Search

We analyze the calibration stability of the 17-yr precipitation radar (PR) data on board the Tropical Rainfall Measuring Mission (TRMM) satellite to develop a precipitation climate record from the spaceborne precipitation radar data of the TRMM and following satellite missions. Since the PR measures the normalized radar cross section (NRCS) over the ocean surface, the temporal change in the NRCS whose variability is insensitive to the sea surface wind is regarded as the temporal change of the PR calibration. The temporal change of the PR calibration in TRMM, version 7, is found to be 0.19 dB decade−1 from 1998 to 2013. The calibration change is simply adjusted to evaluate the NRCS time series and the near-surface precipitation trend analysis within the latitudinal band between 35°S and 35°N. The NRCS time series at nadir and off-nadir are uncorrelated before the calibration adjustment, but they are correlated after the adjustment. The 0.19 dB decade−1 change of the PR calibration causes an overestimation of 0.08 mm day−1 decade−1 or 2.9% decade−1 for the linear trend of the near-surface precipitation. Even after the adjustment, agreement of the results among the satellite products depends on the analysis period. The temporal stability of the data quality is also important to evaluate the plausible trend analysis. The reprocessing of the PR data in TRMM, version 8 (or later), takes into account the temporal adjustment of the calibration change based upon the results of this study, which can provide more credible data for a long-term precipitation analysis. Significance Statement: The stability of long-term data is very important for climate research so that an account of temporal calibration changes in the sensor must be made. In this study, we investigate the calibration stability of the TRMM PR data and evaluate its impact on the precipitation trend analysis. The temporal change of the PR calibration is estimated to be 0.19 dB decade−1. Compensating for this change improves the consistency of precipitation trend analysis between the PR and other precipitation datasets. The reprocessed PR data provide more probable data for long-term precipitation analysis. [ABSTRACT FROM AUTHOR]

A new algorithm is proposed, which estimates two parameters of the particle size distribution (PSD) at each range bin from the global precipitation measurement’s (GPM’s) dual-frequency precipitation radar (DPR) data. The equation that expresses the relationship of the PSD parameters between adjacent range bins is derived. By including the attenuation effect within the bin in the discretized equation, the new algorithm alleviates the double-solution problem when attenuation within the bin is sufficiently large. The stability of the solutions to the equation depends on the value of the mean diameter $D_{m}$ and its gradient with respect to the range in the case of liquid precipitation. There is a critical diameter above which the backward processing of the equation provides a stable or moderately diverging solution, unlike the forward processing that often gives unstable solutions. To provide a set of initial conditions without using the surface reference technique (SRT) in the backward processing, an initialization method using the Hitschfeld–Bordan (HB) attenuation correction method is proposed and tested. The proposed algorithm may provide a tool for investigating the assumptions used in various algorithms. [ABSTRACT FROM AUTHOR]

Assessments on the performance of dual-frequency (13.6/35.5 GHz) Precipitation Radar (DPR) rain retrieval techniques are performed by means other than relying on the surface reference technique. A DPR inversion technique (DPR-IT) with an independent estimate of the differential attenuation, i.e., the difference of attenuation differences between two frequencies over a certain range, is introduced as an alternative to surface reference or iterative methods for resolving the path-integrated attenuation (PIA) information. Retrieval performance of the proposed method is tested to some vertical rain profiles synthesized with arbitrarily defined and disdrometer-measured raindrop-size distribution data. Retrievals of two other DPR-IT-type methods, namely the DPR-IT with sets of preselected surface PIAs at two frequencies from wide ranges and the DPR-IT iterative algorithm, are also considered for the analysis. The comparison of the simulated results obtained from these three methods is presented and discussed. Index Terms--Differential attenuation, dual-frequency rain retrieval, Global Precipitation Measurement (GPM).

