Using mesoscale model winds for correcting wind-drift errors in radar estimates of surface rainfall

For operational radars at middle and high latitudes, even the lowest beam of a scan sequence may be above the melting layer for a considerable proportion of the total range. This means that surface rainfall estimates are inferred from measurements made in the snow and ice. Snow and ice are more susceptible to wind drift than rain because of the low fall speed of around 1 m s(-1). Sampling these wind-induced fall-streak patterns results in a displacement of the radar-rainfall field when compared to ground measurements. To date wind-drift corrections have only been attempted in the rain, but corrections of around 2 km are generally smaller than the resolution of the grid on which the rainfall field is reported. Observational evidence from this study shows fall streaks in the snow can lead to displacements of the order of 10-20 km-a significant effect, especially in colder climates and seasons. In this paper a method for calculating and applying a wind-drift correction between the top of the fall streak and the bright-band height is presented. The forecast wind profile from the mesoscale version of the UK Met Office Unified Model is used to calculate the vertical shear of the horizontal wind. We assume the shear and the fall speed in the layer are constant. Results show that the method is able to reproduce observed displacements from high-resolution radar data in the range-height plane. Assessing the effect of applying the method in plan view shows an improvement in the placement of rainfall at the ground, reflected by an increase in the calculated skill scores. The displacement can be corrected to within 20%. This also shows that whereas wind-drift corrections in the rain have been found to be insignificant, corrections in the snow are not. Applying a wind-drift correction also appears to reduce the variability of vertical profiles of reflectivity (VPR) when they are extracted along a fall-streak path, and lead to a smaller reflectivity lapse rate locally. This would suggest that at least part of the reported variability in VPRs is due to wind effects.