Introduction to Modern Soil Investigation

Soil is the foundation of terrestrial life, supporting agriculture, forestry, urban development, and natural ecosystems. Accurate soil investigation and mapping are critical for sustainable land management, crop yield optimization, environmental conservation, and infrastructure planning. Traditional field-based soil surveys, while essential, are labor-intensive, time-consuming, and often impractical for large or inaccessible areas. Over the past two decades, Geographic Information Systems (GIS) and remote sensing (RS) technologies have fundamentally transformed how soil scientists collect, analyze, and interpret spatial soil data. By combining satellite- or aircraft-based imagery with powerful spatial analysis tools, practitioners can now map soil properties across entire regions with unprecedented speed, consistency, and detail. This article explores the core concepts, techniques, applications, benefits, and future directions of using GIS and remote sensing for soil investigation and mapping.

Understanding GIS and Remote Sensing

What Is GIS?

A Geographic Information System (GIS) is a computer-based framework for capturing, storing, querying, analyzing, and displaying geographically referenced data. At its heart, GIS integrates hardware, software, and data to allow users to visualize patterns, relationships, and trends in spatial information. Layers of data—such as topography, land cover, soil type, climate, and hydrology—can be overlaid and analyzed together. For soil investigations, GIS serves as the central platform where remote sensing outputs, field measurements, and auxiliary datasets are combined to create detailed soil maps and predictive models.

What Is Remote Sensing?

Remote sensing refers to the acquisition of information about an object or area from a distance, typically using sensors mounted on satellites, aircraft, or drones. These sensors detect and record electromagnetic radiation reflected or emitted from the Earth’s surface. Different wavelengths—visible, near-infrared, shortwave infrared, and thermal—reveal distinct soil characteristics. For example, soil color, moisture content, organic matter, and mineral composition all influence spectral reflectance patterns. Remote sensing provides synoptic, repeatable coverage over large extents, making it indispensable for regional soil surveys.

How GIS and Remote Sensing Work Together

While remote sensing generates raw imagery, GIS provides the analytical environment to extract meaningful soil information. Satellite images are first georeferenced and corrected for atmospheric and geometric distortions. Then, in a GIS, spectral indices (e.g., Normalized Difference Vegetation Index, Soil Adjusted Vegetation Index) are calculated, supervised or unsupervised classifications are performed, and spatial interpolation techniques (e.g., kriging, inverse distance weighting) are applied to create continuous soil property surfaces. The integration enables researchers to extrapolate point-based soil observations across landscapes and to monitor changes over time.

Key Applications in Soil Investigation

Mapping Soil Types and Classes

One of the most fundamental applications is the delineation of soil types and taxonomic classes over large areas. Traditional soil mapping relies on intensive field sampling and aerial photo interpretation, but remote sensing accelerates the process. Multispectral imagery from Landsat (30 m resolution) or Sentinel-2 (10–20 m resolution) can be classified using machine learning algorithms to distinguish soil units based on color, texture, and vegetation patterns. For instance, the USDA Natural Resources Conservation Service (NRCS) has integrated remote sensing data into its digital soil mapping efforts, improving the accuracy of Soil Survey Geographic (SSURGO) databases.

Assessing Soil Erosion and Degradation

Soil erosion by water or wind is a major threat to agricultural productivity and ecosystem health. GIS and remote sensing enable spatially explicit erosion risk assessment. The Universal Soil Loss Equation (USLE) and its revisions (RUSLE) can be parameterized using digital elevation models (DEMs) derived from LiDAR or stereo satellite imagery, rainfall data, land cover classifications, and soil erodibility maps. For example, using Sentinel-2 imagery to derive the cover-management factor (C-factor) and a LiDAR-based DEM to compute the topographic factor (LS-factor) allows researchers to map erosion potential at field to watershed scales. Case studies in the Loess Plateau of China and the Sahel region have shown that remote sensing-based erosion maps correlate well with field measurements and help prioritize conservation interventions.

