Winter 2001 snowfall at the ECM1 meteorological site.

The major goal of this focus area is to measure and model the components of the water balance above the mountain front. Determination of water yield from mountain precipitation and snowmelt is vital for estimating the annual water balance in large semi-arid watersheds. Currently there is very limited information about the distribution of hydrologic variables above the mountain front at the spatial and temporal resolutions needed for regional scale hydrologic modeling. Uncertainty in spatially distributed estimates of hydrologic variables introduced by the up-scaling of point measurements and down-scaling of course resolution space-borne remote sensing data motivates research on improving our understanding of the distribution of hydrologic variables at intermediate (0.1km -10 km) resolutions. Three research areas are being pursued in which improved understanding of water, energy and carbon cycling processes (sub-task 1.1.1) are used to improve parameter estimates for spatially distributing hydrologic variables (sub-task 1.1.2). Hydrologic modeling, incorporating the improved inputs and information about input uncertainty, is then used to assess the dependency of output accuracy on parameter and input uncertainty (sub-task 1.1.3).

The three individual research efforts share common primary data acquired in the field as well as common modeling efforts related to snow distribution, melt, runoff and evapotranspiration/sublimation. Similar energy balance, evapotranspiration/sublimation, soil moisture and carbon flux data sets are being collected by TA1 above the mountain front, by TA2 on the desert floor and by TA3 in the riparian zone. An additional data set from a high elevation alpine area has also been leveraged. These data sets are necessary for model development, calibration, and validation, both within these thrust areas and for integration in TA4. The point and basin scale snowmelt models developed and refined by this group are to be integrated into TA4 modeling efforts.

The left side of Figure 1 shows the integration of these subtasks within TA1 and within SAHRA. Between the two individual research efforts there is direct sharing of the common primary data acquired in the field as well as common modeling efforts related to snow distribution, melt, runoff and evapotranspiration/sublimation. Similar energy balance, evapotranspiration/sublimation, soil moisture and carbon flux data sets are being collected by TA1 above the mountain front, TA2 on the desert floor and TA3 in the riparian zone. These data sets are necessary for model development, calibration, and validation, both within these thrust areas and for integration in TA4. The point and basin scale snowmelt models developed/refined by this group are being integrated into TA4 modeling efforts. NASA Regional Earth Science Application Center (RESAC) provides remotely sensed snow products such as spatial snow cover areas (SCA) and snow water equivalent (SWE) information.

Individual projects are:

• Water, energy, and carbon cycling in southwestern subalpine forests
• Spatial distribution of energy balance, snow water equivalence and snowmelt in seasonally snow covered watersheds
• Hydrologic modeling of Alpine snow and runoff

Among the ecosystems present in the semi-arid environment of the southwestern US, the sky island forest is unique has a unique relationship to the sparse surface-water resources available in the region. This ecosystem exists only at the top of mountains because only there does long-term average precipitation input exceed evapotranspiration to the extent that forest vegetation can survive. Sky island forests, therefore, command potentially significant source areas for the water (some originally falling as snow) that ultimately leaves topographically high ground to recharge aquifers in the plains below via mountain-front recharge. Quantifying and understanding water, energy, and related carbon cycling and budgets of this sustainable source of water is of direct relevance to the mission of SAHRA.

The Mount Bigelow project aims to provide an empirically based understanding of the hydro-micrometeorological dynamics of a sky island sub-alpine forest in the southwestern U.S. The fundamental science issues are: characteristics of the surface-atmosphere exchanges of water, energy and carbon; storage of moisture and energy in plants and soil; and partitioning of winter snow and rain between evapotranspiration/sublimation, deep drainage, and the near-surface environmental water resource that sustains the forest. In order to achieve our objective, a network of four below-canopy hydro-micrometeorological stations 3 m tall, and one above-canopy 30 m tall high-resolution eddy correlation tower were deployed within a predominantly Douglas fir/pine second growth forest. This network will operate for a minimum of two years and ideally for the next seven years, in order to capture strong inter-annual climate variability, as well as to leverage on the GEWEX-CEOP related activities as their semi-arid sky island reference site for the larger basin modeling activities. The Mount Bigelow project is the first study to document, analyze, and model the water, energy, and (related) carbon exchanges of the sky island forest ecosystem. The observations have year-round value. Data gathered in winter aids understanding of how water resources are replenished by winter snow and rain. Data collected in spring aids understanding of the partitioning of water between deep drainage and the near-surface environmental water resource that sustains the forest; while data gathered in summer and fall aids understanding of the evolution of the environmental water resource as it is depleted by evapotranspiration but replenished by monsoon storms. Because the sky island forest environment is unique and because the Mount Bigelow study is unique, every measurement made is novel, and every result obtained is potentially publishable.

