Validation of the SMAP Passive Soil Moisture Product in an Arctic Tundra Environment

Over many remote regions, such as the Canadian tundra, little to no information is available for understanding soil moisture conditions due to a lack of monitoring by in situnetworks. SMAP will be essential to monitoring soil moisture conditions within these regions with applications for monitoring permafrost degradation and carbon exchange. Soil moisture products are currently being produced from SMAP over arctic regions, however there are few in situ observation networks available for validation of these products. The objective of this research is to evaluate of the SMAP 36km soil moisture product over an artic tundra region. During the summers of 2015 and 2016, a soil moisture-monitoring network was installed near Trail Valley Creek, approximately 50km north-east of Inuvik, NWT, in the Canadian Arctic Tundra. The network consisted of 10 soil moisture monitoring stations, with Stevens Hydra II probes installed horizontally at depths of 5 and 20cm. At each monitoring site a site and depth specific calibration was conducted. The SMAP soil moisture product is being derived on the EASE-2 Grid, which in lower latitudes, provides a 36x36km product. However, northern latitudes, this grid becomes very elongated, with pixel dimensions closer to 15km wide (in East-West direction) by 85km long, which is not representative of the areal measurement of the emitted microwave radiation. When compared to the network in 2015, there was no correlation between the network measured soil moisture and the SMAP radiometer-only soil moisture product. However, the brightness temperature product is currently available on the polar grid (a 36x36km grid for polar regions). For evaluation, the brightness temperature was modelled using the Community Microwave Emission Model (CMEM) and soil moisture, soil and vegetation characteristics observed from the in situ network. The correlation between the CMEM derived brightness temperature and the SMAP polar grid brightness temperature indicated moderate correlation (r = 0.55). The in situ measurements from the 2016 will be used to verify these results. These modelled results, relative to current SMAP soil moisture product, support the need to derive the soil moisture product on the polar grid for more accurate retrieval of soil moisture.