Water Content Of Snowpack Research Articles

Seasonal prediction of end-of-dry-season watershed behavior in a highly interconnected alluvial watershed in northern California

Abstract. In undammed watersheds in Mediterranean climates, the timing and abruptness of the transition from the dry season to the wet season have major implications for aquatic ecosystems. Of particular concern in many coastal areas is whether this transition can provide sufficient flows at the right time to allow passage for spawning anadromous fish, which is determined by dry season baseflow rates and the timing of the onset of the rainy season. In (semi-) ephemeral watershed systems, these functional flows also dictate the timing of full reconnection of the stream system. In this study, we propose methods to predict, approximately 5 months in advance, two key hydrologic metrics in the undammed rural Scott River watershed in northern California. The two metrics are intended to characterize (1) the severity of a dry year and (2) the relative timing of the transition from the dry to the wet season. The ability to predict these metrics in advance could support seasonal adaptive management. The first metric is the minimum 30 d dry season baseflow volume, Vmin, which occurs at the end of the dry season (September–October) in this Mediterranean climate. The second metric is the cumulative precipitation, starting 1 September, necessary to bring the watershed to a “full” or “spilling” condition (i.e., initiate the onset of wet season storm- or baseflows) after the end of the dry season, referred to here as Pspill. As potential predictors of these two metrics, we assess maximum snowpack, cumulative precipitation, the timing of the snowpack and precipitation, spring groundwater levels, spring river flows, reference evapotranspiration, and a subset of these metrics from the previous water year. Though many of these predictors are correlated with the two metrics of interest, we find that the best prediction for both metrics is a linear combination of the maximum snowpack water content and total October–April precipitation. These two linear models could reproduce historical values of Vmin and Pspill with an average model error (RMSE) of 1.4 Mm3 per 30 d (19.4 cfs) and 25.4 mm (1 in.), corresponding to 49 % and 37 % of mean observed values, respectively. Although these predictive indices could be used by governance entities to support local water management, careful consideration of baseline conditions used as a basis for prediction is necessary.

Modeling Spatial Distribution of Snow Water Equivalent by Combining Meteorological and Satellite Data with Lidar Maps

Abstract An accurate characterization of the water content of snowpack, or snow water equivalent (SWE), is necessary to quantify water availability and constrain hydrologic and land surface models. Recently, airborne observations (e.g., lidar) have emerged as a promising method to accurately quantify SWE at high resolutions (scales of ∼100 m and finer). However, the frequency of these observations is very low, typically once or twice per season in the Rocky Mountains of Colorado. Here, we present a machine learning framework that is based on random forests to model temporally sparse lidar-derived SWE, enabling estimation of SWE at unmapped time points. We approximated the physical processes governing snow accumulation and melt as well as snow characteristics by obtaining 15 different variables from gridded estimates of precipitation, temperature, surface reflectance, elevation, and canopy. Results showed that, in the Rocky Mountains of Colorado, our framework is capable of modeling SWE with a higher accuracy when compared with estimates generated by the Snow Data Assimilation System (SNODAS). The mean value of the coefficient of determination R2 using our approach was 0.57, and the root-mean-square error (RMSE) was 13 cm, which was a significant improvement over SNODAS (mean R2 = 0.13; RMSE = 20 cm). We explored the relative importance of the input variables and observed that, at the spatial resolution of 800 m, meteorological variables are more important drivers of predictive accuracy than surface variables that characterize the properties of snow on the ground. This research provides a framework to expand the applicability of lidar-derived SWE to unmapped time points. Significance Statement Snowpack is the main source of freshwater for close to 2 billion people globally and needs to be estimated accurately. Mountainous snowpack is highly variable and is challenging to quantify. Recently, lidar technology has been employed to observe snow in great detail, but it is costly and can only be used sparingly. To counter that, we use machine learning to estimate snowpack when lidar data are not available. We approximate the processes that govern snowpack by incorporating meteorological and satellite data. We found that variables associated with precipitation and temperature have more predictive power than variables that characterize snowpack properties. Our work helps to improve snowpack estimation, which is critical for sustainable management of water resources.

