A23E-01
SeaFlux: New Methods in Satellite-Derived Air Sea Fluxes
In this talk I will present an overview of the SeaFlux program and its history and goals. A description of the SeaFlux Intercomparison Project and results will be given. Methods and comparisons of satellite-derived air-sea fluxes with in situ research vessels, buoys, and numerical weather prediction models will be described. New methodologies that are being explored with respect to improved global high-resolution air-sea fluxes will be detailed, including improved surface flux models, improved sea surface temperature fields, and possibilities for improved fluxes in coastal regions and at high wind speed regimes will be explored. A focus will be placed on diurnal variability of fluxes, and the effects of such variability on convection over the tropical Pacific as shown by cloud-resolving models will be demonstrated.
A23E-02
Air-sea Fluxes Derived From Satellite Data: Achievements and Perspectives
Time series of satellite data, suitable for retrieval of water cycle components over the ocean, approach lengths that make them attractive to be used for the analysis of inter-annual variability and trends. Additionally, they can serve as an evaluation tool for model based atmospheric reanalyses and climate models. Based on the example of the satellite-derived Hamburg Ocean Atmosphere Parameters and Fluxes from Satellite Data set (HOAPS-3) the presentation will contain some comparisons to ERA40 and control runs of the ECHAM5 climate model to elucidate the current status of similarities and differences between models and observations. The HOAPS-3 data set utilized the NOAA pathfinder sea surface temperature data set and several retrieval schemes for basic variables as near-surface humidity and wind speed applicable to the series of SSM/I instruments. The data set covers a time span from 1987-2005. Satellite based data sets are constructed from a series of instruments flying on successive platforms, e.g. SSM/I on the DMSP series and AVHRR on the NOAA series. To use those data for establishing time series suitable for trend detection a very careful correction of individual instrument and satellite platform errors has to be performed. Examples for those errors are orbit decay of the satellite that changes zenith angles over time and diurnal drift of the satellite platform aliasing in the diurnal cycle. Despite the high quality of some of those corrections a inter- sensor homogenization to a reference platform is unavoidable. The presentation will give a short review on used techniques and their advantages and disadvantages. Finally, the presentation will discuss the idea to use infrared sounding data from the IASI instrument on the MetOp satellite to improve current near-surface humidity and temperature retrievals and ways to include error information to the data sets.
A23E-03
Surface Fluxes and Climate Models
A clear distinction is made between surface flux requirements for climate model evaluation versus forcing, uncoupled ocean or atmospheric models versus partially or fully coupled, and finally air-sea versus air-ice versus ocean-ice fluxes. The forcing of uncoupled ocean climate models for Coordinated Ocean Research Experiments (CORE) is outlined, with emphasis on the role of satellite based data sets. The flux climatologies and inter-annual variability implied by this forcing are presented, and the meaning of model departures from these results is discussed. Focusing on air-sea fluxes, evaluation of model surface fluxes is restricted to fully coupled climate simulations, where, in principle, it could be a valuable tool in attributing model bias to one or multiple component models. Oceanic regions, such as the eastern coasts of subtropical gyres, will be identified as having particularly demanding surface flux requirements. Examples of what have proved to key, and also not so useful, metrics of model performance will be presented.
A23E-04
Ocean surface fluxes as simulated by climate models: How can we make progress?
Surface fluxes of heat and momentum over the ocean are a critical link in our understanding of the climate system, which has suffered from a lack of accurate flux estimates. While progress is being made, direct measurements are still rare and bulk parameterizations remain essential for estimates of turbulent fluxes made either with in-situ or satellite measurements. In this presentation we examine the mean state, variability, and trends of air-sea fluxes simulated by state-of-the-art climate models. Inherent differences between fluxes in coupled ocean-atmosphere models versus atmosphere-only models will be discussed in the context of how observationally-based products can be used for verification. Looking forward, the goals of an emerging project will be described, which strives to further integrate the efforts of a diverse range of scientists in air-sea flux research.
A23E-05
Sampling errors in VOS-based surface air-sea fluxes: estimation, impacts and minimization.
