Abstract The Hewatt transect is a permanently marked area perpendicular to shore running through the protected rocky intertidal zone at Hopkins Marine Station. Species abundance data collected in 1933-37 and 1993-95 showed significant differences for many species during the intervening years. In order to determine how representative the Hewatt transect is of the protected rocky intertidal zone at Hopkins, a 90 x 15 m area surrounding the Hewatt transect was compared to three adjacent areas of the same size. Three organisms were sampled in two or three 15-m transects within each of the four areas. Ten quadrats were counted in each transect. Additionally, the substrate rugosity for each area was measured. For Anthopleura sola, a solitary anemone, there were significant differences among areas (P=.02), but not among transects-within-areas (P=. 16). For Serpulorbis squamigerus, a tube dwelling gastropod, there were no significant differences at either level (P=. 15 and 15). For the alga Endocladia muricata, there were no significant differences among areas (P=.23), but highly significant differences among transects-within-areas (P.001) Substrate rugosity showed no significance between areas (P=. 72). These data indicate that while there are some differences between the Hewatt area and adjacent areas, there is no evidence that the Hewatt area is non-representative of the HMS intertidal. While the intertidal zone is certainly not homogeneous, it is possible to conclude that the changes seen in the 60 year interval were due to significant species abundance change and not stochastic variation. Additionally, the environmental processes that caused these changes are likely to have influenced the distribution of organisms across the intertidal zone. Introduction In the 1930's, Hewatt (1933) conducted an ecological survey of a permanent transect he established in the protected rocky intertidal at Hopkins Marine Station. This study provided extensive information on species distribution and abundance. Sagarin and Gilman (1993) relocated the transect and resampled 19 of Hewatt's original quadrats. Their data showed significant changes for many species. Notably, many species which increased in abundance had geographic distributions with northern boundaries at Cape Mendocino ("southern" species), and the species that decreased in abundance had geographic distributions with southern boundaries at Point Conception ("northern' species). Concurrent increases in sea surface temperatures implicated global climate change as a possible cause of these changes (Barry et al. 1995). This result has important implications for the potential environmental impacts of global warming in intertidal communities. While many anecdotal observations of changes in species abundance have been made, quantitative studies such as this are necessary to provide robust evidence of environmental change. Unfortunately, with limited amounts of reference data such as Hewatt’s, these studies are rare. Furthermore, they are confounded by natural variation. In order to make inferences concerning potentially crucial changes occurring in large areas, it is necessary to understand the spatial and temporal scales at which nature varies. The rocky intertidal is a highly variable environment. A great deal of work conducted in Australia (Underwood and Chapman1996, 1998; Underwood 1998) has focused on this problem, because anthropogenic stress on the southern coastlines of Australia is increasing and may be affecting intertidal communities. The purpose of this research was to better understand spatial patterns prior to attributing causation for temporal changes. Natural spatial variation makes anthropogenic impact difficult to quantify. A number of statistical techniques have been developed to quantify spatial patterns so that changes can be detected (Underwood 1998). A primary conclusion of these studies is natural variation is high and spatial patterns are difficult to quantify. The potential anthropogenic impacts upon the rocky coast of California are also huge, and spatial variation is equally high. Fu (1995) studied organism abundances at various spatial scales and found that different species were patchy at different scales. Considering such high and inconsistent levels of variation, it is appropriate to question how representative the Hewatt transect may be of the intertidal zone at Hopkins Marine Station prior to making inferences about environmental change. This study was designed to address this question by sampling two intertidal invertebrates and one intertidal alga. The animals I chose are both southern species (Barry et al. 1995) that have significantly increased in abundance along the Hewatt transect. This makes them particularly interesting in the context of climate change hypotheses. In order to conclude that southern species are able to "invade" the intertidal zone because of rising sea surface temperatures, it is necessary to show that increased abundances noted along the Hewatt transect are representative of areas elsewhere in the intertidal zone. Two of the study organisms are sessile invertebrates. The anemone Anthopleura sola is abundant and widespread in the mid-intertidal zone. A sessile predator during high tide, it retracts its tentacles and its shell-covered body wall becomes visible when tides are low. Serpulorbis squamigerus is a low-intertidal tube building snail. It is a suspension feeder, spinning webs of mucus to catch particles as they drift by (Morris et al. 1980). In previous years, it was present as a solitary, primarily subtidal species, but recently it has become extremely abundant in the intertidal zone, forming mats of coiled tubes. The third species I sampled was the alga Endocladia. While Hewatt (1933) did not survey algae, some understanding of the variation in algal cover is useful because algae is often the predominate occupant of space in the intertidal, and plays an important role in determining the local animal community. Simons and Leydig (1996) found a 1 ft. downward shift in the Endocladia zone compared to data from the early 1960's (Glynn 1965), as well as associated changes in its assemblage of organisms. This suggests Endocladia may be an indicator species for intertidal change. Materials and Methods: This study was conducted in the sheltered rocky intertidal area west of Bird Rock at Hopkins Marine Station in Pacific Grove, CA, USA (36° 37' N, 121° 54' W). This study site was determined by the location of the Hewatt transect. The transect runs 100 yards perpendicular to shore and is marked by brass bolts. I randomly sampled within an approximately 90 x 15 meter area that contained the Hewatt transect and compared it to three adjacent areas of comparable size and tidal height. The four areas were designated West, Hewatt, East1, and East2 (Fig. 1, Fig. 2). Areas closer to Bird Rock were excluded as they incorporated primarily high intertidal areas. The four study areas are crossed by two permanent channels that run parallel to shore. The zone is also protected along its seaward margin by high rocks, preventing high wave exposure in the interior. Study design and Sampling A preliminary study was conducted to determine the optimal quadrat size for A. sola and S. squamigerus. Both animals were counted in haphazardly placed quadrats of 25-m and 1-m sizes. Precision, a measure of repeatability, was calculated by the following equation: P-SE/X where SE is the Standard Error of the sample and X is the sample mean. For S. squamigerus, 25-m quadrats gave the greatest precision, while for A. sola 1-m’ quadrats were more precise. The alga Endocladia was surveyed in 25-m" quadrats without any comparison to other quadrat sizes. This study used a nested Analysis of Variance (ANOVA) design (Fig. 3). For each species, two or three 15-m transects were sampled in each of the four areas. Ten quadrats were randomly located and counted per transect. For Endocladia and Anthopleura, three transects per area were haphazardly placed through the mid-intertidal zone (approximately 2 - 4 feet above MLLW). Whenever possible, these transects were kept roughly parallel to the Hewatt transect. For Serpulorbis, two transects were haphazardly placed within the low zone (approximately -1 to +1.5 feet above MLLW). There was insufficient space for three non-overlapping low zone transects within each area. Animals were counted if they fell within the randomly located quadrat. Tube snails were counted as number of apertures within each quadrat rather than tubes. Individuals along the edges of a quadrat were counted if their aperture fell inside the quadrat. For A. sola, a relatively large species, animals were counted as halves on the rare occasion that they fell exactly on the edge of the quadrat. For Endocladia, percent cover was estimated using a 25m' quadrat subdivided into 25 5 x 5 cm squares. For each quadrat sampled, the estimated average tidal height was measured using a surveyor's transit and stadia rod. Measurements were cross-referenced with a brass bolt that marks the beginning of the Hewatt transect at 3.82 feet above MLLW. Physical Measurements In addition to ecological sampling, physical topography measurements were taken in each area. A straight line