Ensemble-based assimilation of fractional snow-covered area satellite retrievals to estimate the snow distribution at Arctic sites

Abstract

Abstract. With its high albedo, low thermal conductivity and large water storing capacity, snow strongly modulates the surface energy and water balance, which makes it a critical factor in mid-to high-latitude and mountain environments. However, estimating the snow water equivalent (SWE) is challenging in remote-sensing applications already at medium spatial resolutions of 1 km. We present an ensemble-based data assimilation framework that estimates the peak subgrid SWE distribution (SSD) at the 1 km scale by assimilating fractional snow-covered area (fSCA) satellite retrievals in a simple snow model forced by downscaled reanalysis data. The basic idea is to relate the timing of the snow cover depletion (accessible from satellite products) to the peak SSD. Peak subgrid SWE is assumed to be lognormally distributed, which can be translated to a modeled time series of fSCA through the snow model. Assimilation of satellite-derived fSCA facilitates the estimation of the peak SSD, while taking into account uncertainties in both the model and the assimilated data sets. As an extension to previous studies, our method makes use of the novel (to snow data assimilation) ensemble smoother with multiple data assimilation (ES-MDA) scheme combined with analytical Gaussian anamorphosis to assimilate time series of Moderate Resolution Imaging Spectroradiometer (MODIS) and Sentinel-2 fSCA retrievals. The scheme is applied to Arctic sites near Ny-Ålesund (79 • N, Svalbard, Norway) where field measurements of fSCA and SWE distributions are available. The method is able to successfully recover accurate estimates of peak SSD on most of the occasions considered. Through the ES-MDA assimilation, the root-mean-square error (RMSE) for the fSCA, peak mean SWE and peak subgrid coefficient of variation is improved by around 75, 60 and 20 %, respectively, when compared to the prior, yielding RMSEs of 0.01, 0.09 m water equivalent (w.e.) and 0.13, respectively. The ES-MDA either outperforms or at least nearly matches the performance of other ensemble-based batch smoother schemes with regards to various evaluation metrics. Given the modularity of the method, it could prove valuable for a range of satellite-era hydrometeorological reanalyses.

Towards the assimilation of satellite reflectance into semi-distributed ensemble snowpack simulations

“…From MODIS and S2 spectral Top Of Atmosphere (TOA) radiance products, it is possible to retrieve the snowpack extent as a Snow Cover Fraction by pixel (SCF) and Bottom of Atmosphere (BOA) reflectances which requires to account for the complexity of the radiative transfer in mountainous area (Richter, 1998;Sirguey, 2009). Many studies focus on the assimilation of SCF, showing a strong impact of assimilation in hydrological models (De Lannoy et al, 2012;Thirel et al, 2013;Stigter et al, 2017;Aalstad et al, 2018;Baba et al, 2018). However, SCF is expected to be of less interest for detailed snowpack modelling in alpine terrain, because the information content is limited to the snow line (Andreadis and Lettenmaier, 2006;Toure et al, 2018).…”

Hyper-resolution ensemble-based snow reanalysis in mountain regions using clustering

“…However, basic surface variables in many remote mountain areas remain poorly quantified despite large increases in the capacity of in situ observations, remote sensing platforms and atmospheric model products. Spatial resolutions of 100 m are commonly recommended for modelling of land surface variables such as snow cover or surface temperature in complex terrain (Bierkens et al, 2015;Wood et al, 2011;Baldo and Margulis, 2018) and has come to be known as hyper-resolution (Wood et al, 2011). This is due to the fact that energy and mass fluxes exhibit strong lateral variation due to the effects of topography (Gruber and Haeberli, 2007), and surface/subsurface properties such as vegetation cover (Shur and Jorgenson, 2007), ground material (Gubler et al, 2012) or snow distribution (Zhang, 2005;Liston, 2004) further compound these effects.…”

“…Most continental to global modelling studies operate on regular grids which has placed limitations on model resolutions despite advances in computing power. However, previous efforts using hydrological response units, HRUs (Beven and Kirkby, 1979;Durand et al, 1993;Fiddes and Gruber, 2012), triangular irregular networks (Mascaro et al, 2015;Tucker et al, 2001), or multi-resolution approaches (Baldo and Margulis, 2018) suggest that regular grids are not only expensive but suboptimal as often only subsets of watersheds require detailed model descriptions in order to characterize the system adequately. In addition, deterministic modelling schemes have limitations even at hyper-resolution due to errors in forcing data, particularly regarding fields such as precipitation which suffer from both measurement and modelling biases.…”

“…the European Centre for Medium-Range Weather Forecast, ECMWF, and the North American Land Data Assimilation System, NLDAS), only relatively recently has data assimilation started to be utilized in land surface modelling schemes (Liu et al, 2012); however, it has already shown much promise in the current era of plentiful remote sensing data. In recent years, ensemble-based DA has been successfully applied to the problem of improving snowpack estimates at various spatial scales (Margulis et al, 2015;Aalstad et al, 2018;Magnusson et al, 2017;Griessinger et al, 2016); this is particularly pertinent as it is widely recognized that estimating the spatial distribution of the snow water equivalent (SWE) in mountain regions is currently one of the most important unsolved problems in snow hydrology and in understanding the spatial distribution of other processes dependent on the snowpack mass balance, such as the surface energy balance.…”

Assimilation of Sentinel-2 Data into a Snowpack Model in the High Atlas of Morocco

“…Here we focused on the assimilation of the snow cover area because it is easily retrieved from widely available optical remote sensing observations. However, assimilating the snow cover area only can lead to ambiguous posterior estimations [57]. For instance, if the prior underestimates the snow cover duration at high elevation with respect to satellite observations, this can be corrected by decreasing the air temperature (to reduce melting rate) but also by increasing the precipitation (to increase snow accumulation).…”