This study aims to develop and validate a physically based method for detecting monsoon onset over Celebes Island, Indonesia, a region characterized by complex topography and diverse rainfall regimes. The proposed approach integrates local rainfall variability with regional atmospheric circulation to provide a more robust onset definition. Rainfall data from the Multi-Source Weighted-Ensemble Precipitation version 2 (MSWEP-v2) were validated against six stations from the Indonesian Agency for Meteorology, Climatology, and Geophysics (BMKG), showing strong agreement (r = 0.65–0.98, p < 0.05) and reliable representation of seasonal cycles. Harmonic analysis was applied to classify the region into five subregions based on dominant rainfall cycles. Monsoon onset was initially identified using the cumulative anomaly (CA) method and subsequently refined using backward trajectory analysis of atmospheric moisture sources derived from reanalysis data. The results show that the climatological mean onset occurs in late November, with spatial variability influenced by both moisture transport pathways and topographic conditions. Early- and late-year onsets are associated with moisture transport from the northern and tropical Maritime Continent, whereas mid-year onsets are linked to southern sources. The onset timing is also consistent with a clear reversal of low-level wind direction, indicating the transition between inactive and active monsoon phases. Compared to conventional rainfall-based methods, the integrated approach provides a more physically consistent representation of monsoon onset by incorporating both local and large-scale processes. This improved characterization has important implications for seasonal forecasting, early warning systems, and climate-related decision-making in regions with complex climate variability such as the Maritime Continent.

The dominant seasonal rainfall patterns across the Indonesian Maritime Continent (IMC) and its surrounding oceans (90 degrees E-145 degrees E, 11 degrees S-10 degrees N) were investigated using IMERG V07B satellite precipitation data and rain gauge observations from 2001 to 2023. An empirical orthogonal function (EOF) analysis of monthly rainfall revealed that the leading mode of variability is strongly associated with the Western North Pacific Monsoon (r = -0.83) and sea level pressure over the South China Sea (r = 0.64). Dominant spatial-temporal rainfall patterns were characterized by k-means clustering of the first five EOF modes, collectively explaining more than 50% of the total variance. Five distinct seasonal rainfall regimes were identified and refined through correlation-based reclassification. In addition to two monsoonal regimes in the southern and northern IMC-associated with the Australian monsoon and South China Sea summer monsoon (SCSSM), respectively-three equatorial clusters were discovered. The western equatorial IMC demonstrates a bimodal distribution, with maximum rainfall occurring in November (320.67 f 114.91 mm) and May (204.44 f 57.88 mm), while the eastern equatorial IMC exhibits a peak in June (282.49 f 118.43 mm) and January (203.52 f 105.77 mm). These seasonal contrasts are modulated by monsoonal moisture transport, latitudinal shifts of the Intertropical Convergence Zone (ITCZ), and orographic influences. A distinct regime over the southeastern Indian Ocean was identified, characterized by primary and secondary rainfall peaks in November, July, and March, likely associated with sea surface temperature variability, was also identified. The study's findings serve to update the characterization of seasonal rainfall zonation and peak timing across the IMC, thereby ensuring a more accurate reflection of current climate variability.

El Niño and La Niña events typically affect rainfall patterns and water availability for plants, especially in rainfed and upland farming systems. This study examines year-to-year variations in rainfall associated with El Niño and La Niña events, as well as their impacts on potential planting seasons and the management of food crop planting patterns in Malang Regency. The rainfall data used in this analysis are from the Karangploso and Karangkates climatology stations for the period 2012-2024. The analysis reveals that over the past 13 years, El Niño events have not shifted the duration of the dry or rainy seasons in Lawang Subdistrict, nor have they reduced the rainy-season duration by 1 decade compared to normal conditions in Donomulyo Subdistrict. However, La Niña events can prolong the rainy season by 3-13 decades in Lawang Subdistrict or 2-14 decades in Donomulyo Subdistrict. The average potential planting time at the research location is 210 days in Lawang Subdistrict and 240 days in Donomulyo Subdistrict, posing a significant risk of planting rice across two growing seasons. By selecting adaptive crops and managing planting patterns, it could be possible to plant three times using a rice-corn-beans pattern during the planting seasons in the Lawang Subdistrict or a rice-corn and rice-beans pattern in the Donomulyo Subdistrict. During La Niña events, crop pattern management can be more flexible, and planting intensity can be increased by 3-4 times through effective crop pattern management.