Activities and Results

During the period of September 2001 to June 2002 our research activities were primarily centered around the logistics of getting the eddy correlation tower installed and operational (permission for the tower was granted by the Forest Service in August 2001), and adjusting and refining the hydro-micrometeorological network that was installed in the summer of 2001. The observational system for the eddy correlation tower was successfully field tested and recently installed and operational. These systems are capable of routinely documenting, at the plot scale, the surface-atmosphere exchanges of water, energy, and carbon, the storage of moisture and energy in plants and soil, and the above- and below-canopy micrometeorological variables that control these exchanges and storages in a representative sky island forest ecosystem growing in the semi-arid southwestern U.S. The data from the micrometeorological network show distinct spatial variabilities in terms of below-canopy net radiation, soil and surface temperature, as well as in seasonal differences in the interrelationships of these variables as a function of location and canopy characteristics. The soil core equipment and subsequent analysis of the soil core samples taken at the sites were provided by SAHRA researchers at the US Salinity Lab, who will in turn use the results to enhance their database for semi-arid pedo-transfer functions. Our research is relevant to mountain block recharge studies and to high elevation snowmelt studies, and enables more detailed/accurate representation of the southwestern sky island in the various SAHRA modeling activities. The major activities over this period were:

• Installation of an abbreviated version of the scaffolding tower structure at a site in Tucson. Several sections of the scaffolding tower were installed at an easily accessible site in Tucson. Mounting designs and instrument layout were then tested. The observational system and power systems were also tested at this site. The data acquisition system was customized and installed at the local test site.
• Installation of the 30 m tower on Mt Bigelow. The 30 m tower on was completed in April 2002.
• Redeploying and installing the entire eddy correlation observational system on the 30 m tower on Mt Bigelow, which includes: 1 sonic anemometer, 1 open path Licor H2O/CO2 infrared gas analyzer, 1 4-way radiometer, 2 net radiometers, 3 profile levels with wind, temperature and relative humidity, 2 infrared thermometers (1 canopy and 1 surface), 3 sets of soil temp arrays (7 levels), 6 water content reflectometers, a total of 24 thermocouples installed for bole and xylem temperatures of 6 trees, and 1 rain gauge. The eddy correlation data is logged at 10 Hz and the profile, soil, vegetation data at 15-minute averages.
• Designed, customized and tested data processing routines for flux calculations.
• Acquired and examined eddy correlation data from several forested tall tower sites.
• A hydrometeorological seminar course was designed around the activities of the project and offered during the winter semester. Nine graduate students were engaged in both field and theoretical aspects directly related to this project, and made valuable contributions to our progress.
• In collaboration with the Knowledge Transfer section, a group of 15 high school students completed a vegetation survey of the main eddy correlation site.
• Maintained and regularly download hydro-micrometeorological data from the four micrometeorological stations.
• Installed soil moisture probes and rain gauges at all sites. Multiple water content reflectometers and rain gauges were installed at all sites. Soil cores were taken from these locations and sent to the US Salinity Lab for analysis by SAHRA researchers at that facility.
• Analyzed and prepared data for publication from the micrometeorological network that was installed in the summer 2001.
• Researched innovative processing and delineation of high resolution DEM and vegetation images.
• Tested alpine spatial snowmelt model on a small watershed.
• Collection, quality control and post-processing of the eddy correlation data. The acquisition of 10 Hz eddy correlation data started in June 2002.
• The Mt. Bigelow eddy correlation tower has been incorporated as a reference site for GEWEX-CEOP, and has been participating in the associated planning activities for this endeavor.

