Sanctuary Integrated Monitoring Network
Monitoring Project

Effect of the Moss Landing Power Plant thermal discharge plume on the distribution and behavior of sea otters (Enhydra lutris nereis): a preliminary study

Principal Investigator(s)

  • Gena Bentall
    University of California, Santa Cruz


  • SIMoN
Start Date: November 03, 2005
End Date: March 31, 2006

The Monterey Bay National Marine Sanctuary (MBNMS), which stretches from Marin County to Cambria, includes coastal waters and part of Elkhorn Slough, a large estuary adjacent to the town of Moss Landing. A recent upgrade to the cooling system of the Moss Landing Power Plant (MLPP), currently owned and operated by Duke Energy North America, required permits from the MBNMS and other agencies. The MBNMS is interested in determining the effects, if any, of the modified thermal discharge plume on surrounding marine organisms and their associated habitats.

In 2003 researchers at Moss Landing Marine Laboratories (MLML) began a project to provide the first quantitative evaluation of the ecosystem impacts of thermal discharge into Monterey Bay from MLPP. As the MLML study was in progress, bird researchers noted the presence of sea otters (Enhydra lutris nereis) at the thermal plume, raising questions over the affect of the plume on sea otter distribution and/or behavior. Sea otters have high metabolic requirements which cause them to be tightly linked to habitat quality, in particular, the abundance and distribution of their invertebrate prey. As a result, any affect of the thermal plume on the distribution of invertebrates could potentially affect sea otter distribution, prey selection and health, as many of these invertebrate prey types are known bioaccumulators.

The objectives of this study are as follows:

1) Determine if any spatial bias exists with regards to the distribution of sea otters outside of Moss Landing Harbor, such that they disproportionately utilize the area of the thermal plume from the Moss Landing power plant;

2) Document the behavior of sea otters occurring within the plume area (PA) and compare with behavior of otters in near-bye non-plume areas (NPAs) of comparable bathymetry; and

3) Document the diet of any sea otters feeding within the PA and compare this to the diet of otters feeding in NPAs.

Summary to Date

Data collection was completed in January 2006. All visual surveys were made using a spotting scope on the observation deck of the Monterey Bay Aquarium Research Institute (MBARI). Surveys were made in December 2005 and January 2006.

Spatial Bias:

There was a significant effect of location on the frequency of observations of sea otters in the three study sites (Chi-square = 6.89, df = 2, p = 0.03). Elimination of the site with the greatest mean deviation (NP 1) from the analysis indicated that this result was primarily driven by the infrequency of observations in NP 1, the site most distant from the harbor mouth. When a chi-square test was conducted using linear distance from harbor mouth as a co-variate, the deviation of frequencies from expected was no longer significant. After controlling for the effect of site variation, there was a marginally significant effect of time of day (Chi-square = 5.0, df = 2, p = 0.08), with more otters observed at all three sites during morning (AM) hours. Tide was not a significant predictor of activity overall (Chi-square = 2.78, df = 2, p = 0.25, power = 0.30) although two of the sites (PL and NP2) did show a tendency towards greater utilization at high tide.

Behavior results:

Of the behaviors recorded during scan sessions, 95% were active with feeding being the most frequently observed overall. The results of the two way contingency table showed no significant association between site and behavior category (Chi-square = 5.9, df = 4, p = 0.21, power = 0.48).

Individual data:

Sightings of tagged individual sea otters in the vicinity of the PL and NP sites were extremely rare. On two occasions, marked male sea otters were observed swimming and feeding near the north jetty. Both male otters had been tagged and monitored regularly by the Monterey Bay Aquarium Sea Otter Research and Conservation (SORAC) staff but no longitudinal data relevant to their use of the plume area was obtained during this study.

Foraging Behavior:

Twelve foraging bouts (six PL, six NP) were collected for a total of 114 known outcome foraging dives. Foraging bouts for which <10 dives had been recorded were excluded, leaving ten bouts for analysis. Seven different prey types were recorded, with six out of seven occurring in both PL and NP sites. In the nine bouts where the sex of the foraging otter could be identified, the sex was determined to be male (six sub-adults, three adults). No female otter was identified in any of the sites for the duration of the study.
Discriminant analysis resulted in poor performance at correctly classifying foraging bouts into plume and non-plume site categories (Wilks’ lambda = 0.293, F-ratio = 0.301, df = 8, p = 0.894). A jack knifed re-sampling of the data correctly classified bouts as NP 40% of the time and PL only 20% of the time. Success was improved by limiting the analysis to sub-surface time, surface time and mean number of prey items (jack knifed classifications correct 80% of the time). This analysis suggested that animals feeding at the plume had longer dives and post-dive surface times, during which they handled and consumed relatively more prey items than at non-plume areas.

