MODELING OF ISOTROPIC SEMIVARIOGRAM WITH BOXCOX TRANSFORMATION ON DATA AND USING COKRIGING INTERPOLATION METHOD

• Main Author: Anastasya
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/73031
• Institution: Institut Teknologi Bandung

Isotropic semivariogram is a statistic used to measure the spatial correlation of location pairs separated by a certain distance. The estimation of the experimental semivariogram used is Matheron with the Sturgess lag distance division. Often, data with outliers affect the estimation of semivariogram values. In this study, an isotropic semivariogram model will be determined for peatland data using Ordinary Least Squares parameter estimation and the cokriging interpolation method. Peatland is land formed from the accumulation of decomposed plant remains, both partially decomposed and undecomposed. The peatland data used is from Sumatra Island, obtained from the prims.brg.go.id website. The selected variables are Groundwater Level and rainfall. The Groundwater Level distribution has many outliers, so it will be transformed using Box-Cox transformation. The parameter estimation method used is Ordinary Least Squares. The appropriate theoretical semivariogram model will be validated using the Q1 test method. The suitable semivariogram model for the GWL variable is Spherical, for rainfall is Gaussian, and for the Box-Cox transformed GWL is Exponential. Furthermore, the values at several locations will be estimated using the data interpolation method called Cokriging. The results of the cokriging estimation, with the Box-Cox transformed Groundwater Level as the primary variable, are closer to the actual data by a factor of 22 compared to the primary variable Groundwater Level, which is 180 times the actual data.