Airborne Snow Surveys by State 198098

National Operational Hydrologie Remote Sensing Center

British Columbia

British Columbia

Figure 4 Number of airborne snow survey flight lines by state (1980-1998). National Operational Hydrologie Remote Sensing Center (NOHRSC).

pack, thereby providing an opportunity to collect volume integrated data (e.g., SWE).

Because of the difficulty of making field measurements in snow-covered mountainous regions, remote sensing has been pursued as a means of measuring snow-cover properties. The National Oceanic and Atmospheric Administration's (NOAA's) advanced very-high resolution radiometer (AVHRR) data have been routinely used for classification of snow-covered versus snow-free area (Matson et al., 1986; Matson, 1991; Xu et al., 1993). Differences between the spatial, temporal, spectral, and radiometric resolutions of different remote-sensing instruments result in tradeoffs between instruments for hydrologic applications. Optimization of one type of resolution generally involves some sacrifice in other types of resolution. For example, the Landsat thematic mapper (TM) has a much better spatial resolution than the AVHRR (30 m versus 1 km pixel size, respectively); however, the AVHRR can provide daily coverage of a given point, whereas the TM can only provide biweekly coverage. Development of accurate snow-cover information for areas with steep, variable topography characteristic of the western United States requires higher resolution data than are currently available from operational remote-sensing instruments or improved processing of the current data.

Passive microwave sensors are used to monitor snow, but three problems have limited its application. First, uncertainty in snow texture results in a significant noise in the calibration of a brightness temperature index with measured snow properties. Second, passive microwave imagery has a large pixel size that results in significant mixed pixels over areas with forest, mountains, or lakes. It appears unlikely that snow properties could be "unmixed" in these situations. This restricts operational use to large flat areas such as prairies and tundra. Finally, the signature from snow is indistinguishable from bare ground when the snow is wet, which requires special processing of time series and inference.

For mapping of snow properties of greatest hydrological importance, a synthetic aperture radar (SAR) with some special characteristics is necessary. A SAR is sensitive to many snow parameters such as snow density, depth grain size, free-liquid water content, and snow-pack structures that hydrologists use. It can image day or night in all weather, and it has a fine spatial resolution compatible with topographic variation affecting snow distribution. Experiments as part of the SIR-C/X SAR missions have significantly advanced and demonstrated the capabilities of new multi-frequency and multipolarization SAR—to map both wet and dry snow covers (Shi and Dozier, 1997), to infer snow wetness (Shi and Dozier, 1995), and to estimate snow density and depth and thereby snow water equivalent (Shi and Dozier, 1996).

However, operational satellites do not provide the necessary data. To achieve sufficiently high spatial resolution to measure the variability of SWE in mountainous areas, a multiple-frequency, multiple-polarization SAR is required. At present, the measurement of the spatial distribution of SWE, and total snow volume within a mountainous basin, must be performed by intensive field sampling to attempt to represent the large spatial variability of alpine snowpacks. Logistics and safety limitations generally restrict the number of field samples that may be so obtained (Elder et al., 1991a). Thus the problem of determining the volume and distribution of snowpack water storage within mountain basins remains acute.

Climate Models

While still not an operational tool, the use of high-resolution regional climate models for simulating seasonal and interannual changes in snowpack in areas such as the western United States is quite promising. At a 60-km resolution such a model can reproduce the overall patterns of measured precipitation and snow cover and to some extent year-to-year variations (Figs. 5a and 5b) (Seth et al., 1999). Errors in simulating the actual magnitude of seasonal snow accumulation arise in large part from lack of realism in model topography, with inadequacies in model parameterization also a factor. However, climate modeling offers great promise as a tool that can be integrated with ground-based and satellite observations.

Emerging Technologies

The National Aeronautic and Space Administration's (NASA's) moderate resolution imaging spectrometer (MODIS) will provide near-daily global coverage (comparable

Decern ber-February

June-August

September-November

Figure 1 (Chapter 3) Climatological distribution of tropospheric ozone derived from satellite measurements between 1979 and 2000 (from Fishman et al., 2002). Units Äcontours and Dobson Units (DU). R»gions greater than 40DU h«ve been shaded.

