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.
V.â Harbor Seals A.â Background Harbor seals are widely distributed along the coasts of the North Pacific and North Atlantic. They can be found on exposed coasts and island archipelagos, but frequently inhabit estuaries or coastal lagoons such as Drakes Estero (Bigg, 1981). Intertidal sand banks provide habitat for seals to give birth and suckle their pups or rest during the nonbreed- ing season, and other estuarine areas provide foraging habitat (Wright et al., 2007) and areas where breeding adults engage in underwater display and aquatic mating (Van Parijs et al., 2000; Hayes et al., 2004). B.â What is the Body of Scientific Studies on the Impact of the Oyster Farm on Drakes Estero? Harbor seal research and monitoring projects have been conducted within Drakes Estero over the last 30 years, but none of this research was designed specifically to assess the impacts of mariculture operations. With the exception of individual-based studies carried out in the 1980s (Allen, 1988), research has focused on monitoring changes in abundance at haul- out sites and recording disturbances to hauled-out seals. The nature and intensity of surveys have varied over this period, and sub-site specific data exist only from 1995, when NPS initiated a standardized monitor- ing program. This monitoring program relies heavily upon the efforts of trained volunteers, and has since been integrated into a wider pinniped monitoring program across the San Francisco Area Network of Parks 41
42 SHELLFISH MARICULTURE IN DRAKES ESTERO (Adams et al., 2006). Trends in the abundance of harbor seals across the whole California coast are also monitored less frequently by state and federal agencies (Lowry and Carretta, 2003). The data used to assess annual trends in abundance and distribu- tion of harbor seals within Drakes Estero are from surveys made during the peak pupping season (March 15âJune 1) and molting seasons (June 1âJuly 30) (Hester et al., 2004). During each two-hour survey, hauled-out seals were counted every 30 minutes. The timing of these surveys was determined primarily by the need to provide the most robust estimates of abundance trends. The time and source of all disturbances were also recorded throughout the observation period. Disturbances are listed as head alerts, flush (seals move toward but not into water), and flush to water (seals leave haul-out and enter water). Although a head alert indi- cates a potential for a more serious response, it is less likely to represent an action that depresses fitness or has a negative population-level conse- quence, and may be less informative as a response variable (Jansen et al. 2006). Assessment of population trends relies upon the assumption that observer bias and individual haul-out behavior and, thus detectability, have not changed over this period. Haul-out behavior could change as a function of prey availability or the level and types of disturbance. Collectively, the data from these observer programs suggest that seals using the eight subsites within Drakes Estero are best considered a single unit, within which individual subsite choice may be influenced by factors such as disturbance (Allen, 1988). Mixing occurs between these seals and those at other local colonies, and there is more limited exchange among colonies outside the region. Consequently, any changes in abundance within the Drakes Estero/Limantour colony will result from a complex interaction between broader-scale drivers and local factors, such as dis- turbance. As examples, during the 1998 El NiÃ±o, adult and pup counts were depressed throughout the entire Point Reyes region, and in 2003 a northern elephant seal (Mirounga angustirostris) disrupted the Double Point harbor seal colony in Point Reyes, killing about 40 seals and prob- ably inducing emigration of others, judging from temporally correspond- ing increases in pup counts at Drakes Estero and Bolinas Lagoon (NPS, 2006a; Becker et al., 2009). Abundance trends within Drakes Estero should therefore be considered in relation to wider-scale population trends, but the time series of data is limited to the past 11 years (1997 through 2007), which is not sufficient to make a robust comparison with trends at other sites in the Point Reyes region with even fewer years of standardized count data. There has been one statistical modeling study that tested for poten- tial impacts of mariculture activity on harbor seals. Becker et al. (2009) examined how oyster mariculture activities are related to both interan-
