<![CDATA[Advancements in remote sensing technology have enabled the use of satellite imagery, such as Landsat 8 and HLS-L30, for the spatial and temporal estimation of Land Surface Temperature (LST) with improved resolution. In the context of geothermal exploration, the availability of thermal infrared bands in these datasets facilitates more efficient and cost-effective mapping and identification of surface temperature anomalies, particularly across large and inaccessible areas. This study aims to compare LST estimations derived from Landsat 8 and HLS-L30 imagery using the Mono Window Algorithm (MWA) and Split Window Algorithm (SWA) at 18 geothermal manifestation points within the Mount Ungaran Geothermal Working Area (WKP). A Focal Statistic process was applied to 20 LST datasets, resulting in a total of 100 LST layers. From each layer, LST values were extracted at the 18 manifestation points, producing a total of 1,800 data points. A binary logistic regression analysis was conducted using these LST values alongside those from 20 randomly selected comparison points. The results indicate that the median LST derived from HLS-L30 imagery using the Split Window Algorithm with the minimum Focal Statistic yielded the most optimal performance in classifying geothermal manifestation presence. This method achieved statistical significance (p = 0.028), indicating its capability to effectively distinguish between manifestation and non-manifestation points. However, the pseudo-R² value of 0.107 suggests that the model explains approximately 11% of the variance in the data. These findings underscore the potential application of satellite-based LST analysis in the early detection and assessment of geothermal surface anomalies within WKPs.Keywords : Geothermal, LST, Landsat, HLS-L30, Ungaran]]>