Ozone(ppbv) 0 20 40 60 80 100

SPIRAL

SPIRAL

Figure 5 (Chapter 14) Ozone plume over the Pacific seen during the PEM-aircraft mission in Sept.-Ölet. 19%. (from Fenn et al., 1999).

End date:

8<i>W 7 fcdw 6<Jvtf 3DW ¿5W

1.25 1.5 1.75 2.0 2.25 2.5 2.75 3.0 3.25 3.5 3.75 4.0 4.25 4.5 4.75 5.0 6.0 7.0 8.0 10.0

Height (km)

Fijuiv S (Chnplei 14) Composite of fot"i»ard trajectories from Cuiaba during Ihe 19L)5 SCAH-H Held experiment. A Brazilian varsiOii of ihe (Colorado Stale mesoscate RAMS model was uMtl to provids winds for the University of Sao Paulo kinematic trajectory model (from Lomio c* al., JW9l.

I CO Mixing Ratio (ppbv)

Figure 10 (Chapter 14) (a) MAPS CO, April 1994 (from Christopher et ill., 1998).

-7Rres,

I CO Mixing Ratio (ppbv)

Figure 10 (Chapter 14) (a) MAPS CO, April 1994 (from Christopher et ill., 1998).

Figure 10 (Chapter 14) (b) toincident fixes during April 14194 Spacs Shutrte flight (from Christopher et al., 19980.

MODIFIED RESIDUAL TROPOSPHERIC 03 (DOBSON UNITS)

180 120 -60 0 60 120 J80

Longitud

Figure 12 (Chapter 14) Wave-one pattern in tropospheric ozone apparent in TOMS satellite data, averaged from 2 maps/month during the 1979- 1992 Ninibn? 7 observing period. Wave appears to be present throughout year. Scale is DU (Dohson units). Cf. Figure A1 in Thompson and Hudson (L999).

180 120 -60 0 60 120 J80

Longitud

Figure 12 (Chapter 14) Wave-one pattern in tropospheric ozone apparent in TOMS satellite data, averaged from 2 maps/month during the 1979- 1992 Ninibn? 7 observing period. Wave appears to be present throughout year. Scale is DU (Dohson units). Cf. Figure A1 in Thompson and Hudson (L999).

High Tropical Tropospheric Ozone Column from El-Nino Period N7/T0MS, Oct. 16—Oct. 31, 1982

1S0E 180E

W 150W 12QW 90W )W

SOE 120 E 1S0E 1B0E

80 DV UMD/gsfc

Ffgure 14 (Chapter 14) Tropospheric column ozone (in DU, from modified-residual method, Thompson and Hudson, 1999) during El Kifio-Southern Oscillation (ENSO) of late 1982 (upper panel) as seen in tropical tropospheric ozone map and for September 1997 (lower panel).

Figure 26 (Chapter 16) The brown discoloration resulting from an atmosphere containing nitrogen dioxide (1H5,) being snaded 1|» clouds but viewed against a clear blue sky. Light scattered by particulate matter in lbs atmosphere can cominate light absorbed by NO,, sausing * jrny or blue appearing haze (left side of photograph).

JAN APR JUL OCT

Figure 3 iChapter 21) Two-dimensional (latitude/season) representation of total column ozone as measured by TOMS for the penod 1978 to 1993.

JAN APR JUL OCT

Figure 3 iChapter 21) Two-dimensional (latitude/season) representation of total column ozone as measured by TOMS for the penod 1978 to 1993.

EP/TOMS Total Ozone for Oct 16. 1999

orocji^orocji^j s cji o fo CJI o ocnocnocnocjiocjiocjioui otno

orocji^orocji^j s cji o fo CJI o ocnocnocnocjiocjiocjioui otno

Dobson Units Dark Gray < 100, Red > 500 DU

Figure 5 (Chapter 21) Map of total column ozone diver the Antarctic as determined from TOMS October 16, 1999.

