Regression Downscaling of Coarse Resolution Globsnow Snow Water Equivalent Estimates in the Red River Basin

2020 ◽  
Author(s):  
Margot Flemming ◽  
Richard Kelly
Keyword(s):  
Remote Sensing ◽  
Random Forest ◽  
Spatial Resolution ◽  
Snow Water Equivalent ◽  
Passive Microwave ◽  
Red River ◽  
Random Forest Regression ◽  
Red River Basin ◽  
Snow Water

<p>The spatial and temporal heterogeneity of seasonal snow and its impact on socio-economic and environmental functionality make accurate, real-time estimates of snow water equivalent (SWE) important for hydrological and climatological predictions. Remote sensing techniques facilitate a cost effective, temporally and spatiallyconsistent approach to SWE monitoring in areas where insitu measurements are notsufficient. Passive microwave remote sensing has been used to successfully estimate SWE globally by measuring the microwave attenuation from the Earth’s surface as a function of SWE. However, passive microwave derived SWE estimates at local scales are subject to large errors given the coarse spatial resolution of observations (~625 km<sup>2</sup>).Regression downscaling techniques can be implemented to increase the spatial resolution of gridded datasets with the use of related auxiliary datasets at a finer spatial resolution. These techniques have been successfully implemented to remote sensing datasets such as soil moisture estimates, however, limited work has applied such techniques to snow-related datasets.This study focuses on assessing the feasibility of using regression downscaling to increase the spatial resolution of the European Space Agency’s (ESA) Globsnow SWE product in the Red River basin, an agriculturally important region of the northern United States.</p><p>Prior to downscaling Globsnow SWE, three regression downscaling techniques (Multiple Linear Regression, Random Forest Regression and Geographically Weighted Regression) were assessed in an internal experiment using 1 km grid scale Snow Data Assimilation System (SNODAS) SWE estimates, developed by the National Weather Service’s National Operational Hydrological Remote Sensing Center (NOHRSC). SNODAS SWE estimates for 5-year period between 2013-2018 were linearly aggregated to a 25 km grid scale to match the Globsnow spatial resolution. Three regression downscaling techniques were implemented along with correlative datasets available at the 1 km grid scale to downscale the aggregated SNODAS data back to the original 1 km grid scale spatial resolution. When compared with the original SNODAS SWE estimates, the downscaled SWE estimates from the Random Forest Regression performed the best. Random Forest Regression Downscaling was then implemented on the original Globsnow SWE data for the same time period, as well as a corrected Globsnow SWE dataset. The downscaled SWE results from both the corrected and uncorrected Globsnow data were validated using the original SNODAS SWE estimates as well as in situ SWE measurements from a set of 40-45 (depending on the season) weather stations within the study region. Spatial and temporal error distributions were assessed through both validation datasets. The downscaled results from the corrected Globsnow dataset showed similar overall statistics to the original SNODAS SWE estimates, performing better than the downscaled results from the uncorrected Globsnow SWE dataset. The overall aim of this study is to assess the applicability of regression downscaling as a reliable and reproducible method for local scale SWE estimation in areas where finer resolution data such as SNODAS does not exist. Therefore, the goal is to reproduce the optimal regression downscaling procedure in an area other snow dominated regions across the globe using in situ snow transect data for validation.</p>

Improved Northern Hemisphere Snow Water Equivalent product from passive microwave remote sensing and in situ data

2021 ◽  
Author(s):  
Colleen Mortimer ◽  
Lawrence Mudryk ◽  
Chris Derksen ◽  
Kari Luojus ◽  
Pinja Venalainen ◽  
...  
Keyword(s):  
Spatial Resolution ◽  
Snow Depth ◽  
Snow Water Equivalent ◽  
Microwave Remote Sensing ◽  
Passive Microwave ◽  
Snow Density ◽  
Snow Season ◽  
Snow Water ◽  
Snow Mass

