skip to Main Content skip to Section Navigation in text click here to go to site navigation in text
nms logo for banner graphic banner
click here to go to home pageclick here to go to site search
click here to go to the about the mbnms section click here to go to the visitors information section click here to go to the research and monitoring section click here to go to the resource management issues section click here to go to the education and research section

Overview of Research Program

Research Personnel

Research Activities Panel (RAP)

Technical Reports

Natural Resources

Submerged Cultural Resources

Site Characterization

Monitoring Programs

Research Platforms

Annual Sanctuary Currents Symposium

Map of Regional Marine Research Institutions

Searchable Databases

Tide and Weather Information

Related Regional Sites

 

first gov site link

 
  Research Technical Reports: THE The Effects Of Small-Scale Kelp Harvesting On Giant Kelp Surface Canopy Dynamics In The Ed Ricketts Underwater Park Region  

Cite as: Donnellan, M.D., M.S. Foster. 1999. The effects of small-scale kelp harvesting on Giant Kelp surface canopy dynamics in the Ed Ricketts Underwater Park region. Final Report to the Monterey Bay National Marine Sanctuary and the Cities of Monterey and Pacific Grove.

 

 

The Effects Of Small-Scale Kelp Harvesting On Giant Kelp Surface Canopy Dynamics In The Ed Ricketts Underwater Park Region

Final Report to the Monterey Bay National Marine Sanctuary and the Cities of Monterey and Pacific Grove

Michael D. Donnellan and Michael S. Foster
Coastal Solutions Group
Watsonville, CA 95076

SUMMARY
The effects of hand-harvesting giant kelp (Macrocystis pyrifera) on the surface canopy of kelp forests offshore of Monterey and Pacific Grove were investigated. This study was stimulated by concerns over the possible negative biological effects of recent intensive kelp harvesting by local aquaculture firms on giant kelp, rockfish, and sea otter populations, particularly when harvesting activities are concentrated within a small local area during the winter. We used aerial photographs dating to 1976 and kelp harvest records to address only the question of important harvest effects on giant kelp canopies, not the general effects of kelp harvesting. Three different periods of kelp harvesting relative intensity in this area have occurred since the early 1970s. a period of " moderate harvest" from 1972 - 1985, a period of "low harvest" from 1986 - 1995, and a period of " high harvest" from 1996 to present. Eight surveys were available from each of the periods of 1972 - 1985 and 1986 - 1995, and two surveys were available from the period of 1996 - 1998.

For the years in which aerial kelp surveys were available, hardcopy maps of kelp forest canopies were derived from aerial survey slides taken during the fall period of maximum canopy extent. The resultant maps were scanned into a computer, and image analysis software was used to measure kelp canopy surface area (the actual surface area of kelp canopy fronds floating on the surface), kelp forest spatial extent (the total surface area of a kelp forest canopy contained within its outer perimeter), and kelp canopy density (an index of canopy abundance within a given kelp forest ranging from " dense" to " sparse" ). Using a modified Before - After - Control - Impact (BACI) study design and analysis. of variance (ANOVA) statistical testing, these variables were compared between three harvested areas and one unharvested control area (Hopkins Marine Life Refuge) during each of the three periods of relative harvesting intensity. Since the period of interest (from 1996-1998) had a sample size of only two, we could not make rigorous, meaningful comparisons between this period and the other two periods. However, we were able to use a BACI approach (with two-sample t-tests) to test for a harvesting effect between the period of " moderate harvest" and the period of " low harvest." These comparisons were not directly applicable to the " high harvest" period of interest, but the results may suggest patterns to expect if the current level of kelp harvesting intensity continues. Regression analyses of the amount of kelp harvested during the winter versus the maximum amount of giant kelp surface canopy present during the following year were also performed.

No statistically significant differences in kelp canopies among periods were found using ANOVA analyses, and results of regression analyses were also statistically non-significant. No effects of current kelp harvesting practices on giant kelp canopies were detectable, but statistical power to detect an effect was low given the small sample size (n = 2) of the " high harvest" period and the inherent natural variability of kelp canopies. Therefore, these results do not necessarily indicate that there was not a harvesting effect, only that such an effect was undetectable given the available data. Separate comparisons of kelp canopies between " moderate harvest" and " low harvest ' periods (NOT the period in question) showed significant differences between the two periods at one of the three harvested study sites, but these results may have been confounded by kelp harvesting immediately prior to the aerial surveys. If these results were not confounded, they suggest that harvesting concentrated within a relatively small area may negatively affect giant kelp canopies. Continued yearly aerial surveys during the period of fall maximum kelp canopy are needed in order to resolve this issue.

