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 57
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN 5 Ocean Color INTRODUCTION Marine phytoplankton account for nearly half of the world's total primary productivity. They dominate the structure of the upper trophic levels of the food web and play a critical role in the cycling of biogeochemical properties. Biomass turnover rates for plankton ecosystems are 100 times faster than those for terrestrial systems, leading to a close relationship between upper-ocean ecology and physical forcing. For example, coupled ocean and atmosphere models show that changes in the phytoplankton community structure and the resulting elemental interactions can drastically affect the rate of carbon dioxide (CO2) increase in the atmosphere. Moreover, the large space and time scales associated with ocean biogeochemistry and circulation can be disrupted on far shorter time scales, such as those of El Niñ o/Southern Oscillation (ENSO) events. This coupling of large and small scales leads to the fundamental sampling requirement of global-scale, long time series (decades) at moderate time and spatial scales (days and kilometers). BASIC SCIENCE ISSUES In addition to ocean biogeochemistry and its linkage with climate, the productivity of the ocean is the fundamental limit on the number of fish that can be harvested. Because of the near-collapse of many of the world's fisheries, there is renewed interest in understanding the complex linkage between primary productivity and fisheries production. As with ocean biogeochemistry, the scales are both large and small. However, because coastal zones are often used as nurseries, there is particular interest in the smaller scales associated with the nearshore environments. Another objective focuses on the coastal zone itself as a buffer between human activities on the land and the deep ocean. In addition to physical and biological coupling, there is particular interest in the impacts of coastal pollution, either through direct discharge into the sea or by riverine inputs. Harmful algal blooms (e.g., red tides) can have drastic economic and human health impacts. Given the economic and recreational importance of the coastal zone, there is increasing demand for high-resolution coastal ocean monitoring and prediction. The science issues can thus be summarized as three broad objectives:
OCR for page 58
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Predict the ocean's biogeochemical response to and its influence on climate change; Predict the variability in the structure of the phytoplankton community and its links with higher trophic levels as well as ocean biogeochemistry; and Develop the scientific basis necessary to manage the sustainable resources of the coastal marine ecosystem effectively. Current Approaches in Remote Sensing To address these three objectives, the remote sensing community has developed an evolving suite of satellite sensors to collect measurements of ocean color in the visible portion of the electromagnetic spectrum. The basic measurement of phytoplankton biomass relies on the strong absorption of visible light by chlorophyll (the primary light-harvesting pigment), which peaks near 443 nm (Gordon and Morel, 1983; Kirk, 1994). This absorption characteristic is a robust feature across a broad range of productivity levels in the world's oceans. Atmospheric Correction The challenge for spaceborne sensors is that 80 to 90 percent of a satellite-sensed signal originates in the atmosphere (Gordon and Morel, 1983). Much of this atmospheric signal is Rayleigh (or molecular) scattering, primarily from stratospheric ozone. After accounting for satellite and solar geometry for a particular scene, it is relatively straightforward to make corrections based on knowledge (or estimates) of extraterrestrial solar radiation, ozone concentration, and atmospheric pressure. However, aerosol scattering, primarily from hydrophilic particles in the marine boundary layer, is a much more complex process. It varies strongly as a function of time and location. Because it is not yet possible to make direct measurements of these aerosols and their contribution to atmospheric optical properties, the remote sensing community has relied on an indirect approach. Because the ocean is largely “black” in the red and near-infrared portion of the spectrum, it can be assumed that any radiance measured at these wavelengths originated in the atmosphere and was not backscattered out of the ocean. Relying on ratios between the remaining wavelengths, the spectral dependence of aerosol scattering can be propagated down to the short wavelengths in the blue portion of the visible light spectrum. The atmospheric correction schemes have matured considerably over the past 20 years since the launch of the first ocean color sensor, the Coastal Zone Color Scanner (CZCS), on Nimbus-7. The original atmospheric schemes relied on locating a “clear-water pixel” where chlorophyll concentrations were low and the spectrum of water-leaving radiance therefore well known. The atmospheric correction for this clear-water portion of the image was then extrapolated across the entire scene. Obviously there are serious limitations to this approach; for example, there may not be a low-chlorophyll pixel in the image, or atmospheric properties may change significantly within an image that covers nearly a million square kilometers. The next step was to enable pixel-by-pixel correction, thus eliminating the need for an imagewide correction and a low-chlorophyll region. As analysis and processing of CZCS data continued, it became apparent