- Email: [email protected]
Model analysis of interactive effects of ozone and water stress on the yield of soybean
Environmental Pollution 82 (1993) 39-45
MODEL ANALYSIS OF INTERACTIVE EFFECTS OF OZONE A N D WATER STRESS ON THE YIELD OF SOYBEAN Kazuhiko Kobayashi, a Joseph E. Miller, b Richard B. Flagler c & Walter W. Heck b aNational Institute of Agro-Environmental Sciences, 3-1-1 KannondaL Tsukuba, Ibaraki 305, Japan bUSDA Agricultural Research Service, North Carolina State University, 1509 Varsity Drive, Raleigh, North Carolina 27606, USA CTexas A & M University, College Station, Texas 77843, USA
(Received 17 December 1990; accepted 6 July 1992)
Abstract The interactive effects of ozone and water stress on the yield of soybean (Glycine max (L.) Merr. 'Davis') were addressed with a growth model o f soybean. Two simulations were conducted, using the data from the exposures o f soybean to ozone in open-top chambers under two soil moisture regimes, and the results o f the simulations were compared In the original simulation, soil moisture content was calculated based on a water budget using the actual precipitation and irrigation data. In the modified simulation, the soil water content was given as input data. In this case, soil moisture content was maintained at the same level across the ozone treatments regardless of different water use by the plants. Both simulations included the effect of reduced ozone flux to the leaves due to water stress, whereas only the original simulation included the effect o f mitigated water stress due to reduced water use by the plants under higher ozone concentration. The water stress reduced ozone impact on soybean yield in the orignal simulation on the basis of the ozone dosecrop yield response relationship, but not in the modified simulation. The ozone uptake rate was reduced by water stress in the original simulation, but the relationship between seasonal mean ozone uptake rate and relative yield still showed reduced impact of ozone due to water stress. These results indicated that the alleviation o f water stress by ozone due to reduced plant water use in ozone-treated plots can be a contributing factor in the reduction of ozone impact by water stress. The above conclusion was partly confirmed by the actual data for soil water content, which was significantly lower in the lowest ozone treatment than in the higher ozone treatments. Further experimental and modelling studies are needed to elucidate the mechanism o f the ozone x water stress interaction.
In the National Crop Loss Assessment Network (NCLAN), the Weibull function was adopted for describing the dose-response relationship (Heck et al., 1984) and was used in the economic assessment of crop losses (Adams et al., 1989). However, the NCLAN field experiments found that the dose-response relationship could be affected by other factors, e.g. location, weather conditions, and cultivars (Heagle et al., 1988a). Among the environmental variables, soil moisture is a major factor influencing the impact of ozone on plants. A number of studies have been conducted on the interaction between soil moisture stress and ozone injury. Soil moisture stress can protect plants from ozone injury (Tingey et al., 1982). Tingey and Hogsett (1985) ascribed the protective effects to reduced leaf stomatal conductance due to water stress. The interaction between ozone and soil moisture stress on crop yield has also been reported. Amundson et al. (1986) exposed potted soybean (Glycine max (L.) Merr.) to ozone under two soil moisture conditions, and reported that relative growth reduction due to ozone was less in water-stressed plants than in wellwatered plants. Flagler (1986) conducted ozone exposures of soybean in open-top chambers under two soil moisture regimes for 2 years. Although the interaction between the effects of soil moisture regimes and ozone levels on the seed yield was not significant, ozone-induced yield loss was apparently decreased by water stress in the hot and dry year (Heagle et al., 1987). Similar results have been reported for cotton, Gossypium hirsutum L. (Heagle et al., 1988b; Temple et al., 1985, 1988b), and for alfalfa, Medicago sativa L. (Temple et al., 1988a). On the other hand, Heggestad et al. (1985) reported increased sensitivity of soybean growth and yield to ambient levels of ozone under water stress. The alteration of ozone dose-yield loss relationship by soil moisture stress is important in crop loss assessments; however, most of the dose-response relationships were derived from the NCLAN field experiments under adequate soil moisture conditions (Heagle et aL, 1988a). For better applicability of the NCLAN dose-response relationships, the cause of the interaction between ozone and soil moisture on crop loss needs to be studied. The interaction can occur from the following causes.
INTRODUCTION
Effects of ozone on crop yield have been described by the relationship between ozone dose and yield response. The dose-response relationship is effective in summarizing yield reductions caused by ozone from the field experiments conducted for various crop species. Environ. Pollut. 0269-7491/93/$06.00 © 1993 Elsevier Science Publishers Ltd, England. Printed in Great Britain
39
40
K a z u h i k o Kobayashi et al.
First, water stress can protect plants from ozone damage through reduced stomatal conductance, which means reduction in ozone uptake and effects of ozone. Moser et al. (1988) exposed bush bean planted in special containers to ozone, and drought stress was applied to the plants during the reproductive stage. They found the effects of ozone on growth and yield of bush bean were reduced by drought stress during the reproductive stage, and they assumed that the reduced effects of ozone were the result of decreased stomatal conductance. Second, plants exposed to higher concentrations of ozone use less water than control plants because of smaller plant size and stomatal closure, and the water stress can thus be alleviated. King (1987) suggested from the results of a model simulation that the alleviation of water stress by ozone might be the major factor in the interaction. In fact, Temple (1990) reported that the ozone-injured cotton plants were less stressed by drought, and suggested that ozone alleviated the water stress through stomatal closure and reduced transpiration. Seasonal evapo-transpiration was, however, not changed by ozone, presumably due to increased soil evaporation in higher ozone levels. Temple (1990) also suggested osmotic change in leaves as another possible cause of the drought resistance in the ozone-injured cotton plants. Kobayashi et al. (1990) developed a simple growth model for addressing the effects of ozone and water stress on the growth and yield of soybean. In this paper, soybean growth under various ozone exposure levels and soil moisture regimes in the open-top chamber experiments (Flagler, 1986) was simulated by the model, and resultant interacting effects of ozone and soil moisture stress on the yield were analyzed. MATERIALS AND METHODS Model The model developed by Kobayashi et al. (1990) was based on a simple growth model of soybean (Sinclair, 1986). In the model, daily growth of soybean plants is simulated using environmental data including shortwave radiation, temperature, precipitation, and irrigation. Major growth processes described by the model are leaf growth, dry matter accumulation, nitrogen uptake, water budget, and leaf senescence and abscission linked to seed growth. Among them, the dry matter accumulation and the senescence are affected by ozone. Ozone reduces the light-use efficiency, which is the dry matter increase per unit of radiation intercepted by the canopy, and hence the rate of dry matter accumulation. After seed growth initiation, increase of harvest index is accelerated by ozone. This shortens seed growth duration and accelerates leaf abscission. The effects of ozone on the growth processes are calculated on the basis of specific ozone uptake rate, which is ozone uptake rate per unit leaf area. The ozone uptake rate is calculated from transpiration rate and ozone concentration assuming a proportionality of ozone uptake to transpiration. The model was calibrated to the results of the open-
top chamber experiments (Flagler, 1986). The effects of ozone and water stress on soybean growth and yield were satisfactorily described by the model, and the simulation indicated mitigating effects of water stress on ozone-induced yield loss. See Kobayashi et al. (1990) for further details of the model. Data The open-top chamber experiments, to which the model was calibrated, were conducted in 1983 and 1984. Soybean 'Davis' was exposed to ozone in the open-top chambers (Heagle et al., 1973). The experimental design was a factorial combination of two moisture regimes with either four (1983) or six (1984) ozone levels. Each combination of the treatments was completely replicated in either three (1983) or two (1984) blocks. The ozone levels were charcoal-filtered air (CF), nonfiltered air (NF), and NF to which constant amounts of ozone were added, which are referred to as NF30 and NF60 in 1983, and NF15, NF30, NF45, and NF60 in 1984, where the two digits after 'NF' represent nominal additions of ozone in ppb (10 -9 liters liters ~). Soil moisture regimes were well-watered (WW) and water-stressed (WS). In both years, the WW plots were irrigated with 2.5 cm of water when soil matric potential measured with tensiometers reached -0-03 MPa. The WS plots were irrigated with 2-5 cm of water when leaf water potential measured with psychrometers reached -1.8 MPa (1983) or when soil matric potential estimated from a neutron-prove measurement reached -0.5 MPa (1984). For further details of the experiments, refer to Flagler (1986) and Heagle et al. (1987). Weather data were obtained from the North Carolina State University Weather Data Acquisition System located adjacent to the experimental plots. Daily amounts of precipitation and dates and amounts of irrigation were obtained from the field records. Daily 7-h mean ozone concentrations during the ozone exposure were calculated from the NCLAN data file. Parameter values adopted in the simulation have been described by Kobayashi et al. (1990). Soil moisture was measured with tensiometers daily and with a neutron probe during the dry period for 1983 and 1984. Whereas the tensiometer readings were expressed in MPa, the neutron probe data were converted to the gravimetric soil water content (%) based on a laboratory calibration. Details of the soil moisture measurements have been published (Flagler et al., 1987). The gravimetric soil water content at a depth of 22.5 cm was subjected to Tukey's multiple range test for the effect of the ozone treatment with the day of each measurement as a block. In 1983, 17 measurements were performed, but only 6 in 1984. The tensiometer readings for well-watered plots were averaged over the whole period of measurements for each ozone treatment, and the differences between the means were statistically tested. The number of measurements averaged was 98 (1983) and 105 (1984). Tensiometer data were also available for water-stressed plots, but are not addressed in this paper, since the soil water potential often dropped
Effects of 03 and water stress on soybean
41
ozone treatments. Soil water content was given as input data to the model rather than the result of the calculated water budget as in the original simulation, and the same values of soil water content were used for all ozone treatments in the simulation for each soil moisture regime. In this situation, the simulated plants are subjected to the same level of water stress regardless of their water usage. Thus, the alleviation of water stress at the higher ozone level is eliminated, and the interaction, if any, is ascribed solely to the reduction in the ozone uptake rate due to water stress. The model structure for the original simulation, which was shown in Fig. 1 o f Kobayashi et al. (1990), was changed to accommodate the modified simulation. The modified model structure is shown in Fig. 1 of this paper. In order to simulate actual day-by-day variations in soil moisture conditions, the input data for soil water content were taken from a result of the original simulation which was based on the calculated water budget. The outputs for NF30 treatments in each year were used within each moisture regime. Hence, the modified simulation gave the same results as the original simulation for NF30 treatments, but not for other treatments.