INTRODUCTION
Recent public debates concerning potential regulations for the newly-created Ed Ricketts Underwater Park have addressed the appropriateness of certain "uses" occurring in the region, particularly the harvesting of giant kelp. Various ocean-related user groups have asserted that kelp harvesting operations off the Monterey Peninsula may be causing a decline in the amount of surface canopy in the area extending from the Monterey Coast Guard Breakwater to Lovers Point. Indeed, one aerial photograph (which was displayed in a public forum) of Hopkins Marine Life Refuge and the adjacent area taken during winter/spring 1998 showed a relatively abundant kelp canopy within the reserve (where kelp harvesting is prohibited) and little canopy in the adjacent areas where harvesting occurs. In the spirit of science-based resource management, the Monterey Bay National Marine Sanctuary and the cities of Monterey and Pacific Grove commissioned this report to more thoroughly examine the effects of kelp harvesting on harvested canopies. Using a time series of aerial kelp canopy surveys dating to 1976 and available kelp harvest records, we addressed the following question: Are current kelp harvesting practices causing a detectable effect on the abundance of giant kelp (Macrocystis pyrifera) surface canopy, as measured during the period of fall maximum surface canopy, in the Ed Ricketts Underwater Park region?

BACKGROUND
Kelp has been harvested in the state of California to varying degrees since early this century, and as early as 1914 in Monterey (Scofield, 1959). Ironically, a sample of wet tons of kelp taken from in front of the Del Monte Bathhouse (Monterey) led to the first-ever objection to kelp harvesting, with the claimant stating that it would ruin sardine fishing in Monterey Bay (Scofield, 1959). Kelp harvesting in the Monterey area began in earnest in the early 1930s; 200 tons were taken in 1930 and 500 tons were taken in 1931 (see reviews in Phillips, 1932; Scofield, 1959).

Two different species of kelp (a generic term for large brown algae) have been harvested historically, Nereocystis luetkana (also called bull or bladder kelp) and Macrocystis pyrifera (commonly called giant kelp). Interestingly, the Monterey-area kelp harvest in the early 1930s was exclusively composed of bull kelp, the dominant canopy forming algal species north of Point Sur during that period (Phillips, 1932; Andrews, 1945; Scofield 1959). Phillips (1932) reported that bull kelp was harvested in the Monterey area only along the coast between Point Sur and Point Pinos, implying that this kelp was not abundant enough within southern Monterey Bay to warrant harvesting. Until at least 1962, large stands of giant kelp on the central California coast were apparently limited to protected sites within southern Monterey Bay, northern Carmel Bay, and south of Point Sur (Philips, 1932; Andrews, 1945; Scofield, 1959; McLean, 1962; reviewed in Larson and McPeak, 1995 [p. 3-74 - 3-77]). This distribution is in striking contrast to the present-day; giant kelp is now the dominant kelp as far north as Ano Nuevo Island, except on shallow, exposed rocky reefs. This dramatic shift in relative abundance of bull kelp and giant kelp coincided with, and may have been facilitated indirectly by the reestablishment of the southern sea otter (Enhydra lutris) in its former range (Van Blaricom, 1984).

Kelp has primarily been harvested on a large scale for its chemical byproducts. This harvesting occurs mostly in southern California where giant kelp is more abundant than in central California. Approximately 95 % of the current state-wide harvest of around 100,000 wet tons per year (Leet et al., 1992) is done by Kelco Co., a subsidiary of Monsanto based in San Diego (Spratt, personal communication). Kelco harvests kelp with large ships equipped with mechanical blades that can harvest up to 600 tons per day (Larson and McPeak, 1995 [hereafter cited as CDFG E.I.R.]). Occasionally, Kelco harvests as far north as Spanish Bay off the Monterey peninsula when kelp growth is poor in southern California (Spratt, personal communication). Most of the remaining 5% of kelp harvested is taken by small-scale hand-harvesters for aquaculture purposes (Spratt, personal communication).