that the atmospheric correction schemes had to accommodate multiple scattering by molecules. The first-generation algorithms assumed that a photon would be scattered only once. However, at low Sun angles or at the edge of the sensor swath, the probability of multiple Rayleigh scattering increased substantially. Moreover, post-CZCS sensors— e.g., the Sea-viewing Wide Field of View Sensor (SeaWiFS)—have substantially higher signal-to-noise ratios (SNRs), which means processes such as Rayleigh-aerosol scattering become significant. Atmospheric correction algorithms for the Moderate-resolution Imaging Spectroradiometer (MODIS) incorporate an approach to these issues, and researchers are using algorithms to explore the effects of absorbing aerosols in the stratosphere, especially sulfate aerosols associated with large volcanic eruptions. Based on these processes, a minimal band set for atmospheric correction can be defined. First, bands should be positioned to avoid specific absorption features in the atmosphere such as water vapor and oxygen. Second, at least two bands with some minimum spectral separation are necessary to characterize the spectral trends with sufficient accuracy. Lastly, bands should be placed in the near infrared (NIR) as noted above. A recent report by
OCR for page 59
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN the International Ocean Color Coordinating Group (IOCCG, 1998) proposes that bands near 865 and 750 nm, with perhaps a third band near 710 nm, would be suitable. In-Water Optical Properties Differentiation of all possible dissolved and particulate materials in the ocean would require far more bands with high spectral resolution in the visible wavelengths than are realistically possible. Therefore this discussion focuses on the primary properties of interest: phytoplankton chlorophyll, colored dissolved organic matter (CDOM), and suspended sediment. Sathyendranath et al. (1994) examined these issues for waters with high levels of suspended and dissolved material that are not correlated with phytoplankton abundance, as are often found in coastal zones and occasionally in the open ocean (Nelson et al., 1998). As noted earlier, the shift in ocean color as chlorophyll concentration increases is most pronounced around 445 nm, compared with the weak absorption around 560 nm. Atmospherically corrected water-leaving radiances from these two bands are combined in a ratio. This form is used because the sources of variability and error in the radiances, such as changes in the scattering and backscattering coefficients or the impact of bidirectional reflectance, are greatly reduced when band ratios are used. Although the basic band pair of 443 and 560 (with bandwidths of 20 nm) meets most of the requirements, the addition of a third band at 490 nm (which is easier than the 443-nm band to correct for atmospheric effects) will greatly improve the overall performance, especially for quantifying the effects of suspended sediments. A band that can distinguish between the optical signatures of chlorophyll and CDOM is also required. The divergence between these signatures occurs at wavelengths of less than 440 nm. Because atmospheric correction in the near-ultraviolet is extremely difficult, a compromise band has been used near 410 nm. Again, band ratios are used to estimate CDOM concentrations, much as with chlorophyll. Other Properties Some types of phytoplankton have very distinct spectral characteristics that may allow them to be identified even with limited spectral coverage sensors such as SeaWiFS and MODIS. Trichodesmium, a cyanobacterium that can fix atmospheric nitrogen, is usually found near the ocean surface and is characterized by high backscatter at 550 nm. Its unique pigment composition can be identified in CZCS imagery (Subramanian and Carpenter, 1994), and new algorithms are being developed for SeaWiFS that will quantify their abundance. Nodularia, another cyanobacterium, can also be distinguished in remote sensing of the ocean using a similar approach (Kahru et al., 1994). Coccolithophores, which produce bright calcium carbonate platelets, also change the spectral composition of water-leaving radiance (Balch et al., 1991; Brown and Yoder, 1994). Because they often occur in immense blooms, coccolithophores are easily recognized in ocean color imagery. New algorithms for SeaWiFS and MODIS have been developed to estimate the abundance of coccolithophore platelets. Beyond estimates of standing stocks, ocean color data are being used to formulate estimates of some important ocean fluxes. The most important of these is primary productivity. A comprehensive review of these models can be found in Behrenfeld and Falkowski (1997). The basic models use estimates of incoming solar radiation, chlorophyll, and ocean temperature to estimate photosynthetic rates. The primary difference between these models is in the way the light response is implemented; one class resolves the vertical structure of the fields, and the other uses vertically integrated fields. The best of these models explain roughly 50 percent of the variability observed in field measurements, and model estimates are within about a factor of two of the actual measurements. However, the quality and the spatial distribution of in situ measurements have their own sources of error, so the satellite-based estimates may actually be somewhat better than would be indicated by these studies. Productivity models rely on simple estimates of quantum yield, which quantifies the conversion of absorbed light energy into chemical energy. Light absorption can be relatively well described using estimates of chlorophyll, and most of the variability in the productivity models arises from variations in quantum yield. Changes in species composition and nutrient availability are largely responsible for the variability in quantum yield. Measurements of chlorophyll fluorescence yield from satellites may improve productivity estimates because of the generally
