Gauges Surface and NWP data Working Group summary

Gauges, Surface and NWP data Working Group summary slides June 29, 2005

How to organize? n n n By special emphases/interests Basic areas from Steve V. By individual data type By description of existing ‘systems’ Some combination

Aspects of Gauge data? n n n n Data types, equipment, communications Relative accuracies, reliabilities Effective ranges of coverage, densities, issues of scale Mesonets Quality control strategies, in real-time and in retrospect Historical versus real-time (operational) datasets Archive availabilities

Aspects of Surface data? n Classic meteorologic synoptic variables at observing stations n n n vertical soundings, profiles (observed, remotely-sensed, modeled, ? ); surface transects; time-height plots; etc. climatologic fields categorical (targeted) climatologies hydrologic observations n n n Temperature, moisture, winds, pressure, etc. derived, interpreted stage & flow observations stage mesonets with level loggers Availability/consistency of integrated Surface Analyses n Usage for regime typing and categorization

Aspects of NWP data? n n n Particular models, fields, characteristics Deterministic, ensembled traces Resolutions in time-space Direct model usage vs. diagnostic usage Utility for observation/forecast/updating Downscaling approaches Simplified (process) models Data assimilation systems Rapid refresh models Coupled models (e. g. , atmospheric-hydrologic) Application/Utility-specific requirements for model output

Special Emphases n n n n n Hydromet Inputs over mountains for River Forecasting (PP, TA, HZ) The utility of hydroclimatology in hydrometeorology Fundamental Issues of ‘representative’ scale and resolution NWP & surface analyses for regime typing Gauge QC methods; manual & auto Bias adjustment w/ gauges & r. s. /NWP/surface data (& vice versa) Oklahoma Mesonets Modernizing the Cooperative Network Archival & availability; real-time vs reanalysis NCDC and real-time/meta- data archiving

Areas suggested by Steve n Analysis methods for Gauge data n current methods n n regional approaches n n n e. g. , as mentioned, in the West, NWS River Forecast Centers (RFC’s) have used systems such as MM, with climatologic scaling based on PRISM (Daly et al, at Oregon State University) I know there a number of other regional/local/commercial systems (e. g. , P 3 in Tulsa? ) Gauge QC methods Use of sfc, r. s. , NWP data to identify storm type, characteristics, etc. n n MM, Oklahoma Mesonet, Neron e. g. , research to include regime characterization into analysis algorithms and strategies; one example is an ongoing effort at synoptic regime (targeted) versions of the PRISM climatologies by Daly and the Oregon Climate Service with the NWS Western Region. How to quantify Gauge uncertainty Real-time calibration with Gauge data n RFC datasets as real-time components for Q 2

Ed’s comments on ‘representativeness’ n Does Gauge QC matter? n Importance of QC for model verification n n ETA 12 hour forecasts WRF verification Manual datasets and zeros Business case for QC as a priority

Analysis Methods for QC n Current Methods n Regional example n Mountain Mapper

CNRFC area Topography

Hydromet Input Development in Complex Terrain n Common treatment of Observed & Future, R. T & historical: n n Complications n n Extreme Non-stationarity/heterogeneity of fields Phase changes with elevation Limited effectiveness of remote sensors Advantages n n n Mean Areal Precipitation distribution (MAP) Mean Areal Temperature distribution (MAT) Mean Areal Snow/Freezing Level (‘MAZ’) High Consistency across time-space scales Good Atmospheric Model performance Background Commonality n Climatology (PRISM)

Inversions in the West: data priority pyramid n n n Remote sensors Surface datasets Atmospheric models Climatologies Gauges

Bottom line: Basins

Mountain Mapper (MM) System: Quantitative Precipitation Development n QC-DAILY: QP Estimation (QPE) w/ Observed Data n QC of gauge reports n n neighbor estimations via ‘normalized’ data Gridded estimation of MAPs n RFS ingestion in real-time n n n Historical analysis/calibration Freezing Level Estimation n n problems with discontiguous areas with MAPX Explicit Rain vs Snow delineation in RFS Assist QC of observed data (e. g. . . , tipping bucket) SPECIFY: QP Forecast (QPF) specification VERIFY: QPF/QPE verification ITERATE: MAP Analysis/Update/Iteration

Complimentary Data Types in Daily_QC n Surface Temperature n QC of gauge reports n n n Gridded estimation of Temperature n n n PRISM max/min Temperature climatology grids Display of basin aggregated MATs Implicit Freezing Level Estimation n n Normalize/scale reports with climatology neighbor estimations with differences (vs ratios with PP) Currently: Lapsed for Rain/Snow delineation in RFS Assist QC of observed data (e. g. . . , tipping bucket) Snowmelt energy exchange Freezing (Snow) Level heights n n n ‘observations’ from RUC analysis field PAC-Jet Profilers Simple (flat) Interpolation (no scaling)

