Run analysis of rainfall data affected by urbanization.

Urbanization generally increases the amount of rainfall in urban areas and/or downwind of urban centers. This paper investigates whether urbanization significantly changes the run properties of rainfall. Three run properties, the run-length, run-sum and run intensities of positive (surplus) and negative (deficit) runs, of the annual and monthly rainfall were evaluated by using the theory of runs. Annual rainfall was assumed to be normally of gamma distributed, and the monthly rainfall assumed to follow a two-state Markov chain with stationary, transition probabilities. The methodology was applied to annual and monthly rainfall data at La Porte and three surrounding stations in Indiana. The results indicated that these assumptions are adequate to characterize the annual and monthly rainfall. The following conclusions were drawn: 1) the run properties such as run-lengths, run-sums or run-intensity are not significantly affected by urbanization; 2) run intensity is a better indicator of changes in rainfall than run-lengths or run-sums; and 3) the run properties of both annual and monthly rainfall data at La Porte are not significantly different from those at the surrounding stations, even though La Porte has received increased rainfall in comparison with its environs as a result of urbanization. The run properties of La Porte annual data in the pre- and post-urbanized periods appear to be different. However no significant differences were observed in the run properties of monthly data at La Porte in these two periods.