Abstract
In order to validate its soil moisture products, the NASA Soil Moisture Active Passive (SMAP) mission utilizes sites with permanent networks of in situ soil moisture sensors maintained by independent calibration and validation partners in a variety of ecosystems around the world. Measurements from each core validation site (CVS) are combined in a weighted average to produce an estimate of soil moisture at a 33-km scale that represents the SMAP's radiometer-based retrievals. Since upscaled estimates produced in this manner are dependent on the weighting scheme applied, an independent method of quantifying their biases is needed. Here, we present one such method that uses soil moisture measurements taken from a dense, but temporary, network of soil moisture sensors deployed at each CVS to train a random forests regression expressing soil moisture in terms of a set of spatial variables. The regression then serves as an independent source of upscaled estimates against which permanent network upscaled estimates can be compared in order to calculate bias statistics. This method, which offers a systematic and unified approach to estimate bias across a variety of validation sites, was applied to estimate biases at four CVSs. The results showed that the magnitude of the uncertainty in the permanent network upscaling bias can sometimes exceed 80% of the upper limit on SMAP's entire allowable unbiased root-mean-square error (ubRMSE). Such large CVS bias uncertainties could make it more difficult to assess biases in soil moisture estimates from SMAP.
Highlights
Concerns over recent and projected changes in our planet’s climate have cast a spotlight on the urgent need for atmospheric models accurate enough to provide reliable climate projections under a variety of policy options
We have investigated in situ core validation site (CVS) scaling function biases for four different Soil Moisture Active Passive (SMAP) CVSs
Run results were generated for many values of N outside the range of values included in these multi-N statistics, but it was decided that the statistics collected using this range of N s were likely most representative of the best achievable bias in CVS scaling function soil moisture estimates
Summary
Concerns over recent and projected changes in our planet’s climate have cast a spotlight on the urgent need for atmospheric models accurate enough to provide reliable climate projections under a variety of policy options The development of such models relies on a solid understanding of the global hydrological cycle, of which one key observable over land is soil moisture. The Soil Moisture Active Passive (SMAP) satellite observatory [3], launched in January 2015, was developed to provide frequent global observations of soil moisture using L-band (∼1.4 GHz) microwave measurements collected by on-board radiometer and synthetic aperture radar instruments. The mission requirements for SMAP included a retrieval accuracy placing an upper limit of 0.04 m3/m3 on the overall unbiased root-mean-square error (ubRMSE) This is to be achieved at a spatial scale no greater than 10 km for vegetation water content of less than 5 kg/m2. The L2SMP E product is posted in the EASEGrid-2.0 projection (https://nsidc.org/data/ease/ease grid2.html) [6] [7] at 9 km, with each 9–km soil moisture pixel reflecting data aggregated over a 33–km contributing domain (i.e., primary spatial area contributing to the radiometer brightness temperature response) centered on the pixel, as shown in Figure 3 of [4]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
