MARINE MAMMAL SCIENCE, 14(4):895-902 (October 1998) 0 1998 by the Society for Marine Mammalogy

SEA OTTERS AND BENTHIC PREY COMMUNITIES: A DIRECT TEST OF THE SEA OTTER AS KEYSTONE

PREDATOR IN WASHINGTON STATE

One of the most widely cited examples of the keystone predator hypothesis in community ecology is the sea otter/urchin/algal trophic cascade paradigm, wherein algal communities are thought to flourish in areas occupied by sea otters (Enhydra ftctrzs) because of reduced grazing pressure due to otter pre- dation on dominant consumers. This paradigm has emerged primarily from the published results of studies done in Alaskan rocky habitats (Estes and Duggins 1995) but also in southern California (Ebling and Laur 1988). Our study assesses the geographic generality of the trophic cascade from otters to urchins to algae, by examining community changes associated with the ex- pansion of the more southerly and isolated sea otter population along the outer Olympic Coast of Washington State. The founding members of that popula- tion were introduced from Alaska during 1969-1970 (Jameson et a/. 1982). Since that time, particularly over the last decade, the Washington State sea otter population has expanded dramatically, from the initial translocation of 59 individuals (Jameson et al. 1982) to 100 animals in 1987 (Bowlby et al. 1988) and >300 animals in 1995 (Jameson, National Biological Survey).

Our initial survey of the benthic communities and sea otter habitats along the Olympic coast in 1987 (Kvitek et af . 1989) was a typical space-for-time test of the sea otter/urchin/algal paradigm. The results of that study revealed low invertebrate prey numbers and biomass within the sea otter primary range (mean otters sighted/region/aerial survey >3, after Bowlby et al. 1988), some- what higher numbers within the secondary range (mean otters sighted/region/ aerial survey <3, after Bowlby et af. 1988), and much higher levels of prey resources outside the range, especially to the north (Fig. 1, 2a). Abundance of foliose red algal species eaten by sea urchins (Strongylocentrotus franciscanm) was positively correlated with sea otter abundance, consistent with the paradigm (Fig. 3a; Kvitek et al. 1989). Based on those results, we predicted that if the Washington State sea otter population continued to grow, it would most likely expand to the north, drawn by and depleting the rich prey resources found there, and that if sea otters did move into these northern habitats, there would be significant changes in the benthic algal cover if the otters did reduce the abundance of sea urchins and other invertebrate grazers as expected. The 1987 study established the baseline necessary for a direct test of these sea otter community effects at some point in the future if and when the Washington State otter population expanded its range.

Here we report the results from our 1995 resurvey of the 1987 study sites, as well as two additional sites at the mouth of the Straits of Juan de Fuca (Tatoosh Island and Cape Flattery). Sites previously unoccupied or sparsely populated by sea otters in 1987 (Bowlby et al. 1988, Kvitek et al. 1989) are

895

896 MARINE MAMMAL SCIENCE, VOL. 14, NO. 4, 1998

Figure I. Benthic sampling sites and sea otter ranges for 1987 and 1995 along the Olympic coast of Washington state. Methods and data for determining primary otter range ( > 3 otters observed per survey) and secondary otter range (<1 otter ob- served per survey) from Bowlby e t al. 1988, and Jameson National Biological Survey unpublished data.

NOTES 897

Number of sea otters Prey Prey observed per survey abundance m-2, SE biomass g m2+ SE

I? a, 5 m I - 0

0 1 a .- F m 0

i3 b) 1995

0 c a .-

F m 0

i3

F Makah Bay

2 Anderson Point 2

Pt. of the Arches g Cape Alava Y

Cape Johnson $ Rock 305

Teawhit Head

c)

3

P,

Neah Bay

Tatoosh I. (HC)

Cape Flattery (HC)

P Makah Bay

Anderson Point 2 Pt. of the Arches g

Cape Alava 3 Cape Johnson $

Rock 305

Teawhit Head

0

P,

Figure 2. Invertebrate prey abundance and total prey biomass at benthic sampling sites for 1987 (a) and 1995 (b), related to sea otter abundance. Prey abundance shown by prey type (“Urchins” = Strongylocentrotus franciscanas, “Cucumbers” = Cucumaria spp. and Parastichopas californica, “Other” = molluscs, decapods). HC = high-current areas at Tatoosh Island and Cape Flattery. These two sites sampled in 1995 only.

