Inductive machine learning for improved estimation of catchment-scale snow water equivalent

2015 ◽  
Vol 524 ◽  
pp. 311-325 ◽  
Author(s):  
David Buckingham ◽  
Christian Skalka ◽  
Josh Bongard
Keyword(s):  
Machine Learning ◽  
Snow Water Equivalent ◽  
Catchment Scale ◽  
Inductive Machine Learning ◽  
Snow Water

Catchment-scale evaluation of pollution potential of urban snow at two residential catchments in southern Finland

2013 ◽  
Vol 68 (10) ◽  
pp. 2164-2170 ◽  
Author(s):  
Nora Sillanpää ◽  
Harri Koivusalo
Keyword(s):  
Snow Water Equivalent ◽  
Hydrological Cycle ◽  
Cold Climate ◽  
Road Maintenance ◽  
Catchment Scale ◽  
Small Surface ◽  
Climate Conditions ◽  
Snow Water ◽  
High Level ◽  
Mass Loads

Despite the crucial role of snow in the hydrological cycle in cold climate conditions, monitoring studies of urban snow quality often lack discussions about the relevance of snow in the catchment-scale runoff management. In this study, measurements of snow quality were conducted at two residential catchments in Espoo, Finland, simultaneously with continuous runoff measurements. The results of the snow quality were used to produce catchment-scale estimates of areal snow mass loads (SML). Based on the results, urbanization reduced areal snow water equivalent but increased pollutant accumulation in snow: SMLs in a medium-density residential catchment were two- to four-fold higher in comparison with a low-density residential catchment. The main sources of pollutants were related to vehicular traffic and road maintenance, but also pet excrement increased concentrations to a high level. Ploughed snow can contain 50% of the areal pollutant mass stored in snow despite its small surface area within a catchment.

Snowpack energy balance and changes in the frequency and extremity of rain-on-snow events in the warming climate

2021 ◽  
Author(s):  
Ondrej Hotovy ◽  
Michal Jenicek
Keyword(s):  
Energy Balance ◽  
Air Temperature ◽  
Snow Water Equivalent ◽  
Hydrological Cycle ◽  
Winter Season ◽  
Catchment Scale ◽  
The Future ◽  
Snow Water ◽  
The Impact ◽  
Catchment Runoff

<p>Seasonal snowpack significantly influences the catchment runoff and thus represents an important input for the hydrological cycle. Changes in the precipitation distribution and intensity, as well as a shift from snowfall to rain is expected in the future due to climate changes. As a result, rain-on-snow events, which are considered to be one of the main causes of floods in winter and spring, may occur more frequently. Heat from liquid precipitation constitutes one of the snowpack energy balance components. Consequently, snowmelt and runoff may be strongly affected by these temperature and precipitation changes.</p><p>The objective of this study is 1) to evaluate the frequency, inter-annual variability and extremity of rain-on-snow events in the past based on existing measurements together with an analysis of changes in the snowpack energy balance, and 2) to simulate the effect of predicted increase in air temperature on the occurrence of rain-on-snow events in the future. We selected 40 near-natural mountain catchments in Czechia with significant snow influence on runoff and with available long-time series (>35 years) of daily hydrological and meteorological variables. A semi-distributed conceptual model, HBV-light, was used to simulate the individual components of the water cycle at a catchment scale. The model was calibrated for each of study catchments by using 100 calibration trials which resulted in respective number of optimized parameter sets. The model performance was evaluated against observed runoff and snow water equivalent. Rain-on-snow events definition by threshold values for air temperature, snow depth, rain intensity and snow water equivalent decrease allowed us to analyze inter-annual variations and trends in rain-on-snow events during the study period 1965-2019 and to explain the role of different catchment attributes.</p><p>The preliminary results show that a significant change of rain-on-snow events related to increasing air temperature is not clearly evident. Since both air temperature and elevation seem to be an important rain-on-snow drivers, there is an increasing rain-on-snow events occurrence during winter season due to a decrease in snowfall fraction. In contrast, a decrease in total number of events was observed due to the shortening of the period with existing snow cover on the ground. Modelling approach also opened further questions related to model structure and parameterization, specifically how individual model procedures and parameters represent the real natural processes. To understand potential model artefacts might be important when using HBV or similar bucket-type models for impact studies, such as modelling the impact of climate change on catchment runoff.</p>

