Thermal discharge plume mapping from power plant cooling water
Once-through cooling systems return water 3–10 °C above ambient, leaving a thermal signature that satellite infrared sensors can track across time. Multi-temporal stacking of Landsat TIRS and MODIS data distinguishes chronic discharge from seasonal variation and quantifies plume extent with honest resolution limits.
Sensors
- Landsat 8/9 TIRS: Two thermal infrared bands centred at 10.9 µm and 12.0 µm, 100 m native resolution (resampled to 30 m in products), 16-day revisit per satellite. The split-window pair enables atmospheric correction and sea surface temperature retrieval to roughly ±0.5–1 °C under clear skies. The 100 m resolution floor means plumes narrower than that are not resolvable; thin riverine discharge channels can be missed entirely.
- MODIS (Terra and Aqua): Thermal bands 31 and 32 at 1 km resolution, with Terra and Aqua together providing roughly four overpasses per day. The coarse spatial resolution makes MODIS unsuitable for resolving individual plant plumes in confined estuaries, but its daily revisit is valuable for tracking large coastal plumes over time and identifying seasonal discharge cycles.
- Sentinel-3 SLSTR: Sea and Land Surface Temperature Radiometer, dual-view design with bands at 10.85 µm and 12.0 µm at 1 km resolution for thermal channels, daily global coverage. Accuracy is comparable to MODIS for open-water plumes; the dual-view geometry aids atmospheric correction. Spatial resolution limits its use to larger or offshore plumes.
- ASTER (Terra): Five thermal infrared bands from 8.1 to 11.7 µm at 90 m resolution. ASTER offers better spectral discrimination than TIRS and has been used in published studies to characterise emissivity as well as temperature. Acquisition is on-demand rather than systematic, so archive depth for any given site is uneven and revisit cannot be relied upon for time-series work.
What a thermal plume actually tells you
Once-through cooling draws ambient water, passes it through condensers, and returns it at elevated temperature. For a large coal or nuclear station, that return flow can exceed 50 cubic metres per second at 5–10 °C above intake temperature. The plume spreads under the influence of local currents, tides, and wind, sometimes extending kilometres from the outfall before it dissipates to background.
The satellite measurement is not of the discharge pipe. It is of the surface expression of that plume, integrated over the thermal skin of the water. This matters for interpretation: a plume that dives below the surface thermocline becomes invisible to a radiometer. Conversely, a shallow, well-mixed estuary will show the full spatial footprint of the discharge. Knowing the local hydrography is not optional when reading these images.
The 100 m floor and what falls through it
Landsat TIRS is the workhorse for site-level plume mapping precisely because its 100 m native resolution sits between the coarseness of MODIS and the expense of airborne surveys. For a large coastal plant discharging into open water, 100 m is adequate to delineate the warm core and estimate plume area. For a plant on a river 80 m wide, it is not. The plume may be detectable as an anomalous pixel, but its geometry will be lost.
Atmospheric correction is the other honest caveat. TIRS split-window retrieval degrades under high water vapour or thin cirrus, and errors of 1–2 °C are possible without coincident atmospheric profiles. For regulatory purposes, satellite-derived temperatures should be treated as indicative rather than metrological unless validated against in-situ buoy or sensor data. That validation step is rarely optional if the output is intended to support an enforcement action.
Multi-temporal stacking: separating chronic from seasonal
A single Landsat scene tells you that a plume existed on one clear-sky day. A stack of fifty scenes over five years tells you something operationally different: whether the discharge is continuous, whether it intensifies in summer when ambient water is warmer and regulatory margins tighten, and whether it ceased during known outage periods. That temporal pattern is the analytic product with real enforcement value.
The method is straightforward in principle. Cloud-masked thermal retrievals are co-registered, converted to sea or land surface temperature anomaly relative to a local background polygon, and aggregated into percentile maps. The 90th-percentile anomaly map shows the spatial footprint of chronic maximum discharge; the median map shows typical conditions. Gaps caused by cloud are not trivial at mid-latitudes, where Landsat's 16-day revisit can yield fewer than ten usable scenes per year at persistently cloudy sites. MODIS can fill temporal gaps at the cost of spatial detail.
Seasonal decomposition adds another layer. Fitting a harmonic model to the time series separates the annual temperature cycle of the receiving water body from the superimposed discharge signal. Residuals that persist year-round, regardless of season, are the strongest indicator of chronic thermal loading rather than natural variation.
Regulatory context and the limits of satellite evidence
Many jurisdictions set thermal discharge limits as a maximum permitted temperature rise above ambient, measured at a defined mixing zone boundary. Satellite data can map whether a thermal anomaly extends beyond that boundary, and how often. What it cannot do is provide the continuous, calibrated, legally defensible measurement that a fixed in-situ sensor delivers.
