Real Results with Simulated Data: OSSE Experiments with Data Assimilation of Snow Depth at Basin Scales Inform Mission Design

Snow depth observations can be leveraged with data assimilation (DA) to improve estimation of snow density and snow water equivalent (SWE). A key consideration for mission and campaign design is how snow depth retrieval characteristics (including observation timing/frequency, sampling error) influence SWE accuracy and uncertainty in a DA framework. To quantify these effects, we conduct a series of Observation System Simulation Experiments (OSSEs). We use a process-based model (Alpine3D) to simulate "truth" snow depth, density, and SWE (the Nature Run, NR) across the East River Basin (CO), and then implement a particle filter (PF) assimilation technique to assimilate the depth (with a range of simulated retrieval scenarios) into a different model (Factorial Snow Model 2, FSM2). The experimental setup allows for testing a flexible range of measurement timing and sampling error scenarios representative of remote sensing capabilities. Using the OSSE framework, we test whether our previous findings at the point scale are applicable at the basin scale. Previously, we found that (1) assimilation reduces biases in density and SWE, (2) there is little incremental benefit to SWE accuracy when assimilating more than one depth observation near peak accumulation, (3) SWE estimates are less sensitive to observation timing than sampling error, and (4) more frequent depth observations improve melt-out date timing and reduce SWE uncertainty. At the basin scale, we test whether the PF mostly acts to increase model precipitation inputs, while not systematically shifting other parameter values or forcings. Initial results show that the PF is able to accurately simulate patterns of snow accumulation and melt beneath forest canopies (as represented in the NR), and DA reduces model sensitivity to interception and other canopy parameters (e.g., fractional vegetation, fractional sky view), reducing the need for local observations of site characteristics for modeling.