Report of the Second Workshop

Introduction: Since the first SSIWG, in February 1998 (see SSWIG-1 Final Report), several events of concern to salinity remote sensing have transpired. Satellite proposals and concept papers for joint soil moisture, salinity, and related measurements have been presented to space agencies, described briefly below. Certain of these remain under consideration, either as mission proposals or preliminary design studies. New airborne salinity sensors have been built or are being developed, and science and engineering test flights have been planned, postponed and re-planned. Accordingly, the second SSIWG adopted two central objectives for this workshop: 1) To review recent and emerging satellite proposals, concepts, airborne sensors, field experiments, and climate programs, and 2) from these, to discuss strategies for bringing a salinity measuring satellite mission into being.

Review of recent developments: The HYDROSTAR proposal, joint between U.Michigan and GSFC, was submitted to the NASA Earth System Science Pathfinder (ESSP) in 1998. The mission has been in development for several years, primarily designed for soil moisture science, with salinity as an experimental science objective. Despite the excellent efforts of the proposing team, it was learned in December that it was not among the three missions selected. Meanwhile, during the 1998 summer, NASA requested information and ideas on post 2002 remote sensing programs in earth science. Salinity figured prominently as the top priority for new ocean measurements. At about the same time, the NASA Instrument Incubator Program (IIP) approved conceptual design studies for advanced real aperture and thinned array sensors at JPL and GSFC respectively.

In Europe, salinity and soil moisture satellite designs studies also have been carried out in recent years at ESA, CNES and other institutions. A major development in 1998 was the SMOS (Soil Moisture Ocean Salinity) mission proposal submitted to ESA in November (see http://www-sv.cict.fr/cesbio/smos). On 31 March, 1999, it was announced that SMOS was ranked second among the top five in a short list out of the original 27 proposals. The ESA Programme Board for Earth Observation(PB-EO) is expected to select two (plus one backup) from this short list sometime in May 1999. The prospects for SMOS appear excellent right now, and it may be the first SSS observing satellite mission to be approved.

Agenda synopsis: The first order of business at the SSIWG-2 was a review of various salinity measuring satellite concepts now under study. This began with SMOS, and was followed by reviews of concept studies at NASA GSFC and JPL, and then a low cost alternative proposal by Cal Swift, U. Massachusetts. Next were presentations of science and programmatic concerns, including sea ice and high latitude oceans. Various aircraft sensors and planned field experiments were discussed for both European and U.S. programs. The meeting concluded after a discussion future strategies and activities. Specific recommendations are given at the end of this report.

SMOS satellite: M. Martin-Niera, ESA, described the mission selection process in ESA and noted that if SMOS is approved as the second mission in the queue, launch would be in about year 2005. SMOS is a merger of MIRAS (an ESA technical development program) and RAMSES (a CNES mission proposal). He then briefly described the MIRAS two-dimensional interferometry study that is the conceptual basis of the SMOS sensor. Y.Kerr, the SMOS lead investigator, then outlined the mission objectives and specifications. The sensor is a Y-shaped array, 2-D interferometric, 1.4 GHz, dual polarization radiometer (Figure 1). The orbit will be 6am/pm helio-synchronous at an altitude of 757 km, and the mission is planned for 3-5 year duration. Preliminary image reconstruction test results from an experiment in Avignon were presented. P.Waldteufel explained the interferometry in more detail and described the surface field of view (FOV) pattern (Figure 2). The pattern allows views at multiple incidence angles and pseudo-conical scanning. This makes it possible to develop a multi parameter multi angle retrieval approach requiring heavy processing but reducing measurement noise (< 0.1 psu). The ocean grid resolution can be selected according to accuracy and resolution needs and mapped directly with the inverse transform. It may also be possible to add cross polarization correlators to derive the Stokes-3 term and correct for ionospheric Faraday rotation, but there will be a loss of sampling frequency of the other channels. A.Camps then discussed several engineering issues regarding inverse Fourier transform errors, antenna calibration, and so forth.

SMOS salinity science: J.Font introduced the SMOS ocean measurement perspective. High resolution (minimum 30 km) single pass uncertainties will be large (>1 psu), but are expected to be much smaller (~0.1 psu) at coarser grid scale (i.e. 200 km) using multi view and averaging techniques, assuming independent errors. J.Etcheto discussed the importance of salinity as a circulation tracer and for many other applications such as the carbon cycle in the ocean. A strong correlation between SSS and PCO2 is observed in the tropical Pacific. It was demonstrated how SSS is a better proxy for PCO2 than SST, and that SSS data can be used to improve models of air-sea carbon flux. J.Boutin discussed modeling the uncertainties of SST and wind speed on SSS retrieval errors from SMOS. Using S.Yueh's emission model, she estimated the error per pixel due to sensor noise, SST error and wind speed error, with the rms result of about 0.8 psu, with the highest error source being due to wind speed error. This finding precipitated some debate, however, because the nature of the wind speed (or roughness) effect of 1.4 Ghz brightness is not well understood and will be the subject of future field experiments. J.Font then opened discussion of pre-launch field measurement campaigns from aircraft, ships of opportunity and oil platforms. The Casa Blanca oil platform in the western Mediterranean Sea appears to be an excellent site for radiometer calibration studies and a discussion ensued concerning radiometers in Europe and the U.S. which could be made available for experiments there during the next couple of years.

