Monitoring Coral Reefs from Space

Coral reefs are one of the world’s most biologically diverse and productive ecosystems. However, these valuable resources are highly threatened by human activities. Satellite remotely sensed observations enhance our understanding of coral reefs and some of the threats facing them by providing global spatial and time-series data on reef habitats and the environmental conditions influencing them in near-real time. This review highlights many of the ways in which satellites are currently used to monitor coral reefs and their threats, and provides a look toward future needs and capabilities.

. Rising upper ocean temperatures, as indicated by sea surface temperature (SST), have increased the frequency and intensity of widespread thermal stress events that can cause mass coral bleaching (Eakin et al., in press), and this bleaching is expected to continue into the future (Hoegh-Guldberg et al., 2007, 2008).Rising SST can also result in an increase in infectious disease outbreaks

INTroducTIoN
Coral reefs are one of the world's most biologically diverse and productive ecosystems (Porter and Tougas, 2001).
They provide abundant ecological goods and services (Moberg and Folke, 1999) and       (Bruno et al., 2007) (Liu et al., 2006) from nighttime SSTs (Figure 1a).Over 20 years of NOAA SST data have been reprocessed to produce a consistent data set at 4-km spatial resolution (Kilpatrick et al., 2001).Analysis of the 22-year Pathfinder data set suggests that the water temperatures of most coral reef areas have been increasing at 0.2-0.4°Cper decade (Strong et al., 2009).This warming agrees with studies that indicate corals may need to acclimate their thermal tolerance by 0.2-1.0°C per decade to survive repeated coral bleaching events predicted under future climate scenarios (Donner et al., 2005).
The CRW coral bleaching HotSpot, released in 1996 (Strong et al., 1997) and made operational in 2002 (Liu et al., 2003), was the first coral-specific product developed by NOAA's National Environmental Satellite, Data, and Information Service (NESDIS).Based on the "ocean hot spots" concept introduced by Goreau and Hayes (1994), HotSpots are positive temperature anomalies that exceed the maximum monthly mean (MMM) SST climatology for each 0.5° pixel, thereby identifying currently thermally stressed regions.
These HotSpots are accumulated over a moving 12-week window to produce CRW's Degree Heating Week (DHW) index that is highly predictive of bleaching occurrence and severity.
Significant coral bleaching is expected to occur one to three weeks after reefs begin for identifying resilient reef areas (Maina et al., 2008).

Solar Radiation and Coral Bleaching
A variety of projects over the past decade have examined the use of geostationary satellite data to study aspects of solar radiation over coral reefs (Tovar and Baldasano, 2001;Hansen et al., 2002;Kandirmaz et al., 2004) Figure 2 shows a mock-up of the product that CRW is developing.

Ocean Acidification
As human activities continue to increase carbon dioxide in the atmosphere (CO 2,atm ), much of this CO 2 is absorbed into surface ocean waters.Surface ocean CO 2 (CO 2,aq ) then reacts with water to form carbonic acid, thereby lowering pH; hence, the term "ocean acidification" (Caldeira and Wickett, 2003).This process also consumes carbonate ions, reducing the degree of saturation (Ω sp ) of seawater with respect to calcium carbonate.This may result in reduced coral growth rates and compromise the structural integrity of coral reefs (Manzello et al., 2008;Veron, 2008;Gledhill et al., 2009) (Robinson, 2004;Martin, 2004;Lagerloef et al., 2008) These data are used to study reef carbonate dynamics (Moses et al., 2009), bathymetry in coral reef areas (Hogrefe et al., 2008), geomorphology and biodiversity of coral reefs (Andréfouët and Guzman, 2005;Knudby et al., 2007), and coral reef fish species richness, communities, and fisheries (Mellin et al., 2009;Hamel and Andréfouët, 2010).
They also allowed managers to more effectively integrate in situ and remotely sensed data (Scopélitis et al., 2010).
Given the breadth of literature on this topic, this section will only provide a survey of some important applications, including marine protected area (MPA) management, habitat characterization, and some important biological parameters of coral reefs.

