HAZARDS TO HUMAN HEALTH FROM THE OCEANS
Part I of this report identifies areas where coordinated efforts between the oceanograpic and medical communities will be required to address the risks to human health generated by the oceans and to evaluate the potential consequences of climate change for public health. There are three chapters included in Part I. Chapter 1: ''Climate and Weather, Coastal Hazards, and Public Health," describes how public health is affected by marine processes such as ocean-dependent weather and climate effects, tropical storms, and estuarine and coastal circulation. Chapter 2: "Infectious Diseases," covers the various waterborne marine infectious diseases including bacterial, viral, and protozoal agents of disease. This chapter also examines the effects of weather and climate on vector-borne diseases, such as the increased prevalence of malaria, a disease carried by mosquitoes, following an El Niño event. Finally, in chapter 3: "Harmful Algal Blooms," the various syndromes resulting from exposure to algal toxins are identified and discussed with reference to the ecology and distribution of the specific algae associated with these illnesses.
There have been several recent programs that have highlighted the value of an interdisciplinary approach to the issues described in the following chapters. Brief descriptions of three programs appear in boxes in the appropriate chapter: HEED, in chapter 1; the ENSO Experiment, in Chapter 2; and ECOHAB, in Chapter 3.
There was a problem loading page 16.
Climate and Weather, Coastal Hazards, and Public Health
The Physical Ocean Environment: Circulation and Stratification
Public health policymakers rely on the ocean sciences to help them develop more effective responses to marine hazardous phenomena including tropical cyclones and hurricanes, tsunamis (or long oceanic waves), toxins and pathogens in nearshore and estuarine waters, and ocean-driven weather and climate patterns. These events may either directly cause injury and death or indirectly cause the spread of various types of illness, including waterborne and vector-borne diseases, as well as illnesses associated with toxic algal blooms. The protection of public health requires a thorough understanding of the physical ocean environment for better forecasting and handling of marine disasters. This chapter will describe the major threats to public health and the ocean processes that contribute to them.
Frequently, the ability to anticipate and respond promptly to natural disasters rests on an understanding of weather systems that depend on a complex coupling of the atmosphere, land, and the ocean. The ocean absorbs immense quantities of heat, fresh water, and carbon, and hence act as the "memory" of the atmosphere and land. Although climate is seen more readily by observing the atmosphere, it is the coupled system of the ocean, the atmosphere, and the land masses that determines the evolution of climate over long periods of time. Beyond its role as a reservoir for water, heat, and carbon, the world ocean actively influences the atmosphere, as demonstrated by the El Niño phenomenon. In an El Niño, the normal upwelling of cold water along the Equator fails, and a lens of warm tropical surface water spreads across the eastern Pacific. These warm ocean waters
contribute moisture and energy (in the form of heat) to the atmosphere and bring unusually warm, wet weather to the west coasts of North and South America, and droughts to Australia and southeast Asia.
The ocean also serves as the medium for the culture of phytoplankton, microalgae that produce oxygen as a by-product of photosynthesis and account for about 50% of the earth's primary productivity. Through the consumption of seafood, many human populations depend on this resource. In general, the productivity of the world's oceans, which includes fisheries, is a reflection of the primary productivity of the phytoplankton. However, these phytoplankton also pose public health problems because some species produce toxins that cause various illnesses when they are consumed in seafood, as described in more detail in chapter 3. Hence, it is important to know what properties of the marine environment influence the productivity of phytoplankton.
The continued high productivity of ocean waters is dependent on the renewal of its life-giving resources: oxygen, carbon, minerals, and nutrients. This renewal occurs through the inflow of rivers and stirring of the ocean basins by great global currents that circulate waters from the surface down to the depths of all the major oceans and back to the surface again (Plate I). Without this "overturning" circulation, the surface waters would become depleted of nutrients and the deep waters would become depleted of oxygen, a stagnation observed seasonally in some freshwater lakes and in isolated marine basins and some estuaries.
At the grand scale of the world ocean, the ventilation of the deep water is achieved by cycles of evaporation and precipitation, heat exchange, winds, runoff of fresh water from land, and freezing and thawing of ice. As surface waters are cooled, they become more dense and sink, bringing oxygen-rich water to the deep. This causes stratification, where dense (cold or more saline) waters slip beneath buoyant (warm or less saline) waters. However, this atmospheric-driven sinking occurs only in very narrow, concentrated currents, whereas water rises toward the surface much more broadly. This results in the asymmetric creation of stable, stratified water masses.
Stratification is perhaps the most important physical property of the ocean for life on Earth because it determines the distribution of nutrients and oxygen. Even though the deep waters are only about 0.3% denser than surface waters, this difference is sufficient to segregate water masses and hence the sources and sinks of biological activity. Biological production depends on the input of nutrients into surface waters where there is sufficient sunlight for photosynthesis. Zones of high productivity occur where the nutrients and oxygen enriched deep waters are brought to the surface by physical forces, such as wind-driven upwelling or tides. In some areas, upwelling is driven by the force of the wind on the sea surface, pulling water up from depths of 100 m (328 feet) or more. In coastal waters and estuaries, nutrient input from rivers may be dominant. The lower density of freshwater keeps these nutrients stratified at the surface creating zones of high biological productivity.
Stratification of water masses with different temperatures and salinities also influences climate and local weather systems that in turn affect human health, both through severe storms and through changes in climate that alter the range of agents of infectious disease. This chapter will discuss why physical ocean properties have important implications for public health.
Public Health Problems Caused by Tropical Storms and Other Marine Natural Disasters
Natural disasters involving ocean processes include phenomena such as rain, tropical storms, tsunamis, storm surges, blooms of toxic algae, pathogen contamination of coastal waters, and recurring as well as long-term climate variability. The types of problems faced by the public health system are determined both by geography and the socioeconomic status of the affected country. While the wealthier industrialized nations suffer more economic loss, poorer developing countries often face far greater loss of life, continuing incidence of disease, and longer lasting damage to social and physical structures. The potential for immediate casualties and communicable disease outbreaks often overshadows the more severe and durable long-term impacts on public health, even in developing nations. Both the direct and indirect impacts of marine natural disasters are assessed here with primary consideration given to the impacts of tropical storms, tsunamis, and storm surges. The public health issues arising from the spread of pathogens and harmful algal blooms will be discussed in Chapters 2 and 3, respectively.