HARBOR SEALS 43 nual changes in counts of seals at haul-out subsites closest to mariculture operations and also records of disturbance to hauled-out seals by cultur- ists, using annual oyster production levels as a proxy for mariculture activity. Becker et al. used data from 1997â2007 to assess whether counts of seals during the pupping season (April 15 to May 15) at different subsites within Drakes Estero varied in relation to oyster production levels in Drakes Estero, as well as to broader-scale effects such as density dependence (number of seals at haul-out sites) and the El NiÃ±o-Southern Oscillation (ENSO) events. In the statistical approach used by Becker et al., a statistically significant relationship was found between seal counts and years since the last ENSO event (positively related) and oyster har- vest levels for two of the three subsites in Drakes Estero that are closest to the mariculture operations (negatively related). Neither ENSO alone nor oyster harvest alone significantly related to seal counts on these haul out subsites. The statistical analyses serve as indicators of potential negative inter- actions between oyster harvest and seal attendance at these subsites but do not provide a causal link. Becker and colleagues did not have the official oyster harvest level for 2008 and hence did not include this 2008 datum in the statistical analyses. However, the relationship between the 2008 projected harvest and 2008 seal counts deviates from the pattern of the 11 previous years sufficiently as to call into question whether mari- culture intensity would still be a statistically significant contributor to explaining patterns of seal use of upper-estero haul-outs had the analyses included the full 12 years of data (from 1997â2008). In the paper, Becker et al. (2009) acknowledge the marked deviation of the 2008 data from the previous 11 years. The authors attribute the departure from the previous pattern to new regulations issued by the California Coastal Commission which closed the lateral channel to the oyster farmâs boats during the pup- ping season beginning in 2008 and further explain that this new restriction led to less disturbance of the seals and thus less displacement from their haul-out sites. However, this explanation is misleading because the previ- ous owner of the oyster farm operated under a 1992 agreement to prohibit boat traffic in the main and lateral channels during the pupping season (DOI, 2008) and the current owners maintain that they have voluntarily complied with the 1992 agreement. More generally, the use of annual oyster harvest as a proxy for disturbance rate at the haul-out subsites relies on the assumption that mariculture methods and daily activities have not changed over the 11 years that were analyzed. Some potentially confounding factors include changes in the fraction of the harvest from oysters cultivated using the rack system (mostly more distant from the haul-out sites) and the fraction cultivated in bags placed on sand bars also used by seals as haul-out sites. Changes in boat traffic patterns as a
44 SHELLFISH MARICULTURE IN DRAKES ESTERO result of agreements between the regulatory agencies and the owners of the oyster farm, and level of compliance, could also influence the rate of disturbance independently of the annual level of harvest. Although the NPS seal monitoring program at Point Reyes provides robust data on seal abundance trends, the disturbance data serve mainly as an indicator of a new source of disturbance or a large change in a known source of disturbance. These data have more limited utility in esti- mating changes in the level of disturbance within the estero or in the rela- tive importance of different disturbance sources for the following reasons. First, surveys consider disturbance only of groups of seals that are hauled out, rather than seals that are foraging or mating locally. Second, surveys are conducted towards the middle of the day to capture the peak counts, and about half of the surveys occur during the weekends. This captures sources of disturbance that occur during these times (such as hikers on weekends); whereas other disturbances may occur more frequently on other days or at other times of day. Finally, surveys will miss disturbance events occurring early in the ebb tide, before seals come ashore. These disturbances will not result in flushing but could depress the numbers of seals that haul out at that location at the low tide. The high level of overdispersion (variance greater than the mean and increasing with the mean) in Drakes Estero count data suggests that this may be a common