The in-situ measurement and algorithm group identified four areas on which to focus in the coming months. In doing so, it acknowledged that it seeks to build upon the recent on ongoing effort of the WCRP Working Group on Air-Sea Fluxes (WGASF).
1. The first of these is to identify, assemble, and share existing in-situ data sets to serve as possible input to satellite flux algorithms and as validation data sets for testing methods to derive satellite turbulent fluxes. These data sets would be shared at the highest possible resolution (at the original sampling rates), would include both the basic surface observables and derived fluxes, and would be accompanied by complete documentation. That documentation would specify the flux algorithms used to compute fluxes, the height of the observations, the location, the instrumentation, the sampling, and other specifics needed to fully characterize how the data was acquired.
The goal of the group is to identify, assemble, and share a small number of data sets within the next few months, so time was spent discussing how to identify the most valuable data sets. Quality of the data and completeness of the observations were judged to be a key guideline. Based on this, a small number of high priority data sets were identified as possible candidates: 1) data from deployments of the WHOI IMET buoys; 2) the FSU COAPS WOCE data set; 3) data sets from dedicated air-sea interaction research programs, including TOGA COARE, FASTEX/CATCH, ASTEX/SEMAPHORE/Subduction, SCOPE, JASMINE, GASEX 98, NAURU 99, KWAJEX; 4) the latest versions of the Volunteer Observing Ship (VOS) data set, COADS/WMO47, and 5) data from the Ocean Weather Stations (OWS).
From this list, the group noted that there will be a handful of high quality, comprehensive data sets that have been used for research, are well-documented and ready for the data files to be sent to a Flux Group archive. The TOGA COARE data set from the tropical western Pacific and the Mid-latitude North Atlantic data set resulting from ASTEX, SEMAPHORE, and Subduction were identified as being such data sets.
It was noted that while other data sets may not be as complete or as proven, they may have characteristics that make them particularly valuable and thus worth including in the Flux Group archive. Identified as desirable were data sets that: a) included observations of skin SST, b) included wave observations, particularly wave directional spectra, c) were obtained in high wind conditions, d) were able to support further study of flux algorithms and the physics of air-sea coupling, e) include information of the spatial variability in the upper ocean that would help understand spatial variability in SST, and f) included coincident collection of radiosondes. It was also noted that data more recent than 1996 would be valuable because more satellites data sets were available. Some caution was noted in devoting time and effort to include data sets from near the coast, as some satellite data cannot be used for driving ocean surface turbulent fluxes closer than ~50 km from the coast.
An action item for members of the group was to circulate a form to be made available from the WGASF to more completely develop a data base of candidate data sets and of the metadata associated with them, to prioritize that list, and to work toward submitting before the end of the year these data sets to the Flux Group archive. Another action item was to use the Flux Group web site to coordinate future observational efforts to maximize their benefit to those working on satellite-based retrievals of the turbulent fluxes. It was thought that by posting early notice of future cruises and experiments, opportunities to augment the data to be collected could be identified.
2. The second is to recognize the limitations of the existing data sets and algorithms and to work to remove them. In particular, the group noted that we lack data obtained in high winds and in stable conditions. We also lacked data sets that would allow us to further specify the impact of surface waves and mesoscale heterogeneity (including gustiness) on the turbulent fluxes. As a result, our present data sets and flux algorithms are valid in only part of the parameter space in which we would like to derive turbulent fluxes from satellite data.
Recognition of these limitations leads to the recommendation that future observational efforts be encouraged to remove these limitations, with a particular recommendation for observations to be made in high winds.
3. The group will seek to develop on the website an informational section on flux algorithms. This site will provide the code and supporting references for a representative selection of four flux algorithms: TOGA-COARE, University of Arizona, BVW, and Dupuis. It will also identify and list the references for those algorithms used to compute fluxes in the data sets submitted to the archive. To provide a summary of the differences among these algorithms, Xuben will use a common data set to compute and compare the turbulent fluxes using each of these algorithms. Finally, a test data set of surface meteorological time series will be posted along with the fluxes computed from the four representative algorithms.
4. The final focus area was on the collection of future in-situ data. It was felt that the group should succinctly state its belief that certain types of data are of high priority for validation and production of satellite-derived turbulent flux fields and for further development of flux algorithms. In particular the group planned to develop a statement advocating flux reference sites (well-instrumented surface mooring to be deployed to collect long time series if surface meteorology and air-sea fluxes at select sites around the world's oceans), upgraded VOS on select routes, and further air-sea interaction process studies.
