Seeing Microbes from Space Remote sensing is now a critical resource for tracking marine microbial ecosystem dynamics and their impact on global biogeochemical cycles

Douglas G. Capone and Ajit Subramaniam

icrobiologists studying aquatic or other large bodies of to mean environments are benefiting that they are clean, whereas the of from their use of remote sensing rivers and streams mean they are carrying sus- M technologies. With access to data pended mud or other materials. gathered by remote sensors on This instinctive knowledge forms the basis of aircraft and satellites, microbiologists now can color science. Sunlight that enters the analyze microbial population dynamics with ocean can either be absorbed or scattered back greater breadth and a novel perspective. (Fig. 1). Pure water absorbs most but Remote sensing instruments exploit electro- strongly scatters most light. Thus, open magnetic radiation to study surface processes on ocean with very little material in them earth. Passive sensors use reflected sunlight or appear deep blue, while waters carrying dis- heat being emitted by objects along the earth’s solved organic materials that absorb blue light surface, while active sensors transmit laser or strongly appear brown. Chlorophyll-containing microwaves that are then reflected back to and microalgal assemblages—phytoplankton—ab- recorded by the sensor. Many orbiting satellites sorb both blue and red light, making phyto- currently map ocean properties such as color, plankton-containing waters look . In addi- surface temperature, height, wind velocities, tion, chlorophyll emits a small fraction of the roughness of the ocean surface, and wave absorbed light at a longer wavelength as red height. Plans call for additional sensors for map- fluorescence, providing another signal for re- ping salinity, as well as the vegetation canopy mote sensing. Scientists use these variations in using active light detection and ranging (LI- the color of water to estimate the amount of Douglas G. Capone DAR). Although most remote sensing instru- phytoplankton, suspended sediment, and dis- is the Wrigley Pro- ments are limited to near-surface properties, solved organic material. fessor of Biological subsurface processes often can be inferred. Several satellite-borne sensors, including Sciences in the There are several advantages to using remote the -Viewing Wide Field-of-View Sensor Wrigley Institute for sensing techniques, such as accessing otherwise (SeaWiFS), Moderate-Resolution Imaging Spec- Environmental difficult locations and rapidly mapping large troradiometer (MODIS-Terra, MODIS ), Studies and Depart- swaths of the earth’s surface. For example, sen- Medium Resolution Imaging Spectrometer ment of Biological sors that are aboard polar-orbiting satellites can (MERIS), and the Ocean Colour Monitor (OCM), Sciences, Univer- map the entire surface of the earth each day. measure specific wavelengths of light emanating sity of Southern Each two-minute scene recorded by such a sat- from ocean surfaces (see the online article for a full California, Los ellite sensor, which has a 1-km spatial resolu- list of satellite sensors). Perhaps the best and most Angeles, and Ajit tion, would take 11 years to sample for a ship obvious example of the microbiological exploita- Subramaniam is a traveling at about 20 km per hour. tion of satellite technology is the broad-scale map- Doherty Associate ping of phytoplankton in the oceans (Fig. 2) Research Scientist (http://oceancolor.gsfc.nasa.gov). This technol- at the Lamont Mapping the Oceans Using ogy enables us to describe spatial, seasonal, and Doherty Earth Ob- One form of remote sensing uses the color of interannual patterns and dynamics of oceanic servatory of Colum- water to infer what it contains. For example, plants and is used to infer rates of photosyn- bia University, New many of us consider the deep blue color of thetic carbon uptake, often referred to as pri- York.

Volume 71, Number 4, 2005 / ASM News Y 179 FIGURE 1 being studied by a suite of sensors, in- cluding those that map sea surface tem- peratures (SST), surface wind velocities, and sea surface height (SSH). These nu- trient enrichments result in high densi- ties of diatoms and other eukaryotic microalgae in surface waters. Dia- toms serve as major primary produc- ers in important food webs, contrib- uting substantially to the cycling of silicon in the sea and to the sinking of carbon from surface waters. Diatoms also dominate segments of open ocean that were experimentally fertilized (Fig. 3A). Researchers are developing algorithms to discriminate diatoms from other phytoplankton. Indeed, satellites are being used to detect a range of specific phytoplank- ters. For instance, they are used to map The color of the ocean is recorded by a satellite sensor by measuring the “remote the planktonic marine cyanobacterium

