Sorry, this page is still under construction

A Simulation of Phenology, Growth, Carbon Dioxide Exchange and Yields under Ambient Atmosphere and Free-Air Carbon Dioxide Enrichment (FACE) Maricopa, AZ, for Wheat

ABSTRACT

The impact of increased atmospheric CO2 concentration on the growth and productivity of field grown wheat has been evaluated. Meteorological and soil information from this study were used to validate a model (DEMETER) for simulation of vegetation response to climate change scenarios. The model simulations of phenology, carbon exchange rate, growth and yield for the treatment conditions of the experiment show a reasonable accordance with the experimental data.

Keywords: CO2, Wheat, Modeling, C-Assimilation, Phenology

INTRODUCTION

For the next decades we expect a continuosly rising atmospheric CO2 concentration ([CO2]). To improve our insight into the impact of Global Change on the earth´s vegetation, a sound knowledge of the complex interactions of vegetation, atmosphere and soil is necessary. Recent advances in our understanding of biological and micrometeorological processes have allowed us to write reasonably robust mathematical descriptions of these processes (Goudriaan, 1977; Farquhar et al., 1980; Farquhar et al., 1982). The development of such mathematical descriptions are a prerequisite for more mechanistic approaches to modeling (Long, 1991; Grant et al., 1993) which in turn are essential if we are to have any realistic hope of modeling the complex interactions of entire ecosystems (Grant et al. 1993, Amthor 1993). The effect of increasing [CO2] on wheat and other important agricultural crops has been studied, initially with laboratory and open-top chamber experiments. Although these experimental approaches have provided us with some valuable information, there remaine serious questions concerning their unnatural micrometeorological and soil environments. Beginning in the late eighties, the Free-Air Carbon Dioxide Enrichment (FACE) technique has offered the possibility of studying the complex effects of higher [CO2] in an otherwise natural environment (Hendrey, 1993; Hendrey & Kimball, 1994). In 1992/93 the first FACE Wheat experiment was conducted at Maricopa, AZ, to investigate the complex effects of increasing [CO2] on the world´s most important food crop. Experimental and modeling activities have been included in this international project. This collaboration has allowed us to evaluate a wheat model, developed for European weather conditions and ambient [CO2], with a data set which includes the influence of elevated [CO2] under the most realistic field conditions achieved to date.

MATERIALS AND METHODS

The Free-Air Carbon Dioxide Enrichment (FACE) Wheat-Experiment Maricopa, Arizona 1992/93

The study was conducted in a rural setting about 50 km south of Phoenix, Arizona at the Maricopa Agricultural Center of the University of Arizona. Wheat (Triticum aestivum L. cv Yecora rojo) was planted on 15 December 1992 in an east-west row orientation with 250 mm row spacing. After emergence plant, populations were determined to be about 130 plants m^-2. Immediately after planting, FACE apparatus was installed to enrich the air in four replicate 25 m diameter circular plots. The apparatus consists of rings constructed from 30 cm diameter pipe. The rings had 2.5 m high vertical pipes spaced every 2 m around the periphery. Air enriched with CO2 was blown into the rings and exited through holes in the vertical pipes. Wind direction, wind speed, and CO2 concentration were measured at the center of each ring. This information was used by a computer control system to maintain [CO2] at 550 µmol·mol^-1 at the center of each ring. The design of the system has been described by Lewin et al. (1993) and its performance evaluated by Nagy et al. (1993). The CO2 enrichment system was in operation continuously (24 h a day) from 1 January through 16 May 1993, except for about 2 weeks near the end of January when the enrichment was for 8 hours a day centered about noon, in order to conserve our CO2 supply when rains curtailed deliveries. There were four replicate control rings at ambient [CO2] also installed in the field. The field was irrigated with a subsurface drip irrigation system with the drip tubes spaced 0.50 m apart and at a depth of 0.20 m, with emitters spaced 0.30 m along each tube. One-half of each replicate plot received enough water to meet total evaporative demand (Wet treatment) while the other half plot received 50% of what was applied to the first half (Dry treatment). Nitrogen was broadcast before planting, and followed by three applications thru the irrigation system during plant development (tillering, beginning heading, beginning ripening). A combination of biological and chemical methods were used for pest control. The model can account for losses due to pest or disease, but in this simulation it was assumed that there were no losses due to these influences. Further details of the experimental design and equipment can be found in Wall & Kimball (1993) and Pinter Jr. et al. (1995).

