
In multiple regression, the size of the dependent variable is
assumed to equal the sum of the influences of several independent
variables plus a random error. Consider n observations, each containing
-------
-21-
a value for the dependent variable (denoted by y., where i = l,...,n)
and values for p independent variables (denoted ^y x.., where j = l,...,p
and i = l,...,n). In this study, the y. are the yields and the x. .
include soil fertility measurements and1dummy variables for the treat-
ments. Further, consider the objective of choosing the coefficients p.,
where j = l,...,p, so that the prediction equation, •*
y. = z x..p., (i)
1 j=i XJ J
gives predictions y. close to the measured values of the dependent
variable y.. The p coefficients quantify the influences of the
independent variables on the outcome.
There are various criteria for judging how close the predictions
y. are to the measured values y., one of which is the basis for ordinary
multiple regression. The choice of criterion determines the p coeffi-
cients. Furthermore, the choice of criterion determines how well the
residuals, y. -.y., where i = l,...,n, can be used to choose among
prediction equations with different independent variables. This choice,
in turn, affects all the p coefficients (20).
The criteria used to compare the predictions and the measured values
are functions of the residuals that have the following form,
I P(r./d), (2)
i=l
where
P
r. = y. - I x..p., (3)
- ^ j=1 -Si'
d = scale factor,
p(r./d) = function of r./d.
Two choices of p are to be discussed. They are shown in Figure 5 and
are given by
r\
least-squares: p(u) = u /2 (4)
and 9
u /2, if |u|<0.5
Huber: p(u) = (5)
0.5|u|-0.125, if |u|>0.5.
See Denby and Mallows (21) for a discussion of Huber.
-------
-22-
The most notable difference among the choices of p is the relative
importance placed on large and small residuals. Figure 5 shows that
least-squares places more emphasis on large residuals than does Huber.
Some qualitative differences among the p coefficients and the
residuals that minimize each of these criteria are now clear. Compared
with least-squares , Huber gives predictions that are closer to the
"easy- to-fit" points and further from "hard-to-fit" points. Thus,
Huber gives both more small residuals and larger large residuals.
Further, the p coefficients are influenced less by the "hard-to-fit"
points . Huber gives residuals that are easier to use in choosing the
prediction equation. Missing independent variables and independent
variables that would fit better if reexpressed are diagnosed by plot-
ting residuals vs. the variable in question. When the poorly fitting
residuals are larger, they are easier to see.
These same properties of robust regression, as well as a computa-
tional algorithm, can be obtained by considering the equivalent for
Huber of the normal equations for weighted least-squares. If for each
observation the dependent variable has a different variance given by
a.2, where i = l,...,n, then weighted least-squares is more appropriate
tnan ordinary least-squares. In this case, the criterion to be minimized
is
n 22
I r /a , (6)
and the p coefficients are computed by solving the equation,
p n n
I (I x x w ) p = I x y w (7)
k=l i-1 1J 1K x k i=l 1J x 1
2
where w. = I/a. . For the two criteria discussed, the p coefficients
satis fyxequation 7, where w. = w(r./d) and
least-squares: w(u) = 1 (8)
or
1, if |u|<0.5
Huber: w(u) = (9)
0.5/ u , if |u|>0.5.
With Huber, the weights are data-dependent; that is, they depend on the
size of the residual. In weighted least-squares, the higher the variance
of an observation, the smaller is its weight and, consequently, the
smaller is its influence on the p coefficients. With Huber, the larger
the residual, the smaller is its weight and, consequently, the smaller
is its influence. For Huber, equation 7 must be solved iteratively.
An initial set of residuals is obtained, the corresponding weights are
computed, and equation 7 is solved with these weights. This produces a
new set of residuals on which to base the next iteration.
-------
-23-
LEAST
SQUARES
-2.0
Figure 5. Comparison of effects of residuals on the size of
the criteria: least-squares and Huber.
-------
-24-
Consider finally the choice of scale factor. Denby and Mallows (21)
discuss a procedure for choosing the scale factor for use with Huber.
This procedure is followed in the analysis of the Sand Mountain plots.
The scale factor that results is d = 4.6, which is equivalent to h = 2.3
in the paper by Denby and Mallows.
-------
-25-
SECTION 4
RESULTS AND DISCUSSION
SURFACE AND SUBSURFACE CHARACTERISTICS OF EXPERIMENTAL SITES
The results of the measurements of depth to bedrock and relative
elevation (appendix F) illustrate the greatest difficulty in selecting
large experimental sites (i.e., obtaining an area that is large enough
to accommodate the experimental design and that is uniform in surface
and subsurface characteristics). An example of the three-dimensional
output generated by computer graphics is presented in Figures 6a and
6b. These two figures show that, although the surface of the site
(Figure 6a) appears uniform, the soil depth (Figure 6b) is quite
variable. Because depth of the soil mass influences the water storage
capacity, it is important to account for these differences in an
analysis of within-site yield variability, particularly when rainfall
is abnormal.
Response analysis of surface and subsurface characteristics was
also useful in identifying drainage problems. Drainage ditches
were installed at all sites during the winter and spring of 1977 to
facilitate surface runoff and prevent wet spots. However, because
internal drainage problems cannot be corrected, all plots (except
those planted to pines) are scheduled to be planted to soybeans in
1977 in a further effort to characterize the yield potential of each
plot and to determine the present degree of uniformity in growth and
yield at each site.
EFFICIENCY OF AIR-EXCLUSION SYSTEM
Smoke and Fog Generator Tests
The results of these tests indicate that the air-exclusion system
was least efficient when no plants were present. The smoke flowed
around the air bags, and some smoke became trapped close to the ground
between the bags. However, during the second test, when a canopy had
formed, the system appeared more efficient at removing smoke from the
plot. Although the smoke penetrated the canopy, it was quickly expelled
from the top and sides of the plot. Smoke was expelled vigorously
directly over the air bags and laterally along the edges of the plot
below the plant canopy. Figure 7 shows penetration of the smoke into
the plot. Figure 8 shows residual smoke within the plot immediately
after the fog generator is turned off. Figure 9 shows the smoke being
cleared from the plot directly over the air bag.
SOg Exposure Tests
The purpose of the S02 tests was to quantify the efficiency of
the air-exclusion system. The first test showed that, when a concen-
tration of 3670 |Jg/m3 is present at the edge of the plot, the concen-
tration inside the plot, at the center and just below the canopy, ranges
-------
-26-
200
100
0
Figure 6a. Three-dimensional view of surface characteristics at
site 10 (appendix A gives complete plot layout).
Figure 6b. Three-dimensional view of subsurface characteristics at
site 10 (depth to bedrock).
-------
I
hO
Figure 7. Smoke from fog generator penetrating plant canopy.
-------
CD
I
Figure 8. Residual smoke within plant canopy immediately
after fog generator is turned off.
-------
Figure 9. Smoke from fog generator cleared from plant canopy
after fog generator is turned off.
