Groundwater salinity and rising damp monitoring at mudbrick heritage sites

Rising groundwater carries dissolved salts into mudbrick fabric, where crystallisation cycles fracture walls from within. Sentinel-1 InSAR and Sentinel-2 SWIR bands can detect the precursors before a structure fails.

Sensors

Sentinel-1 SAR (C-band, ESA): Provides interferometric pairs for surface displacement mapping at approximately 5 x 20 m resolution in IW mode, 6-day repeat at mid-latitudes when both satellites are active. Differential InSAR (DInSAR) and time-series methods (SBAS, PS-InSAR) detect millimetre-scale subsidence driven by differential wetting and salt-induced swelling or collapse.
Sentinel-2 MSI (ESA): 10 m visible and 20 m SWIR bands (1610 nm and 2190 nm) allow detection of surface salt efflorescence through elevated SWIR reflectance and suppressed vegetation indices. 5-day revisit with two satellites. Cloud cover over arid Mesopotamian and Egyptian sites is low enough for monthly composites to be reliable through most of the year.
ALOS-2 PALSAR-2 (JAXA): L-band SAR (1.27 GHz) penetrates a few centimetres into dry soil, improving coherence on bare earthen mounds where C-band decorrelates rapidly. Useful for longer-baseline interferograms and for sites where Sentinel-1 coherence is marginal. Spatial resolution in stripmap mode is approximately 3 m.
Landsat 8/9 OLI (USGS/NASA): 30 m SWIR bands (1.61 µm and 2.20 µm) extend the salt-efflorescence spectral record back to 1984 via the full Landsat archive, enabling multi-decade trend analysis of surface salt expression at sites where no in-situ baseline exists.

What rising groundwater actually does to a mudbrick wall

Mudbrick is a capillary system. When the water table beneath a tell rises, moisture wicks upward through the brick matrix, carrying dissolved chlorides, sulphates and carbonates. As the moisture front reaches a drier zone and evaporates, salts crystallise in the pore spaces. The crystallisation pressure, which for sodium sulphate (mirabilite to thenardite transitions) can exceed 10 MPa, is sufficient to disaggregate the mud matrix grain by grain. The process is cyclical: wetting and drying seasons drive repeated crystal growth and dissolution, and each cycle removes material.

This is not a slow aesthetic problem. Structural walls at major Mesopotamian sites have lost measurable volume within single decades. The proximate cause is almost always a combination of irrigation intensification and leaking urban water infrastructure raising the local water table, often by several metres above pre-agricultural baselines. Satellite monitoring cannot stop that process, but it can detect where it is accelerating before a wall face detaches or a mound flank slumps.

What a SWIR band gives away on a salt-crusted surface

Sodium chloride, gypsum and thenardite have characteristic reflectance signatures in the short-wave infrared. At 1610 nm and 2190 nm, salt efflorescence on a mudbrick surface produces distinctly elevated reflectance compared to unaffected mud, which absorbs strongly in those bands when moist and shows a flatter, lower response when dry. Sentinel-2 bands 11 and 12 resolve this contrast at 20 m. At the scale of a major tell, that is sufficient to map the spatial extent of active efflorescence zones and track their seasonal expansion.

The honest limit is spatial: a thin salt film on a 2 m wall face is below detection. What the sensor resolves is the aggregate expression across open mound surfaces, collapsed rubble spreads and the flat aprons around a tell base, where salts often concentrate as capillary moisture spreads laterally. Spectral mixture analysis helps separate salt signal from dry clay and carbonate bedrock, but field calibration at least once per site is advisable. Cloud is rarely the problem in southern Iraq or Upper Egypt; dust haze in spring can degrade optical quality for several consecutive acquisitions.

InSAR displacement maps as a structural early-warning signal

Differential settlement is the physical consequence of uneven moisture loading in a mudbrick mound. Where groundwater reaches the base of a wall at one end but not the other, the wetter section swells slightly, then subsides as salt crystallisation destroys the matrix. Sentinel-1 time-series InSAR, processed with SBAS or persistent-scatterer methods, can resolve line-of-sight displacement rates of roughly 1 to 2 mm per year under good coherence conditions. On bare earthen mounds, coherence degrades faster than on masonry, so six-month to annual displacement maps are more reliable than monthly ones.

The practical output is a displacement velocity map overlaid on the site plan. Zones showing consistent downward movement, particularly if spatially correlated with known wall alignments or with SWIR-detected salt zones, are the priority targets for conservation intervention. ALOS-2 PALSAR-2 L-band data improves coherence on loose earthen surfaces and is worth adding where Sentinel-1 coherence drops below usable thresholds, typically in areas of active surface erosion or recent rainfall disturbance.

One ambiguity worth stating plainly: InSAR measures displacement in the radar line of sight, not vertical settlement directly. Decomposing ascending and descending pass data into vertical and horizontal components requires both geometries, and Sentinel-1 provides both over most of the Middle East and North Africa. Even so, the displacement signal at a mudbrick tell is rarely a clean vertical subsidence; lateral spreading of saturated base material produces a mixed signal that requires careful modelling.