The Wheat Ecosystem Model DEMETER

The wheat ecosystem model DEMETER (Kartschall et al., 1990), originally developed for winter wheat in Europe, was used in this study. There are two current model versions, one on a daily time step base, and a more detailled version on an hourly time step base. The model consists of the following submodels:
- aboveground dynamics (phenology, C-assimilation, respiration, growth, senescence, grain filling),
- main soil processes (H2O-, C-, N-dynamics),
- belowground dynamics (root distribution and activity, water/nutrient uptake), and
- Soil-Vegetation-Atmosphere-Transfer dynamics (exchange of radiation, latent and sensible heat flux (cf. Grossman et al., 1995).
The principal equations for the processes in the model are listed in Appendix A.

Phenology
Phenological development, which can be described with different scales, is primarily temperature dependent (Fig. 1), modified by water and nutrient shortages. The most frequently used scales of phenology were defined by Feekes (1941) and Zadoks et al. (1974) (cf. Tab. 1). Canopy temperature is the driving variable for the rate of development in the hourly based version of the model; in the daily based version of the model a corrected mean daytime air temperature is used (Eqs. 4, 5). As model internal unit of Phenology the Feekes-scale (with the unit [Fe]) is used. The model internal development stage can be transformed to each other comparable scale, eg. to the more frequently used Decimal Code (DC) originally defined by Zadoks et al., (1974). For the comparison of important development stages on both scales see Tab. 1. Phenology plays a major role in activating and controlling all plant-model internal processes (eg. beginning of stem elongation, anthesis, beginning of grain filling, ripening).

Photosynthesis
The Photosynthesis module of the model was developed as a slight modification of the photosynthesis model derived by Long (1991) from the model by Farquhar et al., (1980). The key processes in this module are
- Light absorption by an effective green leaf area for sun and shaded leaves
- Energy transfer to the Rubisco for carboxylation
- Carboxylation of Ribulose1,5bisphosphate by Rubisco.
Photosynthetic Photon Flux Density (PPFD) is calculated from the total solar radiation. PPFD is used in a simple one layer radiation transfer model with exponential extinction as the energy source of photosynthesis. Internal CO2 concentration is calculated using results of the energy and gas exchange module for the hourly version (Goudriaan, 1977) or using the empirical approach of Long (1991) for the daily version (Eq. 17). In the current hourly version there was an unlimited Rubisco avalability assumed. For the daily version the CO2 effect is postulated as a weak perturbance on carboxylation (Eq. 19) and water use efficiency (Eq. 20).

Growth and Respiration
The free assimilates are used in an hierarchic sequence to satisfy the following demands one after the other:
- maintenance respiration (Eq. 6)
- growth respiration (Eq. 7)
- biomass growth (with grain as the main sink of the biomass growth; Eq. 8)
- grain filling directly from assimilate pool, if biomass growth is insufficient for grain filling.

Respiration is divided into four categories
- maintenance respiration
- growth respiration
- both, during night and day,
with different coefficients for energy production efficiency and growth equivalents (Thornley, 1970; Thornley, 1977; Penning de Vries et al., 1974; Penning de Vries, 1975).
The maintenance demands can be fullfilled by using
- free assimilate pool and/or
- translocation of living plant biomass, if demand is unsatisfied by the available pool of free assimilates.
Assimilate consumption for maintenance respiration depends on the total living plant biomass (aboveground and roots) and is driven by canopy temperature (Eq. 6).

Growth respiration is proportional to the synthesis of new biomass dependent upon the growth equivalents of each biomass type with feedback control of synthesis. These processes are described as proportional to the growth respiration with different respiration/growth equivalents (Eqs. 7, 8). The calculated biomass growth is available for
- direct grain filling (after anthesis)
- aboveground biomass growth
- root growth and exudates.

Grain filling
During grain filling assimilate demand and distribution is driven by sink activity and canopy temperature. Sink activity depends on phenological stage (Eq. 30). After anthesis grain becomes the strongest sink for free assimilates and synthesized compounds. The current rate is limited by the following sequence as long as the residual demand is unsatisfied:
- current biomass growth
- available assimilate and synthate pools
- translocation fluxes from photosynthetical and root biomass with lower efficiency compared to the processes listed above.

