Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 25
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN 3 Sea Surface Temperature INTRODUCTION Sea surface temperature (SST) has been used for centuries as a way to trace the origin of surface waters and gain greater knowledge of a location—for example, Benjamin Franklin's maps of the Gulf Stream were used to speed mail between England and North America. High-quality SST fields have been archived since the late 19th century. Surrogate observations from coral isotopic ratios, for example, allow scientists to extend records back several hundred years into the past in some parts of the world's ocean. This makes SST one of the more robust indicators for understanding Earth 's climate. The ability to monitor global and regional surface temperature has improved so that it is now possible to use SST observations as indicators of regional- to basin-scale change, as well as for forecasting stress on the natural flora and faunal assemblages. The National Polar-orbiting Operational Environmental Satellite System (NPOESS) environmental data record (EDR) requirements and goals for SST are aggressive, and their achievement would facilitate the utility of NPOESS observations for climatic research purposes, but attention to the entire observing process is necessary to make such fields generally useful. This chapter reviews the underlying scientific issues, current and future directions, observing strategies, and concomitant needs for processes such as calibration, validation, and data management to maximize the utility of satellite observations of SST. The committee 's findings in Box 3.1 address the present status of space-based measurements and data, as well as future needs in the integrated NPOESS program for research-quality SST data in the study of climate change. BASIC SCIENCE ISSUES Observation of SST in the modern era started with capturing buckets of seawater from over the sides of ships, immersing a mercury thermometer into each bucket, and recording the water temperature. Such measurements were made widely from merchant ships in the 18th and 19th centuries and are still made today. These point measurements were widely separated in time and space, and observers tended to collect seasonal assemblages of measurements and produce large-scale analyses (maps) or thermal analyses based on the assumption that the ocean was a very slowly changing medium. By the mid-20th century, however, it was clear that many parts of the ocean change quickly enough to invalidate such an assumption. In response, observers have tried to increase the density of ship platforms, develop autonomous floats, and use aircraft and spaceborne radiometric sensors to obtain more synoptic views of the surface thermal field.
OCR for page 26
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Box 3.1 Summary and Findings There are two substantive issues associated with the NPOESS environmental data record (EDR) requirements for sea surface temperature (SST) observation. First, the SST requirement has not been modified to be compatible with the results of the Climate Requirements Workshop Report (CRWP) (Jacobowitz et al., 1996). Second, calibration and validation are not a part of the NPOESS specification. The CRWP developed a modified set of ocean observation requirements for SST. It noted that the accuracy objective can be relaxed from 0.1 K1 at pixel resolution to 25 km spatial scale and a one-week temporal scale. It also commented on resolution of diel2 effects and instrumental stability. The CRWP suggested that resolution of the diel cycle would require a constellation of four polar platforms, rather than the three specified by NPOESS, i.e., 3-hour temporal sampling. The CRWP recommended that satellite-based instruments have a demonstrable stability of 0.1 K. 1In space studies of weather and climate, it is customary to denote temperature in the Kelvin, or absolute, scale. For Celsius temperature readings, 0 °C = 273.16 K. 2Diel is defined as a variation over a 24-hour period. Monitoring stability in the SST record during NPOESS missions and in the handoff periods between NPOESS and Earth Observing System (EOS) platforms requires extensive preflight characterization and post-launch validation, as outlined in the text of this chapter. The NPOESS EDRs neither specify such an activity nor suggest how sensor providers will demonstrate on-orbit stability that meets the requirements. On a more general level, there is no strategy either to integrate the lessons learned from EOS into NPOESS, or to provide inter- or intrasystem validation. Interestingly enough, the increasing density of observations has shown that each approach provides a slightly different estimate of the SST because of the peculiarities of each sampling system. This understanding has motivated development of techniques for assimilation that attempt to compensate for such peculiarities and provide surface temperature fields with known characteristics. Generally, SST analyses have been of two types: pattern discrimination and quantitative field estimates. It should be noted that satellite infrared (IR) observations of surface temperature were initiated to support meteorological applications, not to further oceanographic or climatic purposes; later such requirements started to drive accuracy, spectral placement of radiometer windows, and so on. Early satellite-based analyses could discern the edge of the Gulf Stream or the California Current because of the strong surface temperature gradient. However, the estimate of surface temperature might have been accurate only to 1.5 K or so. Eddies, boundary currents, and other mesoscale phenomena were readily identified, even though the accuracy of the estimated temperatures in and around the features was less than that obtained with in situ techniques. Moreover, the level of accuracy was not useful for following large-scale, low-frequency temperature change in the surface ocean, such as might be caused by the El Niño/Southern Oscillation (ENSO), the North Atlantic Oscillation, or other lower-frequency phenomena. During the last two decades, however, the accuracy of SST mapping from satellites has improved so that observations are routinely produced at rms (root mean square) accuracies of 0.6 K or better, thus permitting the observation and study of large-scale, low-frequency fluctuations in the ocean that might be associated with climatic variation or ecosystem change.
