|
The geoland observatory on Food Security and Crop Monitoring produces three
main type of products:
- Crop indicators
- Crop yield forecasts
- Crop area estimates
Crop Indicators
Crop indicators provide a (semi-)quantative measure of the crop status over
the crop growth season and can be provided on grids or pixels or aggregated
to regions or countries. Various approaches can be used to derive crop indicators.
They can be obtained from satellite observations using vegetation indices, crop
models can be applied to simulate biomass development using observed weather
data and METEOSAT can be used to derive the actual crop growth conditions. All
crop indicators provide a (semi-)quantative measure of the crop status but they
do not yet provide a true yield estimate in ton/hectare. Moreover, they must
be carefully interpretated by comparing them to previous years or the long term
average in order to understand their properties and merit.
Vegetation Condition Index from SPOT VEGETATION
|
Dry Matter Productivity based on SPOT VEGETATION
|
Maize biomass based on model simulation using METEOSAT derived input
|
Maize biomass based on model simulation using weather input
|
Crop yield forecasts
Crop yield forecasts provide a quantitative estimate of crop yield in ton/hectare
on the level of districts, regions or countries. The crop indicators are aggregated
to statistical regions and regression techniques are used to determine a relationship
between a time-series of crop indicators and a time-series of past crop yields
(usually available from a national statistical office). The yield forecast should
be supplied with a measure of confidence (derived from the regression analyses)
and can be interpreted unambiguously. The yield forecast values can be provided
in tables and charts or can be visualised spatially.
|
|
|
Example yield forecast for entire Spain which shows that starting
in decade 15 a reasonable yield forecast can be made: forecast quality
values higher then 2 mean that a good quality forecast was obtained.
|
|
|
Example yield forecast for Spain showing the yield forecast at NUTS2
level for decade 20 in 2003. The trend forecast is the yield that is expected
on the basis of the long-term trend, while the actual yield forecast is
based on the crop indicators from the 2003 crop growth season.
|
Crop area estimates
Crop area estimates simply provide an estimate of the total area of a certain
crop grown in a certain area. Together with a yield forecast they can be used
to forecast the total production (in ton) for a region or country. Crop area
estimates over large areas based on satellite imagery have so far been difficult
because of the mismatch between the size of farmers field in many parts of the
world (typically 1-2 hectare) and the pixel resolution of satellites that are
capable of imaging large areas (typically 1 km). Only recently with the advance
of medium resolution satellite sensors (MODIS 250m en MERIS 300m) progress has
been made in carrying out crop area estimates through the growing season.
|
|
|
|
Landsat ETM image of an agricultural area in Russia showing the structure
of fields.
|
MODIS 250M image of the same area, demonstrating that the field structure
can still be largely recognised on the MODIS image.
|
|
|
Comparison between area statistcs and MODIS-derived area
estimates for an area in Russia.
|
|