<p>The European Space Agency Snow CCI+ project provides global homogenized long time series of daily snow extent and snow water equivalent (SWE). The Snow CCI SWE product is built on the Finish Meteorological Institute's GlobSnow algorithm, which combines passive microwave data with in situ snow depth information to estimate SWE. The CCI SWE product improves upon previous versions of GlobSnow through targeted changes to the spatial resolution, ancillary data, and snow density parameterization.</p><p>Previous GlobSnow SWE products used a constant snow density of 0.24 kg m<sup>-3</sup> to convert snow depth to SWE. The CCI SWE product applies spatially and temporally varying density fields, derived by krigging in situ snow density information from historical snow transects to correct biases in estimated SWE. Grid spacing was improved from 25 km to 12.5 km by applying an enhanced spatial resolution microwave brightness temperature dataset. We assess step-wise how each of these targeted changes acts to improve or worsen the product by evaluating with snow transect measurements and comparing hemispheric snow mass and trend differences.</p><p>Together, when compared to GlobSnow v3, these changes improved RMSE by ~5 cm and correlation by ~0.1 against a suite of snow transect measurements from Canada, Finland, and Russia. Although the hemispheric snow mass anomalies of CCI SWE and GlobSnow v3 are similar, there are sizeable differences in the climatological SWE, most notably a one month delay in the timing of peak SWE and lower SWE during the accumulation season. These shifts were expected because the variable snow density is lower than the former fixed value of 0.24 kg m<sup>-3</sup> early in the snow season, but then increases over the course of the snow season. We also examine intermediate products to determine the relative improvements attributable solely to the increased spatial resolution versus changes due to the snow density parameterizations. Such systematic evaluations are critical to directing future product development.</p>

scholarly journals Snowpack water-equivalent estimates from satellite and aircraft remote-sensing measurements of the Red River basin, north-central U.S.A.

1998 ◽  
Vol 26 ◽  
pp. 119-124 ◽  
Author(s):  
E. G. Josberger ◽  
N. M. Mognard ◽  
B. Lind ◽  
R. Matthews ◽  
T. Carroll
Keyword(s):  
Artificial Intelligence ◽  
Remote Sensing ◽  
Linear Regression ◽  
River Basin ◽  
Satellite Observations ◽  
Passive Microwave ◽  
Red River ◽  
General Linear ◽  
North Central ◽  
Red River Basin

Most algorithms to extract dry snowpack water equivalent (SWE) from satellite passive-microwave observations are based on point measurements of SWE or extrapolation of point measurements to the 30 km footprint of the satellite observations. SWE observations on a scale comparable to the satellite observations can be obtained from airborne gamma-ray attenuation techniques from flight lines that are approximately 10 km long. During the winter of 1989, the NOAA National Operational Hydrologic Remote Sensing Center (NOHRSC) flew 92 of these night lines over a 200 × 250 km area of the Red River basin which is located in the north-central part of the United States of America. These observations provide a unique dataset of snow water-equivalent determinations on spatial scales similar to the satellite passive-microwave observations as acquired by the Defense Meteorological Satellite Program (DMSP) Special Sensor Microwave Imager (SSM/I) F-8 satellite. Land-classification determinations from the Advanced Very High Resolution Radiometer (AVHRR) show that the eastern part of the region contains a coniferous forest of varying coverage, while the remainder is farmland or prairie. SSM/I data, including observations from a no-snow case in the preceding fall, the flight-line data and the AVHRR data were all co-registered to a common 20 km grid. The resulting dataset was analyzed using linear regression, artificial intelligence and general linear models. The results showed that the passive-microwave response was similar to the response predicted by Mie scattering theory. A comparison of the three techniques found that the artificial intelligence technique and the general linear model explained significantly more of the variance in the dataset, as evidenced byR2values of 0.97 compared to 0.88 for the linear multiple-regression analysis. Hence, a neural network approach which was continually trained on new datasets as they became available, could provide better estimates of snowpack water equivalent than algorithms based on linear-regression techniques.

scholarly journals Reconstruction of Snow Depth Data at Moderate Spatial Resolution (1 km) from Remotely Sensed Snow Data and Multiple Optimized Environmental Factors: A Case Study over the Qinghai-Tibetan Plateau

2021 ◽  
Vol 13 (4) ◽  
pp. 657
Author(s):  
Pengtao Wei ◽  
Tingbin Zhang ◽  
Xiaobing Zhou ◽  
Guihua Yi ◽  
Jingji Li ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Tibetan Plateau ◽  
Environmental Factors ◽  
Spatial Resolution ◽  
Snow Depth ◽  
Microwave Remote Sensing ◽  
Passive Microwave ◽  
Passive Microwave Remote Sensing ◽  
Depth Data

