<![CDATA[Urban Heat Island (UHI) information plays a crucial role in microclimate monitoring and understanding environmental phenomena. However, the availability of observation station data is often spatially limited. This study aims to estimate Land Surface Temperature (LST) distribution in Buleleng Regency using Cokriging, utilizing the Normalized Difference Vegetation Index (NDVI) as a secondary variable due to its physical correlation with temperature. Addressing the significant scale disparity between LST and NDVI, this study applies Z-score transformation prior to modeling to ensure covariance matrix stability. Experimental variograms were constructed using Sturges' rule for lag distance determination and modeled using the Linear Coregionalization Model (LCM) to maintain a positive definite kriging matrix. Model evaluation using Leave-One-Out Cross-Validation (LOOCV) revealed a strong negative correlation between LST and NDVI (Pearson coefficient of -0.890). The Gaussian model was selected as the best fit, indicated by a low Mean Squared Error (MSE) of 2.7680 and a Mean Absolute Percentage Error (MAPE) of 3.63%. These results demonstrate that integrating NDVI through Cokriging significantly improves spatial estimation accuracy. Furthermore, this study supports the Sustainable Development Goals (SDGs), particularly Goal 13 (Climate Action) and Goal 15 (Life on Land), by providing high-precision environmental data essential for effective climate resilience planning.]]>
Copyrights © 2026