-------
-30-
between 525 and 788 pg/rn3, an 83 percent reduction. However, it
appears that some of the 80% becomes trapped below the canopy, and
the concentration increases slightly. At 15 cm above ground, the
concentration in the plot at several points within 3 m from the edge
of the plot ranged from 655 to 850 (Jg/m3. The concentration at 15 cm
above ground decreased with increasing distance from the source of
anhydrous S02. Concentrations measured along the edge of the plot
inside the air barrier indicate that there is lateral transport of
S02 below the canopy. At the farthest point from the source, in the
southwest corner of the plot, the concentration at 30 cm above ground
was 262
The S02 test that was performed with the fiberglass barrier (30 cm
high) placed between the outer and inner buffer rows and at the north
and south ends of the plot showed similar trends. The S02 concentra-
tion above the plot was 1100 pg/m3. The concentration in the center
of the plot, 5 m from the source and within the canopy, was 158 |Jg/m3,
a reduction of 85 percent. At 4 m from the source, the concentration
within the canopy was 680 |Jg/m3. At that same distance from the source,
but at 7.5 cm above ground, the concentration was 324 jjg/m3, indicating
that some S02 may become trapped below the canopy. There is indication
that some lateral transport occurred in this test also because the con-
centration in the southeast corner of the plot (farthest point from the
source) was 420
The second part of the test, with the air bags reduced to half their
normal length, was complicated because of corresponding sampling points
of the testing area that were closer to the source (i.e., the center of
the plot in the first part of the test was 5 m from the source, whereas
the center of the smaller plot in the second part of the test was 3 m
from the source). Accordingly, the S02 concentration in the center of
the plot climbed to 815 jjg/m3. Trapping and lateral transport was also
indicated by a concentration of 5260 |Jg/m3 below the canopy at the edge
of the plot. When the air bags were inverted to increase air pressure
below the canopy, the concentration within the canopy remained at 1280
|Jg/m3 .
The results of these tests should be interpreted with caution.
Although an 83 to 85 percent reduction in S02 concentration was shown
for the air-exclusion system, the magnitude of that reduction may result
from the high initial concentration that was used in the tests. Also,
the efficiency of the system undoubtedly varies with wind velocity.
Because the wind velocity did not vary significantly during the tests,
an evaluation of that variable is not possible.
Sulfation (Huey) Plate Tests
Sulfation plates were also used to quantify the efficiency of the
air-exclusion system. Results of the analyses are presented in Figure 10'.
The average sulfation rate for the exposure period indicates that S02 in
the exposed plots exceeded that of the air-exclusion plots by 32 percent
at the 25-cm level, 63 percent at the 60-cm level, and 40 percent at the
-------
AIR-EXCLUSION PLOT
.50r-
7/15-8/12
8/12-9/16 9/16-10/8
EXPOSED PLOT
.50i-
.40
to
•o
•^»
CM
U
O
O
.30
.20
.10
7/15-8/12 8/12-9/16 9/16-10/8
i
U)
Figure 10. Monthly sulfation rates on three air-exclusion and
two exposed plots at monitor 9.
-------
-32-
120-cm level. Because the highest level (120 cm) was above the canopy
throughout the exposure period, a large difference was not expected
because exposure should have been about the same for both types of plots.
Likewise, the difference at the shortest height (25 cm) was not expected
to be large because the aerodynamic effect of the canopy should exclude
some of the pollutant, and the amount of S02 that becomes trapped close
to the ground under the canopy would be similar in both types of plots
as indicated by the aforementioned S02 exposure tests. The difference
between the two types of plots at the 60-cm level is of interest because
the smoke tests have shown that the air-exclusion system is effective in
clearing smoke from the region. The percentage difference may be con-
servative because the plates were above the plant canopy for a brief
period of time.
Foliar Sulfur Analysis
The sulfur content of foliage of soybeans (Table 1) grown on the
exposed plots at sites 6 and 9 was significantly higher than that of
soybeans grown on the air-exclusion plots at the same sites. When the
variance for treatments at sites 6, 9, and 10 was pooled, the foliar
sulfur content was significantly higher for exposed plots as compared
with that for air-exclusion plots.
TABLE 1. SULFUR CONTENT OF SOYBEAN FOLIAGE ON AIR-EXCLUSION
AND EXPOSED PLOTS AT MONITORS 6, 9, 10, AND 22
Total sulfur (%)a
Monitor No
6
9
10
22
Air-exclusion plots
0.170
0.150
0.157
0.127
Exposed plots
0.227?;
0.177b
0.157
0.147
a.
b.
Each value
Difference
5 percent
is an average of three plots.
between treatment and control is
level .
significant at the
These results would seem to indicate a much lower efficiency rating
for the air-exclusion system than indicated by the sulfation plates or
the S02 exposure tests. However, because the foliar samples were collected
from the inner buffer rows of the plots, the measurements may not be
representative of sulfur levels present in the interior of the plots.
Also, the contribution to the foliage of sulfur from the soil environment
may be so great as to mask the contribution from the atmosphere.
-------
-33-
ENVIRONMENTAL MONITORING
Sulfur Dioxide
During the 1976 growing season there were 14 occasions on which
S02 concentrations equaled or exceeded 1306 MS/1"3 (0.5 ppm) 1-h average
during daylight hours at sites 6, 9, and 10 near the steam plant (Table 2);
there were six occasions on which concentrations equaled or exceeded
1306 (Jg/™3 f°r 3 h. During the pod-filling stage of growth, when soy-
beans should be most sensitive to yield reduction from foliar injury,
there were four occasions on which concentrations equaled or exceeded
1306 |Jg/m3 3-h average.
The highest total exposure above 262 JJg/m3 during daylight hours
was registered at monitor 10 (Table 3), whereas the highest 1-h average
concentration was registered at site 6 (Table 4). The only occurrence
of foliar injury that was great enough to be distinguished from that
caused by other factors in the environment was observed for the control
plots at site 6 after an exposure on August 13 during which the maximum
1-h average 803 concentration was 2351 MS/1"3 (0.90 ppm). No concentra-
tions of 262 (Ji/™3 or more were registered at site 22.
TABLE 2. NUMBER OF OCCURRENCES OF SULFUR DIOXIDE EXPOSURES
EQUAL TO OR GREATER THAN 1300 |Jg/m3 1-H OR 3-H AVERAGE
1-h average 3-h average
(> 1300 Mg/m3) (> 1300 Mg/m3)
Site Growing season After Aug. 12 Growing season After Aug. 12
(occurrences) (occurrences) (occurrences) (occurrences)
6
9
10
19
4
1
3
6
4
1
1
3
2
0
1
3
2
0
1
1
-------
-34-
TABLE 3. TOTAL EXPOSURE TO SULFUR DIOXIDE CONCENTRATION EQUAL
TO OR GREATER THAN 262 |Jg/m3 DURING DAYLIGHT HOURS3 FOR
1976 GROWING SEASON
Monitor
Exposure Total S02
(h) (ppm) (Mg/ra )
ii
Dose
1 (|Jg m"3 h"1)
6
9
10c
19
65
67
73
16.68
12.12
16.21
43,703
31,754
42,470
2,854,630
2,127,518
3,100,310
a. 6:00 a.m.-7:00 p.m. CDT.
b. Cumulative total above 262 \Jtg/m3.
c. Not calculated because plot was abandoned.