Future projections of High Atlas snowpack and runoff under climate change

Abstract. The High Atlas, culminating at more than 4000 m, is the water tower of Morocco. While plains receive less than 400 mm of precipitation in an average year, the mountains can get twice as much, often in the form of snow between November and March. Snowmelt thus accounts for a large fraction of the river discharge in the region, particularly during spring. In parallel, future climate change projections point towards a significant decline in precipitation and enhanced warming of temperature for the area. Here, we build on previous research results on snow and climate modelling in the High Atlas to make detailed projections of snowpack and river flow response to climate change in this region. We develop end-of-century snowpack projections using a distributed energy balance snow model based on SNOW-17 and high-resolution climate simulations over Morocco with the MIT Regional Climate Model (MRCM) under a mitigation (RCP4.5) and a business-as-usual (RCP8.5) scenario. Snowpack water content is projected to decline by up to 60 % under RCP4.5 and 80 % under RCP8.5 as a consequence of strong warming and drying in the region. We also implement a panel regression framework to relate runoff ratios to regional meteorological conditions in seven small sub-catchments in the High Atlas. Relative humidity and the fraction of solid-to-total precipitation are found to explain about 30 % of the inter-annual variability in runoff ratios. Due to projected future atmospheric drying and the associated decline in snow-to-precipitation ratio, a 5 %–30 % decrease in runoff ratios and 10 %–60 % decrease in precipitation are expected to lead to severe (20 %–70 %) declines in river discharge. Our results have important implications for water resources planning and sustainability of agriculture in this already water-stressed region.

Changes in Vegetation Cover of Yellowstone National Park Estimated from MODIS Greenness Trends, 2000 to 2018

Trends and transitions in the MODerate resolution Imaging Spectroradiometer (MODIS) Normalized Difference Vegetation Index (NDVI) time series at 250-m resolution were analyzed for the period from 2000 to 2018 to understand recent patterns of vegetation change in Yellowstone National Park (USA). Statistical change in the NDVI time series was detected using the “Breaks for Additive Seasonal and Trend” (BFAST) method. This structural change analysis showed that around 30% of abrupt negative shifts in NDVI over the years 2000 to 2018 could be explained by the impacts of recent wildfires. At least one breakpoint could be detected at 12% of the 250-m MODIS pixel locations within the entire YNP study area since the year 2000, but that the majority (about 70%) of NDVI breakpoints detected in vegetation greenness could be not be explained by the impacts of recent wildfires. Positive growing season NDVI trends were detected across most young forests and regrowing (from recent fire disturbance) woodland cover. Evidence further suggested that the 1988 wildfire burns did not pre-dispose vegetation cover in YNP to a higher number of abrupt negative shifts in NDVI since the year 2000. The wildfires of 1988 were associated with significantly higher NDVI recovery trends over the recent 18-year MODIS time series, compared to areas unburned in 1988. Locations on the Northern Range that showed the highest greening trends since 2000 were commonly located in sagebrush steppe-dominated vegetation communities growing at lower than 2500 m elevation. Results from NDVI trend analysis supported the hypothesis that years with relatively high snowpack water content, such as 2007–2008 and 2010–2011, were most closely associated with abrupt negative shifts in NDVI, but no findings strongly supported the supposition that vegetation cover in YNP is changing in greenness in association solely with warming in surface air temperatures or with extreme drought periods across the study region over the past two decades.