distance of 1.88 meters was measured with plumb lines hung from a leveled PVC pipe. The surface distance between those points was determined using fine link chain that conformed to irregularities and cracks in the rock. This length, divided by 1.88, was used as an index of the substrate rugosity. Five replicate lengths were haphazardly placed per area at random compass bearings. Data Analysis Cochran's tests for heterogeneity of variances were performed on the within- transect variances for Serpulorbis, Anthopleura, and Endocladia. For each organism, the test was significant (PS.05), indicating their variances were too heterogeneous to meet the assumptions of ANOVA. To decrease the heterogeneity of variances, Serpulorbis data were log transformed (log (x+1)). Anthopleura data were square root transformed (V(x+.5)). Endocladia data were arcsin transformed (sin (x)). After transformation, variances were no longer heterogeneous. The data were analyzed using a two-factor nested ANOVA. The tests were conducted at q= 10 rather than the more conventional .05 for the following reasons. Two types of statistical error are possible: Type I occurs when a true null hypothesis (Ho) is rejected; its probability is a. Type II, with probability B, is the opposite mistake of retaining an incorrect null hypothesis. Type I errors are more serious in biological experiments, because the acceptance of an incorrect alternative hypothesis will probably not lead to further experiments. By contrast, Type II, or the acceptance of an incorrect null hypothesis, probably will lead to further experiments and the error will be detected. In this study, a larger a was considered appropriate due to the potential result of Type I and Type II errors. With Ho of "no differences among areas," Type I error would conclude the Hewatt transect is not representative of the intertidal at Hopkins Marine Station when it actually is, and no inferences about the generality of environmental processes would be made. Conversely, Type II would suggest the Hewatt transect is representative of the intertidal when it is not, which in this case is an equally serious error. The probability of Type II error, ß, cannot be specified without a specific (quantitative) alternative hypothesis (Hi), but it is clear from Fig. 4 that their probabilities are complementary: an increase in a would decrease ß, and vice versa. For this reason, a more conservative approach that would balance the likelihood of either error was to set a=. 10 rather than the conventional .05. Differences among areas for each species were compared using Tukey's post-hoc tests. Rugosity data were analyzed using a single factor ANOVA. Results Anthopleura densities ranged from 0-23 per m’ (Fig. 5). There were no significant differences between transects-within-areas (P= 16, Table 1), but there were significant differences between areas (P=.02, Table 1). Post-hoc comparisons showed that the Hewatt area did not differ from either of the other three, the only significant differences were between West and East1 and West and East2 (Table 2). Furthermore, 77% of the summed variance components was associated with residual, or among-quadrat, variation. while only 4% was associated with transects-within-areas, and 19% was associated with areas (Fig. 9). Percent cover of Endocladia ranged from 0-93% (Fig. 6). There were highly significant differences among transects-within-areas, especially in West and Hewatt (Fig. 6, Table 1). However, there were no significant differences between areas (P- 23, Table 1). Post-hoc comparisons among areas revealed no significant differences between any of the areas (Table 2). Fisty percent of the summed variance components was associated with among-quadrat variation, 39% was associated with transects-within-areas, and 11% was associated with areas (Fig. 9). Serpulorbis densities ranged from 0-856 individuals per m’ (Fig. 7). There were no significant differences among areas (P= 15, Table 1), or among transects-within-areas (P=.15, Table 1, 2). Sixty-eight percent of the summed variance components was associated with among-quadrat variation, 1 1% was associated with transects-within-areas, and 23% was associated with areas (Fig. 9). For rugosity indices, conformed distances ranged from 1.96 to 3.60 meters, with corresponding ratios of 1.04 and 1.91. Means per area ranged from 1.38 through 1.56 (Fig. 10). There were no significant differences between areas (P=.72, Table 1) Discussion The study attempted to answer the question of how representative, not how identical, the Hewatt transect is of the rocky intertidal at HMS. Considering the high spatial variation in this environment, some differences will always be present. However, in order to be representative, comparisons of the Hewatt transect and other areas should show no more differences than the other areas do when compared among themselves. A consistent pattern of differences between the Hewatt transect and adjacent areas would indicate the transect is not representative at all. Post-hoc Tukey's tests were used because they have the ability to analyze all possible pair-wise combinations of areas. These tests showed no significant differences between any areas for Serpulorbis and Endocladia. For Anthopleura, two out of six possible comparisons were significant and neither of them involved Hewatt. The significant difference (P=.020) noted in the overall ANOVA occurred because West was significantly different from both East l and East2. The significant differences between West and East l/East2 for Anthopleura are most likely related to microhabitat variation. West is located just prior to the seaward rim of the intertidal, and probably receives considerably higher wave exposure. Anecdotally, it is also host to another species of anemone, Anthopleura xanthogrammica. A. xanthogrammica is typical of wave exposed areas (Ricketts et al. 1985). The two East areas, by contrast, receive far less wave exposure than West. The substratum in these areas tends to have many more shallow crevices, where A. sola is most abundant. It is interesting to note that the statistical analyses for differences in substrate rugosity did not differ among areas (Fig. 10, Table 1) The low abundance of Endocladia found in Hewatt (Fig. 6) is due to sampling artifact: nearly all quadrats fell on inappropriate tidal heights. Fig. 7 shows a skewed distribution between quadrat count and average tidal height for each of the four areas. Most of the zero values occur in Hewatt below 2.5' above MLLW. Endocladia is abundant between fairly specific upper and lower boundaries, and the lower boundary is approximately 2.5' above MLLW (Simons and Leydig 1996). If quadrats with zero percent cover extended along the x-axis of Fig. 7, I would conclude that the dearth of Endocladia was unique to Hewatt. However, as the zero values are only found below the lower boundary for the alga, and since many Hewatt quadrats are found at or below this height, it is more likely that portions of Hewatt are simply too low for Endocladia. While these tests show that the Hewatt transect is representative of the intertidal area at Hopkins Marine Station, they raise the general question of spatial heterogeneity in the intertidal. Variation occurs among areas and among transects-within-areas, but based on the components of total variation, the greatest variation occurs between quadrats. This is due to differences in microhabitats that average out at larger spatial scales. While there is a great deal of small scale variability in the intertidal zone at Hopkins Marine Station, there is no consistent pattern suggesting the Hewatt transect is much different from surrounding areas. Considering these results, it is possible to extrapolate that the environmental processes responsible for intertidal change since 1933 10 have acted on the whole intertidal at Hopkins Marine Station. While these processes are far from cle iportant to note that temporal variation across sixty years was far r, it is more significant than spatial variation in this study. This provides compelling evidence that although stochastic variation is always a problem, carefully studied transects can act as indicators for large-scale change. Literature Cited Barry, J.P., Baxter, C.H., Sagarin, R.D., Gilman, S.E.(1995). Climate related, long-term faunal changes in a California rocky intertidal community. Science 267:672-675 Hewatt, H.G., (1934). Ecological studies on selected marine intertidal communities of Monterey Bay. Doctoral thesis, Biological Sciences, Stanford University Fu, I. (1995) Optimal sampling design and spatial scales of variation in the abundance of three intertidal invertebrates. Unpublished student paper, Hopkins Marine Station, Stanford University. Glynn, P. (1965). Ecological studies on the Endocladia muricata - Balanus glandula association in the intertidal zone in Monterey Bay, California. Stanford, California. Morris, R, Abbott, D., Haderlie, E., (1980). Intertidal invertebrates of California. Stanford, California, Stanford University Press. Ricketts, E.F., Calvin, J.C., Phillips, D.W. (1985). Between Pacific Tides 5“" ed. Stanford University Press, Stanford, California, USA. Sagarin and Gilman (1993). Temperature correlated long-term faunal changes in the rocky intertidal. Unpublished student paper, Hopkins Marine Station, Stanford University Sagarin, R., Gilman, S., Baxter, C., Barry, J. (in press) Climate-Related Change in an Intertidal Community over Short and Long Time Scales. Ecological Monographs. Simons, A., Leydig, E. (1996). Long term changes in the Endocladia-Balanus community at Hopkins Marine Station, Pacific Grove, California. Unpublished student paper, Hopkins Marine Station, Stanford University. Sokal, R., Rohlf, F.J. (1995). Biometry: the principles and practice of statistics in biological research. New York: Freeman. Underwood, A.J. (1998). A method for analysing spatial scales of variation. Oecologia 117(4):570-578 Underwood, A.J., Chapman, M.G. (1996). Scales of spatial patterns of distribution of intertidal invertebrates. Oceologia 107:212-224 Underwood, A.J., Chapman, M.G. (1998). Spatial analyses of intertidal assemblages on sheltered rocky shores. Australian Journal of Ecology 23:138-157 Table 1. Analysis of Variance (ANOVA) Anthopleura sola Source Areas Transects(Areas) Within 108 Serpulorbis squamigerous Source Areas Transects(Areas) Within Endocladia muricata Source Areas Transects(Areas) Within 108 Rugosity Index Source Areas 16 Within MS 8.280 1.394 0.913 MS 12.728 4.097 2.364 MS 0.918 0.52 0.059 MS 0.036 0.082 5.939 0.020 1.527 0.156 3.107 0.151 1.733 0.152 0.231 1.766 8.787 0.000 0.460 0.720 Table 2. Results of Tukey's tests. Differences between Area means (row minus column); P-values in parentheses. Anthopleura sola Hewatt East2 0.107 (0.984) East2 Hewatt -0.592 (0.285) -0.699 (0.179) -1.041 (0.042) West -1.121 0.026) -0.422 (0.542) Serpulorbis squamigerus East! East2 Hewatt -1.545 (0.216 East2 -1.629 (0.190) Hewatt -0.084 (0.999) -1.608 (0.196) -0.063 (1.000) West 0.021 (1.000) Endocladia muricata East East2 Hewatt East2 0.081 (0.971) -0.320 (0.376) Hewatt -0.400 (0.217) West -0.025 (0.999) -0.106 (0.939) 0.295 (0.439) Figure Legends Fig. 1. Aerial map of Hopkins Marine Station showing the location of the Hewatt Transect. Dotted lines indicate the approximate locations of the four areas under study: West, Hewatt, East1, and East2. Fig. 2. Photograph of the study site taken at low tide. Fig. 3. Schematic representation of the two factor nested design with areas, transects- within-areas, and quadrat replicates. Fig. 4. The relationship between the F-distribution for Ho, the F-distribution for Hj, a, and B. When Ho is true, the probability of incorrectly rejecting it is a, and when Hi is true the probability of incorrectly rejecting it is ß. Note that without a specified alternative hypothesis, ß is unknown. However, the relationship between a and ß indicates that an increase in a will decrease B. Fig. 5. The distribution of Anthopleura. Each bar represents the mean of one transect. Error bars are Standard Errors. Fig. 6. The distribution of Endocladia. Each bar represents the mean of one transect. Error bars are Standard Errors. Fig. 7. The distribution of Serpulorbis. Each bar represents the mean of one transect. Error bars are Standard Errors Fig. 8. Correlation between Endocladia percent cover per quadrat and average tidal height per quadrat. Lower than two feet above MLLW is considered low zone, which is an inappropriate area in which to sample Endocladia. Fig. 9. Components of variance contributing to total variation seen in the three organisms sampled. Serpulorbis is represented by black, Anthopleura by light gray, and Endocladia by dark gray. Variance is divided into areas, transects-within-areas, and residual, or among-quadrat, variance. Fig. 10. Average substrate rugosity for each area. The height of each bar is the sample mean. Error bars are Standard Errors. FIGURE 1 Monterey Bay 25 km West Beach t Hewatt area 2 Wect arca .larea CLa4 2 ara Hewatt's: K Transeét: goig Loeb Laboratory □ N Bird Rock J Agassiz 2 Seal Rocks Fig. 3 Nested Analysis Intertidal West Hewatt Eastl Transect 1 Transect 2 Transect 3 Quadrat 1 .... Quadrat 10 East2 FIG. 4 :: P=.05 14 12 - 10 0 + Anthopleura Distribution West Area Easti Area East2 Area Hewatt Area Fig. 5 60 50 20 10 0 + West Area Endocladia Distribution Easti Area East2 Area Hewatt Area Fig. 6 20 Serpulorbis Distribution AE 1 West Area Easti Area East2 Area Hewatt Area Fig. 7 100 60 40 20 0 Endocladia Height Distribution 88. v O wdrae Tidal Height (Ft) West Area East1 Area East2 Area V Hewatt Area Fig. 8 90 - 80 70 60 50 5 40 30 - 20 - Variance Components — Serpulorbis - E Anthopleura Endocladia HE Area Transect(Area) Source of Variation Within Fig. 9 Fig. 10 West Area Average Rugosit) Easti Area East2 Area Hewatt Area