OCR for page 27
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Current Applications History plays a role in current applications of satellite SST analyses. Development of algorithms for producing reliable SST data sets from spaceborne infrared radiometers has been pursued by a number of investigators, agencies, and governments since the late 1960s (see reviews by Brown and Cheney, 1983; Abbott and Chelton, 1991). For example, the National Oceanic and Atmospheric Administration (NOAA) (McClain, 1981; McClain et al., 1983; Strong and McClain, 1984; McClain et al., 1985), National Aeronautics and Space Administration (NASA) (Shenk and Salomonson, 1972; Chahine, 1980; Susskind et al., 1984), and Rutherford Appleton Laboratory in the United Kingdom (RAL/UK) (Llewellyn-Jones et al., 1984) have addressed infrared radiometry using a variety of radiation transfer codes, modeled and observed vertical distributions of temperature and moisture, and actual observations. Minnett (1986, 1990) and Barton et al. (1989) summarized the present state of the art for high-quality retrievals from NOAA's Advanced Very High Resolution Radiometer (AVHRR)-class instruments. The current state of the art is limited by radiometer spectral interval placement, radiometer noise, quality of prelaunch instrument characterization, in-flight calibration quality, viewing geometry, atmospheric correction, and characterization of the quality of the SST retrieval, such as contamination by cloud and aerosols. As noted above, SST analyses are generally based on temperature estimates and identification of patterns. Thermal field estimates are used to initialize climate models as well as for diagnostic applications, for example, determining if a boundary current is in the correct location, or if the estimated seasonal temperature range in a region is correct. Similarly, molecular and thermal fluxes have a significant dependence on SST. The ocean-atmosphere CO2 exchange is strongly influenced by SST (Van Scoy et al., 1995), while explosive cyclogenesis off the U.S. East Coast has been associated with the warm waters of the Gulf Stream system (Raman and Niyogi, 1998). ENSO Studies Tracking equatorial Pacific SST anomalies has become a pastime as ENSO impacts have risen in the public consciousness, and maps of equatorial Pacific SST anomalies have become commonplace in the media. 1 From a scientific perspective, SST fields are used to initialize and validate coupled tropical models. Surface Current Pattern and Magnitude Pattern tracing of SST gradients for boundary currents and associated mesoscale structure is used daily by the U.S. Navy, NOAA, and private enterprise for maritime defense and commercial applications. These applications include optimal ship routing, fisheries catch management, antisubmarine warfare, and petroleum exploration and exploitation. Scientific applications fuse other sensor systems, such as altimetry, to obtain estimates of surface layer velocity and transport (Kelly and Gille, 1990; Gõni et al., 1997). Nutrient Estimates The correlated nature of upper-mixed-layer thermal and nutrient fields facilitates the use of SST fields as proxies for areal nutrient concentrations (Kamykowski, 1987). This augments in situ observations needed for model initialization and validation and provides researchers with an approach to determining large-scale nutrient field variations unobtainable by any other means. Coral reef assemblages are sensitive to ambient water temperature, responding to extended periods of abnormally warm water by expelling symbiotic algae. Satellite SST maps have been used to study eastern Pacific episodes of such “bleaching” (Podestá and Glynn, 1997), while NOAA has forecast large-scale impacts on the Great Barrier and Florida Keys reef systems (Montgomery and Strong, 1994). Satellite SST fields are cost-effective sentinels of these events. 1 See, for example, the information available online at <http://www.pmel.noaa.gov/toga-tao/gif/daily/sst_wind_anom_5day.gif >.
OCR for page 28
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Mesoscale Studies During the 1970s, oceanographers started using satellite SST imagery to assist experimental logistics and data analysis in the Somali, California, and Gulf Stream systems (Brown and Evans, 1980, Brown et al., 1985; Abbott and Zion, 1987). The daily or semidaily synoptic coverage makes satellite SST ideal for tracking changes in the oceanic mesoscale, weather target eddies, boundary currents, and squirts and jets. Strategies to determine boundary current loci (Cornillon and Watts, 1987; Olson, 1991; Gõni et al., 1997) have used satellite SST and altimetric systems to good effect. Eddy tracking by SST has been used in all western boundary current systems. Smaller-scale studies of estuarine variability (Framiñan and Brown, 1996) are also facilitated by satellite observations. FUTURE DIRECTIONS Polar-orbiting satellite observing systems sense ocean parameters daily, at best. Cloudiness, sensor swath gaps, and orbital mechanics limit observation frequency. Global observations have also been constrained by sensor, telemetry, and ground-processing technologies. The next decade will see multiple satellite platforms carrying similar observing technologies in Sun-synchronous orbits at differing local times of day (Sun time). Sampling from this combined system should improve retrievals from cloudiness-limited scenes and permit the resolution of diel variations in clear areas. However, assimilation of observations at differing local Sun times necessitates improved understanding of diel cycles in the parameters of interest. For example, current observations from NOAA AVHRR platforms show significant changes in