C21B-0444
Predicting Soil Frost and its Response to Climate Change in Northeastern U.S. Forests
Depth and duration of seasonal snow cover has important effects on temperate forest ecosystems. In the northeastern U.S., recent predictions are that climate warming over the coming century will cause an increase in soil freezing as soils lose the insulation of continuous wintertime snow cover. These studies have also linked soil freezing to elevated nitrate export from soils and streams. In the present study, we used a physically based energy and water exchange model, SHAW (Simultaneous Heat and Water), to predict soil frost and snowpack dynamics at three forested sites in New England: Hubbard Brook (NH), Harvard Forest (MA), and Howland Forest (ME). Results indicate an inverse relationship across all three sites between the depth and duration of the snowpack and soil frost. Simulations were conducted for all three sites with historical weather data for the past 20-40 years, and for future projections (2000-2100) using two different IPCC climate scenarios (A1fi and BI) derived from statistically downscaled GCM simulations. Under both scenarios and at all three sites, SHAW predicted that both the amount of soil frost and the number of extreme soil freezing events will decrease during the 2000-2100 period. In addition, there was no relationship between predicted soil frost, 1966-2000, and observed stream nitrate concentration at Hubbard Brook. These results run counter to existing theories regarding both the impacts of soil frost and the changes that are expected to occur into the future. There was, however, a positive correlation between predicted soil frost and growing season CO2 uptake at Harvard Forest over the 1992-2002 period. This suggests that soil freezing does play an important role in forest biogeochemistry, albeit a different role than that which has been discussed in the literature.
C21B-0445
Subniveal Carbon Flux in a Wyoming Subalpine Ecosystem
Winter soil biological activity may significantly impact annual carbon balance and post-winter processes in subalpine ecosystems. Soil microbial communities are generally active in the subniveal environment due to insulation from the snowpack, allowing winter soil temperatures to remain above freezing. Snowpack characteristics such as snow density and temperature can vary over a season, potentially impacting subniveal processes such as soil respiration because of altered insulation properties. Little is known about the magnitude and variation in subniveal soil CO2 production (due to root and microbial respiration) in subalpine ecosystems. We estimated subniveal CO2 production over a winter season with variable snowpack conditions by sampling forested and open meadow areas in a subalpine ecosystem in the Snowy Range of Wyoming in February (winter) and April (spring). We measured CO2 concentration, oxygen (18O) and carbon (13C) isotope ratios of CO2, snow density, and snow temperature at multiple positions within the snow profile. These data were integrated with a dynamic CO2 diffusion model that incorporates the effects of snow pack properties on CO2 diffusion. The data and model were coupled to (1) partition the contribution of subniveal vs. atmospheric sources or CO2 to total CO2 efflux at the snow surface and (2) estimate subniveal (i.e., soil-derived, biological) CO2 production rates. We evaluated whether the 18O signature of snow and soil water could trace the biological CO2 source. Within the meadow sites, respiration rates were similar between the winter and spring periods. Within the forested sites, soil respiration was higher in the spring compared to the winter. Differences in soil respiration could be due to (1) differences in subniveal CO2 production rates, (2) differences in CO2 diffusion created by changing snowpack conditions, and/or (3) changes in the diffusion or "pumping" of atmospheric CO2 into the snow profile. Yet, the 13C signal was more depleted in the spring, suggesting that spring subniveal CO2 production may be more affected by root respiration and/or microbial substrate use may change. The dual isotope approach (18O and 13C in CO2) was useful for determining source contribution in the winter but less so in the spring because freeze-thaw dynamics of spring snowpack may cause significant exchange of 18O between CO2 and H2O in the snow profile. Since little/no primary productivity occurs during the winter, subniveal soil activity could greatly affect the annual carbon balance of such systems by increasing soil carbon loss, but the degree to which it controls annual, net ecosystem exchange is largely determined by variation in snow properties. Integrating the biological and physical components of winter-driven systems may be critical in determining how changes in winter climate will impact these ecosystems at larger spatial and temporal scales.
C21B-0446
Pervasive coupling between interglacial CO2 concentrations, temperature, and sea level
There is a pressing need to better constrain the impact of anthropogenic climate forcing on global sea level [IPCC, 2007]. We push the benthic isotope record to derive sea level during all interglacial sea-level highstand periods of the last 650 000 years. We demonstrate that a pervasive coupling exists between sea level, atmospheric CO2 concentrations, and Antarctic temperature. This relationship is best explained in terms of a combination of insolation variation and greenhouse-gas concentrations.
C21B-0447
Forest Cover and Topographic Influences on Snow Distribution in a Mixed Hardwood-Conifer Forest of the Northeastern U.S.
Forested landscapes of the northeastern U.S. face increasing pressure from development and recreational uses. Changes in forest structure and the distribution of canopy openings may have measurable impacts on hydrology, particularly in high elevation terrain where gradients in atmospheric inputs are great. Here, we report findings of a pilot study conducted in spring 2007 to examine the effects of forest stand and canopy structure, canopy openings and topography on snow distribution. Our sampling sites are located within on-going studies of canopy development following silvicultural treatments and forest clearings from recreational development (i.e. alpine skiing) in the mixed hardwood-conifer forests of northwestern Vermont. Our findings indicate that snow water equivalent (SWE) was significantly related to key forest metrics including conifer abundance and mean diameter at breast height (dbh). SWE exhibited strong elevational trends over three sampling dates and was more spatially variable on south-facing slopes than on all other aspects. Comparisons between forested and open sampling sites showed no differences in SWE for early (DOY 60 and 68) sampling dates, but differences were significant for later sampling dates (DOY 71 and 82) at high elevation sites. Along ski trail clearings, surveys using ground-penetrating (GPR) radar showed significant differences in SWE for trails covered with man- made snow, relative to those covered with natural snow. Collectively, our findings suggest that (1) mixed forests of the Northeast have limited ability to influence snowpacks through interception when some conifer component exists in the stand, (2) these forests have detectable effects on snowmelt dynamics relative to clearings, (3) topography exerts strong controls on snow distribution, and (4) man-made snow on recreational ski trails introduces an additional layer of variability in snowpack distribution. We speculate on the hydrologic consequences of these findings using a simple, GIS-based model incorporating snow distributing and routing.
