Calibration d'un radar météorologique pour la mesure des précipitations. Premiers résultats en région parisienne

Click to increase image sizeClick to decrease image size

The object of this study is to examine all the possibilities of quantitative precipitation estimation by use of a meteorological radar. We have based our researches on reflectivity data from the Dammartin-en-Goëlle radar (10 cm) and measurements of precipitation intensity from rain-gauge recordings and rain-gauges in the Paris area. Hourly comparisons taken at random (53 stations) between radar hyetographs and rain-gauges recordings have made it possible to highlight several problems encountered in radar measurement (the importance of the Z-R (Marshall-Palmer) relationship, ground echoes, masking effects, distance effects on the radar, uniform calibration procedures) and to derive criteria concerning the choice of a minimum, representative calibration network. We have also conducted initial tests to validate random radar measurement (with uniform, regionalised calibration by a sub-network (64 stations)) with three days' uninterrupted operation of the radar and a dense network of rain-gauges (270). Reconstructed situations effected at points not used for calibration purposes were compared with the actual rain-gauge recordings. For these three days' only the regionalised calibration technique gave results equivalent to those from a conventional (Spline) interpolation method using the calibration sub-network only. Only systematisation of the calibration and validation procedures described to other rain spells and for a more extensive network of recording rain-gauges can lead to definite conclusions. Special estimations of rainfall are also required and here a correctly calibrated radar should be more effective.