Influence Rainy and Dry Season to Daily Rainfall Interpolation in Complex Terrain of Sulawesi

Abstract

Spatial interpolation widely is used to predict rainfall in ungauged location. There are two methods of spatial interpolation, deterministic method such as Inverse Distance Weighting (IDW) method and geostatistics such as kriging. Daily rainfall in the tropical Indonesian maritime region such as Sulawesi is generally uneven. Although, in a location has heavy rainfall, a zero rainfall can be found in adjacent places. It makes choosing suitable spatial interpolation method for ungauged rainfall prediction is difficult. The aim of this research is evaluate the impacts of the rainy season to the interpolation method. Root mean square (RMSE) and correlation is used as statistical parameter to determine the goodness of interpolation method. The result shows that heavy rain event has big impact to interpolation, both IDW and kriging. Log transformation can little improve in RMSE in rainy season, but not recommended in dry seasons.The most biggest deviation of the interpolation result spreads in south west coast, plain at middle of Sulawesi and north part of the Bone Gulf in January and only south west coast of Sulawesi and north part of the Bone Gulf in dry season July.