Application of machine learning techniques for regional bias correction of snow water equivalent estimates in Ontario, Canada

King, Fraser; Erler, Andre R.; Frey, Steven K.; Fletcher, Christopher G.

doi:10.5194/hess-24-4887-2020

Cited by 24 publications

(16 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our models are being used for investigation of sub-grid SWE variability in E3SM's ELM (Caldwell et al, 2019;Bisht et al, 2018), along with our investigation into ecosystem-type approaches for upscaling of SWE. The patterns of our SWE maps illustrate the power of utilizing random forest tools over linear methods of 525 estimating SWE distributions (e.g., Broxton et al, 2019, Revuelto et al 2020, King et al 2020. When compared to linear and GAM models, we found that random forests significantly outperformed those models.…”

Section: Swe Modelling and Predictionmentioning

confidence: 70%

See 1 more Smart Citation

Spatial Patterns of Snow Distribution for Improved Earth System Modelling in the Arctic

Bennett

Miller

Busey

et al. 2021

Preprint

View full text Add to dashboard Cite

Abstract. The spatial distribution of snow plays a vital role in Arctic climate, hydrology, and ecology due to its fundamental influence on the water balance, thermal regimes, vegetation, and carbon flux. However, for earth system modelling, the spatial distribution of snow is not well understood, and therefore, it is not well modeled, which can lead to substantial uncertainties in snow cover representations. To capture key hydro-ecological controls on snow spatial distribution, we carried out intensive field studies over multiple years for two small (2017–2019, ~2.5 km2) sub-Arctic study sites located on the Seward Peninsula of Alaska. Using an intensive suite of field observations (> 22,000 data points), we developed simple models of spatial distribution of snow water equivalent (SWE) using factors such as topographic characteristics, vegetation characteristics based on greenness (normalized different vegetation index, NDVI), and a simple metric for approximating winds. The most successful model was the random forest using both study sites and all years, which was able to accurately capture the complexity and variability of snow characteristics across the sites. Approximately 86 % of the SWE distribution could be accounted for, on average, by the random forest model at the study sites. Factors that impacted year-to-year snow distribution included NDVI, elevation, and a metric to represent coarse microtopography (topographic position index, or TPI), while slope, wind, and fine microtopography factors were less important. The models were used to predict SWE at the locations through the study area and for all years. The characterization of the SWE spatial distribution patterns and the statistical relationships developed between SWE and its impacting factors will be used for the improvement of snow distribution modelling in the Department of Energy’s earth system model, and to improve understanding of hydrology, topography, and vegetation dynamics in the Arctic and sub-Arctic regions of the globe.

show abstract

Section: Swe Modelling and Predictionmentioning

confidence: 70%

Section: )mentioning

confidence: 99%

Spatial Patterns of Snow Distribution for Improved Earth System Modelling in the Arctic

Bennett

Miller

Busey

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The patterns of our SWE maps illustrate the power of utilizing random forest tools over linear methods of estimating SWE distributions (e.g., Broxton et al, 2019;Revuelto et al, 2020;King et al, 2020). When compared to linear and GAM models, we found that random forests significantly outperformed those models.…”

Section: Swe Modeling and Predictionmentioning

confidence: 70%

“…Revuelto et al (2020) used random forests to predict lidar snow depth distribution from several topographic predictors. King et al (2020) used random forests for bias correction of a SWE data assimilation product. Other studies have applied machine learning algorithms using remotely sensed observations as predictors, including brightness temperature, fractional snow-covered area, or the normalized difference snow index (Liu et al, 2020;Bair et al, 2018).…”