C21B-0448
Quantifying the Effects of Storm Track, Topography, and Vegetation on Chemical Loading to a Montane Snowpack, Valles Caldera National Preserve, NM
The spatial variability of chemicals deposited in seasonal snow cover creates difficulty in estimating input of potentially important sources of nutrients and pollutants into terrestrial and aquatic environments. This study quantifies how vegetation, aspect, storm track, and event characteristics control the spatial and temporal differences in snow chemistry and chemical load to a montane snowpack in the Valles Caldera National Preserve, northern New Mexico. This work was conducted in coordination with two simultaneous studies, the first addressing spatial differences in snow accumulation and ablation due to vegetation cover and the second examining the influence of aspect on the transit time of melt water through our study catchments. Building on this coordinated effort, this project will also link broader implications on the role of aspect and vegetation cover through chemical and isotopic analysis. Based on study design, we collected depth, density, stratigraphy and snow chemistry samples from six snow pit locations on approximate monthly intervals between January and April 2007. Snow chemistry samples were analyzed for major anions (Cl-, NO3-, SO42-), major cations (Ca2+, Na+, K+), water isotopes, and biogeochemical nutrients (DOC, DN). Initial analysis of anion data suggests that estimating anion loads with bulk snow samples versus layer distributed samples indicates bulk samples underestimate anion load at peak accumulation (p < 0.05). Coefficients of variation (CV) are used to analyze chemical load variability. CV's of anion loads vary spatially from 27% (Cl-) to 33% (NO3-) and for nutrient loads from 34% (DN) to 39% (DOC) at peak accumulation. Dense vegetation cover appears to strongly influence dissolved organic carbon but has no apparent affect on other solutes suggesting that the vegetation is the source of the increased DOC load rather than enhanced atmospheric deposition on trees. Preliminary storm track analysis indicates distinct NO3-/ SO42- ratios are dependant on depositional wind direction. Storm deposits with westerly components appear to yield higher SO42- concentrations producing lower ratios.
C21B-0449
Ecohydrological controls on snowmelt partitioning in a mixed-conifer sub-alpine forest, Valles Caldera, New Mexico
The processes controlling the partitioning of snowmelt into the various hydrologic pathways are largely not understood. This knowledge gap and the complexity of interactions between the snowpack, vegetation, and the vadose zone water balance motivate comprehensive studies of the terrestrial water balance at hillslope to catchment scales. Instrument clusters, deployed along elevational transects in the Valles Caldera of New Mexico, are being used to study the distribution of rainfall, snowmelt, and soil moisture and associated influences on transpiration rates during the growing season. To fully understand the feedbacks between vegetation structure and water availability, a stratified sampling scheme is being used to evaluate micro-scale gradients in land-surface / atmosphere energy exchange and associated impacts on the distribution of snowmelt and soil moisture. Instrument clusters deployed at multiple elevations are being used to improve understanding of the sensitivity of these feedbacks to climate variability and associated changes in montane snow and vegetation distribution. Here, the potential for increased fire severity, decreased water availability, and changes to the terrestrial carbon budget drive the need to understand these feedbacks. Transects of co-located ultra-sonic snow depth sensors, water content reflectometers, soil thermistors, and sap flow sensors revealed that interception of snowfall by vegetation created substantial variability in snow depth; snow depth at under canopy sites was approximately 40% lower than open areas before snowmelt. Variability in melt fluxes at the sub-meter scale resulted in lower melt rates at under canopy sites as indicated by the persistence of snow at these locations despite the lower accumulation. Despite the lower snow accumulation in the sub-canopy environment, winter soil temperatures remained above freezing. The onset of snowmelt infiltration followed peak snow accumulation by 8 days while vegetation response to water availability in the sub-nivean soil was abrupt, with rapid increases in sap flow immediately following the onset of snowmelt infiltration. Co-located observations of CO2 and H2O vapor flux are also being used to explore the response of vegetation to water inputs and water related stress. http://www.sahra.arizona.edu/valles/
C21B-0450
Long-term changes in snow water equivalent and streamflow in northern Idaho
Concern regarding the hydrologic impacts of climate warming in snow-dominated systems emphasizes the need for long-term analyses of historical snowpack and streamflow data. Several studies have indicated a decline in peak snow water equivalent (SWE), advance of seasonal meltout dates, and corresponding temporal advance in snowmelt-dominated hydrographs in the inland northwest of the U. S. Analyses to date focused primarily on snowpack records less than 50 years and on streamflow records from relatively large basins. There is a need to assess trends over longer time periods and in smaller catchments to advance our understanding of the historical hydrologic trends across a wider range of systems. An analysis of SWE and headwater streamflow trends at two USFS experimental forests in northern Idaho is presented for data collected over the past 70 to 95 years. Both annual and winter precipitation values exhibited a slightly increasing, but non-significant trend. SWE trends indicated a consistent decline for most of the months of February through May. Statistically significant rates of decline ranged from 0.7 to 3.8 mm yr-1. Over the period of record, these declines have resulted in approximately 75% to 50% reductions in SWE at all sites during the melt season. Since the 1950's, annual streamflow at one of the sites showed no significant trend, however the timing of the 50th percentile flow has advanced approximately 12 days, which is comparable to larger basins in the area.