habitat characterization
Mapping has proven to be a valuable tool for understanding the interconnections between coral reefs and associated habitats, such as the role mangroves play as juvenile reef fish nurseries (Mumby, 2006) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008  High-spatial-resolution mapping of mangrove and seagrass species has been explored (Myint et al., 2008;Phinn et al., 2008) and regional seagrass mapping using Landsat data has been assessed for broader applications across the Caribbean (Wabnitz et al., 2008).
A wide variety of remote-sensing platforms and algorithms are available for habitat mapping.A diverse literature exists on habitat mapping using satellite sensors such as Satellite Pour l'Observation de la Terre (SPOT) and Landsat Thematic Mapper (TM) at a spatial resolution of tens of meters (LeDrew et al., 2000), and on using highspatial-resolution sensors such as the commercial Earth-observation satellites IKONOS and QuickBird (Mumby and Edwards, 2002;Andréfouët et al., 2003;Mishra et al., 2006).Even greater accuracies can be obtained by including data from airborne sensors like the Compact Airborne Spectrographic Imager (CASI) either as the prime data source (Mumby et al., 1998;Bertels et al., 2008) or in combination with satellite data (Rowlands et al., 2008).
Perhaps the most ambitious project to characterize the habitats of coral reefs worldwide was the Millennium Coral Reef Mapping Project (MCRMP) (Andréfouët et al., 2002), which used Landsat data to identify geomorphological characteristics of coral reefs on a global scale.Data obtained in this effort have already informed numerous studies on strategic MPA placement, reef condition assessments (Burke and Maidens, 2004), and both climate forcing of reef growth on the Great Barrier Reef and climate change influences on biogeochemical budgets in French Polynesian atolls (Andréfouët et al., 2006).
Seafloor bathymetry and benthic rugosity are important variables to coral reef researchers and managers alike.

Pseudo-bathymetry has been derived
from IKONOS data in a number of studies (Lyzenga et al., 2006), modified by applying nonlinear inversion models (Su et al., 2008), and used to fill the gap between terrestrial digital elevation models (DEMs) and sonar-acquired bathymetry (Hogrefe et al., 2008).In other studies using light detection and ranging (lidar), rugosity maps were developed to illustrate massive stony coral colonies on patch reefs (Brock et al., 2006) and to investigate statistical relationships between rugosity and reef fish communities (Kuffner et al., 2007).
Other studies combined bathymetry with habitat maps derived from remotesensing satellites to drive models that predict coral reef fish distribution (Pittman et al., 2007;Knudby et al., 2008;Knudby et al., 2010).Mellin et al. (2009) advocated a "hierarchy of habitat" for predicting coral reef fish habitats using geomorphology, benthic assemblages, rugosity, and depth as key habitat variables.