-iognormal model

Nonparametric hierarr :al model

Figure 6 (Chapter 33) Comparison of observed and downscaled rainfall fields (July 1997) (from Kang and Ramirez46).

-iognormal model

Nonparametric hierarr :al model

Figure 6 (Chapter 33) Comparison of observed and downscaled rainfall fields (July 1997) (from Kang and Ramirez46).

Figure 1 (Chapter 36) Quebrada San Julian upstream of Caraballeda showing evidence of ne«Bnt debris flows and flash floods. Note the high slope angles, large numbers of debris flow scafs» and abundance of new alluvium and colluvium in the channel bed and fan surface.

Figure 2 (Chapter 36) (4 panels} These scenes show various sections of the Mississippi River near St. Louis before and just after the 1993 floods, whi»h peaked in late July/early August. The imnfw show the area as seetl by the LaudSat Thematic Mapper (TM) instrument. The short-wave inftawd (TM bsnd 5), infrared (TM band 4), and visible green (TM band 2) channels am displayed in the images as red, green, and blue, respectively. Ill fhis combination, barren and/or recently cultivated land appears red to pink, vegetation appears green, water is dark blue, and artificial structures of concrete and asphalt appear dltrrte gray or blaek. Rjfddish areas in the scenes during the Hood show where water had started to recede, leaving barren land.

Figure 2 (Chapter 36) (4 panels} These scenes show various sections of the Mississippi River near St. Louis before and just after the 1993 floods, whi»h peaked in late July/early August. The imnfw show the area as seetl by the LaudSat Thematic Mapper (TM) instrument. The short-wave inftawd (TM bsnd 5), infrared (TM band 4), and visible green (TM band 2) channels am displayed in the images as red, green, and blue, respectively. Ill fhis combination, barren and/or recently cultivated land appears red to pink, vegetation appears green, water is dark blue, and artificial structures of concrete and asphalt appear dltrrte gray or blaek. Rjfddish areas in the scenes during the Hood show where water had started to recede, leaving barren land.

120W

100W

120W

100W

0 100 200 300 400 500 600 700 800

JFM 95 REGCM

110W

120W

100 200 300 400 500 600 700 800

Figure 5 January, February, March 1995 total precipitation (mm) from (a) NCDC+SNOW-SNOWTEL; and (6) simulated by RegCM.

to AVHRR), but at spatial resolutions ranging from 250 m to 1 km. The sensor has two channels in the visible and near-infrared spectral bands at 250-m resolution, five channels in the visible, near-infrared, and short-wave infrared at 500-m resolution, and the remaining 29 MODIS channels have a spatial resolution of 1 km. The MODIS sensor has on-board visible/near-infrared calibrators, while the AVHRR does not; thus one will be able to derive radiances over snow using some of the MODIS sensors. At least one of the visible MODIS sensors will not saturate over snow. This will be an advancement over the AVHRR and TM sensors that experience significant detector saturation over snow and ice targets in the visible channels. The standard MODIS snow and ice products will consist of 500-m or 1-km resolution binary maps of snow and ice cover, respectively, produced on a global, daily basis in most months.

An approach to modeling spatially distributed snowmelt in steep, alpine basins was proposed using net potential radiation, distributed across the basin using a digital elevation model, as the main factor determining relative snowmelt (Elder et al., 1991b). However, to date this approach has only been applied to small, headwater catchments. It is also possible to infer SWE after the fact from measurements of snow-cover depletion. With a time series of snow cover, e.g., from TM, AVHRR, or MODIS imagery, one can tell when the snow cover disappears, i.e., when snow water equivalence goes to zero. Then using a spatially distributed snowmelt model, one can back calculate from the time snow cover disappears at a point, and then infer the starting value of snow water equivalence. This method has been implemented using TM scenes for a small watershed in the Sierra Nevada, California (Cline et al., 1998).

