Data Assimilation to Represent Unmodeled Anthropogenic Processes and Topographic Correction

The present study addresses two crucial challenges faced by land Data assimilation (DA) community namely (a) rendering topographic effect in forward radiative transfer model (RTM) for a reliable direct brightness temperature (Tb) assimilation over hilly terrain, and (b) the benchmarking of soil moisture observation to tackle forcing and model uncertainties. In the first objective, the flat earth model in the conventional forward radiative transfer operator is altered to incorporate the topographic effects such as rotation of polarization plane, change in local incidence angles and shadow pixel effect [Nair and Indu, 2018]. Such a study is very critical for direct Tb assimilation, as one of the key assumptions in DA is an unbiased forward operator. However, previous studies have neglected the bias in forward RTM due to topographic impact. Our study implements a novel approach for topographic corrections in RTM for an efficient direct Tb assimilation over hilly terrain. The proposed RTM formulation with relief correction (Topo) shows improved soil moisture sensitivity with respect to conventional RTM (Flat). The Tb simulations from Topo case indicates a 53 % reduction in Bias with respect to the Flat case. This study further demonstrates that topographic impact is independent of weather conditions. However, due to uncertainty in Tb measurements, it recommends avoiding assimilation during and just after precipitation events owing to the uncertainty in satellite measurements.

The second objective is achieved by assimilating soil moisture products from different sources into the Noah land surface model (LSM). The soil moisture observations for this study stems from European space agency’s (ESA) Climate Change Initiative (CCI) multi-satellite product, Soil Moisture and Ocean Salinity (SMOS), and active sensor based Advanced Scatterometer (ASCAT) data. These products are assimilated into LSM using Ensemble Kalman filter (EnKF) approach [Nair and Indu, 2016]. Results indicate improved model simulations with a reduction in uncertainties induced due to unmodeled process such as irrigation. However, ASCAT based soil moisture has an upper hand to underpin irrigation uncertainties owing to its high spatial resolution. The ESA CCI soil moisture shows superiority in reducing the uncertainty in forcing data. This is owing to the high spatiotemporal coverage of blended soil moisture. This study provides preliminary benchmarking of soil moisture observation on a regional scale.

References:

Nair, A. S., and Indu, J., [2018] A Coupled Land Surface and Radiative Transfer Models Based on Relief Correction for a Reliable Land Data Assimilation Over Mountainous Terrain, IEEE Geoscience and Remote Sensing Letters, (Accepted for inclusion in a future issue). Doi: 10.1109/LGRS.2018.2854908.

Nair, A. S., and Indu, J., [2016], Enhancing Noah Land Surface Model Prediction skill over Indian Subcontinent by Assimilating SMOPS Blended Soil Moisture, Remote Sensing, 8, 976; Doi: 10.3390/rs8120976