The impact of precipitation Bragg backscatter on radar hydrometeorology

This presentation discusses the recent discovery of radar coherent backscatter, and its potential impact on quantitative radar measurements of precipitation. Data in snow and rain have been analyzed. Coherent scatter was found to be pervasive throughout two minutes of data with none of the snow and only up to at most 11% of the rain observations being examples of pure incoherent scatter. On average, 34% in rain and 72% in snow of the total backscattered power arose from coherent scatter. Two sources of coherent scatter are identified, one in the direction of propagation (and backscatter) and the other orthogonal to the beam axis. This latter is generated by spatially correlated precipitation structures acting like diffraction gratings in resonance with the radar wavelength in which the elements of the structures all moving at practically the same velocity. Such diffraction gratings produce fields of maxima and minima of back-scattered intensities which, as they move through the radar beam, produce distinctive power oscillations of frequencies f. This is described by a back-scattered power spectrum, Z(f))of the total backscattered power. This Z(f) allows us to calculate the coherent scatter contribution to the total back-scattered power. It can be computed for stationary antennas even for incoherent radars since Doppler information is not required. The detection and estimation of radar coherent scatter along the beam of a stationary antenna requires pulse to pulse time series of the complex amplitudes at every range bin. The cross-correlation function in neighboring range bins then yields an estimate of the fractional contribution of the coherent along the beam. The reason radar has been used to estimate precipitation is that both the backscattered incoherent power precipitation parameters such as the rainfall rate depend linearly on the particle concentration. This is not the case for coherent backscattered power which goes as the square of the particle concentration. Hence, as will be illustrated in the presentation, the presence of significant coherent backscatter will lead to over-estimates of rainfall and snowfall derived exclusively from simple measurements of the reflectivity factor. There are two ways to deal with this dilemma. The first is to measure and remove it. This is not a trial task because the amount of coherent backscatter can not be measured as an antenna scans. However, such measurements are possible with phase array radar as will be explained in the presentation. The second approach is to minimize its relevance. The best way to do this is to use radar polarization measurements, specifically the combination of Z_DR and differential phase, K_DP. This will also be explained in the presentation. In summary, radar coherent backscatter introduces biases in the radar techniques for estimating precipitation which use the back-scattered power. However, the good news is that it can be dealt with using the right techniques and/or phased array antenna radars so that radar will remain a practical and useful tool for estimating precipitation over large domains in hydrology.