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

2020 ◽  
Vol 24 (10) ◽  
pp. 4887-4902
Author(s):  
Fraser King ◽  
Andre R. Erler ◽  
Steven K. Frey ◽  
Christopher G. Fletcher
Keyword(s):  
Machine Learning ◽  
Linear Regression ◽  
Water Resource Management ◽  
Snow Water Equivalent ◽  
Bias Reduction ◽  
Machine Learning Techniques ◽  
Spatial Coverage ◽  
Snow Water ◽  
In Situ Observations

Abstract. Snow is a critical contributor to Ontario's water-energy budget, with impacts on water resource management and flood forecasting. Snow water equivalent (SWE) describes the amount of water stored in a snowpack and is important in deriving estimates of snowmelt. However, only a limited number of sparsely distributed snow survey sites (n=383) exist throughout Ontario. The SNOw Data Assimilation System (SNODAS) is a daily, 1 km gridded SWE product that provides uniform spatial coverage across this region; however, we show here that SWE estimates from SNODAS display a strong positive mean bias of 50 % (16 mm SWE) when compared to in situ observations from 2011 to 2018. This study evaluates multiple statistical techniques of varying complexity, including simple subtraction, linear regression and machine learning methods to bias-correct SNODAS SWE estimates using absolute mean bias and RMSE as evaluation criteria. Results show that the random forest (RF) algorithm is most effective at reducing bias in SNODAS SWE, with an absolute mean bias of 0.2 mm and RMSE of 3.64 mm when compared with in situ observations. Other methods, such as mean bias subtraction and linear regression, are somewhat effective at bias reduction; however, only the RF method captures the nonlinearity in the bias and its interannual variability. Applying the RF model to the full spatio-temporal domain shows that the SWE bias is largest before 2015, during the spring melt period, north of 44.5∘ N and east (downwind) of the Great Lakes. As an independent validation, we also compare estimated snowmelt volumes with observed hydrographs and demonstrate that uncorrected SNODAS SWE is associated with unrealistically large volumes at the time of the spring freshet, while bias-corrected SWE values are highly consistent with observed discharge volumes.

scholarly journals Improving Snow Water Equivalent Maps With Machine Learning of Snow Survey and Lidar Measurements

2019 ◽  
Vol 55 (5) ◽  
pp. 3739-3757 ◽  
Author(s):  
Patrick D. Broxton ◽  
Willem J. D. Leeuwen ◽  
Joel A. Biederman
Keyword(s):  
Machine Learning ◽  
Snow Water Equivalent ◽  
Lidar Measurements ◽  
Snow Water

Changes in snow storages and their impact on occurrence and extremity of runoff caused by rain-on-snow events

2020 ◽  
Author(s):  
Ondrej Hotovy ◽  
Michal Jenicek
Keyword(s):  
Air Temperature ◽  
Snow Water Equivalent ◽  
Hydrological Cycle ◽  
Winter Season ◽  
Model Performance ◽  
Catchment Scale ◽  
The Future ◽  
Snow Water ◽  
The Impact ◽  
Catchment Runoff

