Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
6 Analysis of Monitoring Efforts As described in Chapter 4, there exists a wide range of current and historical monitoring efforts in the Southern California Bight. Analyzing each of these in turn would be an unrealistic task, but examining only a few in detail might cause us to neglect important insights and patterns that could be derived from a broad survey. This review therefore identifies important conceptual issues, and illustrates them using examples from existing monitoring programs. Many of these issues and examples identity shortcomings of the mon- itoring system and existing programs, and others stress positive develop- ments. The analysis that follows emphasizes that monitoring efforts in Southern California are characterized by a commitment to technical ex- cellence and continued evolution toward more sophisticated and eRective planning and implementation. There is a broad consensus in the mon- itoring community that programs today are, in general, vastly improved over those in effect 10 or more years ago. This progress has highlighted remaining problems and has allowed attention to shift to broader concerns. The willing participation in this case study by all parts of the monitoring community is clear evidence of their interest in continuing to improve monitoring efforts. This chapter focuses on four main topics: 1. institutional objectives and their limitations, 2. technical design and implementation, 3. technical interpretation and decision making, and 116
117 4. the overall allocation and organization of monitoring. Judgments about monitoring's effectiveness in each of these areas are based on the criteria and concepts outlined in Chapter 5. This chapter discusses these concepts more extensively, in light of evidence from specific programs. The panel's analysis of monitoring was based in large part on the written and verbal comments of invited speakers at the fact-finding sessions and further in-depth interviews with members of the monitoring community. The specific comments of these participants in the case study contributed to a consensus about the overall strengths and weaknesses of monitoring in the bight. This consensus is presented here as a series of statements and is amplified in the following sections. The strengths of the monitoring system include: · an established legal requirement for addressing environmental is- sues and problems; · important contributions to environmental decision making; · active links to ongoing research programs; · innovative monitoring program designs and techniques; data; high-quality methods for collecting, analyzing, and interpreting · raw monitoring data of high quality and integrity; · large data sets that have greatly increased understanding of local- ized impacts, particularly of municipal wastewater discharges; and · a few long-term data sets that are valuable for examining large-scale and long-term effects of human activities on the bight. The weaknesses of the monitoring system include: · poorly defined management objectives; · poorly defined monitoring endpoints or decision criteria, especially whole; · sampling designs that do not take into account spatial and temporal scales of natural variability; · reliance on a shotgun approach that measures many parameters, regardless of their relevance to operational' environmental, or public health decisions; · rigidity that does not permit dropping redundant or outdated pa- rameters, incorporating research with defined endpoints, or making adjust- ments in the light of new information; for narrative water quality objectives; · lack of explicit conceptual designs that link monitoring to specific hypotheses or paradigms about the ocean environment; · inability to address regional or cumulative effects in the bight as a
118 over-commitment of resources to well-understood problems; lack of a data management system containing a wide range of data types from all major monitoring programs; · absence of synthesis that provides usable information to managers and other decision makers; and · inability to effectively report the overall status of the resources and water quality in the bight to the public, the scientific community, and policy makers. It should be emphasized that this consensus reflects the judgment of many people actively involved in designing, carrying out, and using the data from monitoring programs. Thus, in spite of the strengths mentioned above, and the fact that monitoring data have been used in decision making, there is evidence that the monitoring system could be more efficient, focused, and comprehensive. INSTITUTIONAL OBJECTIVES AND THEIR LIMITATIONS As described in Chapter 5, the objectives that motivate marine mon- itoring can be considered as a hierarchy or continuum. This begins with broad public concerns about public health and the status of marine re- sources; extends through laws, regulations, and permits; and ends with the specifications of individual monitoring programs. In Chapter 3 the public's concerns were reviewed in the section "Public Concerns for the Bight," while the laws that furnish the regulatory context for monitoring were reviewed in `'The Regulatory Sector." Finally, the structure of effluent limitations and water quality criteria was described in Chapter 4 in "The Monitoring Sector." These objectives influence the design of monitoring programs. They also influence the institutions that oversee the monitoring system. As a result, objectives are expressed explicitly in permits and other documents and implicitly in the behavior of the institutions that regulate monitoring. The following two sections address each of these aspects in turn. Objectives Because of the vast number of parameters that could be measured in the marine environment, monitoring programs require clear and precise objectives. The numeric effluent limitations and water quality criteria in discharge and other permits provide such precision. However, the narrative water quality criteria relating to unacceptable degradation or change do not furnish this level of precision. For example, the NPDES permit for the County Sanitation Districts of Orange County states that marine communities shall not be degraded. 1b monitor degradation in fish
119 communities, a program could legitimately focus on any of the following parameters: diversity, species richness, community trophic structure, relative abundance of numerically dominant species, population sizes of numerically dominant species, population sizes of trophically important species, size-age relationships, · reproductive potential as measured by gonad weight, · mortality of one or more species, incidence of fin rot, tumors, and other abnormalities, or body burdens of specific contaminants. . Although these are all measurable parameters that may be indicators of degradation, they do not define it. 1b design a monitoring program with the objective of ascertaining "degradation," the term must be defined in a meaningful way. Thus, monitoring program objectives should be stated as clear, preferably quantitative, questions or null hypotheses: for example, a program could be designed to determine if the three most abundant fish species within 3 mi of the Orange County outfall had decreased in abundance by more than 50 percent from one year to the next. Such a decrease might be defined as a degradation of these fish populations. One of the most comprehensive efforts to state monitoring objectives in Southern California is an Environmental Protection Agengy (EPA) doc- ument titled Objectives and Rationale for the County Sanitation Dismcts of Orange County 301(h) Monitonng Program. For each program element, objectives of the relevant laws and regulations are stated, and sampling and analysis plans are specified. Objectives are precisely stated for in- fluent, source control, effluent, and solids-handling monitoring. Although objectives for receiving-water monitoring are stated more clearly than ever before, they still contain no quantitative criteria for the kinds or amounts of change that should be monitored for. This is an important shortcoming be- cause receiving-water monitoring focuses directly on determining whether human and ecosystem health objectives are being met. This demonstrates that another level of detail is needed if monitoring in the bight is to consistently provide useful information. It should consist of specific descriptions of the kinds of changes, along with quantitative criteria about the amount of change, that should be monitored for. Hypothetical examples of such objectives, framed as null hypotheses, might be as follows: · The operation of diffusers for the discharge of cooling water will not decrease the monthly average light transmission in the upcoast quarter