Sampling uncertainties in the voluntary observing ship (VOS)-based global ocean-atmosphere flux fields were estimated using different NWP flux products and the procedure of subsampling, simulating VOS-like sampling density in NWP flux products. The highest random sampling errors in surface fluxes were found for the sensible and latent heat flux and range from 30 to 80 Wm-2. Total sampling errors in poorly sampled areas may be higher than random ones by 60%. In poorly sampled subpolar latitudes of the Northern Hemisphere and throughout much of the Southern Ocean the total sampling uncertainty in the net heat flux can amount to 80-100 Wm-2. The largest uncertainties in linear trend estimates are found in the high-latitude North Atlantic and North Pacific as well as the Southern Ocean, where trends can locally show opposite signs when computed from the regularly sampled and undersampled data. Spatial patterns of shorter-period interannual variability, quantified through the EOF analysis, also show remarkable differences between the regularly sampled and undersampled flux datasets in the Labrador Sea and northwest Pacific. In order to minimize sampling errors in VOS fluxes we suggested a methodology of climatological averaging of fluxes based on the application of double-exponential distributions (2ePDF) of sensible and latent surface flux estimates. Application of 2ePDF allows for minimization of sampling errors from 2 to 10 times and provides much more reliable global surface turbulent flux fields. Then, using the same methodology and 125 years (1880-2004) of VOS observations from ICOADS we reconstruct surface ocean- atmosphere heat fluxes over the North Atlantic with monthly resolution in time and variable (2-degree to 5-degree) resolution in space. Produced air-sea flux fields show reasonable minimization of sampling errors and allow for the analysis of regional heat balances and estimation of long-term changes in surface fluxes. Further analysis included computation of monthly anomalies of surface fluxes as well as estimation of the subpolar gyre heat and freshwater budgets. These were computed using two-demensional distributions of surface fluxes in the coordinates of sea-air temperature difference and wind speed. Reconstructed fluxes reveal long-term trends, implying, for example, about 4 W/m2 per decade growing sensible heat fluxes in the Labrador Sea and about 2 W/m2 per decade secular increase in the Central subpolar gyre. Non-secular signals are represented by the decadal-scale and multidecadal (about 40-50 years variability). Decadal scale signal has a clear association with the NAO-like atmospheric circulation variability during 1880-1915 and after 1955, but has a little association with NAO between 1915 and 1955.
A23E-06
Trends and EOF Analyses of Oceanic Evaporation Data Sets
The oceanic fresh water flux (evaporation - precipitation) determines the density of the upper oceans and drives the surface and deep circulations of the oceans. While oceanic precipitation has received much attention due to the success of TRMM and the planned follow-on mission, Global Precipitation Mission (GPM), major research activities in oceanic evaporation has only been revitalized recently. To evaluate the present state of remote sensing oceanic flux products, a comparison between three remotely sensed oceanic evaporation products has been made. These products include the Goddard Satellite Surface Turbulence/Flux data version 2 (GSSTF2), the Japanese Ocean Flux Data Set with Use of Remote Sensing Observations (J-OFURO) and the Hamburg Ocean Atmosphere Parameters and Fluxes from Satellite Data version 2 (HOAPS2) for the period of overlap. Evaporation from the NCEP reanalysis was also included for model comparison. All data sets show global positive trends, with largest in J-OFURO and least in HAOPS2 and NCEP reanalysis. Orthogonal Function (EOF) Analyses revealed an ENSO signal in all data sets, which is second, in variance explained, to a pattern characterized by opposition of tropical to subtropical changes (Xing, 2006). This first mode is interpreted as an enhancement of the tropical Hadley circulation, which is consistent with model results of enhanced hydrologic cycle of warming scenarios (Chiu and Xing, 2004). As all datasets rely on Special Sensor Microwave Imager (SSM/I) data, these spurious trends may be due to inter-satellite calibration of the SSM/Is. A recent release of the SSM/I data (Version 6) has the spurious trends removed. A comparison of the winds from GSSTF2 and the V6 of Wentz's winds show rather different trend patterns. To enable the use for climate studies, we endorsed the call by Wentz and propose the reprocessing of oceanic flux products using the current version (V6) of SSM/I.
A23E-07
Evaporation flux variability in the tropical Pacific from observations and climate models
Turbulent fluxes (evaporation, winds and humidity) at the surface of the ocean from several satellite-derived observational datasets (WHOI, HOAPS, J-OFURO and GSSTF2) are inter-compared and evaluated against in situ measurements from the TAO buoy record in the Tropical Pacific Ocean. Similarly, turbulent flux variability from 10 state-of-the-art climate models that went into the recent IPCC report (AR4) is assessed compared to the observational datasets. We focused on longer (inter-annual) and shorter (intra-seasonal) scales over the period 1989-2000 and on how the relationship between the flux and the state variables is captured in the different datasets and models. For the period of interest, we find that all models overestimate evaporative fluxes in the Warm Pool region by about 30W/m2 whereas the observational uncertainty is 20W/m2 and the bias arises due to exceptionally dry conditions the models consistently display in this region. In the Niño3 region modeled variability over longer time scales better agrees with the observations, although we do not expect models to capture the observed record of El Niño events. Shorter scale variability is not well captured in models, due to unresolved scales such that of the tropical instability waves.
