1 NOHRSC Challenges of using Snow Data Carrie Olheiser Office of Hydrologic Development National...

Post on 06-Jan-2018

218 views 2 download

description

3 NOHRSC Nature of snow observations. Difficult to get “representative” observation. Metadata inconstancies between sources. Data in proper SHEF format including time, amount and duration. Snowfall observations are not always published in a standard format. Timeliness of data Reliably and timing of observations. Adequate training in data collection. Quality controlling the data, who is responsible? Challenges of snow data

transcript

1NOHRSC

NOHRSC Challenges of using Snow Data Carrie Olheiser

Office of Hydrologic DevelopmentNational Weather Service, NOAAU.S. Department of Commerce

National Operational Hydrologic Remote Sensing Center

www.nohrsc.noaa.gov

2NOHRSC

National Snow Analyses (NSA)

Ground

Airborne

Satellite

Multi-sensor Snow Observations

Data ProductsInteractive

MapsTime Series

PlotsText

Discussions

Snow Information Products

Snow Modeling and Data Assimilation

Numerical Weather

Prediction Model ForcingsGridded Snow

Characteristics

U.S.1-km2

Hourly

3NOHRSC

• Nature of snow observations. Difficult to get “representative” observation.

• Metadata inconstancies between sources.• Data in proper SHEF format including time, amount and

duration.• Snowfall observations are not always published in a standard

format.• Timeliness of data• Reliably and timing of observations.• Adequate training in data collection.• Quality controlling the data, who is responsible?

Challenges of snow data

4NOHRSC

• NOHRSC performs 4 automated tests – Reasonable Range– Spatial Consistency– Temporal Consistency– Internal Consistency

• Reasonable range is the only QC test Snowfall data undergoes.

Quality Controlling of Data

5NOHRSC

6NOHRSC

Manual Quality Control Spatial Consistency

These points need to be looked at more closely, not spatially consistent

7NOHRSC

Manual Quality Control Temporal Consistency

Bad Observation

Using data operationally does not allow us to see into the future for temporal quality control.

Unstable SNOTEL ObservationIn early season and late season

8NOHRSC

Manual Quality Control Internal Consistency

• Density Check

9NOHRSC

Station RepresentativenessPoint in Pixel

ESTES PARK RAWS

10NOHRSC

Point in Pixel Consistency ? • Elevation

– Differences in station elevation and 1 km X1 km DEM elevation vary as much as 1500 meters, example station KRWW4, • DEM elevation = 1439• Reported elevation = 2910

• Forest Cover Density– Photos from the Northwest SNOTEL sites and

model forest cover density disconnects.– Example from Caribou Maine and station KCAR.

11NOHRSC

Disconnects between ground survey coop results for the

March 1 to March 4 SurveyForest Density Issues and Stations not representative of the region.

• Caribou, ME• SWE reported March 1,

2004 is 0.038 meters.• KCAR is a Coop Station

located at the Airport in the large wind swept location that is well exposed with a forest density of ~ 6%.

• Caribou, ME SJ snow survey site .

• SWE reported March 1, 2004 is 0.104 meters.

• This station is reported by the USGS and measured at the medical clinic which is a more heavily wooded area.

Assimilation for the region used KCAR a regular reporter inthe season but turns out it was not representative of the region.This was not the only station with similar issues.

12NOHRSC

72 hr Snowfall Observations

13NOHRSC

• It was agreed at the NWS 2004 Cold Regions Hydrology Workshop (Kansas City, 2004 November 15-19) that the NOHRSC would develop web-based, real-time, automated, interactive, snowfall maps for the CONUS.

History of NOHRSC Snowfall Products

14NOHRSC

National Metadata Sources Over 40 different sources of station metadata

Often redundant, often inconsistent!

NEED ONE-STOP SHOPPING FOR STATION METADATA

National Weather Service Databases

NWSLI, MADIS, CSSA (B44’s), NWS-ICAO, NWS-METAR, HADS, NCDC

NRCS SNOTEL and Snow CoursesUSACE New England District Snow SurveysNew York City Dept. Environmental ProtectionFederal Aviation Administration

California Department of Water ResourcesMaine Cooperative Snow SurveyMesoWest (150 + Smaller mesonets)RAWS Remote Automated Weather Stations

Federal and State Agencies

Weather Forecast Offices, River Forecast Centers and Regional Offices

15NOHRSC

Importance of Accurate Metadata • Numerous databases leads to

uncertainties in the station metadata.

16NOHRSC

Willow Creek, Wyoming

17NOHRSC

99 Miles

Boulder Creek, Oregon

18NOHRSC

72 hr SWE Observation

19NOHRSC

Quality Control of Data• Internal Quality Control• Manual Quality Control• Point in Pixel Consistency

20NOHRSC

Internal Quality Control• Behind the scenes 4 QC tests are run

– Reasonable Range– Spatial Consistency– Temporal Consistency– Internal Consistency

• Problems with losing to much data if we have strict requirements in an automated quality control scheme.

• The nature of snow variability makes qc difficult and as much data is needed as possible.

21NOHRSC

• PRODUCTS• Hourly and Daily• 1 km2 Resolution

• 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

22NOHRSC

• It was agreed at the NWS 2004 Cold Regions Hydrology Workshop (Kansas City, 2004 November 15-19) that the NOHRSC would develop web-based, real-time, automated, interactive, snowfall maps for the CONUS.

History of NOHRSC Snowfall Products

23NOHRSC

• NOHRSC has seven (7) new snowfall maps to the NOHRSC selection of Physical Elements on the NOHRSC Interactive Snow Information page.

• Daily total snowfall observations (for previous 24 hours) • 2-day total snowfall observations (for previous 48 hours) • 3-day total snowfall observations (for previous 72 hours) • Daily total snowfall interpolation (for previous 24 hours)• 2-day total snowfall interpolation (for previous 48 hours) • 3-day total snowfall interpolation (for previous 72 hours)• 24-hr Raw Snowfall – Only displays 24 hour totals

NOHRSC Snowfall Products

24NOHRSC

Interpolated Snowfall 24 hrs

25NOHRSC

• Are generated in automated process that uses snowfall data encoded in SHEF and MADIS formats received at the NOHRSC over AWIPS and MADIS Snow Collective.

• The NOHRSC performs only a “reasonable range” quality-control test on the snowfall data.

• Snowfall maps are generated using both– Temporal

• quantization and integration– spatial interpolation

• inverse distance weighting with a search radius of 75km.

NOHRSC Snowfall Products

26NOHRSC

Total Observed Snowfall 24 Hrs

27NOHRSC

Total Observed Snowfall 72 Hrs

28NOHRSC

Total Observed Snowfall 72 Hrs

29NOHRSC

Interpolated Snowfall 48 hrs

30NOHRSC

Interpolated Snowfall 72 hrs

31NOHRSC

Snowfall Text Products

Raw Data Available for Download

32NOHRSC

Cumulative Snowfall Product

• Two stations located a couple kilometers apart. Both reporting snowfall, but MITM2 is reporting zero snowfall.

• Some station do not have a continuous record of snowfall, missing significant events. These products need to be used with a watchful eye!