satellite SST fields as the crossing time drifts later in the day, as one might expect. Geosynchronous platforms now provide subdiel temporally resolved SST fields. Innovative compositing and assimilation approaches can remove much of the data contamination caused by clouds, but here, too, the enhanced sensitivity of the instruments now resolves diel effects. Similarly, use of passive microwave techniques to discern SST variation in predominantly cloudy areas will improve overall understanding of tropical and polar air-sea exchange processes. Coastal Processes High-frequency variability in time and space is typical of nearer-shore processes. Tidal effects, subinertial oscillations, coastal trapped waves, and coupled land-sea effects all contribute to a rich high-frequency spectrum. Significant improvements in temporal and spatial sampling are needed to resolve many coastal processes adequately. Temporally, this indicates a need for geosynchronous Earth orbit (GEO) sensors, while spatially, low Earth orbit (LEO) sensors have the advantage. Observation of coastal processes will probably demand linking of GEO and LEO observations into coherent depictions of phenomena. Such higher-frequency fields would complement systems currently used by linking the higher-frequency variance to the larger scales. Skin Temperature Much progress has been made in the past several years on in situ sensing of skin, or surface radiation, temperature. New technologies applied to this problem (Smith et al., 1996) have provided high accuracy, improved coverage, and diel-resolved observing sets. One can expect that in the future most infrared instrumentation will be validated by comparison with surface radiation temperature, rather than bulk temperature. Bulk-skin temperature differences tell much about the surface heat flux, a topic of major interest to regional and global climate studies. Upper-Ocean Mixed-Layer Temperature A combination of satellite-sensed SST with surface wind observations and in situ profile observations meets minimal requirements for initializing and validating upper-ocean mixed-layer models. Such model systems provide a dynamic framework for the assimilation of satellite and in situ observations and the forecasting of mixed-layer evolution. This will probably be one of the most important applications for satellite SST observations
OCR for page 29
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN in the early years of the 21st century. It is also an ideal framework for implementing regional and basin-scale observing systems. OBSERVING STRATEGY Requirements for SST observations to meet climate, national defense, and science needs have been articulated by a number of groups (WOCE, 1985; IPO NPOESS, 1996; NOAA, 1997). NOAA identified a set of SST observation objectives (NPOESS, 1995) as part of its requirements development process for the converged NPOESS. These objectives push the state of the art in accuracy and temporal and spatial resolution (0.1 K, 0.5 km, 3-hour revisit). They are excellent goals, but achieving them would necessitate a fully operational four-satellite constellation with a cloudless Earth, state-of-the-art radiometers, and a robust data assimilation and validation system. Clearly, such an observing system would fulfill many of the needs articulated in the previous section. However, maintaining such a system would require a robust prelaunch characterization program, postlaunch validation, appropriate data assimilation procedures, and continuing quality assurance of system performance. The major challenge in the short term will be to develop a systematic approach to the use of satellite IR and microwave radiances and derived temperatures from disparate platforms. International collaboration on such systems is recent,2 and the current and nearer-term sensor systems have not been well coordinated. For example, one may find several platforms in orbits with similar overpass times, or systems with a similar SST product but different observing channels, or systems with quite different atmospheric correction approaches. Any one of these issues poses a challenge; we will probably see all of them during the next decade. NASA's and NOAA's Plans During the next 5 years SST observations will be produced by NOAA and the Navy from NOAA's AVHRR system, by NASA from the EOS/MODIS, by the National Space Development Agency (NASDA–Japan) from ADEOS-2/GLI, and by the European Space Agency (ESA) from ENVISAT/Advanced Along Track Scanning Radiometer (AATSR). This is a minimum estimate; other investigators and nations may also develop sensors and analysis systems. Table 3.1 lists some characteristics of known systems that should be in orbit in the next 5 years. Table 3.1 covers passive IR systems; there are also several passive microwave systems being put in place whose composite performance characteristics approach those of the IR systems. TABLE 3.1 Comparison of Low Earth Orbit Infrared SST Observing Systems (1999-2004) System/Attribute NOAA AVHRR EOS MODIS ADEOS GLI ENVISAT AATSR Local Sun time 0730/1430 1030 1030 1000 Spatial resolution (km) 1 1 1 1 Approach NLSST NLSST MCSST Dual Path/MCSST NEΔT (K) 0.12 0.05 0.10 0.025 Validation approach Bulk Skin Bulk Skin Expected SST accuracy (K) 0.55 0.40 0.50 0.30 NOTE: Acronyms are defined in Appendix B. 2 Committee on Earth Observation Satellites (CEOS). 1997. Towards an Integrated Global Observing Strategy, Strategic Implementation Team Scoping Paper. Available online at <http://www.eos.co.uk/ceos-calval/igos/sitscope.htm>. Committee on Earth Observation Satellites (CEOS). 1998. Final Report of the CEOS Analysis Group 1996/1997). Available online at <http://www.smithsys.co.uk/IGOS/Agfinalreport.html>.
