Satellite-based air-sea heat fluxes using GHRSST L4 analysis products Chris Jeffery NOAA National Oceanographic Data Center

2 May 29, 2009Chris Jeffery – GHRSST Users Symposium Air-sea heat exchange Direct measurement requires a complex suite of sensors and high frequency measurements, hence amount of research quality data is small >> Flux parameterization. Measurements of flux-related variables (U, Ta, Ts, qs, qa) are readily available from marine surface weather reports from voluntary observing ships (VOS), buoys/research ships and satellites. BACKGROUND  DATA  FLUXES  CONCLUSIONS Sensible heat fluxLatent heat flux Air density Latent heat of vaporization Sea-air temperature difference Sea-air humidity difference Wind speed Turbulent exchange coefficient Air specific heat capacity

3 May 29, 2009Chris Jeffery – GHRSST Users Symposium Project aims Create satellite derived heat flux products using high resolution SST analyses and other satellite based variables. Evaluate impact of GHRSST products on flux estimates –Through inter-comparison of SSTs and subsequent fluxes. –Comparison with existing flux datasets such as OAFlux, GSSTF, NCEP/NCAR, ECMWF. Initially seek to utilize L4 high resolution SST products to improve estimates of air-sea fluxes. [Try L2(P), L3 in the future?] –Different analysis products for different uses e.g. noise vs. structure. –Which L4 Analyses are the ‘best’ for flux estimates? Validation with fluxes from: TOGA-COARE buoys; NOCS surface flux climatology; SEAFLUX in situ data. BACKGROUND  DATA  FLUXES  CONCLUSIONS

4 May 29, 2009Chris Jeffery – GHRSST Users Symposium Initial setup 4 independent variables needed to calculate Hs and Hl – SST, u10, Ta and qa. –These can all be acquired by remote sensing. –Additional variables derived from these e.g. qs = f(SST). –Also include SWR and LWR to estimate net heat flux (Qnet). Calculate 1  x1  daily latent and sensible heat fluxes for ; produce 3- day, weekly, monthly averages. COARE v3.0 algorithm used - determined to be one of the least problematic for computing ocean surface fluxes [Brunke et al., 2003]. L4 GHRRST datasets used so far [availability]: – NCDC – AVHRR_AMSR_OI [Jun present]. – NCDC - AVHRR_OI [Sep present]. – REMSS – mw_IR_OI [Aug present]. – EUR – ODYSSEA [Jan present]. – UKMO – OSTIA [Mar 2006 – present]. Other inputs (qa, Ta, u10, LWR, SWR) remain fixed. Preliminary comparison with WHOI OAFlux data for 2006 [Yu and Weller, 2007]. BACKGROUND  DATA  FLUXES  CONCLUSIONS

5 May 29, 2009Chris Jeffery – GHRSST Users Symposium Group for High Resolution SST (GHRSST) Example L4 SST differences BACKGROUND  DATA  FLUXES  CONCLUSIONS

6 May 29, 2009Chris Jeffery – GHRSST Users Symposium Additional satellite datasets [Shi and Zang, 2008]. Wind Speed (U10) [Lower left] – Blended SeaWinds (NCDC) [ ]. Air temperature (Ta) [Upper right] and air specific humidity (qa) [Lower right]– derived from NOAA POES AMSU-A/B (NCDC) [ ]. Surface radiative fluxes (ISCCP) can be included to estimate Qnet [ ]. BACKGROUND  DATA  FLUXES  CONCLUSIONS

7 May 29, 2009Chris Jeffery – GHRSST Users Symposium Input data vs. OAFlux (1) BACKGROUND  DATA  FLUXES  CONCLUSIONS OAFlux Blended SeaWinds OAFlux NOAA POES AMSU OAFlux NOAA POES AMSU AVHRR AMSR OI AVHRR OI mw ir OI

8 May 29, 2009Chris Jeffery – GHRSST Users Symposium Input data vs. OAFlux (2) BACKGROUND  DATA  FLUXES  CONCLUSIONS

9 May 29, 2009Chris Jeffery – GHRSST Users Symposium Latent and Sensible heat fluxes (1) Fluxes calculated using Ts = AVHRR AMSR OI BACKGROUND  DATA  FLUXES  CONCLUSIONS

10 May 29, 2009Chris Jeffery – GHRSST Users Symposium Latent and Sensible heat fluxes (2) BACKGROUND  DATA  FLUXES  CONCLUSIONS

11 May 29, 2009Chris Jeffery – GHRSST Users Symposium OAFlux – ‘GHRSST’ Fluxes BACKGROUND  DATA  FLUXES  CONCLUSIONS

12 May 29, 2009Chris Jeffery – GHRSST Users Symposium Latent heat flux time series BACKGROUND  DATA  FLUXES  CONCLUSIONS

13 May 29, 2009Chris Jeffery – GHRSST Users Symposium Sensible heat flux time series BACKGROUND  DATA  FLUXES  CONCLUSIONS

14 May 29, 2009Chris Jeffery – GHRSST Users Symposium Summary and future work Satellite-based input parameters processed and re-gridded to common 1  x1  degree, daily grids. NOAA-COARE v3.0 bulk aerodynamic algorithm used for flux calculations. Five GHRSST L4 SST analysis products used to calculate latent (Hl), sensible (Hs) for Initial comparisons of input data/fluxes with OAFlux for BACKGROUND  DATA  FLUXES  CONCLUSIONS In situ heat flux comparisons e.g. with SEAFLUX data or ship/buoy datasets. Further GHRSST L4 datasets. Investigate other datasets for Ta, Qa, u10. Additional years and different resolutions e.g. sub-daily fluxes. Regional validation – L2/L2P GHRRST datasets.

