Alternative methodology to gap filling for generation of monthly rainfall series with GIS approach

Abstract

ABSTRACT As an alternative to Gap filling in monthly average rainfall series, we attempted to present a methodology for the generation of series only with the observed data available in the rainfall stations present in the study area and its surroundings. For this, a computational tool was developed with a GIS approach, using scripts in the Python language, to automate the study steps. Two calculation alternatives for the mean precipitation, variable Thiessen polygons or variable inverse distance weights (IDW), were considered. Random gaps were imposed from a series of data without gaps allowing us to evaluate the presented methodology. The results of the series calculated according to this methodology were compared to two methods of Gap filling. The behavior of the series was evaluated through the analysis of position and dispersion measurements as well as the temporal behavior by the evaluation of the correlograms and periodograms. The results are found to be satisfactory, which demonstrates the equivalence of the proposal with results found with the gap filling methods under the tested conditions. The differences found between the series were small, which was reflected in the Nash-Sutcliffe Indexes. There were no significant differences between the calculation alternatives by Thiessen polygons or IDW weights.

Keywords

• Hydrology
• Average rainfall
• Geoprocessing