OCR for page 30
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Variations in observing times and validation approaches suggest that the first aspect of an observing strategy must be to understand the diel SST (and skin-bulk temperature) cycle and use such understanding to analyze satellite SST estimates. Second, prior to launch all these systems must be cross-referenced to common standards and characterized by similar approaches. The differing validation approaches suggest that a multinational group should coordinate the joint validation of these missions. Since development of atmospheric correction approaches uses radiative transfer calculations, a common suite of tools should be developed and provided to interested platform operators. Mean cloudiness at midlatitudes varies with season and hemisphere; overall cloud-contaminated pixels can exceed 60 percent over much of the year (R. Evans, Rosenstiel School, University of Miami, personal communication). Strategies to fill in data gaps caused by clouds or other phenomena span the range from linear interpolation to Laplacian relaxation techniques (i.e., techniques involving a differential operator that identifies a particular type of mathematical interpolation), to objective analysis and dynamical interpolation, to blending of IR and microwave-based observations. There are differing approaches to objective analysis and dynamical interpolation using estimated satellite SSTs, brightness temperatures, or radiances. Similarly, we have no validated blending approach for IR and microwave observations. There is no standard approach to solving the “gappiness” problem, which may be caused by temporal sampling characteristics of a system, cloudiness, or a variety of other factors. SST fields are usually tailored for their intended use. For example, weather forecasting requires the best estimate of the current state of the SST fields with a narrow time window, while climate applications might use weekly to monthly fields. An ocean logistical or experimental application might rely on a collection of individual satellite passes. There is no single set of requirements for temporal and spatial resolution, level of accuracy, and acceptable degree of error. Nationally, four operational SST products are obtained using satellite IR observations: an analysis from the Fleet Meteorology and Oceanography Center (FMOC) (Clancy et al., 1992; Cummings et al., 1997); a Naval Oceanographic Office (NAVO) analysis (May et al., 1998); a NOAA Climate Analysis Center (CAC) objective analysis (Reynolds and Smith, 1994); and a NOAA National Environmental Satellite, Data, and Information Service (NESDIS) MCSST analysis (McClain et al., 1983). These products are contrasted in Table 3.2. Most importantly, all these analyses are based on different approaches, use somewhat different data, and are validated against bulk measurements. Each product is produced for particular applications, for example, fleet operations or climate studies, and there are clear differences among them (R. Evans, University of Miami, personal communication). Currently none of these products uses passive microwave radiances for SST estimation, but preliminary results from the Tropical Rainfall Measuring Mission suggest that a composite infrared-microwave approach may offer improvements for tropical SST estimation (F. Wentz, Remote Sensing Systems, personal communication). Current SST products depend strongly on the characteristics of the sensor and the atmospheric correction approach employed. Most atmospheric correction approaches are combinations of brightness temperatures, either linear or nonlinear, with weighting coefficients. Coefficients are determined by regression analysis of the foregoing features versus surface observations or from forward radiative transfer calculations. None of the standard approaches adequately addresses aerosol effects, although the slant path approach used by the Along Track Scanning Radiometer (ATSR) is superior to other techniques. Aerosols remain a problem due to a lack of appropriate characterization and sufficient channels to determine their properties in the satellite data. Upcoming instruments such as MODIS include channels to sense aerosol radiance in the near-IR, so it is possible that substantial progress will be made on this problem over the next few years (see Chapter 7 of this report for a detailed discussion of aerosols). TABLE 3.2 Comparison of National Operational SST Analysis Products Product/Attribute NAVY FMOC NAVY NAVO NOAA CAC NOAA NESDIS Approach Model (OTIS) Range checks Optimal interpolation Range checked Spatial resolution (km) 111 (20 regional) 9 111 50 (14 near United States) Temporal resolution Daily Daily Weekly Daily Contents Satellite/ship/drifter Satellite Satellite/ship/drifter Satellite Validation approach Bulk Bulk Bulk Bulk NOTE: Acronyms are defined in Appendix B.
OCR for page 31
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN IPO/NPOESS Requirements The EDR requirements for SST observations developed as part of the Integrated Program Office (IPO)/ NPOESS Integrated Operational Requirements Document, first version (IORD-1) (IPO NPOESS, 1996) (Table 3.3) are contrasted with NOAA/AVHRR and EOS/MODIS in Table 3.4. From a review of Table 3.4, it is apparent that many of the NPOESS SST requirements are similar to the capabilities of AVHRR and the objectives of MODIS. Some are clear improvements, such as the measurement accuracy and precision objectives. IORD-1 requirements suggest a measurement capability marginally more accurate than that of AVHRR, but with improved pixel navigation, measurement precision, and long-term stability. On the other hand, attainment of the NPOESS objectives would imply an instrument with measurement accuracy and precision exceeding those of the Earth Resources Satellite (ERS)-1/2 ATSR instrument (ATSR accuracy, approximately 0.3 K; precision, 0.1 K, Mutlow et al., 1994; Harris et al., 1995; Mason et al., 1996). Some reviewers of the IORD-1 requirements believe that such levels of accuracy may not be attainable from a satellite-based SST observing system (Jacobowitz et al., 1996). TABLE 3.3 NPOESS Environmental Data Record Requirements for Sea Surface Temperature (IORD-1) Systems Capability Threshold Objective Horizontal Resolution Global, nadir 3 km 1 km Global, worst case 4 km TBDa Regional, nadir 1 km 0.25 km Regional, worst case 1.3 km TBD Horizontal Reporting Interval Horizontal Compliance Measurement Range 271 to 313 K 271 to 313 K Measurement Uncertainty (rms) 0.5 K 0.1 K Measurement Accuracy 0.2 K 0.1 K Measurement Precision TBD 0.1 K Mapping Uncertainty Global, nadir 1 km 0.5 km Global, worst case 3 km TBD Regional, nadir 1 km 0.1 km Regional, worst case 3 km TBD Maximum Local Average 3 h TBD Maximum Local Refresh 6 h TBD SOURCE: Extracted from IPO NPOESS (1996). The updated IORD and otherdocumentation related to the NPOESS program are available onlineat <http://npoesslib.ipo.noaa.gov/ElectLib.htm>. aTBD, to be determined.