120 of the adjacent kelp bed more than X percent below light transmission in the downcoast quarter of the kelp bed · The area around the sewage outfall outside the zone of initial dilution (ZID) exhibiting a change in benthic diversity of X percent ox more shall not increase from year to year. Background diversity shall be defined as that found at reference stations A, B. and C. . The long-term trend of DDT in the muscle tissue of adult Dover sole from the Palos Verdes Shelf shall not increase. Long-term shall mean a period of five years or more, and sampling shall be designed to detect a change in the long-term average of at least 5 percent. These null hypotheses define a specific parameter and the amount of change to be measured. Before actual sampling begins, additional detail relating to confidence limits, background leYels, and over [actors must be decided. In the first hypothesis above, locations (surface, bottom, midwater, water column average), time scales (daily, weekly, monthly averages), and distribution of sampling stations must all be established. These decisions can be made with the support of the technical design tools outlined in Figures 5-1 to 5-4. In contrast to most objectives used as the basis of receiving-water monitoring, the three examples above provide the founda- tion for focused, efficient monitoring programs. In contrast to other major monitoring programs in the bight, the Ma- rine Review Committee (MRC) programs around the San Onofre Nuclear Generating Station (SONGS) were all designed with a specified probabil- ity of detecting definite amounts of change (Chapter 54. This policy was based on predictions of impacts and on a management decision that these amounts of change would be accepted as evidence of power plant impact. There are two impediments to establishing this detailed level of objec- tives: (1) incomplete scientific knowledge (for example, an inability to es- tablish source/receptor relationships), and (2) the institutional environment of monitoring. The environmental effects of all human activities cannot be predicted accurately. Where they cannot, objectives must necessarily remain more subjective, or research should be performed. In other cases, however, environmental effects are well enough understood that reason- ably accurate predictions could be used to design more efficient monitoring programs. The changes that occur in the benthos around municipal waste discharges are a case in point. Changes in community composition, abun- dance, diversity, etc., have been well documented and could be used to develop more ecologically relevant and precise receiving-water objectives. Even where clearer and more quantitative objectives could be developed, however, there may be institutional constraints against implementing them.
121 For example, quantitative receiving water objectives could decrease regula- tory flexibility if they were rigidly interpreted as a measure of compliance and automatically triggered management actions or litigation. Despite these impediments, clearer monitoring objectives would result in beneficial gains in clarity, efficiency, and useful information. These gains would make the effort involved in developing them and integrating them into the regulatory framework worthwhile. In spite of these benefits, a danger of quantitative monitoring objectives is that they may be applied blindly, with little regard for naturally occurring effects. For example, between 1973 and 1977, there was a massive influx of the echiuran worm Listnolobus into benthic communities in the bight (Stull et al., 1986~. This organism's burrowing, respiratory, and feeding activities aerated and reworked sediments throughout the bight. In areas of wastewater impacts (particularly White Point on the Palos Verdes Shelf) these activities reduced apparent impacts from me Los Angeles County outfall. When the worm disappeared, impacts reappeared. Without awareness of this naturally occurring but anomalous and confounding event, the strict application of quantitative criteria would have led to the erroneous conclusion that impacts of wastewater outfalls had decreased and then increased. Institutional Limitations The statutory and regulatory framework within which monitoring is conducted in Southern California has evolved piecemeal over time, and as a result, deficiencies and inconsistencies exist within the institutional structure. These affect not only the way monitoring is carried out but also the quality of the information monitoring can produce. The most important of these limitations are: · lack of attention to nonpermit activities that may have large envi- ronmental impacts; · rigidity and lack of flexibility; and · a piecemeal, permit-by-permit approach to problems that may ac- tually be larger in scope. These limitations will be discussed in turn and illustrated with specific examples. Nonpermit Activities The vast majority of monitoring in the bight is compliance monitoring; that is, it is required as a condition of obtaining a permit. The unstated assumption underlying this policy is that the permitting process addresses
122 all aspects of discharges and other activities that potentially affect the en- vironment. This is not always the case, however, since some large inputs of contaminants are not covered by permits. These include rivers, which con- tain runoff, treated municipal waste water, and upstream discharges; storm drains; fallout of airborne pollutants; and diffuse inputs of hydrocarbons and other contaminants from marinas and harbors. Although rainfall is sporadic in Southern California, winter storms can dump 1 to 3 or more inches of rainfall within 24 hours, washing accumulated contaminants from streets, sidewalks, and other surfaces into rivers and storm drains, where they are carried out to the ocean. The river system in the Los Angeles basin (Figure 1-2) drains a watershed of over 4,100 mi2. During a major storm, the Los Angeles River alone can discharge 65 billion gal of water during a 24-hour period. Additional runoff enters the ocean directly from storm drains. For example, 75 separate storm drains discharge into Mission Bay in San Diego. Many of the industries that discharge into rivers and storm drains operate under National Pollutant Discharge System (NPDES) permits, and there is some monitoring in the Los Angeles basin rivers. However, many river and storm drain inputs are not monitored, and the system as a whole is not managed as a source of contamination. The bight is adjacent to urban areas that are major sources of air pollutants. Aerial fallout to the ocean surface constitutes a significant source of contaminants (e.g., Able 2-2~. The many marinas and harbors are sources of hydrocarbons and other contaminants derived from bilge pumping, sewage discharge, fuel loading and transfer, marine construction and maintenance activities, and ship traffic. Therefore, it is clear that monitoring to satisfy permit requirements does not address all of the large inputs of pollutants to the bight. Inflexibility Because monitoring programs are typically defined in regulatory per- mits, it is difficult to alter them as knowledge accumulates. The lengthy public hearing process required for updating permits has occasionally de- terred permittees from attempting to modify their monitoring programs. In addition, there is a natural reluctance to discard or modify parameters that have traditionally been measured, but which may now be outmoded. As a result, monitoring programs often include outdated or inappropriate mea- surements. Further, procedures that are experimental or in development have been incorporated as routine elements of monitoring, even though the data they produce are not adequate for decision making. Oil and grease (a generic contaminant category including petroleum, synthetic, and biological "oily" materials) are measured throughout the
