Disdrometer calibration using an adaptive signal processing algorithm

Abstract

Disdrometers are considered exotic instruments and provide valuable information. As such, their price tag is also high. Impact disdrometers are instruments that produce an electrical impulse output related to the mass of a rain drop colliding at terminal velocity with a sensor. The produced electrical impulse signal amplitude and energy are related to the drop diameters. This relation is in general non-linear and depends heavily on the type of transducer used. Mechanical structure imperfections and electrical tolerances dictate the need for the individual calibration of each instrument in an attempt to create calibration curves that convert impulse amplitudes to equivalent drop diameters. Conventional calibration techniques using drop towers have been a tedious process to say the least. A proposed alternative calibration technique utilizing an adaptive signal processing algorithm eliminates the need of a single drop calibration. An accumulation rain gauge provides a reference signal to the disdrometer that is used for adaptive training and optimization of a model based calibration function. In this paper we describe a prototype low-cost disdrometer implementation at the University of Central Florida. A prototype impact sensor was built using an array of piezoelectric elements encapsulated in water resistant material. For the data acquisition and processing we use the soundboard of a general purpose computer. The signal processing algorithms and Matlab implementation will be described. Data have been collected and processed and results will be presented. Future plans on developing a low cost disdrometer will also be discussed. The availability of affordable disdrometers will benefit NASA's upcoming GPM program, as well as many other meteorological agencies.