beyond the lower limit for the tensiometer measurement c. - 0 . 0 7 MPa. Simelation
As described in the previous section, the interaction between ozone impact and water stress on the yield of soybean can be caused either by the reduced ozone uptake due to stomatal closure under water stress, by the alleviated water stress under higher ozone concentrations, or by other physiological changes in the sensitivity of the plants to ozone. For assessing the contribution of the former two possible causes of the interaction, two simulations differing in the method of calculating soil water content were conducted. Resulting yield losses were compared between the soil moisture regimes. The other possible causes of the interaction, e.g. osmotic adjustment, were not addressed in this paper.
Original simulation The open-top chamber experiments reported by Flagler (1986) were simulated using actual data on precipitation and irrigation. Soil water content was calculated in the water budget part of the model from the water input data and daily amount of transpiration and soil evaporation estimated in the model. Thus, both causes of the interaction mentioned above are included in this simulation.
Internal ozone dose-yieM loss relationship In addition to the comparison between the two simulations, the effect of water stress on internal ozone dose was examined. The internal ozone dose was expressed by the ozone uptake rate per leaf area, and was calculated
Modified simulation The model simulated a situation in which soil water content is maintained at the same level across the Solar ,('- MODIFIED "~ ~k,WATER BUDGET~ FTSW
radiation Temperature 9 9 , ~ DRY MATTER "~ ~ ........ H~ i I~ I!ii ,.,-~r t = . u w , F(~--~.-~-~--~i . . . . . . . . . ~ Plastochronl I O, =r ~ - " :-~'},~DIR ~:.!;~)~- - ~ index-] [ , : ,K,a, ,,- ,• ..... ~ I:!t. x l " ~ _ ,- . . . . . . . . . . I Tra'nspirable t . . . . : : ~ 1 ~ I~ Vapor SO water [-'" : : ,;-r- P h ~ e l::; i f.q==nrz-taoutrru:'~ : ' pressure , , ', ! . ............. " ' . .... ~ . 1 ~ ,,' :, ' : ~ ~ ' SENESCENCE , ', aer,c,l I 3It" ........ Ld LJ i ii JL ..[" " '
,
: ',
," i
ct
,,
T ,'"'hStoredphoto-I
.-~Transpirati°n Transpiration I.~, ' ....-, "~
," ' ............................................~" ,,C"~r'('EFFECTSOF~
' :
"-,.-'/ ' h
,
,,I1"
o~
]
=
ill
, ,
Ill:
', centration". ] ......
¢'
materials
i
[
rymatter re-
~ ' , ,
~1~(,
I
m°bilizati°n I ~ i ....
"
::~ ', I:'
e !
', I: i
"
: I
'
............................................................,',
. . . . . . . . . . . . . . . l,.f~.l/ Nitrogen I .~N,-~t" S~F~ °:~'. " ~ fixation I L .......... ..v..-.'... 1 Leaf nitr°gen [ ......... ~-....... i
::
LEGEND Flow of
.................
s Ynthate
',
-'- ..... " ............ ~ ~ ' - ' ~ ' " - - ' " J J
,I-~---- ..... ~-I-}
: -"-'-'~ 1103----~con- [ 1/~ ......
!'~l
i ~ . ~ . ~ [ / ~
.q
Absorbed t'~--' I
I
~'~"
ou.co) ,. OZONE
~
" ' ....
i
:
!
F" ,
oration
r~.l~titrogen re_ i" ~FIow of
Auxiliary variableQ
j
~'~ mobilization | .....
information External variable . . . . . . .
-I,,,-
O
Fig. 1. Schematized relational diagram of the model for modified simulation. Only the major relations are shown. Each model component is surrounded by a rounded rectangle. In the modified simulation, daily value of the soil water content is given as the input data. Some of the processes are by-passed depending on the growth phase. Abbreviations are as follows: DIR = daily amount of short-wave radiation intercepted by leaves. EFFO = light utilization efficiency under ozone exposure, FTSW = fraction of transpirable soil water content, HI = harvest index, LAI = leaf area index, OU.co = coefficient for ozone uptake rate, SN.cont= seed nitrogen content, SNF.co = specific daily nitrogen fixation, WUE.co = coefficient for water use efficiency.
42
Kazuhiko Kobayashi et al. 1.1
in the original simulation. The relationship between the seasonal mean intemal ozone dose and the relative yield was examined for each water stress treatment.
1.0
RESULTS Simulated yields in the original simulation are shown in Fig. 2 for both years. The reduction in yield due to water stress was much more pronounced in the hot and dry year of 1983 than in the cooler year of 1984 although the yield for the well-watered plots was higher in 1983 than in 1984. In the original simulation, there was no effect of ozone on the simulated yield for the water-stressed plots in 1983 as shown on a relative basis compared to CF treatment (Fig. 3(A)). Water stress also decreased the impact of ozone on the yield in the original simulation in 1984 (Fig. 3(B)) although to a lesser extent than in 1983. Year-to-year variation in the dose-response relationship is apparent for the well-watered plots, even though the plots were well-irrigated in both years.
1983
A
.
0.9
0.7 "O (9 1o"
I
Orig.
--
Ong. WS
- , , Mod. WW
0.8
Interpretation of the simulation results In the previous paper (Kobayashi et aL, 1990), the model had been calibrated to the results of the open-top chamber experiments (Flagler, 1986). Values of the model parameters including the parameters for nitrogen fixation rate and for the effects of ozone have been determined so as to fit the data. Thus, the model is 'empirical' rather than mechanistic with respect to the parameters used for the 'fitting'. The same caution is needed in a predictive use of the model as is needed in the use of an empirical dose-response relationship, e.g. Weibull function. Under a situation which is quite different from the experiments, different values might be appropriate for the model parameters. Yet a comparison of the simulation results on a relative basis can be as meaningful as a comparison of the dose-response relationships. In this paper, the results of the open-top chamber experiments were simulated with the fitted model, and the relative yields were compared between the two simulations which differed in soil moisture calculation. The analysis can give us a suggestion, if not a confirmation, on the mechanisms underlying the ozone × water stress interaction.
~
- . - Mod. WS
~,~..
O.9
B
"~.? x
1984
tr
--
~%'%% ~" ""'"..\ ,,~
0.7 0.8 I 0.6
...... Mod. VWM
~,~
...... Mod. WS
~\
0.5
0.4 0.02
I
I
I
I
0.04
0.06
0.08
0.10
0.12
Seasonal mean 7-h/d Oa concentration (ppm)
Fig. 3. Comparison between the relative yields calculated in the original and in the modified simulations in 1983 (A) and in 1984 (B) under the two soil moisture conditions. The charcoal-filtered air treatment = 1. Abbreviations as in Fig. 2.