Assessing the capability of the SWAT model to simulate snow, snow melt and streamflow dynamics over an alpine watershed

Summary Snow is an important hydrological reservoir within the water cycle, particularly when the watershed includes a mountainous area. Modellers often overlook water stocked in snow pack and its influence on water distribution, especially when only some portions of the watershed is snow dominated. Snow is usually considered to improve hydrological modelling statistics, but without any regard for the realism of its representation or its influence on the hydrological cycle. This is all the more true when semi-distributed models are used, often considered inadequate for spatially representing such phenomena. On the other hand, semi-distributed models are being increasingly used to realise water budget assessment at a regional scale and such studies should not be realised without a good representation of the snow pack. Lack of field measurements is also a frequent justification for avoiding validating simulated snow packs. In this study, remote sensing data provided by MODIS is combined with in situ data, enabling the validation of the snow pack simulated by the Soil and Water Assessment Tool (SWAT), a semi-distributed, physically-based model, implemented over a partly snow-dominated watershed. Snow simulation was performed without complex algorithms or calibration procedures, using the elevation bands option included in the model and related snow parameters. Representation of snow cover and hydrological simulation were achieved by a standard automatic calibration of the model, over the 2000–2010 period, performed by SWAT-Cup/SUFI2, using six hydrological gauging stations along the fluvial continuum downstream of the snow-dominated area. Results highlight three important points: (i) Set-up of elevation bands over mountainous headwater improved hydrological simulation performance, even well downstream of the snow-dominated area. (ii) SWAT produced a good spatial and temporal representation of the snow cover, using MODIS data, despite a slight overestimation at the end of the snow season on the highest elevation bands. A comparison of the model estimate of snowpack water content with in situ data revealed an underestimation in water content in the lower part of the watershed and a slight overestimation in its upper part. Those errors are linked and originate from difficulties of the model to incorporate very local spatial and temporal variations of the precipitation lapse rate. (iii) Elevation bands brought consistent changes in water distribution within the hydrological cycle of implemented watersheds, which are more in line with expected flow paths.

Quantifying Episodic Snowmelt Events in Arctic Ecosystems

Rapid and extensive snowmelt occurred during 2 days in March 2013 at a low-Arctic study site in the ice-free part of southwest Greenland. Meteorology, snowmelt, and snow-property observations were used to identify the meteorological conditions associated with this episodic snowmelt event (ESE) occurring prior to the spring snowmelt season. In addition, outputs from the SnowModel snowpack-evolution tool were used to quantify the snow-related consequences of ESEs on ecosystem-relevant snow properties. We estimated a 50–80% meltwater loss of the pre-melt snowpack water content, a 40–100% loss of snow thermal resistance, and a 4-day earlier spring snowmelt snow-free date due to this March 2013 ESE. Furthermore, the accumulated meltwater loss from all ESEs in a hydrological year represented 25–52% of the annual precipitation and may potentially have advanced spring snowmelt by 6–12 days. Guided by the knowledge gained from the March 2013 ESE, we investigated the origin, past occurrences, frequency, and abundance of ESEs at spatial scales ranging from local (using 2008–2013 meteorological station data) to all of Greenland (using 1979–2013 atmospheric reanalysis data). The frequency of ESEs showed large interannual variation, and a maximum number of ESEs was found in southwest Greenland. The investigations suggested that ESEs are driven by foehn winds that are typical of coastal regions near the Greenland Ice Sheet margin. Therefore, ESEs are a common part of snow-cover dynamics in Greenland and, because of their substantial impact on ecosystem processes, they should be accounted for in snow-related ecosystem and climate-change studies.

Evolution of groundwater age in a mountain watershed over a period of thirteen years

