Air sea satellite flux datasets and what they do (and don't) tell us about the air sea interface in the Southern Ocean

Air sea satellite flux datasets and what they do (and don't) tell us about the air sea interface in the Southern Ocean Carol Anne Clayson Woods Hole Oceanographic Institution Southern Ocean Workshop Seattle, WA 18 March 2014

Current Satellite/Blended Datasets Goddard product: GSSTF3 Daily, 0.25, input variables and turbulent fluxes; satellite plus Ta from reanalysis 1988-2008; global oceans No current plans to update again funding issues IFREMER version 3 Daily, 0.25, input variables, turbulent fluxes; satellites plus Ta from reanalysis Currently available: 1992 (1999 with QuikSCAT) November 2009; global oceans Japanese Ocean Flux datasets: J-OFURO2v2 Input variables, fluxes, radiation, satellites plus Ta from reanalysis Daily, 1, 1988 2005; global oceans Satellites, JMA model analyses HOAPS3.2 6-hourly, 0.5, global oceans; input variables, precipitation; satellites July 1987 - December 2008 OAFlux Daily, 1, global oceans; blended using reanalysis, in situ, satellites July 1985 current (monthly available back to 1958)

SeaFlux Climatological Data Set Version 1.0 International project under the auspices of the WCRP GEWEX Data and Assessments Panel: to improve our understanding and determination of ocean surface turbulent fluxes Near surface air temperature and humidity Roberts et al. (2010) neural net technique SSM/I only from CSU brightness temperatures (thus only covers 1997 2006) Gap filling methodology use of MERRA variability 3 hour Winds Uses CCMP winds (cross calibrated SSM/I, AMSR E, TMI, QuikSCAT, SeaWinds) Gap filling methodology use of MERRA variability 3 hour SST Pre dawn based on Reynolds OISST Diurnal curve from new parameterization Needs peak solar, precip Uses neural net version of COARE Available at http://seaflux.org 1999 Latent Heat Flux 1999 Sensible Heat Flux

Latent Heat Flux: 1999 2005 W m -2

Sensible Heat Flux: 1999 2005 W m -2

Some comparisons Bias: 2.1 W m 2 Bias: 3.1 W m 2 Std Error: 38 W m 2 Std Error: 13.2 W m 2 Here we do comparison with eddy covariance fluxes from research vessels they are our ground truth

In-situ measured fluxes used in comparisons Southern ocean validation data typically entirely missing

SeaFlux Uncertainty Estimates

Satellite derived humidity Figures adapted from Prytherch et al. (2013) Monthly differences >> accuracy requirements Strong dependence on source of SSM/I data Differences vary: regionally seasonally over longer time scales Biases not easy to relate to in situ data Operational buoy q a may be low quality

Seasonal variability from satellite (Clayson et al. 2013)

Seasonal variability Latent Heat Flux Sensible Heat Flux Yu et al. (2011)

Mean heat fluxes in Drake Passage Flux Product Region relative to Polar Front Annual mean heat flux (W m 2 ) NCEP north 29.5 ± 1.6 south 31.7 ± 1.6 J OFURO north 22.1 ± 1.8 Annual cycle similar for all products south 41.6 ± 1.9 OAFlux north 39.4 ± 1.7 south 56.3 ± 1.9 From Stephenson et al, JGR, 2012 Mean surface heat flux differs substantially

Diurnal variability from satellite (Clayson et al. 2013)

Distributions in the Southern Ocean Sensible Heat Flux Latent Heat Flux Smith et al. (2010)

Comparisons in the Southern Ocean Sensible Heat Flux Latent Heat Flux Smith et al. (2010)

Sensible Heat Flux Extremes

Weather States Weather states defined as in Tseloudis et al. 2013 3 hourly, 2.5 o x 2.5 o boxes Use ISCCP data K-means clustering of joint frequency distributions of cloud top pressure and cloud optical thickness More convection SeaFlux data sorted using the weather states defined by cloud properties Less convection from Tseloudis et al (2013)

Southern Ocean regimes and fluxes

Total heating by regime

Distributions during regimes

Wind speed distributions

Stability differences

Conclusions from joint US CLIVAR/SeaFlux workshop on High Latitudes Acquire more in situ observations (both limited duration focused on physical processes and long time series at given locations) Bulk parameters Direct flux observations (important also for improving bulk flux parameterizations) Improved satellite flux datasets needed (coincident measurements helpful) Increase accessibility of available datasets More flux intercomparisons with standardized methods

Spatial and temporal scales for high latitude processes and the recommended accuracy of related surface fluxes. These accuracies are estimates from a wide range of scientists who require flux products for their research. Bourassa et al. 2013

Need to define what accuracy is for The net heat flux error which would provide a seasonal SST spread less than one half the peak to trough seasonal SST variability (Clayson and Bogdanoff 2013).