Intercomparison of attenuation correction algorithms for single-polarized X-band radars
Intercomparison of attenuation correction algorithms for single-polarized X-band radars
复制标题
DOI：
10.1016/j.atmosres.2017.10.020
复制
发表时间：
2017-08
影响因子：
5.5
通讯作者：
K. Lengfeld;M. Berenguer;D. Torres
中科院分区：
文献类型：
--
作者：
K. Lengfeld;M. Berenguer;D. Torres
由于液态水引起的衰减是雷达观测中最大的不确定性之一。衰减的影响通常与波长成反比，即来自x波段雷达的观测比来自C或s波段系统的观测更受衰减的影响。另一方面，x波段雷达可以以更高的时空分辨率测量降水场，并且由于天线较小，更具移动性和更容易安装。Hitschfeld和Bordan（1954）提出了单极化系统衰减校正的第一个算法（HB），但在误差小（如雷达标定）和衰减强的情况下，该算法会变得不稳定。因此，已经开发出限制衰减校正以保持算法稳定的方法，例如使用表面回波（用于星载雷达）和山地回波（用于地面雷达）作为最终值（FV），或调整雷达常数(C)或系数α。在没有山区回波的情况下，可以使用C波段或s波段雷达的测量值来约束校正。所有这些方法都是基于反射率和比衰减之间的统计关系。另一种校正x波段雷达观测衰减的方法是使用衰减较小的雷达系统的附加信息，例如x波段与C或s波段雷达测量值之间的比率。Lengfeld et al.（2016）提出了一种基于X波段和c波段雷达观测值沿雷达波束的比值等压回归的方法。本文比较了原算法与三种基于反射率与比衰减统计关系的算法，以及两种实现c波段雷达测量附加信息的方法。它们在两个降水事件（一个主要是对流，另一个主要是层状）中的表现表明，限制hbs是避免不稳定所必需的。与垂直指向微雨雷达（MRR）的比较表明，基于统计k- z关系的两种方法fvanda具有良好的性能。该算法似乎对两个系统的校准差异更敏感，并且需要来自C或s波段雷达的额外信息。此外，对五个月的雷达观测的研究检验了每种算法的长期性能。从这项研究中可以得出结论，使用来自较小衰减雷达系统的附加信息可以获得最佳结果。使用这些附加信息的两种算法消除了衰减引起的偏差，并保持了与MRR观测值的一致性。
Attenuation due to liquid water is one of the largest uncertainties in radar observations. The effects of attenuation are generally inversely proportional to the wavelength, i.e. observations from X-band radars are more affected by attenuation than those from C- or S-band systems. On the other hand, X-band radars can measure precipitation fields in higher temporal and spatial resolution and are more mobile and easier to install due to smaller antennas.A first algorithm for attenuation correction in single-polarized systems was proposed by Hitschfeld and Bordan (1954) (HB), but it gets unstable in case of small errors (e.g. in the radar calibration) and strong attenuation. Therefore, methods have been developed that restrict attenuation correction to keep the algorithm stable, using e.g. surface echoes (for space-borne radars) and mountain returns (for ground radars) as a final value (FV), or adjustment of the radar constant (C) or the coefficientα. In the absence of mountain returns, measurements from C- or S-band radars can be used to constrain the correction. All these methods are based on the statistical relation between reflectivity and specific attenuation. Another way to correct for attenuation in X-band radar observations is to use additional information from less attenuated radar systems, e.g. the ratio between X-band and C- or S-band radar measurements. Lengfeld et al. (2016) proposed such a method based isotonic regression of the ratio between X- and C-band radar observations along the radar beam.This study presents a comparison of the originalHBalgorithm and three algorithms based on the statistical relation between reflectivity and specific attenuation as well as two methods implementing additional information of C-band radar measurements. Their performance in two precipitation events (one mainly convective and the other one stratiform) shows that a restriction of theHBis necessary to avoid instabilities. A comparison with vertically pointing micro rain radars (MRR) reveals good performance of two of the methods based in the statistical k-Z-relation:FVandα. TheCalgorithm seems to be more sensitive to differences in calibration of the two systems and requires additional information from C- or S-band radars.Furthermore, a study of five months of radar observations examines the long-term performance of each algorithm. From this study conclusions can be drawn that using additional information from less attenuated radar systems lead to best results. The two algorithms that use this additional information eliminate the bias caused by attenuation and preserve the agreement with MRR observations.