Due to catastrophic declines in wild abalone populations and sustained market demand, numerous abalone aquaculture facilities that require kelp for abalone feed have begun operation since the early 1970s. Four such facilities have operated in the Monterey Bay area since the late 1980s (Monterey Abalone Company, Pacific Abalone Farm, Pacific Mariculture, and U.S. Abalone Inc.). Two of the larger abalone farms, Pacific Mariculture and U.S. Abalone, are based in or near Santa Cruz, and they fulfill most of their kelp requirements from the northern Monterey Bay (D. Ebert, personal communication). Pacific Mariculture contracts an independent kelp harvester, the California Kelp Company, to supply their kelp. Occasionally, the Santa Cruz -based harvesters harvest kelp on the Monterey side of the bay when the Santa Cruz harbor mouth becomes dangerous for boat passage during winter storms or is obstructed by a sand bar during the winter months.

Monterey Abalone Company and Pacific Abalone Farm are based in Monterey harbor and harvest year-round from the Monterey peninsula area (Seavey, personal communication). Grillo Enterprises, a local firm engaged in kelp harvesting for use in the herring-roe-on-kelp fishery in San Francisco Bay, also takes a small amount of kelp from the Monterey area (Larson and McPeak, 1995). During winter, seas become increasingly treacherous as one travels west from Lovers Point to Point Pinos, and southwest of Point Pinos the coastline is exposed to the open ocean and too dangerous to harvest using small boats. Since current kelp hand-harvesting operations are conducted using skiffs equipped with small outboard motors, winter harvesting is concentrated along the 4 kilometers of relatively wave-protected coastline between the Monterey Coast Guard Breakwater and Otter Point. Since the controversy over kelp harvesting in the Monterey / Pacific Grove area began around 1996, harvesters have largely avoided harvesting the highly-visible kelp canopies from the Monterey Breakwater to Hopkins Marine Station (Seavey and Williamson, personal communication), thereby concentrating kelp harvesting between Hopkins Marine Station and Otter Point during the winter months. During the winters (defined here as November through March of 1996/1997 and 1997/1998, 414 tons of kelp were harvested each winter, primarily from this local area. It is this volume of harvesting from a relatively small area, and the related concerns over the potential effects of harvesting from the same kelp plant multiple times, that has led to questions concerning the possible harvesting impacts on kelp forest ecology in this local region.

Numerous studies dating to 1915 have investigated the potential effects of harvesting giant kelp. The 1995 edition of the California Department of Fish & Game's Final Environmental Document, Giant and Bull Kelp Commercial and Sport Fishing Regulations (Section 30 and 165, Title 14, California Code of Regulations, edited by Larson and McPeak) provides an excellent review of these studies, including the effects of harvesting on frond growth and regeneration, holdfast development, individual plant ' survivorship, survivorship of populations of plants, and effects on plants associated with giant kelp (Cameron, 1915; Crandall, 1915; Brandt, 1923; Limbaugh, 1955; Clendennina, 1968; North, 1968; Miller and Geibel, 1973; Kimura and Foster, 1984; Barilotti et al., 1985; McCleneghan and Houk, 1985; Barilotti and Zertuche, 1990; reviewed in CDFG E.I.R., 1995 [p.4- 14 - 4- 18]). Most of this research addressed the effects of large-scale mechanical kelp harvesting, but is also applicable to hand-harvesting practices. The editors of the California Department of Fish and Game Final Environmental Document (CDFG E.I.R., 1995) concluded from these studies that no significant effects are apparent from routine mechanical and hand-harvesting practices as long as individual plants are harvested no more than three times per year. Research in southern and central California has suggested that overharvesting (i.e. cutting the fronds of the same plant four or more times within 12 months) results in decreased yield and reduced plant survivorship, and increases in associated understory plants (Brandt, 1923; North, 1957; Miller and Geibel, 1973).

