Rainfall data constitute an important parameter for studying water resources-related problems. But currently, the density of raingages within the most of areas are not enough to modify required space-time information for the existing flood forecasting models. Remote sensing techniques could provide a more comprehensive overview of the rainfall spatial distribution in a global scale.
Since the mid 1970’s numerous attempts have been made to relate rainfall to observations of cloud from satellites. Unfortunately an accuracy of remote sensing techniques generally is unacceptable for model purposes. The best estimation of rainfield could be done by assimilation of raingages data and satellite information. A variational adjustment method allowed us to do so.
The developed analysis scheme is based on a variation adjustment of ground-based station data on daily precipitation totals and data on a space-time structure of radiation temperature obtained from infrared measurements from geostationary satellite. This method is an attempt to combine the accuracy of measurement of traditional raingages and information about spatial gradients of temperature obtained from IR satellite data for retrieval of rainfall fields. IR information is more suitable for the purpose of operationally estimating precipitation due to the fact that VIS information is not available during a night. The method is applicable to use in tropics and middle latitude over the period when thermal contrast between cloud tops and the surface is substantial.
Half-hourly thermal infrared data from the geostationary satellite Meteosat and conventional raingage information (daily accumulated amounts) have been used in this study. The method is applied on middle latitude region (European part of Russia). The information for 45 days (September 1993, August and September 1994, July 1997) has been used for that analysis.
The method is applicable for middle and high latitudes within a warm period. It is effective within territories with a low density of raingage’s network (less that one raingage on 3600 km2 area).
This method can be used to crop and fire monitoring and predicting river flow. The satellite data would then ensure realistic precipitation fields in data sparse areas. We consider that better results can be obtained if we use the data for summer period.