TABLE 4. MAXIMUM 1-H AND 3-H AVERAGE SULFUR DIOXIDE
CONCENTRATIONS RECORDED AT STUDY SITES 6, 9, 10, AND 19 DURING
1976 SOYBEAN GROWING SEASON
Monitor
6
9
10
19
Date
8/13
8/21
8/14
7/24
Maximum 1-h
average
(pg/m3)
2,351
1,515
2,194
10,318
Maximum 3-h
average
(pg/m3)
1,463
1,097
1,515
4,519
Ozone
Ozone levels were measured at site 9 from July 1 until
September 21, 1976. Some data were lost because of a malfunction
in the recorder. Data are reported in central daylight saving time
(CDT). In reading the charts and tabulating the data, the average
concentration was recorded for each hour beginning at 12 midnight.
The results show that the national primary-secondary air quality
standard for 03 of 160 |Jg/m3 was exceeded on 27 days. The monthly
average of hourly 03 concentration and the monthly average of these
daily maximum concentrations are presented in Table 5.
-------
-35-
TABLE 5. MONTHLY AVERAGE OF HOURLY OZONE CONCENTRATION AND
MONTHLY AVERAGE OF DAILY MAXIMUM OZONE CONCENTRATION
Monthly average concentrations
Period
7/1 to 7/31
8/1 to 8/31
9/1 to 9/21
No. of days
monitored
29.5
29.1
20.8
Hourly ozone
(|Jg/m3)
115.6
88.2
78.2
Daily maximum ozone
(pg/m3)
150.9
135.2
111.7
The highest concentration, which was recorded on five separate days,
was 235 |Jg/m3. These results indicate that the potential for 03 injury
to soybeans exists in the Widows Creek area. A concentration of 196 to
235 |Jg/m3 has been shown to be injurious to a number of plant species.
FOLIAR INJURY, GROWTH, AND YIELD OF SOYBEANS
Foliar Injury
The amount of foliar injury that occurred on the study plots during
the 1976 growing season (appendix G) was small and variable, averaging
about 2 percent or less for all sites during the pod-filling stage of
growth (Table 6). The only chlorosis that could be confidently attributed
to exposure to S02 occurred at site 6 on August 13 during an exposure in
which a maximum 1-h average S02 concentration of 2351 MS/1"3 was recorded.
The percentage of leaf area exhibiting chlorosis for the exposure averaged
1.0 and 6.4 percent for the air-exclusion and exposed plots, respectively.
The maximum percentage chlorosis for samples within treatments was 8.5
and 30.0 for the air-exclusion and exposed plots, respectively. The
levels of chlorosis for the two treatments were significantly different
at the 0.05 level of probability. Differences between treatments for
the other sites or for other exposures were not significant. The data
indicate a trend toward protection of the plots by the air-exclusion
systems.
-------
-36-
TABLE 6. FOLIAR CHLOROSIS OF SOYBEAN ON STUDY SITES DURING
THE POD-FILLING STAGE OF GROWTH
Total leaf area affected (%)
Maximum average Average of ,
Site weekly observation weekly observations
Air-exclusion 0.9 1.0
Exposed 6.4 2.1
Air-exclusion 0.2 0.3
Exposed 2.1 0.9
10
Air-exclusion 0.5 0.2
Exposed 0.3 0.2
22
Air-exclusion 0.3 0.5
Exposed 0.1 0.2
a. Based on maximum for exposed plot.
b. Four weeks of observations.
Height Growth
There were no significant differences in height growth between treat-
ments at sites 6, 10, and 22 (Table 7). The difference at site 19 was
attributed to the extremely shallow soils of the exposed plots. This
site was abandoned, and a new study site was established at monitor 25
at the end of the growing season.
-------
-37-
TABLE 7. AVERAGE HEIGHT GROWTH OF SOYBEANS
BY TREATMENT FOR ALL SITES
Height growth (cm)
Site
No.
6
9
10
19
22
Air-exclusion
104
108
95
110
96
Exposed with
plastic
108
110
98
53
98b
Exposed without
plastic
107b
113b
95b
a.
b.
Height measurements taken the week of August 25 .
Underscored indicates treatments at the same site not
significantly different from each other at the 5 percent level.
Yields
Analysis of variance indicated no significant differences in yield
(Table 8; appendix H) between air-exclusion plots and their comparable
exposed plots (with simulated plastic tubes) at any of the sites except
6 and 19. However, except for the exposed plots without plastic treat-
ment at site 9, higher yields were produced on all sites with the air-
exclusion treatment. The exceptions at site 9 may have resulted from
air-exclusion plots 1 and 2 having been cropped in soybeans in previous
years; plots 10, 16, and 17 were all located on new ground. Jones et
al. (10,11) have observed that continuous cropping in soybeans results
in reduced yields. The differences at site 19 were caused by the shal-
low soil and were reflected in the much reduced height of the plants,
as discussed previously. Although significant differences in yields
occurred between exposed plots with plastic and those without plastic
at sites 9 and 10, there was no consistent relationship in differences.
The differences are attributable to edaphic factors: the pH of exposed
plots without plastic was very low on plots at site 10, which can result
in reduced yields, and the plots of the same treatment at site 9 were
all on new ground, which can result in increased yields.
The results at site 22 indicate that the air-exclusion system does
not adversely affect the growth and yield of soybean plants. However,
the yields at the remote site were considerably lower than those for
the sites in the area exposed to emissions from the power plant. The
lower yields probably reflect the generally lower soil fertility at the
site and less precipitation during the pod-filling stage.
-------
-38-
TABLE 8. AVERAGE SOYBEAN YIELD BY TREATMENT
FOR ALL SITES
Yield (kg/ha)a
Site
No.
6
9
10
19
22
Air-exclusion
3,294
3,006
3,615
3,120
2,707
Exposed with
plastic
2,881
2,910b
3,538b
1,149
2,554
Exposed without
plastic
2,903b
3,320
2,743
a. Oven-dry weight.
b. Underscored indicates treatments at the same
site are not
significantly different from each other at the 5 percent level.
Bean weights were compared by ANOVA for the treatments at site 6
to determine whether the significantly lower yields on the exposed
plots were reflected on smaller beans; the test showed no difference
in averages (Table 9).
TABLE 9. AVERAGE WEIGHT OF 100 SOYBEAN SEED
BY TREATMENT AT MONITOR 6
Sample weight
Treatment (g)a
Air-exclusion plots 14.31
Exposed (plots with plastic) 14.32
Exposed (plots without plastic) 13.73
a. Each mean is an average of 10 replicate samples of 100 beans
each from three sampling rows from three plots.
Because the plants at site 6 were exposed to higher concentrations
of S02 and exhibited more chlorosis, which could be attributed to S02
exposure, it would appear that the reduced yields of the control plots
were caused by S02; however, the data are not conclusive. Part of the
difference in yields within site 6 may have been caused by lower P
-------
-39-
levels in the soils of the exposed plots. The difference between the
exposed plots at site 6 and the plots at the other sites also suggests
a yield reduction, but this may be a result of site differences not
reflected in the available measurements. Furthermore, there was little
difference between the average percentage chlorosis over the entire
pod-filling stage for the exposed plants compared with that for the
air-exclusion plots at site 6 or other sites.
The plots on Sand Mountain differ in ways other than their exposure
to the plume from the Widows Creek Steam Plant. These differences—the
most important of which are in the P, K, Mg, and organic matter contents
of the soil—are suspected to be additional causes of differences in
yield. To separate these effects on the yields, a model for the yields
has been fit to the data.
