Identifying Anomality in Precipitation Processes

Abstract:

Safety, risk and economic analyses of engineering constructions such as storm sewer, street and urban drainage, and channel design are sensitive to precipitation storm properties. Whether the precipitation storm properties exhibit normal or anomalous characteristics remains obscure. In this study, we will decompose a precipitation time series as sequences of average storm intensity, storm duration and interstorm period to examine whether these sequences could be treated as a realization of a continuous time random walk with both “waiting times” (interstorm period) and “jump sizes” (average storm intensity and storm duration). Starting from this viewpoint, we will analyze the statistics of storm duration, interstorm period, and average storm intensity in four regions in southwestern United States. We will examine whether the probability distribution is temporal and spatial dependent. Finally, we will use fractional engine to capture the randomness in precipitation storms.