Monitoring Soil Moisture

Soil moisture is a critical variable for irrigation scheduling, drought monitoring, and hydrological modeling. Active and passive microwave remote sensing—such as from the SMAP (Soil Moisture Active Passive) satellite or Sentinel-1 Synthetic Aperture Radar (SAR)—can estimate surface soil moisture content (0–5 cm depth) with global coverage. In a GIS, these moisture maps can be downscaled using topographic indices (e.g., Topographic Wetness Index) and land cover data to produce field-scale moisture estimates. Farmers and water managers use such products to optimize irrigation and reduce water waste. Moreover, thermal infrared sensors (e.g., Landsat thermal band) provide complementary information by detecting canopy temperature, which relates to plant water stress and root-zone moisture.

Detecting Soil Salinity and Alkali Conditions

Soil salinization affects vast areas of irrigated land, particularly in arid and semi-arid regions. Salt-affected soils exhibit distinct spectral signatures in the visible and near-infrared bands: increased reflectance in the visible range and reduced in the shortwave infrared. Using multispectral indices like the Normalized Difference Salinity Index (NDSI) or Soil Salinity Index (SSI), GIS and remote sensing can map salinity extent and severity. Researchers have combined Landsat time series with field conductivity measurements to map salinity dynamics in the Indus Basin and the San Joaquin Valley, providing actionable information for remediation planning.

Estimating Soil Organic Carbon (SOC) and Fertility

Soil organic carbon is a key indicator of soil health and fertility. Hyperspectral remote sensing (imaging spectroscopy) captures hundreds of narrow spectral bands that can resolve absorption features related to organic matter, clay minerals, and iron oxides. Although spaceborne hyperspectral missions (e.g., EnMAP, PRISMA) are still relatively new, airborne sensors like AVIRIS have demonstrated accurate SOC mapping over agricultural fields. In a GIS framework, these spectral estimates are combined with environmental covariates (topography, land use, climate) to produce high-resolution SOC maps (ISRIC - World Soil Information uses such approaches for global soil property mapping). Farmers can then apply variable-rate fertilization based on these maps, improving efficiency and reducing environmental impact.

Techniques and Data Sources

Satellite Sensors and Platforms

Several satellite missions provide free or low-cost imagery suitable for soil investigations:

  • Landsat (NASA/USGS): 30 m multispectral, 100 m thermal; since 1972, ideal for multi-decadal change analysis.
  • Sentinel-2 (ESA): 10–20 m multispectral (13 bands), 5-day revisit; excellent for vegetation and soil mapping.
  • MODIS (NASA): 250–1000 m, daily coverage; used for regional soil moisture and vegetation monitoring.
  • SMAP (NASA): L-band radar and radiometer; dedicated to soil moisture mapping at 3–9 km resolution.
  • Sentinel-1 (ESA): C-band SAR; all-weather soil moisture and texture estimation.
  • PlanetScope (Planet Labs): 3 m daily multispectral imagery (commercial); suitable for field-scale precision agriculture.

LiDAR for Topographic and Structural Information

Light Detection and Ranging (LiDAR) provides highly accurate digital elevation models (DEMs) with sub-meter vertical precision. These DEMs are essential for computing slope, aspect, curvature, and flow accumulation—variables that strongly influence soil formation, erosion, and moisture distribution. LiDAR can also penetrate vegetation canopy to reveal the bare-earth surface, making it invaluable for soil mapping in forested or shrub-dominated landscapes. Many national mapping agencies (e.g., USGS 3DEP) provide open LiDAR data.

Hyperspectral and Multispectral Imaging

Hyperspectral sensors capture data in hundreds of contiguous narrow spectral bands, enabling detailed characterization of soil mineralogy, organic matter, and moisture. The Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) and the spaceborne PRISMA mission are prominent examples. Processing hyperspectral data requires advanced techniques like continuum removal and matched filtering, but the payoff is the ability to map soil properties that are invisible to broad-band sensors. Multispectral sensors (e.g., Landsat, Sentinel-2) remain more widely accessible and are often sufficient for mapping soil types, moisture, and salinity using indices and classification.

Unmanned Aerial Vehicles (UAVs or Drones)

Drones equipped with multispectral, thermal, or LiDAR sensors offer ultra-high-resolution soil data (<1–10 cm) for small areas (fields to a few square kilometers). This is especially useful for precision agriculture, where farmers need detailed within-field variability maps. UAVs can be deployed on-demand, avoiding cloud cover issues that plague satellite imagery. However, the limited coverage and high operational costs restrict their use to localized studies or validation campaigns.

