In this section, we focus on describing the model structure and the commonly
used crop growth model DSSAT. DSSAT model developed by International Benchmark Sites Network for Agro-technology Transfer (IBSNAT), which consists of
three components: database management module which used to enter, store, and
retrieve data set for model run; land unit module which simulate the effect of soil
processes on crop growth; and a set of crop models for simulating crop growth state
and biomass or yield. The model can simulate crop development process, soil water
balance, carbon and nitrogen processes, and crop management practices (Jones et al.
2003; Thorp et al. 2008). In DSSAT-CSM, different kinds of crops have a single set
of codes; this design feature simplifies the simulation of crop rotations, as Fig. 11.2
shows. Through inputting weather data and physiological/physical parameters in
DSSAT-CSM, users can obtain user-specified objectives. The CSM-CERES is one
set of crop models under DSSAT, such as CERES-rice and CERES-wheat models.
For example, the CERES-wheat model simulates the wheat LAI dynamics, dry
matter development, the water and nitrogen balances of the soil-plant-atmosphere
at a daily step, and the wheat yield.
Data preparation is the first step when using the crop growth model for crop
growth monitoring and yield forecasting and plays a key role for the simulation
result performance. The crop growth model has the input of climate, soil, and crop
management data, and the input data may be different for different complexity
models. For WOFOST, the climate input consists of daily maximum and minimum
temperature, solar radiation, wind speed, vapor pressure, and precipitation. Soil
input consists of soil moisture content of saturated soil, soil moisture content at the
wilting point, and soil moisture content at field capacity (Jones et al. 2003). In
addition, sowing date, the amount of irrigation, and fertilization as crop management
information also need to be collected. These data can be obtained from meteorological station observations and field experiments. For regional applications, the data
preparation can be obtained through the interpolation method. As the crop growth
model has a large number of parameters, parameter estimation based on limited
experimental data is usually considered a key issue that affects the model simulation
results with uncertainty. Therefore, model parameter sensitivity analysis is necessary
Table 11.1 The main crop growth models for crop growth and yield simulation
Model
Country
References
Websites
STICS
France
Brisson et al.
(1998)
https://www6.paca.inra.fr/stics_eng/
CROPSYST USA
Stöckle et al.
(2003)
http://modeling.bsyse.wsu.edu/CS_Suite/
CropSyst/index.html
WOFOST
Netherlands van Diepen et al.
(1989)
http://www.wofost.wur.nl
DSSAT
USA
Jones et al. (2003) https://dssat.net/
EPIC
USA
Williams et al.
(1983)
https://epicapex.tamu.edu/epic/
APSIM
Australia
Keating et al.
(2003)
http://www.apsim.info/
11 Crop Growth Modeling and Yield Forecasting
211
used crop growth model DSSAT. DSSAT model developed by International Benchmark Sites Network for Agro-technology Transfer (IBSNAT), which consists of
three components: database management module which used to enter, store, and
retrieve data set for model run; land unit module which simulate the effect of soil
processes on crop growth; and a set of crop models for simulating crop growth state
and biomass or yield. The model can simulate crop development process, soil water
balance, carbon and nitrogen processes, and crop management practices (Jones et al.
2003; Thorp et al. 2008). In DSSAT-CSM, different kinds of crops have a single set
of codes; this design feature simplifies the simulation of crop rotations, as Fig. 11.2
shows. Through inputting weather data and physiological/physical parameters in
DSSAT-CSM, users can obtain user-specified objectives. The CSM-CERES is one
set of crop models under DSSAT, such as CERES-rice and CERES-wheat models.
For example, the CERES-wheat model simulates the wheat LAI dynamics, dry
matter development, the water and nitrogen balances of the soil-plant-atmosphere
at a daily step, and the wheat yield.
Data preparation is the first step when using the crop growth model for crop
growth monitoring and yield forecasting and plays a key role for the simulation
result performance. The crop growth model has the input of climate, soil, and crop
management data, and the input data may be different for different complexity
models. For WOFOST, the climate input consists of daily maximum and minimum
temperature, solar radiation, wind speed, vapor pressure, and precipitation. Soil
input consists of soil moisture content of saturated soil, soil moisture content at the
wilting point, and soil moisture content at field capacity (Jones et al. 2003). In
addition, sowing date, the amount of irrigation, and fertilization as crop management
information also need to be collected. These data can be obtained from meteorological station observations and field experiments. For regional applications, the data
preparation can be obtained through the interpolation method. As the crop growth
model has a large number of parameters, parameter estimation based on limited
experimental data is usually considered a key issue that affects the model simulation
results with uncertainty. Therefore, model parameter sensitivity analysis is necessary
Table 11.1 The main crop growth models for crop growth and yield simulation
Model
Country
References
Websites
STICS
France
Brisson et al.
(1998)
https://www6.paca.inra.fr/stics_eng/
CROPSYST USA
Stöckle et al.
(2003)
http://modeling.bsyse.wsu.edu/CS_Suite/
CropSyst/index.html
WOFOST
Netherlands van Diepen et al.
(1989)
http://www.wofost.wur.nl
DSSAT
USA
Jones et al. (2003) https://dssat.net/
EPIC
USA
Williams et al.
(1983)
https://epicapex.tamu.edu/epic/
APSIM
Australia
Keating et al.
(2003)
http://www.apsim.info/
11 Crop Growth Modeling and Yield Forecasting
211