Variability and Sensitivity of GPM-retrieved Mass Weighted Diameter over Italy

Variability and Sensitivity of GPM-retrieved Mass Weighted Diameter over Italy.

Leo Pio D’Adderio, Gianfranco Vulpiani, Ali Tokay and Federico Porcù

The National Aeronautics and Space Administration (NASA) and the Japan Aerospace Exploration Agency (JAXA) launched the Global Precipitation Measurement (GPM) mission as a global successor to Tropical Rainfall Measuring Mission (TRMM). The aim of GPM is to measure the precipitation rate as well as to retrieve the parametric form of the rain drop size distribution (DSD) globally. The GPM Core observatory is equipped with Dual-frequency Precipitation Radar (DPR), which operates at Ku- and Ka-band. The DPR DSD retrieval algorithm adopted three-parameter gamma distribution where mass weighted diameter (D_mass) and normalized intercept parameter with respect to the liquid water content (Nw) are derived. At this stage, the third parameter - shape parameter - is set to 2 in combined radar-radiometer algorithm and set to 3 in DPR algorithm.

Both precipitation rate and DSD parameters are retrieved for each footprint of DPR, which operates at three different scanning mode: High Sensitivity (Ka-band only), Normal Scan (Ku-band only), and Matched Scan (Ka-/Ku-band). The latest version (V04) of the GPM DPR retrieval algorithm relates D_mass to the rain intensity R in the form of R=a∙D_mass^b, while the previous version used the k-Z_e relationship, where k is the specific attenuation and Z_e is reflectivity at Ku-band. An adjustment factor, ε^c, has been introduced to take into account the two-way attenuation path along the radar beam and has been derived from k-Z_e relationship (considering the adjustment factor, the shape of R-D_mass relationship becomes R=a∙ ε^c∙ D_mass^b). Two DSD models, for stratifrom and convective precipitations, are used in the DPR retrieval algorithm corresponding to two different R-D_mass relationships. A number of impact disdrometer datasets from various tropical sites were used to derive the R-D_mass relationships. The datasets were collected through TRMM field campaigns in 1990s. Prior to and post GPM Core Satellite launch, the US GPM Ground Validation (GV) program conducted several field campaigns focusing on mid- and high-latitude precipitation. These field campaigns, together with GPM direct data acquisition sites at two NASA centers at Wallops Island, Virginia, and Huntsville, Alabama, provided very rich dataset (disdrometers, rain gauges, radars, etc.). The R-D_mass relationships are derived through sequential intensity filtering technique (SIFT, Lee and Zawadzki, 2005) utilizing the two-dimensional Video Disdrometer (2DVD) and Parsivel² disdrometer (P2) data collected during the GV field campaigns.

The accuracy of D_mass retrieval is determined through a comparative study of ground based and space borne products. The Italian Department of Civil Protection (DPC) manages seven C-band polarimetric radars across the all country with the aim of control and surveillance of the atmospheric phenomena. The DPC provides its own DSD retrieval algorithm to derive D_mass.

This study investigates the variability of D_mass over Italy as derived by the DPR retrieval algorithm applying both the official V04 R-D_mass relationship and the new R-D_mass relationships derived from GV data. A number of DPR passages from July 2015 to December 2016 are analyzed with the aim to cover all the country. The sensitivity in D_mass estimation using tropical based and mid-latitude based R-D_mass relationship is also investigated. The results show a quite good agreement between official and new R-D_mass relationship for the small drop dominated precipitation, while the difference is more marked for significant presence of large drops especially at higher rain rates. Has to be highlighted that the disdrometer ground data shows a marked variability of D_mass for a given R (i.e. at 10 mmh^-1, D_mass can range between 0.8 and 3 mm about, depending on the location). This results in a significant uncertainty of D_mass retrieval when R-D_mass relationship is applied.

The reliability of the D_mass estimation from the GPM DPR retrieval algorithm is analyzed by comparing it with D_mass derived from the Italian radar network data. Two different approaches are employed to derive the D_mass from the radar data. The D_mass generally relies on D_mass-Z_DR relationship, which is derived from the GV disdrometer data. On the other hand, the ground-based radar retrieval approach adopted here is based on a neural network inversion technique. The algorithm, originally developed and validated for S-band radars (Vulpiani et al., 2006; Vulpiani et al., 2009), has been modified for C-band.

Lee, G.W. and Zawadzki, I.: Variability of Drop Size Distributions: Noise and Noise Filtering in Disdrometric Data, 2005, J. Appl. Meteor., 44, 634-652.

Vulpiani, G., F. S. Marzano, V. Chandrasekar, A. Berne, and R. Uijlenhoet, 2006: Polarimetric weather radar retrieval of raindrop size distribution by means of a regularized artificial neural network, IEEE Trans. Geosci. Remote Sens., 44, 3262-3275.

Vulpiani G., S. Giangrande and F.S. Marzano, "Rainfall estimation from polarimetric S- band radar measurements: Validation of a neural network approach", J. Applied Meteor. and Climat., vol. 48, pp. 2022-2036, 2009.