Kelp harvesting clearly has short-term effects: the abundance of kelp canopv and associated invertebrates decreases when kelp is cut and removed from the nearshore environment by harvesting. Kelp fronds grow very little after being severed, and uncut fronds growing up from four feet below the surface (the maximum harvested depth by law) or lower must replace the surface canopy (Crandall, 1915; Brandt, 1923). North (1968) reported that the initial growth rate of kelp plants was retarded after being harvested, but no significant difference was detectable between harvested plants and uncut control plants within a month after harvesting. The rate of surface canopy regeneration following harvesting is variable, however, depending on conditions such as the proximity of the growing subsurface kelp fronds to the surface and plant growth rate (which is primarily dependent on nutrient availability and irradiance levels; see review in CDFG E.I.R., 1995). Based on models and empirical evidence, North (1968) concluded that harvesting of the kelp canopy stimulates growth [by increasing irradiance to subsurface fronds] or retards it, depending on conditions before and after harvesting.

The effects of kelp harvesting on populations of fishes, birds, and mammals are also reviewed in the CDFG E.I.R. (1995, p.4-1 - 4-10). Fish-oriented research has addressed potential harvesting effects on fish eggs, juveniles and adults of various species, sportfishing, and young-of-the-year rockfish in central California (Limbaugh, 1955; Davies, 1968; North and Hubbs, 1968; Quast, 1968; Miller and Geibel, 1973; and Houk and McCleneghan, 1993). The CDFG E.I.R. (1995) concluded from these studies that no significant ecological effects on fish populations occur due to kelp harvesting. Similarly, the CDFG reviewed research on invertebrates and concluded from these studies that no significant effects are apparent on the populations of either motile (e.g., the isopod Idotea and the snails Tegula and Calliostoma [6 species]) or non-motile (e.g.. bryozoans and hydroids) species (Quast, 1968 and Miller and Geibel, 1973). All non motile canopy invertebrates on a given kelp frond are removed by harvesting, but these animals occur in the naturally-ephemeral kelp canopy environment and no doubt settle, grow, and reproduce rapidly relative to kelp harvest frequency (CDFG E.I.R., 1995).

Based on scant available data, the CDFG E.I.R. (1995) concluded that there is no significant effect of kelp harvesting activities on bird or mammal populations. The sea otter (a species protected by state and federal laws) has the closest association with kelp canopies among all marine mammals (CDFG E.I.R., 1995), and is of particular concern in the Monterey area. It has been suggested that kelp harvesting operations occurring in southern Monterey Bay during the winter may further reduce the already sparse kelp canopy available to sea otters for rafting and resting, thereby potentially having a negative impact on the otter population. The question of whether or not kelp canopy is a limiting resource for sea otters has not been rigorously investigated.

Exhaustive reviews of kelp biology and kelp forest communities are available (reviewed in Foster and Schiel, 1985, and CDFG E.I.R., 1995 [p. 3-5 - 3-10, 3-28 - 3-30, 3-44 - 3-51]), and thus will not be summarized here. However. we do emphasize that kelp forests (and therefore their canopies) are well documented to be highly variable in space and time, extremely sensitive to environmental conditions (e.g. storm waves, nutrient availability, etc.), and highly productive. Giant kelp also plays a central role in supporting a species-rich, biologically productive community of other organisms in the nearshore environment.

For the purposes of this study, we defined an " important" effect of kelp harvesting to be a statistically significant decrease in the amount of giant kelp surface canopy in harvested kelp canopies (relative to an unharvested control canopy) after intensive kelp harvesting began. A decrease in the amount of kelp canopy within the harvested areas may be indicative of reductions in plant growth or recruitment, and/or increased plant mortality, thereby constituting a possible important biological or ecological effect. We chose to compare kelp canopies during the period of maximum kelp canopy development (or " fall maximum" ) primarily because most data are available for this time of year. As a result, we could not test whether or not canopy regeneration rates differed between harvested and unharvested areas. However, to test for the potential impact we defined, canopy data available during the fall maximum are ideal; confounding short-term harvesting effects (i.e. canopy loss) are minimized during this time because seas are relatively calm and harvesting within the study area is reduced. Given the opportunities and constraints of these available data, we asked a very specific question: are current kelp harvesting practices causing a detectable effect on the abundance of giant kelp surface canopy, as measured during the period of the fall maximum, in the Ed Ricketts Underwater Park region? We did not investigate related issues of possible effects of harvesting on sea otters, juvenile rockfish recruitment, kelp canopy invertebrates, or other aspects of kelp communities.

