Capabilities and Characteristics of Rainfall Estimates from Geostationary- and Geostationary+ Microwave-Based Satellite Techniques

Heavy precipitation events often occur on time scales incompatible with the limited observing capabilities of current passive microwave-based, low-Earth orbiting satellite platforms, necessitating the use of geostationary satellite-based, infrared imaging systems to provide the rapid time update capability. This study compares the capabilities and characteristics of operationally-oriented geostationary- and geostationary+microwave-based rain rate techniques using a variety of summer season convective systems and western USA coastal winter storms. The techniques include the Auto-Estimator (AE) and the GOES Multispectral Rainfall Algorithm (GMSRA) developed within the National Environmental Satellite Data and Information Service (NESDIS), and a combined geostationary-plus-microwave technique originally developed for NWP data assimilation at the Naval Research Laboratory (NRL). The microwave-based datasets include the various Special Sensor Microwave Imager (SSM/I) instruments, as well as the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) rainrates provided via the TRMM Science Data Information Service (TSDIS). The AE was recently recommended by the National Weather Service (NWS) for operational implementation to support the NWS flash flood watch and warning programs. A major limitation of the AE is the overestimation of rainfall (both with respect to amounts and areal coverage) associated with high, cold-topped cirrus clouds. The GMSRA technique uses combined information from the visible, near-infrared and infrared GOES measurements. Compared with the AE, the GMSRA has lower biases though often does not produce the extreme heavy rainfall that occurs with many convective systems. The combined geostationary-plus-microwave technique produces reasonable estimates for cold top convective systems and has a relatively low bias; this method also tends to underestimate the high rainfall rates similar to GMSRA. A validation of this technique is currently underway with data from a dense rain gauge network. The three techniques perform rather poorly for west coast winter storms, many of which are stratiform in nature and associated with warm cloud tops, often capped with non-precipitating cold-topped cirrus clouds. Additionally, this study will present ideas on how to combine the strengths of all of the various techniques.