C21B-0451
Historical Analysis of Remotely Sensed Snow Properties' Relation to Streamflow
In snowmelt dominated river basins, snow properties near peak accumulation are used to assess spring and summer runoff. Forecast models rely on estimates of the water stored in the snowpack to determine the contribution of snowmelt to runoff. We can measure snow-covered area and albedo from available satellite sources such as MODIS, which provides daily imagery at 500 m spatial resolution, and Landsat, which provides imagery every 16 days at 30 m. In California's Sierra Nevada, daily snow pillow measurements of snow water equivalent are the main source of in situ information. We investigate the relationship of snow-covered area to snow water equivalent in Sierra Nevada watersheds with varying latitudes, orientations, and elevations. Based on monthly unimpaired runoff volumes, we selected a set of years during the Landsat historical record (1985-2007) that encompass 80% of the range of variability in runoff during the last century. We use multiple endmember spectral unmixing to estimate the fraction of snow in each 30 m pixel and the albedo of that fractional snow cover for the American, San Joaquin and Kern watersheds for five years that represent the minimum, quartiles and maximum April, May, and June unimpaired runoff. Recent years have produced similar variability in runoff, and daily fractional snow cover is estimated from MODIS for these years at 500 m resolution. In addition to snow- covered area and albedo, we estimate the spatial distribution of snow water equivalent by blending a hypsometric method with snow-covered area. The correlations of these estimates with streamflow are evaluated for each of the watersheds. http://www.snow.ucsb.edu/
C21B-0452
A statistical estimation of Snow Water Equivalent coupling ground data and MODIS images
The Snow Water Equivalent (SWE) is an important component of the hydrologic balance of mountain basins and snow fed areas in general. The total cumulated snow water equivalent at the end of the accumulation season represents the water availability at melt. Here, a statistical methodology to estimate the Snow Water Equivalent, at April 1st, is developed coupling ground data (snow depth and snow density measurements) and MODIS images. The methodology is applied to the Mallero river basin (about 320 km²) located in the Central Alps, northern Italy, where are available 11 snow gauges and a lot of sparse snow density measurements. The application covers 7 years from 2001 to 2007. The analysis has identified some problems in the MODIS information due to the cloud cover and misclassification for orographic shadow. The study is performed in the framework of AWARE (A tool for monitoring and forecasting Available WAter REsource in mountain environment) EU-project, a STREP Project in the VI F.P., GMES Initiative. http://www.aware-eu.info
C21B-0453
Development and Evaluation of Dimensionless Snowmelt Depletion Curves Using MODIS- Derived SCA
Snow accumulation and melt parameterization at a watershed or mountain-front scale is important for improvements in distributed snowmelt modeling. Dimensionless snowmelt depletion curves, which relate normalized decreases in SCA against normalized decreases in SWE, are an effective way to scale-up snowmelt model. This study involved the development of dimensionless snowmelt depletion curves to parameterize snowpack variability throughout the snowmelt season within specific 500 x 500 m areas (pixels) in the Dry Creek Experimental Watershed in southwestern Idaho and within the Imnaviat Creek Watershed in northern Alaska. MODIS derived SCA was incorporated into the development of depletion curves offering exciting advancements involving the obtainment of SWE at relevant scales. Comparisons of depletion curve parameterizations against ground-based measurements illustrate remarkable similarity. Dimensionless snowmelt depletion curves developed with MODIS-derived SCA offers improvements to the present day methods for prediction snowmelt and therefore snowmelt generated runoff.
C21B-0454
An Assessment of SNODAS Data for Hydrologic Forecasting
The NOAA/NOHRSC SNODAS product is derived by ingesting a multitude of observations into a snow model that is run over a high resolution grid. Due to SNODAS applying all the practical snow observations currently available within it's assimilation strategy, few studies have been able to provide independent validation of the end products. As such, there is still a degree of uncertainty about SNODAS which has limited its utility in fields such as hydrologic forecasting. In this study, a validation, calibration and assimilation of the SNODAS output is undertaken. Using a catchment scale hydrologic model, TOPNET, we assess the utility of the SNODAS data within the problem of streamflow prediction. The snowmelt component of TOPNET is able to operate at different levels of complexity, from simple degree day mode right through to a full surface energy balance. SNODAS output variables such as snow water equivalent, snow temperature and snow melt are used in developing calibration objective functions for the TOPNET model. In a further application, these variables are assimilated directly into the TOPNET model. Results from TOPNET are compared to observed streamflow at the base of the catchment. The study is undertaken in the East Fork of the Carson River basin in California/Nevada, which is the site of the NOAA/NWS/OHD DMIP-2 experiment.
C21B-0455
Potential Biases of April 1 Snow Water Equivalent Records
We compare snow water equivalent (SWE) measurements from snow pillows and snow courses in the Sierra Nevada. Specifically, we use daily measurements from snow pillows to study potential biases when using the April 1 snow course measurements in trend studies. We also study variation in the amplitude and phase of the seasonal cycle in pillow-based SWE profiles. SWE trend studies usually use snow course data because of the length of record, almost always employing April 1 snow course measurements as a surrogate for peak snowpack or for the total amount of precipitation. Questions arise as to whether the April 1 snow course measurements over- or underestimate the peak, thus possibly introducing biases in the trend estimation. Snow pillows provide daily records of SWE across the whole year. At snow pillows we estimate the peak and total amount of precipitation based on the SWE profile. We compare these with April 1 SWE measurements from nearby snow courses. Both amplitude and phase variability in the seasonal cycle contribute to potential biases of April 1 records. We present statistical curve registration summaries of seasonal cycle variation in daily snow pillow records, including relationships between the amplitude and timing of the seasonal peak. This provides further insight on the nature of potential biases.
C21B-0456
Physiographic Variables to Describe Basin Scale Snow Water Equivalent
The Natural Resources Conservation Service has installed automated Snow Telemetry (SNOTEL) stations to replace snow course measurements. These data have been used to explore the relationship between snow water equivalent (SWE) and a variety of physiographic variables for the Colorado River Basin, USA. These variables include location, slope and aspect at different scales, derived parameters to indicate the distance to sources of moisture and proximity to and characteristics of obstacles between these source areas of snow accumulation, and forest density. A weekly time step of SNOTEL SWE data from 1990 through 1999 was used to determine the most relevant variables. The most important variable was elevation. Slope at a medium scale and at a regional scale were postively correlated with SWE. The seasonal variability illustrated the necessity to formulate the regressions for each time step. The interannual variation in the relationship between SWE and physiographic variables partially corresponded with snow accumulation and the El Niño Southern Oscillation cycle, yet the variability across the basin in accumulation trends reduced this correlation. Using the suite of relevant variables, the strength of the relationship to SNOTEL data was examined for a period prior to (pre-1990) and after (post-1999) the initial study period.