sensing reflectance (Rrs)” at specific wavelengths. Rrs is the ratio of light leaving the Trichodesmium that occurs throughout water (Lw) to the light incident on it (ET) and is determined by the absorption (a) and the marine tropics (Fig. 4). In addition backscattering (b ) of light in the (R ϭ L /E ϰ b /(a ϩ b ). b rs w T b b to its role in fixing carbon through pho- tosynthesis, this organism also fixes ni- trogen gas to produce organic nitrogen, mary production. Through such analyses, we making this otherwise scarce nutrient available more fully understand the role of ocean micro- in waters where it resides. Although Trichodes- biota in the global carbon cycle. mium lacks specialized heterocyst cells that some nitrogen-fixing cyanobacteria depend on, it contains proteinaceous vacuoles, or gas vesicles, Satellite Observations Provide a Means providing cells with buoyancy and keeping such for Detecting Specific Organisms populations near the surface where light is plen- tiful. Trichodesmium occurs as individual Interest in learning how to more fully character- trichomes as well as multifilament aggregates, ize phytoplankton populations in the great ex- often referred to as colonies (Fig. 4a and b). panses of the world’s oceans continues to grow. Several optical properties make Trichodes- We now recognize a range of “functional groups” mium an excellent microbe for study by remote of marine autotrophs that are characterized by sensing. For instance, its gas vesicles are very their biogeochemical role in different elemental highly reflective. Furthermore, unlike most eu- cycles. Scientists now can identify and quantify karyotic algae, it contains phycoerythrin as a monospecific blooms of certain microalgae and major accessory pigment. The absorption and cyanobacteria. And, by coupling these measure- fluorescence spectra of this pigment are readily ments with ecosystem modeling, we can esti- distinguished from chlorophyll a. Finally, its mate the broad-scale biogeochemical impacts of near-surface habitat (staying in a zone within these populations. about 50 m from the ocean surface) makes it rel- In many areas, including the oceans around atively easy to observe. Indeed, astronauts aboard Antarctica, along the eastern boundaries of the the space shuttle frequently see and photograph Atlantic and Pacific Oceans, and along the equa- its more extensive blooms (Fig. 4D). Algorithms tor (Fig. 2), a phenomenon referred to as up- that are based on the six color bands detected by welling brings nutrient-rich waters from deep in the SeaWiFS satellite can be used to identify the ocean to the surface. This phenomenon is Trichodesmium-associated chlorophyll (Fig.

180 Y ASM News / Volume 71, Number 4, 2005 A Childhood at the Seashore, a Career Studying Marine Microbiology When Douglas Capone visited rent work is focused on the nitro- Brookhaven National Labora- California in 1997 to interview for gen cycle.” tory, Upton, N.Y. his current job, the dean quizzed He and his colleagues are focus- Capone met his wife, Linda Du- him about a two-year gap during ing on oceanic nitrogen fixa- guay, a marine physiological his undergraduate days. “I was tion—the biological conversion of ecologist who is now a science involved in some radical activities nitrogen gas into biologically use- administrator, during graduate on campus, and I was asked to ful forms of nitrogen. Capone be- school at the University of Miami. leave,” Capone told the dean, who lieves that nitrogen fixation may They have two daughters, 25 and asked in reply, “You’re not going be a key in determining how much 20. Their older daughter works as to do that here, are you?” carbon dioxide tropical oceans an account manager with the Ad- Hardly. Those indiscretions can absorb. visory Board Co. in Washington, during the heady Vietnam War After his two-year excursion D.C., while the younger is a soph- years of the late 1960s and early from college, Capone enrolled at omore at USC. Both focus on 1970s marked a time of wide- the University of Miami, where he business, he says, adding that “the spread protests and plentiful per- received a B.S. degree in biology younger one now has a minor in sonal uncertainties. Capone faced in 1973. He stayed there for his environmental sciences—possibly them as an unenthusiastic premed doctoral studies and, in 1978, re- a concession to us.” student at Seton Hall University ceived a Ph.D. in marine sciences In a typical year, he spends in South , N.J. Like many from the University of Miami about 30–50 days at sea. Capone of his contemporaries, he vehe- Rosenstiel School of Marine and first came to love the ocean while mently opposed the war and thus Atmospheric Science. “Having growing up near the Jersey shore. participated along with other stu- matured a bit, and while finishing His parents, both office workers dent activists who took over an my basic undergraduate degree in in Newark, N.J., took Capone administration building. In 1970, biology, I got into a work-study and his younger sister to the shore Capone dropped out of school, program at the marine laboratory on weekends. By the time Capone worked at several jobs, and then on Virginia Key, near Key started high school, they had hitchhiked around the country. “I Biscayne,” he recalls. “I was just a moved there permanently. was drifting, trying to figure out grunt in the laboratory, doing “When I was young, I loved to where I wanted to go in my life,” Winkler O2 titrations, grinding he recalls. corals and analyzing them—but surf and fish, so a career that kept Soon enough, Capone was the exposure to the researchers me close to the water seemed log- drawn to studying the ocean, and there and their focus on the ical,” he says. “The sea has al- that is how he continues to spend ocean’s microbial world really got ways lured me. It’s been a defining his adult life. Today he directs the my attention.” focus of my activities, both profes- marine environmental biology Before moving to California, sionally and recreationally. Of program at the University of Capone was on the faculty at the course, marine microbiology is Southern California (USC) Wrig- University of Maryland and more than a career—it’s a pas- ley Institute for Environmental served in its center for environ- sion.” When on land, he runs most Studies. His research focuses on mental science in the Chesapeake days (often on the beach!), and the role and importance of marine Biological Laboratory in So- enjoys skiing. But, not surpris- microbes in major biogeochemi- lomons, Md. Earlier, he worked ingly, his favorite vacations are at cal cycles. “We study the natural in the Marine Sciences Research the shoreline. “A busman’s holi- populations of bacteria and ar- Center at the State University of day,” he says. chaea in different environments New York at Stony Brook and Marlene Cimons and try to tease out what they are was a visiting scientist in the de- Marlene Cimons is a freelance writer doing,” he says. “Most of my cur- partment of chemistry at the in Bethesda, Md.