OCR for page 32
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN TABLE 3.4. Comparison of NPOESS Sea Surface Temperature Requirements (IORD-1) with AVHRR Capabilities and MODIS Systems Objectives Systems Capabilities NPOESS Thresholds NPOESS Objectives AVHRR Actual MODIS Objectives Horizontal Resolution Global, nadir 3 km 1 km 4 km 4 km Global, worst case 4 km TBDa 4 km 4 km Regional, nadir 1 km 0.25 km 1 km 1 km Regional, worst case 1.3 km TBD 1 km 1 km Mapping Uncertainty Global, nadir 1 km 0.5 km 2 km 0.1 km Global, worst case 3 km TBD 6 km 0.1 km Regional, nadir 1 km 0.1 km 2 km 0.1 km Measurement Range 271 to 313 K 271 to 313 K 271 to 313 K 271 to 313 K Measurement Uncertainty (rms) 0.5 K 0.1 K 0.7 K 0.4 K Measurement Accuracy 0.2 K 0.1 K 0.2 K 0.1 K Maximum Local Average 6 h 3 h 6 h 12 h aTBD, to be determined. CALIBRATION AND VALIDATION Validation of SST measurements is required over the lifetime of the satellite mission. The validating instruments must be deployed in situations that encompass the entire range of surface temperatures and atmospheric variability. Since no single approach provides a perfect validation measurement, a selection of techniques and instruments is required for an adequate validation data set. The approach includes validation of (1) top-of-the-atmosphere radiances, (2) surface radiances, and (3) surface temperatures. There are three possible methods of validating top-of-the atmosphere radiances: Comparison with other satellite measurements, Comparison with aircraft radiometers flying at lower elevations than the satellite, and Use of radiative transfer modeling to simulate satellite measurements. Comparison with Other Satellite Measurements Surface radiance measurements are usually validated using calibrated spectroradiometers, such as the Marine-Atmospheric Emitted Radiance Interferometer (M-AERI) (Smith et al., 1996), or broadband infrared thermometers. These instruments can be mounted on low-flying aircraft (Saunders and Minnett, 1990; Rudman et al., 1994; Smith et al., 1994), ships (Schluessel et al., 1987; Smith et al., 1996), or fixed platforms. Comparisons of top-of-the-atmosphere measurements obtained with satellite-borne infrared radiometers on different spacecraft have the advantage over the airborne or surface-based instruments described above of comparing the results of measurements by similar instruments. The problems with this approach are (1) the possible changes in the top-of-the-atmosphere radiation field between the two satellite overpasses (resulting from changes in the surface temperature or in the intervening atmosphere), (2) differences in the viewing geometry of the two satellites, (3) differences in the spectral responses of the different satellite instrument channels, and (4) possible less-than-required accuracy level or noise characteristics of the validating instrument. Another possible problem is the potential for undetected in-flight degradation of the validating radiometer. If systematic discrepancies are found, it may not be apparent which satellite sensor is at fault.
OCR for page 33
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Comparison with Aircraft Radiometers A significant advantage of using aircraft radiometers is that data can be taken simultaneously with satellite measurements. However, because of the difference between spacecraft and aircraft speeds, few truly coincident measurements can be made, although within, for example, a 30-minute window of the satellite overpass a large number of validation measurements could be obtained, depending on the time interval selected for data acquisition (Minnett, 1990). Also, aircraft radiometers can in principle be arranged to match the satellite viewing geometry and can be scheduled (again, in principle) to avoid conditions that would make data interpretation difficult (e.g., broken cloud fields). Disadvantages of this technique include the effects of the atmosphere above the aircraft, which can be accounted for by modeling, using an assumed (or measured) temperature and humidity profile, and the accuracy of the aircraft instruments. Candidate aircraft instruments for top-of-the-atmosphere radiance validation of the channels used in SST determination include the MODIS Airborne Simulator (MAS) (King and Herring, 1992) and the High-resolution Interferometer Sounder (HIS) (Bradshaw and Fuelberg, 1993). These instruments are flown typically on the NASA ER-2 research aircraft at a height of approximately 20 km, and under these conditions the spatial resolution is 50 m (MAS) and 2 km (HIS). The noise levels of these instruments are not as low as those for the MODIS infrared channels. For the MAS, the NEΔT (or noise-equivalent delta radiance) is approximately 0.3 K for a target at about 290 K for the 3.7 to 4.0 µm channels and 0.1 to 0.2 K for the 11 to 12 µm channels. However, these levels could be greatly improved (by a factor of 20 if the noise were truly random) by averaging the data down to a typical spatial resolution of about 1 km2. The noise levels in the HIS spectra in the 800 to 1050 cm interval are typically 0.2 to 0.45 mW m−2st−1cm, and these result in an uncertainty of about 0.15 K in the skin SST retrieved from the HIS spectra (Nalli, 1995). Use of Radiative Transfer Models The use of numerical models of radiative transfer through the atmosphere to simulate satellite measurements requires high-quality measurements of the relevant atmospheric properties (temperature and humidity profiles, aerosol characteristics) and emitted radiance at the surface, taken at the time of the satellite overpass. The advantage of this approach is that a large database of measurements can be generated over an extended period of time, representing a large range of atmospheric conditions, surface temperatures, and viewing geometries for a relatively modest outlay. The disadvantages are uncertainties about the accuracies of the atmospheric profiles, generally derived from routine radiosonde measurements (Schmidlin, 1988), and shortcomings in the parameterization of incompletely understood physical processes in the radiative transfer model, such as the anomalous continuum absorption and emission by water vapor and the effects of tropospheric and