PERTURBATION VS' ERROR CORRELATION ANALYSIS PECA

1
THORPEX A GLOBAL ATMOSPHERIC RESEARCH
PROGRAM NOAA LONG-TERM RESEARCH PROGRAM
Scientific Guidance Provided by NOAA THORPEX
Science Steering Committee Presentation
prepared by Z. Toth
2
THORPEX ORGANIZATION
EXECUTIVE OVERSIGHT SCIENTIFIC DIRECTION
INTERNATIONAL LEVEL LINK WITH WMO
International Core Steering Committee Michel
Beland (Co-chair) Louis Uccellini (US
Representat.)
International Science Steering Comm. Co-chaired
by Mel Shapiro Alan Thorpe
REGIONAL (NORTH AMERICAN) LEVEL LINK WITH USWRP
Oversight provided by North American members of
International Core Steering Comm.
North American Science Steering Com Co-chaired
by David Parsons Pierre Gauthier
NOAA LEVEL LINK WITH CORPORATE MANAGEMENT
NOAA THORPEX USWRP Sub-Com. L. Uccellini
(Chair), M. Uhart, M. Colton, and Jack Hayes
NOAA Science Steering Committee Z. Toth (Chair,
Program Manager) 12 NOAA and outside members
3
NOAA THORPEX ORGANIZATIONAL CHART
NOAA THORPEX USWRP Sub-Committee Louis
Uccellini (Chair) NWS Michael Uhart OWAQ Marie
Colton ORA/NESDIS Jack Hayes NWS NOAA THORPEX
Science Steering Committee Zoltan Toth Chair,
Program Manager NOAA/NWS Observations Data
Assimilation Jaime Daniels NOAA/NESDIS Craig
Bishop NRL David Emmitt SWA L.-P.
Riishojgaard JCSDA Thomas Schlatter NOAA/FSL Chri
s Velden CIMSS Forecasting/Predictability Soci
oeconomic Applications Jim Hansen MIT
Rebecca Morss NCAR Jeff Whitaker/T. Hamill
NOAA/CDC Marty Ralph NOAA/ERL George
Kiladis NOAA/AL
4
THORPEX A GLOBAL ATMOSPHERIC RESEARCH
PROGRAM NOAA LONG-TERM RESEARCH PROGRAM PLAN
Based largely on work of NOAA THORPEX Planning
Meeting (October 21-22 2002)

• NOAA NWS
• Zoltan Toth
• Naomi Surgi
• NOAA OAR
• Melvyn Shapiro
• Jeff Whitaker

• Outside NOAA
• Craig Bishop NRL
• David Carlson NCAR
• Ron Gelaro NASA
• Rebecca Morss NCAR
• John Murray NASA
• Chris Snyder NCAR

With further input from NOAA THORPEX Science and
Implementation Team
Acknowledgements D. Rogers, L. Uccellini, S.
Lord, J. Gaynor, W. Seguin
5
NOAAS INVOLVEMENT IN THORPEX

• 1998-99 Discussions started with involvement of
  NOAA scientists
• Apr 2000 First International Meeting
• Mar 2002 First Workshop, International Science
  Steering Committee formed
• Aug 2002 NOAA Tiger Team Meeting
• Oct 2002 NOAA THORPEX Planning Meeting
• Nov 2002 1st Draft NOAA THORPEX Science and
  Implementation Plan
• Jan 2003 NOAA THORPEX Science Steering Committee
  formed
• Feb 2003 Pacific TOST Experiment
• Jun 2003 First NOAA THORPEX Announcement of
  Opportunity
• Sep 2003 25 Full Proposals received

