Publications·November 30, 2024
This UNESCAP working paper maps how Indonesia can harness satellite and ground-based remote sensing to strengthen air-quality management. It synthesizes Indonesia’s pollution burden, the institutional actors (BRIN, BMKG, KLHK, BNPB), what EO (Earth Observation) assets are in use, the status of Pandora spectrometers and the GEMS geostationary mission, and offers a practical set of national and municipal recommendations to move from pilot projects to operational, decision-ready products. Headline framing: Indonesia’s mean PM₂.₅ exposure ≈ 19 µg/m³ (≈3.8× the WHO guideline) with ~106,710 deaths in 2019 attributable to fine particles; Jakarta is repeatedly listed among the most polluted capitals. Recent trends include rising SO₂ and NO₂ in Jakarta (2010–2022) and a 33% jump in coal consumption in 2022—amplifying the case for EO-assisted monitoring and policy evaluation.
Institutions, policy scaffolding, and the monitoring baseline
Who does what. Since a 2021 restructuring, BRIN consolidated national space and atmospheric research (ORPA/ORKM), while BMKG runs forecasting and fire-weather early warning (e.g., SPARTAN). The Ministry of Environment and Forestry (KLHK) operates the regulatory network and communicates the ISPU index; around 38 continuous stations were operating in 2020. BNPB manages disaster response (fires/volcanoes).
Policy architecture. Core instruments include Law 32/2009 on Environmental Protection and Management; the Peatland Restoration Agency (BRG, 2015); and Euro IV fuel/vehicle standards (50 ppm sulphur) moving toward higher stringency. UNEP’s policy scorecard notes Indonesia has actions across all nine assessed sectors (standards, monitoring, waste burning controls, cleaner fuels, etc.). The paper underlines growing public demand for clean air and the need to tie EO outputs directly to enforcement and planning.
What EO is already doing in Indonesia
National satellites. Indonesia has flown multiple LAPAN microsatellites (A1/A3/A4/A5), including SAR capability, providing a domestic EO base for environmental monitoring. Ozonesonde launches (SHADOZ) resumed in 2020 at Watukosek, adding vertical ozone profiles.
Fire & volcano services. KLHK’s SiPongi and BMKG’s SPARTAN provide early warning using MODIS/VIIRS hot spots and meteorology; ToFEWSI adds fire-occurrence forecasts. Magma Indonesia issues volcano advisories, with the report encouraging tighter satellite integration for ash/SO₂ tracking.
Urban & transboundary use cases. A 2020 assessment combining OMI/TROPOMI, ground data and transport modelling traced significant SO₂/NO₂/PM₂.₅ transport into Jakarta from Banten and West Java—evidence that policy must address airshed-scale sources. BRIN’s SRIKANDI platform publishes daily air-quality fields and WRF-Chem forecasts for PM₂.₅, PM₁₀ and gases, already blending satellite inputs.
Research ecosystem. An ESCAP meta-scan tallies 26 Indonesian academic studies using EO (2012–2023) across burning, AOD, PM₂.₅/PM₁₀, NO₂, CO, SO₂ with MODIS, VIIRS, TROPOMI, OMI, MISR and AERONET. Gaps remain: routine PM₂.₅ mapping, SLCP/trace-gas mapping (NO₂, SO₂, HCHO, O₃), and long-term policy evaluation outside major cities.
Pandora and GEMS — what’s installed and what we’re learning
Pandora (PGN/Pandora Asia Network). Three operational Indonesian sites—Bandung, Agam, Pontianak—provide high-frequency columns of NO₂, O₃, SO₂, HCHO and AOD for science and for GEMS validation. Data quality (percent of “high-quality flag” records) varies by site and pollutant—NO₂ often >40–70%, O₃ sometimes <10–15%—due to sun-tracker alignment, power stability and head-sensor connectivity issues. The report recommends systematic QA/QC, spares, and co-locations with reference PM₂.₅/NO₂/SO₂/O₃ monitors to maximise value.
GEMS (geostationary). Case studies show promise and caveats:
Volcanic SO₂ from Mt Dukono (24–29 Apr 2022) is coherently captured by GEMS and consistent with HYSPLIT trajectories/ERA5 winds.
Fire-season aerosols (Sumatra/Kalimantan): GEMS AOD patterns align broadly with CAMS model fields, but AOD variability against AERONET indicates the need for continued validation.
Early correlations with surface PM₂.₅ vary by metric (e.g., AOD–PM₂.₅ r≈0.67; UVAI–PM₂.₅ r≈0.39), reinforcing the role of bias-correction and ML fusion. The paper notes a recent deep-learning bias-correction that raises GEMS NO₂ agreement with Pandora/TROPOMI (r from 0.68→0.88), illustrating the trajectory toward decision-grade geostationary products.
Bottom line: Pandora + GEMS already enable episode forensics (fires, volcanoes) and high-cadence urban NO₂/PM diagnostics, but Indonesia must close uptime, calibration and integration gaps to standardise operational use.
What remote sensing can close — priority knowledge gaps
Unmonitored areas: Produce monthly/annual PM₂.₅ maps (AOD + CTM + statistical bias-correction) to identify hotspots beyond the sparse ground network.
SLCPs & precursors: Operationalise NO₂, SO₂, HCHO, O₃ mapping for source diagnostics and chemistry insights.
Long-term trend & policy evaluation: Use EO + ground data to track impacts of transport/industry fuel standards, coal build-out/retirements, and anti-burning measures.
Transboundary & airsheds: Combine GEMS/TROPOMI with HYSPLIT/WRF to define airsheds and quantify cross-border episodes that demand ASEAN-level action.
Equity mapping: Merge high-res PM/trace-gas fields with population/vulnerability layers to target interventions.
Fire emissions: Shift from hotspot counts to burn-area + FRP for peat/crop-residue emissions to evaluate prevention programmes.
Recommendations — what to do next (national & municipal)
National level
Make EO part of the policy “plumbing”. Regularly update SRIKANDI and institutional websites with open, documented satellite products, APIs, and pre/post-processing guides; publish QA/QC status and version notes.
Invest in people. Fund recurring training (ingest, bias-correction/ML, WRF/CTM, nowcasting/forecasting), university partnerships, and a help-desk; sponsor competitive research that solves operational bottlenecks.
Grow the PGN footprint. Keep Indonesia active in the Pandora Asia Network; expand sites and standardise cal/val with AERONET and regulatory monitors; report uptime/flags transparently.
Integrate for decisions. Fuse Pandora–GEMS–AERONET–ground AQMN in a national EO–AQ platform that outputs near-real-time maps, forecasts and episode analytics with uncertainty metadata.
Link to mitigation portfolios. Use EO to operationalise parts of the “25 Clean Air Measures” (e-mobility, coal/industrial controls, burn prevention, clean energy), linking indicators to programme monitoring.
Municipal level
Densify the ground network (PM₂.₅/NO₂/SO₂/O₃) and add Pandora at strategic urban and downwind sites; co-locate with reference instruments.
Stand up “episode rooms”. Municipal teams using GEMS/Pandora + ground data for alerting, compliance targeting, plume tracking (fires/industrial events), and post-event forensics.
Evaluate & publish. Routinely issue policy-effect reviews (e.g., low-sulphur fuel, traffic measures, waste-burning bans) using EO indicators to build public accountability.
Why this matters regionally
Indonesia’s emissions (peat/forest fires, coal and industry) shape air quality across Southeast Asia. GEMS’ geostationary view and Pandora’s ground truth are ideal for real-time transport tracking and airshed governance, strengthening ASEAN commitments on transboundary haze and enabling evidence-based cooperation.
Key words
Remote sensing; GEMS; Pandora (PGN); PM₂.₅ mapping; NO₂/SO₂/HCHO/O₃; peat-fire emissions; AOD; HYSPLIT/WRF/WRF-Chem; SRIKANDI; airshed & transboundary transport; QA/QC; Indonesia; operational decision support.