Pacific Predictability:  Perspective of a Pacific Predictability:  Perspective of a

NWS Forecast OfficeNWS Forecast Office

Brad ColmanBrad ColmanNOAA/National Weather ServiceNOAA/National Weather Service

Seattle, WashingtonSeattle, WashingtonBrad.Colman@noaa.govBrad.Colman@noaa.gov

22ndnd THORPEX CORE OBJECTIVE THORPEX CORE OBJECTIVE

Contribute to the design and demonstration of interactive forecast systems that allow information to flow interactively among forecast users, numerical forecast models, data-assimilation systems and observations to maximize forecast skill. [As an example, targeted observing strategies incorporate dynamical information from the numerical forecast model itself to identify when, where, and what types of observations would provide the greatest improvement to specific weather forecasts of societal, economic, and environmental interest.]

44thth THORPEX CORE OBJECTIVE THORPEX CORE OBJECTIVE

Develop and apply new methods that enhance the utility and value of weather forecasts to society, economies and environmental stewardship through: (i) user-specific probabilistic forecast products; (ii) the introduction of interactive procedures that make the forecast system more responsive to user needs; (iii) the design of and training in the use of user-specific forecast products. Research will identify and assess the societal/economic costs and benefits of THORPEX recommendations for implementing interactive forecast systems and improvements in the global observing system.

National Digital Forecast DatabaseNational Digital Forecast Database Mosaic forecasts for the Mosaic forecasts for the

entire country or regional entire country or regional domainsdomains

Public, marine, fire Public, marine, fire weather, and other weather, and other programsprograms

Flexible, GIS format, Flexible, GIS format, graphics, etc.graphics, etc.

It’s aggressive and represents a major paradigm shift in the way NWS forecasters produce their forecasts.

National Digital Forecast Database targets a National Digital Forecast Database targets a spectrum of weather information usersspectrum of weather information users

More weather data More weather data

Higher resolution Higher resolution forecastsforecasts

Visual displays of Visual displays of probabilityprobability

User-defined User-defined products create products create business business opportunitiesopportunities

Different Different Products for Products for

Different Different CustomersCustomers

TODAY...RAIN LIKELY. SNOW LIKELY ABOVE 2500 FEET. SNOW ACCUMULATION BY LATE AFTERNOON 1 TO 2 INCHES ABOVE 2500 FEET. COLDER WITH HIGHS 35 TO 40. SOUTHEAST WIND 5 TO 10 MPH SHIFTING TO THE SOUTHWESTEARLY THIS AFTERNOON. CHANCE OF PRECIPITATION 70%.

TODAY...RAIN LIKELY. SNOW LIKELY ABOVE 2500 FEET. SNOW ACCUMULATION BY LATE AFTERNOON 1 TO 2 INCHES ABOVE 2500 FEET. COLDER WITH HIGHS 35 TO 40. SOUTHEAST WIND 5 TO 10 MPH SHIFTING TO THE SOUTHWESTEARLY THIS AFTERNOON. CHANCE OF PRECIPITATION 70%.

The public, emergency The public, emergency managers and city managers and city planners use planners use WWW graphic products WWW graphic products for detailed forecastsfor detailed forecasts

Commercial weather Commercial weather companies & emergency companies & emergency managers use grids to managers use grids to generate tailored generate tailored productsproducts

Radio stations & public Radio stations & public read text forecastsread text forecasts

NDFD Resolution and elementsNDFD Resolution and elements

Spatial resolution:Spatial resolution:

5 km grids for now5 km grids for now

2.5 km grids are being considered2.5 km grids are being considered Temporal resolution:Temporal resolution:

3 hourly for days 1-3 3 hourly for days 1-3

6 hourly for days 4-76 hourly for days 4-7

as available from CPC beyond day 7as available from CPC beyond day 7 Update frequency: every hourUpdate frequency: every hour All forecast parameters: Tx, Tn, T, Td, wind, sky, QPF, All forecast parameters: Tx, Tn, T, Td, wind, sky, QPF,

PoP, precip type, etc.PoP, precip type, etc.