898 MARINE MAMMAL SCIENCE. VOL. 14. NO. 4. 1998

a) 80% 4 1987 Foliose red algae I

$ 6oyoI 1995 Foliose red algae I 4-

a

0% Neah Bay

100%

80% w cn $ 60%

2 E 40% e a

l? 20%

0% Neah Bay

N = 5 0 16

I

Anderson Pt. I

Cape Alava -

1987 Coralline algae

1 1995 Coralline algae

Anderson Pt. 5 0 3 0

Site

< 1

Cape Alava 5 0 3 5

Outside Northern Primary

range 1987 range range

Figure 3. Changes in percent algal cover following sea otter expansion. Foliose red algae increased (a) and coralline algae decreased (b) at Neah Bay and Anderson Pt. where otter numbers have increased (c) and sea urchins decreased between 1987 and 1995. N = number of percent cover quadrats sampled at each site.

now all occupied by much higher numbers of sea otters (Fig. 1, 2b). Current sea otter distribution and abundance data were obtained from the National Biological Service Washington State census (Jameson 1995, unpublished data). Video quadrat techniques were used to resample eight shallow subtidal survey

NOTES 899

sites originally established along the Olympic coast during August of 1987 (Fig. 1) (Kvitek et d. 1989). In addition to the 1987 sites, we also surveyed sites at Tatoosh Island and Cape Flattery (Fig. 1). The original methods em- ployed in 1987 consisted of divers counting and measuring invertebrates and percent algal cover within 1-m2 quadrats deployed along 10-m transect lines. This approach required three weeks of intense field work by a team of six divers. The 1995 resurvey was limited to one week of field time, with only three divers. Given these constraints, video data capture was selected as the most efficient method for a quantitative benthic inventory. This method uti- lized a Sony Hi8 video camera in a light-equipped Light and Motion Stingray housing fitted with two forward-facing PVC pods. These pods were of appro- priate length and position such that when the diver/operator positioned the camera to point vertically down at the sea floor while resting on the pod ends, the video field of view was 0.25 m2 and the pods were outside the field of view. The operator was thus able to move across the seafloor with the camera running continuously, pausing momentarily to push the pods against the sub- strate, capturing 0.25-m2 video images for later analysis. To aid in the later measuring of individuals on the video display, the operator placed a 20-cm scale bar in the field of view of the beginning and ending frames of each quadrat series.

The divers captured the video quadrats along continuous transects, in an effort to duplicate as closely as possible the techniques employed during the 1987 survey. The limits of video resolution required that a smaller quadrat size be used than the 1-m2 in situ quadrats used in 1987. However, a test comparison at a nearby survey site in 1995 (Chibahdehl Rocks) of invertebrate size and abundance data collected using both methods showed no significant difference in results (t-tests, P = 0.32 and 0.24 for abundance and size, re- spectively). Power was not optimal but reasonably high for both of these comparison t-tests as well (power = 0.66 for abundance t-test, 0.67 for size t-test). Chibahdehl Rocks was chosen for this comparison because prey, al- though rare, were sufficiently abundant there to allow evaluation of both size and abundance efficiency for each method (although rarity and high variability of prey probably lowered the power of our method comparisons). For these reasons, we relied entirely on video data for the remainder of the study. Our confidence in the results from this method increased with our qualitative ob- servations of the obvious and dramatic decrease in sea otter invertebrate prey abundance at all sites since 1985.

