Difference between revisions of "Spatial schematization"

From Agri4castWiki
Jump to: navigation, search
Line 1: Line 1:
 +
{{scientific}}
 
The crop models in MCYFS where originally developed to simulate crop yield for a single location (or region) where weather, soil and crop characteristics are assumed homogeneous. To apply the crop models on a large scale the models have to be run for every location with a unique set of weather, soil and crop characteristics.
 
The crop models in MCYFS where originally developed to simulate crop yield for a single location (or region) where weather, soil and crop characteristics are assumed homogeneous. To apply the crop models on a large scale the models have to be run for every location with a unique set of weather, soil and crop characteristics.
  

Revision as of 14:42, 31 January 2014



The crop models in MCYFS where originally developed to simulate crop yield for a single location (or region) where weather, soil and crop characteristics are assumed homogeneous. To apply the crop models on a large scale the models have to be run for every location with a unique set of weather, soil and crop characteristics.

The weather and spatial and temporal crop characteristics are assumed to be homogeneous per grid cell of the 25x25km climate grid. The soil characteristics are assumed to be homogeneous per STU. The STU data are stored in table SOIL_TYPOLOGIC_UNIT. In other words: the models have to be run for every unique combination of grid cell and STU called simulation unit. The simulation units are stored in table SIMULATION_UNIT.

The smallest cartographic unit of the soil map is called a SMU which consists of various STU's. Only the relative amount of a STU within a SMU is known. All unique SMU's are stored in table SOIL_MAPPING_UNIT, the relation between STU's and SMU's is stored in table SOIL_ASSOCIATION_COMPOSITION. When combining the weather, crop and soil data layers we have to intersect the 25x25km climate grid with a map of the SMU's. This results in so-called EMU’s stored in table ELEMENTARY_MAPPING_UNIT.

Still the crop models have to be run for every simulation_unit. The result are aggregated to a values per EMU by a weighted average of the simulation units it consists off. The aggregated results per EMU are stored in the table CROP_YIELD. With the existing relation between EMU's and grid cells the results can be aggregated to grid cells (stored in table GRID_YIELD).

In the Yield Forecasting module the simulated results are compared on a regional scale. The regional yield statistics are normally available at national (level 0) or at provincial level (level 1 or 2). To be able to aggregate model output to administrative regions it is necessary to know the relation between EMU's and administrative regions. Therefore the EMU's are intersected with a map of the administrative regions (NUTS level 3). In this way results can be aggregated to any administrative level. The aggregated simulation results per administrative region are stored in the table NUTS_YIELD.

Spatial schematization. Intersection of weather map (25x25 km grid cells), soil map (SMU's) and administrative regions (NUTS level 3.