About Us | Help Videos | Contact Us | Subscriptions

Testbed design and evaluation of in situ sensors


The Soil Moisture Active Passive Mission (SMAP), launched into orbit in January 2015, is a NASA satellite sensor that maps soil moisture on the land surface at a 30- to 40-km grid scale. Supporting SMAP are more than 15 soil moisture networks around the world, which provide soil moisture measurements at half-hour or hourly intervals and depths ranging from the surface to 2 m.

These networks are designed to support SMAP calibration and validation, but differences between the networks complicate these efforts. For example, existing networks employ a variety of sensors to measure soil moisture, and each sensor type has unique sensitivities and biases. This lack of uniformity hinders internetwork comparisons and large-scale SMAP validation.

To address some of these challenges, the SMAP-MOISST (Marena Oklahoma In Situ Sensor Testbed) was initiated in May 2010 at an Oklahoma State University research station near the abandoned settlement of Marena, OK. In an open access paper in the April 2016 issue of Vadose Zone Journal, a research team compared the performance of several soil moisture sensors deployed at the site, including ones used today in major monitoring networks, as well as new technologies, such as the Cosmic Ray Soil Moisture Observing System (COSMOS) and Passive/Active Distributed Temperature Sensing System (DTS).

Soil moisture sample test bed

The variety of sensors used in existing networks is important to consider because different regions have local preferences and complexities influencing sensor selection. As SMAP validation metrics are being calculated, it is necessary to quantify the impact of these different measurement systems on overall scientific conclusions. In particular, the SMAP mission has an accuracy requirement of 0.04 m3m-3, so in situ sensors would need to have a root mean square error (RMSE) at or below this target. Results from the testbed may also help validate the European Space Agency SMOS (Soil Moisture and Ocean Salinity) satellite soil moisture retrievals.

The current study revealed that all the sensors with the exception of one achieved an RMSE of less than 0.04 m3m-3. However, it was necessary to apply site-specific calibration to most sensors to reach this RMSE.

In addition, in situ stations are often installed in locations that may not represent the larger landscape. To determine if single sensors could estimate the larger spatial scale, a two-year experiment was conducted in which soil moisture samples were collected in eight radial directions at a sampling distance of 50 m between sites.

The experiment showed that some sensors accurately represented the larger landscape with RMSE values below 0.04 m3m-3 while others had RMSE values as high as 0.08 m3m-3. The scientists found that smaller RMSE values were obtained at shallower (2.5 cm) than deeper depths (10 cm), regardless of the sensor. They also observed that for most sensors with a simple scaling function, a single sensor could approximate field-scale soil moisture with RMSE at or below 0.03 m3m-3.

Read the full open access article here