MATERIALS AND METHODS
Two primary raw data sets, kelp harvest records and kelp canopy aerial photos, were used to address the question of the effects of local kelp harvesting on giant kelp surface canopy. Kelp harvest records for bed 220 (Monterey Breakwater to Cypress Point; see Figure 1) were compiled from monthly harvest records in the California Department of Fish & Game (CDFG) archives dating to 1987 (Table 1 and Figure 2). CDFG harvest data were cross-checked for accuracy when possible with officially submitted monthly harvest reports prepared by the individual harvesters. Some discrepancies were found. In these cases, data were modified based on the harvest reports. Bed 220 is much larger than the study area (Monterey Coast Guard Breakwater to Lovers Point in Pacific Grove), so harvest amounts reported could have come from the study area (and some, if not most, surely did during the winter), but not necessarily. As shown in Table 1, Kelco has harvested from bed 220 at least five times since 1987, but they have only harvested that portion of bed 220 between Cypress Point and Point Pinos (Spratt, personal communication). Thus, Kelco harvesting records were irrelevant to the question of effects in the study area and were not used in the analyses.

CDFG harvest records for bed 220 prior to 1987 are unavailable (Spratt, personal communication), so amounts of kelp taken before 1987 had to be estimated based on anecdotal information provided by George Lockwood (personal communication), former proprietor of the Monterey Abalone Farm that operated on Cannery Row from l 972 circa 1989. Lockwood suggested that kelp harvesting during this period ranged from 200 pounds per day or every other day during the early years, and up to 2000 pounds per day or every other day when the business peaked in 1985. Our estimates for this period are probably slight underestimates since we did not include kelp harvested by the CDFG abalone culture facility at Granite Canyon which operated during this same period (E. Ebert, personal communication). The scale of this operation was small (10,000 - 20,000 young abalone), and thus its kelp needs were negligible. Most of the kelp feed for their abalone came from the Granite Canyon area, and kelp was only occasionally harvested (i.e. 8-10 times per year) from the Monterey Breakwater area (Ebert, personal communication).

The kelp canopy aerial photographs were composed of four subsets of aerial surveys. Each of these subsets used low-altitude aerial photography with infrared sensitive film. Years marked with asterisks in Figure 2 denote years in which surveys were available during the period of kelp canopy maximum and of sufficient quality to incorporate in the analyses. The CDFG surveyed kelp canopies regularly from l 974 to 1979, and again in 1989. The surveys done by CDFG in early 1970s were more-or-less exploratory in nature; methods were standardized and refined in the mid-1970s. The quality of the data from the early 1970s was thus questionable and not used in this study. Kelp canopy maps for the area from Seaside to Point Sur were photographed on a near monthly basis from November 1985 to December 1991 (Harrold and Watanabe, unpublished data). These surveys were available already rendered (see definition below) from infrared slides, and one map series per year from the fall maximum was used in the analyses. Van Wagenen filled in the data gaps in the aerial survey time series (when possible) by generously providing access to his personal photographic archives (i.e. 1980. l 981, 1982, 1983, 1984, l 992, 1994, and 1997). All aerial surveys since 1980 were done by the same person using identical techniques, thereby minimizing sampling variation due to photographic technique.

Kelp maps for the area between Lovers Point and the Monterey Breakwater were rendered from the original imagery according to methods used by Van Wagenen (personal communication). In brief, individual photos (slides) were projected by a Minolta color enlarger (MOD- l) onto two contiguous 1: 12000 scale baseline maps of the coastline study area. The shoreline of the baseline map was aligned with the projected image, and magnifying eyeglasses (+4.0) were used to aid "transfer" of kelp surface canopy from image to map. To transfer data from image to map, the portion of the projected image interpreted as kelp canopy was shaded in on the baseline map with a Sanford Vis a Vis wet-erase fine point overhead transparency marker (series 16000). Areas within the greater perimeter of the kelp forest canopy that were devoid of surface canopy were left unshaded. Only the center 75% of the projected image was rendered in order to minimize bias due to distortion. This could be done without loss of information because the aerial photographs were overlapping. The rendered baseline maps that were used in the analyses are shown in Figure 3 (in miniaturized form).