C21B-0457
Size Frequency Distributions for Snow Avalanches
We examine crown size frequencies for two extensive datasets of observations made during operational avalanche control: 10,300 events at Mammoth Mountain, California and 219,000 events from the Westwide Avalanche Network (WAN) which includes ski areas and highway operations. We compare a dozen distributions, and we address observer bias by employing ratio estimates, smoothing functions, and exclusion rules. Knowing that avalanche professionals often do not record small events, we examine both datasets with no exclusions and with a 60 cm exclusion rule. The WAN data are best fit by a power law distribution using the 60 cm exclusion rule. The power law distribution with 60 cm exclusion also fits the Mammoth data, although these data are best fit by a hyperbolic tangent distribution under both the 60 cm exclusion rule and without exclusion. Our findings support past literature showing that power laws provide a good fit for size-frequency relationships across different regions. Power law distributions indicate scale invariance across several orders of magnitude and are consistent with self organized critical systems. Independent of the choice of distribution, we advocate the implementation of probabilistic avalanche forecasts that convey uncertainty to the end-user, unlike deterministic forecasts. We propose the use of cumulative distribution functions (CDFs) as the dependent variables in numerical avalanche forecast models. CDFs allow normalized output for a region or specific path. A user can infer the magnitude of avalanche events for each avalanche path or area of interest from the CDF. We attempt to create a basis for such an implementation in avalanche forecasting.
C21B-0458
No More Snow Pits? Potential to Retrieve Bulk Snow Pack Structure from Transient Barometric Pressure Waves measured from a Prototype Embedded Wireless Sensor Network
Networked micro-sensors have the capacity to enable improvements in ground-based data collection at resolutions that are currently unresolved. Crossbow® Environmental Motes (MEP410 Models) were embedded in an accumulated snow pack in a meadow at Niwot Ridge Long-Term Ecological Research (LTER) C1 site. Motes were mounted to a custom designed deployment structure that constitutes an adaptive sensor "tower" at a maximum height of 142 with sensors deployed at depths of 122cm, 71cm, and 17cm from the ground. Barometric pressures measured from this embedded wireless network indicate low frequency fluctuations over the study period with higher frequency components varying as a function of depth. Mean pressure measured over the analysis period were 707.82 mb (122cm), 702.48 mb (71cm), and 704.09mb (17cm). Results indicate filtering of high frequency components of pressure measured at depth with reduced amplitude and time-lag (phase shift) modulated by changes in snow pack permeability related to density changes.
C21B-0459
Prototyping and Testing a Wireless Sensor Network to Retrieve SWE at High Spatial Resolution
A critical challenge in snow research from space is the ability to obtain measurements at the spatial and temporal resolution to characterize the statistical structure of the space-time variability of the physical properties of the snowpack within an area consistent with the pixel resolution in snow hydrology models or that expected from a future NASA mission dedicated to cold region processes. That is, observations of relevant snow dielectric properties are necessary at high spatial and temporal resolution during the accumulation and melt seasons. We present a new wireless sensor network prototype consisting of multiple antennas and buried low-power, multi- channel transmitters operating in L-band that communicate to a central pod equipped with a Vector Signal Analyzer (VSA) that receives, processes and manages the data. Only commercial off-the-shelf hard-ware parts were used to build the sensors. Because the sensors are very low cost and run autonomously, one envisions that self-organizing networks of large numbers of such sensors might be distributed over very large areas, therefore proving much needed data sets for scaling studies. The measurement strategy consists of placing the transmitters the land surface in the beginning of the snow season which are then run autonomously till the end of the spring and waken at pre-determined time-intervals to emit radio frequency signals and thus sample the snowpack. Along with the sensors, an important component of this work entails the development of an estimation algorithm to estimate snow dielectric properties, snow density, and volume fraction of snow (VF) from the time-of-travel, amplitude and phase modification of the multi-channel RF signals as they propagate through the snow-pack. Here, we present results from full system testing and evaluation of the sensors that were conducted at Duke University using ¢®¡Æsynthetic¢®¡¾ limited-area snowpacks (0.5 by 0.5 m2 and 1 by 2 m2) constructed of various combinations of foam layers of different porosities to simulate heterogeneous distributions of water. The existing sensors are currently being primed for field deployment. Discussion is also presented regarding further technology development including power usage, networking, and distribution and operations in remote regions.
C21B-0460
Embedded sensor network design for spatial snowcover
Scaling point observations of snow water equivalent (SWE) to model grid-element scales is particularly challenging given the considerable sub-grid variability in snow accumulation over complex terrain. In an effort to capture this sub-grid variability and provide spatially explicit ground-truth snow data an embedded snow sensor network was designed and installed in Yosemite and Sequoia National Park, and in the Valles Caldera National Preserve. Extensive snow surveys were used to guide the installation of the network and to relate the observations to more detailed spatial SWE fields. Three years of continuous spatial and temporal data from both Yosemite National Park and the Valles Caldera indicate that accumulation and ablation rates can vary as much as 50% based on variability in topography and vegetation. These snow distribution patterns are especially apparent in the open forests of Yosemite and Sequoia National Parks and the Valles Caldera where vegetation structure largely controls variability in snow distribution. Comparisons of SWE estimates with historical snow course data shows that a single point measurement is a poor estimator of snow depth over a homogeneous area, but 4 or more measurement points can reduce the uncertainty by 50%. Further analyzes indicated that an optimal snow depth network consists of 7 to 10 snow depth sensors. These spatial and temporal measurement arrays will improve remotely sensed and modeled SWE estimates by providing robust, spatially explicit ground- truth values of snowpack states.