POINTS (stations) B A C E D Estimating Precipitation Catch at station with missing report (P ) from surrounding x observations (Pi ) n F Px = Pi i=1 n i=1 ( ) Wi Nx Ni Wi d= distance of separation where Wi = 1 2 d W= station weight N= station seasonal mean P= station precip catch

GRIDS POINTS LUMP B A C E D PATTERN INTEGRATE F Estimate grid boxes as “stations” n Px = i=1 n i=1 Pi ( ) Wi Wi Nx Ni Wi = 1 2 d d= distance of separation

GRIDS POINTS LUMPS B A UPPER C E D PATTERN INTEGRATE LOWER F different for each season/month n Px = i=1 n i=1 Pi ( ) Wi Wi Nx Ni For each MAP area Wi = 1 2 d d= distance of separation

POINTS QPF ----- * * * LUMP (MAP) --------- HRAP GRID Climatology: Seasonal or Monthly Isohyets MAP into SNOW MODEL SOIL MODEL BASIN RUNOFF MODEL (UNIT HYDROGRAPH)

POINTS QPF ----- * HRAP GRID 5500 ft * Rain/snow 20 % SNOW Climatology: Seasonal or Monthly Isohyets 20% MAP into SNOW MODEL LUMP (MAP) --------80% RAIN 80% MAP into SOIL MODEL BASIN RUNOFF MODEL Results in SMALLER Hydrograph (UNIT HYDROGRAPH)

20 % SNOW 5500 ft 80% RAIN Rain/snow elevation approximately 1000 feet o below 0 C level (here, assume 1500 feet) Ers = E v + (( T v - PXTEMP)* (100/L p )) E rs = rain/snow elevation E v = input variable elevation: freezing level or MAT o T v = input variable temperature: 0 C or MAT o PXTEMP = threshold temperature [1 to 2 C] o L p = lapse rate during precipitation ( C/100 m) [wet adiabatic or approx. 55 o C/100 m or 3 o F/1000 ft]

POINTS (stations) Estimating Temperature at a station with a missing report (Tx ) from surrounding observations (Ti ) B A C E D n (T i + ( N i=1 F Tx = Wi d= distance of separation E ix W= station weight N= station monthly mean T= station Temperature Fe= elevation factor E= elevation 1 Wi = ))*Wi - Ni n i=1 where x * d ix+ Fex

Works well but Obstacles include: n Timeliness: n n n Computational Speed Communications Human interactions Software Tools Non-orographic events n Once you’ve got the hammer, everything looks like a nail

Baseline is Improving: ongoing enhancements n Targeted (non-static) backgrounds for normalization n n E. g. , Ongoing development of Smart-PRISM (regime-sensitive) climatologies (Daly, Taylor group at Oregon State) Atmospheric/process model data inclusion Higher Spatial/Temporal resolutions Iterative updating Increased automation

Gauge QC methods/tests 1. 2. 3. 4. 5. 6. 7. 8. Gross range checks Nearest neighbor estimation (x-validation) Gage history Visual inspection of accumulation trace Measuring Hardware: tipping vs. weighing; shielded? Station network characteristics (SNOTEL, ASOS, etc. ) Station Measuring Hardware: tipping vs. weighing Location wrt Freezing (Snow) Level 1. Disqualify tipping buckets

More Gauge QC methods/tests 9. 10. Compare w/ SWE increment at co-located pillow (substitute) Compare with remotely sensed data 1. 2. 3. 11. Consider synoptic regime and surface data (e. g. , post frontal convection? ) 1. 12. 13. 14. NEXRAD overlay IR Satellite views Lightning strikes Adjust deviation allowed from neighbor estimation Reasonableness of Gridded (hrap)/Basin (MAP) renditions of data Consult longer duration results (monthly/seasonal performance w/ monthly-qc) Consistency with temperature, freezing level, soil moisture

Identification of Synoptic Regime n Automated production from NWP/R. S. /surface data n n At what resolutions n n n Adequate skill? Hourly 5 km Recommendations for prototyping

Quantifying Gage Uncertainty n n Research topic Indications from operational archive of meta-data n n Results from multiple tests Utility of Hydrologic application

Real-time role for RFC’s/mesonets in Q 2 n Gauge QC stream as output n n n Availabilities n n n 6/24 hourly 1 hourly progress Real-time Historical Meta-data on QC Validation, Bias adjustment Regional Variations/Optimizations

Platform & Tools for ‘Community’ Model Development? n n n Accessible Extensible Consistent