November 9, 2010 Diurnal Warming and Associated Uncertainties Gary A. Wick NOAA ESRL/PSD New Chair, GHRSST DVWG
Dec 21, 2015
November 9, 2010
Diurnal Warming and Associated Uncertainties
Gary A. Wick
NOAA ESRL/PSDNew Chair, GHRSST DVWG
SST Science Team MeetingNovember 9, 2010
Outline
• The diurnal warming problem
• Recent results and research directions
• Specific results related to uncertainties in physical modeling of diurnal warming
Wick et al.
SST Science Team MeetingNovember 9, 2010
What is the Problem?
• SST varies with time and depth
• Need to account for DW to:– Reference an SST value
to another time• Combination of
observations from different times of the day
– Reference values to a different depth
• Construction of foundation analyses
• Regression against observations at depth
Wick et al.
SST Science Team MeetingNovember 9, 2010
Why the Difficulty?
• Warming is a complicated function of multiple parameters– Models still uncertain– Not all parameters easily measured from space
• Need complete time history of forcing parameters– Sampling of available parameters is not
continuous– All parameters subject to measurement
uncertainties
• Observations/validation still insufficient
Wick et al.
SST Science Team MeetingNovember 9, 2010
A Fundamental Question
Can we estimate diurnal warming with enough skill to improve SST products?
•To what degree is added complexity desired or justified?
– Physical models vs. empirical parameterizations
•At what point is the data insufficient?
•Requires a detailed understanding of uncertainties
Wick et al.
SST Science Team MeetingNovember 9, 2010
Progress to Date
Work within the GHRSST DVWG and independent research has led to:•Improved models•Improved understanding and characterization of DW events•Improved resources for evaluation of models
– Aladin– Tropical Warm Pool+
•Initial diurnal warming analyses
Wick et al.
POSH
• Profiles of Surface Heating (POSH) – F96– Absorption– Reduce accumulated heat/momentum– Structured profiles of temperature within
the warm layer (CG empirical or Kantha/Clayson (WICK) profiles)
Dimensionless DW profileN
onD
im D
epth
(z)
NonDim Heat Content
Comparison throughout the day
DW
EXPERIMENTALPRODUCTS
Diurnal warming 2004-2010
Available on request from 2004
In 2011
Hourly DW available in real timeOn the zone herewith See LeBorgne et al 2010Proceedings EUMETSAT conference, Cordoba
Questions: -Hourly values every hour? Which delay -Or daily files with 24 fields?
SST Science Team MeetingNovember 9, 2010
DVWG Priorities
• Provision of diurnal warming analyses• Guidance on recommended/consensus
approaches• Improved estimates of uncertainties in
diurnal warming products
Wick et al.
Joint DVWG, HL-TAG, ST-VAL Workshop
• When– 28th February to 2nd March 2011
• Where– University of Colorado, Boulder, USA
• Why– To allow more time to address the key scientific
issues– To recognise the high degree of commonality
between each group
Outline format• Three main topics
– High latitude SST estimation• Retrieval, cloud masking, sea-ice analyses, both Arctic and Antarctic
regions– Diurnal variability observation & analyses
• Arctic, SEVIRI, POSH, shallow water, TWP+– SST uncertainty characterisation & SSES
• Uncertainty budgets, next generation radiometers, Argo, SSES• Main plenary talks (20 min), with open discussion (5 min
highlight talks) and working breakout sessions• Open mainly to DVWG, HL-TAG and ST-VAL groups
– 16 people confirmed with another 14 likely to attend– Some oral slots not yet taken
For more information
• Contact– Gary Corlett ([email protected]), Gary Wick (
[email protected]) or Jacob Hoeyer ([email protected])
• On the web– http://www.ghrsst.org/Joint-DVWG,-HL-TAG-and-S
T-VAL-Workshop-2011.html
November 9, 2010
Intercomparison of the Uncertainty in Diurnal Warming Estimates from
Physical Mixed Layer Models
G. A. Wick1 and S. L. Castro2
with C. Merchant, A. Harris, C.A. Clayson, C. Gentemann, and Y. Kawai
1NOAA ESRL2CCAR, Univ. of Colorado
SST Science Team MeetingNovember 9, 2010
Context
• GHRSST Diurnal Variability Working Group– Providing recommended approaches to
estimating the amount of diurnal warming present in satellite observations
– Sub-effort to compare the ability of existing models to reproduce observations of diurnal warming
• Constructing the SST Error Budget– What is the contribution of diurnal warming to
the uncertainty of satellite SST products– Primarily relevant to SST analyses
Wick et al.