Risk is a measurement of the degree of loss (human life, injuries, economic losses, etc.) expected by the occurrence of a disaster. Short- and long-term health consequences are a result of the contributions of many factors, such as:
Direct Impacts on Health
Mortality, the number of deaths caused by natural disasters, is the most common indicator used by the international community to assess the severity of the health impact of a disaster. The number of lives lost provides important statistical data but can be highly misleading in determining the impact on survivors.
High mortality resulting from marine disasters is associated with tsunamis, storm surges, and flash floods resulting from tropical storms with heavy precipitation. For example, a cyclone in the Bay of Bengal in 1970 induced a storm surge causing between 250,000 and 500,000 deaths (Murty et al., 1986; Sommer and Mosley, 1973). The broad range in estimating mortality following the Bangladesh cyclone reflects the lack of reliable data on the consequences of natural disasters in most developing countries. Also, this dramatic loss of life illustrates the high-risk to the population in Bangladesha combination of the frequent occurrence of cyclones and storm surges with an extremely vulnerable and unprepared population. At the other end of the spectrum, accurate forecasting of tropical storms in the Caribbean region and effective evacuation policy and procedure have greatly reduced the loss of lives from recent hurricanes in the United States, although there is still potential for significant loss of life due to tropical storms (Pielke and Pielke, 1997).
The devastating 1998 tropical storm season in the Caribbean illustrates the disproportionate effects of hurricanes on developing and industrialized nations. Hurricane Georges, which struck the Caribbean and U.S. Gulf Coast in September of 1998, was responsible for an estimated 210 deaths in the Dominican Republic while causing fewer than 10 fatalities in the U.S. and Cuba (AP, 1998a).
At the end of October 1998, Hurricane Mitch brought tragedy to Central America when the storm stalled over the coast of Honduras and dropped torrential rains in the highland and coastal regions for several days. The rains caused catastrophic floods and landslides throughout the region, with Honduras suffering the heaviest losses. In Honduras, Nicaragua, Guatemala, and El Salvador there were an estimated 9000 deaths with another 9200 people reported missing. More than half of the casualties occurred in Honduras, where approximately 12,000 people were injured and 1.5 million were affected by the storm and its aftermath (USAID, 1998; UN, 1998). The devastation brought by Hurricane Mitch was not a result of poor prediction, but instead illustrates how natural disasters can result when poverty drives the development and deforestation of vulnerable areas (Copley, 1998; LaFranchi, 1998).
Tsunamis are long oceanic waves caused by earthquakes that displace the seabed. Although tsunamis occur less frequently than tropical storms, some predictions allow vulnerable communities to be warned in advance. As with a storm surge, mortality is high in low-lying coastal areas. On July 17, 1998, a shallow earthquake near the coast of Papua, New Guinea drove a wave onshore, inundating a strip of heavily populated shoreline. The first wave approached 30 feet in height and arrived 9 minutes after the earthquake (Plate II). The sudden influx of water resulted in more than 2000 deaths. Because the earthquake was so close to shore, no form of prediction and warning system could have prevented this loss of life. In many other cases, earthquake epicenters are far from vulnerable shores and warnings are effective. Elaborate warning networks are in place on many Pacific islands and around the Pacific rim.
The human health risks from the Papua, New Guinea tsunami situation were so great that officials declared a state of emergency on the Sissano coast and sealed off an area of about 120 square kilometers (45 square miles) around the lagoon. The devastation of this area forced crowding of displaced residents onto higher land. Such crowding in a wet environment, along with disruptions in the supply of potable water, favors the spread of infectious diseases such as pneumonia, cholera, and malaria. In addition, most of the injuries suffered by survivors were open wounds as a result of the physical force of the tsunami. The greatest danger to the injured therefore was infection because of limited medical care in the immediate aftermath of the disaster.
Although epidemics of waterborne (diarrheal diseases including cholera and typhoid fever) or vector-borne diseases (such as dengue fever and malaria) are a major concern, remarkably few major outbreaks have been scientifically documented in the literature following such natural disasters. Typically, in the Bay of Bengal, storms surges may cause greater problems from the salination of wells and agricultural lands than the contamination of water with pathogens. The absence of anticipated disease outbreaks may reflect several factors. First, water-borne diseases are highly preventable through public and individual environmental health measures. The very fear of devastating outbreaks is an effective incentive to improve otherwise neglected basic sanitation and water control in many countries. Second, dilution of fecal contamination by tropical storm surges in overcrowded and heavily contaminated environments may reduce outbreaks. Finally, the health indicators of the surviving population, in some instances, appears to improve in the case of a disaster such as a storm surge because the death toll is highest among the elderly, children, and the sickgroups with the greatest health problems (Chen, 1973; Sommer and Mosley, 1972).
In the case of vector-borne diseases, tropical storms, floods, and storm surges may either suppress or promote the breeding of the vector and its pathogen. Initially, the influx of water may disrupt insect vector breeding sites and decrease the rodent vector population. Later, however, breeding sites for mosquitoes (the vector for malaria, dengue, and yellow fever), while initially washed away by the
floods or storm surge, increase with the use of residential water pools that build the potential for increased transmission. For example in 1963, Hurricane Flora struck Haiti shortly after the completion of insecticide spraying of the dwellings by the malaria eradication program. The proliferation of breeding sites combined with a disruption of routine control measures resulted in one of the best-documented hurricane-caused outbreaks of malaria, (Mason and Cavalie, 1965).
Indirect Impacts on Public Health
The pursuit of public health encompasses much more than the provision of medical care or the control of communicable diseases. In the constitution establishing the World Health Organization (WHO), health is defined as "a state of physical, mental, and social well-being and not only the absence of disease or infirmity" (WHO, 1946).
The delayed or indirect health impact of marine natural disasters is generally underestimated and under-reported. The long term cost to public health results from the interruption of health services, the permanent damage to infrastructure, the setback in development, and the loss of individual income. In the developing countries, electrical power and potable water shortages or rationing are daily occurrences. Following an ocean-borne disaster, the lack of electrical power (and transportation) has profound and far reaching public health consequences, affecting the operations of hospitals, water plants, and health facilities, as well as degrading the quality of the local environment. As summarized below, loss of these capabilities has the potential to affect public health more profoundly than the immediate impact of an event such as a storm surge.
Disruption of health services: Following a disaster in developing nations, health services experience a decreased ability to respond to normal demands for medical care. Hurricane Gilbert in Jamaica (1988) left a modest toll of 45 people dead, but twenty-two hospitals or health centers were out of service for an extended period of time and 90% of the hospital bed capacity was unavailable for several days to several weeks (Table 1-1; PAHO, 1988; Zeballos, 1993).
Setback in development: The economic impact of natural disasters at a national level is amplified in the health sector for several reasons. Existing resources (medicines and disposable equipment, budget, personnel) are diverted from routine medical care and disease control programs for immediate response to the perceived threats to public health. International assistance, however generous, rarely represents a significant proportion of the material emergency contribution and does not subsidize the future provision of routine care and disease control.
Loss of individual income: Even if the economy of the developing country is not significantly affected by the disaster, the most economically vulnerable population
is likely to sufferpoverty is the major global cause of illness and poor health (Hahn, 1996; McIntyre, 1997; PAHO, 1998a). When a family's income is reduced, there is decreased access to food, medical care, clean water, and other critical services.
Therefore, the mortality and morbidity arising from the immediate impact of marine natural disasters does not necessarily predict the long-term effects on public health. The loss of community services and secondary effects on the economy may have more serious impacts. The challenge to the international community is to help communities establish the infrastructure necessary to improve warning systems and to implement protective and preventive measures.
Forecasting Tropical Storms
The vulnerability of the United States to damages from tropical storms1 is higher now than in the past because of the growth and increased wealth of the coastal population; the population has been increasing at a rate of 4–5% per year (Sheets, 1990). Millions of people live and vacation along the coastline and are exposed to the threat of tropical storm winds, rain, storm surge, and severe
1Tropical storms is used as a general term to describe tropical storms and hurricanes, and cyclones.
weather. During this century, improved forecasts and warnings, better communications, and increased public awareness have reduced the loss of life associated with tropical storms in the United States. However, tropical storm-related damage has increased dramatically. Hurricane Andrew, in 1992, was the most costly natural disaster in U.S. history in terms of physical damage, although the loss of life was relatively low. The higher level of damages in the past decade is not from an increase in hurricane frequency, but reflects inflation, expansion of the coastal population, and the increased wealth of coastal communities (Pielke and Landsea, 1998).
The public's vulnerability is a function of the skill in forecasting the intensity of wind, rain, storm surge, and severe weather near landfall. While specific track prediction models have shown up to a 15% improvement, there has been little improvement in the prediction of intensity change (Elsberry et al., 1992). For this reason, the average length of coastline warned per storm, about 354 miles, has not changed much over the past decade. However, the average preparation costs increased six-fold in the past seven years, from $50M per storm in 1989 to an estimated $300M per storm in 1996 (OFCM, 1997). Unless the rate of forecast improvements can be accelerated, the downward trend of tropical storm casualties is not likely to continue, and the damage will continue to escalate. Track prediction is made more difficult by decadal climate variability, which leads to long-term variation in the frequency, intensity, origins, and paths of hurricanes.
Each improvement in tropical storm forecasting has been achieved by taking advantage of better observations. New strides in our abilities have always paralleled by development of new research tools; from instrumented aircraft, to radar and satellites. Because storms originate in the tropical ocean where few data are available, the scientific community has pioneered mobile observing strategies in order to provide critical observations of the storm's location and strength. One example is the production of high quality images of tropical storms produced by the SeaWiFS satellite (Plate III). These techniques have evolved to include measurements of the upper ocean and atmosphere in the vicinity of the storm. High-quality, high-resolution observations provide essential data used in determining parameters for models of atmospheric, oceanic, or coupled processes. A synergism between observations and models is required to isolate the important physical processes that will allow more accurate forecasts.
The Federal Emergency Management Agency (FEMA) requires coastal communities with limited escape routes to have completed preparation and evacuation before the arrival of gale force winds, typically 24 h to 48 h before landfall. However, an inadequate understanding of the fundamental mechanisms that inflict damage impairs our ability to provide timely warnings. Errors in wind, storm surge, and rainfall forecasts have prevented officials from accurately defining the most vulnerable regions in order to expedite required preparations well in advance of the projected landfall. Hurricane Opal provides an example of this as described below.
During the evening of October 3, 1995, Hurricane Opal was located in the southern Gulf of Mexico [for a summary of Hurricane Opal and its impacts see the National Oceanic and Atmospheric Administration (NOAA) Service Assessment Team Report (NOAA, 1996)]. The storm had been slowly intensifying over the previous three days while drifting slowly over the Gulf of Campeche. Given the small basin size of the Gulf of Mexico, Opal could strike anywhere along the U.S. Gulf Coast within 24 h.
In the middle of the night on October 3, the storm started one of the most rapid deepening and intensification cycles that has ever been observed as it moved at 19 miles per hour (MPH) toward the U.S. Gulf Coast. Within a 5 h period, U.S. Air Force Reserve (AFRES) reconnaissance aircraft measured a steep central pressure drop (939 hPa2 to 916 hPa), estimated surface winds increased to nearly 157 MPH, and the radius of the eye of the storm contracted from 19 to less than 10 miles. This rapid deepening presented the hurricane specialists with a major problem when the storm approached category 5 status within 12 hours of landfall without any means of alerting the public. Fortunately, over the subsequent 6 hours, the storm strength dissipated before reaching landfall later in the afternoon (Plate IV). Despite this weakening, the storm surge and wave activity were greater than anticipated and caused extensive damage along the coast, while wind damage and rainfall were less than forecast for a storm of that strength (NOAA, 1996). In forecasting this storm, why was the surge and the extent of the severe weather greater than predicted, while the wind damage and rainfall were less than predicted?
Crucial unanswered questions concerning the change in tropical storm intensity lie in three major components: (1) upper ocean heat content and the subsurface ocean structures that affect it (Elsberry et al., 1976; Black, 1983; Shay et al., 1992); (2) the inner core dynamics of storms; and (3) the winds at the jet-stream level, which are influential in steering cyclones. Important programs addressing these issues include the Tropical Cyclone initiatives of the Office of Naval Research (Elsberry, 1995) and observations and modeling by NOAA's Hurricane Research Division. These studies have demonstrated that accurate profiles of atmospheric variables obtained from aircraft deployed dropwindsondes are important for track prediction.
Ocean's Role in Modulating Intensity
The ocean's influence on tropical storm pressure and wind variations is dependent on the transfer of heat from the surface waters to the atmosphere. The recent case of Hurricane Opal demonstrated that sudden unexpected intensification
2hPa = hectoPascals.
often occurs within 24 to 48 hours of landfall, when tropical storms pass over warm, oceanic features such as the Gulf Stream, Florida Current, Loop Current or warm core rings in the western North Atlantic Ocean and the Gulf of Mexico (Plate IV). Both sea surface temperatures and the temperature regime of the oceanic boundary layer (defined as the well-mixed upper ocean layer) are needed to assess oceanic regimes where intensification is likely to occur. Sources of warm upper ocean water, carried by currents, provide a nearly continuous source of heat and moisture for moderate to fast-moving tropical storms along the lower boundary (Jacob et al., 1996; Jacob et al., 1998; Shay et al., 1992). This same effect, which occurs over the Gulf Stream, may also have led to significant increases in the surface wind field that devastated South Florida coastal communities during Hurricane Andrew in 1992 (Powell and Houston, 1996). Quantifying the effects of these oceanic features on changes in the surface pressure and wind field during tropical storm passage has far-reaching consequences not only for the research and forecasting communities, but also for the public who rely on the most advanced forecasting systems to prepare for landfall.
Low atmospheric pressure causes sea level to increase underneath the storm. As the storm approaches landfall, cyclonically rotating surface winds on the right side of the eye push water onto the coast, whereas on the left side of the storm center water is driven away from the coastline. This effect, combined with tidal fluctuations in sea level and wind-generated waves, determines the storm surge or elevation of sea level. Flooding caused by the storm surge inflicts significant damage to coastal property, and frequently is responsible for the loss of life during tropical storms.
While hurricane damage in North America has primarily involved economic loss, loss of life has been the major concern with tropical storms in developing countries. As mentioned earlier in this chapter, there were between 250,000 and 500,000 immediate fatalities in the Bay of Bengal during the November 1970 storm surge, which reached 5.6 m (18.4 feet) amplitude (compounded with wind waves, the total water level exceeded 10 m (32.8 feet; Murty et al., 1986). Less easy to quantify were the loss of infrastructure, salinization of agricultural lands, and destruction of the fishing fleets (estimated at 90,000 vessels in the 1970 event). As with Atlantic hurricanes, modern satellite observations permit fairly good predictions of dangerous cyclones, but complete evacuation of the coastal region is close to impossible because most of the country is less than 33 feet above sea level. The population of Bangladesh is growing at a rate of 1.82% per year, thereby increasing the vulnerability of this nation to disastrous storm surges.
The U.S. Weather Research Program (Emanuel et al., 1995) and World Weather Research Program have identified landfalling tropical storms as a major focus of their research programs. It is in the national interest to mitigate damage
that occurs after tropical storms reach landfall. Over the next decade, these issues will have a significant impact on building codes, construction technology, preparedness lead times, and evacuation procedures, all aimed at saving lives and minimizing property damage. Even if the improvement in intensity predictions is only a few percent per year, the benefit-to-cost ratio is high, leading to an improvement over the present state of forecasting.
As people alter the coastal landscape, the impacts of tropical storms will also change. For example, seawalls are built to reduce damage from storm surges or to anticipate sea level rise, but they also decrease the number of functional wetlands and the potential for migration of wetlands in the event of sea level rise. There are fewer natural areas left to sustain the coastal ecosystem, increasing the vulnerability of remaining wetlands and estuaries to the destructive effects of storms. Disruption of these habitats may have consequences for human health by affecting the distribution of disease-causing organisms such as toxic algae and by reducing the productivity of waters upon which many people rely for food. Documentation of such changes is only rarely possible; an exception being the measurements describing the impact of Hurricane Andrew on Florida and Louisiana (Stone and Finkl, 1995).
Estuaries and the Coastal Ocean
The role of estuarine and coastal circulation in public health is largely concerned with the transport, concentration, or dispersion of pathogenic organisms and the contribution of estuaries to marine food webs. In an estuary, water comes from two sources that both define the origin of the pathogens and contribute to the physical processes that affect the distribution of these pathogens. Pathogens derived from human activities enter the estuary through freshwater streams and rivers. Health threatening marine microorganisms (such as vibrios and toxic algae) enter with the influx of seawater from the ocean. Where fresh and salt water meet, the water column stratifies because salty water is more dense and sinks beneath the freshwater. This has an impact on the biological and physical properties of the estuary, with phytoplankton growth being assisted by strong stratification. The flow of waters in and out of estuaries is described below with reference to the effects on organisms that cause human illnesses.