OSIRIS: E.Njoku led off the discussion of the large mesh antenna sensor and satellite concept being pursued at NASA/JPL. OSIRIS is designed with the primary objective of obtaining ocean salinity retrievals with the highest possible measurement accuracy using current technology. The philosophy is to use a real aperture system to attain a low noise and well calibrated multi-channel radiometer system with the necessary ancillary measurements to correct important sources of geophysical error such as SST, surface roughness and ionosphere. Large aperture mesh antenna designs have been developed by aerospace companies for microwave telecommunications, are light weight, and can be adapted to passive microwave radiometry. W.Wilson described more details of the system, which includes a conical scanning ~6 m antenna, and constant incidence angle viewing geometry that will allow forward and backward beams to be averaged (Figure 3). The baseline radiometer will measure both 1.4 and 2.7 GHz (L&S-band) with H&V and HV (Stokes-3) polarization using a flexible design that makes it relatively inexpensive to add channels. An optional L-band radar for wind and sea state correction is also being evaluated. Antenna beam efficiency is expected to exceed 95%. The flexible bandwidth over the ocean can be employed to reduce the rms noise per pixel to~0.1 K (about 0.2-.3 psu salinity error). T.Liu then described a scaled-down version of this concept that is being suggested as an enhancement to the AMSR system on Japan's ADEOS-III mission. Some very preliminary discussions with NASDA have transpired regarding this idea.

S.Yueh presented recent results of a very rigorous simulation study of salinity retrieval errors from OSIRIS. A self consistent analysis of radiometer sensor error, calibration error and geophysical model noise was made using data from satellite wind and SST sensors and ocean model surface salinity (provided by R. Tokmakian) over a two month period (September and October 1996). The same analysis is to be performed for the time period from November 1996 through June 1997 to investigate seasonal effects. The radiometer NEDT is less than 0.1 Kelvin for 100 km integrated footprint. A calibration error of 0.2 K was assumed with a correlation time of 4 minutes. The long correlation scales of calibration error noise leave small signatures that are not easily removed by simple averaging. An error residual table was as follows:

Accumulated errors over 100 km scales for 1-month averages are <0.1 psu in the tropics and 0.1 to 0.2 psu globally (Figure 4, Figure 5). Calibration stability will require more attention to further reduce error due to long correlation scales.

Lower cost alternative: C.Swift presented a preliminary concept designed for open ocean salinity measurements at a coarse resolution (100-200 km) with less complicated hardware and lower cost. The system would comprise of a relatively small aperture antenna pointed at nadir which could be rotated periodically to obtain a cold sky calibration. The emphasis would be on high accuracy at the expense of spatial and temporal resolution. A global map would be sampled 1-3 times per month, similar to satellite altimetry. Various commercially available antenna systems could be adapted to this concept.

Sea ice: D.Cavalieri discussed the issue of melt ponds in floating sea ice that are ambiguous with open water and alias multi-year sea ice concentration retrievals during the Arctic summer. This cannot be resolved with the higher frequency microwave (>18 GHz) data presently available. Lower frequency microwave data would allow discrimination of relatively fresh melt pond water from ambient sea water and significantly improve ice concentration estimates. A new algorithm using C and L bands was proposed that would use the C-band polarization ratio for good ice-water discrimination and the C-L gradient ratio for salt vs fresh water contrast. Per Gloersen reviewed some observations of thin sea ice with simultaneous observations on the surface and using airborne L-band and infrared sensors on board the NASA CV-990 during the Joint US/USSR Bering Sea Experiment, which took place in 1973. The observations showed the capability of observing sea ice thickness up to about 0.5 m. He also described an optical model of sea ice that gave results in agreement with the observations. S. Hakkinnen discussed numerical model studies in the Labrador and Greenland Iceland Seas. She showed that at the average the SSS anomalies resulting from sea ice export through Denmark Strait may reach 0.5psu but they usually do not spread extensively away from the coast of Labrador. Equally large anomalies can develop as a result of decreased Labrador Sea convection which have much more potential to spread over the subpolar gyre (typical strength when they reach the central subpolar gyre is 0.2psu). She suggests that during late sixties both factors, increased ice export and lack of convection, created a lasting SSS anomaly known as 'the Great Salinity Anomaly'.