Range profile data from the surface echo of the Precipitation Radar (PR) onboard the Tropical Rainfall Measuring Mission (TRMM) satellite enable the assessment of the nonuniformity of rainfall. The surface echo profile can be converted to a horizontal pattern of surface echo strength within the PR's footprint at off-nadir angles. Then, using the no-rain surface echo profile over ocean as a reference, the horizontal pattern of the path-integrated attenuation (PIA) can be estimated under moderate to heavy rainfall. Based on the horizontal pattern of the PIA, the nonuniform beam-filling parameter is estimated. However, this technique is applicable only to off-nadir angles and over ocean. Comparisons using ground-based radar data obtained simultaneously with TRMM data verify the effectiveness of this technique. Index Terms--Nonuniform beam filling (NUBF), surface echo, Tropical Rainfall Measuring Mission (TRMM).

The Tropical Rainfall Measuring Mission (TRMM) satellite changed its altitude from 350 to 402.5 km in August 2001. As a result, the level-1 algorithm for the new orbit of the Precipitation Radar (PR) onboard the TRMM has to correct the 'beam mismatch' resulting from the altitude change. Since the PR uses fixed transmission-reception timing, an altitude change of 50 km corresponds to a delay of return signals of 1 pulse repetition interval (PRI). This is not a serious problem if the angle bin of the next pulse is the same as the angle bin of the current pulse. Otherwise, return signals arrive at the PR when the antenna direction shifts to the next angle bin. This is caned a 'beam mismatch.' It affects one pulse sample out of 32 averaged pulse samples. In other words, one 'beam mismatch' pulse sample and 31 normal pulse samples are averaged at the onboard processor of the PR. A new algorithm was added to the PR's level-1 algorithm, 1B21, to eliminate this mismatch sample at the 402.5-km altitude. In this paper, the effect of beam mismatch is estimated for both rain echo and surface echo in terms of the received power and the incident angle dependency. The basic function of the beam mismatch correction algorithm is to estimate the received power of the mismatched pulse. Theoretically, the effective round-trip antenna pattern of a mismatched pulse has a peak right in the middle of the transmission and reception directions with a gain reduction of 6 dB. The new 1B21 algorithm uses the average of the received power of successive angle bins with a 6-dB gain reduction as the power from the mismatched pulse. The effectiveness of the correction algorithm was evaluated using high angular resolution data obtained during external calibration observations and the statistics of the normalized radar cross section of the earth's surface ([[sigma].sup.0]), which is thought to be unchanged. The estimated error was less than 0.2 dB for the rain echo, and a large error of up to 0.5 dB was found at the boundary of the surface based on the error estimation using high angular resolution data. The difference in [[sigma].sup.0] and its angle dependency is explained using a simple surface model. The model results indicate the correction error reaches up to 0.8 dB at the skirts of the surface echo. Index Terms--Algorithm, orbit boost, Tropical Rainfall Measuring Mission (TRMM).

In developing the upcoming Global Precipitation Measurement (GPM) mission, a dual-frequency Ku-Ka-band radar system will be used to measure rainfall in such a fashion that the reflectivity ratio intrinsic to the measurement will be sensitive to underlying variations in the drop size distribution (DSD) of rain. This will enable improved techniques for retrieving rain rates, which are dependent upon several key properties of the DSD. This study examines this problem by considering a three-parameter set defined by liquid water content (W), DSD effective radius ([r.sub.e]), and DSD effective variance ([v.sub.e]). Using radiative transfer simulations, this parameter set is shown to be related to a radar reflectivity factor and specific attenuation in such a fashion that details of the DSDs are immaterial under constant W, and thus effectively represent important variations in DSD that affect rain rate but with a minimal number of parameters. The analysis also examines the effectiveness of including some measure of mean Doppler fall velocity of raindrops ([bar.v]), given that the fundamental properties of a given precipitation situation are uniquely defined by a combination of a drop mass spectrum and drop vertical velocity spectrum. The results of this study have bearing on how future dual-frequency precipitation retrieval algorithms could be formulated to optimize the sensitivity to underlying DSD variability, a problem that has greatly upheld past progress in radar rain retrieval.

Pulse patterns are designed by adopting variable pulse repetition frequency (VPRF) for space missions involving beam scans in the cross-track direction, especially intended for the Global Precipitation Measurement Project (GPM). To cope with large variance in range from a satellite, which is caused by the beam swing and the earth oblateness, a systematic algorithm is proposed, increasing sampling rates. Index Terms--Global Precipitation Measurement Project (GPM), rain radar, spaceborne radar, variable pulse repetition frequency.

