Google Earth Engine Tutorial: Estimate Soil Moisture with Sentinel-1 SAR

Google Earth Engine (GEE) is a powerful platform for geospatial analysis, allowing users to process large-scale satellite data efficiently. Sentinel-1 Synthetic Aperture Radar (SAR) data, with its all-weather, day-and-night imaging capabilities, is particularly useful for soil moisture estimation. This tutorial provides a step-by-step guide to using GEE for analyzing Sentinel-1 data to estimate soil moisture.

1. Accessing Sentinel-1 Data in GEE

First, load the Sentinel-1 dataset using the GEE API. Sentinel-1 provides dual-polarization data (VV and VH) in the Copernicus Open Access Hub. In the GEE code editor, use the following code to filter and select the relevant data:

var dataset = ee.ImageCollection("COPERNICUS/S1_GRD")  
          .filterDate('2023-01-01', '2023-12-31')  
          .filter(ee.Filter.eq('instrumentMode', 'IW'))  
          .select(['VV', 'VH']);

2. Preprocessing Sentinel-1 Data

Before analyzing, ensure the data is calibrated and processed for consistency. Apply a terrain correction and convert backscatter values to decibels (dB) for better interpretation:

var preprocessed = dataset.map(function(image) {  
      return image.select(['VV', 'VH'])  
                 .multiply(10)  
                 .log10()  
                 .rename(['VV_dB', 'VH_dB']);  
    });

3. Calculating Soil Moisture Index

Soil moisture can be estimated using a regression model based on backscatter intensity. The Soil Moisture Index (SMI) is often derived using the formula:
SMI = a * VV + b * VH + c,
where a, b, and c are coefficients calibrated with ground-truth data. Replace these coefficients based on your study area:

var soilMoisture = preprocessed.select('VV_dB')  
                                    .multiply(0.1)  
                                    .add(preprocessed.select('VH_dB').multiply(-0.05))  
                                    .add(5);

4. Visualizing and Exporting Results

Use GEE’s visualization tools to display soil moisture data. Add a map to the GEE interface and export the results as a GeoTIFF or CSV for further analysis:

Map.addLayer(soilMoisture, {min: 0, max: 50}, 'Soil Moisture');  
    Export.image.toDrive({  
      image: soilMoisture,  
      description: 'soil_moisture_export',  
      folder: 'GEE_Exports',  
      fileNamePrefix: 'soil_moisture',  
      region: geometry,  
      scale: 10,  
      maxPixels: 1e10  
    });

5. Validating with Ground Data

Compare your GEE soil moisture estimates with field measurements. Upload a ground data CSV to GEE and perform statistical validation using the meanAbsoluteError or rSquared functions.

FAQ