#21 Interactions between Data,
Observations and Modeling
Comment and Review
Neville Smith
Bureau of Meteorology Research Centre
– Comment on ocean perspectives
• Acknowledge input from ECCO group; GODAE
– Structure, strategic approach
– Priorities, specifics
1
Interaction: Assimilation, estimation,
fusion, blending, synthesis …
Obtain dynamically self-
consistent analyses of the
ocean circulation and
uncertainties: GODAE
Satellite data: fundamental
In situ data: fundamental
Models, assimilation
methods: fundamental
Compute, people
resources: essential
Data access, transport, …
2/8
Climate change product (service):
Estimate of heat change, and how it changed
Water flows along several distinct pathways from the subtropics to
the tropics that depend on intra-annual fluctuation in circulation.
M
C
Boundary Interior
pathways pathways
NGC
UC
(Fukumori et al, 2002) (Lee et al., 2002)
3/8
Climate change product (service):
Guiding observing system design: Koehl and Stammer
Where does SSH need to be observed to properly estimate the
seasonally varying heat transport?
Jan Mar
Aug Dec
4/8
Strategy and Structure
How do we build the needed infrastructure?
[observations, modelling and data & information management]
– Tier 1: Essential/mandatory components (CCRI?)
• Robust, sustained, reliable … but not fixed: “slow” evolution
• Product delivery, climate services: multiple use
– Tier 2: Enhancements (USGCRP?)
• Dynamic, experimental, innovative … but with clear target
Tier 1
– “Grand” challenge elements
• High risk, but even higher potential impact
Part I: The essential components (“enabling”)
Part II: Innovation, challenges
• But maintain balance: infrastructure is poor investment if there is
no innovative, ground-breaking research
5/8
The approach
You can observe, but not understand.
You cannot understand if you do not observe.
Need a clear statement on the fundamental need for a
global climate observing system.
– We do not have a gcos but we should all commit to building
the gcos we require (a comprehensive schedule)
– Observations alone will not answer climate change
questions (Chapter 3).
– Efficiency (return, investment) and effectiveness
(monitoring the important modes of variability and change)
require balance and integration (parts working together as
a whole)
Not so much “interaction” as mutually supportive
and complementary elements for a set of shared
objectives (climate “service” c.f. knowledge delivery)
– The observing system, data assembly and transport,
processing (modelling, assimilation), production (analyses,
forecasts) are the mandatory elements of the system
6/8
Make predictability the
overarching paradigm
Predictability (natural error growth relative to climate
signal) provides a framework for observations,
modelling, assimilation
– The observing system design space/time variability,
interaction/coupling
– Models: resolution, parameterisation, initialisation
“predictable” climate signals
– Data assimilation: statistical models predictable/resolvable
states
– Physical to non-physical: what can be inferred (“predicted”)
and what cannot.
– Global to regional (scales): identifying what impacts (physical,
biological, …) are predictable (deducible, …) and what are not.
A consistent approach to system development and
research
7/8
The CCSP should (among other things) …
(for modelling, observations and information systems)
Have clearer objectives (ref breakout #2)
Include routine ocean (climate) services (analyses, predictions, re-
analyses): refer GODAE, CLIVAR
Elevate data and information management to same level as
observations, modelling
§3.5 does not work
§12.3 is OK as a start but needs high-priority attached to climate “metadata model”
Data and information services: the emerging paradigm
Rearrange as suggested above
§12 is not “Grand Challenges” but infrastructure, enabling framework
Modified p 136 table as framework for developing priority for all
components, not just observations
Embrace applications (value-adding partnerships) within
infrastructure framework
Recognize the synergy with weather and ocean prediction (ref #12
discussion Wednesday)
8/8