We present the first dark-sky map of Thailand, derived from calibrated Visible Infrared Imaging Radiometer Suite (VIIRS) satellite data spanning 2012–2023. Artificial night-sky brightness was classified into 14 levels, with Classes 1–9 defined as potential dark-sky areas where the Milky Way remains visible. International comparisons with the United Kingdom, Chile, and Botswana reveal that Thailand has undergone the steepest decline, losing 15.4% of pristine skies since 2012, while the UK remained stable (+0.8%), Botswana nearly unchanged (−0.7%), and Chile moderately degraded (−5.3%). A correlation analysis shows strong negative associations between potential dark-sky area and both GDP () and population (), while inflation () and unemployment () exhibit weak influence. Five algorithms, including GLM and machine learning models, were tested; among them, the Decision Tree achieved the lowest relative error (), with ensemble methods and GLM performing comparably and Deep Learning being less accurate. By 2023, over 60% of Thais lived under skies too bright to observe the Milky Way by naked eye, and one-fifth were exposed to intensities preventing dark adaptation. Thailand’s rapid transition to LED street lighting after 2015, while energy-efficient, has intensified skyglow. Protecting remaining dark-sky areas requires urgent policies, linking conservation to human health, biodiversity, cultural heritage, and sustainable development.
We present the first dark sky map of Thailand using calibrated zenith artificial night-sky brightness from VIIRS satellite data covering 2012–2023. Results were compared with the United Kingdom, Chile, and Botswana to provide international context. Maps identified 14 classes of dark-sky areas, ranging from pristine sky to cone cell stimulation, with classes 1–9 designated as “potential” dark sky. Influencing factors—including population, GDP, inflation rate, unemployment, time, and country name—were analyzed using mathematical models. The number of potential dark-sky areas was best predicted by a Decision Tree model with a relative error of 0.4% ± 0.3%. Risk assessment suggests that Thailand’s potential dark-sky areas are significantly threatened and could disappear by 2068 without conservation. Since 2012, Thailand has lost 15.4% of its pristine skies, while the UK gained 0.8%. Botswana remained nearly stable (–0.7%), outperforming Chile (–5.3%). By 2023, ~60% of Thai residents had lost naked-eye visibility of the Milky Way, and one-fifth were exposed to very high nighttime light intensities, preventing dark adaptation. These findings highlight Thailand’s rapid dark-sky degradation and emphasize the value of satellite-based monitoring for conservation strategies, sustainable development policies, and international efforts to protect the natural night environment.
We evaluate binary rainfall occurrence classification in southern Thailand using 00 UTC radiosonde profiles from Phuket Airport and rainfall observations at Phuket, Krabi, Takua Pa (Phang-nga) and Nakhon Si Thammarat. Of 2631 soundings from 2011 to 2021, 1487 met quality-control and layer-continuity criteria. Five classifiers (logistic regression, support vector machines, random forests, extreme gradient boosting, and multilayer perceptrons) were evaluated using year-grouped cross-validation (2011–2019) and a chronological holdout (2020–2021). Holdout receiver operating characteristic area under the curve (ROC-AUC) reached 0.845 at Phuket, 0.793 at Krabi and 0.873 at Takua Pa, exceeding station-specific climatological baselines. At Nakhon Si Thammarat, the maximum holdout ROC-AUC was 0.691, indicating reduced spatial representativeness. Tree-based Shapley Additive Explanations (TreeSHAP) identified the K-index as the leading thermodynamic contributor across western stations, with geographically varying secondary contributions. Because daily rainfall windows include pre- and post-sounding periods, separate six-hour post-sounding evaluations (07:00–13:00, 13:00–19:00 and 19:00–01:00 local time) were conducted. Evening discrimination was consistently lower than morning discrimination across station–classifier combinations. A localized terrain–wind proxy provided negligible improvement. These results support temporally validated, station-specific rainfall occurrence classification while highlighting spatial representativeness and lead-time limitations for prospective operational use.