Research shows that remote sensing also allows estimation of several hydrologic variables important for snowmelt modeling. From airborne hyperspectral sensors [currently NASA's advanced visible/infrared imaging spectrometer (AVIRIS)] one can estimate snow grain size, albedo, liquid water content in the surface layer, and subpixel coverage. Using two-frequency, co-polarized synthetic aperture radar, one can map both snow through thick cloud cover and estimate liquid water content accurately. Work on estimation of snow water equivalence is continuing, with promising results from SIR-C/X-SAR, from photogrammetry and from snowmelt modeling with time-series SCA data. These capabilities have been developed largely using experimental sensors that are not included in currently scheduled satellite launches. A fully automated method of subpixel snow cover mapping uses Landsat TM data to map snow cover in the Sierra Nevada and make quantitative estimates of the fractional snow-covered area within each pixel (Rosenthal and Dozier, 1996). Snow fraction estimates from the satellite data can be as accurate as those attainable with high-resolution aerial photography, but they are obtained faster, at much lower cost, and over a vastly larger area.

An important emerging technology is the use of spatially distributed, energy balance snow models to describe the accumulation and ablation of snowpacks. These models organize a wide variety of hydrometeorological and terrain information and permit an improved understanding of snowpack evolution throughout an area of interest. These models have been implemented primarily in well-instrumen ted research basins, where high-quality meteorological measurements are available to drive the models. However, recent applications have extended their use to larger regions, where mesoscale numerical weather prediction (NWP) model analyses are used to drive the models (e.g., Cline and Carroll, 1999). The NWS NOHRSC is currently developing a four-dimensional data assimilation system for snow estimation that will use a spatially distributed, physically based snow energy and mass balance model, mesoscale NWP analyses, and an updating scheme to provide operational SWE estimates for the United States.

Gaps in Measurement and Understanding

The most significant constraint on the development of snow hydrology is the general lack of measurements of SWE. Measurements are very sparse throughout the United States, especially in the eastern United States. Most operational in situ SWE measurements are collected to support empirical models that are used to estimate snowmelt runoff. For this purpose, the location of the measurement does not necessarily have to be representative of the surrounding area—it is the relationship between the SWE at particular sites and runoff that is important. This means that most operational SWE measurements (e.g., snow courses and snow pillows) are not reliable indicators of the distribution of SWE in a given area. Rather they are index sites that have snow cover for much of the season, to better support empirical modeling. Furthermore, there are simply too few measurements available to adequately characterize the spatial variability of SWE. The general lack of SWE measurements imposes a significant constraint on the development of improved remote-sensing procedures and distributed snow models, which requires "ground truth" for validation and parameter estimation.

The spatial variability of SWE and snowmelt processes in models needs to be better understood. This is a key area for future development of snow hydrology. Most physical snow process work has been carried out at the point scale, or at very local scales. However, most hydrologic and atmospheric modeling scales are run at much coarser spatial scales that involve significant variability of important processes. These models must parameterize the subgrid variability in some manner, but little effort has been devoted to this problem in snow hydrology. Improved understanding of the variability of SWE and snowmelt processes is also needed to design improved sampling strategies for field measurements. High spatial resolution remote-sensing observations of SWE and other snowpack characteristics using SAR would significantly improve our understanding of the spatial variability of snow properties.

4 ESTIMATION OF SNOWMELT RUNOFF Historical Approach for Operational Forecasts

Both conceptual and physical approaches have been employed in snowmelt runoff modeling. Conceptual models propose a mathematical relationship between snow-

melt and measured quantities; thus melt can be calculated without treating in detail all of the physical processes and parameters that affect snowmelt. Conceptual models have the benefit of requiring less informational input but suffer from the uncertainty that the conditions hypothesized under different model scenarios are modeled sufficiently.

