Fig 1.
The scheme of the methodology used in this study.
Fig 2.
Raster data visualization used in this study.
Table 1.
Summary of the data type and format.
Fig 3.
Product of the night sky clarity level: a) Distribution of the night sky clarity in Indonesia and several specific areas such as A. Greater Jakarta, B. Bosscha Observatory, C. ITERA Astronomical Observatory (IAO), D. The National Timau Observatory, and b). Average value of the clarity of the night sky in each province in Indonesia.
Fig 4.
Observatory suitability index based on a combination of the Earth’s physical and demographic–economic parameters in Indonesia, with several provinces with high index values such as A. Aceh, B. North Sumatra, C. East Kalimantan, D. Central Sulawesi, and E. Papua.
Fig 5.
Spatio-temporal variation in the average monthly atmospheric index value distribution in 2019–2021.
Fig 6.
Total atmospheric factor index, with several provinces with high index values, such as A. Aceh, B. North Kalimantan, C. East Nusa Tenggara, D. Central Sulawesi, and E. Papua.
Fig 7.
The results of a machine learning-based analysis of suitability levels for constructing observatories: A. displays the results of the random forest (RF) algorithm, B. presents the outcomes of the Gradient Tree Boost (GTB) algorithm, C. illustrates the combined product of both machine learning approaches, and D. shows the statistical results including AUC, validation kappa, and overall accuracy.
Fig 8.
The analysis of agreement between two different approaches to assess the suitability of observatory sites: Multi-criteria and the multi-machine learning approach.
Fig 9.
Combined observatory suitability index distribution zoning based on A). Latitude Zone (L1–L10), B). Longitude Zone (B1–B10) and C. Optimal distribution of observatory scenario: Locations of eighteen recommendations for the construction of a new Indonesian observatory.
Fig 10.
Machine learning statistical evaluation: a. Parameter correlation matrix, b. Predictive power score, c. Variable importance of astronomical observatory and d. distribution value of the existing observatory site to each parameter.
Fig 11.
Bivariate visualization between astronomical observatory suitability index and multi-air pollution risk product.
Fig 12.
Index comparison results for three main observatory locations: a). Aerosol optical depth, b). Night light, c). Wind speed, d). Cloud cover, e). Precipitable water vapor, and f). Rainfall.