Remote sensing of soil moisture in desert regions is possible with microwave sensors (Schmugge 1981) and by
termal techniques (Wetzel & Atlas 1981; England et al. 1983). However, no satellite-based soil moisture
remote sensing system, using microwave techniques, is currently available on an operational or semi-operational
basis for an area the size of the desert locust recession area; thermal remote sensing techniques that use
Meteosat data are still in the research phase.
Remote sensing of meteorological events that cause soil moisture changes relevant for agricultural crop
production and locust population development, can be undertaken from geostationary environmental satellites,
e.g. Meteosat for Africa and the Near-East The high temporal frequency of this type of data in the visible
and thermal infrared wavelengths permits detailed monitoring of weather systems likely to produce the rainfall
necessary to support crop development and initiate locust breeding.
Various schools have developed and tested remote sensing techniques for estimating rainfall quantitatively from
Meteosat data. FAO has been closely involved in these developments as undertaken by the Universities of
Bristol (Barrett 1977,1980; Hielkema 1980; Barrett & Harrison 1986) and Reading (Milford & Dugdale 1987).
The technique developed by the University of Reading, based exclusively on the use of Meteosat thermal
infrared data, which was selected by FAO for implementation on the ARTEMIS system. The technique is
based on the observed linear relationship between the vertical extent of a cloud in the atmosphere, as
evidenced by the temperature of its top, which can be measured by the satellite at frequent intervals, and the
amount of rain which falls from the bottom of the cloud. This technique is particularly suitable for application
in tropical and arid environments where generally over 90% of the annual rainfall comes from strong convective
activity, i.e. strong vertical development of clouds.
Remote sensing of green vegetation biomass is a well developed technique for which several satellites
frequently collect data on a worldwide scale. The technique involves measuring reflected spectral radiance that
results from the interaction between the green-leaf vegetation cover and incident solar spectral irradiance.
These measurements can be made from any altitude above the surface (1 m, 10,000 m, 1,000 km) depending
on the remote sensing system in question (ground-based, aircraft or satellite, respectively). Spectral estimation
of green-leaf biomass generally involves the use of two wavelength regions, the red (0.6-0.7 pm) and near
infrared (0.75-1.1 pm), although the specific wavelengths used often vary slightly between instruments. The
0.6-0.7 pm region correspond to the in vivo red region of chlorophyll absorption and is inversely related to
chlorophyll density. In the 0.75-1.1 pm region, reflectance is proportional to green-leaf density. Ratio
combinations of these two wavelength regions thus contain information related to the chlorophyll-green-leaf
density (Tucker 1979). Use of these two bands for making plant canopy inferences by non-destructive
techniques is facilitated by the proximity of the two bands in the electromagnetic spectrum. This proximity
enables simple band ratios or other combinations to be used to compensate for differences in solar flux
intensities. Linear or ratio combinations of red/near infrared spectral data have been used to quantify a variety
of vegetation types (Tucker 1980; Curran 1983).