C21B-0461
Observations, process studies, and modeling of seasonal snow in New Zealand
Variability of seasonal snow in New Zealand directly affects the sectors of energy, agriculture and tourism, but intensive research and monitoring programs have only been developed recently. This paper will discuss observations, process studies, and modeling of snow in New Zealand: · Observations. Observations of snow in New Zealand have recently been significantly enhanced with the establishment of the National Institute for Water and Atmospheric Research snow and ice monitoring network. This network includes multiple new permanent high elevation stations that measure a large suite of micrometeorological variables as well as snow depth. Additional instrumentation to allow for the accurate measurement of solid and liquid precipitation and SWE are being investigated. We will discuss the establishment of the monitoring network, and outline problems encountered in measuring snow and climate in an extreme maritime alpine environment. · Process Studies. There have only been a few process studies of snow in New Zealand. However, new process studies are needed because there are notable differences between snow in New Zealand and snow in many other countries: Accumulation processes in New Zealand are unique because most of the snow is above treeline and a large amount of snow is redeposited by the wind; melt processes in New Zealand are unique because melt is largest in windy humid conditions when turbulent heat fluxes of sensible and latent heat dominate the radiative heat fluxes. We will summarize results from an intensive measurement campaign aimed at documenting and understanding spatial variability of snow water equivalent. · Modelling. Building an appropriate model to simulate snow in New Zealand requires defining a model with complexity that is justified in light of the available data, the problem we are trying to solve, and processes dominant where the model is applied. Additional model complexity comes at the expense of additional model parameters (many of which are difficult to estimate), and it is difficult to define the necessary complexity a-priori. For these reasons we built a flexible snow model in which processes can be "switched on" and "switched off", and each process is parameterized in the most parsimonious manner imaginable. We will describe application of the snow model to basins in New Zealand. We will also demonstrate how observations, process studies, and modeling of snow in New Zealand contribute to models used to predict streamflow in snowmelt-dominated basins and simulate impacts of climate change on New Zealand's water resources.
C21B-0462
A Micro-Structural Phase-Field Model for Snow Metamorphism and First Experimental Validations using Migrating Air Inclusions in Ice
Snow is a highly porous medium consisting of an ice matrix and porous space containing water vapor. Moreover, snow undergoes metamorphism as heat flow and interface effects induce mass flow and thus profoundly change the microstructure, i.e., the distribution of ice and pores. Reciprocally, this evolution influences the thermophysical, chemical, and mechanical properties of snow. In particular, the microstructure of snow influences the heat conductivity as heat transport consists in (i) heat conduction in the ice and pores, (ii) heat transport related to water vapor diffusion in the pores, and (iii) latent heat release and gain due to phase changes at the ice-pore interfaces Recently, detailed image series of metamorphosing snow using computed X-ray micro-tomography (micro-CT) became available and models for heat conduction through a steady state ice and pore network emerged. We present a phase-field model to solve the coupled heat and mass transport problem including phase-change processes in an evolving ice-pore network. The model considers mass fluxes that are induced by temperature gradients in the snow as well as by curvature effects and handles topological changes of the microstructure implicitly. We apply the model to 3D micro-CT data of snow. The simulations agree qualitatively well with laboratory observations and underline the strong link between microstructure and heat conductivity of snow. In order to validate the model quantitatively and to constrain the model parameters, simpler experiments than snow metamorphism observations by micro-CT are needed. We designed a relatively simple experimental apparatus to observe the migration of air inclusions in ice subjected to a temperature gradient. Considerable insulation and good temperature control at the hot and cold sides of an ice block allow us to impose a nearly constant and mono-dimensional temperature gradient. Small air inclusions can be inserted into the ice for example by drilling. The advantage of using a rather big ice-block with small inclusions instead of the sparse and intricate ice matrix as it would be for snow is that a much better temperature control can be achieved. The migration of the inclusions through the ice for given temperature gradients is observed by digital photography and migration velocities can be computed. We use this experimental data to constrain the phase-field model parameters by comparing it to simulated ice inclusion migrations.
C21B-0463
Land Surface Model Biases and their Impacts on the Assimilation of Snow-related Observations
Some recent snow modeling studies have employed a wide range of assimilation methods to incorporate snow cover or other snow-related observations into different hydrological or land surface models. These methods often include taking both model and observation biases into account throughout the model integration. This study focuses more on diagnosing the model biases and presenting their subsequent impacts on assimilating snow observations and modeled snowmelt processes. In this study, the land surface model, the Community Land Model (CLM), is used within the Land Information System (LIS) modeling framework to show how such biases impact the assimilation of MODIS snow cover observations. Alternative in-situ and satellite-based observations are used to help guide the CLM LSM in better predicting snowpack conditions and more realistic timing of snowmelt for a western US mountainous region. Also, MODIS snow cover observation biases will be discussed, and validation results will be provided. The issues faced with inserting or assimilating MODIS snow cover at moderate spatial resolutions (like 1km or less) will be addressed, and the impacts on CLM will be presented.
C21B-0464
Differences Between the Spatial Organization of Snow Depth Fields in Sub-alpine Forest and Alpine Tundra
In this study, differences in the spatial organization and scale invariance properties of snow depth between a forested environment and an alpine environment are studied. The analysis is based on estimates of the probability distribution function, two-dimensional correlation functions and power spectral densities of high- resolution LIDAR measurements (~ 1 m) obtained for two adjacent study areas of 500 m x 500 m located in the Colorado Rocky Mountains. Both of the areas are located in the Alpine ISA of NASA's Cold Land Processes Experiment (CLPX) and present similar topographic characteristics (e.g., slope and aspect), limiting the differences to vegetation characteristics and the influence of snow redistribution by wind. Furthermore, Fourier filtering techniques and the Turning Bands Method are used for generating synthetic one-dimensional profiles, and isotropic and anisotropic two-dimensional random fields that reproduce the scale invariance properties (i.e., spectral exponents and scale breaks) observed in the snow depth fields. These methodologies are also used to infer explanations for the breaks in the scaling behavior of the spatial distribution of snow depth, as well as the meaning and implication of the spectral exponent values with respect to the variability of the fields. Implications of these results for the design of measurement strategies, interpolation, and modeling are discussed.