123 water column as a part of several wastewater outfall monitoring programs. However, because most oil and grease float, and therefore are rarely found above detection limits in the water column, it is not cost effective to sample there. In addition, dissolved and dispersed oil and grease derive from many other sources, such as oil seeps, bilge pumping, aerial fallout, refinery effluents, stormwater runoff, and even from natural biological sources. Therefore, they are equivocal indicators, at best, of outfall impacts. It was suggested that floating grease balls, which can more directly be related to wastewater outfalls, would be a better indicator. Biological and chemical oxygen demand (BOD and COD, respectively) have traditionally been measured as part of benthic monitoring programs around wastewater outfalls. These parameters were originally included in receiving-water monitoring programs because they were used by sanitary engineers to monitor in-plant sewage treatment processes. There was a consensus among practitioners in the bight that these parameters are less biologically relevant in an open ocean environment and therefore cannot be meaningfully interpreted. It was suggested that measuring organic carbon and carbon flux, ammonia-nitrogen, and total nitrogen would be more ecologically meaningful (see pages 28 294. As a condition of their 301(h) permit, the County Sanitation Districts of Orange County are required to routinely measure a wide range of chemical contaminants, even though many of them are never found in effluent or sediments. This represents a large expenditure of resources where past experience has shouts there is likely to be little contamination. In contrast, in Los Angeles City's Hyperion monitoring program, the search for chemical contamination is more focused. Priority pollutants in the effluent are measured monthly (quarterly for volatile organics), thus providing regular information about what is entering the environment. During the first monitoring year, all priority pollutants are measured in sediments, trawl-caught fish and invertebrates, and sport fish. Contaminants that were not found in the first year are not monitored during the second and third years. In the fourth year, the entire range of priority pollutants is measured again. The city of San Diego is required to monitor suspended solids in the water column around the Point Loma wastewater outfall. However, because sampling stations are near the Point Loma kelp bed, the suspended solids samples sometimes contain larval crustaceans or pieces of kelp, seriously compromising the utility of this outfall plume indicator. More useful approaches here might be to measure light transmission or use sediment traps to determine fluxes of suspended particles in the water column. The location of sampling stations can also be inappropriate. The sam- pling grid around the Point Loma outfall contains a southern control station
124 that is of little or no use as a control because it is close to a dredged ma- terial disposal site and the sediments are predominantly extremely coarse sand. Even assuming that movement of material from the disposal site has not compromised the control station, the unusual sediments will necessarily be associated with a different benthic community, making meaningful come parisons with the outfall stations difficult if not impossible. At the northern end of the sampling grid, the city's permit required sampling a control station called B-2, located in 50 It of water. This station was sampled for years, but was never used in analyses because there were no other stations at this depth. A transect had originally been planned at 50 ft. but all the stations, with the exception of B-2, were located in areas of rocly bottom, where benthic grab sampling was impossible. The city requested that it be allowed to stop sampling El-2 and instead add a control station at 150 ft. This would have been a more efficient use of resources because the sampling grid already included a transect at the 150-ft outfall depth, but lacked a control. Implementing this change in the sampling design required several years and a public hearing, at a cost of wasted sampling effort at B-2 and reduced ability to monitor impacts at 150 ft. As part of its NPDES permit to discharge cooling water from coastal power plants, the Southern California Edison Company is required to monitor for thermal effects on marine resources despite the fact that nearly 20 years of studies have documented the limited nature of these effects. This example is indicative of the lack of clearly defined endpoints in monitoring studies, which hinder reallocation of monitoring resources to unresolved or . . more pressing Issues. Histopathology, tissue analysis for contaminants, and enterococcus measurements have been included as routine parts of monitoring programs, even though many participants in the case study believe they require more research and development before they can provide useful information. The panel stresses that these comments derived from a sincere desire to produce useful information and a frustration with requirements to perform studies whose results are ambiguous or uninterpretable. Several unresolved issues apply to tissue chemistry studies. The basis of presentation of data has not been standardized, making it difficult to interpret and compare results. For example, data may be presented on a dry weight or lipid weight basis, with each method presenting a different picture of contaminant levels. The problem of confounding due to seasonal and reproductive cycles also has not been resolved. In the spring and summer, fishes' reproductive season, fats are mobilized and transferred from the liver to the gonads. This may affect contaminant levels not only in these tissues but in others as well (Cross et al., 1986~. There may be differences in both the timing of reproductive cycles and in tissue chemistry between different species. However, because it is not possible to predict which species will be
'125 abundant enough for tissue chemistry studies at any one time, dischargers are allowed to sample species of opportunity. This means that no two dischargers consistently sample the same suite of species at the same time. It also means that the same discharger will sample different species in successive surveys. Given the unresolved issues related to seasonal cycles and interspecies differences, the lack of consistent target populations makes it extremely difficult to interpret tissue chemistry data and relate them to discharges. The issues of standardization of measurement techniques, seasonal physiological changes, and inconsistent target species also plague histopa- thology studies. In addition, the interpretation of histological changes in marine organisms can be demanding and ambiguous, and it was suggested by several participants that this technique is not yet suitable for routine monitoring. In contrast to these two examples of incompletely developed tech- niques being used as routine monitoring tools, the city of Los Angeles' Hyperion monitoring program includes a microlayer study that is explicitly experimental in design. The permit states that the first-year sampling re- sults will be used to determine the scope and direction of future monitoring. It also defines first-year requirements of an otter trawl sampling program and stipulates that first-year data be used to refine the sampling design for subsequent years. In addition, Hyperion's permit includes specific language that allows for further flexibility as needed (see pages 63 and 65~. These ex- amples suggest that permits can be structured to be flexible and adaptable. This produces two important benefits. First, it allows for improving and refining monitoring programs as data become available. Second, it allows resources to be used more effectively by recognizing that some questions are more appropriately dealt with in a research context than in routine monitoring. Repeatedly collecting the same data over and over again is not always the best way to address unresolved questions about the utility of new technical methods. The Southern California Edison Co. recognized this when it began its program of special studies in the marine environment (see Chapter 4~. The special studies were explicitly experimental in nature because it was understood that it is often difficult to define research programs succinctly enough to make them part of routine monitoring. They produced information that was important in understanding and reducing impacts without becoming a part of routine monitoring activities. On the other hand, Edison personnel pointed out to the case study panel that they found the data from mandated monitoring programs based on conventional measurements to be of relatively little value in managing marine resources.