6
THORPEX OBJECTIVES
INTERNATIONAL PROGRAM SCIENCE GOAL Promote
research leading to new techniques in
Observations (Collect data) Data
assimilation (Prepare initial cond.)
Forecasting (Run numerical
model) Socioeconomic Applications
(Post-process, add value, apply) SCIENTIFIC
RESEARCH MUST ENABLE SERVICE GOALS SERVICE GOAL
Accelerate improvements in utility of 1-14 day
forecasts for high impact weather THORPEX
ANSWER Develop new paradigm for weather
forecasting through Enhanced collaboration Int
ernationally Among different
disciplines Between research
operations Example North American Ensemble
Forecast System (NAEFS)
7
THORPEX OBJECTIVES SYNERGISTIC COLLABORATION
SCIENCE GOAL SHARED WITH ALL PARTICIPANTS,
COMMON THEME Develop new paradigm for weather
forecasting Integrate OBS, DA, FCST,
APPLICATION areas of forecast process All
participants contribute to advancing same science
objective LEVERAGING RESOURCES SCIENTIFIC
RESEARCH MUST ENABLE SERVICE GOALS SERVICE GOAL
DIFFERENT PRIORITIES FOR EACH
AGENCY/NATION Accelerate improvements in utility
of forecasts for high impact weather Severe
weather (Asia?) 1-3 day weather (Europe?)
Global monitoring (NASA?) All participants
share service applications among
themselves COMPLIMENTARY EFFORTS Critical
mass of resources needed Intellectual,
Material, Global observs. NEED INTERNATIONAL
COLLABORATION
8
THORPEX OBJECTIVES

• NOAAS ROLE
• Existing NOAA, USWRP and other programs aimed at
• Short-range forecast problem PACJET, IHOP, Cold
  Season Precip., etc
• Seasonal climate forecast problem CLIVAR,
  GAPP, etc
• THORPEX fills critical gap between short-range
  weather climate programs
• NOAA SERVICE APLLICATION GOAL
• Accelerate improvements in weather forecasts to
  facilitate issuance of skillful
• 3-7 day precipitation forecasts
• 8-14 day daily weather forecasts

• NOAAS ROLE
• Existing NOAA, USWRP and other programs aimed at
• Short-range forecast problem PACJET, IHOP, Cold
  Season Precip., etc
• Seasonal climate forecast problem CLIVAR,
  GAPP, etc
• THORPEX fills critical gap between short-range
  weather climate programs
• NOAAS SERVICE APLLICATION GOAL
• Accelerate improvements in weather forecasts to
  facilitate issuance of skillful
• 3-7 day precipitation forecasts
• 8-14 day daily weather forecasts

9
THORPEX A GLOBAL ATMOSPHERIC RESEARCH
PROGRAM NOAA LONG-TERM RESEARCH PROGRAM PLAN TO
REACH NOAAS SERVICE GOAL 52-page document based
on input from 18 NOAA and outside experts
INTRODUCTION New forecast paradigm SCIENCE
PLAN Major Themes Open Science
Questions Research and Development
Tasks IMPLEMENTATION PLAN Work
Plan Deliverables Performance
measures Education/Outreach Path to
Operations APPENDIX Link with NOAA Strategic
Goals NWS STIP Process
10
WEATHER FORECASTING FOR DAYS 3-14

• Based on guidance from Numerical Weather
  Prediction (NWP) models
• Quality tied with that of NWP model forecasts
• Components of NWP forecasting

• Observing system Collect data
• Data assimilation - Prepare initial conditions
• Forecast procedures
• Run numerical model
• Societal economic applictns
• Post-process, add value, apply

11
TRADITIONAL FORECAST APPROACH

• Each discipline developed on its own
• Disjoint steps in forecast process
• Little or no feedback
• One-way flow of information
• Uncertainty in process ignored

12
STATUS QUO SCENARIO
Substantial resources spent on improving NWP

• Forecast skill improves
• As in any learning process, improvements become
  harder as skill advances
• Maintaining or accelerating rate of improvements
  not possible with current status quo
  approach/resources

Is this acceptable when sensitivity/vulnerability
of society to weather increases?
13
THORPEX SOLUTION

• REVOLUTIONIZE NWP PROCESS
• Invest in major new NWP program gt
• Develop new NWP procedures
• INTEGRATED, ADAPTIVE, USER CONTROLLABLE
• Return Pace of forecast improvement
  maintained/accelerated
• Assess costs and societal/economic benefits of
  new procedures
• Implement operationally most cost effective new
  methods
• Return Enhanced operational capability
• Improved cost effectiveness