biological parameters of coral reef ecosystems
Coral reef cover is a parameter of great interest to researchers and managers.To Researchers either interpret these data directly (Lesser and Mobley, 2007) or extract characteristic spectral derivatives from them (Holden and LeDrew, 1998) to classify live and recently dead coral (Mumby et al., 2004(Mumby et al., , 2001)).For these studies, it is important that the reefs have a clear-water environment and are devoid of brown macroalgae, which are frequently spectrally confused with coral (Hedley and Mumby, 2002).The wider efficacy of these methods for evaluating coral cover is still under investigation (Hedley et al., 2004;Joyce et al., 2004).A third approach to mapping corals has used sonar to discriminate branching corals, such as Acropora spp., from other reef structures (Purkis et al., 2006).
Mass coral bleaching has been successfully detected using remotely sensed observations in cases where the extent of bleaching is pronounced in high-coral-cover reef habitats; Figure 6).
High spatial resolution is key to reliable remote detection of coral bleaching (Andréfouët et al., 2002).Elvidge et al.They would like to see hypoxic or "black water" events that kill fish.These    Stations" available globally, this graph displays the current SST (purple line), the monthly average temperature or "climatology" (blue plus signs), the maximum monthly mean (blue dotted line), and the bleaching threshold (blue solid line).When temperatures reach or exceed the bleaching threshold, thermal stress is accumulated as degree heating weeks (DHWs) and displayed using the right-hand axis (red solid line) with 4°C-and 8°C-weeks marked for reference (red dotted line).A Bleaching Watch is issued when the Hotspot is greater than 0°C but less than 1°C, and a Bleaching Warning is issued when the Hotspot is greater than 1°C-and the DHW less than 4°C-weeks.An Alert Level 1 is declared when DHWs are between 4°C-and 8°C-weeks, and an Alert Level 2 when DHWs are at or above 8°C-weeks.
Coral reefs are one of the world's most biologically diverse and productive ecosystems.However, these valuable resources are highly threatened by human activities.Satellite remotely sensed observations enhance our understanding of coral reefs and some of the threats facing them by providing global spatial and time-series data on reef habitats and the environmental conditions influencing them in near-real time.This review highlights many of the ways in which satellites are currently used to monitor coral reefs and their threats, and provides a look toward future needs and capabilities.This nadir true-color image of australia's great barrier reef was acquired by the multi-angle Imaging Spectroradiometer (mISr) instrument on august 26, 2000, and shows part of the southern portion of the reef adjacent to the central Queensland coast.The width of the mISr swath is approximately 380 km, with the reef clearly visible up to approximately 200 km from the coast.Image courtesy of NASA/GSFC/LaRC/JPL, MISR Team b y c .m a r k e a k I N , c a r l J .N I m , r u S S e l l e .b r a I N a r d , c h r I S T o p h a u b r e c h T, c h r I S e lV I d g e , d w I g h T k .g l e d h I l l , F r a N k m u l l e r -k a r g e r , p e T e r J .m u m b y, w I l l I a m J .S k I r V I N g , a l a N e .S T r o N g , m e N g h u a wa N g , S c a r l a w e e k S , F r a N k w e N T z , a N d d a N I e l z I S k I N Oceanography | Vol.23, No.4 reviews of satellite uses, methods, and applications for coral reefs, including discussions about using remote sensing to monitor key physical parameters that influence the conditions of coral reefs.moNITorINg cor al reeF eNVIroNmeNTal coNdITIoNS Coral reefs are exposed to a number of stressors from their surroundings.Although most satellites were not designed to observe coral reefs, many remote-sensing instruments provide valuable environmental data that are relevant to reef conditions.Calibration and validation data have been used to strengthen these data in shallow, nearshore environments, allowing coral reef managers and researchers to use this information to both understand the role of stressors and to better manage coral reef resources.physical parameters With the growing consequences of global climate change, there is greater need for monitoring impacts in coral reef areas.Increasing anthropogenic concentrations of atmospheric CO 2 have numerous direct and indirect deleterious impacts on coral reefs.Carbon dioxide imparts an important control on the radiative heat balance of Earth's atmosphere, resulting in warming of both the atmosphere and the ocean (IPCC, are central to the socio-economic and cultural welfare of coastal and island communities throughout tropical and subtropical oceans by contributing at least $30 billion (US$) to the global economy when combined with tourism and recreation, shoreline protection, fisheries, and biodiversity services (UNEP-WCMC, 2006).Unfortunately, a range of human activities adversely impacts these valuable resources.Among the key threats are improper fishing activities, land-based sources of pollution, climate change, ocean acidification, and habitat destruction (Dodge et al., 2008; NOAA Coral Reef Conservation Program, 2009).Satellites have the capacity to enhance our understanding of coral reef threats by obtaining global information on environmental conditions in near-real time and by providing spatial and timeseries data relevant to management that are not practically obtained by in situ observations alone.This paper highlights the various ways remote-sensing data are being used to map and monitor coral reefs.It explains some remote-sensing tools commonly used to measure coral reef parameters of interest, how this information aids coral reef managers, some of the limitations of current technologies, and research gaps.

Figure 1 .
Figure 1.Noaa coral reef watch near-real time satellite global 50-km nighttime product suite for February 9, 2009: (a) sea surface temperatures (SST), (b) coral bleaching alert area product that combines hotSpot and degree heating week products.
. In 2009, NOAA developed a suite of experimental surface solar radiation products based on data from Geostationary Operational Environmental Satellite (GOES) systems that were calibrated and validated to provide insolation estimates over oceanic waters (http://www.osdpd.noaa.gov/ml/land/gsip).These products measure total daily global ocean surface insolation.These data have allowed CRW and its partners to develop a new set of coral bleaching products that endeavour to provide a combined measure of satellitederived thermal stress and light as an index of stress on coral photosystems.
photo-degradation of colored dissolved

