Testing some grouping methods to achieve a low error quantile estimate for high resolution (0.25° x 0.25°) precipitation data

Abstract

The study focuses on the estimation of a technique, a method for developing a phenomenon, to obtain a quantile with minimal or low error (AR and R-RMSE bias). To arrive at such a solution, a case study of the Mahanadi River system (Mahanadi Basin) was conducted along with the integration of various techniques available in past and present literature, to come up with a novel solution. Which could answer practical questions in water resource planning and management for addressing a wide range of problems such as meteorological draught analysis, agricultural planning, precipitation forecasting and downscaling, design of water control and conveyance structures, and land-use planning and management. A gridded rainfall data set of resolution 0.25° x 0.25° (1901 – 2017) obtained from IMD Pune is used to calculate the statistics that will be used for the regionalization of precipitation. Other attributes or variables used for regionalization are seasonality measurements and location parameters (latitude, longitude, and elevation). The L-moment statistics are computed from the time series rainfall data and the ratios of the L-coefficient of variance and the L-coefficient of skewness, i.e., the L-moment ratio, are the main components in computing quantile estimates of selected regions for effective regional frequency analysis. To determine potential scenarios for homogeneous regions, the use of seasonal extreme precipitation will serve as a basis for regionalization.