Senescence
Senescence is primarily driven by phenology, modified for stress events (Eq. 22). This monotonic process leads to the irreversible conversion of photosynthetically active biomass into brown biomass. The phenology dependence is weak during the early development stages but increases strongly after anthesis. An additional acceleration by elevated temperature above 25°C is implemented in the model.

Belowground dynamics
The root biomass and length dynamics were observed by Wechsung et al. (1995). There is no comparison made between simulated and observed data in this paper.

Soil Water and Temperature Dynamics
Drip irrigation can provide excellent control of water application to the soil; however due to variations in the depth of the drip emitters it may be difficult to simulate soil water content - particularly in the soil near the soil surface and on a short-time averaged scale (min or h). The soil water dynamic is calculated as the solution of the Richards equation by means of an explicit numerical procedure (cf. Grossman et al., 1995). The soil temperature is calculated as the solution of the heat conduction equation by means of an implicit numerical procedure (Suckow, 1986).

Pests and Diseases
In the field experiment, optimum cultural practices were followed; eg. water (for the wet plots), fertilizer applications, disease and pest control. Therefore, all these influences were negelected for the simulation.

Soil-Vegetation-Atmosphere Exchange processes
A detailed description of energy and water transfer can be found in Grossman et al. (1995).

RESULTS AND DISCUSSION

Because, it was originally developed for a European winter wheat variety (cv Alcedo), the wheat model required changes in parameters for phenological response (Fig. 1) as well as the vernalization demand. Local latitude, elevation and soil parameter sets (Kimball et al., 1993a) were also necessary to run the model. Selected simulation results will be discussed briefly in this section in the same order as in the Method and Materials section. To show the causal chain of the especially investigated CO2 effects, a short description of it is added.

Expected direct and indirect effects of elevated atmospheric CO2 (causal chain)

The elevated [CO2] leads to an increase of leaf internal CO2 concentration (Ci) A partial stomatal closure through higher Ci was presumed to decrease stomatal conductance and transpiration per unit of leaf area (Goudriaan, 1977; Ball et al., 1987; Cardon et al., 1994).

This, Ci induced deficit of latent heat flux per unit leaf area was, in turn, assumed to affect the gas, energy, and water exchange, and to result in higher primary photosynthesis, reduced water demand per unit of leaf area and elevated plant temperatures of the crop under conditions of enriched [CO2] through the following causal chain.

Increases in stomatal resistance leads to lower water consumption per unit leaf area and per unit of fixed CO2. The water use efficiency increases. Linked to this, the specific latent heat deficit in the energy balance per unit leaf area leads to a new equlibrium in the energy exchange, and a higher leaf or canopy temperature can be expected, as calculated in the hourly version of the model as well as observed by Kimball et al. (1993b).

Canopy temperature affects the internal physiological processes of the plant in a complex way: The following can be assumed for increased canopy temperatures: - generally higher activities for most of the biochemical processes - accelerated phenology below the optimum temperature (Fig. 1) - influence (positive) on photosynthesis activity (possibly small compared to the direct effect) - higher maintenance respiration - accelerated growth.

Thus, higher canopy temperature leads to an acceleration of phenological development, a higher level of plant respiration, and a shortened grain filling period for the FACE treatments.

Following the model equations given by Farquhar et al., (1980, 1982), higher Ci increases the carboxylation and depresses the oxygenation activity of Rubisco, so that a primary higher gross rate of photosynthesis is expected (see Eqs. 4, 10). The different limit cases of carboxylation (cf. Farquhar et al., 1980) could not be distinguished in this study. This causal chain also does not contain possible cumulative effects and second order effects, which are expected as sink limited depressions for the assimilates or source limited depressions by nutrient shortage etc. (cf. Stitt, 1991; Stitt, 1993). In the model, the sequence of the effects mentioned above is realized - directly for Ci only - generically, without explicit influence for all following processes using the complexity of the process interactions in the model.