There was a significant relationship between site and diet composition, as measured by the relative frequency of four main prey classes: mussels, non-mussel bivalves (clams), decapods (Cancer crabs) and worms (Chi-square = 8.8, df = 3, p = 0.03). However, the worm prey class (in this case, entirely Urechis sp., or fat innkeeper worms) was only recorded during a single bout, and when worms were excluded from the analysis the difference in diet composition was no longer significant (worm prey class excluded: Chi-square = 3.05, df = 2, p = 0.21). We caution that these results are somewhat equivocal, as the foraging sample size we were able to obtain for this study was small and Chi-square power analysis indicated that the statistical power of this test to detect a significant effect was 22%.

Monitoring Trends

  • Sea otters are using the area within and immediately surrounding the Moss Landing power plant thermal plume in a non-random fashion (Figure 3).
  • Sea otter behavior, and in particular feeding behavior, was generally similar between plume and non-plume areas, although there was some indication that sea otters have slightly longer dives and collect more prey items per dive within the plume area.
  • Plume foragers had a slightly higher frequency of occurrence of prey types associated with hard substrates, such as mussels and Cancer crab species, while non-plume foragers fed more commonly on soft-bottom species, including clams and, in one case, fat innkeeper worms (Figure 4).


The results of this preliminary study provide evidence that sea otters are using the area within and immediately surrounding the Moss Landing power plant thermal plume in a non-random fashion. However, the infrequent observations of sea otters within the non-plume (NP) control site 1, the site most distant from the harbor mouth, was the primary driver of this effect, and we were unable to reject the alternate hypothesis that the observed habitat-use differences simply reflected a decreasing level of utilization as distance from the harbor mouth (and thus distance from the main resting location) increased. In particular, it appears that sea otters tend to emerge from the harbor mouth during feeding bouts, and preferentially feed near the rocky jetties, which likely represent favorable habitat for many of their invertebrate prey. The location of the plume relative to the harbor mouth will render problematic any attempt to distinguish any preference effect specific to the plume from the effect due to proximity to the rocky jetties. Indeed, the closeness of the jetty foraging areas to the plume may increase the probability that foragers discover the plume structure as a feeding patch.

Sea otter behavior, and in particular feeding behavior, was generally similar between plume and non-plume areas, although there was some indication that sea otters have slightly longer dives and collect more prey items per dive (on average) within the plume area. This trend might reflect differences in the types of prey collected in the plume area: while the limited sample size precluded a rigorous testing of diet differences between plume and non-plume sites, visual examination of the raw data suggest that some variation in diet does indeed exist between the site types. Plume foragers had a slightly higher frequency of occurrence of prey types associated with hard substrates, such as mussels and Cancer crab species, while non-plume foragers fed more commonly on soft-bottom species, including clams and, in one case, fat innkeeper worms. The hard substrate provided by the plume structure, as well as the discharge flow likely provides favorable habitat for hard-bottom prey species within a soft-bottom (sandy) habitat. Tagged sea otters feeding within the Moss Landing Harbor and Elkhorn Slough areas have been observed feeding on a high proportion of mussels relative to other prey types (Figure 5).

As individual sea otters are known to specialize in small subsets of the comprehensive population diet it is possible that the proximity of the plume to the harbor allowed otters specializing in the acquisition and handling of mussels and other hard-substrate species to “discover” the somewhat isolated plume-dependent habitat and begin exploiting this resource. Thermal discharge from electric-generating stations in Seal Beach, CA have been shown to attract aggregations of the stingrays (Hoisington and Lowe 2005), with warm water zones hypothesized to be approximating the ray’s disappearing estuarine habitat. In the case of the MLPP thermal plume, the hard substrate of the structure in combination with increased temperature and turbulence may approximate a rocky intertidal habitat favored by some sea otter specialists.