occurrence. Assessments of the relative importance of disturbance from the oyster farm and from other sources are further constrained for two other reasons. First, responses of seals to different types of disturbance may differ. For example, if seals are flushed into the water by a hiker on the beach, it is possible that any perceived risk is reduced or absent once the seals are in the water. In contrast, if seals are flushed into the water by a motor boat, underwater engine noise may result in a continued perception of risk and a stronger response by the seal. Second, the lack of definitive data on spatial and temporal variations in the precise location and behavior of the oyster farm boats prevents any scientific assessment of the authenticity of either the observations of disturbance during the seal surveys (see Box 3), or the counter-evidence provided by DBOC that the boats were absent at these times. For example, âit was not uncommon for DBOC employees to take boats out into the estero after hours to fish,â according to Point Reyes National Seashore Chief Ranger Colin Smith as discussed in the DOI Inspector Generalâs report (DOI, 2008). If this statement is accurate, the work records of DBOC would not provide a complete accounting of motorboat activity that could cause seal disturbances. In summary, research conducted in Drakes Estero confirms that this is an important year-round haul-out site and seasonal pupping area, supporting approximately 20% of the mainland California harbor seal
HARBOR SEALS 45 BOX 3 Harbor Seal Disturbance Data There has been much discussion about the veracity of several observations of disturbance that appear in the NPS harbor seal database and have been referred to in various documents and public testimony prepared by NPS. It is beyond the scope of this study to try to determine the authenticity and reliability of these obser- vations, and the committee does not have the investigative authority that would be required to assess any claims of falsification. The Office of the Inspector General for the Department of the Interior investigated charges about the deliberate misuse of scientific information by Point Reyes National Seashore and issued a report in 2008 (Department of the Interior, 2008). Nevertheless, in an effort to help clarify the issues raised in some of the documents provided by Dr. Corey Goodman and others, the committee provides here some general discussion on the nature of these types of observations and their significance for management. The monitoring program at the Point Reyes National Seashore was developed with the following monitoring objectives: â¢ etermine long-term trends in annual population size and annual D and seasonal distribution of pinniped populations at [the Point Reyes National Seashore] and Golden Gate National Recreational Area (GOGA). â¢ etermine long-term trends in reproductive success of elephant seal D and harbor seal populations through annual estimates of productivity at [the Point Reyes National Seashore] and GOGA. â¢ dentify potential threats (i.e., presence of hikers, motor boats, or air- I planes presence), and estimate degree of threat at harbor seal haul outs in order to identify management needs. (Adams et al., 2006) Monitoring is conducted by trained volunteers and park staff who fill out stan- dard survey forms. Each seal colony is surveyed at least twice per week during the breeding and molt seasons if possible. Seal counts are taken during low to medium tides (ideally 2+ feet MLW or less) between 10:00 and 16:00 because studies have shown that the maximum number of seals are hauled out during that part of the day in the San Francisco Bay region (Risebrough et al., 1978; Fancher, 1979; Al- len, 1988; Stewart and Yochem, 1984 [as cited in Hester et al., 2004]; Allen et al., 1989; Grigg et al., 2002). In Drakes Estero, observers conduct the survey from a single location for monitoring the eight haul-out subsites. The surveys include both counts of seals (with adults and pups recorded separately in the pupping season) and notation of disturbance events. In the 2007 Harbor Seal Monitoring report (Truchinski et al., 2008), disturbance surveys are described as follows: Disturbance tallies were based on disturbance sources rather than the number of subsites or seals affected. Disturbance rates were calcu- lated as the number of disturbance events that occurred during the time period from the first observation to the end of the final observation period. Because the disturbance data were not analyzed for effects on continued