Operationally, the NWS is tasked with forecasting streamflow, floods, and seasonal water supplies in the United States. Various operational data are used by the algorithms making up the NWS River Forecast System (NWSRFS) to produce streamflow and water supply projections that extend several hours to months into the future. Snow accumulation and ablation is modeled in NWSRFS using an empirical approach that is driven using inputs of air temperature and precipitation (Anderson, 1973; Day, 1990). During periods of precipitation, a simplified energy balance approach is used to estimate snowpack state conditions. When no precipitation is occurring, a simple temperature index method is used. The models are spatially lumped, that is, they model the "average" snowpack state conditions over entire basins or subbasins.

Short-term streamflow forecasts are made using NWSRFS and meteorological forecasts of temperature and precipitation. Because meteorological forecasts become less reliable the further out they extend, this method of streamflow prediction is limited to periods of a few hours to a few days. Beyond that time, the short-term meteorological forecasts are blended into the climatologically average conditions. The long-range forecasting component of NWSRFS, called the extended streamflow prediction (ESP) technique, uses present-day streamflow, soil moisture, and snowpack conditions along with a historical time series of precipitation and temperature to estimate streamflow weeks or months into the future.

Two major factors contribute to high uncertainty in estimates of snowpack conditions, particularly in mountainous regions. First, the NWSRFS snow models contain several empirical parameters that must be calibrated in conjunction with the rest of the NWSRFS algorithms. The empirical calibrations are useful, as they help overcome certain problems, such as poorly representative temperature or precipitation measurement sites. The snow models perform best when conditions are near the average conditions used in the calibration. During extreme or unusual conditions, the models often produce spurious results. Second, accurate estimation of precipitation is critical for these models, but this is especially challenging in mountainous regions. Like the distribution of SWE, the distribution of precipitation is typically highly variable in mountain regions, and there are too few measurement sites to adequately represent this variability. The problem is compounded by the fact that precipitation gages do not measure snowfall well. Consequently, large uncertainties in precipitation inputs to the snow models propagate to the estimates of snow state.

Since current snow state conditions in the model serve as initial conditions for streamflow forecasts, the large uncertainties in modeled snowpack state conditions must be reduced where possible prior to extending the model forward in time. This is accomplished by updating the SWE in the snow models with observed SWE.

In the western United States, two data assimilation schemes are currently used to provide SWE updates for the snow models. The first is a relatively simple approach that uses satellite observations of snow cover to indicate areas without snow, plus an interpolation of all available surface- and airborne-based SWE observations. The second data assimilation scheme, SEUS (McManamon et al., 1993), assimilates surface, airborne, and satellite snow information in an optimal interpolation framework. Grid points in a basin are classified and the amount of SWE and snowmelt for each class for each year in a historical record is estimated. The historical mean SWE fields are then used to estimate the actual SWE values from current gridded data.

Spatially Distributed Modeling of Melt and Runoff

While empirical snowmelt runoff models have traditionally been useful for operational runoff volume forecasts, they provide little information on the timing, rate, or magnitude of discharge, and they are inappropriate in situations outside the boundary conditions governing the development of the relevant empirical parameters. Thus they may fail to adequately predict water yield in extreme or unusual years, and they cannot be reliably used in investigations examining snowmelt responses to climate variability and change. These problems, and the increasing importance of understanding intrabasin snowmelt for environmental analysis of such factors as basin ecology (Baron et al., 1993), water chemistry (Wolford et al., 1996), and hillslope erosion (Tarboton et al., 1991) have motivated the development of physically based, spatially distributed snowmelt models in recent years. Such models require information on the spatial distribution of snowpack water storage. But mountain snowpacks are spatially heterogeneous, reflecting the influences of rugged topography on precipitation, wind redistribution of snow, and surface energy fluxes during the accumulation season (Elder et al., 1991a), and no widely suitable method yet exists to directly map SWE or simulate distributed snowmelt in rugged mountain regions. One of the main obstacles to physically based modeling is the compilation of the necessary meteorological and snow-cover data to run, calibrate, and validate such models. For example, basin discharge has frequently been used as the sole physical criterion of model calibration and performance assessment for conceptual snowmelt models. But as it is an integrated response to melt and runoff, basin discharge is not sufficient to discriminate between the effects of the multiplicity of data inputs driving physical models and that distributed snow-cover data are required to assess model performance (Bloschl et al., 1991a, 1991b).