Video tapes were viewed in the laboratory, where benthic invertebrate spe- cies were identified, counted, and measured within each 0.25-m2 video quadrat for all sites. Invertebrate prey biomass was determined based on size/weight regression formulae derived during the 1987 survey project (Kvitek et a/. 1989). Only conspicuous, exposed otter prey items were counted, as divers did not overturn rocks or otherwise disturb the substrate when quadrats were recorded in the field. For this reason, extremely cryptic or hidden otter prey such as octopus and crabs may have been overlooked by our methods [this may account for the apparent paradox of increased otter numbers yet low prey

900 MARINE MAMMAL SCIENCE, VOL. 14, NO. 4 , 1998

at Cape Johnson, Rock 305, and Teawhit Head in 1995 (Fig. 2b)}. Stacking or layering of prey did not occur in any quadrats recorded. Percent cover of substrate type was determined from 0.25-m2 video quadrat frames by super- imposing a random pattern of 50 dots on the image; layering was not ac- counted for, as only the uppermost visible substrate type was recorded.

In the 1987 study prey abundance and biomass had been found to be inversely proportional to sea otter abundance (Fig. 2a). By 1995, prey abun- dance and biomass remained low at the 1987 sea otter sites and had declined substantially at previously sampled sites recently invaded by sea otters (Fig. 2b). Prey abundance and biomass was significantly lower in 1995 than in 1987 for all resampled sites (block-design, repeated measures ANOVA , P = 0.0001). This pattern of reduced prey in areas occupied by sea otters was most evident for sea urchins, which had made up the majority of the prey resource at the northern sites in 1987 (Fig. 2a). Sea urchins had become rare or absent at all sites within the otter range, with one very notable exception, the highly exposed and current-swept mouth of the Straits of Juan de Fuca. Here, Tatoosh Island and Cape Flattery both supported urchin numbers and biomass much higher than at any other location (Fig. 2b). In fact, the values for urchins found at Tatoosh Island in 1995 were higher than for any site sampled in 1987, inside or outside of the sea otter range at the time.

Foliose red and coralline algal cover were measured at four sites in 1987. These sites were chosen to represent each level of otter occupation (or lack thereof) at that time. Only three sites yielded usable percent-cover data in 1995: Neah Bay, Anderson Point, and Cape Alava (Fig. 3). Foliose red algal cover was significantly higher at Neah Bay and Anderson Pt. but lower at Cape Alava in 1995 than in 1987 (Mann-Whitney U tests, P = 0.02). Com- paring 1995 with 1987, coralline cover was significantly lower at Neah Bay (P = O.OOOl), unchanged at Anderson Pt. (P = 0.8), and higher at Cape Alava (P = 0.0004).

As expected, while the otters had increased in numbers within their range since 1987, this range had expanded to the north only (Fig. 1, Jameson, National Biological Service). Also as expected, prey abundance and biomass had declined by an order of magnitude to very low levels at the newly otter- occupied sites on either side of Cape Flattery (Fig. 2). Our results show, how- ever, high prey numbers and biomass at Cape Flattery and Tatoosh Island just off the end of the Cape. This seemingly anomalous persistence of high prey abundance within the sea otter range may be due to the expanding otter population in Washington having skipped over this area (sea otters have rarely been observed at Tatoosh, R. T. Paine, personal communication). Otter avoid- ance of this area may be because of the typically heavy seas and strong tidal currents found here. These sites are well known as areas exposed almost con- tinuously to large waves and strong tidal currents considered a hazard to navigation (Department of Fisheries and Oceans, 1997; U.S. Department of Commerce, 1992). We have noted a similar pattern in our work at Kodiak (Kvitek e t al. 1992), where the only location within the sea otter range that had high prey biomass was Afognak Strait, a site renowned for tidal currents

NOTES 90 1

high enough to be of concern to vessel traffic, and as with the Cape Flattery area, a site at which we could dive only during a narrow slack-tide window. If such patterns (and their explanations) hold up, they will provide a predict- able exception to the generality of the otter trophic cascade paradigm. Fur- thermore, if true, these exceptions may provide rehgia for high-density ag- gregations of large size urchins within the sea otter range. Given the strong positive relationships between urchin density, fertilization success (Mead and Denny 1995) and recruitment (Tegner and Dayton 1977, Duggins 1983), high-current areas may prove to be important local urchin nursery grounds within otter-occupied regions.