The detection of significant differences in diet composition and behavior between the plume and surrounding areas will likely require a more comprehensive study. The incorporation of seasonality and the fluctuating abundances of both sea otters and prey in the region will potentially have profound effects on foraging observations. As indicated by the power analysis, a larger sample of foraging bouts is necessary to increase the probability of detecting an effect. The opportunistic nature of the collection of foraging data often necessitates data collection over comparatively extended time periods in order to obtain a comprehensive sample.

Long-term changes in the distribution of sea otters in the Moss Landing area will likely affect the outcome of any further study. The complete absence of females across all observations is not surprising considering the characteristically male composition of the dynamic group currently occupying the Moss Landing Harbor. Similar “male groups” occur sporadically within the southern sea otter range, most predictably at range fronts, and are known to vary in size from year to year or even season to season (Jameson 1989). The abundance of sea otters in the harbor has fluctuated dramatically over the last five years (Figure 1), and it is likely that the size and composition of this group will be tightly linked to the number of sea otters observed in the general vicinity of the plume (PL and NP sites included).

If indeed sea otters are feeding preferentially in the thermal plume, and on different prey types, what are the implications? The answer is likely dependent on the quality of the prey being consumed. Mussel species are known bio-accumulators of contaminants and pathogens known to infect sea otters. However, links have yet to be made between prey specializations and susceptibility and virulence of these diseases. If mussels associated with the plume structure were found to be carrying a higher contaminant load than mussels from the nearby jetties, otters feeding in the plume may have a higher risk of exposure. The small spatial scale of the plume region would likely limit such a risk to an individual rather than a population-level issue.

Any effect of the MLPP thermal discharge plume on sea otter behavior, movement and/or health is tightly linked to the effect of the discharge and the discharge structure on the invertebrate species on which the sea otters feed. There has been no recent monitoring of either the benthic invertebrates adjacent to the plume or species colonizing the discharge structure and, as such, little is known of the species composition, or the effect of the thermal plume on invertebrate growth and physiology. Research conducted beyond the scope of this preliminary study should necessarily focus not only on a more comprehensive understanding of the effect of the plume on sea otter foraging behavior but, additionally, on the quality of the invertebrate prey colonizing the plume.

Study Parameters

  • Predation
  • Distribution
  • Behavior
  • Habitat association

Study Methods

Study site:

Observational data on sea otter distribution and behavior in and adjacent to the Moss Landing thermal plume were collected from the balcony of the Monterey Bay Aquarium Research Institute (MBARI) building (Figure 2). The outflow from the thermal plume creates substantial turbulence which can clearly be observed from the MBARI location. Our comparative experimental approach (i.e. contrasting the plume area, or PL, with control non-plume areas, or NPs) necessitated the designation of one PL site and two NP sites, NP 1 being the southernmost site and the farthest from the harbor mouth and NP 2 being the northernmost site, nearest the harbor mouth. Each of these sites was designated as a circular patch of sea surface approximately 65 m in diameter. In order to avoid confounding effects due to differences in bottom habitat structure or bathymetric features, we selected NP sites that were at an equivalent depth to the PL site and at 100 m to the north and south (Figure 2). The three sites ranged between 150 and 250 meters from the shoreline. Designated sites were close enough to each other to ensure similar habitat types yet sufficiently distant to ensure that otters feeding within a NP site could not possibly be diving to capture prey from the plume outflow. The presence of the large, dynamic group of sea otters inside the Moss Landing Harbor could potentially create biases associated with the proximity of the harbor mouth to the group’s resting area and the tendency of some otters to feed on prey inhabiting the artificial rocky substrate of the jetty. By choosing one NP site to the north and one NP site to the south, we controlled for these potentially confounding effects. Assessments of distances between sites, bathymetry and proximity to the harbor mouth were confirmed and adjusted using ArcView 3.2.

Testing for Spatial Bias in Sea Otter Distribution

Data collection sessions were conducted by two field personnel between November 3 and January, 2005 for a total of 22 scan days. Each session lasted approximately two hours, and the time of day for each session was randomized to avoid any bias associated with diurnal patterns, and all daylight hours were represented approximately evenly in the final data set. Observations were conducted from shore using high-powered Questar telescopes (50-80x magnification; Questar Corp., Isanti, MN). The observers first identified the PL site (based on visual assessment of surface turbulence), and then used compass bearings in conjunction with laser range finders to identify the two NP sites, based on distance and direction from the observation point. For the first hour of each session, instantaneous scan sampling of sea otter distribution and abundance was conducted (Altmann, 1974). Scans were conducted as follows: one minute scan of NP 1, followed by a one minute scan of PL, followed by a one minute scan of NP 2. In each one minute scan, all sea otters present within the sites were recorded. This sequence was repeated 15x over a one hour period. After the 1 hour scan session, a second hour was spent collecting foraging and telemetry-based data (see below).
For each independent sea otter observed during the shore-based scans, the following information was documented:

• Activity/ Behavior

• Time of Day

• Age-class, sex, ID tag (if present) and reproductive status (pup/no pup), whenever possible

Behavior was categorized hierarchically: first as active or inactive, then further designated using an ethogram of 9 descriptive behavior categories (Gelatt et al. 2002). Prior to analysis, behaviors were condensed to 4 broad categories in order to achieve suitable sample sizes: feeding, swimming, resting (inactive), or other active (all other active behaviors).

Foraging Behavior:

Foraging data were collected by methods identical to those currently in use in all California sea otter studies (Estes et al. 2003b, Tinker 2004, Tinker et al. 2004). After a scan session was completed, a single foraging otter was selected arbitrarily from within the PL site for collection of detailed foraging/diet data (if there was at least one foraging otter available). A second foraging otter was selected arbitrarily from within one of the two NP sites, again conditioned upon at least one subject being available. For both of the two selected focal otters, we used the Questar telescope to collect the following data:

• Start and finish time of data collection

• Precise location (using GPS)

• Sub-surface and surface time intervals for each dive

• Prey identification (to lowest possible taxon) for each dive

• Number of prey items captured per dive

• Prey size estimate (maximum linear dimension), per prey item, per dive

These foraging parameters were recorded for each focal otter for a 30 minute time period (designated as a feeding bout). In cases where a 30 minute bout was not possible, we attempted to record a minimum of 10 consecutive feeding dives.

Supplementary data: Individual trends

The data collection protocols described above provided cross-sectional data on sea otter habitat use, behavior and diet in the PL and NP areas. However, because there were known to be tagged otters in the vicinity of Moss Landing, we took the opportunity to collect longitudinal data from these marked individuals, with the view that such longitudinal data collected over the entire study period might provide further insight into individual patterns of habitat use related to the thermal plume. Accordingly, after the scans and foraging data collection were completed each day, we visually searched the entire vicinity of the plume area for marked individuals: telemetry equipment was also used to detect otters implanted with VHF radio transmitters (Tinker et al. 2004). Additionally, any tagged otters that were observed during scan or foraging sessions were noted. For any tagged otters that were detected, we recorded precise location (using compass, range finder and GPS) and behavior, following established protocols.

Figures and Images

Figure 1. Gena Bentall, a researcher at UC Santa Cruz, used the observation deck at the Monterey Bay Aquarium Research Institute (MBARI) to view otters swimming in and adjacent to the thermal plume generated by the Moss Landing Power Plant (MLPP).

Figure 2. Bentall made observations at the thermal plume and two reference sites within 1 km of the plume.

Figure 3. Data were recorded during observation intervals of 5 min each.

Figure 4. Sea otters use the Elkhorn Slough area to rest and forage.

Figure 5. On calm days, the plume is visible at the surface as a circle (here in the center).

Figure 6. Other animals, such as birds, also appear to use the plume. It is not clear what attracts animals to the plume.

Figure 7. Bentall made observations over a two-month period and the results are expected in the spring of 2006.

Figure 8. Sea otters are now very common near the harbor entrance to Moss Landing.

Figure 1. High counts of independent sea otters within Moss Landing Harbor and Elkhorn Slough from 2001-2005. SP = spring survey, FA = fall survey. Courtesy Brian Hatfield, USGS California biannual sea otter census.

Figure 2. Map of the region of study indicating the location of the three plume-related sites relative to the observer site and the Moss Landing Harbor.

Figure 3. Frequency of observations of sea otters in each of the three study sites. Counts are based on totals for each of the 22 independent scan sessions. Bars represent standard error.

Figure 4. The diet composition of sea otters foraging in the plume and non-plume (NP 1 and NP 2 collectively) sites.

Figure 5. The diet composition of tagged sea otters foraging in Elkhorn Slough and Moss Landing Harbor from 1997-2006. Source: Monterey Bay Aquarium (SORAC)/ US Geological Survey.


  • Bentall and Tinker (2006)
    The effect of the Moss Landing Power Plant thermal discharge plume on sea otter behavior and distribution.
    231 KB PDF