Accurate and consistent rainfall data are essential for climatological analysis and disaster risk management, particularly in regions with complex climatic dynamics such as Indonesia. This study evaluates the performance of Integrated Multi-satellite Retrievals for GPM (IMERG) Final Run version 07B (V07B) against its predecessor, version 06B (V06B), in estimating seasonal rainfall magnitude and peak timing across Indonesia over a complete 20-year period (2001–2020). Results show that IMERG V07B demonstrates significant improvements in accuracy, with lower relative bias (RB) (V06B = +19.04%; V07B = +9.68%) and root mean square error (RMSE) (V06B = 109.92 mm/month; V07B = 102.65 mm/month), along with slightly higher correlation coefficient (CC) (V06B = 0.77; V07B = 0.78). These improvements are consistent across most stations and are attributed to updates in the Goddard profiling algorithm to account more effectively for surface type and orographic influence. Despite substantial differences in monthly rainfall magnitudes between the versions, particularly over oceanic, mountainous, and coastal regions, the annual cycle phase remains relatively consistent. Harmonic fitting further enhances peak timing accuracy, increasing the proportion of correct estimates from 62.69% (V06B) to 71.64% (V07B), underscoring the suitability of V07B for seasonal rainfall analysis. Spatial assessments reveal that V07B generally produces higher rainfall estimates than V06B over land and ocean but lower values in coastal zones. Based on these findings, IMERG V07B is recommended for seasonal zoning and rainfall pattern studies in Indonesia. Further research using sub-daily and near-real-time data is needed to support high-impact applications such as flood early warning systems.

Model simulations for Borneo suggest potential difficulties in capturing the variability of rainfall on regional scales due to complex topography and climate interactions. This study utilizes a high-resolution Conformal Cubic Atmospheric Model (CCAM) to address the impacts of climate change on a regional scale over Borneo by analyzing key parameters such as precipitation, surface temperature, mean sea level pressure, and relative humidity over Borneo for the near future (2023-2033). We also found that the MAM (March-April-May) and SON (September-October-November) seasons show higher rainfall than the DJF (December-January-February) season. Meanwhile, relative humidity tends to decrease in the future compared to previous years. Mean sea level pressure (MSLP) tends to increase over the border of Indonesia and Malaysia in the JJA (June-July-August) season but decreases over southern Borneo in the MAM (March-April-May) and SON (September-October-November) seasons. There is a noticeable increase in temperature from the MAM (March-April-May) to SON (September-October-November) season in the east-south region. However, further research is required to validate these projections.

In recent years, the BMKG (Indonesian Meteorological, Climatological, and Geophysical Agency) has provided premium climate services to PT Berau Coal, including 10-day and monthly rainfall forecasts. These services are crucial for mining operations in preparing work plans, as rainfall can make roads at the mining site slippery, leading to reduced effective working hours, increased pump operating hours, and higher fuel costs. This study evaluates the performance of monthly rainfall predictions for three mining sites (Sambarata, Lati, and Binungan) over the period 2005-2023, based on the availability of rainfall observation data at the mining sites, using the RMSE and Percent of Correct (PoC) methods. The analysis shows that for a lead time of one (six) month, i.e., a forecast given one (six) month in advance, the average RMSE is 94 (95) mm/month. The highest error is observed in January, at 140 (144) mm/month, while the lowest error occurs in August, at 65 (69) mm/month. The PoC values range from 42% to 89%. These results indicate that the climate service forecasts have reasonably good accuracy and can be used as a reference in decision-making, although potential errors, as indicated by the RMSE values, should still be considered.