A distributed snowmelt model consists of two general components: a model for calculating melt at a single point (given a set of prescribed snowpack and meteorological conditions), and a method of developing the requisite snowpack and meteorological data for all points within the basin. Snowmelt modeling efforts require several steps necessary to couple basinwide energy balance snowmelt models with remote sensing and flow routing. The model topographic distribution of solar radiation (TOPORAD) (Dozier and Frew, 1990) uses information on watershed topography (i.e., a digital elevation model) to spatially distribute radiation. Simpler models are used to spatially distribute other energy balance components. These radiation maps are then used as inputs to models to estimate the distribution of snow around the basin prior to snowmelt (i.e., initial conditions for melt), and as inputs to the point snowmelt model.

Climate Change and Variability Issues

Expected changes in global climate are cause for concern for future water resources in the western United States, where limited water supplies are already in great demand, and a constrained water regulatory system struggles to satisfy water users. The main question for the regions is: What does potential climate change mean to snowmelt-dominated western water resources? Certainly the assumption of stationarity of the present snowmelt and streamflow regime must be questioned (Dracup and Kendall, 1990); however, it is not clear what exactly should take its place. One criterion with which many global climate models are run is the anticipated doubling of greenhouse forcing from atmospheric greenhouse gases; this doubling and the resulting modeled climate changes are expected to occur by the mid-twenty-first century, well within the "design life" of many existing water control structures and regulations, and indeed within the careers of the next generation of water planners. For this reason, the problem cannot be easily ignored; however, rapid advances in understanding of the potential effects of climate change on snowmelt-dominated hydrologic systems are necessary before an informed response by water resource planners can be made.

A variety of approaches have been used to estimate the effects of or sensitivity to climate change in snowmelt-dominated basins in the western United States. The methods range from purely statistical techniques (Duell, 1992), regression models, to complex deterministic spatial modeling strategies that explicitly account for every process in as detailed a manner as possible (Leavesley et al., 1992). Potential effects of climate change on snowmelt and streamflow in the west include reduced annual streamflow, earlier peak flows in spring, and less winter precipitation occurring as snow. However, due to the great uncertainty in the input data, and the general inapplicability of the models to forecasts beyond the range of their prior calibration, magnitudes remain largely unknown. There are also issues with the variety of models in use, and few cases where a common approach has been employed in more than one hydroclimatic region.

Two points illustrate the tentative and qualitative nature of work to date and highlight the need for a more rigorous approach. First, examination of the major topographic effects on temperature and precipitation in the western United States shows that only regional-scale climate models are appropriate for use in this part of North America. The results of any modeled climate change scenario for the western United States that has not incorporated reasonable representations of topography must be viewed as spurious at best (Seth et al., 1999). In addition to, and because of, the lack of adequate topographic representation in general circulation models (GCMs), several of the important consequences of topographically controlled precipitation and temperature, such as albedo differences due to the presence of snow at high elevations, and forced uplift and cooling of large-scale airflow over individual mountain ranges are also missing from GCMs. This suggests that it may not be feasible to treat GCM results as even a general trend or rough estimate of realistic climate changes. This is not so much a criticism of GCMs in general, but rather a criticism of the use of their output for modeling the hydrology of basins with topography of which the GCMs are not even aware. Thus accurate quantitative climate change scenarios are conspicuously absent from exercises attempting to predict climate change effects on streamflow in the west.