stratospheric aerosols. The long-term measurement of surface-emitted radiance, or the channel brightness temperatures, at the surface serves to monitor the behavior of the atmospheric correction algorithms and the performance of spaceborne radiometers. The radiance of surface-based measurements has two sources: one is of emitted radiance at the sea surface; the other is the reflected component of the downwelling radiance originating in the atmosphere. The space-based measurement is of this combination, after attenuation by atmospheric absorption and scattering, plus the radiance emitted or scattered by the atmosphere into the sensor field of view. This validation measurement is therefore less direct than a comparison of top-of-the-atmosphere data. Surface measurements can be related to space-based sensor measurements by using a radiative transfer model to estimate the atmospheric attenuation and upwelling and scattered radiation, or by converting the surface measurement to a temperature and comparing it with the surface temperature derived from the space-based measurements. In either case, successful interpretation of these data requires a good description of the atmospheric as well as the surface properties (skin SST, surface emissivity, and wind speed). In the case where surface measurements are converted to temperature, a measurement of the downwelling radiation that is required for derivation of the temperature from the surface measurements can be achieved by pointing the surface radiometer at the sky. Suitable instruments include the M-AERI for use at sea, the AERI, and broadband infrared thermometers (Smith et al., 1996). The M-AERIs have internal blackbody calibration targets
OCR for page 34
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN and so provide a calibrated measurement. They measure the spectrum of infrared radiation in the range of 3.3 to 18 µm with a spectral resolution of about 0.5 cm−1. These spectra can be compared to the spaceborne measurements by multiplying them by the appropriate normalized channel spectral response functions. The M-AERI spectra can also be analyzed to derive surface temperature and emissivity, and, using spectra of sky radiation, the temperature and humidity structure of the atmosphere. Broadband infrared radiation thermometers have an advantage over M-AERIs in that they are smaller and inexpensive. They usually do not have the required accuracy of 0.1 K and have only a simple internal calibration capability, if any. However, recent experience with some types indicates that they may produce useful observations and may be suitable for deployment in larger numbers on platforms of opportunity, especially if combined with a reliable external calibration assembly. In principle, surface temperature thermometers can be deployed in numbers sufficient to provide adequate monitoring of satellite radiometer performance. However, they have a big disadvantage in that their measurement may be decoupled from remote satellite measurement by near-surface temperature gradients. For determining SST, the in situ thermometer is immersed in the water, frequently at depths of 0 to 1 m, and may register a measurement that may differ from the temperature of the radiating skin of the ocean by more than 1.0 K. These gradients are caused by heat exchange between the ocean and the atmosphere (the skin effect; see, e.g., Robinson et al., 1984; Schluessel et al., 1990) or by diurnal heating in conditions of low wind speed and therefore reduced surface mixing (see, e.g., Stramma et al., 1986). Despite this problem, in situ thermometers have been used extensively to validate satellite-measured SSTs (Strong and McClain, 1984; Llewellyn-Jones et al., 1984; Podestá et al., 1997). DATA MANAGEMENT Access to satellite data has been the Achilles' heel of most satellite programs. It had been argued that digital data were difficult to provide, media standards were varied, and end users did not have adequate processing capability to deal with data products. The penetration of high-performance computing into home and office settings, the explosion of Internet connectivity, and the adoption of standards for data products (Hierarchical Data Format and network Common Data Format, for example) have helped to mitigate these issues and have put pressure on data providers to build broad access to satellite products into system design. Recent and future NASA missions such as the Upper Atmosphere Research Satellite (UARS), Sea Viewing Wide Field of View Sensor (SeaWiFS), and EOS all have user-driven data system designs. In contrast, legacy satellite systems such as NOAA AVHRR and ERS ATSR have evolved slowly and lag behind the state of the art for rapid and easy data accessibility. However, there has been a strong commitment to long-term archives from most satellite platform operators (NOAA, NASDA, ESA, and NASA). More recent systems, such as ADEOS and EOS, have built-in user-accessible browsing capabilities and data delivery via the Internet. Currently, access to operational data from the NOAA system is best provided by academic receiving and processing sites (there are a number of Internet-accessible sites in the United States). Academic sites pioneered near-real digital data access for infrared SST observations. Sites at Scripps Institution of Oceanography, the University of Miami–Rosenstiel School, and the University of Dundee have made IR SST products and data available since the late 1970s. A major problem with SST observations and analyzed fields has been documenting system performance over the long term. NOAA now provides documentation for clock drift, sensor faults, algorithm modifications and updates, and processing system changes in a generally accessible form. NASA and NOAA have teamed up in the Pathfinder project to reprocess 15 years' worth of NOAA AVHRR SST observations with state-of-the-art algorithms to provide climate modelers and analysts a consistent set of SST fields. The magnitude of the Pathfinder effort demonstrates the need to design reprocessing capabilities for operational satellite systems. EOS has built reprocessing and maintenance of observations for validation into the Earth Observing System Data and Information System (EOSDIS).