In the modified simulation, on the other hand, soybean yield for water-stressed plots responded to ozone dose quite similarly to well-watered plots in 1983, whereas no effect of ozone was found in the original simulation of water-stressed plots (Fig. 3(A)). Little difference was found between the original and the modified simulations for the well-watered plots. In 1984 (Fig. 3(B)), the difference between the two simulations was much smaller than in 1983. However, the doseresponse relationship for the water-stressed plots in the modified simulation was closer to that for the wellwatered plots than in the original simulation, particularly at the lower ozone doses. The dose-responses for the well-watered plots in the two simulations were almost identical in 1984 (Fig. 3(B)). The relationships between the mean internal ozone dose and relative yield in the original simulation are shown in Fig. 4 for each soil moisture treatment in 1.1
1.0" 0.9'
,o01 "
--1,83 "*"~%-
300
l
- 2001001 0.02
~
-
~,
-
--
1983 WS 1984 W W
a 0.04
0 ~. 0.06
~"XXx\.\" o.io
, - - 1984 WS o.12
Seasonal mean 7-hid 03 concentration (ppm)
Fig, 2, E f f e c t s of" o z o n e o n the s o y b e a n seed yields c a l c u l a t e d
in the original simulation under water-stressed (WS) or wellwatered (WW) conditions in 1983 and 1984.
~-1983
0.8' >~ _~ G) 0.7'
--
"4
rr
0.6 ~
t%, b
1983 W S
--,1984 --'
WW
WW
1984 W S
I
0.5 0.4
""
Orig. WW
- - Orig. WS
2
n
I 4
=
: 6
:
: : : : : ; 8 10 12 14 Seasonal mean 03 uptake rate per unit leaf area (rag m 2 d q)
Fig. 4. Relationship between the seasonal mean ozone uptake rate per unit leaf area, calculated in the original simulation, and the relative yield in 1983 and 1984. Simulated yields for the two soil moisture regimes with the same ozone treatment are interconnected by a dotted line. The charcoal-filtered air treatment = I. Abbreviations as in Fig. 2.
Effects of Os and water stress on soybean Table 1. Mean soil water content from the neutron probe measurements
Ozone treatment
Gravimetric soil water content (°/o)a 1983 WS
CF NF NF15 NF30 NF45 NF60
1984 WW
3.49 a b 4.64 b -6-10 c -5.15 b
8.14 a 9.52 bc --10.69 c -8.68 ab
WS
WW
4.98 b 3-47 ab 4.37 ab 2.82 a 3.67 ab 5.28 b
4-42 a 8.57 b 8-56 b 7.92 b 6.87 ab 6.77 ab
aMean values of 17 (1983) and 6 (1984) neutron probe measurements are shown. WW stands for well-watered plots and WS for water-stressed plots. b Means with the same letter are not significantly different with Tukey's studentized range test. 1983 and 1984. The water-stressed plants had smaller ozone uptake rate than the well-watered plants in the same ozone treatment as shown by the dotted lines in Fig. 4. As a result of the reduced internal ozone dose due to water stress, the discrepancy between the doseresponse curves for water-stressed plants and the curves for well-watered plants in Fig. 4 was somewhat smaller than that for the original simulation in Fig. 3, where the relative yield was plotted against ozone concentration. However, the reduction in the impact of ozone on the yield by water stress was still obvious in 1983. The gravimetric soil water content calculated from the neutron probe measurements is averaged over 17 (1983) and 6 (1984) measurements, and is shown in Table 1 for each ozone treatment. Soil water content was the lowest in the lowest ozone treatments (CF) for both well-watered and water-stressed treatments in 1983, and for well-watered treatment in 1984. Although the soil water content was significantly different a m o n g the ozone treatments for water-stressed treatment in 1984, the relationship between the ozone level and soil water content was obscure. Soil water potential measured with the tensiometers showed no significant effect of ozone for both years (Table 2). Table 2. Mean soil water potential from the tensiometer measurements
Ozone treatment CF NF NF15 NF30 NF45 NF60
Soil water potential (MPa) a 1983WW
1984WW
-0-0151 -0.0158 --0.0153 --0.0141
-0.0124 -0.0141 -0.0100 -0.0127 -0.0109 -0.0097
ab a a a
a a a a a a
a Seasonal mean tensiometer measurements are averaged across three (1983) or two (1984) blocks. Only the results for well-watered plots (WW) are shown. b Means with the same letter are not significantly different with Tukey's studentized range test.
43
DISCUSSION The results of the original simulation clearly showed reduced impact of ozone on soybean yield under waterstressed conditions in 1983. In contrast, the modified simulation, which does not include the effect of mitigated water stress under higher ozone concentrations, resulted in similar dose-response relationships for both well-watered and water-stressed plots. Similar but less clear results were obtained in the 1984 simulation. The contrast between the two simulations indicates that the reduced impact of ozone on soybean yield by soil moisture stress may be caused by the alleviation of water stress by higher ozone doses. The difference between 1983 and 1984 is apparently due to the different weather conditions in the 2 years. The 1983 season was hot and dry with very little rainfall during vegetative growth, whereas the 1984 season was wet through vegetative and early reproductive growth stages (Flagler et al., 1987). The difference in weather conditions may reflect different severities of water stress between the 2 years, and a change in the timing of the ozone exposure during the growing seasons. The neutron probe measurements indicated that the soil water content was lower in the lower ozone treatments. Since the total amount of irrigation was the same for all ozone treatments under water stress (Heagle et al., 1987), the difference in the soil water content for the water-stressed plots in 1983 has likely been caused by the differential water use of the plants as affected by the ozone treatment. The well-watered plots were not irrigated identically across the ozone treatment (Heagle et al., 1987), but the difference in the soil water content (Table 1) may also be a result of the difference a m o n g the ozone treatment in the water use rather than in the water input. Although the differences in the soil water content cannot be translated straightforwardly into the difference in plant water status, the amount of soil water available for plants was not much different a m o n g the ozone levels in well-watered plots as suggested by the tensiometer measurements (Table 2). Reduced ozone flux to leaves under water stress was accounted for by the model. However, the relationships between the internal ozone dose and the yield loss still showed the ozone x water stress interaction in 1983, when the impact of ozone on the yield was clearly decreased by water stress (Fig. 4). This result indicates that the reduced ozone flux to the leaves of waterstressed plants might not be the only factor determining the mitigation of ozone impact by water stress. In the simulation, the stomatal closure of the waterstressed plants reduced the ozone uptake by the leaves and, thereby, the impact of ozone on growth processes. However, the stomatal closure due to water stress also suppressed simulated growth and yield for CF treatments as well as other ozone levels. Thus, on a relative basis compared to the yield for CF treatments, reduced ozone impact due to water stress was partly cancelled out by the yield loss due to water stress in the simulation.
44
Kazuhiko Kobayashi et al.