Volume 71, Number 4, 2005 / ASM News Y 181 FIGURE 2-4

Fig. 2. Global annual average plant biomass map derived from the SeaWiFS satellite. The oceanic biomass is mapped as chlorophyll concentration while the terrestrial biomass is indicated as the normalized difference vegetation index. Fig. 3. A bloom of diatoms induced by open ocean iron fertilization experiment (SOIREE) in the Southern Ocean in February 1999 as detected by the SeaWiFS satellite on 23 March 1999; (B) a SeaWiFS “true color” image showing blooms of Coccolithophores in the North Sea and a diazotrophic cyanobacterium, Nodularia, in the Baltic Sea from July 1999; (C) a photograph of an intense “red ” Noctiluca bloom off Southern California. Fig. 4. Trichodesmium trichomes (A), colonies (B) a surface bloom in the Capricorn Channel and (D) a photo from the space shuttle of NW Australia. The highly reflective features in (A) are gas vesicles. Each filament in (A) is about 10 um in diameter. The colony in B is about 5 mm by 10 mm. The bloom seen in the space shuttle photo extends over 100 km.

182 Y ASM News / Volume 71, Number 4, 2005 FIGURE 5

True- of a Trichodesmium bloom (A), the Trichodesmium derived chlorophyll (B), and modeled N2 fixation (C). From Subramaniam et al., 2002, and Hood et al., 2002, reprinted with permission of Elsevier.

