The National Operational Hydrologic Remote Sensing Center Operational Snow Analysis

Transcription

1 The National Operational Hydrologic Remote Sensing Center Operational Snow Analysis World Meteorological Organization Global Cryosphere Watch Snow-Watch Workshop Session 3: Snow Analysis Products Andrew Rost NOHRSC Director

2 The NOHRSC Mission To support a Weather-Ready Nation by producing the best possible estimate of snow water equivalent using all available data including atmospheric model state variables, satellite imagery, airborne measurements, and ground observations

3 The NOHRSC Mission employing state-of-the-art operational snow, land surface modeling, and data assimilation

4 Outline Snow modeling and data assimilation for the coterminous United States Land surface modeling and data assimilation for Alaska Land surface modeling for Central Asia Future land surface modeling and data assimilation plans for North America Gaps, issues, and potential contributions

5 Snow Modeling and Data Assimilation for the Coterminous United States 2003: Transitioned from linear detrending spatial interpolation model to energy and mass balanced fullphysics snow model Simplified SNOTHERM and Prairie Blowing Snow models Domain: Coterminous US (8M sq km), southern Canada (54N), 30 arc second resolution on lat/lon grid Hourly time steps, four 6-hour model cycles, daily data assimilation (where and when required), no update (reanalysis) period 72-hour forecasts; first 24 hours driven by Rapid Refresh; next 48 hours North America Model

6 Snow Modeling and Data Assimilation for the Coterminous United States Forcings Snow and non-snow precipitation Air temperature Solar radiation Wind speed and direction Relative humidity Principle Model State Variables Snow water equivalent Snow density Snow depth Snow melt Snow sublimation Blowing snow Snow temperature Sensible and latent heat

7 Full-Physics Energy and Mass Balanced Snow Model Forced by down-scaled Rapid Refresh and North America Model numerical weather prediction model data Models mass and energy fluxes between the atmosphere, 3 snowpack layers, and 2 thermally active soil layers Runs on a modest cluster of of four 8-core servers

8 Assimilate All Available In Situ Observations (15M Observations / 37K Stations) The NOHRSC Operates a Real Time National Collective of Observation Collectives

9 Data Assimilation Only Where and When Needed Only Where Observed and Modeled Differences are Spatially Correlated Direct Insertion of 3D Spatially Interpolated Differences Surprising Agreement Between Observations and Modeled Snowpack States

10 6100 In Situ SWE Observation Sites

11 Potentially an Additional 610 In Situ SWE Observation Sites in Canada And the promise of more to come The NOHRSC collaborates closely with Federal, Provincial, and private Canadian data providers

12 Airborne Gamma Surveys Fills in situ observations gaps Partial emancipation from the representativeness of a point debate 10 miles long by 1,000 feet wide (2 square mile area)

13 Airborne Gamma Surveys Atmospheric Radon Cosmic Radiation Potassium, Uranium, and Thorium

14 2500 SWE Gamma Flight Lines And Growing

15 28000 In Situ Snow Depth Observation Sites Supplement in situ SWE observations Most SWE observation sites report infrequently Convert snow depth observations into SWE using modeled snowpack density

16 MODIS Fractional Snow Cover Run T Painter s MODSCAG and MOD09_SPA internally Identifies the edge of the snowpack Allows us to assimilate zero snow observations Constrains the areal extent of modeled snow pack states

17 Benefit Impact on River Forecasting Hydrographs are the grand integrator Evidence that assimilating in situ point observations do not have a negative on modeling RFC Forecast without NOHRSC Data RFC Forecast Updated with NOHRSC Data Simple Lumped Model SWE NOHRSC Modeled SWE Observed Stage Underestimated Forecast Stage Corrected Forecast Stage

18 NOHRSC Stakeholders National Weather Service Federal and State Agencies Private Sector Canadian 13 River Forecast Centers Weather Forecast Offices U.S. Army Corps of Engineers Bureau of Reclamation New York Department of Environmental Protection Natural Resources Conservation Service Department of Transportation Montana Department of Emergency Services San Francisco Public Utilities Commission University of Albany ASRC/CESTM University of Wisconsin Sea Grant Institute National Snow and Ice Data Center FEMA Baron Advanced Meteorological Systems, LLC Weather Channel Meteorlogix, Inc. Merrril Lynch Weather Decision Technologies, Inc. SnowStreet AccuWeather Snow Plow Operators Oppenheimer Campbell Soup Company Snowmobile outfitters Mountaineers General Public Manitoba Department of Natural Resources New Brunswick Department of Natural Resources Alberta Environment BC Hydro British Columbia Ministry of Environment Environment Canada Saint John River Basin Commission

19 PRODUCTS Hourly model time steps 1 km 2 Resolution 30 Hr Forecast (72 FY12) INTERNET Interactive Maps 3D Visualization e.g. Google Earth Time-series loops National/Regional Discussions Text summaries by watershed Point Queries DIRECT FEED Push or Pull Gridded Data Flat Binary or GIS-ready

20 Land Surface Modeling and Data Assimilation for Alaska 2008: Explored operational LIS for AK Domain: 42 x 24 at 0.01 ; 10.1 x 10 6 grid cells (4.6 x 10 6 unmasked) LSMs: CLM 2.0, Noah (2.7.1, now 3.2) LSMs run in the NASA Land Information System modeling framework 9-year spin-up LSMs driven by GDAS initially, now by both NAM and biasadjusted GDAS; i.e. a 4-member ensemble. Primary collaboration with APRFC during spring snow melt. Monthly snow surveys offer data assimilation potential.

21 GDAS+CLM SWE, Advantages of Land Surface Models Models same snowpack states Also models additional water balance variables year round Soil moisture profiles Snow temperature profiles Evaporative fluxes Surface runoff

22 Land Surface Modeling for Central Asia 2009: USGS collaboration for FEWS NET Domain: 70 x 35 at 0.01 ; 24.5 x 10 6 grid cells (22.1 x 10 6 unmasked) on lat/lon grid One hour time step LSMs: CLM, Noah Driven by GDAS only Few quantitative surface snow observations; model states are evaluated via anomaly analysis based on eleven years of retrospective runs.

23 NOAH SWE

24 LIS at NOHRSC: Central Asia

25 Future Land Surface Modeling and Data Assimilation Plans for North America October 2013: National Water Center Initial Operating Capacity ( IOC13 ) Domain: 116 x 55 at ; 91.9 x 10 6 grid cells (37.5 x 10 6 unmasked)

26 Future Land Surface Modeling and Data Assimilation Plans for North America October 2013: National Water Center Initial Operating Capacity ( IOC13 ) Domain: 116 x 55 at ; 91.9 x 10 6 grid cells (37.5 x 10 6 unmasked) LSMs: CLM, Noah Driven by Rapid Refresh and NAM Daily forecasts to 72 hours Hourly history files

27 Future Land Surface Modeling and Data Assimilation Plans for North America October 2014: IOC14 Possible capabilities beyond IOC13: Improved, AoR forcings 30-year reanalysis Update model cycles Snow data assimilation capabilities Additional land surface models Hydrologic routing

28 Gaps, Issues, and Potential Contributions Gaps Improved forcing data required, especially precipitation Better down scaling required (leverage climatology?) Data assimilation required, especially RADAR Improved (more and better) in situ observations required Issues Weather, climate, and hydrology needs differ e.g. atmospheric modelers may not be as concerned as hydrologists by snow under trees Potential contributions NOHRSC snow and land surface models and their IT infrastructure are scalable to global challenges at a reasonable cost Our initial North America effort is resourced with four 8-core Linux servers

29 Questions and Comments