Some results obtained through modelling and field experiments support the possibility that mitigation of water stress by ozone contributes to ozone × water stress interaction. King (1987) developed a model for the impacts of ozone and water stress on crop yield. The model calculated seasonal dry matter production as the product of seasonal transpiration and a coefficient, transpiration efficiency. The model included reduced ozone uptake due to drought stress and reduced water use under high ozone concentrations. The simulation for soybean yield losses suggested that the impact of ozone on the water use by plants, rather than the reduced ozone uptake due to water stress, might be a major factor in the interaction (King, 1987). Heggestad et al. (1988) investigated the effects of ozone on root development of soybean under two soil moisture regimes. They tried to maintain the same soil moisture level across the ozone levels, and hence more water was added to lower ozone treatments, where the plants used more water. Nevertheless, there was an effect of ozone treatment on soil water potential; the lowest soil water potential occurred in the lowest ozone level in the water-stressed plots, as is found in the experiments in this paper (Table 1). Under the lower ozone concentrations, root length density was much greater in water-stressed plots than in well-watered plots, whereas in the highest ozone concentration, root length density was almost the same for both soil moisture regimes. Their data on cultivar 'Williams' indicate significant yield loss due to water stress in lower ozone levels but not in higher ozone levels. These results on root growth and yield might suggest that the plants in lower ozone concentration were subjected to greater water stress than in higher ozone concentration in spite of the authors' effort to maintain the same level of soil water content across the ozone levels. Heggestad and Lee (1990) grew 'Williams' soybean in soil columns placed in greenhouses with charcoal filtered and nonfiltered air, and investigated the effects of ambient ozone and soil moisture stress on water use, growth and yield of the soybean plants. Because of reduced size and earlier maturity, plants in nonfiltered air used less water than plants in charcoal filtered air during the season. Water was added to maintain about the same soil water potential for both filtered and non filtered air plots. Nevertheless, among the water-stressed plots, the plot with charcoal filtered air had lower soil water potential during the experiment and lower soil water content at harvest than the plot with nonfiltered air. This lower soil moisture content in charcoal filtered plot with soil moisture stress might be a contributing factor in the interaction between the effects of ozone and soil moisture regimes on plant growth and yield. Although the interaction was not significant, the combined effect of ozone and water stress was less than the additive effects of both stresses. In contrast to the similar experiment in the open-top chambers (Heggestad et al., 1988), the soil moisture stress did not increase the root development for both ozone regimes. The differential root
response to water stress was attributed to the difference in the soil moisture profile between the experiments (Heggestad & Lee, 1990). In the open-top chamber experiment (Heggestad et al., 1988), there were significant amounts of stored water available at lower soil depth, whereas no water was available at lower depth in the greenhouse experiment (Heggestad & Lee, 1990). Stomatal closure due to water stress has been widely assumed to be a main cause of the reduction of ozone effects by water stress. Moser et al. (1988) exposed bush bean plants in containers to ozone, and imposed drought stress on the bean plants during the early or late reproductive stages. They found reduced effects of ozone on bush bean growth and yield under drought conditions, and assumed that decreased stomatal conductance has resulted in the decreased impacts of ozone based on their preliminary measurement of transpiration. Using a special cultural system, they could control the drought stress conditions in the plant containers rather precisely. Their system attained the drought stress by increasing the distance between root zone and the water table, thereby decreasing the hydraulic conductivity in the cultural medium. The water content in the root media might be decreased by increased water use of the plants under low ozone levels. Hence, the interaction that Moser et al. (1988) found could include the mitigation of water stress from ozone-induced reduction in water use in addition to the reduced ozone uptake through decreased stomatal conductance in the water-stressed plants. Heagle et al. (1988b) exposed cotton to different levels of ozone in open-top chambers under water-stressed conditions or under well-watered conditions. They found significant yield loss due to ozone in the wellwatered regime but not in the water-stressed regime. Although plots for all ozone levels in each soil moisture regime were irrigated with the same amount of water, there were no obvious differences in soil water potential among the ozone levels, partly due to variation between replicates. Therefore, the cause of the apparent interaction between soil moisture stress and ozone is not clear in this case. In summary, the modelling approach in this paper identified reduced water use due to ozone as a contributing factor in ozone x water stress interaction. Although the reduced ozone uptake due to water stress showed some contribution to the interaction, it was not the dominating factor in the case of the severe water stress studied in this paper. Other physiological responses which might be involved in the interaction were not addressed in this paper. Further experimental and modelling studies are needed to elucidate the mechanisms of the interaction between the impacts of water stress and ozone on crop yield. Measurements of ozone uptake rate and internal ozone dose for crop canopies under well-watered or water-stressed conditions will help to better understand the interaction. Experiments in which soil moisture content is precisely maintained at the same level across the ozone treatments could help clarify the mechanisms of the interaction.
Effects of 03 and water stress on soybean ACKNOWLEDGEMENTS We wish to thank D r T. R. Sinclair for his suggestions on the water stress-ozone interaction, D r A. S. Heagle for his help in the use of the field records and other data in the field experiment, Susan Spruill for database management, and D r S. Hasegawa for his discussion on the control of soil moisture content. We also appreciate helpful comments from the reviewers on this manuscript. Major part of this work was performed when the first author was visiting at N o r t h Carolina State University, Department of Botany as a visiting scholar. The visit was funded by the Science and Technology Agency of the G o v e r n m e n t of Japan.
REFERENCES Adams, R. M., Giyer, J. D., McCarl, B. A. & Johnson, S. L. (1989). A reassessment of the economic effects of ozone on U.S. agriculture. J. Air Pollut. Control Assoc., 39, 960-8. Amundson, R. G., Raba, R. M., Shoettle, A. W. & Reich, P. B. (1986). Response of soybean to low concentration of ozone: II Effects on growth, biomass allocation, and flowering. J. Environ. Qual., 15, 161-7. Flagler, R. B. (1986). Effects of ozone and water deficit on growth, yield, and nitrogen metabolism of soybeans. PhD thesis, North Carolina State University, Raleigh, NC, USA. Flagler, R. B., Patterson, R. P., Heagle, A. S. & Heck, W. W. (1987). Ozone and soil moisture deficit effects on nitrogen metabolism of soybean. Crop Sci., 27, 1177-84. Heagle, A. S., Body, D. E. & Heck, W. W. (1973). An opentop field chamber to assess the impact of air pollution on plants. J. Environ. Qual., 2, 365-8. Heagle, A. S., Flagler, R. B., Patterson, R. P., Lesser, V. M., Shafer, S. R. & Heck, W. W. (1987). Injury and yield response of soybean to chronic doses of ozone and soil moisture deficit. Crop Sci., 27, 1016-24. Heagle, A. S., Kress, L. W., Temple, P. J., Kohut, R. J., Miller, J. E. & Heggestad, H. E. (1988a). Factors influencing ozone dose-yield response relationships in open-top field chamber studies. In Assessment of Crop Loss from Air Pollutants, eds W. W. Heck, O. C. Taylor & D. T. Tingey. Elsevier Science, London, pp. 141-79. Heagle, A. S., Miller, J. E., Heck, W. W. & Patterson, R. P. (1988b). Injury and yield response of cotton to chronic doses of ozone and soil moisture deficit. J. Environ. Qual., 17, 627-35.
45
Heck, W. W., Cure, W. W., Rawlings, J. O., Zaragoza, L. J., Heagle, A. S., Heggestad, H. E., Kohut, R. J., Kress, L. W. & Temple, P. J. (1984). Assessing impacts of ozone on agricultural crops: II. Crop yield functions and alternative exposure statistics. J. Air Pollut. Control Assoc., 34, 810-17. Heggestad, H. E. & Lee, E. H. (1990). Soybean root distribution, top growth and yield responses to ambient ozone and soil moisture stress when grown in soil columns in greenhouses. Environ. Pollut., 65, 195-207. Heggestad, H. E., Gish, T. J., Lee, E. H., Bennett, J. H. & Douglass, L. W. (1985). Interaction of soil moisture stress and ambient ozone on growth and yields of soybeans. Phytopathology, 75, 472-7. Heggestad, H. E., Anderson, E. L., Gish, T. J. & Lee, E. H. (1988). Effects of ozone and soil water deficit on roots and shoots of field-grown soybeans. Environ. Pollut., 50, 259-78. King, D. A. (1987). A model for predicting the influence of moisture stress on crop losses caused by ozone. Ecol. Model., 35, 29-44. Kobayashi, K., Miller, J. E., Flagler, R. B. & Heck, W. W. (1990). Modeling the effects of ozone on soybean growth and yield. Environ. Pollut., 65, 33-64. Moser, T. J., Tingey, D. T., Rodecap, K. D., Rossi, D. J. & Clark, C. S. (1988). Drought stress applied during the reproductive phase reduced ozone-induced effects in bush beans. In Assessment of Crop Loss from Air Pollutants, eds W. W. Heck, O. C. Taylor & D. T. Tingey. Elsevier Applied Science, London, pp. 345~4. Sinclair, T. R. (1986). Water and nitrogen limitation in soybean grain production. I Model development. Field Crops Res., 15, 125-41. Temple, P. J. (1990). Water relations of differentially irrigated cotton exposed to ozone. Agron. J., 82, 800-5. Temple, P. J., Taylor, O. C. & Benoit, L. F. (1985). Cotton yield responses to ozone as mediated by soil moisture and evapotranspiration. J. Environ. Qual., 14, 55-60. Temple, P. J., Benoit, L. F., Lennox, R. W., Reagan, C. A. & Taylor, O. C. (1988a). Combined effects of ozone and water stress on alfalfa growth and yield. J. Environ. Qual., 17, 108-13. Temple, P. J., Kupper, R. S., Lennox, R. W. & Rohr, K. (1988b). Injury and yield responses of differentially irrigated cotton to ozone. Agron. J., 80, 751-5. Tingey, D. T. & Hogsett, W. W. (1985). Water stress reduces ozone injury via a stomatal mechanism. Plant Physiol., 77, 944-7. Tingey, D. T., Thutt, G. L., Gumpertz, M. L. & Hogsett, W. E. (1982). Plant water status influences ozone sensitivity of bean plants. Agric. Environ., 7, 243-54.
