The Physical Basis of SST MeasurementsOne (biased) look at progress

Gary A. WickNOAA ESRL/PSD

October 29, 2013

Physical Basis – Priority Topics

• Diurnal Variability• Spatial/Temporal Variability• Skin effect

Physical Basis – Priority Topics

• Diurnal Variability• Spatial/Temporal Variability• Skin effect• SST / Sea Surface Salinity Interactions

SST/SSS Interactions

• Talks today

Skin Effect

• Undertake additional research into the physics and modeling of the skin effect

• Improve our understanding of how errors in model forcing parameters impact skin model errors

• Relevant measurements continue– MAERI and other radiometric time series

• Marginal Ice Zone Processes Experiment - MIZOPEX– Radiometric and in situ SST measurements near melting

ice

Spatial/Temporal Variability

• Develop methods to integrate an improved understanding of the SST variability into the SST retrievals error

• Expand the library of sub-pixel scale SST data sets– Improve understanding of sub-pixel scale variability to better characterize the

contribution of point-to-pixel errors to overall retrieval uncertainty estimates– Explore methodologies to validate representation of spatial variability in SST

products

• Investigate development of a high-resolution μ-wave sensor

Spatial Variability - Activities

• MIZOPEX– UAS-based infrared imagery collocated with

surface observations– Satellite-buoy comparisons

• Frontal Work

L4 SST Analysis Evaluation in the Beaufort Sea

• Castro et al., in preparation

Diurnal Variability

• Continue/expand research into development of diurnal warming models and analysis of satellite observed diurnal warming

• Improve specification of diurnal model uncertainty• Continue and expand research into the role of penetrating

radiation and its relationship to available optical properties in the water column

• Perform additional field observations of diurnal warming events• Develop/enhance links with other communities with interests in

diurnal warming such as the ocean color and the meteorological communities

Diurnal Variability - Progress

• Measurements– Satellite-based

• Diurnal warming matchup database• TWP+• Climatology work

– In situ• Argo• SPURS

• Models

GHRSST XIV, Woodshole, 17-21June 2013

11

A DW dedicated Matchup database

Overview - DW dedicated MDB based on SEVIRI data. - 20 S to 80 N and 80 W to 80 N - June to September 2012.- daily files in ~1 GB netCDF4 format.

Content- Continuous drifting and moored buoy measurements-“sst” and “flx” data in a 5-pixel square box in satellite coordinates-box is centred at the nearest pixel to the buoy position-Model outputs:

- air temperature at 2m, air humidity at 2m, - surface pressure and integrated water vapour- wind speed at 10m

Sonia Péré, Anne Marsouin, Gérard Legendre, Pierre Le BorgneMétéo-France/Centre de Météorologie Spatiale, Lannion, France

GHRSST XIV, Woodshole, 17-21June 2013

12

Conclusion (still many mysteries…)

Anyone interested please contact

Gerard.Legendre@meteo.fr

AATSR, METOP,… Local time

Aims• Assess multiple satellite SST products

over TWP• Quantify diurnal warm-layer events

using satellite data• Assess ≥ 8 diurnal warming models

run using common inputsPeriod

Jan–Apr 2010Location

25°S to 15°N, 90°E to 170°ECollaborators

Bureau of Meteorology, NOAA, Meteo- France, Met Office, WHOI, Uni of Colorado, Uni of Edinburgh, Uni of Arizona, Uni of Miami, JMA

Tropical Warm Pool Diurnal Variability Project (TWP+)

Wick Modified Kantha-Clayson DW

MTSAT-1R Observed DW

Castro Look-Up Table DW

26 Apr 2010Helen Beggs, Bureau of Meteorology

TWP+ Data Set• Satellite SST

• IR: AVHRR (METOP-A, NOAA-17,18 & 19)

• IR dual-view: AATSR (Envisat)• MW: AMSR-E (AQUA)• MW: WindSat (Coriolis)• Geo: JAMI (MTSAT-1R)

• Bureau Regional foundation SST Analysis (RAMSSA)

• In Situ SST (buoys, ships)• Bureau Numerical Weather Prediction

surface flux fields (ACCESS-R)• Bureau Sea State Forecasts

(AUSWAM)• 7 DV Model outputs

Tropical Warm Pool Diurnal Variability Project (TWP+)

MTSAT-1R Observed DW

Castro Look-Up Table DW

26 Apr 2010

ACCESS-R 10m Wind

Diurnal Warming Climatology

• NASA funded project• How large is DW throughout the globe and

how do magnitudes vary with season?• How does frequency vary with magnitude?• What are the spatial scales of DW?• Influences on air-sea fluxes and climate

variability

Diurnal Variability - Progress

• Measurements– Satellite-based

• Diurnal warming matchup database• TWP+• Climatology work

– In situ• Argo• SPURS

• Models

Comparison of Subskin and Foundation Estimates (Argo – SEVIRI)

Warm Count Bias Stdev Bias Stdev

Yes 223 -0.04 0.39 0.23 0.44

No 405 -0.10 0.37 -0.09 0.35

All 628 -0.08 0.38 0.03 0.42

Warm Count Bias Stdev Bias Stdev

Yes 211 -0.02 0.36 -0.16 0.36

No 495 -0.11 0.39 -0.10 0.38

All 706 -0.09 0.38 -0.12 0.39

Subskin SST

Foundation SST

Castro et al., RSE, doi:10.1016/J.RSE.2013.08.042

Comparison of Derived Diurnal Warming

Warm Counts

Bias Stdev

Yes 192 0.00 0.25

No 317 0.02 0.17

All 509 0.01 0.21

Castro et al., RSE, doi:10.1016/J.RSE.2013.08.042

Dedicated ARGO Diurnal Warming Experiment

• Discussed at GHRSST 2013 Science Team Meeting• Dedicated experiment “perfectly doable”