OCR for page 60
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN close relationship between fluorescence yield and quantum yield. Next-generation ocean color sensors, such as MODIS, Global Imager (GLI), and Medium-resolution Imaging Spectrometer (MERIS), include the necessary fluorescence bands at 670, 683, and either 700 or 750 nm with high SNR (Letelier and Abbott, 1996). Variations in water transparency are critical to predicting the depth of the upper ocean mixed layer, as they determine how solar energy is trapped as a function of depth (Denman and Miyake, 1973). Given that the mixed layer plays a critical role in the flux of energy between the atmosphere and the deep ocean, estimates of mixed-layer depth are important for quantifying Earth's energy budget. Operational Issues The spectral features of interest in the ocean are narrow, which leads to a narrow bandwidth requirement for ocean color measurements. Typically, 20 nm is sufficient in the visible range, although the fluorescence bands need to be slightly narrower to avoid atmospheric absorption features. The NIR bands for atmospheric correction can be wider, for example in the 30 to 40 nm range. With high SNR requirements for ocean color, it is also essential that band position be well known and stable over the life of the mission. Calculations of sensor performance must be based on typical values of water-leaving radiance. In the past, these calculations were made at values for top-of-the-atmosphere (TOA) radiances, which in the case of ocean color are up to ten times higher than the signal of interest. Obviously this can lead to spuriously high SNR values. Table 5.1 shows values for noise-equivalent delta radiance (NEΔL) and bandwidths full width at half maximum (FWHM) for SeaWiFS. These NEΔL values are based on typical water-leaving radiances. The atmospheric correction and in-water algorithms described above assume that there are no other sources of radiance reaching the sensor. In the case of ocean measurements, two potential sources must be discussed: sun glint and ocean whitecaps. Sun glint is the specular reflection of sunlight off the ocean surface. The size and exact placement of sun glint in a satellite image depend on (1) wind speed (through its forcing of capillary waves) and (2) satellite viewing geometry. CZCS and SeaWiFS are tilting sensors that can look 20 degrees fore or aft to avoid the glint patch. MODIS cannot tilt, and the glint patch must be masked out. Breaking waves create foam and whitecaps, obviously extremely bright targets on a dark ocean. They are not perfectly white reflectors (Frouin et al., 1996). Early models of breaking waves suggested that they would be an important component of the satellite-sensed radiance at wind speeds greater than 10 m s−1. Recent studies suggest that the effect is much smaller and may not become significant until wind speeds exceed 25 m s−1. As sensor performance improves, processes that once were in the noise of a particular sensor may be part of the signal in its successor. One might question which science requirements are driving the increase in SNR. The basic reason is the low signal emerging from the ocean. For measurements of phytoplankton biomass, we are only interested in those photons that penetrate the sea surface and are backscattered after interaction with suspended particulates (phytoplankton). Thus high SNR will increase our ability to discriminate water types, especially those at low chlorophyll concentration, which make up nearly half of the world's oceans. TABLE 5.1 SeaWiFS Performance Specifications for Eight Channels, Including Bandwidth Full Width at Half Maximum (FWHM) and the Noise-Equivalent Delta Radiance (NEΔL) Wavelength Bandwidth (nm) NEΔL 412 20 9.2 443 20 7.7 490 20 5.6 510 20 4.9 555 20 4.3 670 20 3.1 765 40 1.9 865 40 1.5
OCR for page 61
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Future Directions The basic measurements of phytoplankton chlorophyll, CDOM, and suspended sediment fulfill many of the fundamental requirements for global ocean biological studies. Missions such as SeaWiFS and MODIS fulfill these requirements at the necessary temporal and spatial scales. However, there are emerging scientific issues that will result in new requirements. Phytoplankton optical properties vary over the diel cycle, as does cloud coverage. Present and planned satellite missions are clustered around the time period between 0930 and 1200 (except for MODIS on PM-1, which will sample areas obscured by sun glint in the AM-1 MODIS imagery). Orbit crossing times could be spread more widely around local solar noon (e.g., 1000, 1200, and 1400) to investigate diel variability. This would be especially important for measurements of chlorophyll fluorescence (used to estimate quantum yield), which has a significant diel signal. As noted above, fluorescence represents a new measurement type, but the possibilities for improved estimates of primary productivity using fluorescence-based quantum yield look promising. Hyperspectral remote sensing has been used to study nearshore environments, partly because of the complexity of the optical signal. Early studies such as that of Campbell and Esaias (1983), while not hyperspectral, exploited spectroradiometer