Precipitation mapping at local, regional and global scale

Abstract

It is well established that the Earth's water cycle is accelerating, and extreme precipitation events are becoming more common. While we cannot avoid this issue, we can be better prepared to handle it if we can obtain accurate observations of precipitation for use in short-term and long-term prediction models. Various remote sensing instruments are available to obtain precipitation data. In this research work, mapping precipitation at local, regional, and global scales is studied. The technology of precipitation mapping at these scales is very different and elaborated. Examples of precipitation measurements from these scales are discussed. At the local scale, rain gauges and disdrometers are two prominent instruments that are utilized for precipitation measurement. Precipitation observations captured from these two instruments are introduced. Millimeter wave radars have been previously used in various domains, and extensive research is currently in progress to improve this technology. This research will present the potential of using automobile class radars to obtain local surface precipitation. Since the maximum range of an automobile radar is within a few hundred meters, we can consider the observations to be at a local scale. With the help of signal modeling, methods to obtain the rainfall rate at the millimeter wave band by using radar parameters, such as reflectivity and attenuation are discussed. A simulation tool is developed that generates the radar signals at the millimeter wave frequency band. The various parameters which are used in the signal simulations are explained in detail, and the simulation results are presented. Experiments for mapping precipitation using a current state-of-the-art automobile radar are carried out, and the results are discussed. The reflectivity value obtained from the experiment using automobile radar is compared to the NWS reflectivity mosaic, and the results match within a couple of decibels (dB). Weather radars are remote sensing instruments that provide precipitation observations at a regional scale. They provide data at a large spatial extent. Weather radar observations obtained at various frequency bands for mapping precipitation is discussed with examples. The current networks of instruments and system architectures that provide precipitation information at a regional scale are discussed. The precipitation data obtained from individual automobile radars is considered as a local data point, and precipitation maps at the regional scale are constructed. The system analysis of using a network of automobile radars for mapping precipitation is discussed with the help of simulations. The Dallas-Fort-Worth urban region is considered for the simulation study, and the potential of using millimeter wave radars to create precipitation maps is presented. Three different interpolation techniques, linear, nearest-neighbor, and natural are explored to study the reconstruction of precipitation maps. A system architecture for precipitation mapping using automobile radars is also discussed. The attenuation of radar signals has to be addressed and corrected to obtain accurate precipitation information from radar data. The attenuation correction in weather radars for rain hydrometeors is well studied in the literature, but attenuation correction for snow is limited. This is due to the fact that snow does not attenuate much at lower frequency bands like S and C bands and because the snow particles vary in their particle size distributions and have complex shapes. Theoretical relationships between specific phase and attenuation are developed using signal simulations. This research will introduce a new algorithm that corrects radar signal attenuation in rain and snow. The attenuation correction method developed is applied to X-band and Ku-band radar data, and the results are discussed. It was observed from the data for snow cases that the path integrated attenuation at the X-band reached up to 2 dB and, at the Ku-band, it reached up to 8 dB. Mapping precipitation at a global scale is a challenging task. The Dual Precipitation Radar (DPR) is a spaceborne instrument providing valuable precipitation information at the global scale, but the observations from this instrument suffer from poor spatial and range resolutions. Synthetic Aperture Radars (SAR) are well known for providing high spatial resolution data. In the past, SARs have been deployed on airborne and spaceborne platforms for mapping land cover and constructing surface elevation models. The potential of using SAR for mapping precipitation is not widely explored. In this research, SAR signal simulations are carried out to observe precipitation from spaceborne platforms. The mathematical framework for monostatic and bistatic SAR is discussed. The simulation results for two specific spaceborne SAR architectures are discussed in detail. The variation of precipitation parameters such as velocity and spectral width are studied using simulations. This dissertation presents the roles and challenges of observing precipitation at the three scales, with suggestions for future research.