Image Processing and Analysis Techniques

Raw remote sensing data must be processed to extract soil information:

  • Radiometric and atmospheric correction to convert digital numbers to surface reflectance.
  • Geometric correction to align images to a map coordinate system.
  • Spectral indices: NDVI for vegetation cover, SAVI for soil-adjusted vegetation, NDSI for salinity, CI for color.
  • Classification: supervised (maximum likelihood, random forest, support vector machines) or unsupervised (k-means, ISODATA) to map discrete soil units.
  • Regression modeling: partial least squares regression (PLSR), artificial neural networks (ANN), or random forest to estimate continuous soil properties from spectral bands.
  • Change detection: comparing multi-temporal imagery to identify soil degradation, erosion, or land use change.

Integrating Data in GIS

Data Preprocessing and Harmonization

Effective soil mapping requires a consistent spatial data infrastructure. In a GIS, remote sensing images are integrated with ancillary datasets such as DEMs, climate grids (precipitation, temperature), land use/land cover maps, and existing soil survey polygons. All layers must be reprojected to a common coordinate system and resampled to a uniform resolution (e.g., 30 m for Landsat-based mapping). Outliers and gaps are handled through interpolation or gap-filling algorithms.

Spatial Interpolation and Geostatistics

Since soil properties are continuous in space, interpolation methods are used to predict values at unsampled locations from point observations and remote sensing covariates. Kriging (e.g., ordinary and co-kriging) and regression kriging are popular geostatistical techniques that account for spatial autocorrelation. In regression kriging, the relationship between the target soil property (e.g., SOC) and remotely sensed predictors is first modeled using linear regression or random forest, then the residuals are kriged. This hybrid approach often yields more accurate maps than either method alone.

Digital Soil Mapping (DSM)

The integration of GIS and remote sensing has given rise to digital soil mapping (DSM), also known as predictive soil mapping. DSM uses mathematical models to relate soil observations to environmental covariates (the "scorpan" factors: soil, climate, organisms, relief, parent material, age, and space). Remote sensing contributes covariates like spectral reflectance, vegetation indices, and terrain attributes. Machine learning algorithms (random forest, gradient boosting, neural networks) then generate continuous soil property maps with associated uncertainty. The SoilGrids project is a global example of DSM, providing predictions for soil pH, organic carbon, texture, and bulk density at 250 m resolution using remote sensing and environmental data.

Workflow Example: Mapping Soil Organic Carbon in Agricultural Fields

Consider a scenario where a researcher aims to map SOC across a 50 km² agricultural region. The workflow might involve:

  1. Acquiring Sentinel-2 imagery (cloud-free, growing season) and a LiDAR DEM.
  2. Calculating spectral indices (NDVI, NBR, color indices) and terrain attributes (slope, TWI, curvature).
  3. Collecting 100 soil samples (0–30 cm depth) for laboratory SOC analysis.
  4. Using 70 samples for training a random forest model with satellite and terrain covariates; 30 samples for validation.
  5. Predicting SOC across the entire area and mapping the results in a GIS.
  6. Overlaying the SOC map with land ownership parcels to guide variable-rate lime or fertilizer application.

Benefits and Challenges

Key Benefits

  • Cost and time efficiency: Covering large areas that would require weeks of field sampling takes hours with satellite imagery.
  • Synoptic and repeatable coverage: Satellites revisit the same area every few days, enabling temporal monitoring of soil dynamics (e.g., seasonal moisture, erosion after storms).
  • High spatial resolution: Modern sensors provide sub-meter to 10 m resolution, capturing fine-scale variability important for precision agriculture.
  • Multi-layer integration: GIS allows combination of soil data with topography, climate, and land use for comprehensive analysis.
  • Improved accuracy: Combining field data with remote sensing covariates often yields more accurate maps than field data alone, especially for properties that have strong spectral or topographic signatures.
  • Access to remote areas: Satellite remote sensing enables soil investigation in regions that are difficult or dangerous to access on foot (e.g., conflict zones, wetlands, steep slopes).

Challenges and Limitations

  • Technical expertise required: Image processing, geostatistics, and machine learning demand specialized training. Many agricultural extension services lack this capacity.
  • Data volume and processing complexity: High-resolution imagery and time series generate terabytes of data, requiring powerful computing and storage.
  • Cloud cover and atmospheric effects: Optical remote sensing is hampered by persistent cloud cover in tropical regions. SAR can mitigate this but requires different processing skills.
  • Mixed pixel problems: At moderate resolutions (e.g., 30 m), a single pixel may contain multiple soil types, vegetation, and shadows, complicating spectral analysis.
  • Need for ground truthing: Remote sensing models must be calibrated and validated with field samples. Without adequate ground data, maps may be unreliable.
  • High initial costs: While many satellite datasets are free, acquiring LiDAR, hyperspectral, or high-resolution commercial imagery incurs significant expense. Drone equipment and software also require investment.
  • Soil depth limitations: Optical and radar remote sensing only penetrate the top few centimeters of the soil. Subsurface properties (e.g., subsoil texture, compaction) cannot be directly sensed; they must be inferred from surface indicators and models.

Future Perspectives

Machine Learning and Artificial Intelligence

Advances in machine learning, particularly deep learning (convolutional neural networks, recurrent networks), are enabling more accurate and automated extraction of soil information from remote sensing imagery. For example, CNNs can learn spatial patterns in high-resolution images to classify soil types or estimate SOC without handcrafted features. Transfer learning allows models pre-trained on large datasets (e.g., ImageNet) to be fine-tuned for soil mapping tasks with minimal local data. AI also facilitates automated quality control and outlier detection.

Cloud Computing and Big Data Platforms

Platforms like Google Earth Engine, Amazon Web Services (AWS) Earth, and Microsoft Planetary Computer provide access to massive archives of satellite imagery and pre-built analysis tools. These cloud-based environments eliminate the need for local data storage and high-end computing. Users can run global-scale soil mapping algorithms in minutes. Future soil investigations will increasingly leverage such platforms to process multi-sensor time series for near-real-time monitoring.

Integration with IoT and In-Situ Sensors

The Internet of Things (IoT) is bringing affordable soil sensors (moisture, temperature, EC, pH) into agricultural fields. Combining IoT data streams with satellite imagery in a GIS creates a dense observation network for calibrating remote sensing models. For instance, a network of soil moisture probes can validate SMAP estimates and improve downscaling algorithms. This hybrid approach promises highly accurate, localized soil information updated in real time.

Hyperspectral Satellite Constellations

New satellite missions—such as NASA’s Surface Biology and Geology (SBG) and the planned CHIME (Copernicus Hyperspectral Imaging Mission for the Environment)—will provide routine hyperspectral data at 30 m resolution globally. These datasets will revolutionize soil mineralogy and organic matter mapping, as they will capture the fine spectral features that are currently only available from airborne sensors. The widespread availability of spaceborne hyperspectral data will lower barriers to entry for soil scientists worldwide.

Precision Agriculture and Sustainable Land Management

The ultimate goal of soil investigation is to inform better land management. As GIS and remote sensing become more integrated into farm management systems, variable-rate technologies will apply inputs (fertilizer, irrigation, amendments) based on the high-resolution soil maps. This reduces costs, increases yields, and minimizes environmental pollution. Governments and international organizations are also using satellite-derived soil information to monitor land degradation neutrality (SDG 15.3) and to target conservation programs.

Conclusion

GIS and remote sensing have moved from experimental tools to operational necessities in soil investigation and mapping. They enable practitioners to view soils not as static points but as dynamic, spatially continuous systems. By combining the synoptic view of satellites with the analytical power of GIS, soil scientists can map properties, monitor changes, and support decisions at scales from individual fields to entire continents. The ongoing fusion of machine learning, cloud computing, and new sensor technologies will only deepen this capability, making accurate soil information accessible to a broader community. For anyone involved in agriculture, environmental management, or land-use planning, embracing these technologies is no longer optional—it is essential for a sustainable future.