The crop simulation module of the MCYFS integrates the effect of weather, soil and crop characteristics on crop growth assuming that the influence of other factors like farm management and socio-economic factors is constant.
The heart of the module consists of the BioMA (Biophysical Models Applications) platform (for further information, see the Agri4Cast Resource Portal) in which different crop models simulate crop development. These results are analyzed to identify abnormal situations and are used as input for the yield forecasting module.
The module can be split into four topics:
The crop simulation takes three major factors into account: weather, soil and crop characteristics. For each of these factors input data are required. In order to have the output data also available for administrative regions such as countries or provinces also these should be known to the system. The geographical relation between the weather, soil, crop characteristics and administrative regions is described by the spatial schematization.
The input of weather data is calculated in the weather monitoring module that delivers for each day and each 25x25km grid cell a unique set of weather indicators. Two versions of weather data are used in parallel:
- based on weather observations (OBS) measured at weather stations
- based on model weather calculated by a forecast agency. Different weather models are applied.
The input weather datasets are combined to cover a full crop season and in some cases extend into the future. The following combinations are implemented:
|Abbriviation||Archive simulations||Near real time simulations|
|OBS||Observed weather||Observed weather|
|OBS+OPE|| Observed weather|
extended with 10 day deterministic forecast HRES
extended with analysis HRES
|HIS+OPE|| Analysis HRES|
extended with 10 day deterministic forecast HRES
|HIS+ENS|| Analysis HRES|
extended with 15 day ensemble prediction system
|HIS+ENSEXT|| Analysis HRES|
extended with 32 day ensemble prediction system extended
|HIS+SEAS|| Analysis HRES|
extended with 183 day seasonal forecast model
Each crop has its own crop parameters that influence the response to weather and soil characteristics. They broadly fall into two categories. The first category describes the growth behaviour of each plant like partitioning of assimilates over plant organs and the second category describes the spatial and temporal variation in crop use, for instance which crop variety is used in a particular location and what the average sowing dates are for this location. For crops which are harvested before maturity (green crops) the average harvest dates are specified as well. The parameters of the second category are specified on a 25x25km resolution, similar to the resolution of weather data. Examples of included crops are wheat, grain maize, sugar beets, potatoes and more.
The soil influences the water available for crop growth. First of all it defines the physical limitations for roots (the rooting depth). And secondly soil properties like texture influence how much water the soil can store per meter rooting depth (water capacity). For example a loamy soil can store more water than a sandy soil. Hence the soil determines the maximum amount of soil water available at the start of the growing season (the initial soil water) and how efficient rainfall comes available for a crop. For instance in case of a sandy soil, with a low water capacity, part of the rainfall might be lost because not all water can be stored in the soil. This is especially important for climates with insufficient and irregular distributed rainfall.
To use simulated crop indicators in the quantitative Yield Forecasting module where they are compared to officially reported yields per region (for instance per country), they have to be aggregated to administrative regions. Inside the European window, administrative regions follow the system. Outside Europe (Russia, Kazakhstan, China, India) the system is used. Both systems have regions organized on different administrative levels. The highest level represent the countries and is referred to as level 0. Countries are subdivided in macro regions (level 1, e.g. departements). Macro regions are subdivided in sub-regions (level 2, e.g. districts) and level 3 regions refer to the finest resolution. In the case of Russia and Kazakhstan an extra 0.5 level is added (Okrugs, Oblasts).
The implemented crop models are originally developed to simulate crops for a single location or small region where weather-, soil- and crop characteristics are assumed homogeneous. To apply the crop models on a large scale, the crop models have to be run for every location with a unique set of weather-, soil- and crop characteristics. To find these unique spatial simulation units, the spatial features of weather-, soil- and crop data sets are intersected. Within each spatial simulation units it is assumed there is no spatial variation in the relevant input data. On top of this, the spatial simulation units are intersected with administrative regions to be able to aggregate simulations results to administrative regions (used in the quantitative Yield Forecasting module).