The rendered baseline maps were scanned into a PC with a Hewlett-Packard scanner (model C6260a) as a bitmap file at 350% resolution and imported into Jandel Sigma Scan Pro (v.4.0) image analysis software. Once the baseline maps were imported and the map scale was calibrated, the variables Canopy Area and Planimeter Area were calculated. Canopy Area represented the spatial extent (in hectares) of kelp fronds floating on the sea surface within a defined geographic area (i.e. " gaps" in the kelp canopy were not included in the area measurement). Planimeter Area represented the spatial extent of a kelp forest within the boundaries of its outer canopy perimeter (i.e. " gaps" in the kelp canopy were " filled" and included in the area measurement), and was derived from Canopy Area using pre-determined conventions that filled in the gaps. A third variable, Relative Density Index (R.D.I.) was calculated by dividing Canopy Area by Planimeter Area. This derivation approximates the probability of encountering kelp on the surface at a random point within the perimeter of the kelp forest; R.D.I. values approaching l indicate a " dense" surface canopy and values approaching 0 indicate a " sparse" surface canopy (VanWagenen, personal communication).

Study Design:
The question addressed in this study was evaluated with a modified Before - After - Control - Impact (BACI) approach. The BACI design is currently the most statistically rigorous way to test for "impact" effects (Stewart-Oaten et al., 1986; Schmitt and Osenberg, 1996). In brief, this approach permits testing of an impact hypothesis by comparing the abundance of a variable at a Control site to a potential Impact site both Before the impact begins and After the impact has begun. If the differences between the abundance of the measured variable at the Control and Impact site Before the impact has begun are not statistically different from the differences between the Control and Impact site After the impact has begun, then it would be concluded that there is no impact. The difference values are determined multiple times in both the Before and After periods, and thereby constitute Before and After replicates.

Designation of discrete Before and After periods was not possible with the available data. Instead, Figure 2 suggests three general "periods" of kelp harvesting relative intensity in the Monterey area: a "moderate harvest" period from 1971 - 1985, a "low harvest" period from l 986 - 1995, and a "high harvest" period from 1996 - present. For the purposes of this study, " low harvest" was arbitrarily defined as < 50 tons/year "moderate harvest" as 50 - 250 tons/year, and "high harvest" as > 250 tons/year. According to this convention, the 1989 harvest of 84 tons and the 1990 harvest of approximately 200 tons both fall into the category of "moderate harvest," even though these two years occurred within the period of " low harvest." Therefore, data from these years were not included in the analyses. Ideally, the data from the Before period should be collected before any potential impact begins. For comparison purposes in this study, the period of " low harvest" was used as a " best approximation" of a Before period in which kelp was not harvested, even though this period occurred after a period of " moderate harvest." Conceptually, it is not a serious problem that the Before period occurred after a period of moderate harvesting because giant kelp plants on the central California coast exhibit a relatively fast " turnover" rate (on the order of 1 -2 years). Therefore, it is improbable that prolonged effects on kelp canopies would have occurred due to the relatively small amount of harvest from 1972- 1985.

The study area was divided into one unharvested Control site (Hopkins Marine Life Refuge [HMLR]) and three potential Impact sites (San Carlos, Cannery Row, and Lovers Point) which are open to kelp harvesting (Figure 4). Sites were designated when possible by persistent natural features such as sand channels, and were similar in spatial area for comparability. The HMLR control site was delineated by the reserve boundaries; however, only the main kelp forest directly offshore and surrounding Cabrillo Point was used to calculate Canopy and Planimeter Area. HMLR was chosen as a control site for two important reasons. First, the reserve was established in 1892 and therefore was not harvested during the study period. Second, it is conveniently nestled within the study area and thus influenced by essentially the same range of natural conditions (e.g. storm waves, nutrient availability) as the harvested sites. An important assumption was also made that the HMLR kelp forest was similar in other respects (e.g. depth of bed, substrate, topography, etc.) to the other kelp forests studied. Another control site would have been desirable to help assess canopy variation unrelated to harvesting, but no such site was available.

Testing for a harvesting effect on kelp bed surface canopy was done by analyzing the differences between the variables of Canopy Area, Planimeter Area, and R.D.I. at HMLR and each harvested site among the periods of moderate, low, and high harvest (see Stewart-Oaten et. al., 1986 for detailed account of experimental design). This was achieved by subtracting the raw abundance for a given variable at a Control site from the raw abundance of the same variable at a harvested Impact site. This "difference" value was calculated for the three variables at each impact site for each year within each of the three different periods. Control and Impact sites were sampled (i.e. photographed) simultaneously, and each sampling time was represented in the analysis by only one value for each of the three variables measured, the difference between the Control and each Impact for that time. Thus, the sampling times (17 different years) were used as replicates. For example, the Canopy Area at HMLR in 1976 (3.27 hectares) was subtracted from the Canopy Area at the Cannery Row site in 1976 (5.52 hectares), resulting in a " difference" value of 2.25 hectares; this same calculation was performed for each year in which aerial kelp surveys were available for a total of 17 values.