5B). However, the detection threshold is rela- nous coating gives such mats a high reflectance tively high, about 0.8 mg per cubic meter, which in the green part of the spectrum, allowing sci- represents a substantial bloom. entists to develop algorithms that can distin- Trichodesmium is not the only microorgan- guish Phaeocystis blooms from other phyto- ism that is specifically detectable from space. plankton blooms such as those of diatoms that Coccolithophores, a group of the Prymnesio- also have high reflectances. Phaeocystis and di- phyte algae, form calcium carbonate plates, or atoms play different roles in the carbon cycle at coccoliths, on their outer surfaces. They partic- high latitudes, with the latter helping to draw ipate in carbon cycling by producing both or- down carbon when the bloom sinks. Phaeocys- ganic carbon and particulate inorganic carbon tis is also biogeochemically important because it from dissolved inorganic carbon. Coccolith produces dimethylsulfoniopropionate, which is plates are highly reflective at short wavelengths degraded to dimethyl sulfide that escapes into and turn the water milky blue when found in the atmosphere to join the atmospheric sulfur high concentrations (Fig. 3B). cycle and contributes to cloud formation. Growth of the nitrogen-fixing cyanobacte- In coastal areas, noxious or toxic blooms are rium Nodularia is responsible for the massive recurrent and can interfere with recreational bloom visible seasonally at the center of the and commercial uses of these waters. Often re- Baltic Sea (Fig. 3B). These cyanobacteria have ferred to as harmful algal blooms (HABs) optical properties similar to Trichodesmium, (http://www.whoi.edu/redtide/), they are an proliferate in the late summer, and have been emergent problem. Although there are no cur- mapped using ocean color satellite sensors. rent optically based, organism-specific algo- Such cyanobacterial blooms accumulate at the rithms for mapping these blooms, they often are surface, changing the reflectance of microwaves monospecific and can be spectacular in intensity transmitted by synthetic aperture radar (SAR) (Fig. 3C), making them easy to detect by satel- satellite sensors. These SAR sensors allow us to lite. Coastal management agencies typically use map these blooms even in cloudy conditions and satellite images to identify locations with sub- at spatial resolutions as fine as 10 m. stantial biomass to direct sampling to determine Other monospecific blooms can also be ob- whether they are due to HABs. Scientists at the served from space. Phaeocystis spp., another National Oceanic and Atmospheric Administra- Prymnesiophyte, often form mucilaginous mats tions Coastal Services Center are using remote when cells bloom at high latitudes such as in sensing, in-situ field measurements, and buoy waters surrounding Antarctica. The mucilagi- data to provide a HAB bulletin for the Gulf of

Volume 71, Number 4, 2005 / ASM News Y 183 Mexico coastal management community (http: example, to study toxic cyanobacterial blooms //www.csc.noaa.gov/crs/habf/). in the Baltic Sea and to map oil spills. Poten- tially, it could also be used to follow the disper- sion and microbiological degradation of such Beyond Chlorophyll to Other Products spills. and Indirect Approaches Ocean is used to measure param- Modeling and Scaling eters other than phytoplankton biomass. For instance, satellite imagery of global cloud cover One objective in remotely detecting populations is used to estimate the photosynthetically of microalgae, specific phytoplankters, or func- available radiation (PAR) available at the sea tional groups is to analyze their dynamics as a surface every day. Apart from being key to context for estimating their ecological and bio- photosynthesis, a measure of the penetration of geochemical importance. light in the water column is also an important For instance, we can use remote sensing to mea- measure and K490, the diffuse sure phytoplankton concentrations, then model coefficient at 490 nm, is a standard primary production by the phytoplankton glo- global ocean color product. bally. The growth rate of phytoplankton depends Other algorithms are used or are being devel- mainly on phytoplankton concentration, the oped to estimate particulate inorganic carbon in amount of light the phytoplankton receives for the form of liths from coccolithophores, total photosynthesis, and factors such as temperature suspended solids, particulate organic carbon, and the supply of nutrients to sustain growth. and the absorption due to colored dissolved Since most nutrients in the open ocean come organic matter (CDOM) from highly refractory, from the deep where it is very cold, maps of sea riverborne organic material. Maps of CDOM surface temperatures can be used as a proxy for help us to understand the influence of terrestri- surface nutrient maps. By combining satellite- ally derived organic material on the marine eco- based maps of incident light, phytoplankton system as well as the dynamics of dissolved concentrations, and SST, we can create global organic matter in the open ocean. Dissolved maps of primary production and carbon fixa- organic carbon is the primary substrate for het- tion rates. Similarly, for a diazotroph such as erotrophic marine bacterioplankton, and photo- Trichodesmium, we have developed models to bleaching CDOM conditions it for consumption predict in situ nitrogen fixation based on re- by bacteria. Thus, tracking CDOM may be use- motely sensed biomass and solar flux (Fig. 5C). ful for understanding bacterial activity in sur- However, although Trichodesmium can be face ocean waters. directly detected only at high surface densities, Apart from ocean color, there are other satel- these organisms are found throughout most oli- lite resources which may be exploited for study- gotrophic oceanic warm waters in appreciable ing the distribution and activity of oceanic mi- abundance. Indeed, warm surface waters (Fig. crobes. For example, although Vibrio cholerae 6A) that tend to be depleted of combined inor- cannot be detected directly by satellite sensors, ganic nitrogen such as nitrate (Fig. 6B) are hab- marine plankton are a significant marine reser- itats where one typically finds Trichodesmium voir of these bacteria. In the Bay of Bengal, the and other diazotrophs. Thus, we use SST mea- annual cycle of SST is similar to that of cholera surements from the satellite sensors to estimate cases, and SSH is a good indicator of incursions the area of warm waters (e.g., 25°C) as a proxy of plankton-laden waters. Combined, the two for estimating nutrient-depleted zones, then provide a statistically significant correspon- scale these estimates based on rates of nitrogen dence with the incidence of cholera in Bang- fixation determined on oceanographic cruises. ladesh. SSH is another useful parameter with ecolog- SAR measures the roughness of the sea sur- ical relevance. SSH indicates the depth of the face that can be altered by surfactants or phyto- , the thermal discontinuity between plankton blooms at the surface. SAR, which is warm surface waters and cold, nutrient-rich an active remote sensing technique, can be used deep waters. Through much of the oceans, SSH even when clouds are present and with resolu- inversely correlates with chlorophyll concentra- tions of 10 m. SAR imagery is being used, for tions. That is, normally, a low SSH and shallow