Figure 2 .
Figure 2. prototype of the new Noaa light Stress damage product combining thermal and light stress as a bleaching predictor.
2007; http://coralreefwatch.noaa.gov/satellite/doldrums).Analyses of long-term patterns of winds may provide additional insights into conditions around coral reefs.A recent analysis of satellite Special Sensor Microwave Imager (SSM/I) observations suggests that precipitation and total atmospheric water have increased equally at a rate of ~ 1% per decade over the past two decades (Wentz et al., 2007).A least-squares linear fit of SSM/I wind speed for each 2.5° grid cell was calculated after removing the seasonal variability to provide a decadal trend map of wind speed (Figure 4) and compared with wind trends from the International Comprehensive Ocean-Atmosphere Data Set (ICOADS) from ship-based observations.Winds averaged from 30°S to 30°N increased by 0.04 m s -1 (0.6% per decade), resulting in wet areas becoming wetter (Wentz et al., 2007).Wentz et al. (2007) suggest that two decades may be too short for extrapolating these short-term trends into longer ones; however, continuation of these trends could significantly impact circulation, mixing, and temperature regulation of waters around reefs in the Indo-West Pacific, the heart of coral biodiversity and abundance.chemical parameters Though corals predominantly occur in clear, oligotrophic waters, nutrientrich runoff from land-based sources of pollution (LBSP) often results in algal overgrowth of corals and coral recruits (Adey, 2000).Phosphate pollution can inhibit calcium carbonate (CaCO 3 ) deposition, slowing coral growth (Muller-Parker and D'Elia,

Figure 3 .
Figure 3. Noaa/NeSdIS coral reef watch near-real-time satellite 25-km doldrums experimental product for april 22, 2007, in the Indian ocean region.The color scale indicates the number of days over which the daily mean National climatic data center blended Sea winds remained below 3 m s -1 .
. As mentioned above, salinity at this resolution may be helpful in driving models to provide downscaled salinity to improve our monitoring of alkalinity and, thus, ocean carbon chemistry.It may also help provide information on increased runoff from large rivers as well as production of warm hypersaline waters over extensive shallow banks.Coastal Development Although current satellites cannot identify complex chemical pollutants, it is possible to use satellite observations of human activity to provide a proxy for these other stressors.Understanding the extent of urbanization and human activity can provide a proxy for localized impacts such as pollutants, runoff, fishing, and recreational use of reefs.A prominent indicator of human occupation visible from space is artificial night lighting.Observable features have included biomass burning, massive offshore fisheries, and infrastructure lights.NOAA has processed nighttime lights data acquired by the US Air Force Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) (Aubrecht et al., 2008) in order to integrate the brightness and distance of lights near known coral reef sites as a proxy for human development stresses.Light proximity index (LPI) values have been calculated for the three stressors observable in nighttime lights: human infrastructure, gas flares, and heavily lit fishing boats.The algorithm computes the contribution of lights declining with increased distance from reefs.LPI since 1992 shows increased human settlements near coral reefs (Figure 5), a reflection of the expanding populations and infrastructural development seen in coastal areas in many parts of the world.In contrast, lit fishing boats activity has declined and gas flares near reefs show a more complex pattern with dips in 1994 and 2001, and a peak in 1997.LPI can provide a valuable tool for coastal management, especially in areas where other data on coastal development may be limited.maNagemeNT applIcaTIoNS Measurement of environmental parameters is helpful to coral reef managers because it alerts them of conditions when coral reefs in their jurisdictions are in need of direct monitoring or management action.There are times, however, when in situ data are not sufficient for monitoring, or reef locations are too remote to gather in situ data.In these cases, high-resolution remote-sensing data can provide researchers and resource managers with needed data.The availability of high-resolution satellite data has increased in recent years due to launching of more advanced sensors, application of emerging technologies in sensor designs, and new algorithms.
Analyses of coral reefs using highspatial-resolution (30-m pixel size or less) data span a broad scale of applications, including MPA design and evaluation(Green et al., 2009;Dalleau et al., 2010), study of ecosystem associations, (e.g., coral reefs with seagrass beds and mangroves), and investigations of the ecology of coral reefs and the organisms that live in them (e.g., fish).Although the bulk of these projects are conducted at local scales, there are calls to automate these techniques for use at broader scales(Andréfouët, 2008).NOAA's analysis of MPA characteristics and design in the northwestern Hawaiian Islands and in Puerto Rico has used remotely sensed data in conjunction with in situ measurements of coral habitat and spectral reflectance characteristics to produce benthic habitat maps for shallow-water coral reefs.These baseline habitat maps provide resource managers and researchers with essential information for planning MPAs, monitoring changes, and evaluating MPA effectiveness in both the Pacific(Friedlander et al., 2007(Friedlander et al., , 2008) ) and the Caribbean(García-Sais et al., 2008).