OCR for page 35
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN EVOLUTION STRATEGY When taken as the skin temperature, SST has the desirable property of being related directly to the radiance emitted from the sea surface; that is, it is a geophysical parameter. It can be sensed in both the infrared and microwave parts of the spectrum; however, the radiation temperature is dependent on the wavelength. This dependence offers opportunities to sense temperatures from several effective depths in the near-surface ocean, as well as with differing atmospheric transmissivities. Trade-offs between the infrared and microwave approaches include the accuracy and precision of the temperature estimate, needed spatial resolution, and need for all-weather observations. Historically, IR systems have had lower noise figures and better accuracy than microwave systems. However, improved microwave receivers and new reflector designs suggest that microwave-based approaches are becoming competitive with IR approaches and have the benefit of being all-weather, i.e., not limited by cloudiness. To achieve the same spatial resolution as IR sensors, however, requires very large microwave antennae. IR approaches have also evolved. The ERS ATSR conical scan and dual blackbody calibration design has facilitated the sensing of SSTs with rms errors of 0.3 K and better. Use of slant and nadir look angles is a significant advance in IR determination of SST. It also appears to produce better aerosol corrections, resulting in an improved SST field. These evolving approaches are limited by current knowledge of the surface processes and their diel variations, and the sparseness of skin validation observations. Current SST measurement accuracy is significantly affected by the quality and types of validation measurements (see discussion above). Diel variation in skin and bulk temperatures and related fluxes is not well understood; current research should resolve some of these issues. REFERENCES Abbott, M.R., and D.B. Chelton. 1991. Advances in passive remote sensing of the ocean. U.S. National Report to IUGG. Rev. Geophys. Supplement: 571-583. Abbott, M.R., and P.M. Zion. 1987. Spatial and temporal variability of phytoplankton pigment off northern California during Coastal Ocean Dynamics Experiment 1. J. Geophys. Res. 92(C2): 1745-1755. Barton, I.J., A.M. Zavody, D.M. O'Brien, D.R. Cutten, R.W. Saunders, and D.T. Llewellyn-Jones. 1989. Theoretical algorithms for satellite-derived sea-surface temperatures . J. Geophys. Res. 94: 3365-3375. Bradshaw, J.T., and H.E. Fuelberg. 1993. An evaluation of HIS interferometer soundings and their use in mesoscale analysis. J. Appl. Meteorol. 32: 522-538. Brown, O.B., and R.E. Cheney. 1983. Advances in satellite oceanography. Rev. Geophys. Space Phys. 21(5): 1216-1230. Brown, O.B., and R.H. Evans. 1980. Interannual variability of the Somali Current system during the summer monsoon. (Abstract) EOS 61(32): 574. Brown, O.B., J. Brown, and R. Evans. 1985. Calibration of AVHRR infrared observations. J. Geophys. Res. 90(C6): 11667-11677. Chahine, M.T. 1980. Infrared remote sensing of sea surface temperature. In Remote Sensing of Atmospheres and Oceans, A. Deepak (ed.). New York: Academic Press. Clancy, R.M., J. Harding, K. Pollak, and P. May. 1992. Quantification of improvements in an operational global-scale ocean thermal analysis system. J. Atmos. Ocean Technol. 9: 55-66. Cornillon, P., and D.R. Watts. 1987. Satellite thermal infrared and inverted echo sounder determinations of the Gulf Stream northern edge. J. Atmos. Ocean Technol. 4: 712-723. Cummings, J.A., C. Szczechowski, and M. Carnes. 1997. Global and regional ocean thermal analysis systems. Mar. Technol. Sci. J. 31(1): 63-75. Framiñan, M.B., and O.B. Brown. 1996. Study of the Río de la Plata turbidity front, Part I: spatial and temporal distribution . Cont. Shelf Res. 16(10): 1259-1282. Gõni, G., S. Garzoli, A. Roubicek, D. Olson, and O. Brown. 1997. Agulhas ring dynamics from TOPEX/POSEIDON satellite altimeter data . J. Mar. Res. 55: 861-883. Harris, A.R., M.A. Saunders, J.S. Foot, K.F. Smith, and C.T. Mutlow. 1995. Improved sea surface temperature measurements from space. Geophys. Res. Lett. 22: 2159-2162. Integrated Program Office (IPO), National Polar-orbiting Operational Environmental Satellite System (NPOESS). 1996. Integrated Operational Requirements Document (IORD) I. Joint Agency Requirements Group Administrators. 61 pp. + figures. Jacobowitz, H., B. Chertock, S. Mango, W. Murray, and G. Ohring. 1996. National Polar-orbiting Operational Environmental Satellite System (NPOESS) Climate Requirements Workshop Report. College Park, Md.: University of Maryland. Kamykowski, D. 1987. A preliminary biophysical model of the relationship between temperature and plant nutrients in the upper ocean. Deep-Sea Res. Oceanogr. Abstr. 34: 1067-1079. Kelly, K.A., and S.T. Gille. 1990. Gulf Stream surface transport and statistics at 69°W from the Geosat altimeter. J. Geophys. Res. 95(C3): 3149-3161.