15 May 29, 2009Chris Jeffery – GHRSST Users Symposium Acknowledgements National Research Council Associateship Program. NOAA National Oceanographic Data Center –Ken Casey (Postdoctoral Advisor). NOAA National Climate Data Center –Lie Shi and Huai-min Zhang. BACKGROUND  DATA  FLUXES  CONCLUSIONS

16 May 29, 2009Chris Jeffery – GHRSST Users Symposium Selected references Brunke et al., 2003 : “Which bulk aerodynamic algorithms are least problematic in computing ocean surface turbulent fluxes". Journal of Climate, VOL. 16, pp.“Which bulk aerodynamic algorithms are least problematic in computing ocean surface turbulent fluxes" Curry, J. A. et al., 2004 : “SEAFLUX". Bulletin of the American Meteorological Society, VOL. 85, issue 3, pp pp, doi: /BAMS “SEAFLUX" Donlon, C. J. et al., 2007 : “The global ocean data assimilation experiments high-resolution sea surface temperature pilot project”. Bulletin of the American Meteorological Society, VOL. 88, issue 8, pp , doi: /BAMS Fairall et al., 1996 : “Bulk parameterization of air-sea fluxes for Tropical Ocean-Global Atmosphere Coupled- Ocean Atmosphere Response Experiment". Journal of Geophysical Research, 101 (C2), pp.“Bulk parameterization of air-sea fluxes for Tropical Ocean-Global Atmosphere Coupled- Ocean Atmosphere Response Experiment" Shi, L., and H.-M. Zhang, 2008 : “Sea surface air temperature and humidity retrievals based on AMSU measurement". 16 th Conference on Satellite Meteorology and Oceanography. Fifth Annual Symposium on Future Operational Environmental Satellite Systems- NPOESS and GOES-R.“Sea surface air temperature and humidity retrievals based on AMSU measurement" Yu, L., and R. A. Weller 2007 : “Objectively analyzed air-sea heat fluxes for the global ice-free ocean ( )". Bulletin of the American Meteorological Society, VOL. 88, issue 4, pp, doi: /BAMS “Objectively analyzed air-sea heat fluxes for the global ice-free ocean ( )" Zhang, H.-M., J.J. Bates, and R.W. Reynolds, 2006 : "Assessment of composite global sampling: Sea surface wind speed". Geophysical Research Letters, VOL. 33, L17714, doi: /2006GL "Assessment of composite global sampling: Sea surface wind speed" Zhang, Y et al., 2004 : “Calculation of radiative fluxes from the surface to top of atmosphere based on ISCCP and other global data sets: Refinements of the radiative transfer model and the input data". Journal of Geophysical Research, VOL. 109, D19105, doi: /2003JD “Calculation of radiative fluxes from the surface to top of atmosphere based on ISCCP and other global data sets: Refinements of the radiative transfer model and the input data" BACKGROUND  DATA  FLUXES  CONCLUSIONS

Additional Slides

18 May 29, 2009Chris Jeffery – GHRSST Users Symposium Flux parameterization Measurements of flux-related variables (U, Ta, Ts, qs, qa) are readily available from marine surface weather reports from voluntary observing ships (VOS) and buoys/research ships. Atmospheric re-analyses from numerical weather predication (NWP) centers e.g. NCEP and ECMWF. Problems include: incomplete global coverage, relatively short timescales, systematic bias and random error. Comprehensive global coverage is only possible from an analysis incorporating satellite measurements. Additional slides Sensible heat fluxLatent heat flux Air density Latent heat of vaporization Sea-air temperature difference Sea-air humidity difference Wind speed Turbulent exchange coefficients Turbulent exchange coefficient Air specific heat

19 May 29, 2009Chris Jeffery – GHRSST Users Symposium Group for High Resolution SST (GHRSST) Key variable for air-sea flux calculation, as well as weather prediction and climate [Donlon et al., 2007]. Sea-air humidity difference (qs-qa) is a function of SST. NODC maintains the long term archive – /ghrsst/ GHRSST Products combine several satellite and in situ SST data streams. –designed to provide the best available estimate of the SST. –L2, L3 and L4 SST climate data record analysis products. Multiple data providers using different input data/algorithms to generate analysis products e.g. NCDC, REMSS, UKMO. [GHRSST-PP GMPE ensemble output ] Additional slides

20 May 29, 2009Chris Jeffery – GHRSST Users Symposium Existing problems Flux related variables are available from NWP models e.g. NCEP-NCAR reanalysis and 40-yr ECMWF Re- Analysis (ERA-40). These can contain systematic biases. In some regions (e.g. the Arabian Sea and Bay of Bengal). The biases can be so large that the NWP net heat fluxes have a sign opposite to those of in situ measurements. Biases in NWP products/analyses cause errors in ocean/land simulations. Comprehensive global coverage must incorporate satellite-based measurements Peter Taylor and Bob Weller, OOPC 8 (2003) Additional slides

21 May 29, 2009Chris Jeffery – GHRSST Users Symposium Simple outline SST U10 TA QA LWR SWR Re-grid data Daily averages Apply COARE 3.0 HL HS Qnet Average fluxes Inter- comparisons Compare with other data SST TOGA- COARE NWPOAFlux ISCCP NCDC NODC - GHRSST Comparison data [optional] Additional slides

22 May 29, 2009Chris Jeffery – GHRSST Users Symposium Seasonal latent heat flux (Hl) Additional slides

23 May 29, 2009Chris Jeffery – GHRSST Users Symposium Seasonal sensible heat flux (Hs) Additional slides

24 May 29, 2009Chris Jeffery – GHRSST Users Symposium Regional flux comparisons with OAFlux Additional slides