Plans

• Sustained observations are ongoing at the Mt. Bigelow site and will be maintained using the current observational system (and any additional monitoring systems installed) at the main tower site and three subsidiary sites for at least two full annual cycles. This will allow us to sample the response of the sky island forest ecosystem to inter-annual variability of climate in the southwestern U.S., and as the high altitude, semi-arid observation site for GEWEX-CEOP. Over the next two years we plan to accomplish the following:



• Monitor, analyze, and model the partitioning of winter snow and rain between deep drainage and the near-surface, environmental water resource that sustains the forest. This will involve 1) during the spring months of 2003 and 2004, maintaining the routine observations needed to document the amount and vertical movement of water in the soil at the main tower and three subsidiary sites and the water balance of the Mountain Island Forest ecosystem as a whole, and adding any additional measurements required better to meet this objective; 2) implementing one-dimensional soil-vegetation-atmosphere transfer models with realistic, multi-layer simulation of soil water and energy movement at the main tower and three subsidiary sites, calibrating these models against the available data, and using them to calculate the proportion of water leaving as deep drainage and that retained for use by the forest; and 3) reporting results in presentations and journal articles.



• When there is winter snowfall, document, analyze, and model the evolution in snow and ice cover at four sample locations in the sky island forest ecosystem. This will involve 1) during the winters of 2002/2003 and 2003/2004, maintaining routine automatic observations relevant to snow and ice cover studies and undertaking additional snow surveys, as required, during periods when there is snow present at the main tower and/or three subsidiary sites; 2) implementing one-dimensional snow accumulation/melt models with different complexity at the main ec tower and three subsidiary sites, evaluating their comparative performance and, if required, modifying these models to improve their performance; and 3) reporting results.



• Monitor, analyze, and model the seasonal evolution in the near-surface, environmental water resource that sustains the sky island forest ecosystem, its depletion by evapotranspiration and sublimation and replenishment by snowfall and monsoon storms, and its relationship to plant physiological processes and to the exchange of carbon between the forest and the atmosphere. This will be achieved by: 1) processing and promptly controlling the routine observations from the Mount Bigelow sites to ensure provision of high quality data, 2) analyzing the fast-response data to investigate and define the nature of the turbulent regime at the mount Bigelow site; 3) analyzing the fast response data to provide routine documentation of the surface energy, water, and carbon dioxide fluxes for a typical Mountain Island Forest ecosystem; 4) analyzing the relationship between the surface energy, water, and carbon dioxide fluxes and meteorological, plant physiological, and soil water availability controls on these fluxes; 5) implementing and testing a plot-scale soil-vegetation-atmosphere transfer scheme (selected to be consistent with the preference of the SAHRA modeling group) for a typical Mountain Island ecosystem, and calibrating the parameters therein using the data from the Mt. Bigelow site; and 6) reporting results.



• As the selected high altitude, semi-arid observation site for the GEWEX Coordinated Enhanced Observing Period (CEOP), participate in CEOP by maintaining the collection of observations at the Mt. Bigelow site and providing these in timely manner as quality-controlled data to the CEOP data base. This will involve: 1) controlling the quality of routine observations made at the Mt. Bigelow in a timely manner and publishing these through the CEOP data system; 2) initiating collaborative research at CEOP semi-arid sites; and 3) reporting results.



• To the extent feasible, document over at least two years the difference in the surface exchanges of energy, water, and carbon for fully-grown sky island forest and forest that has recently been burnt. This will involve: 1) vectoring the high-resolution ec data according to wind direction and footprint, and analyzing the respective energy and carbon fluxes; and 2) reporting results.



• The activities listed above will provide SAHRA researchers with high-quality, high-resolution datasets for model validation, and functional relationships and understanding of the interrelationships between the subalpine hydro-micrometeorological variables and the physiological/physiographical characteristics.

Spatial Distribution of Energy Balance, Snow Water Equivalence and Snowmelt in Seasonally Snow Covered Watersheds
R. Bales, B. Davis, S. Fassnacht, N. Molotch (UA-HWR)

