CM Box User GuideMain Page | About | Special pages | Log in

Printable version | Disclaimer | Privacy policy

2.1. The two basic modelling options: grid-based and station-based


Peter Hoefsloot, René Gommes

Two basic modelling options

A crop monitoring program has to be based on data from different locations inside the country or region that needs to be monitored. Technically these monitoring locations are points. Points can be pinpointed at the earth’s surface with coordinates. The coordinate system that is used in crop forecasting is called a geographic coordinate system whereby the coordinates are expressed in Longitude - Latitude pairs.

Example : the geographic coordinates of Pnom Penn (Cambodia) are:

  • Longitude: 104.917445
  • Latitude : 11.558831

Normally expressed as 104.917445, 11.558831 (longitude always first). Read more: Geographic Coordinates

Depending on the circumstances, the crop forecasting network can be based on:

  • An irregularly spaced network based on agrometeorological stations
  • A regularly spaced grid

An irregularly spaced network based on agrometeorological stations

An example of such a network is shown below. The picture depicts rainfall recording stations in Africa. The main advantages of this type of monitoring network are:

  • Input and output of the water balance model can be checked against the situation in the field.
  • Station weather data are more accurate than gridded data, certainly in the vicinity of the stations themselves.
  • Usually requires less computer power due to lower number of monitoring points.
  • Although the water balance output of an irregularly spaced network is irregularly spaced as well, these data can be transformed into grids or images using interpolation techniques. This enables the estimation of water balance parameters for all parts of the monitoring area.
Monitoring stations in Africa

A regularly spaced grid

In this type of network, grid points do not coincide with meteorological stations. It is particularly useful when ground data are scarce or inaccurate and weather data with sufficient accuracy are available on a grid or image basis. The advantages of this type of modeling approach are:

  • Complete coverage of the country/region is possible.
  • Data are relatively easy to get.
  • Input and output are easily visualized.
Grid example for Afghanistan (0.5 degree distance)

Example of worldwide weather grid data availability. Meteorological data with a worldwide extend are offered by the MARS FOOD (European Union) based on outputs of the ECMWF (European Centre for Medium-Range Weather Forecast) atmospherical model. A time series for more than 40 years is now available thanks to the ERA40 reanalysis project. The data can be downloaded in ASCII Grid format. Get the data here. A tool is available to convert the grids into IDA images. This tool (JRC2IDA) can be downloaded here: here.

Data available from MARS FOOD web site
average temperature
maximum temperature
minimum temperature
precipitation sum
evapo-transpiration sum (ES0, bare soil)
evapo-transpiration sum (E0, over water)
evapo-transpiration sum (ET0, Penman-Monteith)
global radiation sum

Which type is best?

Many (sometimes heated) debates have been devoted to which modelling type is best. In many areas it is virtually impossible to work with measured weather data on a near real-time basis (although the real-time WMO GTS network for weather data might be an additional source of data). In some areas, given the timely availability of measured data makes the measured data difficult to ignore. Gaps in measured values can sometimes be filled with gridded data.

A simple rule of thumb could be that whenever good ground data are available, the irregularly spaced network type has a preference over a gridded network.

Retrieved from "http://www.hoefsloot.com/wiki/index.php?title=Chapter7"

This page has been accessed 79,529 times. This page was last modified 14:24, 27 September 2006.


0 Page generated in 0.150710 seconds.