To examine estuarine circulation, the mean circulation over a tidal cycle is considered with the starting assumption that tidal currents cause no net flow. If the tides are small and the river flow is great, the river water flows seaward over a wedge of salt water. Stratification is strong and dense particulate material accumulates at the toe of the salt wedge (Figure 1-1a). If tidal currents are stronger, salt is mixed upward into the upper layer and gets transported seaward. Stratification weakens, but the two-layered flow regime remains: seaward at the surface and landward below, with net transport increased in both layers (Figure 1-1b). The circulation of the estuary then greatly exceeds the river inflow. In this case,
tidal mixing is largely responsible for the quality of the water because it increases the average circulation by as much as 50 times. The rotation of the Earth causes a general displacement of the seaward flowing layer to one bank (direction dependent on location in the northern or southern hemisphere). If tidal mixing is strong, it is possible to observe inflow on one bank and outflow on the opposite bank through the density front at the surface (Figure 1-1c). These density fronts will intersect the shoreline at some point, providing a mechanism for the landward transport of particles. In fact, cysts of dinoflagellates (which can initiate a harmful algal bloom) have been observed to accumulate at the landward edge of these density fronts (Garcon et al., 1986).
Estuaries have diverse geography that affects their circulation, as does the time-variation of forcing by tides, river-flow, and solar heating. Their topography acts to weaken the fresh surface outflow, and strengthen the deep inflow, in deep channels, with the opposite bias over shoals (Valle-Levinson and Lwiza, 1995).
River inflow varies greatly in different estuaries. The ratio of estuary volume to average river inflow per day yields a rough measure of the residence time of water in the system and varies from 579 days (Puget Sound) to 307 days (Chesapeake Bay) to 49 days (San Francisco Bay). These values are subject to great seasonal variation, reflecting changes in seasonal runoff. When weak river flow
rates are coupled with strong nutrient inputs, plankton blooms occur. As the plankton sink through the water column, they are consumed and decomposed, activities that consume and deplete oxygen in subsurface waters. Stratification of the water column acts to restrict mixing and thus prevents the reoxygenation of the depleted subsurface waters (Figures 1-2a and b; Table 1-2). In severe cases,
anoxic conditions develop and cause massive fish kills that may present a health hazard.
Conversely, estuaries also experience river floods, occasionally flushing salt from the river entirely. In this instance, freshwater pathogens (in waste runoff) that would normally die once they encountered saltwater may survive and contaminate seafood. Flushing is also influenced by changes in stratification brought on by the spring-neap tidal cycle in some estuaries, and by wind-driven flow of nearby coastal waters. These events are stochastic, contribute extensively to the effective flushing characteristics of an estuarine system, and occur on time scales commensurate with those of plankton blooms.
Although mean conditions of flow and stratification can be monitored and modeled, this may not always help predict the transport of particulates such as dinoflagellate cysts. In some systems, it is small scale variations in estuarine circulation patterns that appear to be most important in establishing conditions conducive to harmful algal blooms. Small scale variations in flow depend on
stratification, tidal currents, and topographic features that create complex circulation patterns not detected by broadly-spaced current meter arrays. Hence predicting particle transport at any particular time or location will require detailed models and/or extensive real-time monitoring systems.
Estuaries also form the conduit for the transport of high concentrations of land-derived nutrients and pollutants into coastal waters. Despite recent reductions in the input of toxic materials to U.S. waterways, concerns about coastal pollution remain regarding bioaccumulation, ecological and human health effects (NRC, 1994b). Nutrients are essential for the support of fisheries in coastal waters. As mentioned earlier, sometimes these nutrients stimulate large blooms of plankton that can result in oxygen-depleted areas, that are either hypoxic (low oxygen) or anoxic (no oxygen). This problem is particularly severe in the Gulf of Mexico, adjacent to the outflow of the Mississippi and Atchafalaya Rivers, where a low oxygen ''dead zone" forms during the summer months that covers roughly 7000 square miles.
Several mechanisms have been proposed whereby nutrient-laden waters from estuaries mix across coastal waters and the continental shelf: (1) upwelling-favorable winds displace the low-salinity waters offshore, (2) winds from storms yield vertical mixing that homogenizes the water column, and (3) instabilities in flow allow the coastal current to shed eddies into the central shelf region.
In addition to transporting nutrients offshore, these processes also transport minute marine organisms such as dinoflagellates. In the Gulf of Maine, plumes of lower salinity estuarine water have been found to harbor the toxic dinoflagellate Alexandrium tamarense in high concentrations. Upwelling pushes the algae offshore and disperses the bloom while downwelling, when the winds reverse, causes the algae to accumulate at high concentrations along the coast where shellfish beds are more likely to become contaminated (Franks and Anderson, 1992a,b).
Similarly, coastal currents provide a potential route for the transport of toxic dinoflagellates from one region to another. The three mixing processes described above may move phytoplankton from a contaminated coast to offshore waters where currents may carry them to new downstream locations. In this new area, a relaxation of upwelling, associated with a shift in wind direction, will transport the dinoflagellates to coastal waters, hence creating conditions for a toxic bloom in a place where the phytoplankton had never been seen before. This scenario has been used to explain outbreaks of paralytic shellfish poisonings caused by a toxic dinoflagellate in the oceanic bays along the northwest coast of Spain (Fraga et al., 1988). An "upwelling index" (Bakun, 1973), based on meteorological pressure fields, has been used to investigate whether this physical feature can be used to predict blooms of toxic algae in this area. In this way, the use of hydrographic data obtained by studying physical processes may someday allow health officials to anticipate outbreaks of harmful algal blooms before the public is exposed to contaminated seafood.