A23E-08
Introduction of J-OFURO version 2 surface heat flux data set and its analysis over the North Pacific
Japanese Ocean Flux Data Sets with Use of Remote Sensing Observations (J-OFURO) includes global ocean surface heat flux data derived from satellite data and are used in many studies related to air-sea interaction. Recently new surface heat flux data was constructed in J-OFURO as the version 2. In the version 2 many points are improved compared with the version 1. Since we used wind speed and specific humidity data derived from one DMSP/SSMI sensor in the version 1, we obtained two data at most one day. Therefore, there may be large sampling errors for the daily-mean value. In order to escape this problem, multi-satellite data (DMSP/SSMI F08- 15, Aqua/AMSR-E, TRMM/TMI, ERS/AMI and QuikScat/SeaWinds) are used in the version 2. As a result we could improve accuracy and temporal resolution from 3-days mean value in version 1 to daily-mean value in version 2. Also we used an Optimum Interpolation method to estimate specific humidity data instead of a simple mean method. We basically need sea surface temperature (SST), specific humidity and wind speed data for estimation of latent heat flux. In version 1 we used NCEP data (Reynolds and Smith, 1994) as SST data. However, the temporal resolution of the data is based on weekly and considerably low. Recently there are many kinds of global SST data because we can obtain SST data using a microwave radiometer sensor such as TRMM/MI and Aqua/AMSR-E. Therefore, we compared many SST products and determined to use Merged satellite and in situ data Global Daily (MGD) SST provided by Japan Meteorological Agency. A bulk algorithm used for estimation of turbulent heat flux is changed from Kondo (1975) to COASRE 3.0(Fairall et al., 2003). Shortwave and longwave radiation data are based on the ISCCP product and some modifications are carried out for longwave radiation. Finally surface latent and sensible flux data and shortwave and longwave radiation data are extended to1989- 2004. In this presentation we will introduce surface heat flux data in J-OFURO version 2 and comparison and validation results of latent heat flux with data such as GSSTF2, HOAPS and OAFlux etc. Moreover, we analyze variability of surface heat flux over the North Pacific.
A23E-09
Global High Resolution Sea Surface Flux Parameters From Multiple Satellites
Advances in understanding the coupled air-sea system and modeling of the ocean and atmosphere demand increasingly higher resolution data, such as air-sea fluxes of up to 3 hourly and every 50 km. These observational requirements can only be met by utilizing multiple satellite observations. Generation of such high resolution products from multiple-satellite and in-situ observations on an operational basis has been started at the U.S. National Oceanic and Atmospheric Administration (NOAA) National Climatic Data Center. Here we describe a few products that are directly related to the computation of turbulent air-sea fluxes. Sea surface wind speed has been observed from in-situ instruments and multiple satellites, with long-term observations ranging from one satellite in the mid 1987 to six or more satellites since mid 2002. A blended product with a global 0.25° grid and four snapshots per day has been produced for July 1987 to present, using a near Gaussian 3-D (x, y, t) interpolation to minimize aliases. Wind direction has been observed from fewer satellites, thus for the blended high resolution vector winds and wind stresses, the directions are taken from the NCEP Re-analysis 2 (operationally run near real time) for climate consistency. The widely used Reynolds Optimum Interpolation SST analysis has been improved with higher resolutions (daily and 0.25°). The improvements use both infrared and microwave satellite data that are bias-corrected by in- situ observations for the period 1985 to present. The new versions provide very significant improvements in terms of resolving ocean features such as the meandering of the Gulf Stream, the Aghulas Current, the equatorial jets and other fronts. The Ta and Qa retrievals are based on measurements from the AMSU sounder onboard the NOAA satellites. Ta retrieval uses AMSU-A data, while Qa retrieval uses both AMSU-A and AMSU-B observations. The retrieval algorithms are developed using the neural network approach. Training datasets are constructed using co-located AMSU and buoy/ship data. Using the 2002 one-year data we will show that the RMSE for Ta is about 1.94°C. Data description, visualization, sub-setting and downloading in user preferred formats can be obtained at: http:nomads.ncdc.noaa.gov:8085/las/servlets/dataset; ftp:eclipse.ncdc.noaa.gov/pub; and http:www.ncdc.noaa.gov/oa/satellite.html.