Snow depth distribution in the Qinghai-Tibetan plateau is important for atmospheric circulation and surface water resources. In-situ observations at meteorological stations and remote observation by passive microwave remote sensing technique are two main approaches for monitoring snow depth at regional or global levels. However, the meteorological stations are often scarce and unevenly distributed in mountainous regions because of inaccessibility, so are the in-situ snow depth measurements. Passive microwave remote sensing data can alleviate the unevenness issue, but accuracy and spatial (e.g., 25 km) and temporal resolutions are low; spatial heterogeneity in snow depth is thus hard to capture. On the other hand, optical sensors such as moderate resolution imaging spectroradiometer (MODIS) onboard Terra and Aqua satellites can monitor snow at moderate spatial resolution (1 km) and high temporal resolution (daily) but only snow area extent, not snow depth. Fusing passive microwave snow depth data with optical snow area extent data provides an unprecedented opportunity for generating snow depth data at moderate spatial resolution and high temporal resolution. In this article, a linear multivariate snow depth reconstruction (LMSDR) model was developed by fusing multisource snow depth data, optical snow area extent data, and environmental factors (e.g., spatial distribution, terrain features, and snow cover characteristics), to reconstruct daily snow depth data at moderate resolution (1 km) for 16 consecutive hydrological years, taking Qinghai-Tibetan Plateau (QTP) as a case study. We found that snow cover day (SCD) and environmental factors such as longitude, latitude, slope, surface roughness, and surface fluctuation have a significant impact on the variations of snow depth over the QTP. Relatively high accuracy (root mean square error (RMSE) = 2.26 cm) was observed in the reconstructed snow depth when compared with in-situ data. Compared with the passive microwave remote sensing snow depth product, constructing a nonlinear snow depletion curve product with an empirical formula and fusion snow depth product, the LMSDR model (RMSE = 2.28 cm, R2 = 0.63) demonstrated a significant improvement in accuracy of snow depth reconstruction. The overall spatial accuracy of the reconstructed snow depth was 92%. Compared with in-situ observations, the LMSDR product performed well regarding different snow depth intervals, land use, elevation intervals, slope intervals, and SCD and performed best, especially when the snow depth was less than 3 cm. At the same time, a long-time snow depth series reconstructed based on the LMSDR model reflected interannual variations of snow depth well over the QTP.

scholarly journals Evaluation of long term Northern Hemisphere snow water equivalent products

2019 ◽  
Author(s):  
Colleen Mortimer ◽  
Lawrence Mudryk ◽  
Chris Derksen ◽  
Kari Luojus ◽  
Ross Brown ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Northern Hemisphere ◽  
Snow Water Equivalent ◽  
European Space Agency ◽  
Microwave Remote Sensing ◽  
Passive Microwave ◽  
Temporal Correlations ◽  
Surface Snow ◽  
Snow Water ◽  
Surface Observations

Abstract. Seven gridded northern hemisphere snow water equivalent (SWE) products were evaluated as part of the European Space Agency (ESA) Satellite Snow Product Inter-comparison and Evaluation Exercise (SnowPEx). Three categories of datasets were assessed: (1) those utilizing some form of reanalysis (the NASA Global Land Data Assimilation System version 2 – GLDAS; the European Centre for Medium-Range Forecasts interim land surface reanalysis – ERA-land; the NASA Modern-Era Retrospective Analysis for Research and Applications – MERRA; the Crocus snow model driven by ERA-Interim meteorology – Crocus); (2) passive microwave remote sensing combined with daily surface snow depth observations (ESA GlobSnow v2.0); and (3) standalone passive microwave retrievals (NASA AMSR-E historical and operational algorithms) which do not utilize surface snow observations. Evaluation included comparisons against independent surface observations from Russia, Finland, and Canada, and calculation of spatial and temporal correlations in SWE anomalies. The standalone passive microwave SWE products (AMSR-E historical and operational SWE algorithms) exhibit low spatial and temporal correlations to other products, and RMSE nearly double the best performing product. Constraining passive microwave retrievals with surface observations (GlobSnow) provides comparable performance to the reanalysis-based products; RMSEs over Finland and Russia for all but the AMSR-E products is ~50 mm or less. Using a four-dataset ensemble that excluded the standalone passive microwave products reduced the RMSE by 10 mm (20%) and increased the correlation by 0.1; ensembles that contain Crocus and/or MERRA perform better than those that do not. The observed RMSE of the best performing datasets is still at the margins of acceptable uncertainty for scientific and operational requirements; only through combined and integrated improvements in remote sensing, modeling, and observations will real progress in SWE product development be achieved.

Sign in / Sign up

Export Citation Format

Share Document