14
NEW NWP PARADIGM - 1

• INTEGRATED NWP
• Sub-systems developed in coordintation
• End-to-end forecast process
• Strong feedback
• Two-way interaction among components
• Error/uncertainty accounted for at each

Based on better understanding of forecast process
15
NEW NWP PARADIGM - 2

• Integrated
• ADAPTIVE
• Based on more detailed understanding of natural
  processes
• Allows more differentiated, case dependent
  methods/procedures
• Exmples
• Observations Adaptive platform collects data to
  fill gaps due to clouds
• Data assimilation Flow dependent forecast error
  estimates
• Forecasting Situation dependent modeling
  algorithms
• e. g., hurricane relocation
• Applications Probabilistic forecast reflects
  all forecast info gt ultimate adaptation of user
  procedures to weather

16
NEW NWP PARADIGM - 3

• Integrated
• Adaptive
• USER CONTROLLABLE
• Based on
• 2-way interactions (improved forecast process)
• Adaptive approach (better understanding of
  nature)
• Forecast process
• Traditionally driven by FIXED user requirements
• Now responsive to CHANGING user needs
• User needs connected to observational, data
  assimilation, and forecast systems
• Dynamical analysis of nature forecast process
• New, NWP model based tools
• Fully interactive forecast process
• Example User identifies critical forecast
  weather event
• Special observational or forecast procedures
• Improved targeted forecast

17
SCIENCE OBJECTIVE REVOLUTIONIZE NWP PROCESS -
INTEGRATED, ADAPTIVE, USER CONTROLLABLE
TRADITIONAL NWP Each discipline developed
on its own Disjoint steps in forecast
process Little or no feedback One-way flow of
information Uncertainty in process ignored

• NEW NWP
• Sub-systems developed in coordination
• End-to-end forecast process
• Strong feedback among components
• Two-way interaction
• Error/uncertainty accounted for

SERVICE GOAL IMPROVE 3-14 DAY FORECASTS
18
NEW NWP PARADIGM - 4

• Isolated examples exist
• INTEGRATED DEVELOPMENT
• NPOESS instrument/platform design
• Input from OSSE work (data assimilation/forecastin
  g needs considered)
• - North American Observing System initiative
• ADAPTIVE APPROACH
• GFDL hurricane model runs at NWS when needed
• USER CONTROL WSR program at NWS
• Threat of winter storm potential societal
  impact
• Dynamical calculations
• Targeted observations collected
• Targeted data inserted in analysis/forecast
  process
• From the EXCEPTION, THORPEX will make
  interactive, adaptive, user controlled methods
  the RULE

19
NOAA THORPEX PROGRAM OVERVIEW ACTIVITIES

• ANSWER SCIENCE QUESTIONS
• Advance basic knowledge,
• directed explicitly toward NWP applications
• Each task conceived as part of overall program
• DEVELOP NEW METHODS
• Sub-system development
• Academic research
• Cross-cutting activities
• Academic operational centers
• Observing System Simulation Experiments (OSSEs)
• Real-time test and demonstration
• Infrastructure / Core tasks
• Facilitate other activities - Strong agency
  involvement
• THORPEX Data Base
• Operational Test Facility
• RECOMMEND/PREPARE OPERATIONAL IMPLEMENTATION
• Integral part of program

20
SCIENCE QUESTIONS ACTIVITIES

• Observing system (OBS)
• Data assimilation (DA)
• Forecast procedures (FCST)
• Socio-economic Applications (SA)
• Cross-cutting activities
• Core tasks

21
SCIENCE QUESTIONS ACTIVITIES - 1

• OBSERVING SYSTEM
• New in-situ and remote instruments/platforms to
  complement existing network
• Adaptive observing instruments/platforms
• For large data sets
• Super-obing etc prior to OR within data assimil.
• (Joint work with data assimilation)
• Obs. error estimation (correlated/uncorrelated)