Section: Introductionmentioning

confidence: 99%

Spatial patterns of snow distribution in the sub-Arctic

et al. 2022

View full text Add to dashboard Cite

Abstract. The spatial distribution of snow plays a vital role in sub-Arctic and Arctic climate, hydrology, and ecology due to its fundamental influence on the water balance, thermal regimes, vegetation, and carbon flux. However, the spatial distribution of snow is not well understood, and therefore, it is not well modeled, which can lead to substantial uncertainties in snow cover representations. To capture key hydro-ecological controls on snow spatial distribution, we carried out intensive field studies over multiple years for two small (2017–2019; ∼ 2.5 km2) sub-Arctic study sites located on the Seward Peninsula of Alaska. Using an intensive suite of field observations (> 22 000 data points), we developed simple models of the spatial distribution of snow water equivalent (SWE) using factors such as topographic characteristics, vegetation characteristics based on greenness (normalized different vegetation index, NDVI), and a simple metric for approximating winds. The most successful model was random forest, using both study sites and all years, which was able to accurately capture the complexity and variability of snow characteristics across the sites. Approximately 86 % of the SWE distribution could be accounted for, on average, by the random forest model at the study sites. Factors that impacted year-to-year snow distribution included NDVI, elevation, and a metric to represent coarse microtopography (topographic position index, TPI), while slope, wind, and fine microtopography factors were less important. The characterization of the SWE spatial distribution patterns will be used to validate and improve snow distribution modeling in the Department of Energy's Earth system model and for improved understanding of hydrology, topography, and vegetation dynamics in the sub-Arctic and Arctic regions of the globe.

show abstract

“…Yang et al [34] first used Random Forest (RF) to derive a long time series of a snow depth product that was more precise than the Che algorithm output [38]. RF was the most effective at reducing bias in SNOw Data Assimilation System (SNODAS) SWE in Ontario, Canada, with an absolute mean bias of 0.2 mm and RMSE of 3.64 mm when compared with in situ observations [39].…”

Section: Introductionmentioning

confidence: 99%

Snow Depth Fusion Based on Machine Learning Methods for the Northern Hemisphere

Che

Dai

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

In this study, a machine learning algorithm was introduced to fuse gridded snow depth datasets. The input variables of the machine learning method included geolocation (latitude and longitude), topographic data (elevation), gridded snow depth datasets and in situ observations. A total of 29,565 in situ observations were used to train and optimize the machine learning algorithm. A total of five gridded snow depth datasets—Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) snow depth, Global Snow Monitoring for Climate Research (GlobSnow) snow depth, Long time series of daily snow depth over the Northern Hemisphere (NHSD) snow depth, ERA-Interim snow depth and Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2) snow depth—were used as input variables. The first three snow depth datasets are retrieved from passive microwave brightness temperature or assimilation with in situ observations, while the last two are snow depth datasets obtained from meteorological reanalysis data with a land surface model and data assimilation system. Then, three machine learning methods, i.e., Artificial Neural Networks (ANN), Support Vector Regression (SVR), and Random Forest Regression (RFR), were used to produce a fused snow depth dataset from 2002 to 2004. The RFR model performed best and was thus used to produce a new snow depth product from the fusion of the five snow depth datasets and auxiliary data over the Northern Hemisphere from 2002 to 2011. The fused snow-depth product was verified at five well-known snow observation sites. The R2 of Sodankylä, Old Aspen, and Reynolds Mountains East were 0.88, 0.69, and 0.63, respectively. At the Swamp Angel Study Plot and Weissfluhjoch observation sites, which have an average snow depth exceeding 200 cm, the fused snow depth did not perform well. The spatial patterns of the average snow depth were analyzed seasonally, and the average snow depths of autumn, winter, and spring were 5.7, 25.8, and 21.5 cm, respectively. In the future, random forest regression will be used to produce a long time series of a fused snow depth dataset over the Northern Hemisphere or other specific regions.

show abstract

Application of machine learning techniques for regional bias correction of snow water equivalent estimates in Ontario, Canada

Cited by 24 publications

References 45 publications

Spatial Patterns of Snow Distribution for Improved Earth System Modelling in the Arctic

Spatial Patterns of Snow Distribution for Improved Earth System Modelling in the Arctic

Spatial patterns of snow distribution in the sub-Arctic

Snow Depth Fusion Based on Machine Learning Methods for the Northern Hemisphere

Contact Info

Product

Resources

About