Figure 5
Figure 5. light proximity index (lpI) percent change over time (1992-2009) from human settlements, lit fishing boats, and offshore gas flares for the reefs of the world.lpI values are normalized by absolute number of reef points to show percent change over time.colors correspond with stressors and point shapes with the satellite used.
date, there have been three approaches to quantifying changes in the benthic composition of coral reefs using coral and macroalgae cover as key indictors of reef health.A number of scientists have used time-series data to detect changes in overall reflectance that can be attributed to major changes in benthic state.For example, Dustan et al. (2001) analyzed a Landsat time series from Florida and found a change in "temporal texture" associated with the die-off of long-spined sea urchins.These indirect methods are useful, but the data have been difficult to interpret because multiple changes in benthic characteristics may have similar effects.A second approach uses high-resolution optical data to measure changes in reef spectra.

"
SaTellITe TechNologIeS haVe become eSSeNTIal ToolS For moNITorINg coral reeF healTh aNd The INcreaSINg ThreaTS coral reeFS Face arouNd The globe."

(
2004) developed a method for detecting the brightening of reef areas when corals bleach using pairs of high-resolution multispectral images.The technique involves detection of spectral brightening observed when comparing satellite images acquired before the bleaching event with images collected during the bleaching event.This technique has continued to be refined and applied to multiple satellite sensors (Daniel Ziskin, University of Colorado at Boulder, pers.comm., September 27, 2010).However, images like those in Figure 7 only provide a general index of reef bleaching and cannot separate bleaching of corals from bleaching of crustose coralline algae in reef systems, limiting their value for reefs where the community structure is poorly known.Additionally, as long as these approaches require the purchase of costly satellite imagery, they will only be practical for remote areas that cannot be reached by divers.The FuTure oF moNITorINg cor al reeFS From Space Improvements in three major areas are increasing the potential for using remotely sensed data to monitor coral reefs: resolution, spectral bands, and algorithms.Whether it is remote sensing of stressors, environmental parameters, or habitats and their changes, increased spatial and temporal resolution provide new ways that coral reefs can be monitored.Unfortunately, there is often a trade-off between spatial and temporal resolutions, as finer-resolution imagery is frequently obtained through narrower swaths and a corresponding increase in time between repeated observations.However, improved technology, such as off-nadir viewing, and image processing are helping to resolve this issue.Higher-resolution applications also require a larger data set of in situ observations for calibrating and validating changes, especially in highly variable nearshore environments.At the same time, a greater portion of the electromagnetic spectrum is being used in remote sensing.Enhanced optical systems, such as multispectral to hyperspectral sensors covering more of the visible and nearby bands, will be used to detect changes in benthic habitats and will do a better job of separating benthic features from watercolumn properties.Because coral reefs normally exist in shallow, clear waters, many optical bands can "see" the bottom.This attribute is good for habitat measurements, but it creates problems for separating changes in the benthos from changes in the water column.Both enhanced spatial and spectral resolution are essential to resolving these problems and will provide the quantitative data needed to answer many resource-monitoring challenges that currently exist for coral reef ecosystems.None of these improvements are sufficient, however, without considerable work to develop, calibrate, and validate new and improved algorithms.Because satellites can only measure a small set of parameters directly, most of what we monitor from satellites is the result of complex algorithms that derive the parameters of interest.This is also where the combination of multiple instruments and spectral bands show great promise in enhancing our