The model contains seven predictor variables, two variables that
account for conditions unique to particular sets of plots and five
variables that account for the soil fertility. Foliar sulfur measure-
ments , the S02 monitoring results, and the yield measurements themselves
suggest that the exposed plots at site 6 were affected by the power plant;
thus, a variable EXP is included to give the difference in yield resulting
from this exposure. The possibility exists that there are conditions
affecting the yield at site 6 that are not reflected in the measured
factors. A variable SITE6 is included to give the difference resulting
from these conditions. The inclusion of this variable ensures that the
coefficient of EXP reflects only the difference between the three pro-
tected and six unprotected plots at site 6 and not the difference between
site 6 and the other two sites. The soil-content variables included are
the logarithms of the P, K, and Mg and the percentage of organic matter.
Both linear and quadratic terms in the logarithm of the P are included.
The estimated effect of the exposure is -222 kg/ha with -a standard
error of 147 kg/ha. The estimated effect of the unmeasured conditions
peculiar to site 6 is -207 kg/ha with a standard error of 150 kg/ha.
These standard errors are not small enough for these effects to be
clearly distinguished from the random variability in the yields. If
no variable were included to account for the peculiarities of site 6,
then the effect of the exposure would be significant at the 0.01 level.
In other words, comparison of the exposed plots at site 6 with all other
plots on Sand Mountain shows a reduction in yield not accounted for by
the soil fertility variables included in the model.
The model is given by
YIELD (kg/ha) = 4448 - 222 EXP - 207 SITE6 - 1410(LP-2.14)2 , Q,
- 1602 LK + 972 LMG + 391 OM, ^ J
where
LP = base 10 logarithm of the P, kg/ha,
LK = base 10 logarithm of the K, kg/ha,
LMG = base 10 logarithm of the Mg, kg/ha,
OM = percentage organic matter.
-------
-40-
The variable EXP has the value 1 for the exposed plots at site 6 and
the value 0 elsewhere. The variable SITE6 has the value 1 for the plots
at site 6 and the value 0 elsewhere. According to the model, the best
P content is 235 kg/ha. The K content is so high that for these plots
an increase in K has a negative effect on yield. Increases in either Mg
or organic matter have a positive effect on yield. The term LMG - LK
represents an interaction term that can also be expressed as base 10
logarithm of the Mg/K ratio; that is, at low levels of Mg, increased K
levels have a negative effect on yield. Stepwise regression analysis
of the data (appendix I) gave a model that was similar to the least-
squares model in explaining variations in yield among the plots; that
is,
YIELD (kg/ha) = 2882 - 279 EXP - 337 SITE6 - 972(LP-2.14)2
+ 879 L Mg/K + 448 OM (r2 = 0.74).
However, the quadratic term in P was dropped from the regression model
because it was not significant at the 10 percent level. The final model
was
YIELD (kg/ha). = 2878 - 300 EXP - 316 SITE6 + 936 L Mg/K
+ 434 OM (r2 = 0.72).
The interaction term (L Mg/K) was the most significant variable
explaining variation in yields among all the experimental plots. The
Mg/K interaction has been reported by other investigators (22). The
analyses also indicated that, for soils with high OM, larger amounts
of Ca were required to raise the pH than for soils with lower levels
of OM.
The standard errors for equation 10 are 295 for the linear term
in LP, 893 for the quadratic term in LP, 562 for LK, 162 for LMG, and
162 for OM. The standard error for the coefficient of the quadratic
term in P indicates that it is not significantly different from zero.
Excluding this variable would have two consequences: (1) two large
negative residuals appear that seem attributable to excess P; (2) for
site 6, the effect of increasing P on yield becomes less, and, to
compensate, the effect of the exposure becomes large enough to be
significant at the 0.05 level.
The goodness of fit of the model is determined by examining the
residuals, the differences between the actual yields, and the yields
predicted by the model. First, the residuals are plotted vs. variables
not in the model to determine whether these variables should be included.
Second, the residuals are plotted vs. the variables in the model to
determine whether a transformation would give a better fit. Third, the
residuals are plotted vs. the fitted yields to determine whether the
fit is good at the ends of the range of yields. Because the model is
a least-squares fit, the residuals, considered as a vector, are perpen-
dicular to the variables in the model. For this reason, more sensi-
tivity is obtained by plotting vs. variables that have been adjusted to
remove their dependence on the variables in the model. For variables
not in the model, the usual correlation formula applied to the points
plotted gives partial correlations.
-------
-41-
Figures 11 through 13 show the residuals plotted vs. the soil pH,
the base 10 logarithm of the Ca content of the soil, and the index of
injury due to disease and insects (each variable adjusted to remove its
dependence on the variables in the model). Figure 11 suggests that a
weak positive dependence on pH holds except for the four points furthest
right, which have pH values of 6.7, 6.7, 6.6, and 6.5. The negative
residuals for these four plots may be explained by something other than
excessive pH. Thus, pH was not included in the model. Figure 12 suggests
that a weak positive dependence on Ca holds except for a few points.
Figure 13 shows no dependence on the insect-disease index.
Figures 14 through 17 show the residuals plotted vs. LP, LK, LMG,
and OM, each adjusted for the variables in the model except itself.
Figure 14 shows no dependence. Figure 15 shows that the most negative
residual has very low K, perhaps so low that the negative dependence
on K that is part of the model does not hold. Results of a decrease in
the influence of this point on the model are discussed below. Neither
Figure 16 nor Figure 17 shows any dependence.
Often the poorest fit occurs for the largest or smallest predicted
values. To allow this to be checked, Figure 18 shows the residuals
plotted vs. the fit. It shows no evidence of poor fit at the ends of
the range.
Some data points may not fit the model very well, not because of
the random variability in the yields, but because of effects not included
in the model. Least-squares fits are unduly influenced by such points.
The procedure suggested by Denby and Mallows (21) gives a sequence of
robust fits that are influenced less and less by such points.
-------
-42-
0.00
a
3'
YRES
(kg/ha)
-500.00
-1.00
0.00
PHRES
1.00
Figure 11. Yield residuals (YRES) vs. soil pH residual (PHRES)
adjusted for variables in model.
0.00
YRES
(kg/ha)
-500.00
-0.40
0.00
LCARES
(log kg/ha)
0.40
Figure 12. Yield residuals (YRES) vs. log calcium residual
(LCARES) adjusted for variables in model.
a. number of multiple observations at same value.
-------
-43-
0.00
YRES
(kg/ha)
-500.00
-40.00
0.00
INDRES
I
40.00
Figure 13. Yield residuals (YRES) vs. insect-disease index
residual (INDRES) adjusted for variables in model.
0.00
YRES
(kg/ha)
500.00
• • •
• • •
• •
2°
• • • • •
•
• • fl • •
•
•
—
1 \
-0.40 0.00 0.40
LPRES
(log kg/ ha)
Figure 14. Yield residuals (YRES) vs. log phosphorus residual
(LPRES) adjusted for other variables in model.
a. number of multiple observations at same value.
-------
-44-
0.00
YRES
(kg/ha)
-500.00
-0.20 0.00 0.20
LKRES
(log kg/ha)
Figure 15. Yield residuals (YRES) vs. log potassium
residual (LKRES) adjusted for other variables
in model.
0.00
YRES
(kg/ha
-500.00
I
-0.40
0.40
0.00
INGRES
(log kg/ha)
Figure 16. Yield residuals (YRES) vs. log magnesium
residual (INGRES) adjusted for other variables
in model.