World Ocean data base: S. Levitus gave a review of the ocean data archives and the data archeology product. The quantity of salinity observations in the data base has more than tripled to about 1.3 million samples since the first World Ocean Atlas was published in 1984 (WOA-84). Other modern (US Navy) data are being added, and an additional million Nansen casts are expected to be retrieved from Russian archives, many in the Southern Ocean. Historical ship of opportunity surface observations are being researched as well, and may number 1 to 2 million samples.

Ship of Opportunity data archive: T. DelCroix discussed the volunteer observing ship (VOS) surface salinity data in the Pacific that has been collected by France from New Caledonia. The highest priority has been the warm pool region. Surface bucket data were obtained from 1970-1995, and automated thermo-salinographs (TSGs) were phased in after 1991. The 1999 VOS-TSG lines are given in Figure 6. Analysis of these long term records has shown that seasonal precipitation governs SSS in the ITCZ and SPCZ with about a 3 month lag, and that ENSO variations are dominant near the equator and are primarily caused by advection. Information and/or data can be obtained from the following internet sites:

Validated TSG

Real time TSG data (rough validation):

TAO and PIRATA: A. Busalacchi reported that salinity sensors are deployed on some tropical moored buoys in the Pacific and Atlantic. Partly in support of the TRMM satellite, a N-S TAO line along 165E is equipped, and 140W and 95W lines may be added by the end of 2000 with NOAA/OGP funding. Japanese TRITON moorings will have sensors in the far western Pacific. PIRATA moorings in the Atlantic are all equipped with salinity sensors.

OGCM studies: R. Tokmakian presented results from the 1979-1997 simulation of the parallel ocean circulation model (POCM-4C) at 1/4o resolution using ECMWF momentum, heat and freshwater fluxes for surface boundary conditions. (See www.scivis.nps.navy.mil/~rtt). A number of comparisons between model and ocean data, such as sea level, were presented. POCM-4C surface salinity variability was significantly greater than the earlier POCM-4B simulation that neglected surface freshwater fluxes.

Surface salinity applications and analysis: R. Reynolds reported on recent work by C. Maes where vertical salinity profiles in the tropical Pacific were estimated with a method using EOF modes augmented with surface salinity. The results indicate that with such techniques, surface salinity can be very useful in initializing tropical ocean and climate models. Reynolds then outlined a recommended approach to computing optimal interpolation (OI) surface salinity maps combining satellite and in situ data. The approach is based on his SST analyses and would use in situ data in a bias correction loop in a similar manner to the SST processing.

Studies of ocean observation satellites in Japan: Y. Ito from the Earth Science Technology Organization (ESTO) reviewed the activities of a JAMSTEC committee to evaluate priorities for ocean remote sensing and utilization of planned missions. The committee is chaired by Prof. Y. Toba. Important parameters were identified that are currently being observed from space, and surface salinity and sea ice thickness were acknowledged as important parameters that are not available yet and are technically difficult.

Thinned Array designs: D. Levine described a next-generation 2-D imaging radiometer development task at GSFC. The sensor will consist of an array of antenna patch elements that can be configured to a number of different thinned array geometries, such as a cross, “ T” or “ Y” shape, to evaluate the respective image reconstruction capabilities.

SLFMR-2: J. miller reported that next generation models of the Scanning Low Frequency Microwave Radiometer (SLFMR) are being purchased by NRL and NOAA for coastal salinity mapping studies using light aircraft. Various coastal plume studies are planned or proposed for the next few years on the east and west coast. He also proposed a plan to mount a radiometer on a long pier at Duck, North Carolina to test salinity retrievals and wind speed and roughness effects.

High precision airborne sensor: W. Wilson described the radiometer built at JPL with a focus on high precision measurements. It includes L and S band channels using horn antennas to achieve low noise (~0.1 K) and high beam efficiency. An active radar channel can be added to aid testing geophysical models, with a focus on the roughness effects. The instrument will fly on a C-130 and initial flights are planned for June 1999 in the western Atlantic off Virginia and North Carolina, in collaboration with GSFC.

Gulf Stream field experiment: An extensive experiment was originally planned for fall 1998 and was cancelled because of NASA aircraft mechanical problems. It has been re-scheduled for late August to early September 1999. (See http://nemo.gsfc.nasa.gov/~howden/SSS.html). The objectives are to test open ocean salinity measurement capability and improve retrieval algorithms. S. Howden outlined the experimental design, shown in Figure 7. The NASA P-3 will carry the SLFMR and ESTAR istruments for salinity, a radar wave spectrometer, a wind scatterometer, a sea foam camera and IR radiometer. Surface observations will be made with the R/V Cape Henlopen including upper layer temperature, salinity, current profiles and meteorological data, as well as surface temperature and salinity from volunteer ships M/V Oleander and M/V Godafuss, and from a few WOCE surface drifters to be deployed at the time. A flight of opportunity across the Atlantic sub polar gyre on a transit to Greenland is also contemplated to test cold water retrieval accuracy.