SST Science Team MeetingNovember 9, 2010
Approach
• Diurnal warming models tested for both idealized and real forcing
• Models initially considered– COARE warm layer model– Modified Kantha-Clayson– POSH– Generalized Ocean Turbulence Model (GOTM) – k-
epsilon approach• Tested models with common solar penetration model• Used common skin layer treatment• Used common vertical grid• Fluxes computed with COARE model and used in
other models
Wick et al.
SST Science Team MeetingNovember 9, 2010
Approach
• Diurnal warming models tested for both idealized and real forcing
• Models initially considered– COARE warm layer model– Modified Kantha-Clayson– POSH– Generalized Ocean Turbulence Model (GOTM) – k-
epsilon approach• Tested models with common solar penetration model• Used common skin layer treatment• Used common vertical grid• Fluxes computed with COARE model and used in
other models
Wick et al.
SST Science Team MeetingNovember 9, 2010
Idealized Forcing
• Goal to compare behavior at low wind speeds and assess sensitivity to factors including solar penetration model and environmental conditions
• Constant wind speed from 0.5 – 10 m/s• Peak insolation from 50 – 1000 W/m2
• Conditions representative of tropics, mid-latitudes, and high-latitudes
• Models run for 5 days
Wick et al.
SST Science Team MeetingNovember 9, 2010
Model Simulations
Wick et al.
• Tropical conditions with u = 3 m/s, Qspeak = 800 W/m2
SST Science Team MeetingNovember 9, 2010
Modeled Warming at the Skin
• Tropical conditions• Results shown at 13:30 on day 3 of simulation• Warming computed relative to 25-m depth
Wick et al.
Modified Kantha-Clayson COARE
SST Science Team MeetingNovember 9, 2010
Modeled Warming at 1-m Depth
Wick et al.
• Tropical conditions• Results shown at 13:30 on day 3 of simulation• Warming computed relative to 25-m depth
Modified Kantha-Clayson COARE
SST Science Team MeetingNovember 9, 2010
Sensitivity to Other Fluxes
Results shown for Modified Kantha-Clayson model
Wick et al.
Tropics Mid-latitudes
SST Science Team MeetingNovember 9, 2010
Sensitivity to Other Fluxes
Results shown for Modified Kantha-Clayson model
Wick et al.
Mid-latitudes – Tropics
SST Science Team MeetingNovember 9, 2010
Sensitivity to Solar Absorption
Modified Kantha-Clayson model, tropical conditions
Wick et al.
3-Band from Fairall 9-Band
SST Science Team MeetingNovember 9, 2010
Sensitivity to Solar Absorption
Modified Kantha-Clayson model, tropical conditions
Wick et al.
3-Band – 9-Band
SST Science Team MeetingNovember 9, 2010
Real Forcing
• Goal to evaluate absolute accuracy relative to observed warming and sensitivity to temporal resolution of forcing parameters
• ETL cruise database from the R/V Ronald H. Brown– Validation against sea-snake near-surface temperature– Data courtesy C. Fairall
• R/V Ronald H. Brown cruises with the CIRIMS– Validation against skin temperature observations from
the CIRIMS– CIRIMS data courtesy A. Jessup
• Models forced with continuous meteorological observations from the ship
Wick et al.
SST Science Team MeetingNovember 9, 2010
Real Simulations
• Example of visual comparison shown here• Models ability to reproduce observations varies
notably with conditions
Wick et al.
Skin Validation
SST Science Team MeetingNovember 9, 2010
Absolute Accuracy
• Models run for entire set of cruises and differences relative to observations binned as a function of local solar time
• Results shown for wind speeds less than 4 m/s• Mean bias can be reduced to small levels but RMS differences
of O(1K) remain even with full forcing data
Wick et al.
Modified Kantha-Clayson COARE
SST Science Team MeetingNovember 9, 2010
Impact of Forcing Resolution
• Simulations re-run for Modified Kantha-Clayson model and forcing data sampled at 6-hour intervals
• Significant degradation observed in both bias and RMS relative to observations
Wick et al.
Full Resolution Interpolated 6-Hourly Data
SST Science Team MeetingNovember 9, 2010Wick et al.
Conclusions
• Modeled diurnal warming highly variable at lowest wind speeds
• Significant sensitivity to solar penetration model
• Warming also exhibits some sensitivity to background fluxes
• Bias in warming predictions can be largely removed with tuned models
• RMS uncertainty in predicted warming on O(1K) at low wind speeds