Data Assimilation to extract Soil Moisture Information from SMAP Observations.

Area of Science:

• Earth Science
• Hydrology
• Remote Sensing

Background:

• Accurate soil moisture data is crucial for understanding land surface processes and climate.
• The Soil Moisture Active Passive (SMAP) mission provides valuable satellite-based soil moisture observations.
• Assimilation of SMAP data into land surface models can improve soil moisture estimates.

Purpose of the Study:

• To compare different methods for assimilating SMAP soil moisture retrievals into the NASA Catchment model.
• To evaluate the impact of neural network (NN) and physically-based retrievals with various bias correction strategies.
• To assess the extraction of independent information from SMAP observations.

Main Methods:

• Assimilated NN and physically-based SMAP soil moisture retrievals into the NASA Catchment model over the contiguous United States.
• Employed global and localized bias correction methods for SMAP retrievals.
• Validated model performance against in situ measurements from SMAP core validation sites (CVS).

Main Results:

• Assimilation of NN retrievals without bias correction improved correlations and reduced ubRMSE against in situ data.
• Global bias correction methods showed potential for extracting more independent information but were vulnerable to retrieval errors.
• Localized bias correction yielded slightly lower skill improvements compared to global methods.

Conclusions:

• Different SMAP data assimilation strategies offer varying degrees of soil moisture improvement.
• Global bias correction can enhance information extraction but requires robust quality control.
• Careful consideration of bias correction and land model recalibration is essential for optimal assimilation results.