Sea Ice experiment: An aircraft experiment was proposed with the objective of testing a new algorithm developed to discriminate between melt ponded ice and sea water using a combination of C- and L-band radiometers. Another objective is to test the retrieval of open ocean SSS at polar latitudes. The plan is to overfly the Arctic Ocean from Thule, Greenland with the Wallops P-3 in July 2000. The ice-free ocean measurements will be made during the transit flights.

Field experiment subgroup: An open discussion ensued regarding coordination among various experiments and platforms in the US and Europe. One of the outstanding issues remains the sea state effect on 1.4 GHz brightness temperatures. A field experiments subgroup was organized informally to advance these collaborations. The group includes S.Howden, C.Koblinsky, D.LeVine, J.Miller, C.Swift, W.Wilson, S.Yueh, J.Font, J.Etcheto, J.Boutin, A.Camps. It was suggested that D.Burrage from Australia, who was not present, be included also. A sea-ice team consists of D.Cavalieri, P.Gloersen and C.Swift.

Space agency programs: A general discussion took place regarding the various space agency programs and how progress toward a salinity mission fits within those programs.

ESA Earth Explorer Opportunity Missions (EEOM) is the program to which SMOS is proposed. As noted, final selections are expected to be announced in late May 1999. Phase A/B studies start late in 1999 and last about 9 months. Phase C/D/E for the second mission will take about 4 years, with anticipated launch 2004-2005.

NASA Instrument Incubator Program (IIP) focuses on new instruments or components requiring hardware development or design and analysis studies. IIP provides funds for the OSIRIS design study at JPL and the Thinned Array design study at GSFC.

NASA New Millennium Program (NMP) sponsors space flight validation of breakthrough technologies that will significantly benefit future space science and Earth science missions (http://nmp.jpl.nasa.gov/). To facilitate frequent opportunities for flight validation, the NMP expects to launch, on the average, one Deep Space and one Earth Observing flight approximately every 18 months. There are presently no salinity related projects supported by NMP.

NASA Post-2002 Request For Information (RFI) entitled "Concepts for Science and Applications Missions in the Post-2002 Era" was issued in May 1999 to initiate strategic planning for earth science activities and missions beyond the first Earth Observing System series. Based on the position papers received and subsequent reviews, NASA is now anticipating a salinity/soil moisture mission in the next decade and is planning long term budgets accordingly. (*Subsequent to the SSIWG-2 meeting, NASA initiated a review of the option of splitting this program into separate salinity and soil moisture missions). Funding for such missions would come through competitive proposals to the ESSP program (see below).

NASA Earth System Science Pathfinder (ESSP) solicits proposals every two years for satellite measurements that are outside the scope of presently approved earth science missions (see http://essp.gsfc.nasa.gov/). HYDROSTAR was proposed to ESSP last year, and the next solicitation will be in 2000. Proposals can address any earth science discipline, from which two missions and one alternate are selected based on scientific priority and technical readiness. Missions must be completed and launched within strict time and budget limits.

NASDA (Japan space agency) programs were briefly discussed by Y. Ito. Among several earth observing satellite programs, the primary mission is ADEOS II to be launched November 2000 and will carry AMSR and SeaWinds. The follow on to ADEOS II is a high priority for the JFY 2000 budget with launch in 2005 or 2006. A TRMM follow-on mission, under joint study with NASA. Other missions were mentioned. As noted above, some preliminary discussions have occurred with JPL concerning possible salinity sensors on NASDA satellites.

Strategies and Recommendations: The final phase of the workshop was a general discussion of outstanding issues and future strategies. From this, the following recommendations were put forward:

1. The SSIWG-2 endorsed the SMOS mission as a pioneering experiment for salinity and sea ice studies, and recommended final approval by ESA PB-EO. Pending this outcome, the SSIWG-2 recommended that scientific participation by U.S. and other international investigators be expanded, and that research funding from appropriate national agencies and institutions be pursued. The group also suggested specific studies of the SMOS sensor salinity retrieval error budget, particularly the unknowns in the 2-D image reconstruction error.

2. SSIWG-2 members formed an international Field Experiments Subgroup to plan and coordinate various in situ and radiometer measurement campaigns from aircraft and fixed platforms in the U.S. and Europe (see above).

3. The recommendation was made to study the scientific and technical merit of a low cost, small aperture, high precision salinity mapping mission with coarse spatial resolution devoted to large scale, open ocean, science objectives.

4. It was recommended, as a research activity, that the above concept be included among others in a set of trade-off studies to examine various mission concepts for cost and technology risk, and scientific gain, using a facility such as the Integrated Mission Design Center (IMDC) and Goddard.