According to the last IPCC report (2013), agriculture is among the Climate-Change most vulnerable sectors. Crop models are the right tool to evaluate agricultural climate impacts.
The negative climate impact would be higher in regions with low-income family agriculture. This type of agriculture strongly depends on weather conditions. However, climate impacts affect also well funded agriculture. Climate risks usually relate to losses due to climate variability and extreme events.
Climate impact assessments, based on simulations from crop models, might help to increase climate resilience.
Water scarcity and consequent irrigation limitations is one of the most important agricultural climate risks. Improving irrigation management therefore is imperative. Irrigation is expensive, but absolutely necessary in many tropical and subtropical areas. Farmers in water scarcity areas must adapt irrigation management to actual crop needs and soil properties.
Agricultural models simulate crop growth, water use and other variables that facilitate understanding the climate effects on agricultural processes and therefore decision making.
The models can be classified as Empirical, Functional and Mechanistic, as shown in the following Table.
|Obtained from regressions between local variables, weather, etc.||Combining pphysical laws and semi-empirical relationships||Based on the physical laws of the soil-water-plant-atmosphere system|
|Valid only for the conditions and places where they were obtained||Realiable for most practical conditions, prior local calibration||Valid for any weather and soil conditions|
|Require few variables, generally available||Need semi-detailed information, available in many places||Need detailed information, usually not available|
“Mechanistic” or physical-based models rely on the physical laws of all processes involved. They simulate crop growth, evapotranspiration, water and nutrient uptake by roots, soil water movement and many other processes. However these crop models are usually complex. They need many input variables for their simulations.
There are also empirical models, which generally use statistical relations. These statisitcial relationships are difficult to extrapolate. Finally, the so-called “functional” models, combine both approaches. These crop models are simple in terms of implementation, while retaining an adequate technical basis. The latter have been the most used .
However, the model CROPSYST has been more often considered for simulating agricultural climate risks in Europe. Other important crop models are the Dutch mechanistic model WOFOST, regularly used by JRC for crop forecasts in Europe, as well as the French functional model STICS.
Although they have been extensively used in climate impact assessments, functional models have limitations simulating climate effects in water resources.
DSSAT models follow the “cascade approach” for soil-water movement simulation. This is a simplification that ignores other contributions to water balance such as capillary rising and runoff.
The “cascade approach” does not consider upward water flux in sites with shallow water tables. This approach does not simulate the impact of excessive rains either. In such cases crop models based on Richards equations should be used. Models such as SWAP are more suitable.