The mean differences between Control and Impact sites within each of the three periods ("moderate harvest", "low harvest", and "high harvest") were tested against each other using a one-way, 3-level Model I analysis of variance (ANOVA) for each measured variable (Canopy Area, Planimeter Area, and R.D.I.). A kelp harvesting effect would be indicated by statistically significant differences among the three harvest periods. Since the " high harvest" period of interest (from 1996- 1998) had a sample size of only two, comparisons involving this period will not be informative until more data is collected. Therefore, a more powerful BACI approach (with one-tailed, two-sample t-tests) was also used to test for a harvesting effect between the period of " moderate harvest'' and the period of " low harvest." These comparisons were not directly applicable to the "high harvest" period of interest, but the results may suggest patterns to expect if the current level of kelp harvesting intensity continues.

BACI testing does not distinguish a harvesting effect from any other environmental effects that might have occurred coincident with harvesting. Therefore, regression analyses of each measured variable versus the amount of kelp harvested the previous winter for each site were also done. Again, this test does not assess causation but can be used as circumstantial evidence of harvesting effects. The independent variable chosen to regress against each kelp forest variable was the amount of kelp harvested from bed 220 during the previous winter (defined as November - March). This choice was logical because most of the kelp harvested during the winter would have likely come from within the study area, and therefore should correlate more closely to effects observed within the study area than would total yearly harvest from bed 220.

All statistical analyses were done using Systat v. 7.0.1.

RESULTS
Annual variability of the fall maximum kelp canopy was quite high from 1976 1998. Nevertheless, Canopy Area, Planimeter Area, and R.D.I. measurements exhibited close correlation among all study sites (Table 2 and Figure 5). Note that the lines connecting data points in Figure 5 are for illustrative purposes only and do not include seasonal variation between sampling periods. Missing datapoints in the trend-line of Figure 5 represent years in which aerial surveys were not available (i.e. 1978, 1979, 1993, 1995, and 1996). Canopy Area, Planimeter Area, and R.D.I apparently declined between 1991 and 1997, although this can't be confirmed due to lack of data (Figure 5). This possible decline could not have been due to kelp harvesting since large harvests were not made until 1996, and Canopy Area, Planimeter Area, and R.D.I. at unharvested HMLR also declined during these years. Environmental factors, possibly conditions associated with frequent and persistent El Nino events, may have caused consistently low surface canopies during this time.

As with the canopy abundance data, some of the "difference" values calculated from the raw abundance data also displayed high variability (Table 3 and Figure 6). In particular, the Lovers Point and San Carlos sites exhibited substantial within-period variation of Canopy Area and Planimeter Area during the periods of " moderate harvest" (1976 - 1985) and "low harvest" (1986 - 1995). This variation was not problematic in the analyses, however, since all " difference" values within a given harvesting period were averaged, but this variation did result in high standard errors for some of these averages (especially for the " high harvest" period due to small sample size). ANOVA comparisons of the average "difference" values among the three periods at each Impact site (for each variable) were not statistically significant (P> 0.05; Table 4 and Figure 7), indicating that no effect of kelp harvesting on the three measured variables was detectable. However, statistical power to detect such an effect was low (Power < 25 desired Power = 80), resulting, in part, from the small sample size of period 3 (1996 1998; n = 2). Given the natural variability of the kelp canopy, this small sample size would allow the detection of only extremely large harvesting effects.

BACI analyses of the " moderate harvest" and " low harvest" periods with one tailed. two-sample t-tests allowed more powerful hypothesis testing of harvesting effects. Most comparisons yielded non-significant results, but Canopy Area at the San Carlos study site exhibited a significant difference between periods (P = 0.036; Table 5). This indicates that the average abundance of kelp canopy was depressed during the " moderate harvest" period of 1976 - 1985, and then increased (on average) during the "low harvest" period from 1986 - 1995. However, the result for Canopy Area at San Carlos period of " low harvest." These comparisons were not directly applicable to the high harvest "period of interest, but the results may suggest patterns to expect if the current level of kelp harvesting intensity continues.

