Gap-filling mountain regions of the Snow CCI SWE product using Bayesian snow reanalysis
Abstract ID: 3.13009 | Accepted as Talk | Talk/Oral | TBA | TBA
Colleen Mortimer (0)
Sun, Haorui (2), Fang, Yiwen (3), Margulis, Steven A. (2), Mudryk, Lawrence (1), Derksen, Chris (1), Marin, Carlo (4), Barella, Riccardo (4), Nagler, Thomas (5), Schwaizer, Gabriele (5)
Colleen Mortimer (1)
Sun, Haorui (2), Fang, Yiwen (3), Margulis, Steven A. (2), Mudryk, Lawrence (1), Derksen, Chris (1), Marin, Carlo (4), Barella, Riccardo (4), Nagler, Thomas (5), Schwaizer, Gabriele (5)
1
(1) Climate Research Division, Environment and Climate Change Canada,, Toronto, Canada
(2) Department of Civil and Environmental Engineering, University of California, Los Angeles, Los Angeles, USA
(3) Zhejiang University-University of Illinois at Urbana-Champaign Institute,, Haining, China
(4) EURAC Research, Bolzano, Italy
(5) ENVEO – Environmental Earth Observation IT GmbH, Innsbruck, Austria
(2) Department of Civil and Environmental Engineering, University of California, Los Angeles, Los Angeles, USA
(3) Zhejiang University-University of Illinois at Urbana-Champaign Institute,, Haining, China
(4) EURAC Research, Bolzano, Italy
(5) ENVEO – Environmental Earth Observation IT GmbH, Innsbruck, Austria
The ESA CCI Snow water equivalent (SWE) product does not provide estimates over complex terrain because the grid spacing of the input passive microwave (PMW) data is incompatible with the scales of SWE variability, SWE exceeds the limit of PMW differencing methods, and in situ snow depth observations used to constrain the retrieval are insufficient to meaningfully capture elevation and topographic controls on snow depth distribution. We address this gap within the Snow CCI product line by assimilating Snow CCI snow cover fraction (SCF) (1 km resolution) into a previously developed Bayesian reanalysis approach (Margulis et al. 2016) to reconstruct SWE. Initial implementations over four mountain domains in western North America (Tuolumne and Aspen in the US, Bow and Lajoie in Canada) for water years 2001 – 2019, showed that the SWE reanalysis framework using the Snow CCI SCF product performs similarly to previous implementations in other regions and using higher resolution SCF estimates. However, there were systematic differences in performance compared to a reference analysis that assimilated LandSat-derived SCF (30 m resolution), in part attributed to the differing resolutions of the two input SCF products, although viewing geometry also played a role. Analysis showed that both the number of SCF images and their characteristics, such as zenith angle and cloud cover fraction, significantly affect the accuracy of SWE estimation. Finally, performance was worse in the Canadian domains where we applied meteorological downscaling and prior uncertainty estimates which were developed for the Western US. These aspects need to be better understood prior to large-scale implementation of the method. We will present results of the initial tests in Western North America and preliminary results from experiments designed to understand the impact of SCF spatial resolution, viewing angle geometry, and meteorological forcing. Margulis, S. A. et al. (2016) J. Hydromet., doi: 10.1175/JHM-D-15-0177.1
| N/A | ||||||||
|