MODEL ANALYSIS OF INTERACTIVE EFFECTS OF OZONE A N D WATER STRESS ON THE YIELD OF SOYBEAN Kazuhiko Kobayashi, a Joseph E. Miller, b Richard B. Flagler c & Walter W. Heck b aNational Institute of Agro-Environmental Sciences, 3-1-1 KannondaL Tsukuba, Ibaraki 305, Japan bUSDA Agricultural Research Service, North Carolina State University, 1509 Varsity Drive, Raleigh, North Carolina 27606, USA CTexas A & M University, College Station, Texas 77843, USA
(Received 17 December 1990; accepted 6 July 1992)
Abstract The interactive effects of ozone and water stress on the yield of soybean (Glycine max (L.) Merr. 'Davis') were addressed with a growth model o f soybean. Two simulations were conducted, using the data from the exposures o f soybean to ozone in open-top chambers under two soil moisture regimes, and the results o f the simulations were compared In the original simulation, soil moisture content was calculated based on a water budget using the actual precipitation and irrigation data. In the modified simulation, the soil water content was given as input data. In this case, soil moisture content was maintained at the same level across the ozone treatments regardless of different water use by the plants. Both simulations included the effect of reduced ozone flux to the leaves due to water stress, whereas only the original simulation included the effect o f mitigated water stress due to reduced water use by the plants under higher ozone concentration. The water stress reduced ozone impact on soybean yield in the orignal simulation on the basis of the ozone dosecrop yield response relationship, but not in the modified simulation. The ozone uptake rate was reduced by water stress in the original simulation, but the relationship between seasonal mean ozone uptake rate and relative yield still showed reduced impact of ozone due to water stress. These results indicated that the alleviation o f water stress by ozone due to reduced plant water use in ozone-treated plots can be a contributing factor in the reduction of ozone impact by water stress. The above conclusion was partly confirmed by the actual data for soil water content, which was significantly lower in the lowest ozone treatment than in the higher ozone treatments. Further experimental and modelling studies are needed to elucidate the mechanism o f the ozone x water stress interaction.
In the National Crop Loss Assessment Network (NCLAN), the Weibull function was adopted for describing the dose-response relationship (Heck et al., 1984) and was used in the economic assessment of crop losses (Adams et al., 1989). However, the NCLAN field experiments found that the dose-response relationship could be affected by other factors, e.g. location, weather conditions, and cultivars (Heagle et al., 1988a). Among the environmental variables, soil moisture is a major factor influencing the impact of ozone on plants. A number of studies have been conducted on the interaction between soil moisture stress and ozone injury. Soil moisture stress can protect plants from ozone injury (Tingey et al., 1982). Tingey and Hogsett (1985) ascribed the protective effects to reduced leaf stomatal conductance due to water stress. The interaction between ozone and soil moisture stress on crop yield has also been reported. Amundson et al. (1986) exposed potted soybean (Glycine max (L.) Merr.) to ozone under two soil moisture conditions, and reported that relative growth reduction due to ozone was less in water-stressed plants than in wellwatered plants. Flagler (1986) conducted ozone exposures of soybean in open-top chambers under two soil moisture regimes for 2 years. Although the interaction between the effects of soil moisture regimes and ozone levels on the seed yield was not significant, ozone-induced yield loss was apparently decreased by water stress in the hot and dry year (Heagle et al., 1987). Similar results have been reported for cotton, Gossypium hirsutum L. (Heagle et al., 1988b; Temple et al., 1985, 1988b), and for alfalfa, Medicago sativa L. (Temple et al., 1988a). On the other hand, Heggestad et al. (1985) reported increased sensitivity of soybean growth and yield to ambient levels of ozone under water stress. The alteration of ozone dose-yield loss relationship by soil moisture stress is important in crop loss assessments; however, most of the dose-response relationships were derived from the NCLAN field experiments under adequate soil moisture conditions (Heagle et aL, 1988a). For better applicability of the NCLAN dose-response relationships, the cause of the interaction between ozone and soil moisture on crop loss needs to be studied. The interaction can occur from the following causes.
INTRODUCTION
Effects of ozone on crop yield have been described by the relationship between ozone dose and yield response. The dose-response relationship is effective in summarizing yield reductions caused by ozone from the field experiments conducted for various crop species. Environ. Pollut. 0269-7491/93/$06.00 © 1993 Elsevier Science Publishers Ltd, England. Printed in Great Britain
39
40
K a z u h i k o Kobayashi et al.
First, water stress can protect plants from ozone damage through reduced stomatal conductance, which means reduction in ozone uptake and effects of ozone. Moser et al. (1988) exposed bush bean planted in special containers to ozone, and drought stress was applied to the plants during the reproductive stage. They found the effects of ozone on growth and yield of bush bean were reduced by drought stress during the reproductive stage, and they assumed that the reduced effects of ozone were the result of decreased stomatal conductance. Second, plants exposed to higher concentrations of ozone use less water than control plants because of smaller plant size and stomatal closure, and the water stress can thus be alleviated. King (1987) suggested from the results of a model simulation that the alleviation of water stress by ozone might be the major factor in the interaction. In fact, Temple (1990) reported that the ozone-injured cotton plants were less stressed by drought, and suggested that ozone alleviated the water stress through stomatal closure and reduced transpiration. Seasonal evapo-transpiration was, however, not changed by ozone, presumably due to increased soil evaporation in higher ozone levels. Temple (1990) also suggested osmotic change in leaves as another possible cause of the drought resistance in the ozone-injured cotton plants. Kobayashi et al. (1990) developed a simple growth model for addressing the effects of ozone and water stress on the growth and yield of soybean. In this paper, soybean growth under various ozone exposure levels and soil moisture regimes in the open-top chamber experiments (Flagler, 1986) was simulated by the model, and resultant interacting effects of ozone and soil moisture stress on the yield were analyzed. MATERIALS AND METHODS Model The model developed by Kobayashi et al. (1990) was based on a simple growth model of soybean (Sinclair, 1986). In the model, daily growth of soybean plants is simulated using environmental data including shortwave radiation, temperature, precipitation, and irrigation. Major growth processes described by the model are leaf growth, dry matter accumulation, nitrogen uptake, water budget, and leaf senescence and abscission linked to seed growth. Among them, the dry matter accumulation and the senescence are affected by ozone. Ozone reduces the light-use efficiency, which is the dry matter increase per unit of radiation intercepted by the canopy, and hence the rate of dry matter accumulation. After seed growth initiation, increase of harvest index is accelerated by ozone. This shortens seed growth duration and accelerates leaf abscission. The effects of ozone on the growth processes are calculated on the basis of specific ozone uptake rate, which is ozone uptake rate per unit leaf area. The ozone uptake rate is calculated from transpiration rate and ozone concentration assuming a proportionality of ozone uptake to transpiration. The model was calibrated to the results of the open-
top chamber experiments (Flagler, 1986). The effects of ozone and water stress on soybean growth and yield were satisfactorily described by the model, and the simulation indicated mitigating effects of water stress on ozone-induced yield loss. See Kobayashi et al. (1990) for further details of the model. Data The open-top chamber experiments, to which the model was calibrated, were conducted in 1983 and 1984. Soybean 'Davis' was exposed to ozone in the open-top chambers (Heagle et al., 1973). The experimental design was a factorial combination of two moisture regimes with either four (1983) or six (1984) ozone levels. Each combination of the treatments was completely replicated in either three (1983) or two (1984) blocks. The ozone levels were charcoal-filtered air (CF), nonfiltered air (NF), and NF to which constant amounts of ozone were added, which are referred to as NF30 and NF60 in 1983, and NF15, NF30, NF45, and NF60 in 1984, where the two digits after 'NF' represent nominal additions of ozone in ppb (10 -9 liters liters ~). Soil moisture regimes were well-watered (WW) and water-stressed (WS). In both years, the WW plots were irrigated with 2.5 cm of water when soil matric potential measured with tensiometers reached -0-03 MPa. The WS plots were irrigated with 2-5 cm of water when leaf water potential measured with psychrometers reached -1.8 MPa (1983) or when soil matric potential estimated from a neutron-prove measurement reached -0.5 MPa (1984). For further details of the experiments, refer to Flagler (1986) and Heagle et al. (1987). Weather data were obtained from the North Carolina State University Weather Data Acquisition System located adjacent to the experimental plots. Daily amounts of precipitation and dates and amounts of irrigation were obtained from the field records. Daily 7-h mean ozone concentrations during the ozone exposure were calculated from the NCLAN data file. Parameter values adopted in the simulation have been described by Kobayashi et al. (1990). Soil moisture was measured with tensiometers daily and with a neutron probe during the dry period for 1983 and 1984. Whereas the tensiometer readings were expressed in MPa, the neutron probe data were converted to the gravimetric soil water content (%) based on a laboratory calibration. Details of the soil moisture measurements have been published (Flagler et al., 1987). The gravimetric soil water content at a depth of 22.5 cm was subjected to Tukey's multiple range test for the effect of the ozone treatment with the day of each measurement as a block. In 1983, 17 measurements were performed, but only 6 in 1984. The tensiometer readings for well-watered plots were averaged over the whole period of measurements for each ozone treatment, and the differences between the means were statistically tested. The number of measurements averaged was 98 (1983) and 105 (1984). Tensiometer data were also available for water-stressed plots, but are not addressed in this paper, since the soil water potential often dropped
Effects of 03 and water stress on soybean
41
ozone treatments. Soil water content was given as input data to the model rather than the result of the calculated water budget as in the original simulation, and the same values of soil water content were used for all ozone treatments in the simulation for each soil moisture regime. In this situation, the simulated plants are subjected to the same level of water stress regardless of their water usage. Thus, the alleviation of water stress at the higher ozone level is eliminated, and the interaction, if any, is ascribed solely to the reduction in the ozone uptake rate due to water stress. The model structure for the original simulation, which was shown in Fig. 1 o f Kobayashi et al. (1990), was changed to accommodate the modified simulation. The modified model structure is shown in Fig. 1 of this paper. In order to simulate actual day-by-day variations in soil moisture conditions, the input data for soil water content were taken from a result of the original simulation which was based on the calculated water budget. The outputs for NF30 treatments in each year were used within each moisture regime. Hence, the modified simulation gave the same results as the original simulation for NF30 treatments, but not for other treatments.