National (NDFD) vs Local (LDFD) DatabasesNational (NDFD) vs Local (LDFD) Databases

National products and National products and services are derived from services are derived from a local digital databasea local digital database

Database kept current Database kept current through coordinated local through coordinated local updatesupdates

Standard formats of Standard formats of gridded and derived gridded and derived graphical productsgraphical products

Follow standard time and Follow standard time and space conventionsspace conventions

National products and National products and services are derived from services are derived from a local digital databasea local digital database

Database kept current Database kept current through coordinated local through coordinated local updatesupdates

Standard formats of Standard formats of gridded and derived gridded and derived graphical productsgraphical products

Follow standard time and Follow standard time and space conventionsspace conventions

Interactive Forecast Preparation System (IFPS)Interactive Forecast Preparation System (IFPS)

Numerical weather prediction inputsNumerical weather prediction inputs• Relatively coarse horizontal and vertical presentationsRelatively coarse horizontal and vertical presentations• ““SmartInit” process to downscale and generate SmartInit” process to downscale and generate

sensible elementssensible elements• Integration of point MOS guidanceIntegration of point MOS guidance

Forecaster inputs and adjustmentsForecaster inputs and adjustments• Generally iterative (w/o fresh model data)Generally iterative (w/o fresh model data)• Graphical editing toolsGraphical editing tools• ““SmartTool” scriptsSmartTool” scripts

Grid editing and coordinationGrid editing and coordination• Basic building block for the weather element grids is 5 Basic building block for the weather element grids is 5

x 5 km spatial and 1 hour temporalx 5 km spatial and 1 hour temporal• On-line chatOn-line chat• Poor objective techniques for some parameters like Poor objective techniques for some parameters like

sky cover, precip type, etc.sky cover, precip type, etc.

Considerations:Considerations:• System is predominately deterministic System is predominately deterministic

in presentationin presentation No gradation in spatial scale Day 1 No gradation in spatial scale Day 1 Day Day

77 Minimal gradation in temporal scaleMinimal gradation in temporal scale

• Strong motivation was to present Strong motivation was to present “realistic” weather“realistic” weather

• ““point-and-click” interface drove point-and-click” interface drove system to very fine scalesystem to very fine scale

• Growing client baseGrowing client base• Forecast process is resource intensiveForecast process is resource intensive

Separating processes and scales…Separating processes and scales…

48-h forecasts from 5 Ensemble Members Initialized 0000 UTC, Monday,

8 January 2001

Sailboats that broke free from their mooring lines litter the East Beach area of Santa Barbara, California, at sunrise following an overnight storm.

Importance of Upstream Initial Conditions

Probabilities in NDFD (proposed)Probabilities in NDFD (proposed)

• Percent probability of frozen precipitationPercent probability of frozen precipitation

• Percent probability of freezing precipitationPercent probability of freezing precipitation• Percent probability of thunderstormsPercent probability of thunderstorms• Exceedance forecasts for preset confidence levels:Exceedance forecasts for preset confidence levels:

Max QPFMax QPF Max wind speedMax wind speed Max wave heightMax wave height Extreme max temp (summer)Extreme max temp (summer) Extreme min temp (winter)Extreme min temp (winter) Max snow amountMax snow amount

Critical QuestionsCritical Questions(1)(1) Given the current NWS operational framework, what is the optimal way to incorporate Given the current NWS operational framework, what is the optimal way to incorporate

uncertainty information into the forecast process?uncertainty information into the forecast process?

(2)(2) On a day-to-day basis, what confidence information can be provided to the forecaster On a day-to-day basis, what confidence information can be provided to the forecaster to help them decide whether to follow a new solution? How much detail to include?to help them decide whether to follow a new solution? How much detail to include?

(3)(3) What can be done to separate out forced or climate signals from transients? Can the What can be done to separate out forced or climate signals from transients? Can the two parts be intelligently presented?two parts be intelligently presented?

(4)(4) What can be learned from existing NWP methodologies as far as benefits and What can be learned from existing NWP methodologies as far as benefits and limitations of a particular approach to presenting uncertainty information? What limitations of a particular approach to presenting uncertainty information? What method leads to the most specific and decisive actions?method leads to the most specific and decisive actions?

(5)(5) To what extent can the existing NWS forecast system accommodate growing demands To what extent can the existing NWS forecast system accommodate growing demands for uncertainty information?for uncertainty information?

(6)(6) Quantitative precipitation forecasts are among the most challenging. Can we skillfully Quantitative precipitation forecasts are among the most challenging. Can we skillfully determine limits, exceedance values, etc.? determine limits, exceedance values, etc.?

(7)(7) What are appropriate measures to assess the skill of forecasts on these spatial and What are appropriate measures to assess the skill of forecasts on these spatial and temporal scales?temporal scales?