Landfill leachate pond extent and contamination mapping from multispectral data
Multispectral and SAR data can map leachate pond extent, track post-rainfall expansion, and flag subsurface migration at active and closed landfills, giving regulators a timestamped record that ground visits alone cannot provide.
Sensors
- Sentinel-2 MSI: 10 m resolution in visible and near-infrared bands, 20 m in shortwave-infrared (SWIR) bands 11 and 12 (1.6 µm and 2.2 µm). Five-day global revisit at the equator, shorter at higher latitudes. SWIR reflectance and red-edge bands are sensitive to dissolved organic matter, iron-rich effluent and turbidity anomalies in open water.
- Landsat 8/9 OLI-TIRS: 30 m multispectral resolution including SWIR bands at 1.6 µm and 2.2 µm, plus a 100 m thermal band (TIRS) that can detect thermal stratification in leachate ponds. Sixteen-day revisit per satellite, eight days with both Landsat 8 and 9 combined. Archive extends to 1972 for historical baseline construction.
- Sentinel-1 SAR (C-band): 5.4 GHz C-band synthetic aperture radar, 10 m ground range resolution in Interferometric Wide swath mode. Six-day repeat cycle enables coherence change detection regardless of cloud cover or darkness. Subsurface moisture changes from leachate migration alter soil dielectric properties, producing measurable coherence loss in interferometric pairs.
- PRISMA hyperspectral: Italian Space Agency hyperspectral imager covering 400–2500 nm in 239 bands at approximately 30 m spatial resolution. Enables full spectral curve matching for dissolved iron species, humic acids and other leachate-specific chromophores that broadband sensors can only approximate with indices.
What leachate looks like from 700 km up
Landfill leachate is chemically distinctive. High concentrations of dissolved organic carbon, ammonia, iron (commonly Fe²⁺ and Fe³⁺), heavy metals and humic substances give leachate ponds a spectral signature that differs measurably from clean water, agricultural runoff and natural ponds. Iron-rich leachate absorbs strongly in the blue and green wavelengths and reflects in a narrow band around 700–750 nm, producing a reddish-brown appearance that Sentinel-2's red-edge bands (Band 5 at 705 nm, Band 6 at 740 nm) are well-placed to capture.
The SWIR channels are equally informative. Clean water absorbs almost all incident radiation beyond 1.4 µm, so any non-trivial SWIR reflectance from a pond surface indicates suspended solids, algal biomass or surface films. Sentinel-2 Band 11 (1.6 µm) and Band 12 (2.2 µm) routinely flag turbid or organically loaded water bodies that appear unremarkable in true colour. This is not a novel idea: published work in Remote Sensing of Environment has used SWIR-based indices to discriminate contaminated from uncontaminated water at industrial sites for over a decade. The physics is well-established; applying it systematically to landfill inventories is the operational gap.
Tracking pond expansion after rainfall
A single Sentinel-2 image gives a snapshot. A time series gives a story. After significant rainfall events, leachate ponds can expand by tens of metres within 24 to 72 hours as percolating water mobilises stored contaminants and overwhelms liner capacity. At 10 m resolution with a five-day revisit, Sentinel-2 can capture this expansion cycle reliably in temperate and semi-arid climates, provided cloud cover does not obscure the acquisition.
Cloud cover is the honest constraint. In humid tropical environments, consecutive cloud-free Sentinel-2 acquisitions over a specific landfill can be separated by weeks rather than days. Landsat's thermal band adds a complementary dimension: leachate undergoing anaerobic decomposition generates heat, and TIRS can detect pond surfaces that are anomalously warm relative to surrounding soil or clean water, even through thin cirrus that defeats optical spectral analysis. Neither sensor eliminates the cloud problem; together they reduce it.
Temporal differencing of pond boundaries, derived from a water index such as the Modified Normalised Difference Water Index (MNDWI, using green and SWIR bands), produces a polygon layer showing net pond expansion or contraction between any two acquisition dates. Over a multi-year archive, this record becomes a compliance instrument: it shows whether a pond grew beyond its permitted boundary and precisely when.
Below the surface: SAR coherence and subsurface migration
The most dangerous leachate is often the leachate nobody can see. Where liner integrity fails, contaminated liquid migrates laterally through substrate before it ever reaches the surface as standing water. By the time a surface expression appears, the plume may already have reached a drainage channel or aquifer recharge zone.
Sentinel-1 interferometric SAR offers a partial window into this process. When the dielectric properties of near-surface soil change, such as when dry substrate becomes saturated with leachate, the phase relationship between two SAR acquisitions separated by six days breaks down. This coherence loss appears as a decorrelated patch in an interferogram. It is not a direct chemical measurement; it is a moisture and dielectric anomaly. Distinguishing leachate migration from ordinary rainfall infiltration requires contextual analysis: proximity to the waste boundary, spatial pattern relative to topographic gradient, and temporal correlation with known pond levels. The method flags candidates for ground investigation rather than delivering confirmed contamination maps.
Published studies using Sentinel-1 coherence over landfill sites have demonstrated detection of subsurface anomalies at depths of up to roughly one metre in dry sandy soils, with sensitivity declining sharply in clay-rich or already-moist substrates. C-band penetration is limited compared with L-band systems, so Sentinel-1 is most useful in sites with permeable, relatively dry cover soils.
Where PRISMA hyperspectral adds resolution in chemistry, not pixels
Broadband multispectral indices are proxies. They correlate with contamination load but cannot identify specific compounds. Italy's PRISMA satellite, operational since 2019, provides 239 contiguous spectral bands from 400 to 2500 nm at 30 m spatial resolution. This allows spectral unmixing and absorption-feature matching against known spectral libraries for humic and fulvic acids, iron hydroxides, and certain heavy-metal-bearing mineral phases that precipitate at pond margins.
The practical limitation is tasking. PRISMA is a pointed instrument with a 30 km swath and no continuous global coverage. Acquiring imagery over a specific landfill requires a tasking request, and the archive is far shallower than Sentinel-2's. PRISMA is best used for periodic deep characterisation of high-priority sites rather than routine temporal monitoring. Think of it as the laboratory-quality spectrometer that validates what the broadband sensors are flagging on a daily basis.
Honest limits of the spectral approach
Spectral methods detect what reaches the sensor. A pond covered by floating waste, dense algal mats, or a synthetic liner cap produces a spectrum dominated by those surface materials, not by the leachate beneath. Closed landfills with capped ponds may show no spectral anomaly at all, even when subsurface migration is active.
Spatial resolution sets a minimum detectable pond size. At 10 m, Sentinel-2 reliably delineates ponds larger than roughly 200 to 400 m² (two to four pixels), but smaller drainage sumps and seeps fall below the detection threshold. Landsat's 30 m pixel is more limiting still. Pond edges are spectrally mixed, so boundary precision degrades at sub-pixel scales. For regulatory purposes, pond area estimates derived from satellite data carry an uncertainty of at least one pixel width on each edge, which should be stated explicitly in any enforcement submission.
Finally, spectral indices indicate contamination probability, not concentration. Converting a SWIR anomaly into a milligrams-per-litre figure for dissolved organic carbon or iron requires in-situ calibration samples collected concurrent with satellite acquisition. Without that ground truth, the satellite product is a screening tool, not a substitute for water quality analysis.
From pixels to a regulatory record
The practical output of a leachate monitoring programme is a time-stamped, georeferenced record of pond extent and spectral anomaly intensity, updated at each cloud-free acquisition. Paired with the site's permitted boundary polygon, this record can automatically flag exceedances and generate alerts. SAR coherence layers add a second evidence stream for subsurface migration candidates. PRISMA acquisitions, scheduled around significant rainfall events or permit reviews, provide periodic chemical characterisation depth.
Satellize applies this multi-sensor workflow for government clients requiring defensible evidence for enforcement action or permit compliance review. The analytical approach draws on the same open-constellation data and published spectral methods that underpin the company's Tonga crop-estimation programme, adapted to the specific spectral and temporal requirements of contaminated-water detection. A regulator receives not raw imagery but a structured change log: dates, coordinates, pond-area estimates with uncertainty bounds, and anomaly scores referenced to a site-specific baseline.
Typical figures
| Spatial resolution (optical) | 10 m (Sentinel-2 visible/NIR), 20 m (Sentinel-2 SWIR), 30 m (Landsat OLI), 30 m (PRISMA hyperspectral) |
| Spatial resolution (SAR) | 10 m ground range (Sentinel-1 IW mode) |
| Revisit cadence | 5 days (Sentinel-2, equatorial); 8 days combined Landsat 8+9; 6 days (Sentinel-1) |
| Spectral bands used | Sentinel-2: B5 705 nm, B6 740 nm, B8A 865 nm, B11 1.6 µm, B12 2.2 µm; Landsat TIRS: 10.9 µm; PRISMA: 400–2500 nm (239 bands) |
| Minimum detectable pond area | Approximately 200–400 m² (Sentinel-2); approximately 1,800 m² (Landsat); smaller features require sub-pixel spectral unmixing with reduced confidence |
| Cloud cover limitation | Optical sensors fully blocked by cloud; SAR unaffected. Effective optical revisit in humid tropics may fall to weeks between usable acquisitions |
| SAR coherence sensitivity depth | Up to ~1 m in dry sandy soils; substantially less in clay or moist substrates (C-band, 5.4 GHz) |
| Archive depth | Sentinel-2: from 2015; Landsat: from 1972; Sentinel-1: from 2014; PRISMA: from 2019 (tasked only) |
| Delivery latency | Sentinel-1/2 Level-1 data typically available within 3–6 hours of acquisition via Copernicus Data Space; processed analytics within 24–48 hours depending on pipeline |
| Delivery formats | GeoTIFF polygon layers, GeoJSON change logs, PDF compliance reports with uncertainty bounds |
Analytics Satellize can run
| Pond extent time series | MNDWI thresholding on Sentinel-2 SWIR and green bands, with per-scene cloud masking using Scene Classification Layer | GeoJSON polygon archive with area estimates and ±1-pixel uncertainty bounds at each acquisition date |
| SWIR contamination anomaly map | Band ratio and spectral index analysis (B11/B8A, B12/B8A) referenced to a site-specific clean-water baseline derived from pre-operational imagery | Raster anomaly score layer (GeoTIFF) per acquisition, with exceedance flags relative to permitted boundary polygon |
| Iron-rich effluent probability map | Red-edge reflectance ratio (B5/B4) and SWIR composite, calibrated against published spectral signatures of Fe²⁺/Fe³⁺-bearing water bodies | Classified GeoTIFF with confidence tiers: probable, possible, background |
| Post-rainfall expansion alert | Automated differencing of consecutive cloud-free MNDWI pond-extent polygons; alert triggered when expansion exceeds a client-defined threshold in hectares | Email or API alert with before/after imagery thumbnails and area-change figure |
| SAR coherence anomaly layer | Sentinel-1 six-day interferometric coherence differencing over the landfill footprint; decorrelation patches outside the pond boundary flagged as subsurface migration candidates | GeoTIFF coherence-loss map with candidate migration pathway polygons, delivered as a GIS layer for field prioritisation |
| PRISMA deep-characterisation report | Spectral unmixing and absorption-feature matching against humic acid, iron hydroxide and clay mineral spectral libraries; tasked acquisition scheduled around permit review dates | PDF report with per-pixel dominant chromophore maps and annotated spectral curves for pond margin zones |
| Multi-year compliance record | Full archive stack analysis combining Sentinel-2 and Landsat, normalised to a common spatial grid, with boundary exceedance events tabulated chronologically | Structured PDF compliance dossier with timestamped evidence suitable for regulatory submission |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.