beyond the lower limit for the tensiometer measurement c. - 0 . 0 7 MPa. Simelation
As described in the previous section, the interaction between ozone impact and water stress on the yield of soybean can be caused either by the reduced ozone uptake due to stomatal closure under water stress, by the alleviated water stress under higher ozone concentrations, or by other physiological changes in the sensitivity of the plants to ozone. For assessing the contribution of the former two possible causes of the interaction, two simulations differing in the method of calculating soil water content were conducted. Resulting yield losses were compared between the soil moisture regimes. The other possible causes of the interaction, e.g. osmotic adjustment, were not addressed in this paper.
Original simulation The open-top chamber experiments reported by Flagler (1986) were simulated using actual data on precipitation and irrigation. Soil water content was calculated in the water budget part of the model from the water input data and daily amount of transpiration and soil evaporation estimated in the model. Thus, both causes of the interaction mentioned above are included in this simulation.
Internal ozone dose-yieM loss relationship In addition to the comparison between the two simulations, the effect of water stress on internal ozone dose was examined. The internal ozone dose was expressed by the ozone uptake rate per leaf area, and was calculated
Modified simulation The model simulated a situation in which soil water content is maintained at the same level across the Solar ,('- MODIFIED "~ ~k,WATER BUDGET~ FTSW
radiation Temperature 9 9 , ~ DRY MATTER "~ ~ ........ H~ i I~ I!ii ,.,-~r t = . u w , F(~--~.-~-~--~i . . . . . . . . . ~ Plastochronl I O, =r ~ - " :-~'},~DIR ~:.!;~)~- - ~ index-] [ , : ,K,a, ,,- ,• ..... ~ I:!t. x l " ~ _ ,- . . . . . . . . . . I Tra'nspirable t . . . . : : ~ 1 ~ I~ Vapor SO water [-'" : : ,;-r- P h ~ e l::; i f.q==nrz-taoutrru:'~ : ' pressure , , ', ! . ............. " ' . .... ~ . 1 ~ ,,' :, ' : ~ ~ ' SENESCENCE , ', aer,c,l I 3It" ........ Ld LJ i ii JL ..[" " '
,
: ',
," i
ct
,,
T ,'"'hStoredphoto-I
.-~Transpirati°n Transpiration I.~, ' ....-, "~
," ' ............................................~" ,,C"~r'('EFFECTSOF~
' :
"-,.-'/ ' h
,
,,I1"
o~
]
=
ill
, ,
Ill:
', centration". ] ......
¢'
materials
i
[
rymatter re-
~ ' , ,
~1~(,
I
m°bilizati°n I ~ i ....
"
::~ ', I:'
e !
', I: i
"
: I
'
............................................................,',
. . . . . . . . . . . . . . . l,.f~.l/ Nitrogen I .~N,-~t" S~F~ °:~'. " ~ fixation I L .......... ..v..-.'... 1 Leaf nitr°gen [ ......... ~-....... i
::
LEGEND Flow of
.................
s Ynthate
',
-'- ..... " ............ ~ ~ ' - ' ~ ' " - - ' " J J
,I-~---- ..... ~-I-}
: -"-'-'~ 1103----~con- [ 1/~ ......
!'~l
i ~ . ~ . ~ [ / ~
.q
Absorbed t'~--' I
I
~'~"
ou.co) ,. OZONE
~
" ' ....
i
:
!
F" ,
oration
r~.l~titrogen re_ i" ~FIow of
Auxiliary variableQ
j
~'~ mobilization | .....
information External variable . . . . . . .
-I,,,-
O
Fig. 1. Schematized relational diagram of the model for modified simulation. Only the major relations are shown. Each model component is surrounded by a rounded rectangle. In the modified simulation, daily value of the soil water content is given as the input data. Some of the processes are by-passed depending on the growth phase. Abbreviations are as follows: DIR = daily amount of short-wave radiation intercepted by leaves. EFFO = light utilization efficiency under ozone exposure, FTSW = fraction of transpirable soil water content, HI = harvest index, LAI = leaf area index, OU.co = coefficient for ozone uptake rate, SN.cont= seed nitrogen content, SNF.co = specific daily nitrogen fixation, WUE.co = coefficient for water use efficiency.
42
Kazuhiko Kobayashi et al. 1.1
in the original simulation. The relationship between the seasonal mean intemal ozone dose and the relative yield was examined for each water stress treatment.
1.0
RESULTS Simulated yields in the original simulation are shown in Fig. 2 for both years. The reduction in yield due to water stress was much more pronounced in the hot and dry year of 1983 than in the cooler year of 1984 although the yield for the well-watered plots was higher in 1983 than in 1984. In the original simulation, there was no effect of ozone on the simulated yield for the water-stressed plots in 1983 as shown on a relative basis compared to CF treatment (Fig. 3(A)). Water stress also decreased the impact of ozone on the yield in the original simulation in 1984 (Fig. 3(B)) although to a lesser extent than in 1983. Year-to-year variation in the dose-response relationship is apparent for the well-watered plots, even though the plots were well-irrigated in both years.
1983
A
.
0.9
0.7 "O (9 1o"
I
Orig.
--
Ong. WS
- , , Mod. WW
0.8
Interpretation of the simulation results In the previous paper (Kobayashi et aL, 1990), the model had been calibrated to the results of the open-top chamber experiments (Flagler, 1986). Values of the model parameters including the parameters for nitrogen fixation rate and for the effects of ozone have been determined so as to fit the data. Thus, the model is 'empirical' rather than mechanistic with respect to the parameters used for the 'fitting'. The same caution is needed in a predictive use of the model as is needed in the use of an empirical dose-response relationship, e.g. Weibull function. Under a situation which is quite different from the experiments, different values might be appropriate for the model parameters. Yet a comparison of the simulation results on a relative basis can be as meaningful as a comparison of the dose-response relationships. In this paper, the results of the open-top chamber experiments were simulated with the fitted model, and the relative yields were compared between the two simulations which differed in soil moisture calculation. The analysis can give us a suggestion, if not a confirmation, on the mechanisms underlying the ozone × water stress interaction.
~
- . - Mod. WS
~,~..
O.9
B
"~.? x
1984
tr
--
~%'%% ~" ""'"..\ ,,~
0.7 0.8 I 0.6
...... Mod. VWM
~,~
...... Mod. WS
~\
0.5
0.4 0.02
I
I
I
I
0.04
0.06
0.08
0.10
0.12
Seasonal mean 7-h/d Oa concentration (ppm)
Fig. 3. Comparison between the relative yields calculated in the original and in the modified simulations in 1983 (A) and in 1984 (B) under the two soil moisture conditions. The charcoal-filtered air treatment = 1. Abbreviations as in Fig. 2.
