Polarimetric X‐band weather radars for quantitative precipitation estimation in mountainous regions

Abstract

In order to minimize the economic and social impact of hazardous weather in the Alps, three X‐band dual‐polarization radars were deployed in the southeast of France during the RHYTMME project in 2011–2013 and more recently have been incorporated formally into the conventional radar network already operating S‐ and C‐band radars. In this mountainous region, the radars' beam shielding by the complex terrain, the very different altitudes of the radars, the significant attenuation at X‐band and the low density of rain‐gauges have made the integration of these radars into the operational service a challenging exercise. In this article, the framework used to evaluate the corrections applied to the volume data produced by the radars is presented and the residual errors related to the partial beam blockages, the attenuation by rain, the wet radome, and the melting layer are carefully studied and disentangled. The results of the analysis suggest that: (a) the beam blockage correction requires the use of a much higher resolution digital elevation model (DEM) in mountainous regions than the 250 m resolution DEM currently uses, (b) the dual‐pol reflectivity (Zh) attenuation correction used at Météo‐France performs well in the rain but the differential attenuation on Zdr is overestimated, (c) the attenuation of wet snow is underestimated and a mean error of 2 dB is induced by the melting layer in spring and autumn, (d) the wet radome, as well as producing significant attenuation on Zh, induces differential attenuation on Zdr. The attenuation in the melting layer is further investigated taking advantage of the height difference between the radar sites. A new specific attenuation coefficient for wet snow is suggested and it is shown to reduce the seasonal bias on three particular events observed in spring and autumn.