184 Y ASM News / Volume 71, Number 4, 2005 thermocline indicates a proximal FIGURE 6 source of deep nutrients resulting in high chlorophyll. However, in certain areas of the oligotrophic ocean, such as the southwest North Atlantic Ocean, high SSH is coincident with high chloro- phyll. This unusual finding appears to indicate an area of the ocean where higher levels of nitrogen fixation may support higher algal biomass despite a paucity of nutrients. Remote sensing can also provide in- sights on large-scale controls of micro- biological processes in the ocean. Throughout much of the surface ocean, dissolved iron is available only in trace concentrations (less than nanomolar). A primary source of iron is through deposits of atmospheric dust. Nitrogen- fixing microorganisms, particularly those that also carry out photosynthe- sis, need relatively high quantities of iron to satisfy the demands of special- ized enzymes. Satellite sensors have Distributions of SST (A) concentrations of nitrate (B) and aerosol optical thickness (AOT) been used to estimate the atmospheric (C) in the global ocean. (D) A true color SeaWiFS image of a major dust storm off aerosol optical thickness, a measure of northwest Africa. SST and AOD are derived from the NOAA AVHRR satellite, nitrate from the particulate burden in the atmo- the / WOCE data set. sphere, globally (Fig. 6C). The most intense deposits of dust, and hence iron, into the oceans occur in the tropical North At- mation being gleaned from each separate ap- lantic from the Sahel desert of Africa (Fig. 6D). proach. The ocean optics field is moving toward de- ploying more sophisticated sensors that will likely improve our ability to resolve and quan- Bright Prospects amid Looming tify ocean microbes and processes associated Uncertainties with their activity. For instance, the goals of the Earth System Science Pathfinder program, Paralleling development of remote sensors is the which is sponsored by the National Aeronautics deployment of passive and active sensors on and Space Agency (NASA), include an explicit buoys to monitor microbial populations and effort to better understand phytoplankton phys- other characteristics, including nutrient concen- iology and study functional microbial groups trations, pigment spectra, and photosynthetic from space. rates (e.g., by fast-repetition-rate fluorometry). The proposed NASA Physiology LIDAR- Perhaps the most exciting prospect from a mi- Multispectral (PhyLM) mission would charac- crobiological standpoint would be to link re- terize key upper-ocean carbon cycle compo- mote sensing with other key technology devel- nents, including phytoplankton carbon biomass opments affecting marine molecular ecology. and physiology. This mission would use a mul- For example, soon instruments will be deployed tispectral sensor capable of measuring up to 40 on buoys capable of detecting specific organ- wavebands in the , visible, and near isms, genes, their expression, and metabolic path- . This instrument thus could measure ways. Then, linking “sea truth” data from such phytoplankton fluorescence for a biomass index sites with information being gathered by satel- in regions that are too optically complex for lites will greatly enhance the value of the infor- inversion algorithms. This mission would also