C21B-0465
Distribution and Energy Balance of Shallow Patchy Snow in Complex Terrain
The extent and energy balances of patchy snow are of great interest when satellite images are used in basin scale snow melt modeling. Problems of accuracy arise when subgrid variability of snow covered area (SCA) occurs on a scale of tens of meters inside a pixel with a scale of hundreds of meters. Warming climate has and is predicted to continue to cause snowlines to rise in lower elevation mountainous areas resulting in deep snowpacks being replaced with shallow subgrid fractional snow cover. This study used basal snow temperatures to show the extent of snow cover and differential melt timing on two aspects of a semi arid watershed in the foothills of Boise, Idaho. As part of the study on the energy balance of shallow patchy snow, distributed temperature sensing (DTS) was used to capture basal snow temperatures across both north and south facing aspects of the Upper Dry Creek experimental watershed in Idaho during a 36 hour melt period following a storm event in March of 2007. Consistent temperatures on the north face reflect a continuous full snow coverage while snow depth decreased by 10 cm. Temperatures on the south facing aspect show initial patchy snow patterns followed by a nearly complete melt by the end of monitoring due to greater solar heating. The instrument used for DTS measurements detects Ramen scattering to yield temperatures with an accuracy of „b0.1„aC and a spatial resolution of 1 meter for the entire length of the cable. The study showed the technology to be a substantial advantage over single point measurements when characterizing soil temperatures in complex terrain with differential solar exposure and melting times.
C21B-0466
Terrain and forest shelter effects on snowcover energetics, patterns of snow deposition, snowmelt and runoff over a semi-arid mountain catchment
In mountainous regions, topographic and vegetation structure control snowcover energetics, patterns of snow deposition, meteorological conditions, snowmelt and runoff. A topographically distributed snow accumulation and melt model (ISNOBAL) is coupled to a wind field and snow redistribution model to simulate the development and ablation of the seasonal snowcover over a small mountainous catchment, the Reynolds Mountain East basin in southwestern Idaho, USA. Simulations were conducted for 9 water years (1984, 1986, 1987, 1989, 1992, 1997, 2001, 2004 and 2006) representing a range of wet and dry, warm and cold conditions. Simulated patterns of snowcover energetics, mass balance and surface water input for each water year are shown. Discussion and analysis of how distributions of snowcover energetics, development, melt and ablation response to dry and wet, and warm and cold snow season conditions is presented. Snow redistribution in relation to drift and scour zones, and vegetation shelter under the range of conditions presented is analyzed. The details of select events, such as mid-winter melt, rain-on-snow, early spring wind and high radiation are evaluated and discussed. This research will improve our understanding of how the structure of topography and vegetation influences snowcover distribution, energetics, hydrology and water resources in snow and wind-dominated mountainous regions.
C21B-0467
Quantifying the Effects of Forest Canopy Cover on Snow Accumulation and Ablation at a Continental, Mid-latitude Site, Valles Caldera National Preserve, NM
Basin scale estimates of water resource quantity and quality in snow-dominated systems, typically based on point measurements such as snow courses and SNOTEL sites, are complicated by the interrelated forcings of topography and vegetation on snowpack accumulation and ablation. Although the effects of forest density on snowpack are relatively well characterized in northern systems, the interactions in the higher solar-radiation environments found at mid-latitude sites are not well understood. In this study we quantify the effects of forest density and geometry on snow accumulation and demonstrate how this information can be used with remote sensing to improve estimates of snow distribution. We measured snow depth and density through a continuum of forest canopy densities in the Valles Caldera National Preserve of New Mexico during the early season, peak accumulation, and melt season of spring 2007. We measured snowpack properties in six snow pits, representing a variety of aspects and vegetation densities, and took a total of 1,350 measurements of snow depth in 20 plots of various canopy densities corresponding to specific Landsat Enhanced Thematic Mapper pixels. By comparing depth data to canopy cover as measured in the field and by satellite, we quantify the interactions between snow and forest vegetation at the 30-m scale. Relationships developed at the pixel scale are used to develop algorithms for basin-scale estimation of snow distribution using the newly available National Land Cover Data - Forest Cover dataset. Initial results show significant correlations between snow depth and both field-measured and remotely sensed forest canopy density. To determine the physical processes underlying these correlations we used statistical models to separate and evaluate the effects of canopy interception, solar radiation shading, enhanced longwave radiation, and wind redistribution on snowpack depth. We also found significant differences in snow depth corresponding to the density of vegetation in neighboring areas, indicating that vegetation influences snow depth beyond the canopy and canopy-fringe. This study is part of a larger effort to quantify the water balance in the Valles Caldera, which also comprises efforts to determine the isotopic signature and chemical load of winter precipitation, and to determine the residence time of water in the Redondo Peak massif using these chemical tracers in conjunction with the depth estimates of this study.
C21B-0468
Spatial Scale for Modelling Blowing Snow on the Canadian Prairieis
Blowing snow transports and sometimes sublimates much of the seasonal snowfall in the Prairies of western Canada. Snow redistribution is an important feature of Prairie hydrology as deep snowdrifts provide a source of meltwater to replenish ponds and generate streamflow in this dry region. The spatial distribution of snow water equivalent in the spring is therefore of great interest for Prairie hydrology. A test of the appropriate spatial scale for modelling blowing snow redistribution and sublimation was conducted at St Denis National Wildlife Area in the rolling, internally drained prairie pothole region east of Saskatoon, Saskatchewan, Canada. A LiDAR based DEM and LANDSAT based vegetation map were available for this region. A coupled complex windflow and blowing snow model was run with ~250,000 6 m x 6 m grid cells to produce spatially distributed estimates of seasonal blowing snow transport and sublimation. The calculation was then aggregated, using 7 landscape units that represented the major influences of surface roughness, topography and fetch on blowing snow transport and sublimation. Both the distributed and aggregated simulations described similar end of winter snow water equivalent with substantive redistribution of blowing snow from exposed sparsley vegetated sites across topographic drainage divides to the densely vegetated pothole wetlands. Both simulations also agreed well with snow survey observations. While the distributed calculations provide a fascinating and detailed visual image of the interaction of complex landscapes and blowing snow redistribution and sublimation, it is clear that blowing snow transport and sublimation calculations can be successfully aggregated to the spatial scale of the major landscape units in this environment.