This study examined the diurnal variation in rainfall in Kalimantan, focusing on precipitation accumulation (PA), precipitation frequency (PF), and precipitation intensity (PI), using rain gauge data from 103 stations (2016-2022) and Integrated Multi-Satellite Retrievals for Global Precipitation Measurement Version 07 B (IMERG V07B) for an extended period (2002-2022). The results showed that PA was more closely aligned with PF than with PI, indicating that PF played a dominant role in determining total precipitation across Kalimantan. Diurnal peaks of PA and PF were observed in coastal regions during the late afternoon to early evening (14-19 local standard hours [LST]), whereas inland and mountainous regions experienced these peaks in the early morning (00-03 LST). In contrast, PI exhibited greater spatial and temporal variability, reflecting the influence of localised convective processes. Cluster analysis of PA, PF, and PI identified four distinct regions based on peak time and amplitude, showing a strong dependence on topographic features and distance from the coastline. Short-duration rainfall events (< 3 h) were dominant across all parts of the island, with the highest percentages in mountainous areas. In contrast, long-duration events (> 6 h) tended to dominate in inland flat regions, although they were still occasionally observed in mountainous areas during the early morning hours. These patterns were driven by onshore rainfall propagation, as observed in the IMERG time cross-sections, where rainfall moved inland from the southwestern coast between 18 and 00 LST and persisted over mountainous areas for 1-3 h before propagating northeastward between 03 and 08 LST. Offshore propagation observed in the northwest and southeast regions of Kalimantan between 00 and 06 LST further highlights the role of low-level wind patterns and local atmospheric convergence in modulating rainfall distribution. This study enhances our understanding of the complex diurnal rainfall patterns in Kalimantan.

Mixed Rossby-gravity (MRG) waves are key equatorial disturbances that modulate convection and rainfall across the Maritime Continent, yet their representation in regional models remains underexplored. Accurate simulation of MRG waves induced vertical structure is critical for improving forecasts of tropical weather variability and associated hydrometeorological hazards. This study evaluates the Weather Research and Forecasting (WRF) model's ability to simulate vertical atmospheric profiles during an MRG wave event on the southern coast of West Java from 26 to 29 September 2022, using radiosonde observations at 06, 12, and 18 local time (LT). Eleven model configurations were evaluated, differing in domain schemes (two vs. three nesting steps), vertical resolution (33, 45, 60, 80, and 100 levels), and input data sources (Real-Time Global Forecast System [GFS] vs. Final Operational Global Analysis [FNL]). All model configurations used in the simulation have a fixed physics scheme parameterization. The simulations were compared with radiosonde observations and evaluated statistically using the correlation coefficient (R) and Normalized Mean Absolute Error (NMAE). The analysis demonstrates that the WRF model effectively captures MRG wave dynamics by simulating key atmospheric variables, including pressure (P), temperature (T), relative humidity (RH), zonal (U) and meridional (V) wind anomalies in strong agreement with observations. P is well represented, exhibiting the highest R (0.81), whereas RH is the lowest (0.21), likely reflecting the model's inability to capture fine‑scale observed moisture variations. Configurations that utilized a two-step nesting domain and the FNL input demonstrated the best performance, achieving higher R values and lower NMAE. Input data had a notable impact on model performance: the FNL analysis improved R by ~36% and reduced NMAE by ~12% compared to GFS, likely due to FNL's assimilation of observational data, which reduces uncertainty. Moreover, a domain scheme with a smaller outer domain and fewer nesting steps also improved R by ~36% and reduced NMAE by ~12%, suggesting that simpler domain configurations help limit error propagation. Additionally, increasing the vertical resolution from 33 to 100 levels enhanced the simulation of MRG wave structures, improving R by ~45% and reducing NMAE by ~22%. These findings enhance the understanding of MRG wave dynamics and offer valuable insights for improving regional weather forecasting.

Ketahanan pangan merupakan isu penting yang menjadi perhatian negara-negara di dunia. Proyeksi pada tahun 2020, sekitar 690 juta orang berada dalam kondisi kelaparan. Diperkirakan angka penduduk malnutrisi berjumlah lebih dari 760 juta pada tahun 2021, dan akan mencapai lebih dari 800 juta pada tahun 2030. Salah satu faktor yang berpengaruh terhadap ketersediaan dan produksi pangan adalah variabilitas dan perubahan iklim. Hal ini dapat berpotensi mengakibatkan frekuensi kejadian cuaca eksrim meningkat yang dapat memperburuk kondisi ketahanan pangan global terutama di daerah-daerah yang rentan dengan kelaparan dan malnutrisi. Aksesibilitas pangan, ketersediaan pangan dan pemanfaatan pangan yang merupakan merupakan variabel ketahanan pangan yang rentan terhadap variabilitas Iklim (García-Díez dkk, 2021). Selain itu, kebijakan dan strategi pemerintah juga berperan penting dalam stabilitas produksi pangan dalam menghadapai tantangan ini. Kemajuan sains dan teknologi dalam bidang klimatologi seperti pemanfaatan satelit cuaca, radar cuaca dan semakin luasnya jaringan pengamatan iklim dapat menjadi modal untuk dapat meningkatkan pemahaman tentang karakteristik iklim. Melalui pemahaman yang baik tentang iklim, diharapkan dapat dijalankan berbagai strategi dalam mewujudkan ketahanan pangan dalam pertanian dan perikanan, seperti program pertanian-perikanan berkelanjutan, pengembangan pertanian-pertanian pintar, dan diversifikasi produksi pangan. Kemajuan teknologi komputasi berkinerja tinggi juga memiliki peranan dalam memprediksi iklim dan cuaca ekstrim lebih akurat. Hal ini dapat menjadi masukan bagi para pemangku keputusan seperti pemerintah pusat dan daerah dalam menerapkan kebijakan ketahanan pangan.