-------
-45-
0.00
YRES
(kg/ha)
-500.00
-0.40
0.00
OMRES
(Percent)
0.40
Figure 17. Yield residuals (YRES) vs. percent organic matter
residual (OMRES) adjusted for other variables in
model .
0.00
YRES
(kg/ha)
-500.00
J_
2800.00
3200.00
YFIT
(kg/ha)
3600.00
Figure 18. Yield residuals (YRES) vs. fitted yields (YFIT).
-------
-46-
The model given by robust fitting is
YIELD = 4928 - 256 EXP - 236 SITE6 - 1332 (LP-2.14)2
- 1602 LK + 972 LMG + 391 OM.
This model does not differ from the least-squares model enough to change
the conclusions.
Often the most valuable output from a robust fit is a set of residuals
containing unusually large values for the unusual data points. In this
case, the residual for the plot with lowest K changes from -352.1 to -465.8.
The two residuals that decreased the most were those for the two plots with
lowest K, and the two residuals that increased the most were those for the
two plots with highest K. This shows the effect of decreasing the influence
of the observation with lowest K. Additionally, the robust residuals do
suggest that some other plots are unusual, perhaps because of soil varia-
tions within sites. The least-squares and robust residuals are shown in
Table 10. This information is useful in designing future experiments for
these plots to minimize the bias that may be introduced by differences in
site factors and location.
-------
- -47-
TABLE 10. LEAST-SQUARES AND ROBUST RESIDUALS
Site
6
6
6
6
6
6
6
6
6
9
9
9
9
9
9
9
9
9
10
10
10
10
10
10
10
10
10
Plot
1
2
3
13
15
16
9
20
22
1
2
7
15
13
14
10
16
17
1
2
3
4
5
6
17
18
19
Least-squares
34.0
5.2
28.7
150.4
-42.9
-140.8
91.7
74.1
15.8
56.0
-352.1
-88.2
148.0
-311.2
-163.9
245.2
209.9
266.4
-69.1
199.9
230.80
-162.9
126.1
70.0
169.9
18.3
138.9
Robust
2.3
4.7
-7.0
191.9
3.5
-115.9
65.0
-85.4
-1.8
-8.2
-465.8
-85.4
134.6
-239.3
-123.5
223.5
186.7
296.7
-72.6
220.6
246.9
-156.8
-103.6
-39.8
-176.1
50.3
114.1
-------
-48-
REFERENCES
1. U.S. Environmental Protection Agency. Effects of Sulfur Oxides in
the Atmosphere on Vegetation. Air Quality Criteria for Sulfur
Dioxide (Revised Chapter 5), Ecological Research Series EPA-R3-73-030,
1973. 43 pp.
2. Zahn, R. Effects of Sulfur Dioxide on Vegetation: Results of
Gas Exposure Experiments. Staub., 21:56-60, 1961.
3. Katz, M., and F. E. Lathe. Summary. In: Effect of Sulfur Dioxide
on Vegetation, National Research Council of Canada, 1939. N.R.C.
No. 815, pp. 429-447.
4. Barrett, T. W., and H. M. Benedict. Sulfur Dioxide. In: Recogni-
tion of Air Pollution Injury to Vegetation: A Pictorial Atlas,
J. S. Jacobson and A. C. Hill, eds. Air Pollut. Control Assoc.,
Pittsburgh, Pa., 1970. pp. C1-C17.
5. Jones, H. C., D. Weber, and D. Balsillie. Acceptable Limits for Air
Pollution Dosages and Vegetation Effects: Sulfur Dioxide. Presented
at the 67th Annual Meeting of Air Pollut. Control Assoc., Denver,
Colo., June 9-13, 1974. 31 pp.
6. Hill, G. R., Jr., and M. C. Thomas. Influence of Leaf Destruction
by Sulfur Dioxide and by Clipping on the Yield of Alfalfa. Plant
Physiol., 8:223-245, 1933.
7. Brisley, H. R., and W. Jones. Sulfur Dioxide Fumigation of Wheat
with Special Reference to Its Effect on Yield. Plant Physiol.,
25:666-681, 1950.
8. Brisley, H. R., C. R. Davis, and J. A. Booth. Sulfur Dioxide
Fumigation of Cotton with Special Reference to Its Effect on Yield.
Agron. J., 51:77-80, 1959.
9. Davis, C. R. Sulfur Dioxide Fumigation of Soybeans: Effect on
Yield. J. Air Pollut. Control Assoc., 22:964-966, 1972.
10. Jones, H. C., J. R. Cunningham, S. B. McLaughlin, N. T. Lee, and
Shirley S. Ray. Investigation of Alleged Air Pollution Effects on
Yield of Soybeans in the Vicinity of the Shawnee Steam Plant. TVA
Report E-EB-73-3, 1973. 36 pp.
11. Jones, H. C., Frances P. Weatherford, W. S. Liggett, Jr., and
J. R. Cunningham. Effect of Foliar Injury Caused by Exposure
to Sulfur Dioxide on Yield of Soybeans - Results of a Large-Scale
Field Investigation. Manuscript in preparation.
12. Quarles, J. R. Reevaluation of the Three-Hour SQ_ Standard.
Fed. Regist., 38(178):25680-25681, 1973.
-------
-49-
13. Mandl, R. H., L. H. Weinstein, D. C. McCune, and M. Keveny. A
Cylindrical, Open-Top Chamber for the Exposure of Plants to Air
Pollutants in the Field. J. Env. Qual., 2:371-376, 1973.
14. Howell, R. K., E. J. Koch, and L. T. Rose. Field Assessment of
Air Pollution Reduced Soybean Yield Losses. Araer. Soc. Agron.
Abstr., 1976. p. 84.
15. Todemann, A. R., and T. D. Anderson. Rapid Analysis of Total
Sulfur in Soils and Plant Material. Plant Soil, 35:197-200,
1971.
16. Huey, N. A. The Lead Dioxide Estimation of Sulfur Dioxide Pollution.
J. Air Pollut. Control Assoc., 18:610-611, 1968.
17. U.S. Public Health Service. Determination of Sulfates in Atmospheric
Suspended Particulates: Turbidimetric Barium Sulfate Method. Publ.
No. 999-AP-ll, 1965.
18. Barr, A. J., J. H. Goodnight, J. P. Sail, and J. T. Helwig. A Users
Guide to SAS76. Sparks Press, Raleigh, North Carolina, 1976. 329 pp.
19. Hoaglin, D. C., and R. E. Welsch. MIT-SNAP--An Interactive Data
Analysis System. Massachusetts Institute of Technology, Cambridge,
Massachusetts. Preliminary manual.
20. Beaton, A. E., and J. W. Tukey. The Fitting of Power Series, Meaning
Polynomials, Illustrated on Band-Spectroscopic Data. Technometrics,
16:147-185, 1974.
21. Denby, L., and C. L. Mallows. Two Diagnostic Displays for Robust
Regression Analysis. Technometrics, 19:1-13, 1977.
22. Mugwira, L. M., and K. I. Patel. Soybean Growth and Composition
as Affected by K, Ca, and Mg Rates and Corn Rotation. Commun.
Soil Sci. Plant Anal., 7:319-330, 1976.
