Heap leach pad moisture and saturation mapping
Uneven moisture distribution across heap leach pads signals wasted lixiviant, reduced metal recovery, and structural risk. Sentinel-1 SAR backscatter, Landsat-9 SWIR reflectance, and ALOS-2 L-band penetration map saturation gradients that ground sensors miss.
Sensors
- Sentinel-1 C-band SAR (ESA): 10 m spatial resolution, 6-day repeat at mid-latitudes (12-day single-satellite). VV and VH polarisation backscatter responds to surface dielectric constant, which rises sharply with moisture content. Cloud-independent. Penetration limited to the top few centimetres of ore; cannot see moisture deeper in the pile.
- Landsat-9 OLI-2 SWIR bands (USGS/NASA): 30 m resolution, 16-day revisit. SWIR-1 (1.57 µm) and SWIR-2 (2.22 µm) reflectance drops measurably as surface moisture rises, because liquid water absorbs strongly at these wavelengths. Useless under cloud cover and sensitive only to the surface centimetre or so. Best for detecting broad saturation zones and ponding.
- ALOS-2 PALSAR-2 L-band SAR (JAXA): 3–10 m resolution depending on acquisition mode, 14-day revisit. L-band (1.27 GHz) penetrates 20–50 cm into dry coarse ore, giving a moisture signal from deeper within the pile than C-band. Particularly useful for coarser-crushed or run-of-mine heaps where surface readings are misleading.
- EMIT (Earth Surface Mineral Dust Source Investigation, NASA/JPL): 60 m resolution, non-systematic revisit from ISS orbit. Covers 380–2500 nm in 285 bands at ~7.4 nm sampling. Hydroxyl and water absorption features at 1400 nm and 1900 nm can distinguish surface moisture states and secondary mineral phases (gypsum, jarosite) that indicate reagent interaction zones. Archive growing since 2022.
What a heap leach pad is actually asking you to measure
A heap leach pad works by trickling a lixiviant, typically dilute sulphuric acid for copper or a cyanide solution for gold, through a crushed ore pile stacked on an impermeable liner. Recovery depends on the solution reaching every tonne of ore. Where it does not, metal stays in the ground. Where it pools, the pad risks piping failures or liner stress. The engineering problem is that a pad covering several square kilometres, stacked in multiple lifts, cannot be adequately monitored by point sensors alone. Drip emitters fail. Compaction creates preferential flow paths. Fines migration blocks drainage.
Remote sensing does not replace geotechnical instrumentation. What it does is give you a spatially continuous picture of surface and near-surface moisture states at intervals that match operational decision cycles, something a grid of tensiometers cannot do. The key measurable is the dielectric constant of the ore-solution mixture, which governs both microwave backscatter and near-infrared reflectance. Both change detectably as volumetric water content shifts.
C-band backscatter: fast, frequent, honest about its depth limit
Sentinel-1's C-band (5.405 GHz) is the workhorse here. Published studies in the peer-reviewed remote sensing literature consistently show that VV-polarisation backscatter increases by 3–8 dB as surface volumetric moisture rises from near-dry to saturated conditions in granular media, though the precise relationship depends on ore particle size and surface roughness. At 10 m resolution and a 6-day repeat cycle over most mining latitudes, Sentinel-1 provides a near-operational monitoring cadence at zero data cost.
The honest limit: C-band penetrates only a few centimetres into the ore surface. A pad that looks uniformly moist on the surface may be bone-dry 30 cm down, or vice versa. Interpretation must account for the fact that recent irrigation wets the surface without indicating full-profile saturation. Change detection between consecutive passes, rather than single-image absolute values, is generally more reliable for identifying anomalous dry or wet zones.
SWIR reflectance and what it adds to the radar picture
Landsat-9's SWIR bands exploit a different physical principle. Liquid water absorbs strongly at 1.57 µm and 2.22 µm, so a wet ore surface reflects less energy in those bands than a dry one. The Normalised Difference Moisture Index (NDMI), calculated from near-infrared and SWIR-1, is a published and widely validated approach for surface moisture mapping in soils and unconsolidated materials. Applied to heap leach pads, it produces a relative moisture map at 30 m resolution.
Cloud cover is the obvious constraint. In the Atacama or Nevada, that is rarely a problem. In tropical or high-altitude sites with frequent cloud, Landsat-9 revisit of 16 days can mean weeks of data gaps during wet seasons, precisely when over-saturation risk is highest. This is where Sentinel-1, being radar, earns its keep. The two datasets are complementary rather than redundant: SWIR confirms surface state, SAR monitors through cloud.
EMIT adds a third layer. Its full-spectrum coverage identifies not just moisture but the secondary mineral assemblages that form when acid lixiviant interacts with ore. Jarosite and gypsum have distinct absorption features in the 2.1–2.5 µm range. Their spatial distribution can reveal where reagent has actually reacted versus where solution has merely wetted the surface without chemical contact.
L-band for coarse ore: looking deeper into the pile
Run-of-mine heaps and coarsely crushed ore create a surface that is structurally different from finely crushed material. Large voids between boulders mean the surface backscatter signal is dominated by structure rather than moisture. C-band, with its short wavelength, scatters off the surface geometry. ALOS-2 PALSAR-2's L-band, at roughly 23 cm wavelength, penetrates further into the void space and interacts with moisture at greater depth, typically 20–50 cm in dry coarse material, less in wet or fine-grained ore.
