I am leading a group of investigators currently consisting of one post-doctoral associate and six graduate students all poised to make major contributions in the field. Equipped with unique research instruments, which include a Doppler polarimetric weather radar on wheels and a two-dimensional video disdrometer unit, have managed to secure competitive funding to support my group’s experimental research component through participation in high-profile international field experiments. My most recent work has focused almost exclusively on the development of experimental methodologies and technologies, which will offer physical validation of precipitation remote sensing techniques and cloud resolving model parametrizations. Physical understanding of the uncertainties inherent in remote sensing data and models, and the resultant capabilities and limitations, is important for a number of applications ranging from real-time weather and flood forecasting to climate monitoring. As part of this effort my research group with our mobile instrumentation have recently participated in NASA’s Keys Area Microphysics Project, Wallops Winter Precipitation Experiment, while currently we are participating in the International H2O Experiment (IHOP) in Western Oklahoma. The synergy of rainfall observations from a wealth of research instruments available in these field campaigns is a fundamental building block for our investigations on precipitation microphysics and radar/radiometer rainfall measurements, which would help improve satellite rainfall estimation and quantitative precipitation forecasting.
I am member of the science team for the joint US/Japan Tropical Rainfall Measuring Mission (TRMM), which launched the first ever satellite equipped with radar dedicated to measuring precipitation around the globe. The mission’s primary goal is to observe tropical rainfall and determine how it affects the global climate. In support of TRMM, I have developed a system for using the TRMM satellite radar to test and then calibrate ground based weather radar systems within its path. Those radars are used by weather service agencies to measure rainfall and issue warnings for floods and flash floods. With the hundreds of radar sites around the globe being under the TRMM umbrella, keeping all radars properly calibrated would be unfeasible due to its cost and labor intensity. I developed an approach for monitoring the calibration of hundreds of those systems, using a single platform, the TRMM radar (www.engr.uconn.edu/~gracp). Our research has shown calibration variations among radar sites do exist. An investigation of 20 NWS radar sites in the southern United States, where flash floods are common, determined that systematic radar rainfall differences among adjacent radars are significant owing to calibration differences. By finding fluctuations like this real-time flood forecasting by radar can be improved. It is a technology the National Weather Service is taking seriously. Negotiations are underway to give the agency use of the system.
But the majority of the globe is not covered by ground radar, so precipitation in those areas is feasible only from space-based platforms. With funding through NASA’s New Investigator Program I have led my group in (1) developing a truly combined radar/radiometer rain-profiling algorithm based on a physically based approach, and (2) investigating efficient statistical procedures for near continuous overland rainfall estimation through combination of less definitive precipitation observations from geo-stationary satellite platforms and auxiliary lightning data. As part of this effort, and working jointly with the private sector and the National Observatory of Athens, we have developed a long-range lightning detection system called ZEUS, which advances precipitation monitoring from NASA’s meteorological satellites. The system consists of six VLF radio sensors positioned around Europe (see sifnos.engr.uconn.edu for details on the system) detecting primarily cloud-to-ground lightning occurring within Europe, North Africa, and surrounding waters with great location accuracy (5-10 km) and time resolution (1 millisecond). The system is valuable in several ways. Not only is it relatively cost efficient, but also by combining lightning detection with satellite observations, high-resolution rainfall estimates over large-scale areas are improved. Our in press paper in Journal of Hydrometeorology shows that combining information from this system with satellite infrared observations in a rain retrieval algorithm is more accurate (reduces regional biases by 30% and the RMS in convective rainfall by 20%) than using satellite data alone. Our continuing research concentrates on demonstrating ZEUS potential on improving flash flood predictions and climate analysis in lieu of weather radar data, and on improving its lightning location accuracy over ranges exceeding the networks periphery (>3,000 km). My long-term goal is to make ZEUS global. I figure it would take a minimum of 25 additional receivers to cover the entire world. Now it is just a matter of finding the funding. My next step this year is to cover the Africa continent with three additional receivers, a proposal that is in preparation for submission to NSF’s Water Cycle Research program this June.
Finally, in collaboration with colleagues in Europe am conducting research on error propagation of remotely sensed rainfall estimates in flood forecasting. This is a fairly new research direction am taking, where contrary to the traditional notion for flood forecasting uncertainty, errors are accounted in both the process of estimating rainfall from radar (or satellite) and in the modeling of the rainfall-runoff transformation. We have developed a Bayesian framework for the assessment of this uncertainty interaction that is consistent with the limitations of a hydrologic model and rain data available, and allows a direct quantitative comparison between model predictions obtained using rainfall inputs from remotely sensed data and those from a considered ground “truth” dense rain gauge network. So far our investigations have focused on radar data from a mountainous region in Italy where complex topography amplifies radar errors making the problem of flood forecasting with use of radar more challenging but simultaneously extremely important considering the large flash flood potential of those regions. Most recently, I have expanded this research to investigate the implications on flood forecasting associated with various passive microwave satellite-sampling scenarios, which is relevant to the planed international Global Precipitation Measurement mission research needs.