OCR for page 36
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN King, M., and D. Herring. 1992. The MODIS Airborne Simulator (MAS). The Earth Observer ( November-December). Llewellyn-Jones, D.T., P.J. Minnett, R.W. Saunders, and A.M. Zavody. 1984. Satellite multichannel infrared measurements of sea-surface temperature of the N.E. Atlantic Ocean using AVHRR/2. Q.J.R. Meteorol. Soc. 110: 613-631. Mason, I.M., P.H. Sheather, J.A. Bowles, and G. Davies. 1996. Black body calibration sources of high accuracy for a space-borne infra-red instrument, the Along Track Scanning Radiometer. Appl. Opt. 35: 629-639. May, D., M. Parmeter, D. Olszewski, and B. McKenzie. 1998. Operational processing of satellite SST retrievals at the Naval Oceanographic Office. Bull. Am. Meteorol. Soc. 79(3): 397-407. McClain, E.P. 1981. Multiple atmospheric-window techniques for satellite derived sea surface temperatures. Oceanography from Space. Volume 13, J.F.R. Gower (ed.). New York: Plenum. McClain, E.P., W.G. Pichel, C.C. Walton, Z. Ahmed, and J. Sutton. 1983. Multi-channel improvements to satellite derived global sea surface temperatures. Proc. XXIV COSPAR. Adv. Space Res. 2(6): 43-47. McClain, E.P., W.G. Pichel, and C.C. Walton. 1985. Comparative performance of AVHRR-based multichannel sea surface temperatures . J. Geophys. Res. 90(C6): 11587-11601. Minnett, P.J. 1986. A numerical study of the effects of anomalous North Atlantic atmospheric conditions on the infrared measurement of sea surface temperature from space. J. Geophys. Res. 91(C7): 8509-8521. Minnett, P.J. 1990. The regional optimization of infrared measurements of sea-surface temperature from space. J. Geophys. Res. 95: 13497-13510. Montgomery, R.S., and A.E. Strong. 1994. Coral bleaching threatens oceans, life. (Abstract) EOS 75: 145-147. Mutlow, C.T., A.M. Zavody, I.J. Barton, and D.T. Llewellyn-Jones. 1994. Sea-surface temperature-measurements by the Along-Track Scanning Radiometer on the ERS-1 Satellite—early results. J. Geophys. Res. 99(C11): 22575-22588. Nalli, N.R. 1995. Sea surface skin temperature retrieval using the High-Resolution Interferometer Sounder. MS Thesis. University of Wisconsin-Madison. National Oceanic and Atmospheric Administration (NOAA). 1997. Climate Measurement Requirements for the National Polar-orbiting Operational Environmental Satellite System (NPOESS): Workshop Report, Herbert Jacobowitz (ed.), Office of Research Applications, NESDIS-NOAA, February. 77 pp. Olson, D.B. 1991. Rings in the ocean. Annu. Rev. Earth Planet. Sci. 19: 283-311. Podestá, G.P., and P.W. Glynn. 1997. Sea surface temperature variability in Panamá and Galápagos: extreme temperatures causing coral bleaching. J. Geophys. Res. 102: 15749-15759. Raman, S., and D.S. Niyogi. 1998. Mesoscale analysis of a Carolina coastal front. Boundary-Layer Meteorol. 86: 125-145. Reynolds, R., and T. Smith. 1994. Improved global sea surface temperature analysis using optimum interpolation . J. Climate 7: 929-948. Robinson, I.S., N.C. Wells, and H. Charnock. 1984. The sea surface thermal boundary layer and its relevance to the measurement of sea surface temperature by airborne and space borne radiometers . Int. J. Remote Sensing 5: 19-46. Rudman, S., R.W. Saunders, C.J. Kilsby, and P.J. Minnett. 1994. Water vapour continuum absorption in mid-latitudes: aircraft measurements and model comparisons. Q.J.R. Meteorol. Soc. 120: 795-807. Saunders, R.W., and P.J. Minnett. 1990. The measurement of sea surface temperature from the C-130. MRF Internal Note No. 52. Meteorological Research Flight, Royal Aerospace Establishment, Farnborough , Hampshire, U.K. 16 pp. Schluessel, P., H-Y. Shin, W.J. Emery, and H. Grassl. 1987. Comparison of satellite-derived sea-surface temperature with in-situ skin measurements. J. Geophys. Res. 92: 2859-2874. Schluessel, P., W.J. Emery, H. Grassl, and T. Mammen. 1990. On the bulk-skin temperature difference and its impact on satellite remote sensing of sea surface temperatures. J. Geophys. Res. 95: 13341-13356. Schmidlin, F.J. 1988. WMO International radiosonde intercomparison phase II, 1985. Instrument and Observing Methods, Report No. 29. WMO/TD 312. Geneva: World Meteorological Organization. 113 pp. Shenk, W.E., and V.V. Salomonson. 1972. A multispectral technique to determine sea surface temperature using NIMBUS II data. J. Phys. Oceanogr. 2: 157-167. Smith, A.H., R.W. Saunders, and A.M. Zavody. 1994. The validation of ATSR using aircraft radiometer data over the Tropical Atlantic. J. Atmos. Ocean Technol. 11: 789-800. Smith, W.L., R.O. Knuteson, H.E. Revercombe, W. Feltz, H.B. Howell, W.P. Menzel, N.R. Nalli, O.B. Brown, J. Brown, P.J. Minnett, and W. McKeown. 1996. Observations of the infrared radiative properties of the ocean—implications for the measurement of sea-surface temperature via satellite remote sensing. Bull. Am. Meteorol. Soc. 77(1): 41-51. Stramma, L., P. Cornillon, R.A. Weller, J.F. Price, and M.G. Briscoe. 1986. Large diurnal sea surface temperature variability: satellite and in situ measurements. J. Phys. Oceanogr. 16: 827-837. Strong, A.E., and E.P. McClain. 1984. Improved ocean surface temperature from space—comparisons with drifting buoys. Bull. Am. Meteorol. Soc. 65(2): 138-142. Susskind, J., J. Rosenfield, D. Reuter, and M.T. Chahine. 1984. Remote sensing of weather and climate parameters from HIRS2/MSU on TIROS-N. J. Geophys. Res. 89(C6): 4677-4697. Van Scoy, K., K.P. Morris, J.E. Robertson, and A.J. Watson. 1995. Thermal-skin effect and the air-sea flux of carbon dioxide—a seasonal high-resolution estimate. Global Biogeochemical Cycles 9: 253-262. World Ocean Circulation Experiment (WOCE). 1985. WOCE Global Air-Sea Interaction Fields, U.S. WOCE Technical Report No. 1, W.G. Largee (ed.). 36 pp.
Representative terms from entire chapter: