What Are the Effects of Nutrient Over-Enrichment?
KEY POINTS IN CHAPTER 4
This chapter explores the impacts of nutrient over-enrichment and finds:
The productivity of many coastal marine systems is limited by nutrient availability, and the input of additional nutrients to these systems increased primary productivity.
In moderation in some systems, nutrient enrichment can have beneficial impacts such as increasing fish production; however, more generally the consequences of nutrient enrichment for coastal marine ecosystems are detrimental. Many of these detrimental consequences are associated with eutrophication.
The increased productivity from eutrophication increases oxygen consumption in the system and can lead to low-oxygen (hypoxic) or oxygen-free (anoxic) water bodies. This can lead to fish kills as well as more subtle changes in ecological structure and functioning, such as lowered biotic diversity and lowered recruitment of fish populations.
Eutrophication can also have deleterious consequences on estuaries even when low-oxygen events do not occur. These changes include loss of biotic diversity, and changes in the ecological structure of both planktonic and benthic communities, some of which may be deleterious to fisheries. Seagrass beds and coral reefs are particularly vulnerable to damage from eutrophication and nutrient over-enrichment.
Harmful algal blooms (HABs) harm fish, shellfish, and marine mammals and pose a direct public health threat to humans. The factors that cause HABs remain poorly known, and some events are entirely natural. However, nutrient over-enrichment of coastal waters leads to blooms of some organisms that are both longer in duration and of more frequent occurrence.
Although difficult to quantify, the social and economic consequences of nutrient over-enrichment include aesthetic, health, and livelihood impacts.
Nutrient enrichment can have arange of effects on coastal systems. On occasion, in some ecosystems moderate nutrient enrichment can be beneficial because increased primary production can lead to increased fish populations and harvest (Jørgensen and Richardson 1996; Nixon 1998). Far more often, when nutrient enrichment is sufficiently great, the effects are detrimental. In some cases, even small increases in nutrient inputs can be quite damaging to certain types of ecosystems, such as those particularly susceptible to changed conditions (e.g., coral reefs).
Direct and indirect ecological impacts of nutrient enrichment include increased primary productivity, increased phytoplankton biomass, reduction in water clarity, increased incidences of low oxygen events (hypoxia and anoxia), and changes in the trophic structure, trophic interactions, and trophodynamics of phytoplankton, zooplankton, and benthic communities. Harmful algal blooms may become more frequent and extensive. Coral reefs and submerged macrophytic vegetation, such as seagrass beds and kelp beds, may be degraded or destroyed. Fish kills may occur, and more importantly, subtle changes in ecological structure may lead to lowered fishery production. Generally, nutrient over-enrichment leads to ecological changes that decrease the biotic diversity of the ecosystem.
The ecological effects of nutrient over-enrichment can have societal impacts as well, although the economic consequences are generally difficult to quantify. These include aesthetic impacts, such as loss of visually exciting coral reefs and seagrass beds, as well as production of noxious odors and unappealing piles of algal detritus on beaches. Fishery resources can be damaged or lost. Human health is threatened by accumulation of toxins in shellfish. Property can be devalued. This chapter summarizes the societal impacts of nutrient enrichment.
Increased Primary Productivity
As discussed earlier in this report, eutrophication is a process of increasing organic enrichment of an ecosystem where the increased rate of supply of organic matter causes changes to that system (Nixon 1995). This increased rate of supply is driven by primary productivity. Primary productivity is affected by a variety of factors, including light availability, nutrients, and grazing mortality. The interplay of these factors determines how a coastal marine ecosystem will respond to nutrient additions. (These and other factors that determine an estuary’s sensitivity to eutrophication are discussed in detail in Chapter 6.) For many systems, primary productivity is limited largely by nutrient availability, and in these systems
increasing the nutrient input increases the primary productivity rate and often the phytoplankton biomass. As explained in Chapter 3, in the majority of coastal systems—at least in the temperate zone—nitrogen is the element most limiting of primary productivity; consequently, rates of primary production and standing stock of phytoplankton biomass are often directly related to nitrogen inputs.
As noted in a previous report from the National Research Council (NRC 1993a), planners and managers now are often at a disadvantage because “no guidelines exist by which to determine whether coastal marine ecosystems are in fact eutrophic.” That report goes on to recommend that coastal eutrophication be judged by some measure of the relationship between phytoplankton biomass (as represented by chlorophyll concentrations) and trophic status, the same approach that is generally used by limnologists for freshwater lakes. Adoption of such an approach would lead to the conclusion that “few estuaries are oligotrophic, many are mesotrophic, and many are extremely eutrophic” (NRC 1993a).
Other authors have suggested similar approaches. For instance, Jaworski (1981) has suggested a lake-based framework of nutrient-loading guidelines that, if met, would tend to keep most estuaries from becoming eutrophic. However, demonstrable harm from human-increased nutrient loading to estuaries has occurred in some systems even when the loadings were low enough not to be called eutrophic by these standards (NRC 1993a).
Nixon (1995) suggested another set of guidelines—these based on measured rates of primary production—for determining whether an estuary is eutrophic. In this classification scheme, estuaries with productivity between 300 and 500 g C m−2 yr-1 would be considered eutrophic, while those with productivities greater than 500 g C m−2 yr-1 would be considered hypereutrophic. These guidelines, too, lead to the conclusion that many estuaries are eutrophic or even hypereutrophic.
Increased Oxygen Demand and Hypoxia
Eutrophication is accompanied by an increased demand for oxygen. Some of this increased oxygen demand is due to the greater respiration of the increased biomass of plants and animals that are supported in the nutrient-loaded ecosystem. Much of it is often due to respiration of bacteria (in both the water column and sediments) that consume the organic matter produced by the greater plant production. If the loss of oxygen caused by increased respiration is not offset by the direct introduction of additional oxygen by photosynthesis or mixing processes, then hypoxia or anoxia occurs. Biologists generally refer to the situation where some oxygen is present but where dissolved oxygen levels are less than or
equal to 2.0 milligrams per liter (mg l−1) as hypoxia. Anoxia is the complete absence of oxygen.
Hypoxia and anoxia are more likely to occur in summer because warming of the water column can lead to stratification and the formation of a barrier that prevents the introduction and mixing of oxygen from surface waters. Also, the solubility of oxygen decreases and oxygen demand (respiration rate) generally increases as temperature increases.
As noted earlier, many studies of the biological impacts of reduced dissolved oxygen concentrations have used 2.0 mg l−1 as the cut off for designating conditions as hypoxic (e.g., Pihl et al. 1991, 1992; Schaffner et al. 1992), because below this threshold there are severe declines in the diversity and abundance of species in the systems. There is evidence, however, that 2.0 mg l−1 may not be a universal threshold. For example, the results of a study of biological resources in Long Island Sound, New York, revealed that 3.0 mg l−1 was the threshold level for finfish and squid (Howell and Simpson 1994). A study of the benthic community in Corpus Christi Bay, Texas, indicated that dissolved oxygen concentrations less than 3.0 mg l−1 should be the operational definition of hypoxia in that system (Ritter and Montagna 1999), and that a single value of dissolved oxygen as a water quality standard for estuarine waters may not be appropriate.
Many states have standards for dissolved oxygen levels in aquatic systems that are well above the limits used to define hypoxia. The Florida Department of Environmental Protection mandates that the average level of dissolved oxygen that must be maintained in marine waters designated for the commercial harvest of shellfish, recreation, and for the maintenance of healthy fish and wildlife is greater than 5.0 mg l−1 in a 24-hour period and never less than 4.0 mg l−1. Although this level may seem conservative, in the absence of detailed information for a system it may be appropriate.
The occurrence of hypoxic and anoxic bottom waters, particularly in the coastal zone, has become a major concern in recent years because it appears that the frequency, duration, and spatial coverage of such conditions have been increasing, and this increase is thought to be related to human activities (Diaz and Rosenberg 1995). Zones of reduced oxygen can disrupt the migratory patterns of benthic and demersal species, lead to reduced growth and recruitment of species, and cause large kills of commercially important invertebrates and fish (NRC 1993a). Such conditions can also lead to an overall reduction in water quality, thereby affecting other coastal zone activities such as swimming and boating. Reports of a “dead zone,” an extensive area of reduced oxygen levels covering an expanse originally of some 9,500 km2 in the Gulf of Mexico (Rabalais et al. 1991), have focused attention on the problem of coastal zone hypoxia. By
the summer of 1999, the hypoxic area in the Gulf of Mexico had grown to an area of 20,000 km2 (Rabalais personal communication).
Researchers studying the Chesapeake Bay have said since the 1980s that the occurrence of hypoxic and anoxic bottom waters has increased in association with nutrient inputs (Taft et al. 1980; Officer et al. 1984). More recent studies examined pollen distribution, diatom diversity, and the concentration of organic carbon, nitrogen, sulfur, and acid-soluble iron in sediment cores from the mesohaline portion of the bay (Cooper and Brush 1991). The cores represented a 2,000-year history of the bay. Changes in the concentration of organic components and pollen abundance coincided with the new settlement by Europeans in the late 1700s. This period was marked by major land clearing in the watershed, which likely promoted increases rates of sedimentation, mineralization, and nitrification, and an increase in agricultural activity and the use of manures. Analysis of the sediment cores indicated a shift in the phytoplankton community from centric to pennate diatoms for this time period, and this was interpreted as evidence of increased nutrient input to the bay. This historical perspective indicates a role for nutrients in the occurrence of hypoxia in Chesapeake Bay. As discussed in Chapter 5, the input of nutrients to Chesapeake Bay has probably accelerated even more in the last several decades due to increased use of inorganic fertilizer and increased combustion of fossil fuels and the resulting atmospheric deposition of nitrogen.
The northern Adriatic Sea and northern Gulf of Mexico are two other coastal systems that have experienced increasing episodes of hypoxia (Justic et al. 1993; Turner and Rabalais 1994). Both systems are affected by river flow, the Po River in the case of the former and the Mississippi River in the latter. In both systems researchers have documented a seasonal increase in primary productivity in surface waters that was related to nutrients and river flow; this increase was followed by hypoxia in the bottom waters. The hypoxia onset, however, lagged peak river flow in the Gulf of Mexico and Adriatic Sea by two and four months, respectively. This difference in the lag period was ascribed to greater depth of the water column in the Adriatic Sea and differences in the downward flux of organic matter. Again, the evidence showed that the introduction of new nutrients in the river flow contributed to the development of hypoxia in these systems, but stratification of the water column was a necessary condition.
There also is evidence that increased nutrient loading has contributed to the occurrence of hypoxia in Florida Bay and the Florida Keys (Lapointe at al. 1990; Lapointe and Clark 1992). The most severe cases of hypoxia were found in the canal and seagrass systems closest to the discharge areas. Increased nitrogen levels were associated with increased growth of nutrient-limited phytoplankton, whereas high levels of soluble reactive P
were associated with increased growth of macroalgae and tropical seagrasses. Lapointe and Matzie (1996) showed that episodic rainfall events led to higher submarine discharge rates that were followed within days by hypoxic oxygen levels.
Shifts in Community Structure Caused by Anoxia and Hypoxia
The occurrence of hypoxic and anoxic bottom waters may also lead to shifts in benthic and pelagic community structure due to the mortality of less mobile or more sensitive taxa, reduction of suitable habitat, and shifts in predator-prey interactions (Diaz and Rosenberg 1995). Hypoxia plays a major role in the structuring of benthic communities because species differ in the sensitivity to oxygen reduction (Diaz and Rosenberg 1995). The response of species to reduced oxygen availability also depends on the frequency and duration of these events. With short bouts of hypoxia, some large or very motile species are able to adjust to or move away from the stress.
Hypoxia tends to shift the benthic community from being dominated by large long-lived species to being dominated by smaller opportunistic short-lived species (Pearson and Rosenberg 1978). In addition, recurring hypoxia may limit successional development to colonizing communities. In such systems more organic matter is available for remineralization by the microbial community. This can decrease the amount of energy available for benthic recruitment when hypoxia and anoxia disappears. Zoo-plankton that normally vertically migrate into bottom waters during the day may be more susceptible to fish predation if they are forced to restrict their activity to the oxic surficial waters. Roman et al. (1993) concluded that the vertical distribution of copepods in the Chesapeake Bay was altered by the presence of hypoxic bottom waters. Moreover, an hypoxic or anoxic bottom layer may constitute a barrier that de-couples the life cycle of pelagic species (e.g., diatoms, dinoflagellates, and copepods) that have benthic resting stages (Marcus and Boero 1998).
In a controlled eutrophication experiment (Doering et al. 1989), the structure of the zooplankton community was affected by the presence or absence of an intact benthic community. In the absence of an intact benthic community, holoplanktonic forms, especially higher level predators, dominated, whereas meroplanktonic forms were more evident in the presence of an intact benthic community. Although the data did not identify the mechanism behind these shifts, the differences likely reflected alterations in the coupling of the benthic and pelagic environments (nutrient as well as life cycle linkages) (Marcus and Boero 1998).
Changes in predator-prey interactions in the water column can also lead to shifts in energy flow. Increased fish predation on zooplankton can
release grazing pressure on the phytoplankton and increase the deposition of organic matter to the sediments. If the duration and severity of the hypoxia is not sufficient to cause mortality of the macrobenthos, the increased supply of organic matter to the benthic system could fuel the growth of benthic fauna and demersal fish populations at the expense of pelagic fisheries. On the other hand, extended hypoxic and anoxic events could lead to the demise of the macrobenthos and the flourishing of bacterial mats. The loss of burrowing benthic organisms that irrigate the sediments and the presence of an extensive bacterial community may alter geochemical cycling and energy flow between the benthic and pelagic systems (Diaz and Rosenberg 1995). For example, the flux of nitrogen out of the sediments is affected by the rates of nitrification and denitrification, and these processes depend on the naturally oxic and anoxic character of the sediments.
Changes in Plankton Community Structure Caused Directly by Nutrient Enrichment
Nutrient over-enrichment can also change ecological structure through mechanisms other than anoxia and hypoxia. Phytoplankton species have wide differences in their requirements for and tolerances of major nutrients and trace elements. Some species are well adapted to low-nutrient conditions where inorganic compounds predominate, whereas others thrive only when major nutrient concentrations are elevated or when organic sources of nitrogen and phosphorus are present. Uptake capabilities of major nutrients differ by an order of magnitude or more, allowing the phytoplankton community to maintain production across a broad range of nutrient regimes. A decrease in silica availability in an estuary and the trapping of silica in upstream eutrophic freshwater ecosystems can occur as a result of eutrophication and thus nitrogen and phosphorus over-enrichment. This decrease in silica often limits the growth of diatoms or causes a shift from heavily silicified to less silicified diatoms (Rabalais et al. 1996). Given these changes in the cycling of nitrogen, phosphorus, and silica, it is no surprise that the phytoplankton community composition is altered by nutrient enrichment (Jørgensen and Richardson 1996).
The consequences of changes in phytoplankton species composition on grazers and predators can be great, but in general these are poorly studied. As noted by Jørgensen and Richardson (1996):
Any eutrophication induced change in the species composition of the phytoplankton community which leads to a change in size structure of the phytoplankton community will potentially affect energy flow in the entire ecosystem. Thus, eutrophication can, at least in theory, play an
important role in dictating whether the higher trophic levels in a given system are dominated by marketable fish or by jellyfish. . . . Little is actually known about the effects of eutrophication on the size structure of the phytoplankton community under various conditions, but this is an area that warrants further research.
In particular, a change from diatoms toward flagellates, which may tend to result during eutrophication as the silica supply is diminished, may be deleterious to food webs supporting marketable forms of finfish (Greve and Parsons 1977). On the other hand, if the silica supply remains high enough, moderate eutrophicaton can encourage more growth by diatoms and lead to higher fish production (Doering et al. 1989; Hansson and Rudstam 1990).
Looking beyond the major nutrients, it is also evident that phytoplankton species have variable requirements for nutritional trace elements or have different tolerances for toxic metals (Sunda 1989), and the effects of these elements can be affected by dissolved organic matter (DOM) concentrations. One example of this effect is seen with copper, which is highly toxic to marine organisms and is often significantly elevated in harbors and estuaries due to anthropogenic inputs. Copper is strongly bound to organic chelators in seawater, and this lowers copper’s biological availability and consequently its toxicity (Sunda 1989). Moffett et al. (1997) studied copper speciation and cyanobacterial distribution and abundance in four harbors subject to varying degrees of copper contamination from anthropogenic sources. Cell densities of cyanobacteria, one of the most copper-sensitive groups of phytoplankton, declined drastically in high copper waters compared to adjacent unpolluted waters. Because of the variability in the concentrations of the natural organic ligands that bind the copper, relatively small changes in the total copper concentration (7 to 10 times) among the study sites were associated with much larger (greater than 1000 times) changes in the free Cu2+ activity, the biologically available form.
The bioavailability of metals such as iron also can be affected by human activities, including nutrient pollution and the resulting eutrophication, and this in turn can affect phytoplankton species composition and thus ecosystem structure and function. Iron is an essential element for algae, and is required for electron transport, oxygen metabolism, nitrogen assimilation, and DNA, RNA, or chlorophyll synthesis. Organic ligands are needed to keep iron in solution at the pH of seawater, as iron hydroxides have extremely low solubility and tend to transform into stable crystalline forms that do not directly support algal growth. DOM plays a critical role in enhancing the bioavailability of iron in seawater. A variety of factors can affect DOM levels in estuaries and coastal systems, but in general
eutrophication results in higher DOM levels—due to higher levels of primary production, leakage of DOM from phytoplankton, release as phytoplankton are eaten or decompose—with concomitant changes in iron availability.
Although the potential impacts of nutrient enrichment on phytoplankton community structure in the field seem obvious, there are few well-documented examples. This is because the changes in community structure often are gradual and easily obscured by fluctuations in other controlling factors, such as temperature, light, or physical forcings.
Long-term data sets offer another insight into possible changes since they allow sustained trends to be detected in spite of short-term variability caused by weather or other environmental forcings. For example, a 23-year time series off the German coast documented the general enrichment of coastal waters with nitrogen and phosphorus, as well as a four-fold increase in the nitrogen:silicon and phosphorus:silicon ratios (Radach et al. 1990). This was accompanied by a striking change in the composition of the phytoplankton community, as diatoms decreased and flagellates increased more than ten-fold. Other data from nearby regions showed a change in the phytoplankton species composition accompanying a shift in the nitrogen:phosphorus supply ratio along the Dutch coast (Cadée 1990), as well as increased incidence of summer blooms of the marine haptophyte Phaeocystis after a shift from phosphorus-limitation to nitrogen-limitation (Riegman et al. 1992). Nutrient status, particularly phosphorus-limitation, is now believed to be a major factor driving colony formation in this genus. Experiments performed with cultures of Phaeocystis demonstrate that free-living solitary cells outcompete the more harmful colonial forms in ammonium- and phosphate-limited conditions, whereas colonies dominate in nitrate-replete cultures. This suggests that free-living Phaeocystis cells would be prevalent in environments that are regulated by regenerated nitrogen, whereas colonial forms would require a nitrate supply and thus would be associated with “new” nitrogen such as that supplied by pollution.
Another long-term perspective on nutrient enrichment on phytoplankton community structure is seen in recent data examining the abundance of dinoflagellate cysts in bottom sediments of Oslofjord, Norway (Dale et al. 1999). Dinoflagellate cysts are an important group of micro-fossils used extensively for studying the biostratigraphy and paleoecology of sediments. In this study, dinoflagellate cyst records were analyzed from sediment cores that covered a period of anthropogenic nutrient enrichment that began in the mid- to late-1800s, was heaviest from 1900 to the 1970s, and then diminished from the mid-1970s to the present. Over the period of nutrient and organic enrichment, cyst abundance in the sediments doubled and a marked increase in one species in particular,
Lingulodinium machaerophorum (=Gonyaulax polyedra), from less than 5 percent to around 50 percent of the assemblage was noted. In the core considered most representative of general water quality in the inner fjord (Figure 4-1), these trends reversed back to pre-industrial levels during the 1980s and 1990s when improved sewage treatment took effect. Other changes in the phytoplankton community no doubt occurred that were not revealed with this approach, but the cyst record nevertheless demonstrates substantial changes in the abundance and composition of a major phytoplankton class.
Although at times changes in community structure are directly the result of nutrient enrichment, sometimes they are an indirect result of other changes caused by increased nutrients. For instance, a change in the phytoplankton community in the form of selection for different species can be caused directly by increased nitrogen. On the other hand, a change in phytoplankton community structure can be caused indirectly by increased nitrogen, because higher levels of nitrogen increase productivity, which increases dissolved organic carbon, which in turn causes changes in the community structure. Generally it is difficult to determine whether community structure changes are direct or indirect.
Harmful Algal Blooms
Among the thousands of species of microscopic algae at the base of the marine food web are a few dozen that produce potent toxins or that cause harm to humans, fisheries resources, and coastal ecosystems. These species make their presence known in many ways, ranging from massive blooms of cells that discolor the water, to dilute, inconspicuous concentrations of cells noticed only because of the harm caused by their highly potent toxins. The impacts of these phenomena include mass mortalities of wild and farmed fish and shellfish, human intoxications or even death from contaminated shellfish or fish, alterations of marine trophic structure through adverse effects on larvae and other life history stages of commercial fisheries species, and death of marine mammals, seabirds, and other animals.
“Blooms” of these algae are sometimes called red tides, but are more correctly called HABs, and are characterized by the proliferation and occasional dominance of particular species of toxic or harmful algae. As with most phytoplankton blooms, this proliferation results from a combination of physical, chemical, and biological mechanisms and interactions that are, for the most part, poorly understood. HABs have one feature in common, however; they cause harm either due to their production of toxins or to the way the cells’ physical structure or accumulated biomass affect co-occurring organisms and alter food web dynamics. This descrip-
tor applies not only to microscopic algae but also to benthic or planktonic macroalgae that can proliferate and cause major ecological impacts, such as the displacement of indigenous species, habitat alteration, and oxygen depletion. The causes and effects of macroalgal blooms are similar in many ways to those associated with harmful microscopic phytoplankton species.
HAB phenomena take a variety of forms. One major category of impact occurs when toxic phytoplankton are filtered from the water as food by shellfish that then accumulate the algal toxins to levels that can be lethal to humans or other consumers. These poisoning syndromes have been given the names paralytic, diarrhetic, neurotoxic, and amnesic shellfish poisoning (PSP, DSP, NSP, and ASP). A National Research Council report (NRC 1999b) summarized the myriad human health problems associated with toxic dinoflagellates. In addition to gastrointestinal and neurological problems associated with the ingestion of contaminated seafood, respiratory and other problems may arise from toxins that are released directly into seawater or become incorporated in sea spray. Whales, porpoises, seabirds, and other animals can be victims as well, receiving toxins through the food web from contaminated zooplankton or fish.
Another type of HAB impact occurs when marine fauna are killed by algal species that release toxins and other compounds into the water or that kill without toxins by physically damaging gills. Farmed fish mortalities from HABs have increased considerably in recent years, and are now a major concern to fish farmers and their insurance companies. The list of finfish, shellfish, and wildlife affected by algal toxins is long and diverse (Anderson 1995) and accentuates the magnitude and complexity of the HAB phenomena. In some ways, however, this list does not adequately document the scale of toxic HAB impacts, as adverse effects on viability, growth, fecundity, and recruitment can occur within different trophic levels, either through toxin transmitted directly from the algae to the affected organism or indirectly through food web transfer. This is because algal toxins can move through ecosystems in a manner analogous to the flow of carbon or energy.
Yet another HAB impact is associated with blooms that are of sufficient density to cause dissolved oxygen levels to decrease to harmful levels as large quantities of algal biomass fall to the sediment and decay as the bloom declines. Oxygen levels can also drop to dangerous levels in “healthy” blooms due to algal respiration at night. Estuaries and nearshore waters are particularly vulnerable to low dissolved oxygen problems during warm summer months, especially in areas with restricted flushing.
One of the explanations given for the increased incidence of HAB outbreaks worldwide over the last several decades is that these events are
a reflection of pollution and eutrophication in estuarine and coastal waters (Smayda 1990). Some experts argue that this is evidence of a fundamental change in the phytoplankton species composition of coastal marine ecosystems due to the changes in nutrient supply ratios from human activities (Smayda 1990). This is clearly true in certain areas of the world where pollution has increased dramatically. It is perhaps real, but less evident, in areas where coastal pollution is more gradual and unobtrusive. A frequently cited dataset from an area where pollution is a significant factor is from Tolo Harbor in Hong Kong, where population growth in the watershed grew six-fold between 1976 and 1986. During that time, the number of observed red tides increased eight-fold (Lam and Ho 1989). The underlying mechanism is presumed to be increased nutrient loading from pollution that accompanied human population growth. A similar pattern emerged from a long-term study of the Inland Sea of Japan (Box 4-1).
Both the Hong Kong and Island Sea of Japan examples have been criticized, since both could be biased by changes in the numbers of observers through time, and both are tabulations of water discolorations from algal blooms, rather than just toxic or harmful episodes. Nevertheless, the data demonstrate that coastal waters receiving industrial, agricultural, and domestic effluents, which frequently are high in plant nutrients, do in fact experience a general increase in algal growth.
Nutrients can stimulate or enhance the impact of toxic or harmful species in several ways. At the simplest level, toxic phytoplankton may increase in abundance due to nutrient enrichment but remain as the same relative fraction of the total phytoplankton biomass (i.e., all phytoplankton species are stimulated equally by the enrichment). In this case, we would see an increase in HAB incidence, but it would coincide with a general increase in algal biomass. Alternatively, some contend that there has been a selective stimulation of HAB species by nutrient pollution. This view is based on the nutrient ratio hypothesis (Smayda 1990), which argues that environmental selection of phytoplankton species has occurred because human activities have altered nutrient supply ratios in ways that favor harmful forms. For example, diatoms, the vast majority of which are harmless, require silicon in their cell walls, whereas most other phytoplankton do not. As discussed in Chapter 3, silica availability is generally decreased by eutrophication. In response to nutrient enrichment with nitrogen and phosphorus, the nitrogen:silicon or phosphorus:silicon ratios in coastal waters have increased over the last several decades.
Diatom growth in these waters will cease when silicon supplies are depleted, but other phytoplankton classes (which have more toxic species) can continue to proliferate using the “excess” nitrogen and phosphorus. The massive blooms of Phaeocystis that have occurred with increasing frequency along the coast of western Europe are an example of this phe-
A prominent example of the link between eutrophication and increased HABs is seen in a long-term data set of red tides in the Inland Sea of Japan. As Japanese industrial production grew rapidly in the late 1960s and early 1970s, pollution of coastal waters also increased. Currently, the number of visible red tides increased from 44 per year in 1960 to more than 300 a decade later, matching the pattern of increased nutrient loading from pollution (Figure 4-2). Japanese authorities instituted effluent controls in the early 1970s through the Seto Inland Sea Law, resulting in a 70 percent reduction in the number of red tides that has persisted to this day (Okaichi 1997).
FIGURE 4-2 Changes in the number of visible red tides in the Inland Sea of Japan, 1960-1990 (Okaichi 1997; used with permission from Terra Scientific Publishing).
nomenon (Lancelot et al. 1998). Other examples include the fish killing blooms of Chattonella species, which have caused millions of dollars of damage in the Seto Inland Sea, though the frequency and severity of these outbreaks has decreased since pollution loading was reduced and eutrophication abated somewhat (Okaichi 1997).
Another frequently cited example of the potential linkage between HABs and pollution involves the recently discovered “phantom” dinoflagellate Pfiesteria. In North Carolina estuaries and in the Chesapeake Bay, this organism has been linked to massive fish kills and to a variety of human health effects, including severe learning and memory problems (Burkholder and Glasgow 1997). A strong argument is being made that nutrient pollution is a major stimulant to outbreaks of Pfiesteria or Pfiesteria-like organisms because the organism and associated fish kills have occurred in watersheds that are heavily polluted by hog and chicken farms and by municipal sewage. The mechanism for the stimulation appears to be two-fold. First, Pfiesteria is able to take up and use some of the dissolved organic nutrients in waste directly (Burkholder and Glasgow 1997). Second, this adaptable organism can consume algae that have grown more abundant from nutrient over-enrichment. Even though the link between Pfiesteria outbreaks and nutrient pollution has not been fully proven, the evidence is strong enough that legislation is already in various stages of development and adoption to restrict the operations of hog and chicken farms in order to reduce nutrient loadings in adjacent watersheds. Pfiesteria has thus provided the justification needed by some agencies to address serious and long-standing pollution discharges by nonpoint sources, which heretofore have avoided regulation.
Degradation of Seagrass and Algal Beds and Formation of Nuisance Algal Mats
Many coastal waters are shallow enough that benthic plant communities can contribute significantly to autotrophic production if sufficient light penetrates the water column to the seafloor. In areas of low nutrient inputs, dense populations of seagrasses and perennial macroalgae (including kelp beds) can attain rates of net primary production that are as high as the most productive terrestrial ecosystems (Charpy-Roubaud and Sournia 1990). These perennial macrophytes are less dependent on water column nutrient levels than phytoplankton and ephemeral macroalgae, and light availability is usually the most important factor controlling their growth (Sand-Jensen and Borum 1991; Dennison et al. 1993; Duarte 1995). As a result, nutrient enrichment rarely stimulates these macrophyte populations, but instead causes a shift to phytoplankton or bloom-forming benthic macroalgae as the main autotrophs. Fast-growing micro- and
macroalgae with rapid nutrient uptake potentials can replace seagrasses as the dominant primary producers in enriched systems (Duarte 1995; Hein et al. 1995). The biotic diversity of the community generally decreases with these nutrient-induced changes (Figure 4-3A&B).
Over the last several decades, nuisance blooms of macroalgae (“seaweeds”) in association with nutrient enrichment have been increasing along many of the world’s coastlines (Lapointe and O’Connell 1989). Phytoplankton biomass and total suspended particles increase in nutrient-enriched waters and reduce light penetration through the water column to benthic plant communities. Epiphytic microalgae become more abundant on seagrass leaves in eutrophic waters and contribute to light attenuation at the leaf surface, as well as to reduced gas and nutrient exchange (Tomasko and Lapointe 1991; Short et al. 1995; Sand-Jensen 1977). Ephemeral benthic macroalgae have light requirements that are significantly less than either seagrasses or perennial macroalgae, and also can shade perennial macrophytes such as seagrasses and contribute to their decline (Markager and Sand-Jensen 1990; Duarte 1995). These nuisance algae are typically filamentous (sheet-like) forms (e.g., Ulva, Cladophora, Chaetomorpha) that can accumulate in extensive thick mats over the seagrass or sediment surface, and this can lead to destruction of these submerged aquatic seagrass systems. Massive and persistent macroalgal blooms ultimately displace seagrasses and perennial macroalgae through shading effects (Valiela et al. 1997). The nuisance algae also wash up on beaches, creating foul-smelling piles.
In addition to shading, seagrass distribution in eutrophic waters is influenced by increased sediment sulfide concentrations resulting from decomposition in anoxic organic-rich sediments. Elevated sediment sulfide has been shown experimentally to reduce both light-limited and light-saturated photosynthesis, as well as to increase the minimum light requirements for survival (Goodman et al. 1995). Both effects interact with increased light attenuation to decrease the depth penetration of seagrasses in eutrophic waters.
Decreased photosynthetic oxygen production at all light levels also reduces the potential for oxygen translocation and release to the rhizosphere, and creates a positive feedback that reduces sulfide oxidation around the roots, further elevating sediment sulfide levels. In Florida, chronic sediment hypoxia and high sediment sulfide concentrations have been associated with the decline of the tropical seagrass Thalassia testudinum (Robblee et al. 1991). Sulfide also may reduce growth and production of seagrasses by decreasing nutrient uptake and plant energy status, as has been shown for salt marsh grasses (Bradley and Morris 1990, Koch et al. 1990).
Declines in seagrass distribution caused by decreased light penetra-
tion in deeper waters or changes in community composition prompted by the proliferation of benthic macroalgae in shallower waters will have significant trophic consequences. Seagrass roots and rhizomes stabilize sediments, and their dense leaf canopy promotes sedimentation of fine particles from the water column. Loss of seagrass coverage increases sediment resuspension and causes an efflux of nutrients from the sediment to the overlying water that can promote algal blooms. Seagrasses also provide food and shelter for a rich and diverse fauna, and reduced seagrass depth distribution or replacement by macroalgal blooms will result in marked changes in the associated fauna (Thayer et al. 1975; Norko and Bonsdorff 1996).
In addition, where mass accumulations of macroalgae occur, their characteristic bloom and die-off cycles influence oxygen dynamics in the entire ecosystem. As a result, eutrophic shallow estuaries and lagoons often experience frequent episodic oxygen depletion throughout the water column rather than the seasonal bottom-water anoxia that occurs in stratified, deeper estuaries (Sfriso et al. 1992; D’Avanzo and Kremer 1994). Benthic macroalgae also uncouple sediment mineralization from water column production by intercepting nutrient fluxes at the sediment-water interface (Thybo-Christesen et al. 1993; McGlathery et al. 1997) and can outcompete phytoplankton for nutrients (Fong et al. 1993). Except during seasonal macroalgal die-off events in these shallow systems, phytoplankton production is typically nutrient-limited and water column chlorophyll concentrations are uncharacteristically low despite high nutrient loading (Sfriso et al. 1992).
Coral Reef Destruction
Coral reefs are among the most productive and diverse ecosystems in the world. They grow as a thin veneer of living coral tissue on the outside of the hermatypic (reef-forming) coral skeleton. The world’s major coral reef ecosystems are found in nutrient-poor surface waters in the tropics and subtropics. Early references to coral reef ecosystems preferring or “thriving” in areas of upwelling or other nutrient sources have since been shown to be incorrect (Hubbard, D. 1997). Rather, high nutrient levels generally are detrimental to “reef health” (Kinsey and Davies 1979) and lead to phase shifts away from corals and coralline algae toward dominance by algal turf or macroalgae (Lapointe 1999). For example, some offshore bank reefs in the Florida Keys that contained more than 70 percent coral cover in the 1970s (Dustan 1977) now have only about 18 percent coral cover; turf and macroalgae now dominate these reefs, accounting for 48 to 84 percent cover (Chiappone and Sullivan 1997). Reduced
herbivory, either caused by disease or overfishing, also can lead to increases in macroalgal cover and reduced coral abundance.
Because the growth rates of macroalgae under the high light intensities and warm temperatures found in coral reef waters are highly dependent on the concentration of the growth-limiting nutrient (typically either nitrogen or phosphorus), even slight increases in ambient dissolved nutrient concentrations can lead to expansion of algae at the expense of coral. Standing stock concentrations are not the best indicators of nutrient availability because they do not take into account turnover times of nutrient pools (Howarth 1988), but in a comparative sense they can provide information on trends in ambient nutrient conditions. In a review of eutrophication on coral reefs in Kaneohe Bay, Hawaii, fringing reefs in Barbados, and the Great Barrier Reef, Bell (1992) found that macroalgae tended to dominate coral communities at reported dissolved inorganic nitrogen (DIN) concentrations above 1.0 μM and 0.2 μM soluble reactive phosphorus (SRP). Macroalgae also have became the dominant space-occupying organisms on some carbonate-rich reefs in the Caribbean (Lapointe et al. 1993) and in localized areas near the Florida Keys (Lapointe and Matzie 1996) even though there were only small increases in measured DIN and SRP concentrations. The growth rate of the coral reef macroalga Dictyosphaeria cavernosa , which overgrew Kaneohe Bay in the 1960s due to sewage nutrient enrichment (Banner 1974), has recently been shown to be nitrogen-limited and to achieve maximum growth rates at concentrations as low as 1.0 μM DIN (Larned and Stimson 1996). That nutrient enrichment at these low levels enhances macroalgal growth and triggers such dramatic ecological changes underscores the extreme sensitivity of these oligotrophic ecosystems to even slight increases in nutrient enrichment.
Increased macroalgal cover on reefs inhibits the recruitment of corals and leads to second-order ecological effects. For instance, macroalgal blooms can lead to hypoxia and anoxia of the reef surface, reducing the habitat quality needed to support the high diversity of coral reef organisms and potentially important grazers.
The recent recognition that the “global nitrogen overload problem” (Moffat 1998) now affects remote areas via atmospheric pathways (Vitousek et al. 1997) illustrates how nitrogen enrichment can potentially impact even remote reef locations on earth and contribute to coral “stress.” There is some evidence that nitrogen availability, in addition to temperature, light, and other environmental factors, may influence the “coral bleaching” (i.e., loss of zooxanthellae) phenomenon that has expanded globally in recent years (D’Elia et al. 1991). Over a six-year period, Fagoonee et al. (1999) found that the zooxanthellae density of corals in Mauritius varied considerably, correlating most significantly with nitrate concentration of the water column.
Myriad other direct and indirect effects of coastal eutrophication are known to affect coral reefs. Direct effects include decreased calcification associated with elevated nutrients. This is presumably at least in part due to shading by increased macroalgal biomass, since calcification is a direct function of photosynthesis by zooxanthallae. Indirect effects on coral reefs can stem from increased phytoplankton biomass that alters the quality and quantity of particulate matter and optical properties of the water column (Yentsch and Phinney 1989). For instance, reduced light levels associated with turbidity on reefs in Barbados depressed larval development and maturation in the coral Porites porites (Tomascik and Sander 1985). Another indirect effect occurs when nutrient enrichment enhances predator species; for example, outbreaks of the “Crown-of-Thorns” starfish in the South Pacific, which preys on living coral tissue, have been related to the stimulatory effects of nutrient runoff on their larval development (Birkeland 1982).
Disease and Pathogen Increases
The occurrence of microbial pathogens in the marine environment that is of concern to human health generally is associated with environmental contamination by human sewage and not nutrients per se (NRC 1993a). However, one group of pathogens, the Vibrios, has been identified as autochthanous members of the microbial community in brackish estuarine and coastal waters (Colwell 1983). In laboratory studies, the growth rate of Vibrio cholerae has been positively correlated with organic enrichment (Singleton et al. 1982). Another species V. vulnificus has been identified as a dominant member of the heterotrophic bacterial community of the Chesapeake Bay (Wright et al. 1996). It is possible, therefore, that eutrophication promotes the growth of these pathogens under field conditions. Research has also revealed an association of V. cholerae and V. parahaemolyticus with zooplankton, particularly copepods, to whose surfaces the bacteria attach (Kaneko and Colwell 1975; Huq et al. 1983). Colwell (1996) has suggested that phytoplankton blooms may be the ultimate cause of some outbreaks of V. cholerae, by fueling the growth of copepods. The increased abundance of copepods with their associated Vibrio flora could provide the dose necessary to cause cholera, and it could be worthwhile to determine if higher level predators of copepods (e.g., fish) could become contaminated by Vibrios and transmit the pathogens to human consumers.
The impacts of nutrient enrichment can be measured not only in eco-
logical terms but also in economic terms. The economic impacts are a measure of either the damages (i.e., lost value) from nutrient enrichment or the benefits of the improvements that result from reducing or reversing this process. To measure value, economists use the concept of “willingness to pay” (WTP) for improvements in, for example, water quality, or “willingness to accept” (WTA) compensation for environmental degradation (Freeman 1993; Smith 1996).1 Because of the difficulty of measuring WTA, most empirical studies estimate WTP measures of value. WTP represents the amount of money that an individual would be willing to give up or pay to secure an environmental improvement, which reflects how much the improvement is “worth” to the individual.2
Note that the amount the individual is willing to pay may differ substantially from the amount actually paid. For example, many recreational opportunities are available free of charge.3 This does not mean, however, that the opportunity has no value. As long as an individual would have been willing to pay for that opportunity (e.g., for access to the marine resource), it has a positive economic value (i.e., it generates gross benefits for that individual) even if the individual does not pay for it. Likewise, if a water quality improvement increases the amount an individual would have been willing to pay for a recreational opportunity, that improvement has a positive economic benefit even if the individual did not actually pay for the improvement. A primer on the concept of economic value and its application to the valuation of coastal resources is available from National Oceanic and Atmospheric Administration (NOAA) (Lipton and Wellman 1995). This handbook is specifically designed to provide coast resource managers with an introduction to economic valuation.
Types of Economic Value
There are a number of ways to classify economic value (Freeman 1993). A fundamental distinction can be drawn between use value and non-use value (Freeman 1993; Smith 1996). Use value is the value an individual derives from directly using a resource. In the context of marine
resources, this includes value derived from the use of the resource for swimming, recreational fishing, commercial fishing, wildlife viewing, boating, and beach uses (Bockstael et al. 1989). In contrast, non-use value is the value derived from the resource even though it is not currently used. Non-use values include (1) existence value, which is the value derived simply from knowing that the resource exists and is maintained, (2) bequest value, the value that the current generation received from knowing that the resource will be available for future generations, and (3) option value, which derives from preserving the resource so that the option of future use is retained (Smith 1996). This categorization encompasses a very broad definition of economic value. In particular, even when a resource is not currently being used by humans, it will still have economic value if individuals are willing to pay to preserve it despite the lack of current use. Thus, the ecological function of an ecosystem will have economic value as long as individuals value that ecological function (for whatever reason).
A second categorization of values that also encompasses this broad definition of value is the distinction between the value of market and nonmarket goods. Market goods are goods and services that are traded in the market at a given price. For these goods, the price that people actually pay for the good or service serves as a reasonable proxy for how much they would be willing to pay (at the margin) for the good and hence serves as a useful proxy for its value. Market goods are the source of commercial value for many resources (e.g., commercial fisheries).
Non-market goods, on the other hand, are not bought and sold in markets and hence have no observable price. Water quality is a classic example of a non-market good. While individuals value water quality improvements (in the same way that they would value an increase in the consumption of shellfish, for example), they cannot go to the market and “purchase” additional water quality for a given price (in the same way that they can purchase an additional pound of shellfish). Without a price that reveals a minimum amount that individuals would be willing to pay for the water quality improvement, some other mechanism for estimating the value of the improvement must be found. A large amount of research has gone into the development of non-market valuation techniques (Box 4-2).
Alternative Valuation Techniques
A wide variety of techniques4 exist for estimating the dollar value of
For goods that are bought and sold in markets (e.g., commercial fish), the price of the good reveals the price people are willing to pay for one more unit of that good (i.e., what they think an additional unit of the good is “worth”). Some goods, such as water quality, do not have a market price, and as a result economists have had to look for other indicators of their value. For example, a study by Boyle et al. (1998) used differences in house prices across lakes of different water clarity to infer a measure of the value of improved water clarity. If people value water clarity and other factors remain the same, they should be willing to pay more for a house on a lake with greater clarity.
Using statistical techniques applied to data on house prices, water clarity, and other house characteristics for sales around 36 lakes in Maine, the researchers found that for a select group of lakes, a 1-meter reduction in visibility, a measure of water quality resulted in a reduction in the average property value between $6,001 and $7,629 (in 1998 dollars; Boyle et al. 1998). They then used the property value information to calculate the impact on property owners.
Similarly, economists have developed methods for inferring the value of water quality improvements from observations about recreational travel behavior. If individuals value water quality, they should be willing to travel farther or more often to sites with better water quality. Since increased travel involves increased costs, this increased travel reveals how many people are willing to pay for the improved water quality. For example, using survey data on visits to 11 public beaches on the western shore of Maryland and estimates of associated travel costs, Bockstael et al. (1989) applied a travel cost model to estimate the amount that individuals are willing to pay for additional visits to the beach. In a second-stage analysis, they then estimated the relationship between the value of additional visits and water quality, measured by the loadings of nitrogen and phosphorus. By varying water quality and calculating the corresponding change in the value of visits, the individual value of a water quality improvement could be estimated. Finally, from the individual estimates for an average user and information on the number of users, aggregate estimates of the value of a water quality improvement could be obtained. Applying this technique to the Chesapeake Bay, Bockstael et al. (1989) found that on average a 20 percent reduction in nitrogen and phosphorus inputs near the beach would generate benefits of $34.6 million (in 1984 dollars) from increased public beach use on the western shore.
improvements in environmental quality. The applicability of these techniques varies. Some are applicable only to improvements that directly affect market goods (e.g., commercial fisheries), while others are designed to estimate the value of improvements in non-market goods (e.g., recreational fishing). While different categorizations exist, valuation techniques are generally classified as either (1) indirect, or revealed prefer-
ence, approaches or (2) direct, or stated preference, approaches. Revealed preference approaches infer values from observed behavior. Stated preference approaches estimate values based on survey responses to questions about hypothetical scenarios.
There are a number of revealed preference approaches to valuation, including (1) derived demand/production cost estimation techniques, (2) cost-of-illness method, (3) the averting behavior (or avoidance cost) method, (4) hedonic price method, and (5) travel cost method (NRC 1997a). Each is designed to capture a particular component of economic value, and hence is applicable to a particular subset of environmental impacts. Thus, each provides only a partial measure of value; none is capable of providing a measure of the total economic value of a water quality improvement.
The derived demand approach is applicable when water quality serves as an input into the production of a marketed good. For example, ambient water quality or the amount of submerged aquatic vegetation can affect the stocks of commercial fish species. Thus, changes in water quality change the supply conditions for these species and hence the profits derived from related commercial fisheries. Under the derived demand approach, the change in profits is a measure of the value of the environmental improvement. Application of this technique requires documentation of the relationship between water quality and supply. In some cases, analyses simply attempt to estimate the impact of a discrete event, such as a Pfiesteria outbreak or HAB, on revenues, using a “before” and “after” comparison. For example, Lipton (1998) estimated that the commercial seafood industry in Maryland lost $43 million in sales in 1997 as a result of the public’s concern about Pfiesteria. However, this approach can both over- and under-estimate the economic impacts of such events. It over-estimates them because it does not account for cost savings due to lower production levels or for the ability of consumers to substitute other products (which raises revenues for the producers of these substitute products). As with all revealed preference approaches, it under-estimates the impact of such outbreaks because it does not capture losses in non-use value (NRC 1997a).
For more continuous water quality problems such as nutrient enrichment, the impact on supply can be estimated using econometric methods, provided there is sufficient variability in the environmental indicator either over space or over time. Examples of studies of this type include Lynne et al. (1981), who estimated the relationship between marsh characteristics and productivity of the blue crab fishery on the Gulf Coast of Florida, and Kahn and Kemp (1985), who estimated the impact of submerged aquatic vegetation on the striped bass population. More recently, Diaz and Solow (1999) estimated the effect of hypoxia on the brown and
white shrimp fisheries in the Gulf of Mexico, but failed to find any significant effect of hypoxia on these fisheries during the study period (1985–1995). It is possible that in this case, the natural variability in the underlying data is sufficiently large that it is difficult to detect changes attributable to hypoxia. Alternatively, it may be that discernible effects do not exist at current hypoxia levels, even though the effects could become significant if conditions worsen (Diaz and Solow 1999).
The derived demand approach is designed to capture the effects of environmentally induced supply shifts on producers. In contrast, the cost-of-illness method can be used to estimate the economic impact of environmental events that affect human health. It measures the benefits of a pollution reduction by estimating the possible savings in direct out-of-pocket expenses resulting from the illness (e.g., medicine, doctor, and hospital bills) and the lost earnings associated with the illness (NRC 1997a). It does not account for any discomfort or other health-related impacts that are not avoided through medical treatment, or for any expenditure undertaken to prevent the illness.
Expenditures undertaken to reduce or avoid the damaging effects of pollution are termed “averting” or “avoidance” costs. Reductions in these costs (i.e., cost savings) attributable to water quality improvements constitute a partial measure of the economic value of that improvement. For example, the Environmental Protection Agency (EPA) estimated the benefits of reduced nitrogen loadings to estuaries by calculating the pollution control costs that could be avoided if loadings were reduced (EPA 1998b). They estimated that in the east coast, Tampa, and Sarasota estuaries, the Regional NOx State Implementation Plan will save $237.8 million (in 1998 dollars) in pollution control costs.
Hedonic pricing models are most applicable to measuring the benefits of environmental improvements to property owners. The basic principle underlying this approach is that the amount consumers are willing to pay for a house or piece of property depends on the characteristics of that house, including the perceived environmental characteristics of the surrounding areas (i.e., those that are of concern to prospective purchasers) (Rosen 1974). Thus, a house on an estuary with high water quality should sell for a higher price than an otherwise comparable one on an estuary with low water quality. The increase in the price of the house provides an estimate of the value of the difference (or, equivalently, an improvement) in water quality. A recent study of lakefront property owners in Maine found that housing prices were significantly affected by water clarity (Box 4-2). Although this study was based on lakes rather than coastal waters, it suggests that homeowners do value water clarity improvements and are willing to pay for them. However, measures of willingness to pay based on hedonic studies of this type
capture only the value to the property owner. In particular, they do not capture the value to other users of the marine resource (e.g., recreationists) who do not own property near the resource.
The most commonly used method for estimating the value of improved water quality to recreationists is the travel cost method. Although recreationists may not pay an access fee for use of a recreational resource, they incur costs in the form of travel costs (including the opportunity cost of their time) and other out-of-pocket expenses. These costs can be viewed as the price paid for the recreation trip. Individuals who live farther from a site must pay a higher price (in the form of higher travel costs) and hence would be expected to demand less (i.e., take fewer trips). Thus, by relating travel costs to the number of trips taken, the demand for trips and the associated willingness to pay for them can be estimated. Improved water quality should increase the value of a recreational trip, which should in turn result in more frequent trips or trips from farther away. The changes in demand that result from a water quality improvement can be used to infer the economic value of the improvement.
There is a large theoretical and empirical literature on application of travel cost methods to the valuation of changes in environmental quality, particularly improvements in water quality. In a study of the benefits of improvements in water quality in the Chesapeake Bay, Bockstael et al. (1989) used the travel cost method to examine the impacts on three activities: beach use, boating, and fishing. For beach use and boating, water quality was represented by the level of nutrient enrichment as measured by the total input of nitrogen and phosphorus. For fishing, water quality was proxied by the catch rates for striped bass. From their travel cost study, they estimated the fishing, boating, and swimming benefits of a hypothetical 20 percent improvement in water quality in the bay to be in the range of $18 to $55 million per year in 1987 dollars. Bockstael et al. (1989) also used contingent valuation to estimate the benefits of water quality improvements in the bay (see discussion below).
Single-site travel cost models do not account explicitly for the possibility of substitutability across sites. As water quality at one site changes, users may switch to other sites, even if switching leads to increased travel costs. Random utility models are designed to model the choices that individuals make among alternative sites, as determined by environmental quality, distance, and other site characteristics. By examining the tradeoffs that individuals are willing to make between environmental quality and travel costs, estimates of the value of environmental improvements can be derived. In one study, Kaoru et al. (1995) applied a random utility model to the estimation of the value of reductions in nitrogen loadings in the Albemarle and Pamlico Sounds of North Carolina. Their estimates of the benefits to an individual of a 36 percent decrease in nitro-
gen loadings ranged from $0.12 to $11, depending on the assumptions made. While this study does not provide aggregate measures of value, it does highlight the sensitivity of estimates to individual characteristics and the level of model aggregation.
As noted above, by focusing on a particular user population, each of the revealed preference approaches provides a partial measure of the value of improved water quality. In particular, by their nature, they can capture only the use value of an environmental resource. In some cases, non-use value may constitute a significant portion of total economic value. The only techniques currently available for estimating total economic value, including non-use value, are stated preference approaches. The most commonly used stated preference approach is the contingent valuation method, although alternative methods such as conjoint analysis are being increasingly applied to environmental valuation (NRC 1997a).
The basic instrument used in contingent valuation method is a survey asking hypothetical questions about consumers’ willingness to pay. Although open-ended questions have been used, there is a consensus that a dichotomous choice format is preferable (Arrow et al. 1993; NRC 1997a). With a dichotomous choice format, individuals are asked whether they would be willing to pay a specific amount (say, $20) for a given improvement in environmental quality. If they answer “yes,” then their WTP exceeds $20. If they answer “no,” then their WTP is less than $20. By varying the specified amount and using statistical analysis, an estimate of mean WTP can be derived (Freeman 1993; Smith 1996). The mean values can then be aggregated to the population level to provide estimates of aggregate value. Bockstael et al. (1989) use a contingent valuation survey to estimate the willingness to pay to make water quality in the Chesapeake Bay suitable for swimming. They surveyed both users and non-users of the Chesapeake Bay. The average willingness to pay was $121 (in 1984 dollars) for users and $38 for non-users. For the total population in the District of Columbia and Baltimore standard metropolitan statistical area, the aggregate value of the improvement is estimated to be between $65.7 and $116.6 million (in 1987 dollars).
Although the contingent valuation method has the advantage of allowing estimation of total economic value, it has been the subject of much controversy (Diamond and Hausman 1994; Hanemann 1994). Concerns about its use stem from the potential for biased estimates as a result of the hypothetical nature of the survey responses. Much research has been devoted to testing or correcting for, or designing methodologies that eliminate, potential biases (Carson and Mitchell 1993; Smith 1996). The potential bias can be reduced through careful survey design and pre-testing.
Because neither revealed nor stated preference approaches provide an ideal technique for measuring environmental values, economists have
Each of the various techniques available to estimate economic impacts provides only a partial estimate of the benefits of water quality improvement. Aggregating benefits across categories of use to get a total measure of value is difficult for at least two reasons. Aggregation requires that the various studies consider identical water quality changes. This is rarely the case if the studies have been done independently, and maintaining comparability can be difficult even within a given study. For example, in estimating the benefits of water quality improvements using the travel cost method, Bockstael et al. (1989) considered a 20 percent reduction in nitrogen and phosphorus loadings when estimating values relating to beach use and a 20 percent increase in catch rates when estimating values derived from sport fishing. Since the “goods” that are being valued are not necessarily the same (i.e., a 20 percent reduction in nutrient input does not necessarily lead to a 20 percent increase in catch rates), combining the value estimates across these categories is problematic. In addition, in aggregating estimates across categories, care must be taken to avoid double counting. For example, the contingent valuation estimates of Bockstael et al. (1989) should capture both use and non-use values. These cannot be combined with other partial measures of use value without double counting.
Despite these difficulties, a recent NOAA study (Doering et al. 1999) attempted to combine the results of a number of valuation studies to estimate the wetlands benefits of reduced nutrient loadings to the Gulf of Mexico. The total benefit of wetlands restoration from reduced loadings are derived from both market goods (commercial fishing and fur trapping) and non-market goods. The non-market goods generated both use value (from recreation, sport fishing and waterfowl hunting) and non-use value (existence and bequest value). Combining estimates of these various components of value from previous valuation studies, the researchers estimated that the value to residents of the Mississippi drainage basin of restoring 100,000 acres of coastal wetlands would range between $11.8 and $40 billion (in 1999 dollars). The largest benefits are derived from non-use values, followed by the value from recreational fishing and ecological services provided by wetlands. The NOAA study also considered erosion control benefits of reduced nutrient loadings. Because of limited data, however, the study makes no attempt to put a dollar value on the other economic impacts of eutrophication (e.g., impacts on swimming, boating, or commercial fishing).
Even though there has been considerable research on the ecological impacts of nutrient pollution, to date research on the economic impacts has been very limited. There have been a few key studies, using a variety of valuation techniques, but each captures only a subset of benefits relating to a specific use or user group. Aggregating these across uses, as well as across locations, is difficult, and neither national nor regional estimates of total economic impact exist. Some of the studies that have been done show large economic impacts in some cases or under some assumptions, and rather modest impacts in other cases.
The literature clearly shows that people value improvements in water quality and are willing to pay for those improvements, even though market-based measures of value significantly under-estimate total value since a large component of
total value stems from non-market uses of the marine resources (e.g., for swimming, recreational fishing, boating, wildlife viewing). Furthermore, the studies that have been done have consistently concluded that non-use value comprises a significant share of total value. In other words, people are willing to pay significant amounts for water quality improvements even when they do not directly use the waterbody. Although non-use values are particularly difficult to measure and the techniques used to measure them are controversial, available data is consistent in showing that non-use values are important. Thus, estimates of economic impacts that fail to include non-market benefits, including non-use values, are likely to significantly underestimate total economic impacts.
recently developed methodologies that combine information from both types of approaches in an effort to improve benefit estimates (Cameron 1992; Adamowicz et al. 1994, 1997; Englin and Cameron 1996; Huang et al. 1997). Combining the two approaches has the potential to increase the reliability of the estimates, although a possible inconsistency in the use of joint estimation exists (Huang et al. 1997). It also allows a decomposition of total value into use and non-use values. A study by Huang et al. (1997) combines stated and revealed preference models in the estimation of the willingness to pay for improvements in water quality in the Albemarle and Pamlico Sounds of North Carolina. While they do not provide measures of aggregate value, their results do show that non-use value constitutes a significant portion (over half) of total value. Bockstael et al. (1989) also found significant non-use values in their study. This suggests that ignoring non-use values will lead to significant under estimation of the total value of a water quality improvement (Box 4-3).