Soil moisture drives many hydrological and ecological processes including runoff, stream flow, forest productivity, nutrient cycling, drought stress,

Photo Credit

USDA Forest Service photo by Adrienne Nottingham

Soils at the Monongahela National Forest in West Virginia. Many existing soil moisture data products do not account for soil moisture in the rooting zone, deep within the soil profile.

 insect and disease outbreaks, and wildland fire probability and severity. Land managers therefore need high quality and timely information about soil moisture to inform their management decisions. Currently, the desired high resolution spatial and temporal soil moisture datasets (i.e., less than 1 kilometer in resolution) are not widely available. Land surface models can also be used to estimate soil moisture, but they require extensive computational capacity to be run at large scale and with temporal frequency. Many existing soil moisture data networks include agricultural sites or sites with slight slopes, and do not accurately represent the complex terrain of many forests. To develop high-resolution remote sensing products and land surface models, we need in-situ soil moisture measurements to calibrate and test the models.

This pilot study at the Fernow Experimental Forest established on-site soil moisture measurements across a variety of complex terrain conditions. In the future, similar networks could be installed in other experimental forests across the country. These new measurements will help calibrate new, more accurate models that forest managers can use to inform decisions that affect the ability for forests to provide clean drinking water, fiber for energy and construction, and carbon storage, among other ecosystem services.

To test the feasibility of a broader network of soil moisture measurements across experimental forests, researchers are establishing pilot study at the Fernow Experimental Forest in Parsons, West Virginia in cooperation with the USDA Natural Resources Conservation Science  Soil and Plant Sciences Division and the USDA-Agricultural Research Service Remote Sensing Lab in nearby Beltsville, Maryland.

The pilot consists of four small watersheds spread across 2.5 kilometers, which is roughly the size of a pixel from a downscaled NASA Soil Moisture Active Passive (SMAP) product. Each small watershed is about 500 meters across, which is equal to about four pixels in imagery from NASA-NISAR, a satellite launched in 2024. Within each watershed, at least 8 soil moisture arrays will be deployed at variable hillslope positions and aspects. This nested design enables testing remote sensing and modeling estimates of soil moisture at multiple scales and can be implemented in phases.

Once soil moisture sensors are installed in at least one watershed, an unmanned aerial system (UAS) equipped with an L-band radiometer will be flown over the pilot to compare remotely sensed spatial data to in-situ measurements. The L-band radiometer on the UAS is the same technology deployed on the NASA-SMAP and NISAR satellites. We will also install a tower mounted P-band radiometer that is capable of estimating soil moisture to a depth of 80 centimeters. This instrument will provide high resolution temporal estimates of soil moisture in the roots zone.

Expected Outcomes

The data produced by the network would allow researchers to better parameterize and test watershed scale eco-hydrology models, regional land surface models, and remote sensing products. Improved accuracy of these models and remote sensing products would inform land management decisions, particularly in forests with complex terrain. Replicating this network across experimental forests would allow researchers to improve models across a greater range of forest types.