The second point of the argument is that hydrologic simulations are severely limited by the excessive need to calibrate snowmelt runoff models to achieve a good fit to a basin hydrograph. Under the assumption of stationarity, such calibration is acceptable if runoff is the only variable of interest. Unfortunately, there is absolutely no reason to assume that if climate changes the hydrologic characteristics of a basin will remain stationary. If internal basin hydrologic characteristics are also of interest, then most existing snowmelt runoff models cannot be used since they do not provide any information on internal basin processes. In fact, most of the models depend on sufficient aggregation of internal processes so that accurate outflow predictions may be made. Distributed parameter models appear to address the internal basin processes to a larger degree, but often require more extensive data.

In most studies to date of the potential effects of climate change on snowmelt runoff, the "effect" receiving the greatest attention is the basin hydrograph. Internal basin characteristics should be given greater attention for two reasons. First, basins themselves are important, and we should be concerned about how they may be affected by changing climate. Second, we need to begin incorporating the notion of nonstationarity of basin properties into hydrologic models. For example, climate-change-induced changes in the extent and characteristics of forest cover within a basin might have a greater effect on both the timing and magnitude of snowmelt runoff than changes in temperature or precipitation. Changes in the proportion of rainfall/snowfall could lead to important changes in sediment transport, mass wasting, and other geomorphic characteristics of a basin, which would affect appropriate selection and use of hydrologic models. The hydrochemistry of streams and rivers is largely dependent on the interaction of water with various basin component; thus intrabasin biological and geomorphic changes due to climate change could have important implications for stream water chemistry. Improvements in distributed-parameter hydrologic models and particularly in methods of collecting sufficient input data for these models will certainly be necessary before internal basin processes can be adequately examined.

REFERENCES

Anderson, E. A., National Weather Service River Forecast System: Snow Accumulation and Ablation Model, NOAA Technical Memorandum, NWS, Hydro-17, U.S. Department of Commerce, Washington, DC, 1973. Baron, J. S., L. E. Band, S. W. Running, and D. Cline, The effects of snow distribution on the hydrologic simulation of a high elevation Rocky Mountain watershed using Regional HydroEcological Simulation Systems, RHESSys, Eos, Trans. Am. Geophys. Union, 74(43), 237, 1993.

Bloschl, G., D. Gutknecht, and R. Kirnbauer, Distributed snowmelt simulations in an Alpine catchment 1. Model evaluation on the basis of snow cover patterns, Water Resour. Res., 27(12), 3171-3179, 1991a.

Bloschl, G., D. Gutknecht, and R. Kirnbauer, Distributed snowmelt simulations in an Alpine catchment 2. Parameter study and model predictions, Water Resour. Res., 27(12), 31813188, 1991b.

Cline, D., K. Elder, and R. Bales, Scale effects in a distributed snow water equivalence and snowmelt model for mountain basin, Hydrol. Process., 72(10-11), 1527-1536, 1998.

Cline, D., and T. Carroll, Inference of snow cover beneath obscuring clouds using optical remote sensing and a distributed snow energy and mass balance model, J. Geophys. Res. Atmos., 104( D16), 19631-19644, 1999.

Day, G. N., A Methodology for Updating a Conceptual Snow Model with Snow Measurements, NOAA Technical Report, NWS 43, U.S. Department of Commerce, Washington, DC, 1990.

Dozier, J., and J. E. Frew, Rapid calculation of terrain parameters for radiation modeling from digital elevation data, IEEE Trans. Geosci. Remote Sensing, 28, 963-969, 1990.

Dracup, J. A., and D. R. Kendall, Floods and droughts, in P. E. Waggoner (Ed.), Climate Change and U.S. Water Resources, Wiley, New York, pp. 243-267, 1990.

Duell, L. F. W., Jr., Use of regression models to estimate effects of climate change on seasonal streamflow in the American and Carson River Basins, California-Nevada, in Hermann, Raymond, ed., Managing water resources during global change: 28th Annual Conference, American Water Resources Association, Reno, Nev., November 1992, Proceedings, p. 731-740.