In the modified simulation, on the other hand, soybean yield for water-stressed plots responded to ozone dose quite similarly to well-watered plots in 1983, whereas no effect of ozone was found in the original simulation of water-stressed plots (Fig. 3(A)). Little difference was found between the original and the modified simulations for the well-watered plots. In 1984 (Fig. 3(B)), the difference between the two simulations was much smaller than in 1983. However, the doseresponse relationship for the water-stressed plots in the modified simulation was closer to that for the wellwatered plots than in the original simulation, particularly at the lower ozone doses. The dose-responses for the well-watered plots in the two simulations were almost identical in 1984 (Fig. 3(B)). The relationships between the mean internal ozone dose and relative yield in the original simulation are shown in Fig. 4 for each soil moisture treatment in 1.1
1.0" 0.9'
,o01 "
--1,83 "*"~%-
300
l
- 2001001 0.02
~
-
~,
-
--
1983 WS 1984 W W
a 0.04
0 ~. 0.06
~"XXx\.\" o.io
, - - 1984 WS o.12
Seasonal mean 7-hid 03 concentration (ppm)
Fig, 2, E f f e c t s of" o z o n e o n the s o y b e a n seed yields c a l c u l a t e d
in the original simulation under water-stressed (WS) or wellwatered (WW) conditions in 1983 and 1984.
~-1983
0.8' >~ _~ G) 0.7'
--
"4
rr
0.6 ~
t%, b
1983 W S
--,1984 --'
WW
WW
1984 W S
I
0.5 0.4
""
Orig. WW
- - Orig. WS
2
n
I 4
=
: 6
:
: : : : : ; 8 10 12 14 Seasonal mean 03 uptake rate per unit leaf area (rag m 2 d q)
Fig. 4. Relationship between the seasonal mean ozone uptake rate per unit leaf area, calculated in the original simulation, and the relative yield in 1983 and 1984. Simulated yields for the two soil moisture regimes with the same ozone treatment are interconnected by a dotted line. The charcoal-filtered air treatment = I. Abbreviations as in Fig. 2.
Effects of Os and water stress on soybean Table 1. Mean soil water content from the neutron probe measurements
Ozone treatment
Gravimetric soil water content (°/o)a 1983 WS
CF NF NF15 NF30 NF45 NF60
1984 WW
3.49 a b 4.64 b -6-10 c -5.15 b
8.14 a 9.52 bc --10.69 c -8.68 ab
WS
WW
4.98 b 3-47 ab 4.37 ab 2.82 a 3.67 ab 5.28 b
4-42 a 8.57 b 8-56 b 7.92 b 6.87 ab 6.77 ab
aMean values of 17 (1983) and 6 (1984) neutron probe measurements are shown. WW stands for well-watered plots and WS for water-stressed plots. b Means with the same letter are not significantly different with Tukey's studentized range test. 1983 and 1984. The water-stressed plants had smaller ozone uptake rate than the well-watered plants in the same ozone treatment as shown by the dotted lines in Fig. 4. As a result of the reduced internal ozone dose due to water stress, the discrepancy between the doseresponse curves for water-stressed plants and the curves for well-watered plants in Fig. 4 was somewhat smaller than that for the original simulation in Fig. 3, where the relative yield was plotted against ozone concentration. However, the reduction in the impact of ozone on the yield by water stress was still obvious in 1983. The gravimetric soil water content calculated from the neutron probe measurements is averaged over 17 (1983) and 6 (1984) measurements, and is shown in Table 1 for each ozone treatment. Soil water content was the lowest in the lowest ozone treatments (CF) for both well-watered and water-stressed treatments in 1983, and for well-watered treatment in 1984. Although the soil water content was significantly different a m o n g the ozone treatments for water-stressed treatment in 1984, the relationship between the ozone level and soil water content was obscure. Soil water potential measured with the tensiometers showed no significant effect of ozone for both years (Table 2). Table 2. Mean soil water potential from the tensiometer measurements
Ozone treatment CF NF NF15 NF30 NF45 NF60
Soil water potential (MPa) a 1983WW
1984WW
-0-0151 -0.0158 --0.0153 --0.0141
-0.0124 -0.0141 -0.0100 -0.0127 -0.0109 -0.0097
ab a a a
a a a a a a
a Seasonal mean tensiometer measurements are averaged across three (1983) or two (1984) blocks. Only the results for well-watered plots (WW) are shown. b Means with the same letter are not significantly different with Tukey's studentized range test.
43
DISCUSSION The results of the original simulation clearly showed reduced impact of ozone on soybean yield under waterstressed conditions in 1983. In contrast, the modified simulation, which does not include the effect of mitigated water stress under higher ozone concentrations, resulted in similar dose-response relationships for both well-watered and water-stressed plots. Similar but less clear results were obtained in the 1984 simulation. The contrast between the two simulations indicates that the reduced impact of ozone on soybean yield by soil moisture stress may be caused by the alleviation of water stress by higher ozone doses. The difference between 1983 and 1984 is apparently due to the different weather conditions in the 2 years. The 1983 season was hot and dry with very little rainfall during vegetative growth, whereas the 1984 season was wet through vegetative and early reproductive growth stages (Flagler et al., 1987). The difference in weather conditions may reflect different severities of water stress between the 2 years, and a change in the timing of the ozone exposure during the growing seasons. The neutron probe measurements indicated that the soil water content was lower in the lower ozone treatments. Since the total amount of irrigation was the same for all ozone treatments under water stress (Heagle et al., 1987), the difference in the soil water content for the water-stressed plots in 1983 has likely been caused by the differential water use of the plants as affected by the ozone treatment. The well-watered plots were not irrigated identically across the ozone treatment (Heagle et al., 1987), but the difference in the soil water content (Table 1) may also be a result of the difference a m o n g the ozone treatment in the water use rather than in the water input. Although the differences in the soil water content cannot be translated straightforwardly into the difference in plant water status, the amount of soil water available for plants was not much different a m o n g the ozone levels in well-watered plots as suggested by the tensiometer measurements (Table 2). Reduced ozone flux to leaves under water stress was accounted for by the model. However, the relationships between the internal ozone dose and the yield loss still showed the ozone x water stress interaction in 1983, when the impact of ozone on the yield was clearly decreased by water stress (Fig. 4). This result indicates that the reduced ozone flux to the leaves of waterstressed plants might not be the only factor determining the mitigation of ozone impact by water stress. In the simulation, the stomatal closure of the waterstressed plants reduced the ozone uptake by the leaves and, thereby, the impact of ozone on growth processes. However, the stomatal closure due to water stress also suppressed simulated growth and yield for CF treatments as well as other ozone levels. Thus, on a relative basis compared to the yield for CF treatments, reduced ozone impact due to water stress was partly cancelled out by the yield loss due to water stress in the simulation.
44
Kazuhiko Kobayashi et al.