Climate Variability and Global Climate Change
When climate change persists over long time-scales (greater than one year), the consequences for human health tend to be more serious. Regional drought, for example, can be withstood for a limited time (more so in industrialized countries) but after a prolonged period leads to famine and displacement of populations. In drought-stricken areas, higher temperatures change regional rain patterns and affect agricultural productivity, thus disrupting local food supplies. In other areas, climate variability may bring increased rainfall. Dependent on the region, these changes may alternatively increase or decrease agricultural productivity and the potential for outbreaks of waterborne diseases. Also, higher temperatures could increase or decrease the range and abundance of insect and rodent vectors of disease, possibly spreading diseases such as rift valley fever and malaria to new areas. Finally, climate change could lead to an increase in heat-related deaths, including deaths from respiratory diseases caused by air pollution which is expected to be more severe with longer, warmer summers. Although death rates increase at both temperature extremes (heat waves and extreme cold), heat-related deaths are predicted to more than offset a reduction in winter mortality (Pearce et al., 1995).
Temperature and rainfall patterns, while most often thought of as atmospheric phenomena, actually involve the interaction of the atmosphere with the ocean and the land, particularly for long-term climatic changes. In this coupled system, the ocean serves as the major reserve of heat and moisture. Attention has centered on sea-surface temperature, however, it is actually the available heat content of the upper-ocean that counts. The size of this reservoir is in part determined by salinity stratification and ice cover as well as by temperature fields. These variables, which affect the capacity of the ocean to absorb and transmit changes in atmospheric conditions, are key factors in assessing the risks posed by climate change and variability.
El Niño/Southern Oscillation
(ENSO) and the North Atlantic Oscillation
The El Niño/Southern Oscillation (ENSO) and the North Atlantic Oscillation (NAO) serve as examples of recurring weather patterns that take place on time scales longer than one year. El Niño recurs every 3–7 years, when the prevailing easterly (westward) winds of the tropical Pacific fail. This suppresses the upwelling of cold, nutrient-rich water along the central and eastern equatorial Pacific and releases a pool of warm water from the western end of the Equator. This pool of warm water propagates eastward across the equatorial Pacific towards the western hemisphere.
Sometimes the human health consequences of ENSO weather are severe. The immediate results of a warmer sea surface are increased rainfall in the eastern
Pacific and decreased rainfall in the Asian sector. Australia typically experiences severe drought. The yearly migration of the Asian monsoons also correlates with ENSO, bringing drought to some parts of Africa and India. The sea surface temperature (SST) anomalies in the central equatorial Pacific correspond to an atmospheric response that propagates along a great-circle path over North America yielding increased rain and storminess in the southwestern and southeastern U.S. Changes in temperature and rainfall due to ENSO have been postulated to lead to outbreaks of malaria and cholera, however, this proposed link is controversial. This issue is covered in greater detail in the following chapter on infectious diseases.
NAO is a mode of variability in the atmosphere over the North Atlantic. It has a very broad spectrum of time-scales, from days to centuries. It was discovered in atmospheric pressure records by 18th century missionaries in Greenland who observed that cold winters there often occurred when Scandinavian winters were mild, and vice versa. The NAO is in part a strengthening or weakening of the Icelandic low-pressure center that dominates North Atlantic weather, a statistical result of changes in the path and intensity of wintertime storms that grow over the northern Atlantic Ocean. The NAO also correlates with a large stratospheric vortex that is centered over the North Pole (Perlwitz and Graf, 1995; Thompson and Wallace, 1998).
The NAO index, which describes the waxing and waning of this phenomenon, correlates strongly with many weather variables relating to human health. Temperatures and precipitation in northern Europe, northwest Africa, and the Middle East are particularly affected. The precipitation and river-flow rate in the Tigris-Euphrates is correlated with the NAO (which corresponds to about 70% of the observed variability). Regions like this, with limited fresh-water supply, are sensitive to this degree of change. Since the 1960s the positive NAO phase has corresponded to droughts in southern Europe and the Mediterranean (Hurrell, 1995) and decreased rainfall in Morocco (Lamb and Peppler, 1987).
Global Warming, Global Change: Gradual and Abrupt Climate Change
Recent weather records reveal that over the past century the climate we have experienced is at least as warm as any century since 1400 AD, possibly due to increases in greenhouse gases. A trend towards a warmer world is emerging from the complex spectrum of natural variability (Nicholls et al., 1995). The global-average surface temperature in 1997 was the warmest of the century (Figure 1-3) and probably the warmest of the past 1000 years, while the past 8 years have included the 3 warmest years since at least AD 1400 (Mann et al., 1998). In 1998, each month set a new record for globally averaged surface temperature.
However, this change is far from uniform. A pattern of response "modes" appears to be involved, in which warming is concentrated in northern Asia (which has seen up to a 3.5 ¹C warming in average wintertime surface air temperatures
between 1980 and 1997) with lesser warming in western North America, while large regions of the North Pacific and North Atlantic Oceans and their neighboring shores have actually cooled since the 1960s. Many climate variables are affected. Water transported in streams and rivers in North America has increased significantly, along with precipitation (Nicholls et al., 1995).
Warming of the lower atmosphere is particularly rapid over far northern land masses such as central Asia and western Canada. In Alaska, the front of the Columbia glacier has retreated 8 miles inland over the past 16 years, and no longer reaches Glacier Bay. As global surface temperatures have increased by 0.3–0.6 ¹C during the last century, the maximum recent warming has occurred in winter over the high mid-latitudes of the Northern Hemisphere (Nicholls et al., 1995). This warming has been especially marked for the period from 1975–1994,
which correlates with unusual ENSO activity (Nicholls et al., 1995) and analysis of paleoclimate indicators suggest that 1990, 1995 and 1997 were warmer than any of the last 500 years (Mann et al., 1998). The warming trend is expected to continue, with brief warming and cooling events expected as part of the natural variability. Although there is variability in outcomes from computer models of the coupled atmosphere/ocean/land system, some predict that in a world with twice the current atmospheric CO2 levels there will be a nearly 50% increase in high northern latitude precipitation (minus evaporation) combined with significant warming (Manabe and Stouffer, 1994). The predicted global-average warming over the next 50 years (to the year 2050) ranges from 0.7 to 2 ¹C, with values much greater than this over land and at high latitudes. Although such change is still very speculative, it is the common outcome of these models.