data to derive chlorophyll retrievals based on spectral curvature. More sophisticated algorithms (Lee et al., 1994) have been applied to Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) data to derive chlorophyll and CDOM concentrations, as well as bottom depths in shallow waters. As discussed above, the existing algorithms for the interpretation of remotely sensed ocean color signals are based on empirical correlations between the ratios of water-leaving radiances at only a few different wavelengths. These algorithms use a simplified parameterization of the composition of seawater in terms of chlorophyll concentration alone. A more rigorous approach to understanding remote sensing reflectance begins with determining the inherent optical properties of each of the various optically significant components of seawater. That knowledge enables one to model and understand the roles played by each component in determining the bulk inherent and apparent optical properties of the ocean, including remote-sensing reflectance. Such an approach leads naturally to the development of mechanistic remote-sensing models rather than to correlation models derived from statistical analysis of (usually incomplete) field data. These hyperspectral algorithms may allow us to characterize the types of flora in the ocean based on their pigment composition. The study of coastal ocean processes will require far more intensive spatial and temporal sampling than the open ocean because of its small characteristic scales. For example, tidal forcing is an important component of the coastal environment, and Sun-synchronous orbits will shift this high-frequency variability into lower frequencies. Geostationary platforms that could sample small regions every 15 minutes could be used to resolve such key processes. OBSERVING STRATEGY Time and Space Sampling Requirements The time and space variability of the ocean must be convolved with the biological time scales to derive an appropriate sampling strategy. Using ship, buoy, and drifter observations, the characteristic time scales for phytoplankton are on the order of 3 to 4 days (with shorter scales in more productive areas, such as the coastal ocean). When combined with typical circulation properties, this leads to characteristic spatial scales of about 20 to 30 km. Given the patterns of clouds that will obscure a large fraction of any one image (average cloudiness over the ocean is about 70 percent), we must have higher resolution to obtain this effective resolution. Chelton and Schlax (1991) have shown that for typical patterns of cloudiness, one can expect to achieve temporal resolution with reasonable errors of about 3 weeks. Although this is longer than the characteristic scale noted earlier, this does not mean that useful information cannot be obtained from typical satellite sampling of 1 km (nadir resolution) with 2-day global repeat coverage. Instead, the effective resolution requirement depends on the scientific questions being addressed. The Chelton and Schlax (1991) analysis assumes that spatially and temporally uniform error fields are required. This may be the
OCR for page 62
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN case for science questions involving long time series of global processes, but many mesoscale studies, such as the effect of coastal upwelling, may not have such stringent requirements. Thus the nominal 1-km, 2-day global coverage sampling strategy is a reasonable compromise for a wide variety of scientific problems. The last part of any sampling strategy is the resilience to temporal gaps in the data record. That is, must continuity be maintained absolutely? Aside from operational requirements, which are not the focus here, most science questions involving ocean color can tolerate some gaps in the record. Given that the dominant ocean climate signal is the ENSO, which recurs approximately every 3 to 5 years, a gap longer than about 6 months runs a serious risk of missing this critical phenomenon. NASA's Plans Since the success of CZCS (launched in 1978), NASA has pursued many avenues to obtain science-quality ocean color measurements on a global basis. In 1997, Orbital Sciences, Inc., successfully launched SeaWiFS aboard Orbimage-2 as part of a data purchase agreement with NASA. SeaWiFS should provide measurements through 2002. With its bilinear gain, SeaWiFS will collect data over both the dark ocean and the bright land. MODIS, launched in 1999 aboard the Earth Observing System (EOS) AM-1 platform, is a multipurpose sensor, measuring radiance in the visible portion of the spectrum for both land and ocean applications as well as infrared measurements for surface temperature, atmospheric sounding, and cloud imaging. MODIS has considerably higher SNR than SeaWiFS and also has some 250-m resolution bands for land-surface imaging. Because MODIS does not tilt, a second MODIS will be launched in 2000 on PM-1 to provide global, 2-day coverage, albeit at different orbit crossing times. Given the high level of science involvement in both the SeaWiFS and MODIS missions, these sensors meet the basic science requirements and provide new capabilities to explore new scientific questions. In 1997, NASA began to reformulate its plans for the second series of EOS missions. Currently, NASA does not have any concrete plans for ocean color measurements after MODIS on PM-1, although it is likely that some plan will emerge during the second series planning. The question remains whether this mission will meet the science requirements. Moreover, there are new directions that the ocean science community wishes to