46 SHELLFISH MARICULTURE IN DRAKES ESTERO BOX 3â Continued the seal count data in this report, all actual disturbance data were used for analysis regardless of the quality of the associated seal count data (Truchinski et al., 2008). The important point for the current controversy is the last sentence which ex- plains that the disturbance data are treated differently than the population count data. Specifically, the count data were screened to ensure that there was a high likelihood of obtaining reliable, maximum counts at a given site on a given day. Since disturbance events are not dependent on observing the maximum number of seals, the disturbance data are not screened to exclude less experienced ob- servers or suboptimal viewing conditions. For example, disturbances are recorded throughout the survey and not just when the tide level is +2.0 feet (MLW) or lower. Seals will haul out at low to medium tides even when the sand bars are still sub- merged (see Figure 8). Typically, the highest number of seals will haul out when the sand bars are exposed. Limiting counts to these lower tide conditions provides a more consistent index of the abundance of seals in Drakes Estero for monitoring long term fluctuations in abundance. This filter was also important for Becker et al. (2009) in which only the qualifying count data were used in the statistical analysis. The differences in these two types of data (disturbances and counts) are reflected in the standard survey sheets. The seal counts survey form includes information on the low tide level while the disturbance survey form does not. Perhaps the most important confounding factor in any monitoring database, but especially one staffed by volunteers, is the potential for simple recording errors, such as date, time, or tide level. Such an error would generally have little or no effect on the overall trends identified in the database, but would make it difficult to reconstruct the exact events recorded during any individual survey. It is not possible for the committee to resolve the controversy over individual survey sheets, but the focus on these observations highlights how this type of monitoring program is best utilized to indicate potential disturbance problems (that might result in decreased population (Lowry and Caretta, 2003). The remote nature of the estero, combined with an absence of marine predators and other pinnipeds, make this an important habitat for harbor seals. This is reflected in the use of the estero by harbor seals for breedingâDrakes Estero, along with the Double Point colony, consistently accounts for a large fraction of the pups at Marin County haul-out sites (Truchinski et al., 2008). The observations of disturbance recorded in the NPS database cannot be reliably used to infer impacts of mariculture, relative importance of different sources of disturbance, or impacts on seal fitness. The disturbance observations that have been collected as part of the monitoring program serve to demon- strate that there are multiple sources of human and natural disturbances to seals hauled-out on sand bars in Drakes Estero, but they do not per- mit rigorous determination of which sources of disturbance, if any, have greater population-level consequences.
HARBOR SEALS 47 use of a haul-out habitat) rather than to quantify them definitively. The latter would require a data collection system that could be independently verified, such as time and date stamped photographs. This verification is especially important in circum- stances where there is an indication of a source of disturbance that could lead to a regulatory action, as was the case with disturbances attributed to DBOC. FIGURE 8.â Harbor seals in Orkney rest on a submerged haul-out site in the early phase of the tidal cycle. Photograph provided by Paul Thompson. C.â What Effects Can Be Directly Demonstrated by Research Conducted in Drakes Estero Itself? None of the scientific research projects within Drakes Estero was designed specifically to assess whether the oyster farm operations were impacting the local harbor seal population, and this constrains attempts to draw definitive conclusions about potential impacts. Analyses of moni- toring data found a correlation between seal counts and years since the last ENSO event and oyster harvest levels at two haul-out sites within the upper estero (Becker et al., 2009), but this cannot be used to infer cause and effect. Consequently, research that has been conducted within Drakes Estero cannot be used either to directly demonstrate any effects of the oyster farm on harbor seals or to demonstrate the absence of potential effects.
