The value of high-resolution measurements of soil moisture has been recognized for decades – including benefits like improved weather forecasting, better agricultural yields and prediction, and greater confidence in IED detection – yet remote sensing of soil moisture remains elusive and ground-based measurements are sparse and impractical to gather.
One of the most promising technologies for remote sensing of soil moisture is synthetic aperture radar (SAR). Its high spatial resolution and sensitivity make it ideal for detecting soil moisture, but the challenge has always been the backscatter signal created by ground vegetation. Under a contract from DARPA (read press release), OmniEarth is layering SAR data with its imagery-based land classification to provide an accurate and precise measure of soil moisture by identifying that clutter (and removing it) from the SAR moisture measurements.
How SAR Measures Soil Moisture
At an elementary level, SAR detects moisture in soil by transmitting microwave pulses to the ground and then measuring the strength of the signals that bounce back from Earth (backscatter). The amount of signal reflection is dependent upon the physical properties of the targets. In the case of soil moisture, we are essentially measuring the difference in the dielectric constants of water and soil. Dry soil absorbs radar signals, while water reflects nearly all radar signals.
Several studies have shown that SAR has an excellent capacity to measure soil moisture. But, keep in mind, all objects reflect radar to some degree or another. Plant material, especially, interferes with measurement of soil moisture because of the large amount of water that make up plants. Most of the historical studies have been conducted on bare soil. NASA’s Soil Moisture Active Passive (SMAP) mission is an exception, but it shows soil moisture at a very coarse level – showing trends rather than actionable data. To achieve the results needed for precision agriculture or military mobility, one needs much greater spatial resolution.
The efficacy of the system depends on the accuracy of the land classification data. For agricultural applications, such as corn or soy fields which have excellent homogeneity across an area, accuracy is expected to be fairly high. For areas of mixed landscaping, such as an area of mixed trees, shrubs and grass, development will take longer.
The Importance of Precision Soil Moisture Knowledge
Most existing remotely sensed observation systems, especially agricultural yield models, are based on visual cues – meaning that the crops are already suffering from too much or too little water by the time the damage can be seen by satellite or aircraft. Improved soil moisture knowledge will give growers the opportunity to adjust irrigation based on soil moisture, rather than plant health (or sickness).
Similarly, systems like OmniEarth’s current water efficiency products are based primarily on visual cues. But we have already begun feeding the prototype SAR results into our platform with good results. This will not only provide greater accuracy to the system, but give water managers additional chances to save water by helping customers better understand how much – or little – water is needed to maintain a healthy yard.