126 Permit-by-Permit Approach The existing regulatory framework necessarily forces monitoring into a permit-by-permit approach to environmental problems in the bight. This results in monitoring programs that look at each activity in isolation from all others. Taking monitoring results at face value requires making two related and scientifically dangerous assumptions. The first is that there are no cumulative, overlapping, or interactive effects. The second is that the measurements taken to document the effects of a particular activity reflect that activity and no others. Neither of these assumptions is especially robust, as several examples will make clear. The County Sanitation Districts of Orange County carry out a mon- itonng program around their wastewater outfall. Within or very near the sampling grid are other biological and physicaUchemical patterns that in- teract with the effects of the outfall. On the eastern edge of the sampling grid is an active EPA interim-designated, dredged material disposal site for dredged material from upper Newport Bay. This dumpsite is in temporary use, and many of the contaminants found in the outfall effluent are also found in the dredged material. Just inshore of the outfall is the mouth of the Santa Ana River, which seems to be associated with a plume of modified sediments that affect benthic community patterns in the sampling grid. On the western edge of the sampling grid is a region of elevated contaminant levels of unknown origin. The permit-by-permit approach makes it more likely that these potentially confounding influences will be disregarded when designing a monitoring program for the Orange County outfall. The city of Los Angeles and the County Sanitation Districts of Los Angeles and Orange counties all carry out fish trawling programs around the Hyperion, White Point, and Orange County wastewater outfalls, re- spectively. These sampling programs are used to independently assess the effects of each outfall on fish populations in the region of the outfall. However, it is likely that at least some portion of the studied fish popula- tions moves throughout the entire area. This means that, for example, the monitoring program at White Point may actually also be measuring some effects of Hyperion and Orange County. The city of Los Angeles' trawl sampling program in Santa Monica Bay is designed to document effects of the Hyperion outfall on fish populations. However, the Southern California Edison Company and Los Angeles De- partment of Water and Power also operate coastal power plants in Santa Monica Bay. Entrainment of large numbers of fish larvae by cooling wa- ter intakes and impingement of adults may affect fish population sizes and community structure in the bay. In addition, some of the species monitored in the trawling program may spend part or all of the juvenile phase of their
127 life cycle in harbors and marinas in and around the bay. This example illustrates that patterns in fish populations (particularly population size and community structure) measured by the Hyperion monitoring program may actually reflect the effects of a suite of impacts, some of them occurring on other life stages than those targeted by the monitoring program. Other sources of effects were not incorporated into the design of the Hyperion trawling program despite the outfall's close proximity to coastal power plants; permitted and accidental discharges from oil refineries, stormwa- ter drains, and nonpoint sources of pollution; marinas; and contaminated juvenile habitats. The permit-by-permit approach to establishing monitoring programs also leads to important inconsistencies among monitoring programs. Some of these reflect the fact that permits were written at different times, with more recent permits incorporating more up-to-date knowledge. However, other inconsistencies reflect differences in approach or expertise among the regional water quality control boards and EPA Region IX personnel. As discussed more completely below, such inconsistencies make it difficult to develop an integrated view of impacts and trends in the bight as a whole. Specific examples of inconsistencies among monitoring programs in- clude the following: · The city of Los Angeles has a flexible approach to measuring pri- ority pollutants in sediments and organisms, whereas the County Sanitation Districts of Orange County measure priority pollutants regularly. Bawl sampling around wastewater outfalls is usually conducted quarterly or semiannually, but trawl sampling around coastal power plants is conducted every two months. . . The city of Los Angeles conducts offshore water quality sampling weekly because its discharge is near areas of intense water-contact recre- ational areas, whereas the County Sanitation Districts of Orange County conduct offshore water sampling monthly for some parameters and quar- terly for others. . No two dischargers consistently use the same organisms for tissue chemistry measurements. . The city of San Diego is not required to conduct trawl, rig fishing, or tissue chemistry studies, although other dischargers are required to do so. However, trawls are performed on a voluntary basis to contribute to a regionwide assessment of fisheries resources. TECHNICAL DESIGN AND IMPLEMENTATION This section summarizes the extent to which monitoring programs in the Southern California Bight meet the criteria for technical design presented in Figures 5-1 to 5-4. The discussion is organized around
128 · the issues of statistical design of monitoring plans, · the establishment of field and laboratory procedures that ensure valid, high quality measurements, and data management strategies. Statistical Design There is still room for improvement in how statistical tools-quantita- tive null hypotheses, statistical models, quantification and partitioning of variability, optimization analyses, and power tests, for example are applied to program design. These tools are beginning to be applied to monitor- ing programs in the bight, and the EPA has produced 301(h) guidance documents that provide instructions for their use; however, lack of clear quantitative objectives prevents effective application. New monitoring tools can be properly applied only in the context of clear statements of manage- ment needs and the questions and/or hypotheses that resect them. The following examples illustrate this point: · Power tests can estimate the likelihood that a sampling plan will detect a change, such as an increase in the diversity of the benthic infaunal community of 0.1, 0.2, 0.3, 0.4, etc. Without guidance from regulations or ecological theory about what specific amount of change is important, it is still possible to perform power tests for a wide range of possible changes, then choose the sampling plan that is most likely to detect change (any change). However, a more useful approach would be to decide that a specific increase of 0.3, for example, is a strong indicator of outfall enrichment effects, then use power tests to design a sampling plan with a high probability of detecting that precise amount of change. · Measurements of background variability can be extremely useful in designing efficient sampling plans. In spite of this, great time and expense could be wasted attempting to measure variability on all scales (e.g., feet to hundreds of miles and days to decades). However, if managers determine that only present effects within 6 mi of an outfall are of interest, other variability scales can be Reemphasized. If managers are also interested in change from year to year, annual background variability would [become relevant. If managers are interested in longer-term trends more than 10 years, for example then interannual variability on that time scale would become relevant. · There has been discussion in the development of 301(h) monitoring plans about the proper number of "replicate" benthic grabs to take at each station. This discussion has used the results of technical design tools such as power analysis. Even these tools cannot resolve the issue because there is no one right number of replicates to collect. The proper number depends on the questioners) being asked, the amount of predicted change sampling
129 should detect, and the sources of variability that could obscure monitoring results. This last point deserves further discussion because of the mistaken assumption that the same number of replicates is appropriate for all situa- tions. As one example, if concern is focused on the difference in diversity inside and outside the ZID boundary at one point in time, then a different number of grabs at each station may be required than if the concern is about how the relationship between diversity inside and outside the bound- ary changes over five years. Further, if concern is focused on how diversity inside the ZID changes over five years in response to changes in the output of suspended solids, then another number of grabs might be appropriate. Some of the deficiency in the consistent and proper use of technical design tools in monitoring programs in the Southern California Bight stems from the incorrect use of statistical concepts. Lao such important concepts are "significance" and "replication." Portions of permits and regulations state that a particular activity shall not cause a "significant" alteration, change, decrease, or degradation in some physical, chemical, or biological parameter. The California ocean plan (State Water Resources Control Board, 1987) defines a "significant" difference as '`a statistically significant difference in the means of two distributions of sampling results at the 95 percent confidence level." The problem with this definition is that it provides no guidance in determining how large a change is of importance and should therefore be detected by a monitoring program. This is because virtually any change can be a statistically significant difference, depending on the intensity of sampling. Thus, a monitoring program with a low intensity of sampling will find only large changes to be statistically significant, while one with a high intensity of sampling could find even minuscule changes to be statistically significant. Permits and regulations should replace the word "significant" with another such as "meaningful" or "important" and then define the terms clearly. There is an emphasis on replication in Southern California Bight monitoring programs but no equivalent awareness that replication has at least two different meanings, and that many aspects of a sampling plan can conceivably be replicated. Replication is loosely used to refer to the process of collecting repeated measurements, samples, or comparisons. However, in a stricter sense, it refers to the process of repeating entire experimental treatments. In addition, a sampling plan may have many levels of sampling, any and all of which may be repeated. For example, a monitoring program set up to determine whether benthic infaunal diversity inside the ZID is decreasing over time with respect to diversity outside the ZID might include: . several stations inside the ZID I,
130 · several stations outside the ZID, · one or more "replicate" grabs at each station, and · several sampling periods over time. This sampling plan thus includes replicate grabs at each station, repli- cate stations within each area, and replicate times or surveys during which all stations are sampled. Depending on the resolution desired, technical design tools such as power and optimization analysis might indicate differ- ent numbers of "replicates" at each level of sampling (e.g., two grabs per station, five stations per area, and nine surveys over time). When different kinds of "replication" are not clearly distinguished, monitoring programs tend to emphasize repeated samples at a single place and time. A balance has to be struck between extensive replication of all samples and spreading limited sampling resources over other levels of a sampling plan. Field and Laboratory Procedures Many field and laboratory procedures are of commendable quality in Southern California monitoring, where an attempt is made to use state- of-the-art methods, particularly in the larger programs. In addition, an emphasis on improving monitoring methods has resulted in standardization of invertebrate taxonomy, benthic grab sampling techniques, and chemical analysis procedures. Monitoring programs at the municipal wastewater discharges benefit directly from research carried out at SCCWRP. New questions and new methods of sampling and analysis have been incorporated quickly into ongoing monitoring programs. Although monitoring methods are state of the art, they may not al- ways be adequate to address monitoring objectives. Such an example was described above with reference to tissue chemistry and histopathology stud- ies. As another example, public health surveillance methods are not precise enough to detect brief episodes of mild illness among swimmers due to bac- terial or viral agents in marine waters. In addition to the epidemiological problems, studies of putative viral agents are hampered by lack of culture techniques. There is growing recognition that there may well be a better indicator of fecal contamination than the coliforms (i.e., the enterococcus group), and health agencies are actively acquiring information to assess these new indicators. Because epidemiological studies are expensive to perform and marine epidemiological studies often yield equivocal results, especially when performed on a small-scale local basis, state and federal public health and water quality agencies have been reluctant to fund such studies.
131 Data Management Data management is vitally important to monitoring efforts because it determines the final accessibility and utility of the data. Data management should include quality control procedures that ensure data accuracy at every step from initial collection to final analysis and reporting. It should also include methods for making the data readily available in usable formats to those responsible for analyzing and examining them. Another important but little-recognized aspect of data management is the importance of specifying data tabulation methods, structures, and handling procedures before a sampling program starts. This allows data to be collected and processed in ways that are appropriate to their final use, dissemination, and storage. This specification of data management procedures at the beginning of a program can save significant effort and money that would otherwise be spent correcting errors in raw data, analyses, and reports. At present, there is a wide variety of approaches to marine monitoring data management in the bight. In spite of this variety, the panel found that the major monitoring programs all have well-developed and active systems for ensuring the accuracy and quality of their raw data. These data are continually reviewed and updated when necessary. The following examples are representative of data management approaches in the bight. The 301(h) programs configure their data in the National Oceano- graphic Data Center (NODC) format and are now required to submit monitoring data to the EPA Ocean Data Evaluation System (ODES). ODES, designed to provide ready access to 301(h) data, has recently be- come fully operational and includes formal quality control procedures. However, not all historical outfall monitoring data are in digital format. For example, the Los Angeles County sanitation districts have computer- ized past monitoring data from the White Point outfall, whereas such data from the County Sanitation Districts of Orange County are available only in written reports. Data from the California Cooperative Oceanic Fisheries Investigation (CalCOFI) program are in NODC format and are available in published data reports. The Southern California Edison Company maintains its own data base for a wide range of monitoring data. The National Marine Fisheries Service and the California Department of Fish and Game have fisheries monitoring data available on magnetic tape; however, these agen- cies do not maintain user-oriented data bases to provide access to these data. Scientists at the Scripps Institution of Oceanography monitor temper- ature and wave energy and provide these data on magnetic tape on request. Data from smaller studies (e.g., Los Angeles Harbor, Marina del Rey) are typically stored on floppy disks or on consultants' computer systems. The city of San Diego and the County Sanitation Districts of Orange
132 County have initiated analogous programs to centralize and automate their in-house data management procedures. These systems provide computer- ized data entry functions that automatically perform quality control checks on a range of raw data. Validated data are stored in a centralized data base and a set of menu-driven options allow users to update and extract data. Additional menu options permit users to automatically produce stan- dardized regulatory reports and automatically format data for submission to ODES. Finally, the systems incorporate links to a variety of analytical tools, such as spreadsheets and analysis and graphics software. The taxonomic efforts of the Southern California Association of Marine Invertebrate Taxonomists (SCAMIT) and the ODES data base represent important steps in setting consistent standards for standardization, quality control procedures, error checking, and digital formats for monitoring data. However, there is currently no easily accessible, user-oriented data base sys- tem to provide access to analysts interested in integrating data from several different kinds of studies. Such a system would greatly facilitate attempts to study regional and longer-term questions related to environmental effects in the bight. There are two prototypes for such a system, each with its own strengths. These are the operational environmental data base developed by the En- vironmental Research Group of Southern California Edison and ODES. Both systems are unusual in that they include extensive quality control procedures and on-line documentation and are designed to permit data analysts to use menu-driven routines to readily extract data needed for analyses. Southern California Edison's system was designed to perform the following functions: . store corrected and updated archival versions of important data sets so that all analysts access the same version of the data; · store important data sets in a data base management system that provides the ability for easy extraction, updating, and manipulation of data; · provide comprehensive on-line documentation of methods, error corrections, data characteristics and peculiarities, and publications for each data set; · provide automated browse, search, retrieval, and reporting facili ties; · provide flexible links to the Statistical Analysis System (SAS) and other data analysis systems; and · allow easy addition of novel data types to the system. This system is fully operational and contains a wide variety of mon- itoring studies in standardized formats, thus facilitating comprehensive analyses. These studies currently include:
133 · benthic infauna and sediment data from monitoring programs at San Diego, Los Angeles city and county municipal wastewater outfalls; · Southern California Coastal Water Research Project's (SCCWRP's) 198-ft (60-m) survey; · Scripps' shoreline temperature data for the west coast of the United States, and wave energy and wave direction database; · California Department of Fish and Game sportfish catch; National Marine Fisheries Service commercial fish catch data; benthic infaunal and sediment data from the Bureau of Land Management (BLM) study of the bight; · complete impingement data for all Southern California Edison coastal power plants; · data from bightwide ichthyoplankton studies and fish trawl studies performed for Southern California Edison; and · selected Marine Review Committee studies. This system is proprietary and is not accessible to scientists outside of Southern California Edison. It does, however, illustrate that such com- prehensive databases can be constructed. The main strength of Southern California Edison's system is that it contains a wide range of data from biological and physical oceanographic studies that are bightwide in scope. The experience of constructing this database substantiated the fact that locating, acquiring, correcting, and standardizing disparate data sets is a significant effort. The other system that points the way toward bightwide data manage- ment is ODES. ODES is intended as a national database to contain 301(h) monitoring data, which includes (among others) benthic infauna and sedi- ment chemistry, otter trawl, water quality, and other data types. It includes a wide range of menus that assist users in extracting and combining data from different studies, in performing common types of analyses, and in creating maps and graphics. In addition, ODES provides for extracting raw data for analysis with other software packages. Despite its strengths, ODES has shortcomings that restrict its utility and that must be corrected in any future system that successfully provides access to a range of monitoring studies. There is widespread dissatisfaction with ODES within the Southern California monitoring community. This dissatisfaction results from the difficult and labor intensive procedures required to format data for submission to ODES. It also stems from the lengthy wait required for feedback to requests for new species codes and answers to technical questions. There is therefore a long delay between the initiation of the submission process and the final availability of the data. Users of ODES have access only to the analysis and reporting routines that have been programmed into the system. While many of these are
134 very useful, they do not cover the full variety of approaches required for a comprehensive analysis of monitoring data. Requests for additional analytical tools must wait until they can be programmed into the system, since ODES does not allow users to directly access other analysis systems. Users can, however, extract data from ODES and download them to their own computer systems. Another shortcoming is that when new data types are encountered, ODES must be reprogrammed to accept them, a process that can take several months. In contrast, data base systems that are designed for adaptability use table-driven data definition approaches to allow for rapid modification of data base structures. ODES provides the ability to combine data from more than one study in order to perform regional or national analyses. However, in practice this capability is severely limited because ODES lacks an aggressive program to update data sets in the system and to standardize taxonomy among data sets. Experience in the bight has shown that such taxonomic updating and standardization is crucial if data sets are to retain their utility and if different studies are to be combined. Species names, particularly of benthic invertebrates, change continually over time as scientists adjust taxonomic affinities. Thus, for data sets to remain current, even historical data must be updated regularly. Autonomic standards invariably differ among different studies. This is true even when efforts are made to use common standards. Thus, in order for data from two or more studies to be combined, careful attention must be paid to reconciling these superficial dissimilarities. As a result of the lack of such updating and standardization procedures, only analyses that do not depend on merging or matching species data can be performed with ODES. Such analyses include those using derived variables such as diversity indices, total abundance, numbers of species, or summaries of higher taxonomic groups. TECHNICAL INTERPRETATION AND DECISION MAKING The ultimate goal of monitoring is to provide data and information to support informed decision making. In this section, the technical inter- pretation of data obtained in monitoring programs and its use in decision making are addressed. Some examples show that monitoring data have been adequately interpreted and used in decision making. Overall, how- ever, considering the effort that has been put into data collection, no comparable effort and expense has been devoted to translating that data into useful information and using it in decision making. In spite of the shortcomings in the interpretation and decision-making process (reviewed below), it is important to recognize that monitoring in- formation has played a significant role in many far-reaching management
