Can gridded data represent extreme precipitation events?

Abstract

The analysis and characterization of extreme precipitation at regional scale requires data at high temporal and spatial resolution due to the abrupt variations of this variable in time and space. In recent years there has been an increasing demand for comprehensive regular high-resolution (both in time and space) gridded datasets from different sectors, including hydrology, agriculture and health which are severely affected by extreme events. One of the main shortcomings of gridded datasets is that extreme events can be smoothed during the interpolation process. Heavy rainfall events can be very local and, hence, interpolation with neighboring stations may lead to an underestimation of the precipitation amounts. In this work we study the capability of a high-resolution daily precipitation gridded dataset over Spain (we refer to this dataset as Spain02, Herrera et al 2010) to characterize extreme precipitation. A dense network of 2756 quality-controlled stations was selected to develop the Spain02 grid with a regular 0.2º horizontal resolution covering the period from 1950 to 2003. We study both upper percentiles and the extreme indicators commonly used to characterize extreme precipitation regimes. We also show the performance of the gridded dataset to capture both the intensity and the spatial structure of severe precipitation episodes which constitute characteristic ephemerides of extreme weather in the Iberian peninsula. The results are compared to the 25 Km E-OBS grid (Haylock et al 2008) developed in the ENSEMBLES project, which is the best daily dataset for the whole Europe to date.