Regionalised spatiotemporal rainfall and temperature models for flood studies in the Basque Country, Spain

摘要

A spatiotemporal point process model of rainfall is fitted to data takenfrom three homogeneous regions in the Basque Country, Spain. The model is thesuperposition of two spatiotemporal Neyman–Scott processes, in which raincells are modelled as discs with radii that follow exponential distributions.In addition, the model includes a parameter for the radius of storm discs, sothat rain only occurs when both a cell and a storm disc overlap a point. Themodel is fitted to data for each month, taken from each of the threehomogeneous regions, using a modified method of moments procedure thatensures a smooth seasonal variation in the parameter estimates.Daily temperature data from 23 sites are used to fit a stochastictemperature model. A principal component analysis of the maximum dailytemperatures across the sites indicates that 92% of the variance isexplained by the first component, implying that this component can be used toaccount for spatial variation. A harmonic equation with autoregressive errorterms is fitted to the first principal component. The temperature model isobtained by regressing the maximum daily temperature on the first principalcomponent, an indicator variable for the region, and altitude. This, togetherwith scaling and a regression model of temperature range, enables hourlytemperatures to be predicted. Rainfall is included as an explanatory variablebut has only a marginal influence when predicting temperatures.A distributed model (TETIS; Francés et al., 2007) is calibrated for aselected catchment. Five hundred years of data are simulated using therainfall and temperature models and used as input to the calibrated TETISmodel to obtain simulated discharges to compare with observed discharges.Kolmogorov–Smirnov tests indicate that there is no significant difference inthe distributions of observed and simulated maximum flows at the same sites,thus supporting the use of the spatiotemporal models for the intendedapplication.