135 decisions in the Southern California Bight. Water quality and bacteriologi- cal monitoring data from Santa Monica Bay documented the severity and extent of nearshore contamination from sewage discharges in the 1940s and 1950s. These data helped make the case for construction of offshore outfalls in 1957 and 1959 that dramatically reduced nearshore sewage con tamination. In 1977, the California Department of Fish and Game closed the abalone fishery from Palos Verdes Point to Dana Point. This decision was based on monitoring surveys and catch data. As another example, scientists of NOAA's Ocean Assessments Division have used data from SCCWRP and the municipalities to evaluate environmental conditions relating to the body burdens of chlorinated hydrocarbons in coastal marine organisms (Mearns and O'Connor, 1984; Matta et al., 1986; and Mearns and Van Ness, 1987~. The inability of the city of San Diego's Point Loma wastewater treat- ment plant to consistently meet bacterial standards contained in the 1983 California ocean plan (State Water Resources Control Board, 1983) for offshore kelp beds contributed to a decision by the city to extend its outfall farther offshore. Earlier monitoring data generated by Southern California Edison Company showed that unacceptably large numbers of fish were being taken into cooling-water intakes of power plants. As a result, intakes were redesigned with velocity caps and other changes to reduce entrain- ment and impingement. Monitoring data were then used to confirm that the design changes were effective. Data generated over the last eight years by the Marine Review Com- mittee on the environmental impacts of SONGS will be used to make decisions about changes in the design or operation of the cooling-water sys- tem. These data will also be used to support the development of mitigation measures to offset Impacts documented through monitoring. The recently released first-year report for the 301(h) monitoring program performed by the County Sanitation Districts of Orange County resulted in adjustments to the districts' permit. In addition, the data in the report suggested that no changes were needed in the waste discharge or treatment processes. By far the greatest effort in data interpretation between the 1950s and the present has been the work of SCCWRP scientists. Starting with the 1973 report on conditions in the bight and implications for management (SCCWRP, 1973), their periodic reports and scientific journal publications have become internationally recognized. Although their work has included much more than evaluation of routine monitoring data, it has resulted in improved monitoring methods and in quality control activities that increase the reliability of the data. In fact, the scientific publications of the majority of SCCWRP scientists are cautious, if not silent, on interpretation of moni- toring data with respect to regulator actions. Instead, their interpretations
136 generally focus on environmental conditions and, to a somewhat lesser extent, on possible impacts of pollutants. On a smaller scale, the Channel Islands National Park monitoring program has generated data since 1981 from diving surveys at 14 stations, conducted primarily by volunteers. These data are used to make decisions about visitor access, harvesting of resources, and development of the park resource. As another example, the program conducted by Occidental College for Southern California Edison was originally related to monitoring the effects of waste heat discharge from coastal power plants. It has also yielded useful resource information on a sedentary reef fish community. This latter example demonstrates that if data were made available scientists would find monitoring programs useful for filling in information gaps about marine resources. In many instances, the use of monitoring data is not as clearcut as in the examples just cited. In some cases, it is difficult to document whether decisions were based on monitoring results, particularly when decisions were made not to change existing procedures. In some instances, disagreements about the interpretation of data can hamper the ability to make resource management decisions. For example, during the 1940s and 1950s, major differences of opinion among scientists working on sardines hindered implementation of the management measures needed to protect this fishery resource (Baxter, 1982~. Scientists from the U.S. Bureau of Commercial Fisheries contended that year-class size was independent of the size of the spawning stock and that catch size therefore had no effect on stock size in subsequent years. California Department of Fish and Game scientists believed that there was a strong link between year-class size and spawning stock size. By the time the disagreement was resolved in 1966 in favor of the Department of Fish and Game, the fishery had collapsed. Complicating such scientific uncertain is the fact that the societal implications of resource decisions can be quite extensive. Thus, decisions based on limited data impose risks that managers have to weigh against expected benefits and the time constraints of required actions. For exam- ple, decisions involving the economic livelihood of fishermen who harvest pelagic fish stocks may require a decade to correct if the result of the decision is not as expected. In fact, a decade or more may sometimes be required to produce a signal sufficient to determine if the decision was correct. In addition to scientific uncertainty, institutional limitations can limit the effective use of monitoring information in decision making. All too frequently, data reports sent to regulatory agencies are not subjected to thorough scrutiny and summarized for policy makers and the public. This is because the human resources and budgets of the regulatory agencies are
137 inadequate to interpret the growing masses of data generated each year and translate them into information useful to environmental managers and policy makers. Dischargers and other permitters often perform extensive analysis and interpretation of monitoring data. However, their reports are usually too lengthy and detailed to be readily accessible to policy makers and the public. In most cases, budgets earmarked for data analysis and interpretation by both the regulatory agencies and the permittees are judged inadequate. It was the consensus of the case study participants that monitoring data were incompletely synthesized and inadequately used in decision making. This is unfortunate because many monitoring reports contain extensive data sets that are not available in scientific journals even though they are peer reviewed to rigorous standards. In spite of this, some are suspect because the quality and quantity of the reviews are not documented. A statement at the beginning of such reports documenting the review process would have a favorable payoff in building confidence among the aware lay public who are trying to sort out technical issues. There are some exceptions to this generalization (for example, Matta et al., 1986) that provide both data, frequently from monitoring programs, and analysis of data. These are widely distributed and are cited in many regulatory documents such as the 301(h) decision documents. Another institutional limitation derives from the differing responsibili- ties of the various regulatory agencies involved in managing monitoring ac- tivities. The EPA acts primarily as an enforcement and compliance agency. The state of California, through the State Water Resources Control Board has primary responsibility for the development of ocean policy in general, represented by the California ocean plan (State Water Resources Control Board, 1987~. Evaluation of monitoring data is one part of the process of developing this policy and the specific regulatory actions intended to implement it. The state board establishes overall policy and the regional water quality control boards determine individual permit requirements. Both the EPA and the regional boards believe that most monitoring programs are well planned, well executed, and yield data that are useful in demonstrating compliance and in documenting regulatory changes. The state board, however, has the additional responsibility of identifying bene- ficial uses of marine resources and establishing water quality objectives to protect those uses. The state board staff believe that the question, "~e beneficial uses being protected?" is of more fundamental importance than mere compliance, but that monitoring data are not presently adequate to answer this question. As explained in the next section, this may be because the available monitoring data are not sufficient to fully address this broader question and because the specific questions are not asked precisely enough to guide monitoring efforts.