-------
-50-
APPENDIX A
PLOT ARRANGEMENT AT EACH SITE AND DIRECTION FROM STEAM PLANT
-------
CO
i
Ul
LEGEND
O Essex Soybean Plot
T.
E
^-6.1 m-*\
37.2m2
Plot (Equipped with air-exclusion
system)
Figure Al. Plot arrangement, monitor 6.
-------
-52-
22
24
cc
o"§
LEGEND
Essex Soybean Plot
k
6
w~
•0
t
-6.1m-*|
37.2m2
Plot (Equipped with air-exclusion
system)
o
.M
60
4-1
O
3
60
•H
-------
221
POWER PLANT(3.2km)
N
A
4LF
10
4LF HLF 4LP
LEGEND
O Essex Soybean Plot
T
37.2m2
Plot (Equipped with air-exclusion
system)
LO
Figure A3. Plot arrangement, monitor 10.
-------
24
23
22
18
15
O
II
12
8
S02
MONITOR
f1!?! I1??!
r'TTi
LEGEND
O Essex Soybean Plot
•
*
37.2m2
Plot (Equipped with air-exclusion
system)
Figure A4. Plot arrangement, monitor 19.
-------
-55-
7
O
6
O
I
O
LEGEND
O Essex Soybean Plot
37.2m2
Plot
S02
MONITOR
CM
O
-p
•H
C
O
e
a
-------
-56-
APPENDIX B
SOIL CHARACTERISTICS, AVERAGE SEASONAL PRECIPITATION, AND TEMPERATURE
JACKSON COUNTY1
The soils at monitors 6, 9, and 10 are of the Hartsells fine sandy
loam type of the eroded, undulating phase. Most of this type of soil
occurs in relatively small areas in close association with other Hartsells
soils. The soil is similar to the undulating phase, but is more eroded.
One-fourth to three-fourths of the original surface soil has been eroded.
Both .external and internal drainage are good to excellent. The soil has
good moisture-absorbing and moisture-holding qualities and good tilth.
It is well suited to a wide range of crops and produces excellent yields
under good management. Sandstone is found at a depth of 35 to 75 cm.
The expected average annual yield per hectare under improved management
practices are 4.4 metric tons of soybean hay and 544 kg of lint cotton.
The soil at monitor 19 is of the Hanceville fine sandy loam type
of the eroded rolling phase. This type of soil occurs in scattered
areas on Sand Mountain. It is closely associated with Hartsells soils
and resembles them in texture and friability. However, it has a browner
surface soil and red subsoil. It is derived from weathered products of
sandstone and, in places, slate. One-half to three-fourths of the original
surface soil has been eroded. It has a 10- to 12.5-cm brown, fine sandy
loam surface soil and a red, friable, fine sandy clay subsoil. In places
plowing brings subsoil material to the surface. However, despite erosion,
the productivity of this soil has not been seriously impaired. It is very
responsive to good management, and the loose, deep subsoil absorbs and
retains moisture well. The expected average annual yield per hectare under
improved management practices are 4.2 metric tons of soybean hay and 390 kg
of lint cotton.
The seasonal average precipitation and temperature for the growing
season for Jackson County are presented in Table B.I.
TABLE B.I. SEASONAL AVERAGE PRECIPITATION AND TEMPERATURE FOR JACKSON
COUNTY
Precipitation Temperature
Season (cm)
Spring
Summer
Fall
37.5
35
22.5
15.5
19.4
15.4
From: Soil Survey - Jackson County, Alabama, Series 1941, No. 8.
Issued 1954. U.S. Dept. of Agriculture, Soil Conservation Service,
Alabama Dept. of Agriculture and Industry, Alabama Agricultural
Experiment Station. U.S. Government Printing Office, Washington, DC.
222 pp.
-------
- -57-
DEKALB COUNTY1
The soils at monitor 22 are in Hartsells fine sandy loam of the
eroded, undulating phase. With this type of soil, runoff is slow, and
internal drainage is rapid. The moisture-holding capacity is moderate.
This soil has been under cultivation for a number of years. Because
the surface soil and subsoil materials are friable, the plow layers
differ very little from the original surface soil. The depth to
bedrock ranges from 0.91 to 1.5 m. When dry, the plow layer varies
from light gray to very pale brown. The soil has very good to excellent
workability, excellent tilth, and very good moisture-absorbing qualities.
The expected average yields per hectare under improved management practices
are 6.7 metric tons of soybean hay and 705 kg of lint cotton. The ^seasonal
average precipitation and temperature for the growing seasons for DeKalb
County are presented in Table B.2.
TABLE B.2. SEASONAL AVERAGE PRECIPITATION AND TEMPERATURE FOR DEKALB
COUNTY
Precipitation Temperature
Season (cm) (°C)
Spring 37.6 14.9
Summer 35.0 24.6
Fall 22.5 15.2
From: Soil Survey - DeKalb County, Alabama, Series 1951, No. 3.
Sept. 1958. U.S. Dept. of Agriculture, Soil Conservation Service,
Alabama Dept. of Agriculture and Industry, Alabama Agricultural
Experiment Station, U.S. Government Printing Office, Washington, DC.
108 pp.
-------
-53-
APPENDIX C
PERCENTAGE ORGANIC MATTER,
pH, AND CONTENTS OF P, K, Mg, AND Ca OF THE SOILS OF THE EXPERIMENTAL SITES
Monitor No. Plot No.
6 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
9 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
PH
6.4
5.6
6.0
6.2
6.3
6.0
6.2
6.0
4.9
5.2
5.2
5.4
5.2
5.5
6.0
6.7
6.1
6.2
5.9
5.3
6.3
5.0
5.6
6.0
6.4
6.6
6.3
6.4
6.0
6.5
6.0
5.9
6.2
5.4
6.3
6.7
6.7
6.5
5.6
6.1
6.1
6.7
5.9
Organic
matter
(%)
1.46
1.86
2.24
2.44
2.92
2.85
2.99
3.33
2.37
2.28
1.80
1.58
1.88
2.06
1.36
1.64
2.16
2.44
2.68
2.15
3.12
2.14
2.55
3.16
1.64
1.98
2.40
2.23
1.76
1.61
1.79
2.20
1.98
1.84
1.72
1.50
1.50
1.90
2.08
1.83
1.40
1.43
1.92
Phosphorus
(kg/ha)
85
137
120
117
162
129
121
136
57
98
117
124
132
179
91
90
162
163
141
144
212
78
129
152
92
130
188
164
330
336
400
420
205
285
375
316
318
288
292
281
228
354
100
Potassium
(kg/ha)
221
203
222
271
304
273
268
285
240
229
211
204
190
228
164
176
208
238
264
192
358
191
239
284
119
104
181
192
181
228
200
292
181
170
234
262
263
298
223
167
169
181
273
Magnesium
(kg/ha)
254
159
141
132
163
177
495
181
121
117
106
187
47
122
185
167
192
186
198
132
193
103
144
185
336
30
42
42
39
52
41
49
39
36
38
62
62
73
41
36
36
43
49
Calcium
(kg/ha)
1455
1120
1670
1890
2250
2070
3100
2340
1545
1224
1189
1358
876
1050
1670
2380
1555
1980
1890
1825
2660
1568
1776
1970
2443
2480
2450
2390
2252
2410
1890
2080
2360
2340
2162
2420
2420
2387
1950
2320
2342
2285
1640
-------
-59-
APPENDIX C
(Continued)
Monitor No. Plot No.