Effects of elevated atmospheric CO2 concentration on selected processes of plant physiology

Phenology
Phenological development exhibited small but growing differences between simulations for the four treatments. Water deficits induced accelerateddevelopment in the dry plots, by direct water stress as well as indirectly through increased canopy temperature (Eqs. 4, 5) caused by latent heat deficits. There is also a clear CO2 effect on the model results in the form of higher canopy temperatures for the [CO2] enriched plots. For the FACE Wet plot the acceleration amounted to the time of ripening reduced by 5 days compared to the CONTROL Dry treatment. Thus, in the simulations there were accelerated senescence and a shortened grain filling period. These results are similar to those recorded in the field (Figs. 2, Tab. 2).

Photosynthesis
The simulated diurnal courses of canopy CO2 exchange (Eq. 9, cp. Figs. 3) show for three plots (CW, FD, FW) a good agreement with the data observed by Garcia et al., 1993. Deviations for all treatments accure immediately after sunrise and before sunset. These can be explained because a weakness in the current version of the model is under conditions of rapidly rapidly changing light. As a possible reason for these relatively high deviations between observed and calculated data, a general overestimation of the PPFD effect on photophosphorylation for unsaturated light can be assumed. For the nearly light saturated conditions, the simulated phosphorylation response should reflect the real situation. Additional larger differences were found for the CONTROL Dry version (ambient CO2 ; 50% of well watered irrigation), beginning immediately after sunrise. The model probably underestimated the effect of water shortage on stomatal closure, and ultimately on photosynthesis and gas exchange. Another reason could be the overestimation of water supply, possible by overestimated - soil water contents and/or - root biomass/activity during this day.

Growth and yields
Higher photosynthetic activity and the reduction in water use per unit living biomass led to higher biomass in both FACE treatments compared to the CONTROL treatments (Figs. 4, Tab. 3). The grain filling normally takes place (and in this experiment also) under relatively high solar radiation, leading to the highest additional warming of the canopy caused by elevated [CO2]. The main reason for the partial reduction of the advantages under higher [CO2] compared to the CONTROL plots during grain filling can be assumed the acceleration of physiological processes driven by the elevated canopy temperature. Thus [CO2] induced higher rates of phenological development after anthesis and the higher temperature by itself leads to a more rapid transformation from green to brown biomass and shortend the grain filling period. Therefore the ratios between FACE and CONTROL of final grain yields for the Wet plot was lower than this ratio for living biomass at anthesis (Tab. 4). Especially during the late stages of development, the plant physiology, from the view point of agronomy and ecology, seems to be negatively affected by elevated CO2 as a result of complex interactions and higher order negative feedbacks. This demonstrates the complexity of the so called ’’CO2 fertilizer effect’’.

CONCLUSIONS

Both hourly and daily time-step versions of the model were able to simulate the behaviour of the wheat canopy in good agreement with observations. More detailed information about energy, water and gas exchange is made available by the hourly version, but the main macrovariables describing plant development, C-assimilation, grain yield, etc., have been reflected in the reduced daily version too. To study detailed processes, the hourly version is preferred. For regional aspects and aggregated tasks, the daily version is more practical. The model-based description of the combined interaction of different processes and the comparison of experimental and simulated results gives a good understanding of the CO2 effect on wheat, and, more generally, on vegetation. The assumed generic causal chain has been found, in principle, in both real and model behaviour. These interactions seem to be realistic for C3 plants, but the current quantative relations should be modifiedfor different species. The results of experiment and simulation show the complex interactions of the positive and negative effects of elevated [CO2], as seen from the perspective of ecology and agronomy. A tendency to higher water use efficiency, biomass production, and final yields was found under higher [CO2].

ACKNOWLEGDEMENTS

This research was supported by the Federal Science and Technology Ministry (BMFT) of the Federal Republic of Germany within the Projects BMFT/BEO 0339626-93 and BMFT/USF 01LK9107-5-91 and by the Agricultural Research Service, U. S. Department of Agriculture.

We also acknowlegde the helpful cooperation of Dr. Roy Roschkolb and his staff at the Maricopa Agricultural Center. The FACE apparatus was furnished by Brookhaven National Laboratory, and we are grateful to Mr. Keith Levin, Dr. John Nagy, and Dr. George Hendrey for assisting in ist installation and consulting about ist us.

This work contributes to the Global Change of Terrestial Ecosystems (GCTE) Core Research Programme, which is part of the International Geosphere-Biosphere Programme (IGBP).

Thomas Kartschall's homepage
PIK-FACE homepage