pursue, especially as the U.S. Global Change Research Program (USGCRP) begins to focus on linkages between ecosystem processes and climate change. With the success of SeaWiFS and the anticipated success of MODIS, there will be considerable pressure to maintain at least MODIS-quality ocean color measurements. NPOESS Plans The Department of Defense and NOAA have developed a set of operational requirements for ocean color measurements as part of the environmental data records (EDRs) process (see Table 5.2). These EDR requirements will be met by the Visible and Infrared Imaging Sensor (VIIRS); it has not been decided whether VIIRS will consist of a single package like MODIS, or if it will consist of separate sensor modules. These requirements were reviewed briefly in a climate workshop (NOAA, 1997). The present threshold and objective requirements for ocean color are difficult to evaluate for several reasons. First, only one of the planned National Polar-orbiting Operational Environmental Satellite System (NPOESS) orbits will be able to make measurements of sufficient quality (crossing time 1330); 0530 and 0930 are simply too early in the day. Second, the requirements are reported in terms of chlorophyll rather than as radiances. Since chlorophyll errors will be a function of the various atmospheric correction and in-water algorithms, it will simplify sensor comparisons if they are based on normalized water-leaving radiances (which have the atmospheric correction) and TOA radiances (which do not). Third, many important sensor specifications are absent: for example, spectral bandwidth, band stability, digitization rate, bright target recovery, band-to-band registration, and so forth are not specified. This was done largely because the NPOESS EDR process was based on final data products, not on a specific sensor design. While this approach may result in a lower-cost sensor and also does not lock NPOESS into a particular technical implementation, it is based on an assumption that the quality of a data product is independent of the sensor design. As was learned with SeaWiFS, this may not be a realistic assumption. Fourth,
OCR for page 63
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN TABLE 5.2 NPOESS Environmental Data Records for Ocean Color Ocean Color/Chlorophyll Threshold Objective Horizontal resolution Global, worst case 2.6 km 1 km Regional, worst case 1.3 km 0.1 km Mapping accuracy Global, worst case 3 km 0.5 km Regional, worst case 3 km 0.1 km Measurement range 0.05-50 mg/m3 0-100 mg/m3 Measurement precision 20% 10% Measurement accuracy ±30% ±30% Refresh 48 h 12 h Turbidity Sensing depth Surface Horizontal resolution 1.3 km 0.25 km Mapping accuracy TBDa 0.5 km Measurement range TBD 0-100 mg/l Measurement precision TBD 0.1 mg/l Measurement accuracy ±30% ±0.1 mg/l Refresh 48 h 24 h there are no explicit requirements for long-term stability. Although there is an implicit expectation that the data product must continue to meet its threshold performance, experience has shown that new platforms are replaced only when critical sensors fail. The slow degradation of a noncritical data product such as ocean color may result in a long period of scientifically useless data. More importantly, there is no explicit plan for periodic calibration and validation, as there is with sensors such as SeaWiFS and MODIS. The committee 's assessment of the current status and future NPOESS plans for the observational measurement of ocean color can be found in Box 5.1. aTBD, to be determined. SOURCE: IPO NPOESS (1996). The updated IORD and other documentationrelated to the NPOESS program are available online at <http://npoesslib.ipo.noaa.gov/ ElectLib.htm>. International Plans Other nations are planning ocean color sensors; notable among these are the Global Imager on the Japanese ADEOS-2 spacecraft and MERIS on the European ENVISAT. Both of these missions are strongly science-driven, and although there are differences in performance and objectives, both are similar to MODIS. Figure 5.1 compares the performance of SeaWiFS, the Japanese Ocean Color Temperature Scanner (OCTS) on ADEOS-1, MODIS, GLI, and MERIS. Although GLI and MERIS are highly capable sensors, it will be difficult to use them as a basis for developing a strategy for long time series of ocean color for studies of climate-related processes, given the uncertainty of international planning. Therefore, the most prudent strategy is one based on a combination of research and operational missions. DATA PRODUCTS The critical data products are discussed earlier in this chapter. The fundamental set includes phytoplankton chlorophyll, CDOM, and suspended sediment. To obtain these products requires a complicated set of atmospheric correction algorithms. Because these two classes of algorithms are still the subject of research, it is difficult to conceive how they can be “purchased” as part of a sensor package. Ocean algorithms will continue to evolve, especially as the time series lengthen, new problems arise, and scientific understanding improves. Examples of possible improvements are in accounting for absorbing aerosols in the stratosphere, bidirectional reflectance off the ocean surface, and in-water algorithms based on inherent optical properties. Therefore a program of active research to improve and evaluate algorithms must also be continued. NASA has a history of supporting research to improve data sets, although it generally focuses on data sets collected by NASA missions. The NASA/NOAA Pathfinder data sets were a notable exception, with the two agencies collaborating to produce high-quality data sets for Earth science.