Figure 6 .
Figure 6.Image differencing, or spectral brightening, of Nikunau Island, kiribati.The difference image (far right) shows a gold color where presumably bleaching has occurred.
managers would also like to know when changes in water quality cause harmful algal blooms or the spread of bacterial mats across the seafloor.They would like to know when diseases are spreading through ecosystems.Some of these capabilities are likely to be available over time scales of a few years while others will likely take decades and require significant advances in the development of new sensors and algorithms.Some of these needs will be answered by advancing the use of remotely sensed and in situ data to initiate and support numerical models.These models estimate environmental parameters (as was done for ocean acidification) or impacts, such as changes in community composition.In many cases, these models integrate multiple data sets to provide new estimates of unmeasured parameters.These estimates may serve as solutions to scientific questions that inform management needs or may be a bridge until new sensors are available.The development and launch of new satellite sensors is a slow process, and agencies involved need to better incorporate the needs of resource managers into their instrument and satellite development processes.At the same time, we need to identify key parameters that must be collected with high quality and continuously to understand long-term changes to coral reefs and the water quality and oceanographic conditions influencing them.Coral reefs are important and valuable ecosystems.Satellite technologies have become essential tools for monitoring coral reef health and the increasing threats coral reefs face around the globe.We need to continue to advance the tools available and the science behind them in order to make these tools as useful as possible.As the anthropogenic threats to coral reefs continue to mount, we need to continue our diligent use of all tools at our disposal to keep these valuable resources healthy.ackNowledgemeNTS Thanks to the many colleagues and collaborators who have contributed to the development of satellite remotesensing technology and its application to coral reef ecosystems.In particular, we thank Serge Andréfouët and additional reviewers for their thoughtful evaluation of manuscripts.We are also grateful to the funding agencies and organizations that have supported the work, including the National Oceanic and Atmospheric Administration, the National Aeronautics and Space Administration, the Australian Research Council, and the World Bank/Global Environment Facility.The manuscript contents are a cl aSSroom acTIVIT y uSINg SaTellITe Sea SurFace Temper aTureS To predIcT cor al bleachINg Don't miss the related Hands-On Oceanography activity from oceanography 22-2.This activity illustrates how temperature influences coral bleaching and how remote sensing is used to monitor coral health.It can be presented with different levels of complexity to upper middle school through early college students.Download the full activity from hT Tp://www.ToS.org/haNdS-oNOceanography Vol.22, No.2 254 Oceanography June 2009 255 Remote Sensing and Sea Surface Temperature Remote sensing is the technique of measuring a property of an object without touching it.Each day, NOAA uses satellite remote sensing to monitor characteristics of Earth's surface from space.The Advanced Very High Resolution Radiometer (AVHRR) sensor on NOAA satellites measures SST by detecting the infrared radiation given off by the ocean surface.Infrared radiation is of lower energy than visible radiation on the electromagnetic spectrum and, although it cannot be seen, it can be felt as heat.For example, it is infrared radiation that you feel as heat coming off embers or a hot stove even when there is no visible glow.You feel infrared radiation on a sunny day.SST varies daily, seasonally, and among different locations.Measuring it is an important component of weather forecasting, climate prediction, and managing natural resources; we can also use it to observe and understand the conditions around coral reefs.NOAA Coral Reef Watch Products NOAA Coral Reef Watch monitors satellite measurements of global-ocean SST to predict areas where coral bleaching might occur.SST data from NOAA polar-orbiting satellites are presented in 0.5-degree (~ 50 km) pixels, twice per week, and display nighttime temperature calibrated against patterns in buoy data at 1-m depth (Figure 1A).Thermal stress occurs when corals are exposed to temperatures warmer than their usual range.For each month and for each pixel, a seven-year mean SST or "climatology" was calculated from historical satellite data to provide the "usual" conditions for each location through the annual cycle.The SST anomaly is the difference between the measured SST and the climatology for that time of year (Figure 1B).Areas in purple and blue are cooler than normal, while areas in yellow and orange are warmer than normal.It is important to recognize that most corals do not live right at the ocean surface where satellites measures temperature.Although the corals often experience different temperatures from the sea surface, the nighttime temperature anomaly is generally consistent from the surface through the range of depths in which coral reefs are found (to 100 m).As such, the nighttime SST anomaly is an effective measure of whether the conditions experienced by corals are "normal" or not.The Coral Bleaching Hotspot and Degree Heating Weeks (DHW) products are derived from SST to specifically pinpoint coral bleaching.The Hotspot product shows areas where the current SST is above the average temperature of the warmest month for each pixel (Figure 2A).When the Hotspot reaches the bleaching threshold value of 1°C, the temperatures in that region are high enough to cause coral bleaching (Hoegh-Guldberg, 1999; Berkelmans, 2002), shown as yellows and oranges in the Hotspot image.Widespread bleaching occurs when temperatures get hot and stay hot.The DHW product (Figure 2B) measures accumulating thermal stress over the past 12 weeks by summing any Hotspots at or above 1°C and expressing them in units of °C-weeks.A DHW of 4°C-weeks represents enough accumulated thermal stress to cause ecologically significant bleaching, and a DHW of 8°C-weeks indicates that widespread bleaching and mortality are likely.Satellite data are summarized at 190 "Virtual Stations" around the world (http://coralreefwatch.noaa.gov/satellite/current/experimental_products.html)that simulate data-collecting buoys in the water.For each station (such as the one pictured in Figure 3), the purple line shows SST time series, the blue plus signs (+) are the monthly climatology values (the historical monthly means), the dotted blue line represents the warmest monthly mean temperature found in the climatology for that pixel, and the solid blue line is the bleaching threshold temperature.The DHW data are shown by a solid red line, relative to the right-hand axis, with the key values of 4°C weeks and 8°C weeks indicated

