Rain microstructure from polarimetric radar and advanced disdrometers

Abstract

The use of polarimetric techniques in weather radars has in recent years captured a great deal of interest within the operational weather radar community (e.g., Ryzhkov et al. 2005). These systems often operate in three different frequency bands, namely S, C and X-bands, corresponding to approximately 3, 5 and 11 GHz, respectively, and furthermore, make use of dual-linear polarizations at horizontal and vertical states, as proposed back in the late 70s by Seliga and Bringi (1976). Such dualpolarization techniques are particularly useful at low elevation angles because the rain medium consists of highly oriented non-spherical particles (i.e., rain drops), with their axis of symmetry oriented along the vertical, and as such, will give rise to differences in their forward and back scatter amplitudes between vertical and horizontal polarizations. The retrieval or the estimation of rainfall rates from the back scatter polarization measurements requires fundamentally some knowledge of the rain microstructure. By this, we mean primarily, (a) drop size distribution, (b) drop shapes and their variations due oscillations, (c) drop orientation angles and (d) drop fall velocities. The dual-linear polarization methods make use of the fact that the drops, particularly those larger than 1 mm diameter, are non-spherical in shape and that these shapes are diameter- dependent. By making use of the fact that the forward scatter amplitudes and the back scatter reflectivities will, therefore, show polarization dependence, it becomes possible to retrieve information on the drop size distribution (DSD) from the polarimetric measurements and hence estimate the rainfall rates more accurately than the conventional weather radars which use single polarization (e.g., Doviak and Zrnic 1993). Rain microstructure can also in parallel be determined from the socalled disdrometers. These are instruments which are designed to measure or infer the DSDs within a given sensor volume. The conventional systems, such as the impact-type Joss-Waldvogel disdrometers (JWD; Joss and Waldvogel 1967), make certain assumptions on the drop diameter dependence on fall velocities and do not allow for nonzero drop canting angles. Other types of disdrometers such as microrain radars (MRR) operate on the Doppler principle to derive the DSD from the fall velocity spectra. There are also laser-based optical devices (such as the Parsivel; Lffler-Mang and Joss 2000), which does not image the particle but rather gives the maximum width and velocity of particles binned in a 32 32 matrix. The most advanced system of all, at least at present, is the 2-dimensional video disdrometer (abbreviated to 2DVD) whose measurement principles and specifications have been described in detail in Chap. 1. It measures the size, shape, orientation and fall velocity of each individual hydrometeor falling through its sensor area; i.e., it measures all four primary parameters of rain microstructure. Moreover, all four parameters are measured by a direct method through imaging techniques using fast line scan cameras. For this reason, we give here several examples which utilize measurements from this instrument, most of which have only recently been published. The examples are given to highlight certain important properties/features of the rain microstructure. For other disdrometer data and evaluation, readers are referred to publications by other authors, as listed in Table 1. Table 1. Examples of advanced disdrometers and references for detailed Instrument Reference Parsivel Lffler-Mang and Joss (2000) Dual-beam spectro-pluviometer Hauser et al. (1984) 2-dimensional video disdrometer Randeu et al. (2002) Kruger and Krajewski (2002) Schnhuber et al. (2007) Micro-rain radar() Peters et al. (2005) Precipitation Occurrence Sensor System () Sheppard (1990) Sheppard and Joe (1994) () These instruments estimate the drop size distribution from the measured Doppler power spectra As for intercomparison between instruments, Krajewski et al. (2006) have given DSD, rainfall and velocity-diameter data taken during intense precipitation which show some discrepancies in certain cases. Here we deal with mostly 2DVD and one other, very promising, instrument, namely POSS (referred to in Table 1). This instrument is similar to the MRR in that it derives the DSD from the Doppler power spectra, but it operates in a bistatic mode and retrieves the DSD within the common volume defined by the transmit and receive antenna patterns.