The scientific questions that have been addressed during the current reporting period have been aimed at improving the spatial and temporal continuity of remotely sensed snow water equivalence (SWE) estimates, which are limited due to cloud cover, and improving our understanding of the uncertainty associated with the up-scaling of point measurements of snow water equivalence. These efforts apply directly to the overall goal of the project, which is to develop an advanced set of tools for accurately estimating spatially distributed energy balance variables, snow accumulation and melt in seasonally snow covered catchments. Spatial and temporal continuity improvements were achieved by using spatially distributed temperature estimates to define the snow cover extent under clouds in the Salt-Verde basin, Arizona. The spatial continuity of the SWE estimates was improved with a mean error of 21%, indicating that the snow-covered area (SCA) beneath clouds can be determined with some success. Improved understanding of the uncertainty associated with the up-scaling of point SWE measurements was obtained by using intensive field observations to validate spatially distributed SWE estimates derived from the interpolation of the Natural Resource Conservation Service's (NRCS) Snow Telemetry station (SNOTEL) network. Field observations were collected at 7 sites during the 2001 and 2002 snowmelt seasons in the headwaters of the Rio Grande. Results showed that, at maximum SWE accumulation, SNOTEL stations overestimated the SWE across the study area domain. The overestimation of the SNOTEL stations resulted in an overestimation of spatially distributed SWE across the Rio Grande headwaters, with greater overestimation at peak SWE accumulation and increased accuracy as the melt season progresses. The data set collected during the field campaigns of 2001 and 2002, as well as the improved spatially distributed SWE estimates will provide a comprehensive ground truth estimate of SWE for modeling efforts in other thrust areas and sub-tasks, that can be used to validate/evaluate model-derived SWE.

Given the gaps in the hydrologic data acquisition needed for the goals of sub-task 1.02, the next logical step is to acquire and spatially distribute the remaining required hydrometeorological variables for the snowmelt modeling in the Rio Grande headwaters and Tokopah basin. The next step involves running a pair of snowmelt models.

1. A physically based energy balance snowmelt model based on the equations of SNTHERM.89 (Jordan, 1991).


2. A parameterized snow melt model based on the equations of the modified SRM (Brubaker et al., 1996) in which snowmelt is computed based on the radiation and temperature fluxes between the atmosphere and the snowpack.

The two models are being used in the two different basins so that model output uncertainty can be compared for the following 4 cases:

• High-resolution physically based modeling with low-input uncertainty
• Low-resolution physically based modeling with high-input uncertainty
• High-resolution parameter based modeling with low-input uncertainty
• Low-resolution parameter based modeling with high-input uncertainty

Outputs of spatially distributed snowmelt and SWE from both modeling efforts will be validated/evaluated based on the SWE data already collected during the 2001 and 2002 snowmelt seasons for the Rio Grande and during the 1997 and 1998 snowmelt seasons for the Tokopah Basin. This effort will afford an estimate of the change in uncertainty when the techniques are applied at different scales and with different input data accuracies (i.e., the change in uncertainty when the model is applied at the research basin scale versus the operational basin scale).

Hydrologic Modeling of Alpine Snow and Runoff
R. Bales (UA-HWR), S. Fassnacht (UA-HWR), Miller (LBL), B. Nijssen (UA-HWR), K. Dressler (UA-HWR)

Activities and Results

Two research questions are being addressed under this research: 1) What is the impact of incorporating spatial maps of snow properties (from sub-task 1.1.2) into energy balance and mass balance models on runoff potential from alpine snowmelt; and 2) Which model parameter(s) are most crucial for accurate streamflow estimates when used at different spatial scales, and parameter grouping schemes (e.g. Hydrologic Response Unit vs. Grid Response Unit). Both questions have been addressed using the Precipitation Runoff Modeling System (PRMS) and will continue to be addressed using the NOAH Land Surface Model (LSM), forming a link with TA4 in the integrated modeling framework. Key results include a decision on which interpolation method is optimal for assimilation of SWE data at the scale of the Colorado River Basin and the conclusion that vegetation representation from differing sources is significant in accurate prediction of cumulative streamflow in the Salt River Basin using the PRMS model. Snowmelt modeling, basin scale hydrologic modeling and model parameter improvement efforts will be directly integrated with the modeling efforts of TA4 to address the various established scenarios.

Plans

Over the next two years we plan to:

• Assimilate the currently available snow data products into hydrological models and evaluate the impact of this assimilation on the runoff potential and streamflow prediction. The hypothesis is that assimilation of these data products will lead to better snowmelt forecasts and will enable improved management of water resources in the Southwestern U.S. Emphasis will be placed on the Salt-Verde River system in the Colorado Basin and the Upper Rio Grande River Basin. To facilitate up-scaling of results to the SAHRA region at large, current modeling studies will implement the NOAH LSM as part of SAHRA's integrated modeling framework.



• Improve snow process parameterizations, in the context of NOAH LSM, based on the results of detailed snow process studies. In this way, the research will provide a link between the high resolution, location-specific process studies and the basin-wide, medium resolution modeling efforts undertaken by TA4.