OCR for page 64
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN Box 5.1 Findings Some of the basic science requirements for ocean color remote sensing will be met by the VIIRS sensor on NPOESS. The EDRs for chlorophyll and suspended sediment are congruent with those developed for SeaWiFS. However, there is no EDR for CDOM (see Table 5.1), an important component of the ocean optical environment and its biogeochemistry. Moreover, there are many other aspects of overall system performance that must be met. The plans for VIIRS are difficult to evaluate because many critical elements are incomplete or not yet determined. This stems from the EDR approach, which proceeds from the definition of the data product requirements rather than specifications of the technology. These missing elements include calibration and validation, data archiving, sensor characterization, algorithm development, and technology insertion. Climate research requires more than hardware in orbit; it requires a complete, integrated program of sensor engineering and testing, as well as scientific research. The first series of EOS sensors will explore new capabilities in ocean color remote sensing. The first series of EOS platforms (AM-1 and PM-1) will meet the basic science requirements for ocean color, as will SeaWiFS. MODIS, with additional bands and better performance, will support the development of new data products and testing of new concepts. For example, chlorophyll fluorescence bands, which are part of every upcoming ocean color sensor (MODIS, MERIS, and GLI), will not be included in VIIRS. These bands will significantly improve estimates of primary productivity in the upper ocean. Although NASA may continue such measurements in a research context, there is no mechanism to insert these new requirements into the operational NPOESS system. The present plans (as they have been defined) do not ensure that a continuous time series of ocean color data will be collected that is suitable for climate research. There is the possibility of a gap between the first series of EOS missions and the first NPOESS mission, and it is probable that any “gap filler” will be less capable than MODIS. However, if AM-1 lasts only until 2005, then there may be a significant drop in coverage until the launch of VIIRS in 2009. This gap may be eliminated by a sensor on the second EOS series, but the quality of such a sensor and its associated data products is still being defined. NASA has committed to continuity of ocean color measurements, perhaps through an early flight of VIIRS, which is being studied as part of a joint NASA/NOAA bridging mission to fly in the 2005 to 2009 time frame. While the ocean color measurements in NPOESS may be useful for long time series, considerably more planning and coordination are required. If dynamic continuity is to be achieved, programs such as Sensor Intercomparison and Merger for Biological and Interdisciplinary Oceanic Studies (SIMBIOS) must be extended and maintained to include the operational missions as well as the research missions. Required activities include adequate prelaunch characterization and testing, as well as ongoing data product validation and analysis. Improvement is needed in coordinating NASA's technology innovation and climate research with the Integrated Program Office's (IPO's) commitment to long time series of operational missions. NASA's technology plans are moving so rapidly that it is unlikely that any new technology will become part of NPOESS, given the mismatch in schedules. An explicit plan must be developed for ocean color research to transfer to the operational systems. The fact that ocean ecosystems change on decadal scales in response to climate change makes their observation especially challenging. The commitment of NASA and IPO/NPOESS to long time series is a critical component of an observing strategy, but the present plans are incomplete. It is not simply a case of defining better requirements for NPOESS ocean color observations, though this is needed. A complete program of research, analysis, and technical innovation, with a long-term commitment to the measurements, is required. There is an opportunity to achieve these scientific goals without an enormous financial investment. The research missions begun by NASA have established a strong scientific basis for ocean color as well as a sound technical basis. NPOESS represents an opportunity to ensure long-term observations. The challenge will be to blend active research, technology innovation, and continuous measurements to develop an observing system for climate research.