22
SCIENCE QUESTIONS ACTIVITIES - 2

• Observing system
• DATA ASSIMILATION
• Improve techniques
• Forward models, transfer codes
• Thinning of data
• Treatment of data with correlated errors
• Advanced methods to use flow dependent covariance
• 4DVAR research, e.g., continual update of error
  covariance
• Ensemble based techniques
• Treatment of model errors
• Adaptive observing techniques
• Quick use of targeted data (pre-emptive
  forecasting)
• Methods in the presence of
• Strong non-linearities
• Model error
• Effectiveness of targeted data in
  analyses/forecasts
• Effect on climatological applications of data

23
SCIENCE QUESTIONS ACTIVITIES - 3

• Observing system
• Data assimilation
• FORECAST PROCEDURES
• Initial ensemble perturbations (Joint with data
  assimilation)
• Role of non-modal behavior
• Separate model related error from initial value
  errors
• Systematic vs. random errors
• Atmospheric features most affected
• Critical model features responsible for different
  errors
• Improve model formulation to reduce errors
  (Coupling techniques)
• Techniques to account for remaining uncertainty
  in ensembles (physics, etc)
• Adaptive modeling and ensemble techniques

24
SCIENCE QUESTIONS ACTIVITIES - 4

• Observing system
• Data assimilation
• Forecast procedures
• SOCIO-ECONOMIC APPLICATIONS
• Probabilistic forecasting
• Statistical post-processing
• New procedures for intermediate and end users
• Add-on costs of new THORPEX NWP process
• Cost of data from multi-use satellite platforms
  (Joint with Observtns.)
• Incremental societal/economic benefits of new NWP
  process
• New NWP verification measure
• Societal aspects of new adaptive NWP procedures
• Equitable use of NWP resources, how adaptive
  procedures applied nationally and internationally

25
CROSS-CUTTING ACTIVITIES

• Integrating NWP procedures from four sub-systems
• Observing System Simulation Experiments (OSSEs)
• Data needs of NWP
• What variables/resolution/accuracy required
• Instrument/platform neutral assessment
• What instruments/platforms can provide data needs
• Existing and new in-situ remote platforms
• Adaptive component to complement fixed network
• Most cost effective solution
• Relative value of improvements in four
  sub-systems
• Improvements in which sub-system offer best
  return?
• Reallocation of resources
• Test of proposed operational configurations
• Major field program if needed
• Cost/benefit analysis - Select most cost
  effective version

26
CORE TASKS

• Needed for efficient research planned
  operations
• Strong agency involvement
• THORPEX data base (observations, forecasts)
• Information Technology challenge
• High data volume
• Transmission
• Storage of data
• Foster collaboration in critical areas
• Workshops (Societal and economic impacts)
• Joint proposals Interdisciplinary collaboration
• Critical in past programs like FASTEX
• Test-bed Pathway from research to operations
• Formal procedure for researchers to follow
• Melting pot for new ideas
• Venue for cross-cutting activities

27
NOAA THORPEX PROGRAM OVERVIEW - DELIVERABLES
DELIVERABLES New forecast techniques Observing,
data assimilation, forecasting, application
tools Accelerated forecast improvements Integrate
d, adaptive, user controllable NWP Cost
effective operational system Based on
cost/benefit analysis Enhanced user interface
COSTS Research Grant Program Integrated program
- Four sub-areas cross-cutting
activities Operational Test Facility Simulated
forecast process Database Real-time
test/implementation Data transmission,
Computations, Training
OVERALL MEASURE OF SUCCESS
SOCIO-ECONOMIC BENEFITS MUST OUTWEIGH
OPERATIONAL COSTS
28
LINK WITH NOAA MISSION GOAL

• NOAAS 3rd MISSION GOAL sounds like excerpt
  from THORPEX doc.
• NOAA will provide integrated observations,
  predictions, and advice for decision makers to
  manage environmental resources.
• Mission strategies and measures of success
• directly correspond with
• THORPEX Sub-program areas
• NOAA MISSION STRATEGY THORPEX FORECAST
  COMPONENTS
• Monitor and Observe Observations
• Understand and Describe Data Assimilation
• Assess and Predict Forecasting
• Engage, Advise, and Inform Socio-economic
  Applications
• Different Line Offices responsible for various
  forecast components
• NEED FOR NEW MATRIX MANAGEMENT CONCEPT FOR
  INTEGRATION

29
LINK WITH NWS STIP PROCESS

• National Weather Service (NWS)
• NOAAs operational weather forecast
  provider
• NWS Science and Technology Infusion Plan (STIP)
• Operational requirements should motivate all
  service oriented research
• Research must have thread to operations
• Credible path to operational implementation
• SCIENTIFIC RESEARCH MUST ENABLE SERVICE GOALS
• THORPEX seeks advanced knowledge on two fronts
• Nature (atmospheric and related processes)
• Forecast procedures (OBS, DA, FCST SA
  techniques)
• Integrating knowledge from two areas leads to new
  forecast paradigm of
• INTEGRATED, ADAPTIVE, AND USER CONTROLLABLE FCST
  PROCESS

30
THORPEX A GLOBAL ATMOSPHERIC RESEARCH
PROGRAM OVERVIEW OF NOAAS THORPEX-RELATED
ACTIVITIES

• ACCOMPLISHMENTS
• Contributed to International Science Plan
• Contributes to forming THORPEX International
  Program Office (Under WMO auspices in
  Geneva)
• Contributes to North American Implementation Plan
• Formed NOAA THORPEX Science Steering Committee
• Developed NOAA THORPEX Long-Term Research Plan
• Issued First NOAA THORPEX Announcement of
  Opportunity (AO)
• ONGOING EFFORT
• Evaluation of research proposals in response to
  AO
• Atlantic Regional Campaign
• OUTSTANDING ISSUES
• Funding for AO unresolved
• Funding for Operational Test Facility (FTO) needed

31
NORTH AMERICAN ENSEMBLE FORECAST SYSTEM PROJECT

• GOALS Accelerate improvements in
  operational weather forecasting
• through Canadian-US collaboration
• Seamless (across boundary and in time) suite of
  products
• through joint Canadian-US operational ensemble
  forecast system
• PARTICIPANTS Meteorological Service of Canada
  (CMC, MRB)
• US National Weather Service (NCEP)
• PLANNED ACTIVITIES Ensemble data exchange (June
  2004)
• Research and Development -Statistical
  post-processing
• (2003-2007) -Product development
• -Verification/Evaluation
• Operational implementation (in phases,
  2004-2008)
• POTENTIAL PROJECT EXPANSION / LINKS
• Shared interest with THORPEX goals of
• Improvements in operational forecasts
• International collaboration
• Expand bilateral NAEFS in future

32
BACKGROUND MATERIAL
33
COSTS/DELIVERABLES

• Costs
• Research program
• Integrated concept need to fund research in all
  four areas of NWP
• Operational implementation
• Deliverables
• New observing, data assimilation, forecasting,
  application tools to implement integrated,
  adaptive, user controllable NWP
• Acceleration in current NWP improvements

Socio-economic benefits must outweigh operational
costs
34
NOAA THORPEX OBJECTIVES

• 1) Develop new forecast procedures leading to
• Improved operational NWP forecasts and
• Develop/adapt cost/benefit tools to measure
  resulting societal impact
• ULTIMATE MEASURE OF SUCCESS
•  
• The overall success of the NOAA THORPEX program
  will be measured in a unique and comprehensive
  way. The program will be considered successful if
  the newly developed cost/benefit analysis tools
  (point 3 above) indicate that the forecast
  improvements (point 2) due to the new THORPEX
  procedures (point 1) can be achieved
  operationally in a cost-effective manner. That
  is, the incremental economic and societal
  benefits associated with the use of the new
  THORPEX forecast procedures outweigh their
  implementation and maintenance costs.
•  

35
NEED FOR COLLABORATIVE PROGRAM

• Interdisciplinary research
• Different groups/agencies/nations need to
  collaborate
• Integrated approach to NWP 4 sub-systems
• Practical goal Research Operations
• Challenging program
• Need critical mass of resources
• Intellectual
• Material
• Synergistic activities
• Priorities of other agencies may be different
• Common overarching THORPEX themes
• Complementary efforts
• Leveraging of resources
• Global data and all NWP methods universally
  needed
• INTERNATIONAL PROGRAM HIGHLY DESIRABLE