The movement direction of propagating convective systems originating from both inland and offshore over the north coast of West Java in Indonesia is determined primarily by the prevailing wind. However, the role of land composition over the western part of the Indonesian Maritime Continent (IMC) is also expected to affect the development of propagating convective systems by enhancing upward motion. This hypothesis is tested using a Weather Research and Forecasting model incorporating convection-permitting with 3 km spatial resolution to simulate the heavy rainfall event during the 2002 Jakarta flood. We addressed the influence of land on the local circulation, particularly in the area surrounding Jakarta, by replacing the inland over the western IMC (96°–119°E, 17°S–0°) with a water body with an altitude of 0 m. We then compared the results of model simulations with and without land. The results show that land-sea difference forcing has a significant role in enhancing upward motion and generates a deep convective cloud in response to the land-based convective system, which then continuously and rapidly propagates offshore due to the cold pool mechanism. This land effect mainly triggers gravity waves and also results in early morning convection over coastal regions.

Understanding rainfall variability and trends across Indonesia is crucial for developing strategies to mitigate the impacts of climate change. This study examines the spatial-temporal variability and trends in rainfall across Indonesia using long-term (1981-2023) Climate Hazards Group InfraRed Precipitation with Stations (CHIRPS) v2.0 data. Rainfall variability was quantified using monthly and daily rainfall indices, while trends were examined with the Mann-Kendall test and the Theil-Sen method. Evaluation against 67 rain gauge stations showed high reliability of CHIRPS v2.0 data (CC > 0.7; RB < +/- 25%), supporting its use for long-term analysis. The highest interannual variability, characterised by a coefficient of variation (CV) greater than 30%, is observed in southeastern Sulawesi, Maluku and southwestern Papua. In contrast, the interannual lowest variability (CV < 15%) occurs in Sumatra and Kalimantan, which also exhibit the lowest precipitation concentration index (PCI < 10) and precipitation concentration degree (PCD < 0.2). These patterns reflect strong influences of topography and geographic position on rainfall distribution. Trend analysis shows a general increase in annual rainfall across most of Indonesia, with significant rises (> 20 mm/year, p < 0.1) in central-northern Sumatra, southeastern Sulawesi, Maluku and southwestern Papua, while northern Papua near the Pacific margin shows significant declines towards drier conditions. Monthly rainfall trends show strong positive changes in June and November and negative trends in September, reflecting heterogeneous monsoonal responses to climate variability. In addition to changes in total rainfall, significant upward trends are observed in the frequency and intensity of extreme rainfall events across parts of Sulawesi, Maluku and Papua, consistent with regions exhibiting the highest CV. The greatest increases in extreme rainfall frequency (R50mm), intensity (RX1day) and amount (R99p) are observed in Southern Sulawesi. The spatial disparities of rainfall trends and extremes across Indonesia emphasise the need for region-specific adaptation strategies.

Malaria continues to pose a major public health burden in Indonesia, with transmission patterns varying widely across regions. Western provinces have made significant strides toward elimination, while persistent transmission remains in eastern zones, particularly Papua. Climatic variability is a key driver of malaria dynamics, yet its integration into routine surveillance and prediction systems remains limited. To address this gap, this study introduces framework that integrates malaria surveillance data (2019–2025) with climate observations (BMKG, Copernicus ERA5) and demographic data (BPS) to enhance early warning systems and prediction accuracy. Three models were developed to capture distinct malaria transmission dynamics: deep learning architecture (LSTM-GRU-Attention) for endemic zones in Central Papua, hurdle model for elimination zones in Lampung, and univariate time-series model for Mimika Regency as a high-burden hotspot. Results indicate that zone-specific approaches substantially improve predictive performance, with the elimination zone model achieving higher accuracy (SMAPE 19.05%) compared to the endemic zone model (SMAPE 172.21%). These findings highlight the importance of tailoring modeling strategies to transmission intensity and demonstrate how integrating climatic and epidemiological data can strengthen surveillance and accelerate progress toward malaria elimination in Indonesia.