<p>Seasonal snowpack significantly influences the catchment runoff and thus represents an important input for the hydrological cycle. Changes in the precipitation distribution and intensity, as well as a shift from snowfall to rain is expected in the future due to climate changes. As a result, rain-on-snow events, which are considered to be one of the main causes of floods in winter and spring, may occur more frequently.</p><p>The objective of this study is 1) to evaluate the frequency, inter-annual variability and extremity of rain-on-snow events in the past based on existing measurements and 2) to simulate the effect of predicted increase in air temperature on the occurrence of rain-on-snow events in the future. We selected 59 near-natural mountain catchments in Czechia with significant snow influence on runoff and with available long-time series (>35 years) of daily hydrological and meteorological variables. A semi-distributed conceptual model, HBV-light, was used to simulate the individual components of the water cycle at a catchment scale. The model was calibrated for each of study catchments by using 100 calibration trials which resulted in respective number of optimized parameter sets. The model performance was evaluated against observed runoff and snow water equivalent. Rain-on-snow events definition by threshold values for air temperature, snow depth, rain intensity and snow water equivalent decrease allowed us to analyze inter-annual variations and trends in rain-on-snow events during the study period 1980-2014 and to explain the role of different catchment attributes.</p><p>The preliminary results show that a significant change of rain-on-snow events related to increasing air temperature is not clearly evident. Since both air temperature and elevation seem to be an important rain-on-snow drivers, there is an increasing rain-on-snow events occurrence during winter season due to a decrease in snowfall fraction. In contrast, a decrease in total number of events was observed due to the shortening of the period with existing snow cover on the ground. Modelling approach also opened further questions related to model structure and parameterization, specifically how individual model procedures and parameters represent the real natural processes. To understand potential model artefacts might be important when using HBV or similar bucket-type models for impact studies, such as modelling the impact of climate change on catchment runoff.</p>

scholarly journals Modelling the spatial variability of snow water equivalent at the catchment scale

2007 ◽  
Vol 11 (5) ◽  
pp. 1543-1550 ◽  
Author(s):  
T. Skaugen
Keyword(s):  
Spatial Distribution ◽  
Spatial Variability ◽  
Gamma Distribution ◽  
Snow Water Equivalent ◽  
Temporal Correlation ◽  
Catchment Scale ◽  
Snow Season ◽  
Dynamical Parameters ◽  
Snow Water ◽  
Alpine Catchments

Abstract. The spatial distribution of snow water equivalent (SWE) is modelled as a two parameter gamma distribution. The parameters of the distribution are dynamical in that they are functions of the number of accumulation and melting events and the temporal correlation of accumulation and melting events. The estimated spatial variability is compared to snow course observations from the alpine catchments Norefjell and Aursunden in Southern Norway. A fixed snow course at Norefjell was measured 26 times during the snow season and showed that the spatial coefficient of variation change during the snow season with a decreasing trend from the start of the accumulation period and a sharp increase in the melting period. The gamma distribution with dynamical parameters reproduced the observed spatial statistical features of SWE well both at Norefjell and Aursunden. Also the shape of simulated spatial distribution of SWE agreed well with the observed at Norefjell. The temporal correlation tends to be positive for both accumulation and melting events. However, at the start of melting, a better fit between modelled and observed spatial standard deviation of SWE is obtained by using negative correlation between SWE and melt.

scholarly journals Investigating the development of shallow snowpacks on arable land, using comprehensive field observations and spatially distributed snow modelling

2017 ◽  
Vol 49 (1) ◽  
pp. 41-59
Author(s):  
Torsten Starkloff ◽  
Jannes Stolte ◽  
Rudi Hessel ◽  
Coen Ritsema
Keyword(s):  
Overland Flow ◽  
Snow Water Equivalent ◽  
Arable Land ◽  
Model Performance ◽  
Field Observations ◽  
Catchment Scale ◽  
Spatially Distributed ◽  
Snow Water ◽  
Southern Norway ◽  
Agricultural Areas

Abstract Shallow (<1 m deep) snowpacks on agricultural areas are an important hydrological component in many countries, which determines how much meltwater is potentially available for overland flow, causing soil erosion and flooding at the end of winter. Therefore, it is important to understand the development of shallow snowpacks in a spatially distributed manner. This study combined field observations with spatially distributed snow modelling using the UEBGrid model, for three consecutive winters (2013–2015) in southern Norway. Model performance was evaluated by comparing the spatially distributed snow water equivalent (SWE) measurements over time with the simulated SWE. UEBGrid replicated SWE development at catchment scale with satisfactory accuracy for the three winters. The different calibration approaches which were necessary for winters 2013 and 2015 showed the delicacy of modelling the change in shallow snowpacks. Especially the refreezing of meltwater and prevention of runoff and infiltration of meltwater by frozen soils and ice layers can make simulations of shallow snowpacks challenging.

Sign in / Sign up

Export Citation Format

Share Document