Combining the two signals: where the methods agree is where the risk is real

Neither InSAR displacement nor SWIR salt mapping is individually conclusive. A displacement anomaly might reflect seasonal soil moisture change in surrounding agricultural land rather than structural failure in the tell itself. A SWIR salt anomaly might be a natural sabkha deposit unrelated to groundwater rise. The diagnostic power comes from spatial and temporal co-occurrence: zones that show both accelerating subsidence and expanding salt efflorescence, season after season, are very likely experiencing active salt-crystallisation damage.

Published work on Egyptian sites, including studies using Sentinel-1 and Sentinel-2 data over Luxor and the Nile Delta, has demonstrated this combined approach at operational scale. The Mesopotamian context adds complexity because many tells are surrounded by irrigated fields that create strong moisture gradients at the mound boundary, producing edge-effect displacement signals. Masking the immediate agricultural perimeter and focusing the InSAR analysis on the mound interior reduces false positives substantially.

Honest limits and what they mean for a conservation programme

Satellite monitoring at this scale is a triage tool, not a substitute for ground investigation. A 20 m SWIR pixel cannot tell you which specific wall face is failing, and a 5 x 20 m InSAR pixel cannot resolve the displacement of a single column or arch. What the satellite record provides is a site-wide, time-continuous picture that no ground team can match in coverage or revisit frequency. The appropriate use is to prioritise where ground teams deploy, not to replace them.

Archive depth is a genuine asset. Sentinel-1 data runs from 2014, Sentinel-2 from 2015, and Landsat from 1984. For a site where conservation records are sparse, a multi-decade spectral time series of surface salt expression can reconstruct the history of groundwater encroachment with reasonable confidence. That historical baseline is often the most persuasive evidence when arguing for infrastructure intervention, such as drainage works or irrigation reform, with national water authorities.

Satellize runs combined InSAR and multispectral salt-index analytics on open Sentinel archives for heritage clients. The workflow is the same class of analysis applied in the Tonga crop-estimation programme: systematic index derivation, change detection and structured reporting, adapted here for conservation rather than agricultural decision-making. A site assessment typically begins with an archive review before any new tasking is considered.

Typical figures

InSAR spatial resolution (Sentinel-1 IW)	~5 x 20 m ground range; resampled to 10–20 m for displacement products
InSAR minimum detectable displacement rate	~1–2 mm/year (time-series methods, good coherence); single-pair precision ~5–10 mm
Sentinel-1 revisit	6 days at mid-latitudes with both satellites; 12 days with one
Sentinel-2 SWIR resolution	20 m (bands 11 and 12 at 1610 nm and 2190 nm)
Sentinel-2 revisit	5 days (two-satellite constellation)
ALOS-2 PALSAR-2 resolution (stripmap)	~3 m; L-band (1.27 GHz) improves coherence on loose earthen surfaces
Landsat SWIR archive depth	1984 to present (Landsat 5 TM onwards); 30 m resolution
Minimum mappable salt efflorescence area	Approximately 400 m² aggregate at 20 m SWIR resolution; sub-pixel mixing models extend sensitivity
Typical cloud cover limitation	Low in Mesopotamia and Upper Egypt; dust haze can affect 2–4 consecutive Sentinel-2 acquisitions in spring
Delivery formats	GeoTIFF displacement velocity maps, GeoPackage salt-index change layers, PDF site assessment report

Analytics Satellize can run

Salt efflorescence extent map	Sentinel-2 SWIR band ratio (B11/B8A or B12/B8A) with spectral mixture analysis to separate salt, dry clay and carbonate	Seasonal GeoTIFF layer showing salt-affected area fraction per 20 m pixel, with change map versus baseline year
Surface displacement velocity map	Sentinel-1 SBAS or PS-InSAR time-series processing; ascending and descending pass decomposition for vertical/horizontal separation	Annual displacement velocity GeoTIFF with uncertainty estimate; flagged anomaly zones as GeoPackage polygons
Combined risk-priority layer	Spatial intersection of persistent displacement anomalies with salt-index exceedance zones; ranked by co-occurrence frequency across seasons	Site-wide risk-priority GIS layer with zone classifications (high/moderate/low) for conservation team field deployment
Multi-decade salt-expression trend	Landsat 5/7/8/9 SWIR time series from 1984; annual composites with cloud/dust masking; Mann-Kendall trend test per pixel	PDF report with pixel-level trend charts and a map of statistically significant salt-increase areas over the archive period
Groundwater moisture-front proxy index	Sentinel-2 SWIR-based soil moisture index (e.g. SWIR transform of surface reflectance) applied to mound apron and surrounding fields to track lateral moisture gradient	Monthly raster stack and summary statistics for the 500 m buffer zone around the site perimeter
L-band coherence supplement	ALOS-2 PALSAR-2 interferometric coherence map to identify zones of Sentinel-1 coherence loss and extend displacement monitoring coverage	Coherence comparison GeoTIFF and flagged coverage-gap report, with recommendation on acquisition scheduling

Sources

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.