In this century, the occurrence of the two strongest El Niño events (1982/3 and 1997/8), and the occurrence of the strongest positive phase of the NAO (1972–1995) have led to speculation that global warming influences these events. This recent NAO has yielded the strongest Icelandic low observed in the past 120 years, with a northward displacement and intensification of storms and strengthened deep convection in the ocean (Dickson et al., 1996). Both ENSO and NAO affect analysis of global warming through their direct impact on surface temperatures. Hence there is an interconnection between global warming, ENSO, and NAO which complicates the prediction of future events.
Effects of global warming on other weather patterns such as tropical storms are difficult to predict. Nevertheless, recent analyses suggest that there are no global historical trends in tropical storm number, intensity or location and current thermodynamic models predict a modest (10–20%) increase in maximal potential storm intensity for a doubled CO2 climate that is small compared with natural variations (Henderson-Sellers et al., 1998). However, the continuing rise in sea level, described below, will contribute to the impact of tropical storms through the elevation of the base for storm surges (NRC, 1998a). Potential effects of global warming on major weather events like ENSO may also influence tropical storms. ENSO shifts the regions of storm activity and frequency in the eastern and northwest Pacific and decreases the frequency of storms during the warm phase of ENSO in the North Atlantic region (Henderson-Sellers et al., 1998).
Risk of waterborne infectious diseases and vector-borne diseases also is likely to be influenced by climatic changes and ENSO events as discussed in Chapter 2 of this report. In addition, there is a risk of increased morbidity and mortality among vulnerable populations if global warming disrupts normal weather patterns and causes temperature extremes (hot or cold waves), regional flooding, and severe storms. In some areas, however, climate change may result in milder weather patterns, resulting in lower morbidity and mortality. Some of these health concerns, as well as the effects of climate change on marine ecosystems, have been addressed by the collection of marine disturbance event data in the HEED program (Box 1-1).
The rise in sea level, estimated at 1.2–2 cm per decade over the last century. threatens human health through the magnitude of the storm surges that, even in the past, have had a devastating effect in low lying countries like Bangladesh (NRC, 1998a). Sea level rise results from the expansion of water as it warms, the melting of polar ice, and the solid-Earth adjustment3 (rebounding from the weight of the last glaciation). Sea level rise has already had a noticeable effect on storm-surge occurrence at low-lying coasts and islands. Because most of the non-oceanic water is stored in the polar ice sheets, changes in the stability of these sheets could have a significant impact on sea level (Nicholls et al., 1995). There has been concern that the stability of the West Antarctic Ice Sheet is vulnerable to global warming and may collapse, causing a significant effect on global sea level. It is not yet possible to predict the likelihood of collapse, but the stability of the West Antarctic Ice Sheet is currently under intensive study. During the winter of
3 The upward or downward movement of the Earth's crust following the melting of the ice sheets, which "unloaded" huge weight. Glacial rebound is still being observed and has a major effect on observed sea level.
1997/98, the ice cover in the Weddell Sea was less than in any winter since satellite observations began in the early 1970s.
Abrupt change in ocean circulation has been advanced as a possible outcome of warming and increased fresh-water loading of the high latitude oceans (Broecker, 1997; Manabe and Stouffer, 1994; Rahmstorff, 1995). The great meridional overturning circulation described at the start of this chapter is thought to have been stable over the past 1000 years. Paleoclimate studies show that it was greatly altered in intensity and depth distribution during the last glacial period, and suffered frequent oscillations during other glacial periods. Computer models of the coupled ocean/atmosphere/land system suggest that we may now be entering a period of instability. Although still very speculative, such events could lead to dramatic changes in the climates of northern Europe and the Middle East over time-scales of 10 to 50 years.
The Importance of Preparedness of the Health Services
The most effective way to reduce the immediate cost in lives and human suffering from a natural disaster caused by a shift in weather patterns as occurs during an El Niño, is to improve the promptness and quality of the response of health and medical services. Coastal areas subject to surges, tropical storms, or tsunamis require a higher level of preparedness of both the emergency medical services and the health sector at large. Time and money invested on hospital contingency planning, simulation exercises, and training, not only of the first responders, but also of the entire health services, should ameliorate the outcome of ocean-driven disasters. Evidence for such progress may be seen in the recent response to the 1997/1998 El Niño. Although the actual health impact is still being evaluated by the affected countries and compiled at the international level by the Pan American Health Organization (PAHO)/World Health Organization (WHO), it is striking to note the sharp decrease in international appeals by Latin American countries compared to the 1982/83 El Niño. This suggests an improvement in the local response, a result of 20 years of increased national health preparedness and training.
Scientifically accurate and timely forecasting alone will not prevent high death tolls or decrease damages, unless the warning is transmitted, disseminated, and acted upon locally. The last decades' cyclones in the Indian subcontinent were marked by unnecessary loss of lives due to unheeded alarm. Apparently, part of the population either did not receive or fully appreciate the official warning and in other instances the design of the shelters did not accommodate the cultural norms of the community, i.e., did not include dividing walls to separate men and women (Talukder et al., 1992). Finally, even improved predictions and effective reporting are useless if no evacuation plans are available. The lower mortality associated with hurricanes on the coasts of the United States in the last 30 years is the result of several factors: improved warning systems, compliance
with more stringent building codes, a successful policy of massive evacuation in the event of a serious storm, and fewer landfalling hurricanes (Pielke and Pielke, 1997).
Gradual climatic changes and the possible and widely anticipated impact on disease transmission require approaches other than the traditional disaster planning aimed at improving medical readiness. Operational research on the cost effectiveness of surveillance techniques and control measures are needed to monitor and respond to disease outbreaks, natural disasters, such as floods and storms, and heat waves. Malaria, dengue, and cholera are increasing pandemics, regardless of any causal effect of the El Niño, and even the known effective control measures have yet to be implemented. Protection of the population in the United States and other developed countries in the event of climatic disturbances requires a frontal and decisive reduction of the transmission in developing countries as well. The problem is global, therefore the solution must involve the cooperation of the international community.