Sub-seasonal to seasonal (S2S) prediction has emerged as an important tool in anticipating climate variations over shorter timescales, from a few weeks to several months ahead. This research undertakes multiple evaluations of verification results derived from various deterministic and probabilistic forecasting approaches at the S2S scale, employing diverse techniques accessible within the Python tool named Xcast. Developed as a proficient utility, Xcast is a tool capable of utilizing statistical and machine learning methodologies to rapidly and effectively process various gridded climate data. The study conducts a comparative analysis of several methods, including multiple linear regression (MLR), extreme learning machine (ELM), and probabilistic output extreme learning machine (POELM). The assessment employs blended rain data from rain posts and Global Satellite Mapping of Precipitation (GSMaP), alongside S2S the European Center for Medium-Range Weather Forecasts (ECWMF) forecast data—both data are on a 10-day time scale with the period from 1996–2021 tailored for Indonesia region. The research domain employs a condition whereby the tercile probability is determined by data points that accumulate rainfall of over 50 mm per 10 days, with a 30

The availability of surface rainfall data with high spatial -temporal resolution is needed to understand the diurnal rainfall characteristics in the Indonesian Maritime Continent (IMC) to improve the accuracy of weather and climate models in this region. This study validates the accuracy of the Final Run product of IMERG data version 06B (V06B) and version 07A (V07A), which have a resolution of 0.1 degrees - 30 min for diurnal rainfall analysis over IMC. Validation was conducted for precipitation amount (PA), precipitation frequency (PF), and precipitation intensity (PI), by recording 302 automatic rain gauges (RG) every 10 min from January 2014 to September 2021. IMERG V06B and V07A perform well in observing diurnal PA and PF but struggle in observing diurnal PI compared to RG observations. This has been determined by the correlation coefficient (CC) values of IMERG V06B (V07) data, which are 0.76 (0.72) for PA, 0.77 (0.76) for PF, and 0.21 (0.13) for PI. The IMERG data align to an extent with RG observations for the PA value, having a small relative bias (RB). The results also display a systematic PF and PI values error. The high false alarm ratio (FAR) of IMERG data suggests rain detection errors, leading to overestimated PF values. Additionally, the underestimation of PI values is due to the limitation of IMERG data in observing extreme and small-scale rain events. About 86.10% (78.14%) and 81.45% (81.78) of IMERG V06B (IMERG V07A) data show a peak time difference of <3h for PA and PF when compared with RG observations. Overall, IMERG V06B performs better than V07, possibly due to inaccurate orbits of GPROF data, removal of SAPHIR satellite observations, and inter -calibration issues with CORRA and GPCP data. Continuous monitoring and data improvements are necessary to improve the accuracy and reliability of IMERG data in detecting diurnal rainfall patterns in the IMC region.

The impact of interaction SS with MJO on rainfall along Java Island to East Nusa Tenggara has been studied using daily rainfall observation data from 2140 weather observation stations, which are also equipped with GSMaP, OLR, and ERA5 data. There were 61 SS events found, 17 during active MJO and 44 in inactive MJO. Rainfall increases more in the western study area when the active MJO and more in the eastern study area when the inactive MJO is due to SS. SS reduces rainfall in the Jakarta region, regardless of the activity or inactivity of the MJO. SS causes extreme rainfall to only occur in a small part of certain areas, so it tends to significantly reduce the possibility of extreme rainfall. In the southern part of the IMC, SS predominates over MJO in supporting increased water vapor transport. Large-scale synoptic circulations like the monsoon, SS, and MJO can interact with local land-sea wind circulation to produce spatiotemporal variability in rainfall on Java Island. Rainfall mostly increases in the afternoon and decreases in the morning when SS occurs, whether there is MJO or not. Convective instability analysis indicates that SS increases precipitation, most likely by lowering CIN and raising VIMFC.