OCR for page 65
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN FIGURE 5.1 Noise-equivalent radiance as a function of wavelength for several ocean color sensors. The distribution shows the improvement in sensor performance from the first ocean color sensor (CZCS) to the upcoming sensors of MODIS, MERIS, and GLI. The S-GLI sensor is planned for ADEOS-3. Acronyms are defined in Appendix B. One particular area where ocean color algorithms need attention is the estimation of water column productivity. Although fluorescence should improve these estimates, there is no plan within NPOESS to continue these measurements. Therefore, it is probable that productivity estimates will have to rely on indirect estimates of the photoadaptive state. This may include estimates of mixed-layer depth, nutrient concentration through the use of nutrient and temperature climatologies, and so forth. These models may be more robust than simple statistical correlations, but the additional uncertainties that arise from these new parameterizations will pose new challenges. CALIBRATION AND VALIDATION SeaWiFS and MODIS have extensive onboard calibration systems to monitor sensor performance. These systems include monitoring of sensor response through lunar and solar calibration, as well as monitoring of spectral stability (for MODIS). Although these systems (and in the case of lunar calibration, orbital maneuvers)
OCR for page 66
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN will monitor sensor performance, subtle changes may result in significant changes in derived data products. NPOESS will maintain some level of calibration, but neither the requirements nor the approach have been defined. As discussed in this and earlier chapters, onboard monitoring is necessary but not sufficient. A rigorous program of vicarious calibration through the use of in situ measurements is also necessary. Both SeaWiFS and MODIS have ongoing programs of field studies in support of vicarious calibration. Extensive cruises and moored buoys are part of this program. These programs are necessarily costly, given the need to ensure that the in-water measurements are themselves well calibrated. Thus there are recurring questions about alternative approaches to such vicarious calibration. However, there is no suitable alternative if the requirement is to maintain data quality for climate research. NPOESS has not defined a plan for vicarious calibration. Neither NASA nor NPOESS has yet committed to maintaining dynamic continuity between sensors. After AM-1 and PM-1, it is not clear what type of ocean color sensors NASA will fly, or how they will overlap with the first EOS series. The transition to the operational VIIRS sensors on NPOESS is also not yet planned. The risk is that an enormous investment will be made, yet it will not be possible to intercalibrate the data sets. Dynamic continuity was challenging for the Pathfinder data sets, where identical sensors were flown and the sensors and their associated algorithms were simpler. Given the low signal, ocean color requires precise and accurate knowledge of calibration. Ground networks (including process studies, moorings, and drifters) provide well-calibrated data sets that are essential for interrelating successive satellite sensors. Lastly, insertion of new technology developed by NASA or NPOESS must be part of any continuous time series. It is not apparent how new technology infusion will be accommodated in NPOESS after the final design for VIIRS is selected. NASA has recently started the SIMBIOS (Sensor Intercomparison and Merger for Biological and Interdisciplinary Oceanic Studies) program, which is designed to gather in situ information necessary to develop a consistent time series of ocean color from multiple satellite sensors. Although it initially is focused on U.S. sensors, the IOCCG (working under the auspices of the Committee on Earth Observation Satellites) is studying a possible international program that would incorporate all the ocean color research missions. There is no firm plan to continue this activity beyond 2001. Programs such as SIMBIOS require extensive prelaunch sensor characterization studies. Tests for SeaWiFS and MODIS included polarization, spectral response, scan mirror reflectivity, and others. These activities are costly, and there is frequently considerable pressure to eliminate or scale back such tests to lower costs or preserve schedule. However, one lesson learned from any time series is that such characterization is essential. Moreover, the data from such tests must be preserved and documented with as much care as the original satellite data. Since such tests are part of sensor construction, they are sometimes treated simply as engineering data and are preserved only in written form. NPOESS contractors are supposed to maintain this information in an accessible form. For MODIS, these tests are conducted with the cognizance of