20
21
22
23
24
10 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
22 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
pH
5.4
5.8
5.9
6.0
6.0
6.2
6.1
5.4
4.8
4.8
4.8
5.3
6.0
6.0
6.0
6.0
6.3
6.2
6.1
5.9
6.5
5.9
6.1
6.1
5.4
6.1
5.8
5.5
5.6
6.1
6.2
6.1
6.1
5.9
5.9
5.7
6.0
6.0
6.2
5.9
5.9
5.7
6.0
6.1
Organic
matter
(%)
2.47
1.82
1.54
1.54
1.58
1.78
1.47
1.30
1.61
1.42
1.59
1.43
1.61
1.82
1.78
1.72
1.96
1.65
1.60
1.74
1.65
1.50
1.57
1.61
2.07
1.75
1.78
1.68
2.20
0.96
1.17
0.96
0.98
0.89
0.92
1.04
1.27
0.98
0.83
1.03
1.03
1.03
1.13
1.01
Phosphorus
(kg/ha)
104
79
66
62
127
246
169
121
148
115
152
270
222
248
190
137
272
224
163
168
176
143
158
93
172
147
174
160
112
71
54
75
105
64
76
71
103
109
90
122
82
79
74
83
Potassium
(kg/ha)
176
144
140
144
208
232
212
172
183
170
194
252
212
250
234
220
310
255
196
278
258
184
252
190
308
203
220
208
270
244
204
215
275
226
250
242
347
378
214
415
228
208
202
215
Magnesium
(kg/ha)
61
29
28
28
215
228
198
110
49
29
26
83
42
35
52
39
53
67
105
93
64
200
215
209
135
25
144
116
122
54
70
71
85
71
52
49
106
78
53
103
47
40
55
53
Calcium
(kg/ha)
1195
1610
1390
1493
2690
1279
920
710
528
450
406
580
2442
1850
1955
2085
1580
775
620
1025
1195
813
1165
1232
835
1300
850
718
890
745
1050
987
1230
962
675
684
1436
1525
883
1600
693
613
870
824
-------
-60-
APPENDIX C
(Continued)
Monitor No. Plot No.
15
16
17
18
19
20
21
22
23
24
pH
6.1
6.0
5.9
5.8
5.6
5.7
5.9
6.1
5.6
5.7
Organic
matter
(%)
1.01
1.03
1.03
0.89
1.20
1.24
1.20
1.00
1.00
1.20
Phosphorus
(kg/ha)
83
62
93
67
74
28
81
92
69
62
Potassium
(kg/ha)
215
225
225
188
206
229
230
240
222
195
Magnesium
(kg/ha)
53
66
52
49
49
48
63
70
43
42
Calcium
(kg/ha)
824
789
736
738
675
774
833
885
701
566
-------
i -61-
APPENDIX D
LIST OF EQUIPMENT AND VENDORS
1. Anemometer
2. Charcoal
3. Chemiluminescent Ozone
Monitor
4. Dayton Wheel Blowers
5. Dyna Fog Smoke Generator
6. Krene Plastic Tubes
7. Sulfur Gas Analyzer,
Model SA 160-S
8. Philips S02 Monitor,
Model PW9700
Belfort Instrument Co.
Baltimore, Maryland
Ba rneby-Cheney
Columbus, Ohio
McMillan Electronics Corp.
Houston, Texas
W. W. Grainger, Inc.
Birmingham, Alabama
Westfield, Indiana
Livingstone Coating Co.
Charlotte, North Carolina
Meloy Laboratories, Inc.
Springfield, Virginia
Phillips Electronics Instruments, Inc.
Mount Vernon, New York
-------
-62-
APPENDIX E
PLANTING AND HARVESTING DATES FOR SOYBEANS DURING 1976
Monitor
6
9
10
19
22
Planting
date
5/27
5/26
5/26
5/27
5/27
Harvesting
date
10/21, 10/22
10/19, 10/21
10/27, 10/28, 11/3
11/4
10/18
-------
APPENDIX F
DEPTH TO BEDROCK, AVERAGE RELATIVE ELEVATION, AND PERCENT SLOPE OF PLOTS AT EXPERIMENTAL SITES
Monitor 6
Depth
to
Plot bedrock
Ho. (cm)a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
a.
b.
63
55
58
68
63
66
60
52
78
33
70
56
60
45
55
72
46
78
55
65
70
76
60
86
Average of
Avg.
relative
elev.
(cm)6
265
279
268
250
246
236
220
195
185
159
133
104
302
236
224
201
197
195
183
169
151
133
105
73
Slope
CM
2.85
0.25
3.30
6.20
4.45
4.20
4.10
5.20
3.10
1.10
1.75
2.65
3.05
2.55
3.75
2.90
4.50
4.65
4.30
4.30
4.50
5.00
3.25
3.85
three measurements per
Average elevation of
Monitor 9
Depth
to
bedrock
(cm)*
67
85
87
>90
87
78
>90
>90
82
>90
87
87
>90
>90
>90
>90
>90
82
>90
88
>90
85
>90
>90
plot.
Avg.
relative
elev.
Co.)6
88
96
99
98
101
87
120
139
128
122
130
159
162
177
180
177
174
185
192
209
218
212
209
221
Slope
«)
3.95
1.15
4.50
1.30
3.20
1.45
6.30
4.60
2.00
4.30
2.35
3.45
3.90
0.36
3.30
3.60
3.95
3.30
0.10
0.60
0.65
0.65
0.65
1.85
Monitor 10
Depth
to
bedrock
(on)*
80
78
67
82
77
72
77
82
77
65
67
67
78
80
<90
82
82
77
88
70
80
62
68
80
Avg.
relative
elev.
(cm)6
188
207
235
259
296
270
238
215
188
212
185
181
212
244
229
204
185
180
171
218
204
192
197
198
Slope
OS)
2.65
1.60
2.05
1.90
3.15
0.55
2.00
1.55
0.50
10.30
4.10
0.60
0.70
0.05
0.95
0.90
1.60
0.35
1.85
1.60
0.65
1.20
3.65
1.35
Depth
to
bedrock
(cm)8
68
67
67
75
67
80
78
77
87
77
77
>90
82
77
73
78
73
>90
87
77
88
70
83
87
Monitor 22
Avg.
relative
elev.
(cm)6
122
116
124
136
125
134
151
136
139
153
145
168
186
180
171
197
186
203
220
217
210
214
227
226
Slope
(%)
0.20
0.00
0.15
0.40
0.20
0.55
0.55
0.70
0.90
1.65
0.65
0.35
1.05
0.95
2.60
1.45
3.95
1.30
0.95
0.10
2.95
4.10
0.50
2.90
Monitor 25
Depth
to
bedrock
(cm)a
88
>90
87
88
>90
77
77
82
77
>90
80
77
82
83
68
77
83
77
82
75
78
88
>90
>90
Avg.
relative
elev.
(cm)6
82
85
101
111
102
107
119
125
136
159
153
156
136
110
110
134
130
125
151
162
168
194
183
174
Slope
«)
2.46
0.08
0.79
1.54
1.87
2.64
1.87
1.58
2.00
1.75
2.25
3.25
5.92
1.62
1.53
1.82
1.91
1.04
2.04
2.16
2.88
2.12
2.58
1.21
front and rear of plot.
-------
-64-
APPENDIX G
SULFUR DIOXIDE EXPOSURES AND ESTIMATED PERCENT FOLIAR
CHLOROSIS FOR THE 1976 GROWING SEASON
S02 concentration
(Mg/m3)
Site Date 1-h
6 8/13^ 2351
8/16 1881
8/21 1332
9/17 2743
9 7/22 1097
8/21 1515
9/21 862
10 7/7 1646
7/28 1776
8/14 2194
8/21 444
19 7/7 993
7/10 2482
7/14 2142
7/23 914
7/24 10318
8/16 1881
8/22 2377
9/15 1881
3-h
1463
914
914
2090
784
1097
784
1149
993
1515
340
836
1228
1541
862
4519
888
1593
1071
Estimated chlorosis
(%)
Treatment Min Max
1
2
1
2
1
2
3
1
2
1
2
0 8.5
0.6 30.0
No new effects observed
No estimates available
No effects observed
0 3.8
0 7.5
No estimates available
No effects observed
No effects observed
0 4.0
0.1 0.8
0.1 4.2
No new effects observed
No effects observed
No effects observed
No effects observed
No effects observed
No effects observed
0.2 3.8
0 3.0
0.1 1.8
0 4.5
No estimates available
Avg
0.9
6.4
0.7
0.7
0.5
0.3
1.0
1.4
1.0
0.8
0.6
a. 1 - air-exclusion; 2 -
plastic between rows.
b. Observations made on 8,
exposed
fll.
with plastic
between rows; 3 - exposed wi
thou
-------
-65-
APPENDIX H
YIELDS FOR EXPERIMENTAL PLOTS, 1976 GROWING SEASON, WIDOWS CREEK STEAM PLANT
Air-exclusion
Site Plot
6 1
2
3
X
9 1
2
7
X
10 1
2
3
X
19 1
3
7
X
22 1
6
7
X
Yield*
(kg/ha)
3136
3359
3387
3294
3103
3015
2901
3006
3610
3733
3501
3615
3153
2917
3291
3120
2850
2673
2599
2707
Standard
deviation
114
399
102
112
108
166
290
83
95
151
30
95
134
346
91
154
245
165
92
105
Exposed (plastic)
Plot
13
15
16
15
13
14
17
18
19
13
15
20
12
13
18
Yield3
(kg/ha)
2821
2955
2867
2881
3239
2597
2894
2910
3442
3490
3682
3538
1393
767
1286
1149
2703
2301
2657
2554
Standard
deviation
118
513
97
56
328
136
286
262
223
245
140
104
533
35
41
273
378
144
178
180
Exposed (no plastic)
Yield3
Plot (kg/ha)
9 2745
20 3080
22 2883
2903
10 3380
16 3344
17 3235
3320
4 2912
5 2628
6 2689
2743
Standard
deviation
99
88
102
137
155
259
33
62
100
65
37
122
a. Average of yields of three inner rows of plot.
-------
-66-
APPENDIX I
DATA SET USED IN REGRESSION ANALYSIS
Site Plot
6 1
2
3
13
15
16
9
20
22
9 1
2
7
15
13
14
10
16
17
10 1
2
3
4
5
6
17
18
19
Type"
1
1
1
2
2
2
3
3
3
1
1
1
2
2
2
3
3
3
1
1
1
3
3
3
2
2
2
Yield
(kg/ha)
3136
3359
3387
2821
2955
2867
2745
3080
2883
3103
3015
2901
3239
2596
2894
3380
3344
3235
3610
3733
3502
2912
2628
2689
3442
3490
3623
P
(kg/ha)
85
137
120
132
91
90
57
114
78
92
130
400
293
317
288
252
291
229
245
169
121
148
115
152
143
158
93
K
(kg/ha)
221
206
222
191
164
176
241
193
192
119
104
201
223
262
298
170
167
169
232
212
173
183
169
194
184
252
191
Mg
(kg/ha)
226
159
141
47
185
167
121
132
103
34
30
41
41
62
73
36
36
36
229
198
110
49
29
26
201
215
210
Ca
(kg/ha)
1457
1110
1666
876
1666
2369
1545
1831
1571
2447
2482
1891
1952
2421
2394
2343
2325
2273
1293
919
711
529
451
407
815
1171
1232
PH
6.4
5.6
6.0
5.2
6.0
6.7
4.9
5.3
5.0
6.4
6.6
6.0
5.6
6.7
6.5
5.4
6.1
6.1
6.2
6.1
5.4
4.8
4.8
4.8
5.9
6.1
6.1
OMb(%)
1.46
1.86
2.24
1.88
1.36
1.64
2.37
2.15
2.14
1.64
1.98
1.79
2.08
1.50
1.90
1.84
1.83
1.40
1.78
1.47
1.30
1.61
1.42
1.59
1.50
1.57
1.61
INDC
245
236
246
249
230
244
258
281
283
246
238
254
259
273
266
254
255
268
207
206
203
169
172
169
222
213
207
a. 1 - air-exclusion; 2 - exposed with plastic between rows; 3 - exposed
without plastic.
b. Percent organic matter.
c. Index of insect-disease injury.
-------
-67-
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-77-122
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Experimental Air Exclusion System for Field Studies of
S02 Effects on Crop Productivity
5. REPORT DATE
November 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
H.C. Jones, N.L. Lacasse, W.S. Liggett, and Frances
Weatherford
8. PERFORMING ORGANIZATION REPORT NO.
E-EP-77-5
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Division of Environmental Planning
Tennessee Valley Authority
Chattanooga, TN 37401
10. PROGRAM ELEMENT NO.
1NE625B
11. CONTRACT/GRANT NO.
79 BDL
12. SPONSORING AGENCY NAME AND ADDRESS
Office, of Energy, Minerals, § Industry
Office of Research $ Development
U.S. Environmental Protection Agency
Washington, DC 20460
13. TYPE OF REPORT AND PERIOD COVERED
Milestone
14. SPONSORING AGENCY CODE
EPA/600/17
15. SUPPLEMENTARY NOTES
This project is part of the EPA-planned and coordinated Federal Interagency
Energy/Environment R§D Program.
16. ABSTRACT
The Tennessee Valley Authority (TVA) characterized and quantified relationships
among sulfur dioxide (S02) exposure, symptomatology of injury, and yield of
soybean crops, which are sensitive to S02 and economically important to the
southeastern United States. Characterization included analysis of soil fertility,
pH, soil depth, slope, and content of organic matter. Regression analysis was used
to identify site factors that might affect yield and mask the effects of S02
exposure; results of the analyses were used to control or eliminate those factors.
TVA designed, constructed, and tested an air-exclusion system that permits the
comparison of plants exposed to S02 with plants at the same site that were protected
from S02 exposure. Tests showed the system to be as much as 85 percent efficient
in excluding or reducing S02 concentrations to subthreshold levels during exposure.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Ecology
Environments
Earth atmosphere
Control technology:
energy resource ex-
traction
Coal cleaning
Fuel:
Coal
6F; 10A; 10B;
97A; 97B
18. DISTRIBUTION STATEMENT
RELEASE UNLIMITED
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
------- |