Abstract: Attenuation of signals transmitted by radars due to rain drops and other precipitation particles is a factor that limits the accuracy of signal power measurements of radars, especially at higher frequencies such as X-band. For a long time, X-band radars were excluded from meteorological applications because of the attenuation problem. Recent advances in dual-polarimetric radar research show that the accumulative attenuation (A, in dB) due to rain can be compensated using the differential phase measurement (¦dp, in deg). This had led recently to the resurgence of X-band radars in meteorological applications since the rain attenuation can be quantified and moreover, the cost of such radars is much lower as compared with long wavelength radars. However, the estimation of the attenuation in mixed phase (rain mixed with wet ice) still remains a great challenge. The accuracy of the estimation is affected by many unknown factors in the atmosphere as well as in the radar system itself.
We approach this estimation problem by identifying those unknown factors whose uncertainties can be greatly reduced through modeling. First, we target improving the attenuation estimation for rain by using a fast non-linear least squares method to accurately estimate the coefficient (±) in the A-¦dp linear relationship model. This improvement can reduce the uncertainties due to temperature, rain drop shapes, and to a lesser extent large variations in the characteristic drop diameter of the rain drop size distribution (DSD). Second, we provide a preliminary method to separately estimate attenuation caused by rain and wet ice (mixed phase) particles along the beam.
We evaluate the fast non-linear least squares method for correcting the reflectivity at horizontal polarization (Zh) and the differential reflectivity between horizontal and vertical polarizations (Zdr) using simulation and operational X-band radar data. In the simulation, we apply our method to radar variables generated from simulated mono DSD profiles and variable DSD profiles. We evaluate the performance of our method under ideal and noisy environments. It shows that our method is able to adjust the coefficient according to the change in temperature and drop shapes with very fast convergence. Both Zh and Zdr are corrected to a very satisfying degree of accuracy. Operational X-band radar data are obtained from National Institute of Earth Science and Disaster Prevention (NIED), Japan and the first generation radar network of the Collaborative and Adaptive Sensing of the Atmosphere (CASA) Engineering Research Center (ERC). Our method is accurate on NIEDs data when compared against ground truth calculated from in situ disdrometer measurements. Our method is shown to be robust and meets the real-time operation requirements for CASA. It is worth noting that estimation of the specific attenuation at horizontal polarization (kh) and the specific differential attenuation between horizontal and vertical polarizations (khv) is independent of any system biases in the h- or v-channel of radars using our method.
We evaluate our preliminary method to separately estimate rain and wet ice attenuation using a CSU-RAMS supercell simulation case. The retrieved rain and wet ice specific attenuation fields are in close correspondence to the true fields calculated from the simulation. The wet ice attenuation field is useful in studying the k-Z relationship for wet ice, which can help improve the profiling algorithms used in Tropical Rainfall Measuring Mission (TRMM) or Global Precipitation Measurement (GPM).
Adviser: V.N. Bringi Co-Adviser: V. Chandrasekar Non-ECE Member: Graeme Stevens, Atmospheric Science Member 3: V. Chandrasekar, Electrical and Computer Engineering Addional Members:
