Gulf of Maine Circulation Modeling: Prospects for Skill 6 July 2005

1
Gulf of Maine Circulation Modeling Prospects
for Skill 6 July 2005

• Daniel R. LynchDartmouth CollegeHanover NH

2
Points of Departure

• Science
• People
• Data
• Problems

3
Science

• Well-Established
• Physical Quantities
• Equations
• Algorithms for solutions

• Distributed across Academe

4
(No Transcript)
5
People

• At least 3 different communities
• Theory
• Observation
• Simulation the third science
• Algorithms
• Systems
• Non-Local
• Incentives Advancement of Learning

6
Data

• Unprecedented new abundance
• Sampling in (x, y, z, t) necessarily sparse
• Real Data is site-specific, event-specific
• Necessary to relate theory to facts
• By itself, hopelessly incomplete
• Interpolation
• Extrapolation
• Interpretation
• Relation to other information

7
Problems

• Local
• Not alligned with political boundaries
• Distributed across agencies
• Regulatory context

8
Status Quo

• Science generic, roaming
• Data local, incomplete
• People distributed across academe
• Advancent of knowledge, not application
• Problems regulatory context, local, political
  boundaries

The Key Challenge is Organizational
9
Regional Progress

• 1993 RARGOM Workshop lt-------------
• RMRP
• MWRA Mass. Bays
• GLOBEC
• EcoHAB
• Sea Grant (x4)
• GoMOOS
• Multiple NOAA programs
• Canadian Companions

10
Data and State Estimation
11
Time of Occurrence (Ocean)
State Estimation
Future
(Now)
Past
Time of Availability (Information)
12
Time of Occurrence (Ocean)
Forecast
Nowcast
Hindcast
Time of Availability (Information)
13
Time of Occurrence (Ocean)
Forecast
Nowcast
Hindcast
All Data
Time of Availability (Information)
14
Time of Occurrence (Ocean)
Forecast
Nowcast
Hindcast
All Data
Time of Availability (Information)
15
Time of Occurrence (Ocean)
Forecast
Model Data Product
Nowcast
All Data
Hindcast
Time of Availability (Information)
16
Time of Occurrence (Ocean)
Forecast
Nowcast
Hindcast
Data Used
Bell
Time of Availability (Information)
17
Time of Occurrence (Ocean)
Forecast
Nowcast
Hindcast
Data Used
Bell
Publication
Time of Availability (Information)
18
The Well-Posed Problem(The Mathematical Standard)

• Theory
• The Data
• Necessary and Sufficient
• initial state, simultaneous
• boundary conditions (deep ocean, cross-shelf
  transports)
• forcing (atmospheric fluxes, rivers)
• Parameters (bottom, surface roughness)
• All roads lead to Rome
• (small DX)

19
The Well-Posed Problem(The Mathematical Standard)

• Theory
• The Data
• Necessary and Sufficient
• initial state, simultaneous
• boundary conditions (deep ocean, cross-shelf
  transports)
• forcing (atmospheric fluxes, rivers)
• Parameters (bottom, surface roughness)
• All roads lead to Rome
• (small DX)

• Actual

Nonnecessary Insufficient DX is finite
20
The Well-Posed Problem(The Mathematical Standard)

• Theory
• The Data
• Necessary and Sufficient
• initial state, simultaneous
• boundary conditions (deep ocean, cross-shelf
  transports)
• forcing (atmospheric fluxes, rivers)
• Parameters (bottom, surface roughness)
• All roads lead to Rome
• (small DX)

• Actual

Nonnecessary Insufficient DX is finite
There is never a well-posed problem in nature
21
Poorly Posed Problems

• Must make up what is not known but necessary
• Use the data you have, deduce what you need
• Criterion credibility
• Credibility implies a Prior Estimate
• mean and variance

22
What is Truth?

23
What is Truth?
ed
em
Data
Model
Misfit d
24
What is Truth?
ep
ed
em
Prediction
Data
Model
Misfit d
Truth real but unknowable Errors unknowable
Prediction a credible blend Data
Model Blend Invokes statistics of ed , em
Prediction Error blend of statistics of ed ,
em, d
25
What is Truth?
ep
ed
em
Prediction
Data
Model
Misfit d
Skill Misfits Small, Noisy Unknown Inputs
Small, Smooth ep grows with time
Truth real but unknowable Errors unknowable
Prediction a credible blend Data
Model Blend Invokes statistics of ed , em d
26
Examples

• SAB resolve the data or burn it
• Great Bay Hi-resolution Lagrangian exchange
• ECOHAB Results hindcast trajectories
• Georges Bank Real-time Wind Forecast Error

27
A Data-Assimilative System
28
Resolution
29

• The difference between high-resolution and
  low-resolution forward simulations

30
Inverse Error with Low Resolution-- DA cannot
make up for Inadequate Resolution --
31
Estuarine Resolution
32
Boundary deduced from Interior Data
33
Data Assimilative Hindcast
34
Mean Separation Rate 1.78 km/day
35
The 2005 Prior
36
The 2005 Hindcast
37
The 2005 Hindcast

• Data-Assimilative
• Real Time
• At-Sea
• Limited-area
• Hindcast of complete cruise
• May 9- 18

38
The 2005 Hindcast
39
The 2005 Hindcast
40
The 2005 Hindcast
41
The 2005 Hindcast
42
The 2005 Hindcast
43
Who Painted the Bays Red?
44
Who Painted the Bays Red?
45
Frontal Dispersion - Forecast
46
Frontal Dispersion - Forecast
47
Frontal Dispersion - Hindcast
48
Frontal Dispersion - Hindcast
49
Key Challenges

• Organizational

50
Recommendations

• Accept
• Organizational progress must occur
• Scientific progress must continue in parallel
• Modeling is its own science
• Focus on
• the modelers, not the tools
• energizing the science community
• do not try to change the scientific culture
• organize the use of the Gulf of Maine as a
  laboratory
• enabling scientific progress on practical
  problems
• Circulation Modeling as initial baseline
• Invent
• no new organizations
• one new task Gulf of Maine Modeling Roundtable
• Insist on its standing in science and
  regulatory communities
• Do not distort the University Mission - Announce
  a new one

51
Interagency Agreement

• Establish a GoM Modeling Roundtable within a
  standing organization
• Service Populations Science, Engineering,
  Management
• Spread the cost among agencies
• Govenance
• full time staff
• board of overseers
• regular users group
• Outreach workshops on model products and
  strategies
• Focused publications on Gulf of Maine
• Archive
• software
• data
• simulation results

52
How to Recognize Success?

• Honor thy precursors
• Broad Participation
• Data Standards
• Accumulation of
• Data
• Software
• Reports
• Papers
• Progress is enabled
• People are enabled

People are your Investment