48 SHELLFISH MARICULTURE IN DRAKES ESTERO D.â What Effects Can Reasonably Be Inferred from Research Conducted in Similar Ecosystems? Research on interactions between marine mammals and mariculture in other areas has focused on finfish farming and the economic and ecolog- ical impacts that result from seal depredation (e.g., Nash et al., 2000). There has been no research conducted in similar ecosystems that has directly assessed the impact of shellfish mariculture on harbor seals or indeed any other seal populations. Nevertheless, potential conflicts between marine mammals and shellfish mariculture have been recognized, and Wursig and Gailey (2002) highlight the need to consider potential loss of feeding and breeding habitat from shellfish and finfish farms, particularly given predicted increases in these facilities in in-shore environments. There has, however, been research on responses of harbor seals to dis- turbance from other sources that can inform assessments of disturbance from shellfish mariculture. These studies have focused on impacts upon groups of seals that are already ashore at haul-out sites, with disturbance sources that include people and dogs (Allen et al., 1984; Brasseur and Fedak, 2003), recreational boaters (Lewis and Mathews, 2000; Lelli and Harris, 2001; Johnson and Acevedo-Gutierrez, 2007), commercial shipping (Jansen et al., 2006), industrial activity (Seuront and Prinzivalli, 2005), and aircraft (Perry et al., 2002). Depending upon the intensity and proxim- ity of the disturbance source, a harbor sealâs response can vary from an increase in vigilance, through movement within the haul-out site towards the water, to flushing into the water (Allen et al., 1984). Once groups are flushed into the water, some seals may return to the same or nearby haul-out sites, but counts typically do not return to pre-disturbance levels within the same tidal cycle, particularly if disturbance occurs after low tide (Allen et al., 1984; Suryan and Harvey, 1999). Several studies have explored how the likelihood of a response by seals varies according to the proximity of the disturbance source. This depends at least partly upon the source of disturbance, for example where stationary boats elicit a stron- ger response than boats moving along a predictable route (Johnson and Acevedo-Gutierrez, 2007). The mean distance at which seals are flushed into the water by small boats and people ranges between 80 m and 530 m, with some disturbances recorded at distances of over 1,000 m (Appen- dix D). These empirical studies have been used to underpin zonation of marine protected areas, for example where a 1.5-km buffer exists around harbor seal haul-out sites in the Dutch Wadden Sea to exclude recreational disturbance (Brasseur and Fedak, 2003) and where a 500-m exclusion zone around breeding and molting haul-out sites has been included in the mariculture industryâs best practice guidelines in Shetland, United Kingdom. The 100-yd (91-m) buffer between seal haul-out sites and mari- culture activities, designated in the current Special Use Permit issued to
HARBOR SEALS 49 DBOC, follows the guidelines of the National Marine Fisheries Service (http://www.nmfs.noaa.gov/pr/pdfs/education/viewing_northwest. pdf) for adherence to the Marine Mammal Protection Act. Nevertheless, there are few studies on the impact on seals as a function of distance of approach by various human activities. Allen et al. (1984), in a study of disturbance at Bolinas Lagoon, found that disturbances at a distance of 100 m or less were more likely to cause harbor seals to leave haul-out sites. Nonetheless, seals responded to boats (power and non-power) by leaving the haul-out in 20 out of 43 observations of boats at a distance greater than 100 m. All of these studies assessed the likelihood of disturbing seals that are ashore on their haul-out sites. Because seals often remain in the vicinity of their intertidal haul-out sites at high tide (Allen, 1988), a full assessment of potential disturbance should also consider whether disturbance dur- ing other phases of the tidal cycle affects the sealsâ haul-out behavior. The only study to do this was an experimental study in the Dutch Wadden Sea, in which 13 harbor seals were satellite tagged to assess the impact of recreational boat traffic passing over their haul-out sites at high tide. During the experiment, tagged seals showed a 50% reduction in use of the area compared to use of the same area in years with less recreational boat traffic, and these disturbances also appeared to influence diving behavior (Brasseur and Fedak, 2003). Some oyster rack and oyster bag areas within Drake Estero are located within 500 m of sand flats used by harbor seals as haul-out sites. Based upon the findings in the studies outlined above and the informal obser- vations of biologists who study seals, visits to these areas by oyster farm workers can be expected to lead to the short-term disturbance of any seals using these haul-out areas at the time. Depending upon visibility and wind conditions, disturbance may also occur at greater distances. Furthermore, the work by Brasseur and Reijnders (2001) suggests that seals could be disturbed before they come ashore if boats pass through haul-out areas at high tide. It would be challenging to design a study that could demonstrate whether or not short-term responses to disturbance have long-term population consequences for harbor seals, and no stud- ies of this kind have yet been conducted anywhere. This would require long-term study of known individuals, and high-quality data on those individualsâ exposure both to disturbances and to other potential envi- ronmental stressors. In the absence of additional research, a precautionary approach to management would seek to reduce types of disturbance that affect behavior during the breeding season to avoid potential population effects that would only be evident with long-term monitoring.