TY - Generic T1 - Using Snow Water Equivalent Reconstruction for Operational Use: Two Case Studies T2 - 84th Annual Western Snow Conference Y1 - 2016 A1 - Edward H. Bair A1 - Karl Rittger A1 - Carrie Vuyovich A1 - Bruce McGurk A1 - Jeff Dozier AB -

In the snow water equivalent (SWE) reconstruction method, the snowpack is built up in reverse from downscaled energy balance forcings and satellite-based estimates of fractional snow covered area (fSCA). For melt modeling, we have developed a full energy balance model called ParBal that relies on satellite-based measurements and does not need ground-based inputs. Using ParBal to compute the melt, SWE reconstruction has been shown to be accurate when compared with a variety of validation sources, most recently by comparison with Airborne Snow Observatory (ASO, Painter et al., 2016) measurements in the Upper Tuolumne Basin, CA USA (Bair et al., 2015; Bair et al., in review). The major disadvantage is that reconstructed SWE estimates are only available retrospectively. While this limitation cannot be overcome, we present two case studies where reconstructed SWE can be used in an operational context. 1) We run ParBal as a melt only model, examining seasonal and daily melt without reconstructing the snowpack. Although we cannot reconstruct snow on the ground in this manner, we can potentially make near real-time estimates of snowmelt if these daily melt estimates are then fed through a routing model and compared with streamflow measurements. ParBal can be driven by: a) Global Data Assimilation System (GDAS, Kleist et al., 2009) meteorological forcings, now available at 1/9° spatial resolution and at 3 hr intervals with almost no latency; and b) Moderate Resolution Imaging Spectrometer Snow Covered Area and Grain Size (MODSCAG, Painter et al., 2009) snow cover data, available daily with one or two day latency. 2) We use machine learning techniques to build statistical relationships between remotely-sensed products, such as the ones mentioned in 1), and reconstructed SWE in Afghanistan. This approach allows real-time SWE prediction in remote areas that previously had little or no baseline for comparison

JF - 84th Annual Western Snow Conference CY - Seattle, Washington UR - /files/PDFs/2016Bair.pdf ER - TY - Generic T1 - Comparison and Error Analysis of Reconstructed SWE to Airborne Snow Observatory Measurements in the Upper Tuolumne Basin, CA T2 - 83rd Annual Western Snow Conference Y1 - 2015 A1 - Edward H. Bair A1 - Karl Rittger A1 - Jeff Dozier A1 - Robert E. Davis KW - Airborne Snow Observatory KW - reconstruction KW - Sierra Nevada AB -

We present scientific and computing improvements to our new reconstruction model compared to a previous model. Snow water equivalent (SWE) reconstruction involves building a snowpack up in reverse, from melt out to peak SWE, given estimates of melt energy and fractional snow covered area (fSCA). The model was initially tested at an energy balance site on Mammoth Mountain, near the Upper Tuolumne basin, where nearly all the output could be verified and the disappearance of snow was known precisely. A full energy balance version of the model, that computes melt hourly, accurately estimated SWE and most energy balance terms. We were not able to verify a key parameter, the snow albedo, possibly because of low snow depths that affected our ability to measure solar radiation reflected by the snow. A net radiation/degree-day version of the model, that computes melt daily, underestimated SWE, probably because of its coarse daily time step. We then used SWE from the Airborne Snow Observatory (ASO) in 2014 as ground truth to verify the energy balance reconstruction model. Given the high spatial and spectral resolution of ASO measurements, we assume they are an accurate ground truth. Compared to the Snow Data Assimilation System (SNODAS) and the Advanced Microwave Radiometer 2 (AMSR2), reconstruction (with two different fSCA inputs) was by far the most accurate, with a bias of 36-40 mm at the basin wide maximum SWE, and an RMSE of 32-43 mm for all ASO measurement dates. The AMSR2 SWE estimates were generally too low, probably because of deep snow and interference from the canopy. SNODAS tended to overestimate SWE, perhaps because of an overreliance on measurements from snow pillows which are purposely located at heavy snow sites. Finally, we tested the reconstruction model over snow pillows located across the entire Sierra in 2014. The bias was 3 mm and the RMSE 140 mm at the maximum SWE accumulation, showing significant lower than a previous model.

Presentation slides (PDF)

JF - 83rd Annual Western Snow Conference CY - Grass Valley, California UR - files/PDFs/2015Bair.pdf ER - TY - Generic T1 - Snow Water Equivalent Estimates in the Hindu Kush and the Sierra Nevada Using Passive Microwave and Reconstruction T2 - 82nd Annual Western Snow Conference Y1 - 2014 A1 - Edward H. Bair A1 - Jeff Dozier A1 - Carrie Voyuvich A1 - Robert E. Davis KW - Afghanistan KW - AMSR-E KW - passive microwave KW - reconstruction KW - Sierra Nevada AB -

Accurate measurement of spatially distributed snow water equivalent (SWE) in mountain watersheds is perhaps the most significant problem in snow hydrology. We examine SWE measurements from two techniques. The first uses passive microwave estimates, provided by the National Snow and Ice Data Center, from the AMSR-E sensor aboard the Aqua satellite. Passive microwave (PM) has been used to estimate SWE for decades, and while it is subject to numerous problems, it is the only source of global real-time SWE estimates. Recently, SWE Reconstruction has been shown to be accurate at estimating basin-wide SWE in the Sierra Nevada. Reconstruction combines a melt model with snow covered area measurements to retroactively build the snowpack, from disappearance back to its peak. Reconstruction can only be used retrospectively, so it cannot be used to estimate today’s SWE. Thus, we use Reconstruction of prior water years to better understand the strengths and weaknesses of PM SWE estimates. Our test case is California’s Sierra Nevada, where we have full natural flow estimates and a large network of SWE sensors for comparison. Our application area is the Hindu Kush range in Afghanistan, where there are neither stream flow nor ground-based SWE measurements. Both regions are snowmelt dominated and subject to drought. Using annual SWE estimates from Reconstruction for verification, our results show that annual PM SWE estimates are biased in California and Afghanistan. In California, SWE estimates from AMSR-E are too low, by up to 10×. In the Amu Darya, one of the largest basins in Afghanistan, SWE estimates are too low by about 2×, while in 5/8 basins, PM SWE estimates are consistently too high. In terms of ranks, PM performs poorly, having low Spearman rank correlation coefficients. An exception is the Amu Darya, where the Spearman correlation coefficient is 0.81 for the eight years studied. We examine potential sources of error. Consistent with previous studies, we find that PM error is caused by shallow snow, deep snow, and forest cover. The explanation for the relatively low bias in SWE and relatively high correlation of rank for the Amu Darya appears to be a snowpack that was neither shallow nor exceptionally deep in a region that is nearly devoid of tree cover.

JF - 82nd Annual Western Snow Conference T3 - Proceedings of the Western Snow Conference CY - Durango, Colorado UR - sites/westernsnowconference.org/PDFs/2014Bair.pdf ER -