The practical use case for regulators is screening and prioritisation, not prosecution. A persistent anomaly in multi-year Landsat composites justifies deploying in-situ monitoring or commissioning an inspection. It can also support compliance audits by cross-checking operator-reported discharge volumes against the observed plume geometry: a plant claiming reduced output should show a smaller or cooler plume. Discrepancies between reported and observed thermal signatures are the kind of flag that satellite analysis is well placed to raise.
Ecological consequence and the secondary signal
Thermal plumes alter dissolved oxygen, accelerate algal growth, and can shift species distributions in the receiving water body. Satellite data can track some of these secondary effects. Sentinel-2 and Landsat OLI multispectral bands can detect chlorophyll-a anomalies and turbidity changes in the plume footprint, providing a proxy for biological impact that complements the thermal measurement. The two signals together, thermal extent from TIRS and chlorophyll anomaly from OLI, give a more complete picture of ecological loading than either alone.
Satellize runs this kind of paired thermal-optical analysis as part of its emissions and pollution monitoring work. The approach is the same one applied in the Tonga crop-estimation programme: open-constellation data, processed systematically, with outputs calibrated to what the physics actually supports rather than what a client might prefer to see.
Archive depth and what it makes possible
Landsat's continuous archive runs from 1972 for the MSS era and from 1982 for thermal bands (Landsat 4 and 5 TM). Landsat 8 TIRS data, with its improved noise-equivalent temperature difference of approximately 0.4 K, is available from 2013. That decade-plus of consistent TIRS data is long enough to assess whether a plant's thermal footprint has grown as generating capacity expanded, or shrunk as cooling tower retrofits were installed.
MODIS adds daily coverage back to 2000. The combination of MODIS for temporal density and Landsat for spatial precision is a published and well-tested approach in the remote sensing literature. Neither dataset requires a commercial licence, which means the archive is accessible for retrospective enforcement analysis without data acquisition costs.
Typical figures
| Spatial resolution (Landsat TIRS) | 100 m native thermal, resampled to 30 m in standard products |
| Spatial resolution (MODIS thermal) | 1 km (bands 31/32) |
| Spatial resolution (Sentinel-3 SLSTR thermal) | 1 km |
| Spatial resolution (ASTER TIR) | 90 m |
| Revisit (Landsat 8 + 9 combined) | 8 days at equator; fewer usable scenes per year due to cloud |
| Revisit (MODIS Terra + Aqua combined) | ~4 overpasses per day globally |
| Temperature sensitivity (Landsat 8/9 TIRS) | Noise-equivalent temperature difference ~0.4 K; retrieval accuracy ±0.5–1 °C under clear skies with atmospheric correction |
| Minimum detectable plume | Plumes wider than ~200–300 m reliably resolved by Landsat TIRS; narrower riverine discharges may be detectable but not geometrically resolved |
| Archive depth | Landsat TIRS: 2013–present; MODIS: 2000–present; Landsat TM thermal: 1982–2013 |
| Data access | Open, no licence fee (USGS EarthExplorer for Landsat; NASA LAADS DAAC for MODIS; Copernicus Data Space for Sentinel-3) |
Analytics Satellize can run
| Plume extent map (per scene) | Thermal anomaly thresholding: background SST/LST estimated from upstream or offshore reference polygon, plume defined as pixels exceeding background by a user-set delta (typically 1–2 °C) | GeoTIFF or vector polygon layer per Landsat scene, attributed with date, plume area (km²), and peak anomaly (°C) |
| Multi-year thermal anomaly composite | Cloud-masked TIRS time-series stacked and summarised as median, 90th-percentile, and frequency-of-exceedance rasters | Set of three GeoTIFF composites covering user-defined period, with accompanying summary statistics table |
| Seasonal discharge pattern report | Harmonic decomposition of monthly median anomaly time series to separate annual thermal cycle from persistent discharge signal | PDF report with time-series charts, seasonal amplitude maps, and narrative interpretation of chronic versus seasonal loading |
| Regulatory mixing-zone exceedance flag | Spatial intersection of plume polygon with operator-declared mixing-zone boundary; exceedance recorded when warm anomaly extends beyond boundary | Tabular exceedance log (date, exceedance area, peak temperature at boundary) delivered as CSV or GIS layer for each scene in the archive |
| Paired thermal and chlorophyll-a anomaly layer | Landsat OLI band-ratio chlorophyll-a retrieval (bands 3 and 4) co-registered with TIRS thermal anomaly for the same acquisition | Dual-layer GeoTIFF with thermal and optical anomaly co-displayed, suitable for ecological impact screening |
| Discharge-versus-reported-output cross-check | Plume area and peak anomaly compared against operator-reported generation output or fuel consumption data; statistical correlation and residual analysis | Quarterly summary report flagging scenes where observed thermal signature is inconsistent with reported plant status |
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.