We compile a unique 13-year record of groundwater age for 11 springs in Sagehen basin, a watershed in the Sierra Nevada Mountains of California, USA. Chlorofluorocarbon (CFC), sulfur hexafluoride (SF6), and tritium (3H) data collected in prior studies from 1997 to 2003 are re-interpreted and combined with new data collected in 2009 and 2010. The age record is analyzed to explore the potential value of groundwater age monitoring for (1) providing additional constraints on the age distribution in mixed-age samples, and (2) identifying temporal changes in groundwater recharge. Motivation for this study is provided by a lack of knowledge of how groundwater recharge and discharge (stream baseflow) in mountain watersheds might respond to climate change, and a resulting need to better understand mountain aquifer residence times, storage, and recharge. Piston-flow ages for the springs generally range from 10 to 50yr. The plausibility of different age mixing models is tested by comparing observed temporal variations in age with those simulated using simple numerical models, and by examining plots comparing the concentrations of different age tracers. We find that most spring waters are best characterized by a bimodal mixing model consisting of a new (<1yr old) fraction and a fraction that is older, but still modern (recharged after 1950). Identification of this mixing model would not have been possible without data from multiple age tracers and data from multiple years. Computed mean ages vary substantially (often by 3–7yr) between sampling events for most springs, including those with ages of 20–50yr. Mean age variations are likely controlled by variations in the magnitude of the new fraction, which is positively correlated with annual snowpack water content. Most springs show overall upward trends in mean age for the sampling period, consistent with decreasing recharge rates in response to diminishing snowpack. Groundwater age monitoring appears to be a potentially valuable tool for both constraining the age distribution in sampled waters and investigating recharge transients.

Winter Climatic Controls on Spring Snowpack Density in the Western United States

For numerous climate studies, snowpack density is used to determine snow water equivalent from snow depth (or the reverse) or to determine snow surface albedo through the characterization of aging snow covers. In addition, high spring snowpack water content (and thus density) can act as a catalyst for wet avalanches. Surprisingly, there are few empirical studies that focus on spring snowpack density. In this study, spring snowpack densities in the western United States are statistically related to four variables that characterize the antecedent winter conditions: (1) mean air temperature for days without snowfall, (2) the fraction of precipitation falling as snow, (3) total precipitation, and (4) mean snowfall density. Areal composite regression analysis for the western United States indicates a highly significant (p = 0.005) positive relationship between winter precipitation total and April 1 snowpack density. This relationship weakens in lower elevation regions and coastal regions where warmer winter temperatures are conducive to more frequent rain events and melt events which affect snowpack density and ablate snow cover. These empirical results are supported by a simple snowpack model. The significant positive relationship between precipitation and density is likely due to increased densification rates through gravitational compaction from the presence of greater snow water equivalent resulting from more snowfall.

Realizations of Daily Weather in Forecast Seasonal Climate

Stochastic daily weather time series models (‘‘weather generators’’) are parameterized consistent with both local climate and probabilistic seasonal forecasts. Both single-station weather generators, and spatial networks of coherently operating weather generators, are considered. Only a subset of parameters for individual station models (proportion of wet days, precipitation mean parameters on wet days, and daily temperature means and standard deviations) are found to depend appreciably on the seasonal temperature and precipitation outcomes, so that extension of the single-station models to coherent multisite weather generators is straightforward. The result allows stochastic simulation of multiple daily weather series, conditional on seasonal forecasts. Example applications of spatially integrated extreme daily precipitation and snowpack water content are used to illustrate the method.

Estimating the Uncertainty in Spatial Estimates of Areal Snow Water Equivalent

With the increased demand for water in the United States, particularly in the West, it is essential that water resources be accurately monitored. Consequently, the National Weather Service (NWS) maintains a set of conceptual, continuous, hydrologic simulation models used to generate extended streamflow predictions, water supply outlooks, and flood forecasts. A vital component of the hydrologic simulation models is a snow accumulation and ablation model that uses observed temperature and precipitation date to simulate snow cover conditions. The simulated model states are updated throughout the snow season using snow water equivalent estimates (estimates of the water content of snowpack) obtained from airborne and ground-based snow water equivalent data. The National Weather Service has developed a spatial geostatistical model to estimate the areal snow water equivalent in a river basin. The estimates, which are obtained for river basins throughout the West, are used to update the snow model. To facilitate accurate updating of the simulated snow water equivalent estimates generated by the snow model, it is necessary to incorporate measures of uncertainty of the areal snow water equivalent estimates. In this research, we derive the expression for the mean-squared prediction error of the areal snow water equivalent estimate and illustrate the methodology with an example from the Upper Colorado River basin.

