Spatial sampling optimization of precipitation with multivariated geostatistics

This proposal will allow answering the following questions: How informative are the current precipitation monitoring networks in Ecuador? What are the optimal monitoring sites to achieve adequate representation of the spatial variability of precipitation? Traditionally, precipitation is observed with in-situ rainfall. Their observations are interpolated to obtain a continuous precipitation map. The quality of the map, and therefore of knowledge about the variability of precipitation, depends on both stages: sampling and interpolation. The problem lies where the rainfall is located. Despite the existence of other sources of information for precipitation such as radar and satellites, in-Situ rainfall remain one of the most reliable sources of information and therefore continue to strengthen themselves. In this context, it is essential to know if the location of the rainfall of the national network, which have been gradually densified since the 60s, allows the spatial variability of precipitation to be properly captured. This proposal will propose a spatial sample design for the continental Ecuador that allows optimal monitoring the spatial variability of precipitation. Multivariated geostatistics methods will be used in three stages: 1) Regionalization to identify regions with homogeneous precipitation processes; 2) the covariable analysis to improve the spatial prediction of precipitation; 3) and the selection of the most informative sampling sites in each region. The expected results will deepen knowledge about the behavior of precipitation, both in its spatial variability and in its relationship with covariables. In addition, the selected monitoring sites will improve the representation of the spatial precipitation variability. The results will be useful for public institutions (INHAMI, stage, CG-PAUSTE) and research groups to make an adequate deployment of their precipitation monitoring networks.

14th RESEARCH PROJECT COMPETITION