The characteristics of atmospheric variables over the southern coast of West Java in the presence of Australian Monsoon (AUM) and Mixed Rossby-Gravity (MRG) waves were investigated by conducting a dedicated radiosonde observation campaign from 26 to 29 September 2022 at Pameungpeuk Station (107.7°E, 7.6°S). The vertical profiles of pressure, temperature, relative humidity, and wind in the presence of the AUM and MRG waves were obtained and analyzed. The Global Navigation Satellite System - Radio Occultation (GNSS-RO) and the latest ECMWF climate reanalysis (ERA5) data were used to identify the MRG waves features using Hovmöller, space-time spectral, and wavelet analyses. Furthermore, spatial and time-series analyses were performed to study the wave propagation. The results showed the characteristics of atmospheric variables over Pameungpeuk Station in the presence of the AUM and MRG waves. We found that the presence of AUM significantly reduced the relative humidity, particularly in the region from 2 to 10 km altitude, and enhanced the average wind speed during the campaign compared to the wet season. Moreover, we found the reduction of low-level pressure, the enhancement of temperature and relative humidity in the mid-troposphere, and the weakening of the easterly and southerly winds during the dry phase approaching the wet phase of the MRG waves over the SOUTHERN COAst of West Java following the wave structure and propagation.

This study focuses on the evaluation of operational rainfall prediction models utilized by the Meteorological Climatological and Geophysical Agency (BMKG) for predicting the onset of seasons within distinct Zones of Season (ZOMs). The models employed include the NCEP Climate Forecast System Version 2—CFSv2, the European Centre for Medium-Range Weather Forecast’s fifth-generation—ECMWF SEAS5, the North American Multimodel Ensemble—MME, and a hybrid model—HyBMG, which incorporates Autoregressive Integrated Moving Average—ARIMA—and Wavelet ARIMA—WARIMA. These models encompass both dynamical (CFSv2, MME, ECMWF) and statistical (ARIMA, WARIMA) formulations. By employing the Taylor Skill Score (SS) for evaluation, the research identifies the best-performing model for each ZOM and determines the number of ZOMs dominated by each model across initial releases from January to December. SEAS5 (cor and raw), followed by CFSv2 (cor and raw), dominates predictions in almost the entire territory of Indonesia in some initial time, while in few location and specific initial, ARIMA and WARIMA also show best performance in predictions.

The impact of the southerly surge’s interaction with the MJO on rainfall in this study was investigated using daily rainfall data from 2140 weather-observation stations. The southern surge, which coincided with the MJO, enhanced rainfall in the western research region, with Yogyakarta seeing the greatest increase at 4.69 mm/day. Meanwhile, the southern surge that occurred without the MJO increased rainfall in the eastern region, with West Nusa Tenggara seeing the greatest rise at 3.09 mm/day. However, the southerly surge has the effect of lowering rainfall in Jakarta, reaching −2.21 mm/day when the MJO is active and −1.58 mm/day when the MJO is inactive. The southerly surge causes extreme rainfall to only occur in a small part of certain areas, so it tends to significantly reduce the possibility of extreme rainfall. In the southern part of the Indonesian maritime continent, the southerly surge predominates over the MJO, supporting increased water vapor transport. Rainfall mostly increases in the afternoon and decreases in the morning when the southerly surge occurs, whether there is the MJO or not. Convective instability analysis indicates that SS increases precipitation, most likely by raising vertically integrated moisture flux convergence, with a correlation coefficient value of 0.82.

Jakarta, a vast urban sprawl, undergoes rapid economic development accompanied by a vast population, land-use change expansion, and increased energy demand for transportation and industrial activities, resulting increase in PM2.5 concentrations. Efforts take place by governments to address urban air pollution by establishing air quality monitoring networks. However, the number of networks is still sparse and insufficient to record long-term series PM2.5 data. Thus, providing real-time continuous measurement of PM2.5 concentration remains challenged due to the inadequacy of ground-based measurements and expensive maintenance. Without long-term series PM2.5 data, assessing high-risk air pollution exposure is hard to quantify. This study is the first to estimate PM2.5 concentration using a combined MLR and RF in Jakarta megacity, Indonesia. The daily 24-h averaged ground-level PM2.5 concentrations are estimated by using meteorological parameters such as relative humidity (RH), visibility, 24 h average temperature (TAV), minimum temperature (TMIN), maximum temperature (TMAX), and dew point (DP) in 2016–2020. MLR and RF models were developed for dry, wet, and overall seasons. Using cross-validation (CV) with 75