New Technologies for Ocean Environmental Observation
As the problems facing life on Earth multiply, our ability to monitor them is also increasing rapidly. There has been a rapid increase in technological capacity, but a lag in implementation of these new tools. Improvements in our understanding of ocean processes will depend on a greater commitment to using new technologies to improve the baseline data used in constructing and calibrating models. Some of the current projects in ocean observations are described below.
Thanks to solid-state electronics, a large family of drifting and moored sensor packages is now at work monitoring the 3-dimensional ocean: its velocity, movement of fluid particles, temperature, salinity, and dissolved oxygen. PALACE floats, for example, now roam the ocean by the hundreds, drifting with water masses at great depth, and rising periodically to the surface to transmit their data to a satellite. The 3-dimensional ocean, with its patchy, laminated structures of biology, chemistry, and physics can now be monitored by drifting and moored sensors, and monitored from space using remote sensing. The launch of SeaWiFS in August 1997 gives us the first satellite dedicated to the global imaging of oceanic surface biological activity since the loss of the Coastal Zone Color Scanner in 1986. Comprehensive monitoring of coastal pollution and global primary productivity is now possible. Tracking of tropical cyclones and prediction of their path and intensity is vastly improved by satellite imagery, infrared images showing the sea-surface temperature, and the altimetric measurements of the Ocean Surface Topography Experiment (TOPEX)-Poseidon satellite, which measure the background surface currents of the ocean. The new network of long-range
Doppler radars of the U.S. weather service aids in this process as storms approach the coast.
Integration of these new observations into computer models of the ocean and atmosphere improves both predictions and understanding of the underlying dynamics of the system. The nearly exponential increase in computing power over recent decades gives us tools that may soon resolve processes of moderate scalelike regional plankton bloomswithin the context of a global model of the coupled ocean/atmosphere. However, a recent NRC report has identified a lack of available computing resources for the testing and application of climate models (NRC, 1998b).
Chemical and Biological Sensors
Biological and medical research is rapidly improving our ability to make rapid measurements of chemical and biological substances in fluids. This technology, applied to environmental measurements of the ocean, will give us high resolution 3-dimensional maps of key biological and chemical components over time. Although this technology can be prohibitively expensive, transfers of medical technologies developed for blood analysis to marine applications are feasible. In the past, only dissolved oxygen and salinity measurements have been widely collected by small electronic sensors. However, now there is active development of sensors for dissolved CO2 (to better than 1 PPM4) and nutrients, which in the past required elaborate analysis of retrieved water samples. Biological monitoring with fluorometers and light-transmission sensors has been possible for some time, but new "bioprobes" for a wide range of substances are being developed aggressively. Fiber-optic chemical sensors can detect a range of variables in extremely small fluid samples (organic vapors, dissolved oxygen, CO2, and pH; Ferguson et al., 1997; Tabacco et al., 1998). Methods for measuring specific complex biological molecules are now under development.
One of the challenges in developing new instrumentation lies in the special requirements of the ocean environment. In hospital applications, sensors for blood pH, O2, and sugar must be replaced or recalibrated on time scales of one day or less. This would be impossible in the long term deployments typically used in oceanographic research. A discussion of sensor development, particularly those relating to oceanic carbon, is given in NRC (1993).
Collection of data from the physical, chemical, and biological monitoring systems described above creates a parallel need for comprehensive, structured databases. The efficient utilization of this information should be optimized by carefully designed, query-driven retrieval systems.
4 PPM = Parts per million.
Marine natural disasters present challenges to both public health officials and scientists who must work together to minimize the impact of these events on human health. In some cases, this involves implementing existing technologies and preventive measures that have proven capacity to reduce human suffering. However, it is still impossible to predict and prepare for all emergencies as the recent tsunami in New Guinea and Hurricane Mitch in Central America have demonstrated. Furthermore, there are indications that the global climate is changing, affecting weather and the ocean alike. Change is nonlinearits individual components interact in surprising ways. When human interventions occur on a back-drop of global climate change, the net effects become all the more complex and difficult to predict. The implications of the current warming trend are controversial, but this in itself argues for vigilance in monitoring changes in physical and biological systems so that predictions can be improved and problems can be identified before they become emergencies.
Potential strategies for the future include programs for public health and scientific monitoring for climate change and the health effects of climate change.
1. Support for international ocean research programs. Programs such as CLIVAR (Climate Variability and Predictability Programme) and GOOS (Global Ocean Observing System) help meet some of the needs for global monitoring. GOOS, for example, has a major initiative called "Health of the Oceans," which examines the coastal ocean and human health.
2. Closer cooperation and exchange of information across borders. Improved communication between emergency and disaster coordinators in the Western Hemisphere can help to identify common problems and solutions. In this area, the efforts of the PAHO/WHO to establish Internet links among professionals in the Americas need to be strengthened and expanded.
3. Emphasis on improving the public health infrastructure. In developing countries, storm resistant health facilities would improve medical and hospital disaster preparedness. Contingency planning and training also improve preparedness as well as support for the adoption of strict standards for hurricane and flood resistance. This emphasis on preventative measures to reduce the impacts of disasters was recognized by the Scientific and Technical Committee convened by the International Decade of Natural Disaster Reduction in June, 1998 at the World Bank in Washington, D.C. (PAHO, 1998c).
4. Establish baseline observations of the physical ocean and its ecosystems to monitor global change. Change can only be evaluated in the context of past experience. Therefore, our ability to understand future events will depend on the quality of the observations and analyses of that are collected now. This will require the implementation of newly developed technologies and the establishment of databases to design and evaluate models.
The next decade will be a crucial period for monitoring climate. The uncertain predictions of climate models will be tested and the changes in the environment may be more striking and heterogeneous than currently predicted. Achievement of the above recommendations will require a cooperative effort among the social, political, and economic sectors of the community, both local and internationally; potentially through close coordination between the scientific community and various local and international agencies such as the U.S. Agency for International Development and PAHO/WHO.