Volume 71, Number 4, 2005 / ASM News Y 185 be expected to determine particle between their industrial and natural sources. from which phytoplankton carbon, particle Aura also measures vertical distributions of abundance, and particle size spectra can be esti- greenhouse gases such as methane, water vapor, mated. Elsewhere, there is an increasing focus and ozone that will provide insights about their on remote sensing of the coastal ocean, with production by microbial and physical processes plans to use geostationary satellites and hyper- in tropical regions. The Orbiting Car- spectral sensors. A geostationary satellite can scan bon Observatory, a satellite mission planned for the coast of the United States several times a day, launch in 2007, will measure carbon dioxide, allowing us to understand tidal effects and trans- another greenhouse gas that is affected by mi- port and diurnal changes, and would enable us to crobial and human activity. reduce interference from coastal convective Although NASA has future plans for ocean clouds. color sensors aboard satellites, there are no bud- The upper troposphere and lower strato- geted science follow-on ocean color missions to sphere are regions of study for instruments on the Earth Observing System. This lapse may be NASA’s Aura satellite, one of a triptych of due, in part, to scientists not effectively explain- Earth-observing missions. Those instruments ing their needs for future missions. Because re- measure materials affecting air quality, includ- mote sensing is now a critical resource for track- ing carbon monoxide, nitrogen dioxide, sulfur ing ecosystem dynamics, any lapses in key dioxide, ozone, and aerosol particulates, pro- satellite assets will greatly constrain our ability viding global maps that can help to distinguish to detect and document these changes.

ACKNOWLEDGMENTS We thank Ed Carpenter for discussions over the years and Claire Mahaffey for her thoughtful comments on the manuscript. Norman Kuring of NASA Goddard Space Flight Center produced the beautiful satellite images (Fig. 2A, 3B, and 5A). The photographs in Fig. 3C, 4A, and 4B are from Peter Franks (Scripps Institution of ), Kaori Ohki (Fukui Prefectural University), and Hans Paerl (University of North Carolina), respectively, with permission. We thank the SeaWiFS Project (Code 970.2) and the Goddard Earth Sciences Data and Information Services Center/Distributed Active Archive Center (Code 902) at the Goddard Space Flight Center, Greenbelt, MD 20771, for the production and distribution of the SeaWiFS data. We also acknowledge the Ocean Sciences division of NSF and the Ocean Biology and Biogeochemistry Program at NASA for sustained financial support. This is LDEO contribution number 6748.

SUGGESTED READING Behrenfeld, M. J., E. Boss, D. A. Siegel, and D. M. Shea. 2005. Carbon-based ocean productivity and phytoplankton physiology from space. Global Biogeochem. Cycles 19:GB100610.1029/2004GB002299. Boyd, P., et al. 2000. A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization. Nature 407:695–702. Capone, D. G. 2001. Marine nitrogen fixation: what’s the fuss? Curr. Opin. Microbiol. 4:341–348. Coles, V. J., C. Wilson, and R. R. Hood. 2004. Remote sensing of new production fuelled by nitrogen fixation. Geophys. Res. Lett. 31:10.1029/2003GL019018. Hood, R., S. Subramaniam, L. May, E. Carpenter, and D. Capone. 2002. Remote estimation of nitrogen fixation by Trichodesmium. Deep-Sea Res. Part II 49(1–3):123–147. Iglesias-Rodriguez, D. M., C. W. Brown, S. C. Doney, J. Kleypas, D. Kolber, Z. Kolber, P. K. Hayes, and P. G. Falkowski. 2002. Representing key phytoplankton functional groups in ocean carbon cycle models: Coccolithophorids. Global Biogeo- chemical Cycles 16(4):1100. Lobitz, B., L. Beck, A. Huq, B. Wood, G. Fuchs, A. S. G. Faruque, and R. Colwell. 2000. Climate and infectious disease: Use of remote sensing for detection of Vibrio cholerae by indirect measurement. Proc. Natl. Acad. Sci. USA 97:1438–1443. Moran, M. A., W. M. Sheldon, Jr., and R. G. Zepp. 2000. Carbon loss and optical property changes during long-term photochemical and biological degradation of estuarine dissolved organic matter. Limnol. Oceanogr. 45:1254–1264. Sathyendranath, S., L. Watts, E. Devred, T. Platt, C. Caverhill, and H. Maass. 2004. Discrimination of diatoms from other phytoplankton using ocean-colour data. Marine Ecol. Progress Ser. 272:59–68. Subramaniam, A., C. W. Brown, R. R. Hood, E. J. Carpenter, and D. G. Capone. 2002. Detecting Trichodesmium blooms in SeaWiFS imagery. Res. Part II 49 (1-3):107–121. Subramaniam, A., S. Kratzer, E. J. Carpenter and E. So¨ derba¨ck. 2000. Remote sensing and optical in-water measurements of a cyanobacteria bloom in the Baltic Sea. Sixth International Conference on Remote Sensing for Marine and Coastal Environments, Charleston, SC, Veridian ERIM International.

186 Y ASM News / Volume 71, Number 4, 2005