JAXA makes ALOS-2 data available through its research and commercial channels. Tasking is not free, and the 14-day revisit is slower than Sentinel-1. For operational monitoring, ALOS-2 is best used periodically to calibrate the shallower C-band signal rather than as the primary data stream.
From backscatter to a moisture map: the analytic chain
Raw SAR backscatter is not a moisture map. The processing chain involves terrain correction using a digital elevation model of the pad (important because pad geometry changes as new lifts are added), radiometric calibration to sigma-naught, and then either empirical regression against any available in-situ moisture measurements or a semi-empirical dielectric model such as the Dobson or Topp equations, which relate dielectric constant to volumetric water content for granular media. Published values exist for common ore types but site-specific calibration improves accuracy substantially.
Change detection is often more operationally useful than absolute moisture estimation. A map showing which zones have dried relative to the previous acquisition, or which zones have become anomalously wet after a rainfall event or emitter failure, is actionable without requiring precise absolute calibration. Fusion of SAR change and SWIR reflectance change, co-registered and differenced, reduces false positives from surface roughness changes alone.
Satellize runs this processing chain on open-constellation data, including Sentinel-1 and Landsat-9, and can add commercial ALOS-2 tasking on client licence. The Tonga crop-estimation programme established the organisation's baseline for multi-sensor moisture-index fusion; the same spectral and backscatter physics apply to ore-pile surfaces, though the calibration targets differ.
Honest limits and what ground truth must still provide
No satellite sensor currently resolves individual drip emitters or sees moisture variation within a single ore lift. The spatial resolution floor of 10 m for Sentinel-1 means anomalies smaller than roughly 30 m across may be missed or blended with adjacent pixels. Sub-pad liner integrity cannot be inferred from surface moisture alone. A wet surface patch near the pad toe could indicate a liner leak, a blocked drain, or simply an over-irrigated zone; satellite data narrows the list but does not close it.
Temporal latency is real. Sentinel-1 data is typically available within 1–3 hours of acquisition through the Copernicus Data Space, but processing to a calibrated moisture-change layer adds time. For a fast-moving saturation event, the 6-day revisit may be too slow to catch the onset. Dense in-situ sensor networks and satellite monitoring are complements, not substitutes.
Finally, ore chemistry matters. High-iron ores and metallic mineralisation affect dielectric properties independently of moisture. Sites with strong mineralogical variability across the pad surface will show backscatter patterns that partly reflect mineralogy rather than moisture. EMIT's mineralogical mapping can help separate these effects, but the ambiguity should be acknowledged in any delivered product.
Typical figures
| Primary SAR resolution (Sentinel-1 IW mode) | 10 m range × 10 m azimuth (after multi-looking) |
| SAR revisit (Sentinel-1, mid-latitudes) | 6 days (two-satellite constellation); 12 days single satellite |
| SWIR resolution (Landsat-9 OLI-2) | 30 m |
| SWIR revisit (Landsat-9) | 16 days |
| L-band resolution (ALOS-2 PALSAR-2, stripmap) | 3–10 m depending on mode |
| L-band penetration depth (dry coarse ore) | 20–50 cm (published range for granular media) |
| EMIT spectral range | 380–2500 nm, ~285 bands at ~7.4 nm sampling; 60 m GSD |
| Minimum detectable moisture anomaly (SAR change detection) | Approximately 3–8 dB backscatter shift; corresponds to meaningful volumetric moisture change in granular media but is ore-type dependent |
| Sentinel-1 archive depth | From 2014 (Sentinel-1A launch) |
| Typical data latency (Sentinel-1 to processed layer) | 1–3 hours acquisition to raw data; additional processing time site-dependent |
Analytics Satellize can run
| SAR moisture-change map | Multi-temporal Sentinel-1 sigma-naught differencing with terrain correction; empirical or semi-empirical dielectric calibration | GeoTIFF layer per acquisition showing relative moisture change from previous pass; flagged anomaly zones in GIS-compatible format |
| SWIR surface saturation index | Normalised Difference Moisture Index (NDMI) from Landsat-9 OLI-2 NIR and SWIR-1 bands | 30 m raster per cloud-free acquisition; time-series chart of mean index by pad zone |
| Fused SAR and SWIR moisture composite | Co-registered multi-sensor fusion; change detection across both data streams to reduce false positives from surface roughness variation | Combined confidence-weighted moisture-state map; PDF report with anomaly summary |
| L-band deep-moisture profile (periodic) | ALOS-2 PALSAR-2 backscatter modelling for sub-surface moisture in coarse ore; comparison against C-band surface signal to estimate moisture gradient | Supplementary GIS layer; interpretation note on surface-versus-depth moisture state |
| Secondary mineral and reagent-interaction zone map | EMIT spectral unmixing targeting hydroxyl and sulphate absorption features (jarosite, gypsum) at 1400 nm, 1900 nm, 2100–2500 nm | Mineral-phase map at 60 m resolution; annotated zones where reagent interaction is spectrally confirmed |
| Dry-zone and over-saturation alert | Threshold-based anomaly detection on rolling SAR change stack; alert triggered when zone exceeds defined deviation from baseline | Email or API alert with coordinates, acquisition timestamp, and magnitude of anomaly |
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.