Modeling measurement errors when estimating snow water equivalent

Recently the National Weather Service has developed a geostatistical model to generate areal estimates of snow water equivalent (the water content of snowpack) in various river basins throughout the Western USA. The areal snow water equivalent estimates are used to update the hydrologic simulation models used to produce extended streamflow forecasts in the USA. Currently, when the snow water equivalent estimates are obtained, it is assumed that the ground-based and airborne data used in the spatial model are measured without error. In this research, we illustrate how to modify the spatial prediction model to account for measurement errors in the observed airborne and ground-based snow data, derive the expressions necessary to model the measurement errors, and illustrate the modeling techniques on data collected in the Upper Colorado River basin.

Debris Flows and Hyperconcentrated Floods Along the Wasatch Front, Utah, 1983 and 1984

The Wasatch Front north of Salt Lake City has a recurrent history of debris flows. Two distinctly different types of climatic events—summer thunderstorms and rapid spring snowmelts—have triggered debris flows. Heavy autumn rains, exceptional winter snowpacks, and cool and extended springs followed by sudden warming characterized weather during 1983 and 1984. As a result of these climatic events, large amounts of water were rapidly added to the landscape, resulting in several hundred landslides, many of which subsequently mobilized into debris flows and (or) hyperconcentrated floods. Reconnaissance during June of 1983 revealed many hillsides with recently active landslides that had moved slightly but had not mobilized, or had only partially mobilized, as debris flows. The presence of these partly-detached landslides prompted an investigation of the potential for future debris flows. Wieczorek and others (1983) developed a technique for evaluating the potential for debris flows and hyperconcentrated floods to reach the canyon mouths, based on three assumptions: 1) debris flows would occur in response to a large influx of water to the soils, either from a thunderstorm or rapid snowmelt; 2) the partly-detached landslides were the most likely sources for future debris flows; and 3) debris-flow material incorporated from channels would be proportional to that from landslide sources. The climatic and debris-flow events of 1984 were sufficiently similar to those of 1983 that we could evaluate these assumptions. Slightly fewer debris flows occurred in 1984 than in 1983, consistent with a slightly lower snow-pack water content and slightly slower rate of melting. Debris flows occurred near the rapidly receding snowline, apparently as a result of the large influx of water into the soils there. On an areal basis, partly-detached landslides identified in 1983 were statistically significant as sources for debris flows in 1984. In the few documented cases, the contribution of material scoured from channels was variable but very high and appeared significantly influenced by the geomorphic condition of the stream channel.

Composition of snowmelt and runoff in northern Michigan

Snowmelt and runoff were studied during the 1982-1983 and the 1983-1984 winters at the University of Michigan Biological Station, which is located near the northern tip of Michigan's lower peninsula. The first 50% of the snowpack acidity was released in meltwater and rainwater equal to 25% of the original snowpack water content. Interaction between the meltwater and the litter layer produced large changes in the concentrations of most species. Runoff to two streams had high SO/sub 4//sup 2 -/ and very low NO/sub 3//sup -/ concentrations. It is concluded that most of the NO/sub 3//sup -/ is either biologically utilized or retained in the watershed, even during the early snowmelt period at this site. 18 references, 5 figures, 3 tables.

HYDROMETEOROLOGICAL SIGNIFICANCE OF RIME ICE DEPOSITS IN THE COLORADO ROCKIES1

ABSTRACT Observations and measurements were made during storm periods at mountain top sites in the Colorado Rockies during the winter of 1981‐1982. On average, liquid clouds occurred with and without snowfall for, respectively, 75 percent and 15 percent of the observations. On average, the rime ice deposition rates were measured to be seven times the precipitation rates. The rime deposits were estimated to contribute 10 percent to the water content of the snowpack. Combining this figure with the estimated rime deposits on falling snow crystals (up to 50 percent by mass), up to 60 percent of the water content of the annual snowpack in the mountains near Steamboat Springs, Colorado, could be due to rime ice deposits.