Network design for drought monitoring by geostatistical techniques

Monitoring of hydrometeorological events characterized by high variability in time and space requires the optimal selection of the number and location of the gauging stations within the area under study. A renewed interest about this topic has risen because of the increasing diffusion of automatic data acquisition networks, which forces to review the design criteria of traditional networks in order to reduce installation and management costs of the new gauges, as well as to better monitor in real time extreme hydrometeorological events, such as floods and droughts. In the present work, a methodology has been developed to assess the network capability to properly describe drought events, by verifying that the Root Mean Square Error (RMSE) of the Standardized Precipitation Index (SPI) in locations without gauging stations is lower than an acceptable value. The proposed approach, based on the use of kriging technique, has been applied to the automatic data acquisition network of rainfall stations managed by the Regional Hydrographic Service of Sicily. The results of the analysis allow to determine and compare the RMSE related to different network configurations in order to select a suitable network of stations to be adopted for drought monitoring in Sicily region.