Finally, removal of urchin grazers by sea otters was most likely responsible for the rise in foliose red algal cover at the recently occupied Neah Bay and Anderson Pt. sites (Fig. 3 ) . The greatest change in algal cover occurred at the site which experienced the greatest decline in urchin abundance, Neah Bay. Foliose red cover increased greatly, while coralline crust dropped from virtually 100% to 42% as a result of being covered by the foliose plants.

Our results are consistent with, and extend the geographic generality of, Estes and Duggins's (1999 conclusion that where sea otters are established, urchin herbivory will typically not be a dominant force structuring nearshore benthic communities.However, the hypothesis that high-current areas may provide a predictable exception to this rule will require further testing.

ACKNOWLEDGMENTS

We thank K. Kiest, T. Jacobs, and G. Galasso for their help in the field. Reviews by J. Estes and G. Green added substantially to the quality of the manuscript. Primary funding for the project came from the Olympic Coast National Marine Sanctuary, NOAA. Additional support was provided by the National Science Foundation OPP- 932 1504, California State University Monterey Bay, and Moss Landing Marine Labs.

LITERATURE CITED

BOWLBY, C. E., B. L. TROUTMAN AND S. J. JEFFRIES. 1988. Sea otters in Washington: distribution, abundance, and activity patterns. Washington State Department of Wildlife, Olympia, WA. Final Report to National Coastal Resources Research and Development Institute, Newport, Oregon. Present address: USGS Biological Re- sources Div., EPA, 200 SW 35th St., Corvallis, OR 97333.

DEPARTMENT OF FISHERIES AND OCEANS. 1997. Canadian tide and current tables 1997. Volume 5. Juan de Fuca Strait and Strait of Georgia. Department of Fisheries and Oceans, Ottawa, O N (Canada), Hydrographic Service.

DUGGINS, D. 0. 1983. Starfish predation and the creation of mosaic patterns in a kelp-dominated community. Ecology 64: 161 0-1 619.

EBLING, A. W., AND D. R. LAUR. 1988. Fish populations in kelp forests without sea otters: Effects of severe storm damage and destructive sea urchin grazing. Pages 169-191 in G. R. VanBlaricom and J. A. Estes, eds. The community ecology of sea otters. Springer-Verlag, Berlin.

ESTES, J. A., AND D. 0. DUGGINS. 1995. Sea otters and kelp forests in Alaska: Gen- erality and variation in a community ecological paradigm. Ecological Monographs 65:75-100.

902 MARINE MAMMAL SCIENCE, VOL. 14, NO. 4, 1998

JAMESON, R. J., K. W. KENYON, A. M. JOHNSON AND H. M. WIGHT. 1982. History and status of translocated sea otter populations in North America. Wildlife Society Bulletin 1O:lOO-107.

KVITEK, R. G., D. SHULL, D. CANESTRO, E. BOWLBY AND B. TROUTMAN. 1989. Sea otters and benthic prey communities in Washington State. Marine Mammal Sci- ence 5:266-280.

KVITEK, R. G., J. S. OLIVER, A. R. DEGANGE AND B. S. ANDERSON. 1992. Changes in soft-bottom prey communities along a gradient in sea otter predation. Ecology 7 3:4 1 3 4 2 8 .

MEAD, K. S., AND D. W. DENNY. 1995. The effects of hydrodynamic shear stress on fertilization and early development of the purple sea urchin Strongylocentrotus par- puratus. Biological Bulletin 188:46-56.

US. DEPARTMENT OF COMMERCE. 1992. United States coast pilot 7: Pacific Coast., National Oceanic and Atmospheric Administration, National Ocean Service. USA.

TEGNER, M. J., AND P. K. DAYTON. 1977. Sea urchin recruitment patterns and im- plications of commercial fishing. Science 196:324-326.

RIKK G. KVITEK, California State University, Monterey, California, U.S.A.; e- mail: rikkkvitek@monterey,edu; PAT J. IAMPIETRO, Moss Landing Marine Lab- oratories, Moss Landing, California, U.S.A.; C. EDWARD BOWLBY, Olympic Coast National Marine Sanctuary, Port Angeles, Washington, U.S.A. Received 21 July 1997. Accepted 4 November 1997.