Soil moisture mapping for cable trench reinstatement and settlement monitoring
Sentinel-1 C-band radar and NASA SMAP L-band radiometry track soil moisture dynamics along cable trenches, exposing differential settlement risk before conduit joints fail. Surface-layer retrieval is well-established; estimating moisture at burial depth requires modelling and carries real uncertainty.
Sensors
- Sentinel-1 (ESA, C-band SAR, 5.405 GHz): Interferometric Wide Swath mode delivers 10 m ground range resolution with 6-day repeat at mid-latitudes (12-day for a single satellite). VV and VH backscatter coefficients respond to surface volumetric soil moisture in the top 2-5 cm. Cloud-independent, day-night capable. Saturates over very wet or densely vegetated surfaces.
- SMAP (NASA, L-band radiometer, 1.41 GHz): 36 km effective resolution brightness temperature retrievals, 2-3 day global revisit. L-band penetrates light vegetation and retrieves moisture from roughly the top 5 cm of bare or lightly vegetated soil. Volumetric soil moisture accuracy is approximately ±0.04 m³/m³ unbiased RMSE over open terrain. Spatial resolution is far too coarse for individual trench routes; SMAP is most useful for regional seasonal context and calibrating finer-resolution SAR retrievals.
- ALOS-2 PALSAR-2 (JAXA, L-band SAR, 1.27 GHz): L-band SAR at 3-10 m resolution in spotlight and stripmap modes. Longer wavelength penetrates vegetation canopy and probes soil to roughly 10-30 cm depth depending on moisture content, giving better sensitivity to sub-surface conditions than C-band. Revisit is 14 days. Commercial tasking required; not freely available like Sentinel-1.
- Sentinel-2 (ESA, multispectral): 10-20 m multispectral imagery at 5-day revisit. Short-wave infrared bands (Band 11 at 1610 nm, Band 12 at 2190 nm) are sensitive to surface soil moisture through reflectance changes. Optical only: cloud cover frequently interrupts acquisitions, which is a significant limitation in temperate climates during the wet seasons when settlement risk is highest.
Why trenching changes everything about local drainage
Cutting a trench through undisturbed soil destroys the natural capillary structure that took decades to consolidate. Backfill, however well compacted, is looser and more permeable than the surrounding matrix. Water infiltrates preferentially along the trench corridor. In clay-rich soils, that preferential wetting produces localised swelling during wet periods and shrinkage cracks during dry ones. The result is cyclical vertical movement that is spatially abrupt: a few centimetres of differential settlement between the disturbed strip and the undisturbed shoulder is enough to stress conduit joints and splice housings.
The problem is worst where the route crosses soil-type boundaries, runs through areas with a seasonally high water table, or passes through made ground with variable compaction history. These are not rare edge cases. Most long-distance cable routes encounter all three.
What the radar backscatter actually measures, and where it stops
Sentinel-1 C-band backscatter responds to the dielectric constant of the soil surface, which is dominated by volumetric water content. Empirical and semi-empirical models, including the widely used Oh and Dubois models, can convert backscatter to volumetric moisture estimates over bare or sparsely vegetated surfaces. The penetration depth at C-band is roughly 2-5 cm in moist mineral soils, shallower still when the surface is wet. That is well above typical cable burial depths of 0.6-1.2 m for direct-buried fibre or 0.45-0.9 m for ducted installations.
ALOS-2 PALSAR-2's L-band reaches deeper, perhaps 10-30 cm, but cable burial depth is still well beyond direct retrieval. To estimate moisture at depth, practitioners apply pedotransfer functions that relate surface moisture observations to profile moisture using soil texture, bulk density and antecedent precipitation as inputs. The uncertainty in that extrapolation is substantial: errors of 0.06-0.10 m³/m³ or more are realistic in heterogeneous soils. Any settlement-risk product derived from depth-moisture estimates should be presented as a relative risk index, not a precise volumetric figure.
Vegetation complicates matters further. A grass cover over a reinstated trench attenuates and scatters the radar signal before it reaches the soil, reducing sensitivity. VH/VV polarisation ratios help separate volume scattering from surface scattering, but the correction is imperfect. Bare or freshly seeded reinstatements are the most tractable cases.
Shrink-swell mechanics and the conduit joint failure pathway
Clay minerals, particularly smectites and illites, absorb water into their crystal lattice and expand. The linear shrinkage of high-plasticity clays can exceed 15% between field capacity and air-dry states. Across a trench width of 0.5 m in a clay soil with a plasticity index above 40, a single wet-dry cycle can produce several millimetres of differential vertical movement between the trench fill and the undisturbed shoulder.
Conduit joints are designed with limited angular deflection tolerance. Repeated cycling concentrates stress at the joint, particularly where the trench crosses a soil boundary and one side of the joint is in expansive clay while the other is in gravel or sand. Monitoring soil moisture dynamics over multiple seasonal cycles identifies which sections of a route are experiencing the largest moisture swings, and therefore the highest cumulative mechanical stress, before any surface expression of settlement is visible to a site inspector.
Building a route-level risk index from multi-sensor time series
The practical workflow starts with a Sentinel-1 time series covering at least one full annual cycle, ideally two or three years to capture inter-annual variability. Backscatter values are extracted along a buffer of 20-50 m either side of the trench centreline. A change-detection algorithm flags segments where the moisture signal in the trench corridor diverges significantly from the surrounding undisturbed soil, which is the signature of anomalous infiltration into the backfill.