Some results obtained through modelling and field experiments support the possibility that mitigation of water stress by ozone contributes to ozone × water stress interaction. King (1987) developed a model for the impacts of ozone and water stress on crop yield. The model calculated seasonal dry matter production as the product of seasonal transpiration and a coefficient, transpiration efficiency. The model included reduced ozone uptake due to drought stress and reduced water use under high ozone concentrations. The simulation for soybean yield losses suggested that the impact of ozone on the water use by plants, rather than the reduced ozone uptake due to water stress, might be a major factor in the interaction (King, 1987). Heggestad et al. (1988) investigated the effects of ozone on root development of soybean under two soil moisture regimes. They tried to maintain the same soil moisture level across the ozone levels, and hence more water was added to lower ozone treatments, where the plants used more water. Nevertheless, there was an effect of ozone treatment on soil water potential; the lowest soil water potential occurred in the lowest ozone level in the water-stressed plots, as is found in the experiments in this paper (Table 1). Under the lower ozone concentrations, root length density was much greater in water-stressed plots than in well-watered plots, whereas in the highest ozone concentration, root length density was almost the same for both soil moisture regimes. Their data on cultivar 'Williams' indicate significant yield loss due to water stress in lower ozone levels but not in higher ozone levels. These results on root growth and yield might suggest that the plants in lower ozone concentration were subjected to greater water stress than in higher ozone concentration in spite of the authors' effort to maintain the same level of soil water content across the ozone levels. Heggestad and Lee (1990) grew 'Williams' soybean in soil columns placed in greenhouses with charcoal filtered and nonfiltered air, and investigated the effects of ambient ozone and soil moisture stress on water use, growth and yield of the soybean plants. Because of reduced size and earlier maturity, plants in nonfiltered air used less water than plants in charcoal filtered air during the season. Water was added to maintain about the same soil water potential for both filtered and non filtered air plots. Nevertheless, among the water-stressed plots, the plot with charcoal filtered air had lower soil water potential during the experiment and lower soil water content at harvest than the plot with nonfiltered air. This lower soil moisture content in charcoal filtered plot with soil moisture stress might be a contributing factor in the interaction between the effects of ozone and soil moisture regimes on plant growth and yield. Although the interaction was not significant, the combined effect of ozone and water stress was less than the additive effects of both stresses. In contrast to the similar experiment in the open-top chambers (Heggestad et al., 1988), the soil moisture stress did not increase the root development for both ozone regimes. The differential root
response to water stress was attributed to the difference in the soil moisture profile between the experiments (Heggestad & Lee, 1990). In the open-top chamber experiment (Heggestad et al., 1988), there were significant amounts of stored water available at lower soil depth, whereas no water was available at lower depth in the greenhouse experiment (Heggestad & Lee, 1990). Stomatal closure due to water stress has been widely assumed to be a main cause of the reduction of ozone effects by water stress. Moser et al. (1988) exposed bush bean plants in containers to ozone, and imposed drought stress on the bean plants during the early or late reproductive stages. They found reduced effects of ozone on bush bean growth and yield under drought conditions, and assumed that decreased stomatal conductance has resulted in the decreased impacts of ozone based on their preliminary measurement of transpiration. Using a special cultural system, they could control the drought stress conditions in the plant containers rather precisely. Their system attained the drought stress by increasing the distance between root zone and the water table, thereby decreasing the hydraulic conductivity in the cultural medium. The water content in the root media might be decreased by increased water use of the plants under low ozone levels. Hence, the interaction that Moser et al. (1988) found could include the mitigation of water stress from ozone-induced reduction in water use in addition to the reduced ozone uptake through decreased stomatal conductance in the water-stressed plants. Heagle et al. (1988b) exposed cotton to different levels of ozone in open-top chambers under water-stressed conditions or under well-watered conditions. They found significant yield loss due to ozone in the wellwatered regime but not in the water-stressed regime. Although plots for all ozone levels in each soil moisture regime were irrigated with the same amount of water, there were no obvious differences in soil water potential among the ozone levels, partly due to variation between replicates. Therefore, the cause of the apparent interaction between soil moisture stress and ozone is not clear in this case. In summary, the modelling approach in this paper identified reduced water use due to ozone as a contributing factor in ozone x water stress interaction. Although the reduced ozone uptake due to water stress showed some contribution to the interaction, it was not the dominating factor in the case of the severe water stress studied in this paper. Other physiological responses which might be involved in the interaction were not addressed in this paper. Further experimental and modelling studies are needed to elucidate the mechanisms of the interaction between the impacts of water stress and ozone on crop yield. Measurements of ozone uptake rate and internal ozone dose for crop canopies under well-watered or water-stressed conditions will help to better understand the interaction. Experiments in which soil moisture content is precisely maintained at the same level across the ozone treatments could help clarify the mechanisms of the interaction.
Effects of 03 and water stress on soybean ACKNOWLEDGEMENTS We wish to thank D r T. R. Sinclair for his suggestions on the water stress-ozone interaction, D r A. S. Heagle for his help in the use of the field records and other data in the field experiment, Susan Spruill for database management, and D r S. Hasegawa for his discussion on the control of soil moisture content. We also appreciate helpful comments from the reviewers on this manuscript. Major part of this work was performed when the first author was visiting at N o r t h Carolina State University, Department of Botany as a visiting scholar. The visit was funded by the Science and Technology Agency of the G o v e r n m e n t of Japan.
REFERENCES Adams, R. M., Giyer, J. D., McCarl, B. A. & Johnson, S. L. (1989). A reassessment of the economic effects of ozone on U.S. agriculture. J. Air Pollut. Control Assoc., 39, 960-8. Amundson, R. G., Raba, R. M., Shoettle, A. W. & Reich, P. B. (1986). Response of soybean to low concentration of ozone: II Effects on growth, biomass allocation, and flowering. J. Environ. Qual., 15, 161-7. Flagler, R. B. (1986). Effects of ozone and water deficit on growth, yield, and nitrogen metabolism of soybeans. PhD thesis, North Carolina State University, Raleigh, NC, USA. Flagler, R. B., Patterson, R. P., Heagle, A. S. & Heck, W. W. (1987). Ozone and soil moisture deficit effects on nitrogen metabolism of soybean. Crop Sci., 27, 1177-84. Heagle, A. S., Body, D. E. & Heck, W. W. (1973). An opentop field chamber to assess the impact of air pollution on plants. J. Environ. Qual., 2, 365-8. Heagle, A. S., Flagler, R. B., Patterson, R. P., Lesser, V. M., Shafer, S. R. & Heck, W. W. (1987). Injury and yield response of soybean to chronic doses of ozone and soil moisture deficit. Crop Sci., 27, 1016-24. Heagle, A. S., Kress, L. W., Temple, P. J., Kohut, R. J., Miller, J. E. & Heggestad, H. E. (1988a). Factors influencing ozone dose-yield response relationships in open-top field chamber studies. In Assessment of Crop Loss from Air Pollutants, eds W. W. Heck, O. C. Taylor & D. T. Tingey. Elsevier Science, London, pp. 141-79. Heagle, A. S., Miller, J. E., Heck, W. W. & Patterson, R. P. (1988b). Injury and yield response of cotton to chronic doses of ozone and soil moisture deficit. J. Environ. Qual., 17, 627-35.
45
Heck, W. W., Cure, W. W., Rawlings, J. O., Zaragoza, L. J., Heagle, A. S., Heggestad, H. E., Kohut, R. J., Kress, L. W. & Temple, P. J. (1984). Assessing impacts of ozone on agricultural crops: II. Crop yield functions and alternative exposure statistics. J. Air Pollut. Control Assoc., 34, 810-17. Heggestad, H. E. & Lee, E. H. (1990). Soybean root distribution, top growth and yield responses to ambient ozone and soil moisture stress when grown in soil columns in greenhouses. Environ. Pollut., 65, 195-207. Heggestad, H. E., Gish, T. J., Lee, E. H., Bennett, J. H. & Douglass, L. W. (1985). Interaction of soil moisture stress and ambient ozone on growth and yields of soybeans. Phytopathology, 75, 472-7. Heggestad, H. E., Anderson, E. L., Gish, T. J. & Lee, E. H. (1988). Effects of ozone and soil water deficit on roots and shoots of field-grown soybeans. Environ. Pollut., 50, 259-78. King, D. A. (1987). A model for predicting the influence of moisture stress on crop losses caused by ozone. Ecol. Model., 35, 29-44. Kobayashi, K., Miller, J. E., Flagler, R. B. & Heck, W. W. (1990). Modeling the effects of ozone on soybean growth and yield. Environ. Pollut., 65, 33-64. Moser, T. J., Tingey, D. T., Rodecap, K. D., Rossi, D. J. & Clark, C. S. (1988). Drought stress applied during the reproductive phase reduced ozone-induced effects in bush beans. In Assessment of Crop Loss from Air Pollutants, eds W. W. Heck, O. C. Taylor & D. T. Tingey. Elsevier Applied Science, London, pp. 345~4. Sinclair, T. R. (1986). Water and nitrogen limitation in soybean grain production. I Model development. Field Crops Res., 15, 125-41. Temple, P. J. (1990). Water relations of differentially irrigated cotton exposed to ozone. Agron. J., 82, 800-5. Temple, P. J., Taylor, O. C. & Benoit, L. F. (1985). Cotton yield responses to ozone as mediated by soil moisture and evapotranspiration. J. Environ. Qual., 14, 55-60. Temple, P. J., Benoit, L. F., Lennox, R. W., Reagan, C. A. & Taylor, O. C. (1988a). Combined effects of ozone and water stress on alfalfa growth and yield. J. Environ. Qual., 17, 108-13. Temple, P. J., Kupper, R. S., Lennox, R. W. & Rohr, K. (1988b). Injury and yield responses of differentially irrigated cotton to ozone. Agron. J., 80, 751-5. Tingey, D. T. & Hogsett, W. W. (1985). Water stress reduces ozone injury via a stomatal mechanism. Plant Physiol., 77, 944-7. Tingey, D. T., Thutt, G. L., Gumpertz, M. L. & Hogsett, W. E. (1982). Plant water status influences ozone sensitivity of bean plants. Agric. Environ., 7, 243-54.