We examine a problem, recently opened up by Jameson and Kostinski [1], of statistics for radar rain echoes in a nonstationary case such as the observation of rain by a moving antenna beam. We study this problem under the framework of the standard fluctuation theory of Marshall and Hitschfeld [2]. We obtain the exact nonstationary probability density function (pdf) for an average of n mutually independent samples of the received power in the square-law detection, and an approximate nonstationary pdf in the logarithmic detection. We find that the concept of bias does not appropriately describe the effect of antenna beam movement during the sampling. We introduce a new quantity to assess the effect of antenna beam movement on rain observation by radars.

The need to understand the complementarity of the radar and radiometer is important not only to the Tropical Rain Measuring Mission (TRMM) program but to a growing number of multi-instrumented airborne experiment that combine single or dual-frequency radars with multichannel radiometers. The method of analysis used in this study begins with the derivation of dual-wavelength radar equations for the estimation of a two-parameter drop size distribution (DSD). Defining a 'storm model' as the set of parameters that characterize snow density, cloud water, water vapor, and features of the melting layer, then to each storm model there will usually correspond a set of range-profiled drop size distributions that are approximate solutions of the radar equations. To test these solutions, a radiative transfer model is used to compute the brightness temperatures for the radiometric frequencies of interest. A storm model or class of storm models is considered optimum if it provides the best reproduction of the radar and radiometer measurements. Tests of the method are made for stratiform rain using simulated storm models as well as measured airborne data. Preliminary results show that the best correspondence between the measured and estimated radar profiles usually can be obtained by using a moderate snow density (0.1-0.2 g/[cm.sup.-3]), the Maxwell-Garnett mixing formula for partially melted hydrometeors (water matrix with snow inclusions), and low to moderate values of the integrated cloud liquid water (less than 1 kg/[m.sup.-2]). The storm-model parameters that yield the best reproductions of the measured radar reflectivity factors also provide brightness temperatures at 10 GHz that agree well with the measurements. On the other hand, the correspondence between the measured and modeled values usually worsens in going to the higher frequency channels at 19 and 34 GHz. In searching for possible reasons for the discrepancies, it is found that changes in the DSD parameter [Mu], the radar constants, or the path-integrated attenuation can affect the high frequency channels significantly. In particular, parameters that cause only modest increases in the median mass diameter of the snow, and which have a minor effect on the radar returns or the low frequency brightness temperature, can produce a strong cooling of the 34 GHz brightness temperature.

The Global Precipitation Measurement (GPM) Dual-frequency Precipitation Radar (DPR) has operated in full scan (FS) mode in both the Ku-band and Ka-band since May 2018. This FS mode can enable about 245 km full swath dual-frequency processing for the first time, whereas previous algorithms enabled dual-frequency processing in the narrower inner swath, having a width of approximately 125 km. This paper describes the classification (CSF) module in newly developed DPR level 2 V06X experimental algorithms corresponding to the FS mode. The CSF module classifies precipitation as three major types, stratiform, convective, and other, and provides bright-band (BB) information. One-month statistics show that a significant jump occurs in the Ka-band precipitation type counts at the edges of the inner swath due to the different sensitivity of Ka-band radar in Ka-band Matched Scan (Ka-MS) mode compared with that in Ka-band High Sensitivity (Ka-HS) mode. However, precipitation type counts indicate that dual-frequency processing works well not only in the inner swath but also in the outer swath. Statistics on BB counts show a significant improvement in BB detection, especially in the outer swath, when dual-frequency processing is performed. In addition, the V06X Ku-band algorithm solves two problems related to the CSF module: (a) appearance of a very large near-surface precipitation rate of stratiform precipitation reclassified by the slope method and (b) a rare case of misjudging BB peak as an upper part of surface echo. The data structures of GPM DPR algorithms were drastically changed in V06X. The new data structures introduced in V06X will be used in V07A and later. In this regard, the V06X CSF module outlined in this paper is a prototype of future versions of each CSF algorithm. [ABSTRACT FROM AUTHOR]