C21B-0469
Simple and Computationally Efficient Modeling of Surface Wind Speeds Over Heterogeneous Terrain
In mountain catchments wind frequently is the dominant process controlling snow distribution. The spatial variability of winds over mountain landscapes is considerable producing great spatial variability in mass and energy fluxes. Distributed models capable of capturing the variability of these mass and energy fluxes require time-series of distributed wind data at compatible fine spatial scale. Atmospheric and surface wind flow models in these regions have been limited by our abilities to represent the inherent complexities of the processes being modeled in a computationally efficient manner. Simplified parameterized models, such as those based on terrain and vegetation, though not as explicit as a model of fluid flow, are computationally efficient for operational use, including in real time. Recent work described just such a model that related a measure of topographic exposure to wind speed differences at proximal locations with varied exposures. The current work used a more expansive network of stations in the Reynolds Creek Experimental Watershed in southwestern Idaho, USA to test extension of the previous findings to larger domains. The stations in the study have varying degrees of wind exposure and comprise an area of approximately 125 km2 and an elevation range of 1200 - 2100 masl. Subsets of site data were detrended based on the relationship derived in the prior work to a selected standard exposure to ascertain and model the presence of any elevation-based trends in the hourly observations. Hourly wind speeds at the withheld stations were then predicted based on elevation and topographic exposure at each respective site. It was found that reasonable predictions of wind speed across this heterogeneous landscape capturing both large-scale elevation trends and small-scale topographic variability could be achieved in a computationally efficient manner.
C21B-0470
Impurities in Snow: Effects on Spectral Albedo of Prairie Snowpacks
While extensive research on soot in snow has been done in the Polar Regions, there remains a lack of observations addressing the effect of soot on snow albedo in North American prairie snowpacks which causes uncertainty to the overall global effect that soot in snow has on climate. Measurements of snow impurities in freshly fallen prairie snowpacks in northwestern Iowa and central Texas collected from February 28 - March 5, 2007 and April 6, 2007, respectively. Two significant snowfall events occurred in northwestern Iowa during the study; the second snowfall event produced the most severe blizzard conditions in northwestern Iowa in the last thirty years. An unusual snowfall event in central Texas offered a unique sampling opportunity Several types of sites were sampled during the field campaign; this includes: frozen lakes with minimal human impact, agricultural fields impacted by agricultural dust, and human impacted sample sites. At twelve sites in northwestern Iowa samples were collected on multiple days and for both snow events to examine changes in snow impurities over time. At all site locations snow samples, temperature, density, and grain size were recorded. Snow reflectance and snow radiance was collected at a subset of the sites with an ASD VNIR Spectroradiometer (350 - 1500 nm). Snow impurities of light-absorbing particulate matter were measured by filtering the meltwater through a nuclepore 0.4 micrometer filter. Impurity concentration was determined by comparing the filters against a set of standards. A photometer will provide a more exact determination of snow impurities in the near future. Preliminary soot observations indicate prairie snow pack concentrations ranging from 1 ngC/g to 236 ngC/g with an average of 61.4 ngC/g. These measurements are within range of previously published values in the Arctic and can lower snow albedo. Differences in soot concentrations were observed between the two Iowa snowfall events. Impurity concentrations measured for the central Texas snowfall were higher than either of the Iowa snowfalls. As expected, spectral albedo was found to decrease with increasing impurities.
C21B-0471
Comparison of Algorithms for Incoming Atmospheric Long-wave Radiation
Early season snowmelt can be dominated by the long-wave radiation balance. While numerous algorithms exist for predicting incident atmospheric long-wave radiation under clear (Lclr) and cloudy skies, only a handful comparisons have been published to assess the accuracy of the different algorithms. Virtually no comparisons have been made for both clear and cloudy skies across multiple sites. This study evaluates the accuracy of twelve algorithms for predicting incident long-wave radiation under clear skies, ten cloud correction algorithms, and four algorithms for all-sky conditions using data from twelve sites across North America and China. Data from three sites were combined with publicly available data from nine sites in the AmeriFlux network. Clear sky algorithms that excelled in predicting Lclr were the Dilley-O'Brien, Prata, and Angstrom algorithms. Root mean square difference (RMSD) between predicted and measured Lclr averaged 22 to 23 Wm-2 for these three algorithms across all sites. Cloud-correction algorithms of Kimball, Unsworth-Monteith and Crawford described the data best when combined with the Dilley clear-sky algorithm. Average RMSD across all sites for these three cloud corrections was 24 Wm-2. Based on the results, the recommended algorithms can be applied with reasonable accuracy for a wide range of climates, elevations and latitudes.
C21B-0472
Development of a hyperspectral method for detecting the radiative impact of desert dust on alpine snow.
Springtime dust storms regularly deposit hematite-rich dust upon the San Juan Mountains of southwestern Colorado. Dust deposited upon the snow surface reduces the visible albedo of snow. The high spectral resolution of AVIRIS (NASA-JPL Airborne Visible Infrared Imaging Spectrometer) makes it applicable to quantitative detection of the grain size and visible albedo of snow, and thus an effective tool for evaluation of the radiative effects of dust. An inversion technique, ND-Model, derives the spatial distribution of optical snow grain radii via continuum normalization of the 1.03 micrometer ice absorption feature. The Scaled Integral Dust Index (SIDI) provides an index of the shortwave radiative impact of dust by taking the integral of the spectral reflectance of dusty snow's deviation from the visible spectral reflectance of pure snow of the same spherical optical grain size as modeled by DISORT (Discrete Ordinate Radiative Transfer Model). Because dust is hypothesized to increase snow grain size, it was expected that retrievals values of dust index (SIDI) and snow optical grain size would be positively correlated. Initial applications of ND-Model and SIDI to AVIRIS data from the San Juans instead yielded a negative correlation between grain size and dust index. The presence of hematite-rich dust at the surface may violate the clean-snow, pure-pixel assumptions of ND- Model, changing the scaled area of the 1.03 μm ice absorption feature and retrieved optical snow grain size. Such a problem within the grain size model would affect the detected correlation between dust index and optical grain size. It is hypothesized that, in the presence of a critical amount of dust within the optical depth of the snowpack, the signals for grain size and dust concentration effectively are crossed. Here, we investigate the possible invalidation of ND-Model via spatially explicit topographic radiation modeling, analysis of in-situ observations, and GIS.
C21B-0473
Surface Hoar Formation in Complex Terrain
Vapor exchange processes between the snow surface and the lower atmosphere lead to mass sublimation or deposition. While the former usually occurs during day, the latter often co-occurs at night with the formation of surface hoars. These crystals, when buried after snow falls, often form failure planes for avalanches. Understanding surface hoar formation and modeling the underlying physics would then substantially improve avalanche warning and reduce hazards in the alpine area. The present experimental study is based on measurements obtained at 2540 m a.s.l. above Davos, Switzerland, during winter 2007 and it focuses on the formation, development and destruction of surface hoar crystals. Mass changes were measured with a box in the field, turbulent fluxes were estimated by means of two sonic anemometers coupled with two fast gas analyzers, snow parameters were collected as well as meteorological parameters in situ and at synoptic scale. Measurements clearly indicate that turbulent vapour fluxes (obtained at three meters above the surface) are entirely responsible for the mass gain in the snow pack and thus of the formation and growth of surface hoars. The analysis of the meteo-data suggest that local katabatic winds from nearby slopes during nights of surface hoar development significantly contribute to the turbulent fluxes measured near the surface and thus to the growing of surface hoars. From the modeling side, the bulk-approach as implemented in the snow cover model SNOW-PACK agrees well with the eddy correlation, and the mass balance measurements, thus resulting in a correct prediction of surface hoar formation, for the cases examined.
C21B-0474
Lateral Water Movement in a Melting Snowcover: Implications for Hydrologic Modeling
Catchments convert vertical inputs of precipitation to lateral fluxes. Quantifying the integral processes of moisture redistribution within a catchment is a key challenge for hydrologic modelers, yet the lateral movement of meltwater within a snowpack is commonly ignored. We present data that suggests that the processes governing lateral redistribution of meltwater routing in the snowpack during a snowmelt event are important to the total flux of water out of a watershed. Initial dye tracer field studies near Mores Creek Summit, Idaho, have shown that meltwater movement on moderate angle slopes (<35°) through a ripe, melting snowpack is highly mobile with flow directions that are largely dependant on snowpack characteristics. These experiments have shown considerable downslope movement of meltwater with a dependance on diurnal cycles. Due to the spatial variability of snowpack stratigraphy, there is also a resultant variability in meltwater volume movement over small areas. The experiments showed the dye tracer infiltrated vertically through the snowpack until it reached a stratigraphic boundary, where it began to accumulate and move laterally downslope. These layers within the snowpack stratigraphy were signified by an increase in liquid water content, grain size, and density. The tracer moved large distances laterally along these boundaries, unless conditions allowed for it to resume vertical infiltration. Initial dye emplacement experiments showed a downslope water movement velocity on the order of 1 m h-1. The results presented here are preliminary observations from field work performed during the spring 2007 snowmelt event.
C21B-0475
Modeling Spring Snowmelt Dynamics in a Northern Rockies Basin using a Modified Temperature-Index Model
Water resources and ecological processes in snow-dominated basins are directly forced by spring snowmelt processes. Here we investigate snowmelt volume and timing in the Middle Fork Flathead River Basin, a large (2903 km2) basin located west of the continental divide in the northern Rockies. We developed a spatially distributed snow melt model which employs a degree-day approach modified to incorporate solar radiation influences on melt. Moderate Resolution Imaging Spectroradiometer (MODIS) snow-cover products are piped into the model to determine snow covered area (SCA). Snowpack and climate variables were determined using Snowpack Telemetry (SNOTEL) data and other climate station data. A radiation proxy can be calculated without the need for additional data, improving on the model while maintaining its simplistic need for meteorological inputs. Climate and solar radiation data were input to the model and combined with the SCA data derived from MODIS snow-cover products to generate snowmelt volumes during the melt period for each study year. Incorporation of solar radiation effects elucidates the processes driving melt and allows for visualization of the spatial variations in snowmelt throughout the watershed.
C21B-0476
MODIS-based Snow Cover Variability of the Upper River Grande Basin
Snow cover and its spring melting in the Upper Rio Grande Basin provides a major water source for the Upper to Middle Rio Grande valley and Elephant Butte Reservoir. Thus understanding the snowpack and its variability in the context of global climate change is crucial to the sustainable water resources for the region. MODIS instruments (on Terra and Aqua) have provided time series of snow cover products since 2000, but suffering with cloud contaminations. In this study, we evaluated four newly developed cloudless snow cover products (less than 10%) and four standard products: daily (MOD10A1, MYD10A1) and 8-day (MOD10A2, MYD10A2), in comparison with in situ Snowpack Telemetry (SNOTEL) measurements for the hydrological year 2003-2004. The four new products are daily composite of Terra and Aqua (MODMYD10DC), multi-day composites of Terra (MOD10MC), Aqua (MYD10MC), and Terra and Aqua (MODMYD10MC). The standard daily and 8-day products can classify land correctly, but had fairly low accuracy in snow classification due to cloud contamination (a average of 39.4% for Terra and 45% for Aqua in the year 2003-2004). All the new multi-day composite products tended to have high accuracy in classifying both snow and land (over 90%), as the cloud cover has been reduced to less than 10% (~5% for the year) under the new algorithm . This result is consistent with a previous study in the Xinjiang area, China (Wang and Xie, 2007). Therefore, MOD10MC (before the Aqua data available) and MODMYD10MC products are used to get the mean snow cover of the Upper Rio Grande Basin from 2000 to 2007. The snow depletion curve derived from the new cloud-free snow cover map will be used to examine its effect on stream discharge. http://www.utsa.edu/LRSG/