Figure 1 .
Figure 1.Global sea surface temperature (SST) and SST Anomaly for September 2, 2005.These two NOAA Coral Reef Watch products are released twice weekly.Global nighttime satellite SST (A) shows warmer waters near the equator and cooler waters near the poles; the global SST Anomaly product (B) indicates where temperatures are warmer or cooler than normal for the same time period.

Figure 2 .
Figure 2. Western Hemisphere Coral Bleaching Hotspot and Degree Heating Weeks for September 2, 2005.Two NOAA Coral Reef Watch products are derived from SST to specifically pinpoint coral bleaching: the Coral Bleaching Hotspot product (A) indicates where SSTs are warmer than the warmest month climatology and if the bleaching threshold for each location is exceeded, while the Degree Heating Weeks product (B) accumulates bleaching-level thermal stress over the previous 12 weeks.

Figure 3 .
Figure 3.Time Series Graph for Lee Stocking Island in the Bahamas.One of 190 "VirtualStations" available globally, this graph displays the current SST (purple line), the monthly average temperature or "climatology" (blue plus signs), the maximum monthly mean (blue dotted line), and the bleaching threshold (blue solid line).When temperatures reach or exceed the bleaching threshold, thermal stress is accumulated as degree heating weeks (DHWs) and displayed using the right-hand axis (red solid line) with 4°C-and 8°C-weeks marked for reference (red dotted line).A Bleaching Watch is issued when the Hotspot is greater than 0°C but less than 1°C, and a Bleaching Warning is issued when the Hotspot is greater than 1°C-and the DHW less than 4°C-weeks.An Alert Level 1 is declared when DHWs are between 4°C-and 8°C-weeks, and an Alert Level 2 when DHWs are at or above 8°C-weeks.

Figure 4 .
Figure 4.Western Hemisphere Doldrums for September 2, 2005.Using satellite wind data from several satellites, the Coral Reef Watch Doldrums product indicates for how long the daily average wind speed was less than 3 m s-1 (about 7 mph).Turquoise indicates doldrumscondition persistence for the past four days, blues for between four and 12 days, and yellow to orange colors for the most recent two weeks to a month.with dashed red lines.Yellow-to-red colors along the x-axis indicate periods when bleaching alerts were issued.NOAA Coral Reef Watch is developing additional satellite products to monitor other conditions related to a coral bleaching event, such as an extended period of low wind (Doldrums) product that uses data from several satellites to indicate how long the average wind speed was below a critical threshold (3 m s-1, or about 7 mph; Figure4).This information is important

Table 1 .
remote sensing platforms and sensors relevant to the monitoring of coral reefs and associated habitats.Adapted fromMumby et al. (2004)and the Directory of Remote Sensing Applications for Coral Reef Management by the Remote Sensing Working Group of the Global Environment Facility Coral Reef Targeted Research & Capacity Building for Management Program Symbology: 3 indicates well-established, ?3 indicates fairly well-established, ?indicates experimental, 3* indicates data should be used in conjunction with acoustic sonar devices, blank indicates not currently possible.

Table 1 ,
continued.remotesensingplatforms and sensors relevant to the monitoring of coral reefs and associated habitats.Adapted fromMumby et al. (2004)and the Directory of Remote Sensing Applications for Coral Reef Management by the Remote Sensing Working Group of the Global Environment Facility Coral Reef Targeted Research & Capacity Building for Management Program Symbology: 3 indicates well-established, ?3 indicates fairly well-established, ?indicates experimental, 3* indicates data should be used in conjunction with acoustic sonar devices, blank indicates not currently possible.