138 OVERALL ORGANIZATION OF MONITORING The preceding description and analysis of monitoring efforts in the Southern California Bight show that monitoring has achieved important successes. It has documented the extent of impacts from point sources such as power plants and wastewater outfalls. It has tracked the improvement of gross contamination in areas such as Los Angeles Harbor and the beaches of Santa Monica Bay. Longer-term studies, such as those carried out at the White Point outfall by the County Sanitation Districts of Los Angeles, have provided valuable insights into how human impacts interact with natural disturbances. However, the same analysis shows that the existing monitoring system does not address all important sources of impacts (e.g., storm drains). In addition, Figure 5-5 shows that many important resources are affected by more than one kind of human or natural perturbation. In spite of this, there are no monitoring programs that focus on resources by integrating data about the cumulative effects of more than one kind of perturbation. This is because the monitoring system derives predominantly from a fo- cus on regulating specific human activities, rather than managing natural resources. Finally, Figure 5~ shows that many contaminants and other sources of change act on time and space scales much larger than those of the typical monitoring program. As a result, the existing monitoring system has difficulty resolving bightwide patterns of change that may be just as important as the localized impacts that are the current focus of monitoring. In Chapter 5, four questions were identified as being especially per- tinent to evaluating the overall success of monitoring in the bight. These were as follows: objectives? Does monitoring address clearly stated management and societal · Does monitoring address the major environmental problems facing the bight? · Do the spatial and temporal scales of monitoring reflect those of the major environmental problems? · Are monitoring resources allocated effectively both within and among monitoring programs? The foregoing analysis provides the basis for answering these questions. In each case, the summary answers below are focused on assessing the performance of the monitoring system as a whole, rather than on individual monitoring programs. Objectives As described previously, there are different kinds of objectives that
139 motivate monitoring, from the broad concerns of the public to the detailed specifications of individual monitoring programs. These objectives can be classified as those pertaining to the effects of specific activities (e.g., dredging), to the overall status of important resources (e.g., kelp beds), and of the bight as a whole. Because of the institutional structure of the regulatory and permitting system, only the first of these is addressed in any detail by the existing monitoring system. In Figure 5-5, this can be represented as looking only at each row in isolation, ignoring both columns and the matrix as a whole. While objectives relating to measuring and managing the impacts of individual activities may not always be clearly stated, they nevertheless are the unmistakable focus of permits and monitoring programs. In contrast, important concerns about the status of resources and the bight as a whole are not manifested in the more detailed objectives that structure monitoring programs. Major Environmental Problems There can be no arguing with the fact that monitoring addresses many of the major environmental problems facing the bight. However, it is also clear that the existing monitoring system cannot address other problems that are just as pressing. These include nonpermitted sources, such as storm drains and atmospheric input of contaminants. They also include cumulative impacts stemming from the action of more than one kind of human or natural perturbation on a single resource. Finally, the existing monitoring system cannot adequately assess the existence and importance of large-scale and long-term environmental trends in the bight. The importance of these other environmental problems is a result of two major trends in the bight. First, increasing population and attendant utilization of marine resources have magnified the potential for cumulative and large-scale impacts. Sources of contamination and perturbation are more numerous and more closely spaced than in the past. Second, the existing monitoring and management system has been remarkably successful in removing gross pollution from the bight. As a result, concerns about cumulative impacts and subtle changes over time have become relatively more important. Spatial and Temporal Scales As a general rule of thumb, the spatial and temporal boundaries of a monitoring program should match those of the phenomena it is attempting to monitor. As Figure 5-6 shows, the spatial and temporal boundaries of existing monitoring programs match those of some but by no means all of
140 the relevant processes in the bight As a result, the existing monitoring system has only a limited ability to resolve trends and changes occurring on larger time and space scales. Such trends and changes can be natural, in which case they represent a moving background against which human impacts must be compared. Large-scale changes can also result from human impacts that by their nature cannot be restricted to one location (e.g., DDT contamination). The CalCOFI program (e.g., Chelton et al., 1982) and the Bureau of Land Management study of benthic communities in the bight (e.g., Thompson and Jones, 1987) provide examples of the abilitr of larger- scale sampling programs to describe important patterns that cannot be detected by point-source monitoring programs. Because monitoring occurs throughout the bight, the existing monitoring system has the potential for measuring events on larger time and space scales. However, this potential cannot at present be fully realized because separate monitoring programs are not sufficiently coordinated and integrated. Allocation of Monitoring Resources Despite the large amount of time and money (at least $17 million per year) spent on monitoring in the bight, it is not possible to perform all the monitoring that would be desirable given unlimited resources. The available resources should therefore be allocated based on criteria that prioritize environmental problems and impacts. Such a process should be based in part on an overall assessment like that summarized in Figure 5-5. At present, this is not possible. Each monitoring program is developed in- dependently, and its scope and cost are established in negotiations between the permitted and the regulatory agencies. As a result, some problems re- ceive a disproportionate share of monitoring resources while others receive little or none. SUMMARY The analysis of monitoring in the Southern California Bight led to conclusions and insights about individual programs and about the moni- toring system as a whole. In general, monitoring programs in the bight use state-of-the-art methods and produce accurate and reliable data. In addition? monitoring data have contributed to many important decisions related to pollution abatement and the management of natural resources. In general, monitoring has been successful in identifying and quantifying the impacts of such point-source activities as wastewater outfalls and coastal power plants. In spite of these successes, the panel found several shortcomings, some
141 related to the execution of individual programs and some to the institutional structure of the monitoring system as a whole. The most important of these were: . storing efforts; poorly stated objectives that provided insufficient guidance for mon · inability to monitor the effects of activities falling outside the existing permit structure; · inflexibility that inhibits needed adaptability; · overemphasis on a permit-by-permit approach to monitoring and environmental decision making, thus limiting the ability to monitor cumu- lative and large-scale impacts; · insufficient use of statistical design tools in the development of sampling and measurement plans; and · lack of a bightwide data management system to support integration and synthesis of data from different studies. The panel performed a preliminary synoptic assessment of environ- mental problems in the bight. This assessment, combined with the analysis of individual programs, led to important conclusions about the structure of the overall monitoring system. Because the existing system focuses on in- dividual permitted activities, it is unable to foster the higher level planning and coordination needed to assess cumulative and larger scale environmen- tal problems. In addition, the focus on individual human activities makes it difficult to focus on important resources that are affected by more than one type of impact. As a result, it is difficult to draw conclusions about the status of the bight as a whole and about whether beneficial uses of the marine environment are being protected.