Elder, K., R. E. Davis, and R. C. Bales, Terrain classification of snow-covered watersheds, Proc. Eastern Snow Conf, 48, 39^19, 1991a.

Elder, K., J. Dozier, and J. Michaelsen, Snow accumulation and distribution in an alpine watershed, Water Resour. Res., 27, 1541-1552, 1991b.

Hart, D., and F. Gehrke, Status of the California Cooperative Snow Survey Program, in 58th Western Snow Conference Proceedings, Sacramento, CA, 1990, pp. 9-14.

Leavesley, G. H., M. D. Branson, and L. E. Hay, Using coupled atmospheric and hydrologic models to investigate the effects of climate change in mountainous regions, in Hermann, Raymond, ed., Managing water resources during global change: 28th Annual Conference, Americal Water Resources Association, Reno, Nev., November 1992, Proceedings, p. 691700.

Lin, H. F., F. K. Hare, and K. P. Singh, Influence of the atmosphere, in M. G. Wolman and H. C. Riggs (Eds.), Surface Water Hydrology: Boulder, Colorado, Geological Society of America, The Geology of North America, 1990, pp. 11-53.

McManamon, A., T. L. Szliga, R. K. Hartman, G. N. Day, and T. R. Carroll, Gridded snow water equivalent using ground-based and airborne snow data, in Proceedings of the 50th Eastern Snow Conference, Quebec City, 1993, pp. 75-81.

Matson, M., C.F. Roeplewski, and M.S. Varnadore, 1986: An Atlas of Satellite-Derived Northern Hemisphere Snow Cover Frequency, National Weather Service, Washington D.C., 75 p.

Matson, M., 1991: NOAA satellite snow cover data, Palegeography and Paleoecology, 90: 213 280

Paulson, R.W., E.B. Chase, R.S. Roberts, and D.W. Moody, compilers, National Water Summary 1988-89: Hydrologic Events and Floods and Droughts, U.S. Geological

Survey Water Supply Paper 2375, U.S. Government Printing Office, Washington, DC, 1991, p. 591.

Rosenthal, W-, and J. Dozier, Automated mapping of montane snow cover at subpixel resolution from the Landsat Thematic Mapper, Water Resour. Res., 32(1), 115-130, 1996.

Seth, A., R. C. Bales, and R. E. Dickinson, A framework for the study of seasonal snow hydrology and its interannual variability in the alpine regions of the Southwest, J. Geophys. Res., 104(D18), 22117-22135, 1999.

Shi, J. and J. Dozier, Inferring snow wetness using C-band data from DIR-C's polarimetric synthetic aperture radar, IEEE Trans. Geosci. Remote Sensing, 33(4), 905-914, 1995.

Shi, J., and J. Dozier, Estimation of snow water equivalence using SISR=C/X-SAR, in Proceedings IGARRS 96 , IEEE No. 96Ch35875, 1996, pp. 2002-2004.

Shi, J., and J. Dozier, Mapping seasonal snow with SIR-C/X-SAR in mountainous areas, Remote Sensing Environ., 59, 294-307, 1997.

Shi, J. and J. Dozier, Estimation of Snow Water Equivalence Using SIR-C/X-SAR Image Data, Proceedings Progress in Electromaftietic Research Symposium, p. 1041, 1995.

Tarboton, D. G., M. J. Al-Adhami, and D. S. Bowles, Preliminary comparisons of snowmelt models for erosion prediction, Proc. Western Snow Conf, 59, 79-90, 1991.

Wolford, R. A., R. C. Bales, and S. Sorooshian, Development of a hydrochemical model for seasonally snow-covered alpine watersheds: Application to Emerald Lake Watershed, Sierra Nevada, California, Water Resour. Res., 32(4), 1061-1074, 1996.

Xu, H., J. O. Baily, E. C. Barrett, and R. E. J. Kelly, Monitoring snow area and depth with integration of remote-sensing and GIS, Int. J. Remote Sensing, 14(17), 3259-3268, 1993.

0 0

Post a comment