The crop models of the MCYFS are implemented in the BioMA Framework. Some models are generic and can simulate various crops (WOFOST, CropSyst). Other models are designed for a specific crop (WARM for rice, CANEGRO for sugarcane). The following combinations of regional windows, crops and input weather datasets are implemented in the production lines:
|Window||Crop model||Simulated crop||Input weather|
|Europe (EUR)||WOFOST||Winter wheat|| OBS|
|Russia Kazachstan (RUK)||WOFOST||Winter wheat|| ERA+HIS|
|China (CHN)||WOFOST||Winer wheat|| OBS|
|WARM||Rice (early season)|
|Rice (middel season)|
|Rice (late season)|
simulates daily plant growth. The model calculates intercepted light and convert this into plant matter (potential gross photosynthesis). When there is not enough soil water available the potential plant matter is reduced (actual gross photosynthesis). After substracting maintenance costs (maintenance respiration) the newly formed plant matter is distributed over different plant organs: roots, stems, leaves and storage organs (grains/tubers). Depending on the age of the plant ( which is driven by temperature) the different organs receive different shares. During the conversion from plant matter into organs part of it is lost as construction costs (growth respiration).
Early in the season, most of the plant matter is invested into roots, stems and leaves. With more leaves also more light can be intercepted and plant growth increases rapidly. Towards the end of the season all or most of the plant matter is converted into storage organs and leaves start to decay therefore reducing light interception and thus reducing plant growth. Grains are ripening.
is a user-friendly, simplified daily time step model for simulation of growth and development of paddy rice crops. The model accounts for all the main processes which characterize this peculiar system in a balanced way. It simulates biomass growth, leaf evolution, phenology events, floodwater effect on the vertical thermal profile, blast disease, cold-shock induced spikelet sterility during the pre-flowering period and hydrological peculiarities of paddy soils.
is a user-friendly, conceptually simple but sound multi-year multi-crop daily time step simulation model. The model has been developed to serve as an analytic tool to study the effect of cropping systems management on productivity and the environment. The model simulates the soil water budget, soil-plant nitrogen budget, crop canopy and root growth, dry matter production, yield, residue production and decomposition, and erosion. Management options include: cultivar selection, crop rotation (including fallow years), irrigation, nitrogen fertilization, tillage operations (over 80 options), and residue management.
simulates sugarcane crop growth and development from daily weather data, cultivar and soil properties, and management input data. It simulates canopy development at the tiller and leaf level, radiation capture from leaf area index, the water balance using soil-plant-atmosphere continuum principles, biomass accumulation following a radiation use efficiency/respiration approach and biomass partitioning to different plant components, including stalk sucrose, using a sourcesink approach and affected by physiological age, temperature and water stress.
Calibration of crop parameters
Calibration of crop parameters within the MCYFS is the process of (locally) tuning crop parameters in such a way that observed (measured) crop data match the simulation results as close as possible while keeping parameters values within realistic ranges.
Crop growth model results such as 'above ground biomass', 'biomass of storage organs' and 'crop development stage' provide insight in the crop season conditions. The primary result are available for each spatial simulation unit. To answer questions like:
'What is the average simulated above ground biomass in France for winter wheat?'
an area weighted aggregation to 25x25 km grid resolution and administrative regions is applied. The simulated yields must be available for administrative regions in order to make a regional forecast. The relation between simulations units, 25x25 km grid and administrative regions is described by the spatial schematization. Because crop yield simulations are based on observed weather or various ECMWF forecasted models (HIS, OPE, EPS, MON and SEA), separate aggregation products are available for each of these 6 alternative weather data sources. For the grid resolution additional indicators are calculated; especially indicators that have a link with the weather indicators and crop indicators. For instance the number of heat waves around a crop development stage or the rainfall around sowing.
Long term average
Comparing current year indicators with long term average indicators can reveal early or late phenological stages or more or less than normal biomass accumulation. In this way anomalies in regional weather conditions and the consequences for crop growth are detected early, providing a basis for alarm warnings or relief measures. (van Diepen and van der Wal, 1995).
Long term average values are calculated over the period 1975 until last year. Averages are calculated for simulated yields based on observed weather and based on forecasted weather (HIS model); both in 6 resolutions: EMU resolution, 25x25 km grid resolution and 4 levels of administrative regions.