– Majority of new floats use iridium communications– Float could repeatedly sample near-surface ocean for

period of 3-5 days– Willing to explore sampling closer to surface

• Group to develop plans for experiment of opportunity– Explore forecasting capability

Diurnal Variability - Progress

• Measurements• Models

– Clayson and Bogdanoff– Weihs and Bourassa– Karagali and Høyer– Coupled modeling– Scanlon et al.– Wick– …

GHRSST DVWG Focus

• Desired approaches/requirements for diurnally resolved SST– Provision of diurnal warming estimates

• Temporal frequency• Representative depth• Representative spatial scale

– Direct provision of models/parameterizations– Provision of error characteristics for existing models

and predictions– Increasing available validation data– Understanding of basic physics

Scanlon et al., Ocean Science, in press

• Compared simulated near-surface temperature profiles from the Kantha-Clayson model with observations from SkinDeep

• Individual profile comparison difficult, but reasonable agreement in average sense

• Base scheme performed well• Provides some estimate of

errors with depth

Real-Time Diurnal Warming Estimates Forced with GFS Model Inputs

Gary A. WickNOAA ESRL/PSD

Motivation

• Compensation for presence of diurnal warming in SST analyses

• Facilitation of further comparison and validation of models

Components

• Model Inputs– GFS analysis fields, 6 hourly, 0.5 degree

• Wind stress• Radiative and turbulent fluxes

– Wave Watch III Wave Model• Wave period, direction, and significant wave height

• Models initially evaluated– Kantha Clayson with wave effects– COARE– Parameterizations

Sample Forcing Fields

Logistics

• Models initialized based on SST• Fluxes interpolated to model time step• Model run globally for 2 days with output

taken from the second day• Warming taken as Tsubskin - Tdepth

• Warming compared against SEVIRI observations

Kantha Clayson Stokes Drift ModelDiurnal Warming Evolution, 00-23 UTC, 21 March 2013

SAMPLE MODEL RESULTS

Comparison of Warming with SEVIRI Observations

• SEVIRI warming computed as SSThour – SSTFoundation

• SSTFoundation computed as nighttime average

• Initial comparison in terms of warming distribution– Tests models ability to reproduce

range of diurnal warming– Less sensitive to errors in model

wind stress• Supplementary absolute

comparisons to follow

SEVIRI-Derived Diurnal Warming1400 UTC, 21 March 2013

Comparison of Observed and Modeled Diurnal WarmingSEVIRI Domain – 21 March 2013, 1400 UTC

K-C Stokes Drift K-C Stokes Enh (2) K-C Stokes Enhanced

COAREWick K-C Blended Gentemann

Comparison of Observed and Modeled Diurnal WarmingSEVIRI Domain – 21 March 2013, 2200 UTC

K-C Stokes Drift K-C Stokes Enh (2) K-C Stokes Enhanced

COAREWick K-C Blended Gentemann

TIME EVOLUTION OF WARMING DISTRIBUTIONS FOR MODELS

Comparison of Observed and Modeled Diurnal WarmingSEVIRI Domain – 21 March 2013

Kantha Clayson Model with Stokes Drift

0200 UTC 0600 UTC 1000 UTC

1800 UTC1400 UTC 2200 UTC

Comparison of Observed and Modeled Diurnal WarmingSEVIRI Domain – 22 March 2013

Kantha Clayson Model with Stokes Drift

0200 UTC 0600 UTC 1000 UTC

1800 UTC1400 UTC 2200 UTC

Summary

• Automated daily computation of global diurnal warming from NWP forcing coming on line for application to SST analyses

• Distribution of diurnal warming reasonably well reproduced

• Some instances of persistent warming, but not entirely without support

• Display, distribution, and further evaluation of products to follow

Extra Slides

Gentemann ParameterizationDiurnal Warming Evolution, 00-23 UTC, 21 March 2013

COARE ModelDiurnal Warming Evolution, 00-23 UTC, 21 March 2013

Scanlon et al., Ocean Science, in press

• Compared simulated near-surface temperature profiles from the Kantha-Clayson model with observations from SkinDeep

Comparison of Warming with SEVIRI Observations

• SEVIRI warming computed as SSThour – SSTFoundation

• SSTFoundation computed as nighttime average• Initial comparison in terms of warming distribution

– Tests models ability to reproduce range of diurnal warming

– Less sensitive to errors in model wind stress• Supplementary absolute comparisons to follow

Conclusions: The potential utility of Argo data for diurnal warming applications has been a key question for the DVWG

Results demonstrate remarkable consistency between the estimates of the subskin and foundation temperatures and corresponding diurnal warming from SEVIRI and unpumped Argo. This lends support to both products and associated methodologies.

Unpumped Argo floats provide accurate estimates of diurnal warming. They constitute a very valuable independent data set for L4 SST validation and diurnal warming studies, suggesting we advocate for more unpumped Argo floats.

Castro et al., RSE, in press