SMAP provides the seasonal envelope: it establishes when regional soil moisture is rising or falling, giving context for interpreting the Sentinel-1 anomalies. Sentinel-2 SWIR imagery adds a cross-check during cloud-free periods and helps identify surface ponding or runoff patterns that are not visible in radar data alone.
The output is a segmented risk index along the route, expressed as a relative score based on moisture variability amplitude and the number of wet-dry cycles detected per year. Segments scoring above a threshold are flagged for priority ground inspection. This does not replace inspection; it directs it. A 200 km cable route cannot be walked every quarter. Satellite-derived prioritisation reduces the inspection burden to the sections where the physics suggests the greatest cumulative stress.
Honest limits: what this method cannot do
No current spaceborne sensor directly measures soil moisture at cable burial depth. That is a physical constraint, not a data-quality issue. The pedotransfer approach is a modelled inference, and its accuracy depends heavily on having reliable soil texture data for the route corridor. Where soil surveys are coarse or absent, the depth-moisture estimate degrades significantly.
Dense urban reinstatements are largely inaccessible to this method. Radar backscatter over tarmac and concrete is dominated by surface roughness and double-bounce returns from structures, not soil dielectric properties. The approach is most reliable for rural and semi-rural routes across agricultural or grassland terrain with minimal built cover over the trench.
Cloud cover interrupts Sentinel-2 optical acquisitions, sometimes for weeks at a time in temperate climates. The SAR sensors are unaffected by cloud, but the optical cross-check is intermittent. Finally, a 10 m Sentinel-1 pixel is large relative to a trench width of 0.3-0.6 m; the backscatter signal is a spatial average that includes undisturbed soil on either side. Narrower trenches in heterogeneous terrain are harder to isolate.
From data to a maintenance decision
Satellize runs this workflow as a recurring analytics layer, delivering a GIS-compatible risk-segment file after each Sentinel-1 acquisition cycle. The Tonga crop-estimation programme established the operational pattern: open-constellation time series, processed to a client-ready output, without requiring the client to manage satellite access or processing infrastructure.
For a telecoms operator, the practical next step is to supply the trench centreline geometry and any available soil survey data for the route, then commission a baseline moisture-variability assessment covering the first post-reinstatement wet season. That baseline determines which segments warrant quarterly monitoring and which can be reviewed annually. Inspection budgets go where the physics points.
Typical figures
| Sentinel-1 spatial resolution (IW mode) | 10 m ground range x 10 m azimuth (multi-looked product) |
| Sentinel-1 revisit interval | 6 days at mid-latitudes (two-satellite constellation); 12 days single satellite |
| SMAP radiometer resolution | ~36 km effective (Level 3 composite); unsuitable for individual trench-scale mapping |
| SMAP revisit interval | 2-3 days global |
| ALOS-2 PALSAR-2 resolution | 3-10 m (mode-dependent); 14-day revisit; commercial tasking required |
| Soil moisture retrieval depth (C-band SAR) | Approximately 2-5 cm in moist mineral soil |
| Soil moisture retrieval depth (L-band SAR) | Approximately 10-30 cm, moisture-dependent |
| SMAP volumetric moisture accuracy | ±0.04 m³/m³ unbiased RMSE (open terrain, published specification) |
| Sentinel-1 archive depth | From April 2014 (Sentinel-1A launch); systematic global coverage from 2016 |
| Typical deliverable format | GeoPackage or GeoTIFF risk-segment layer, CSV anomaly log, PDF interpretive report |
Analytics Satellize can run
| Trench-corridor moisture anomaly map | Sentinel-1 VV/VH backscatter time series; change detection relative to surrounding undisturbed soil buffer using empirical surface soil moisture inversion | GeoTIFF layer per acquisition epoch, showing backscatter anomaly magnitude along route segments |
| Seasonal moisture variability index | Annual amplitude of soil moisture signal extracted from Sentinel-1 time series, normalised against SMAP regional seasonal envelope | Segmented GIS polyline with variability score per 100 m route section; updated annually |
| Differential settlement risk ranking | Composite index combining moisture variability amplitude, soil shrink-swell potential (from published soil texture data), and number of wet-dry cycles detected per year | Priority-ranked segment list in GeoPackage format with supporting PDF narrative |
| Depth-moisture profile estimate | Pedotransfer function applied to surface moisture retrievals using available soil texture and bulk density inputs; uncertainty bounds stated explicitly | CSV table of estimated volumetric moisture at burial depth per segment, with ±uncertainty range; flagged as modelled inference |
| Surface ponding and runoff anomaly detection | Sentinel-2 SWIR band ratio (B11/B12) change detection during cloud-free windows; cross-referenced against Sentinel-1 wet anomalies | Alert polygon layer identifying surface water accumulation over or adjacent to trench corridor |
| Inspection prioritisation schedule | Threshold-based flagging of segments exceeding agreed moisture-variability or anomaly-frequency criteria, updated after each Sentinel-1 pass | Quarterly ranked inspection list in CSV and PDF, with GPS coordinates of highest-priority sections |
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.