Performance Comparison of an Energy Budget and the
Performance Comparison of an Energy. Budget and the Temperature Index-Based (Snow-17) Snow Models at SNOTEL Stations Fan Lei 1, Victor Koren 2, Fekadu Moreda 3, and Michael Smith 2 1 Riverside Technology, inc 2 Hydrology Laboratory, Office of Hydrologic Development, NWS/NOAA 3 MHW, Inc
Motivation Ø Improve National Weather Service (NWS) water resources forecasts by using energy budget models of snow accumulation and melt.
Background Ø Simple degree-day, conceptual lumped model is currently used to model snow accumulation and melt for NOAA/NWS operational river forecasting. Ø Energy-budget snowmelt models are physically more consistent and they require no (or much less) calibration. Ø Without reliable driving fields of meteorological data, the application of energy-budget snowmelt models is limited so far. Ø Emerging meteorological data may lead to better performance of energy-budget snowmelt models. Ø NWS Office of Hydrologic Development (OHD) is conducting research on transitioning from conceptual to energy-budget snowmelt modeling to improve current operational river forecasts.
Model Description (1) Ø SNOW-17 • A current operational snowmelt component in the NWS River Forecast System (NWSRFS), Developed by Anderson (1973, 1976); • Uses air temperature as index of major snow processes; • Model performs well after calibration; • Being tested in distributed mode (HL-RDHM, Moreda et al. , 2005)
Model Description (2) Ø Energy-Budget Snowmelt Model (EBSM) • • • One layer model linked to multilayer soil/vegetation scheme (a version of Eta-LSS, Koren et al [1999]); Energy forcings are described by meteorological fields, including: surface air temperature, surface downward short wave flux, surface downward long wave flux, surface wind and surface humidity; Model does not include conceptual type parameters, no (or very little) calibration is needed.
Nevada Carson River Basin Pacific California Ocean Test Basin Ø Considering snow data availability, Carson River Basin is selected as the test basin. Carson River Basin elevation (units: m)
Data (1) SNOpack TELemetry (SNOTEL) ground measurements • Hourly temperature, precipitation since 1997; • Daily snow water equivalent. Ø North American Regional Reanalysis (NARR) • Based on National Centers for Environmental Prediction (NCEP)'s mesoscale Eta forecast model and Eta Data Assimilation System (EDAS); • 3 -hourly 2 m air temp. , 2 m relative humidity, surface downward long wave radiation, 10 m surface wind, precipitation; • 0. 375 degree (about 32 km) resolution. Ø
Data (2) Ø GEWEX [Global Energy and Water Cycle Experiment] Continental Scale International Project (GCIP) and GEWEX America Prediction Project (GAPP) Surface Radiation Budget (SRB) Data l Re-processed hourly averaged surface downward short wave flux; 1/8 degree (about 16 km) resolution. Ø North American Land Data Assimilation System (NLDAS) l Surface albedo, Leaf Area Index (LAI), (Greeness FRACtion) GFRAC, Soil type, vegetation type, etc; l 1/8 degree (about 16 km) resolution across North America; l Some of the parameters are adjusted in energy-budget snow melt model. l
Experiment Design 1999 water year was selected for experiments, based on data availability and quality; Ø Snow Water Equivalent (SWE) was selected as main snow property; Ø Extracted NLDAS data are used as EBSM model parameters. Ø • LAI and GFRAC are manually adjusted to match the sites land cover. Ø Extracted NARR data, SNOTEL ground measured Temp. & Precip. were applied as model inputs; • NARR Temp. are adjusted for elevation. Ø Both models are run to generate: l l l Point SWE simulations, Basin SWE simulations (on going), Basin outlet hydrographs (in plan).
Experiment Design (2) Accumulated Precipitation from NARR and SNOTEL Precip-NARR
Results: Observed and simulated SWE using Snotel precip. SWE-Measured EBSM-TSnotel SN 17 -TSnotel EBSM-T 2 m-NARR SN 17 -T 2 m-NARR
Results: Observed and simulated SWE using NARR precip. SWE-Measured EBSM-TSnotel SN 17 -TSnotel EBSM-T 2 m-NARR SN 17 -T 2 m-NARR
Discussion Ø The two models show reasonable agreement with each other and with the ground measurements, given reasonable temperature and precipitation data. Ø Both models are very sensitive to temperature especially during accumulation periods. Ø The experiments indicate that with the elevation adjustment, the temperature data interpolated from NARR may be used to drive the EBSM, although some model fitting may be needed. Ø Given the highly spatially-variable nature of precipitation in mountainous areas, special treatment is necessary or other more reliable data sources need to be explored.
- Slides: 13