the science team. Validation consists of estimating the quality of the derived data products, including the effects of sensor noise and processing algorithms. Estimation of the impact of sampling errors is also part of the validation process. SIMBIOS is supporting some activities in this regard, although the focus is on the algorithm part of the error budget. One of the MODIS team activities is assessment of algorithm errors as well. There are no identified plans for data product validation as part of NPOESS. EVOLUTION STRATEGY NASA has added three programs to develop and test new technology. Although there are no specific plans yet in the area of ocean color, some proposed candidate technologies include measurements from geostationary orbit to study high-frequency events, hyperspectral measurements to examine phytoplankton pigment groups, fluorescence imagers to improve productivity models, and lidar to study mixed layers and pigment fluorescence. The Integrated Program Office for NPOESS (IPO/NPOESS) is planning to reserve some platform resources to accommodate new technology, but this plan is not well defined at present, nor is there a mechanism to exploit these resources. NASA has moved strongly toward a program of rapid technological development in order to lower costs and develop more capable sensors in response to improvements in scientific understanding. Many of these missions
OCR for page 67
ISSUES IN THE INTEGRATION OF RESEARCH AND OPERATIONAL SATELLITE SYSTEMS FOR CLIMATE RESEARCH: I. SCIENCE AND DESIGN have short time scales, completing the planning and design phase in 1 to 2 years, followed by a relatively short mission. NASA technology programs such as the New Millennium Program and the Instrument Incubator Program are not focused on improving the capabilities of NPOESS. If NPOESS planning takes 5 or 6 years, it is not clear how NASA technology will be incorporated into NPOESS. Operational missions require several years of proven spaceflight to increase confidence in the sensor design as well as demonstrate the utility of the data set. REFERENCES Balch, W.M., P.M. Holligan, S.G. Ackleson, and K.J. Voss. 1991. Biological and optical properties of mesoscale coccolithophore blooms in the Gulf of Maine. Limnol. Oceanogr. 36: 629-643. Behrenfeld, M.J., and P.G. Falkowski. 1997. Photosynthetic rates derived from satellite-based chlorophyll concentration . Limnol. Oceanogr. 42: 1-20. Brown, C., and J.A. Yoder. 1994. Coccolithophorid blooms in the global ocean. J. Geophys. Res. 99: 7467-7482. Campbell, J.W., and W.E. Esaias. 1983. Basis for spectral curvature algorithms in remote sensing of chlorophyll . Appl. Opt. 22: 1084-1093. Chelton, D.B., and M.G. Schlax. 1991. Estimation of time-averages from irregularly spaced observations: With application to coastal zone color scanner estimates of chlorophyll a concentrations. J. Geophys. Res. 96: 14669-14692. Denman, K.L., and M. Miyake. 1973. Upper layer modifications at Ocean Station Papa: Observations and simulation. J. Phys. Oceanogr. 3: 185-196. Frouin, R., M. Schwindling, and P.-Y. Deschamps. 1996. Spectral reflectance of sea foam in the visible and near-infrared: In situ measurements and implications for remote sensing of ocean color and aerosols. J. Geophys. Res. 101: 14361-14371. Gordon, H.R., and A.Y. Morel. 1983. Remote Assessment of Ocean Color for Interpretation of Satellite Visible Imagery. A Review. New York: Springer-Verlag. 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. International Ocean Color Coordinating Group (IOCCG). 1998. Minimum requirements for an operational ocean-colour sensor for the open ocean. IOCCG Rept. No. 1, Dartmouth, Nova Scotia, Canada. Kahru, M., U. Horstmann, and O. Rud. 1994. Satellite detection of increased cyanobacteria blooms in the Baltic Sea: Natural fluctuation or ecosystem change. Ambio 23: 469-472. Kirk, J.T.O. 1994. Light and Photosynthesis in Aquatic Ecosystems. New York: Cambridge University Press. Lee, Z., K.L. Carder, S.K. Hawes, R.G. Steward, T.G. Peacock, and C.O. Davis. 1994. Model for the interpretation of hyperspectral remote-sensing reflectance . Appl. Opt. 33: 5721-5732. Letelier, R.M., and M.R. Abbott. 1996. An analysis of chlorophyll fluorescence algorithms for the Moderate Resolution Imaging Spectrometer (MODIS). Remote Sensing Environ. 58: 215-223. 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. Nelson, N.B., D.A. Siegel, and A.F. Michaels. 1998. Seasonal dynamics of colored dissolved material in the Sargasso Sea . Deep-Sea Res. 40: 931-957. Sathyendranath, S., F.E. Hoge, T. Platt, and R.N. Swift. 1994. Detection of phytoplankton pigments from ocean color: Improved algorithms . Appl. Opt. 33: 1081-1089. Subramanian, A., and E.J. Carpenter. 1994. An empirically-derived protocol for the detection of blooms of the marine cyanobacteria Trichodesmium using CZCS imagery. Int. J. Remote Sensing 15: 1559-1569.
Representative terms from entire chapter: