Skip to main content
ARS Home » Research » Publications at this Location » Publication #257050

Title: Using Agricultural in Situ Soil Moisture Networks to Validate Satellite Estimates

item Cosh, Michael
item Prueger, John
item Jackson, Thomas

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 8/1/2010
Publication Date: 9/27/2010
Citation: Cosh, M.H., Prueger, J.H., Jackson, T.J. 2010. Using Agricultural in Situ Soil Moisture Networks to Validate Satellite Estimates. Remote Sensing and Hydrology 2010 Symposium. 2010 CDROM.

Interpretive Summary:

Technical Abstract: The validation of soil moisture remote sensing products is generally based upon the in situ networks which are in non-representative locations. Soil moisture sensors have until recently, been added to existing precipitation networks which have different requirements for representativeness, such as a full sky view and uniform distribution. These stations are generally located along field boundaries or in non-representative sites with regards to soil type or soil moisture. In addition, it is difficult to install permanent sensors into a tilled acreage, such as an agricultural field. An initial attempt at a realistic agricultural network has been developed in the Walnut Creek watershed near Ames, Iowa. Small temporary soil moisture stations are installed within the corn and soybean fields which dominate the region. Land owners and operators are able to move the stations if necessary, such as during planting and harvesting or field tillage. This network design results in a non-continuous, but representative watershed average during active growing seasons. Begun in 2006, nine stations have been recording the surface soil moisture (~5 cm), which is commonly used in the validation of the AMSR-E instrument. Statistical comparisons of the network performance over the years and through the seasons reveal a consistent soil moisture regime. The AMSR-E soil moisture products are compared to the Walnut Creek network during its deployment. There is a strong correlation between the remote sensing signal and the in situ network, but there is a significant bias.