Interpolación espacial de la precipitación pluvial en la zona de barlovento y sotavento del Golfo de México

this study was conducted in the digital agro-maps laboratory of the central gulf region of the national research institute for forestry, agriculture and livestock in 2007. the objective was to compare four rain interpolation methods using the historic precipitation data of september from the period 1961 to 2003 of 147 climatic stations belonging to the national weather service. these stations cover a region of the windward and leeward direction in mexico. in order to define a suitable method, four interpolation methods were compared: inverse of the distance, krigging, co-krigging and thin plate smoothing spline. data was analyzed to define if there is relationship of rainfall with altitude, longitude and distance from the ocean. the method of thin plate smoothing spline was superior since it gave the lowest mean square error followed by the method of krigging and co-krigging using gaussan method. it is important to mention that with the co-krigging method, altitude was used as a covariate despite that there was not a strong correlation of this variable with precipitation. the last option for rainfall data interpolation was the inverse of the distance with optimized values.