BACI testing does not distinguish a harvesting effect from any other environmental effects that might have occurred coincident with harvesting. Therefore, regression analyses of each measured variable versus the amount of kelp harvested the previous winter for each site were also done. Again, this test does not assess causation but can be used as circumstantial evidence of harvesting effects. The independent variable chosen to regress against each kelp forest variable was the amount of kelp harvested from bed 220 during the previous winter (defined as November - March). This choice was logical because most of the kelp harvested during the winter would have likely come from within the study area, and therefore should correlate more closely to effects observed within the study area than would total yearly harvest from bed 220.

All statistical analyses were done using Systat v. 7.0.1.

RESULTS
Annual variability of the fall maximum kelp canopy was quite high from 1976 1998. Nevertheless, Canopy Area, Planimeter Area, and R.D.I. measurements exhibited close correlation among all study sites (Table 2 and Figure 5). Note that the lines connecting data points in Figure 5 are for illustrative purposes only and do not include seasonal variation between sampling periods. Missing datapoints in the trend-line of Figure 5 represent years in which aerial surveys were not available (i.e. 1978, 1979, 1993, 1995, and 1996). Canopy Area, Planimeter Area, and R.D.I apparently declined between 1991 and 1997, although this can't be confirmed due to lack of data (Figure 5). This possible decline could not have been due to kelp harvesting since large harvests were not made until 1996, and Canopy Area, Planimeter Area, and R.D.I. at unharvested HMLR also declined during these years. Environmental factors, possibly conditions associated with frequent and persistent El Nino events, may have caused consistently low surface canopies during this time.

As with the canopy abundance data, some of the "difference" values calculated from the raw abundance data also displayed high variability (Table 3 and Figure 6). In particular, the Lovers Point and San Carlos sites exhibited substantial within-period variation of Canopy Area and Planimeter Area during the periods of " moderate harvest" (1976 - 1985) and "low harvest" (1986 - 1995). This variation was not problematic in the analyses, however, since all " difference" values within a given harvesting period were averaged, but this variation did result in high standard errors for some of these averages (especially for the " high harvest" period due to small sample size). ANOVA comparisons of the average "difference" values among the three periods at each Impact site (for each variable) were not statistically significant (P> 0.05; Table 4 and Figure 7), indicating that no effect of kelp harvesting on the three measured variables was detectable. However, statistical power to detect such an effect was low (Power < 25 desired Power = 80), resulting, in part, from the small sample size of period 3 (1996 1998; n = 2). Given the natural variability of the kelp canopy, this small sample size would allow the detection of only extremely large harvesting effects.

BACI analyses of the " moderate harvest" and " low harvest" periods with one tailed, two-sample t-tests allowed more powerful hypothesis testing of harvesting effects. Most comparisons yielded non-significant results, but Canopy Area at the San Carlos study site exhibited a significant difference between periods (P = 0.036; Table 5). This indicates that the average abundance of kelp canopy was depressed during the " moderate harvest" period of 1976- 1985, and then increased (on average) during the "low harvest" period from 1986 - 1995. However, the result for Canopy Area at San Carlos was probably confounded by short-term harvesting effects if kelp harvesting was localized in this area year-round during the 1970s and early 1980s.

None of the slope coefficients obtained from regression analyses of canopy with the previous winter's harvest were statistically significant (P>0.05; see Table 6 and Figure 8), indicating an undetectable harvesting effect. Two clusters of data points are evident in Figure 8, based on two primary levels of winter harvest (harvest greater than 400 tons and harvest less than 100 tons). No relationship between winter harvest and any of the three variables measured at each site was apparent when winter harvest was less than 100 tons. Note that the slope coefficient at HMLR displayed essentially the same downward trend as the other sites even though HMLR was not harvested; this result, though statistically non-significant, would not be expected if harvesting had an effect. Also, note the extremely wide 95% confidence intervals in Figure 8 due to data from the " high harvest" period. As with the ANOVA results, however, statistical power to detect a significant trend was low due to only 2 samples taken during the "high harvest" period. Some comparisons failed a test of regression assumptions and the results were not reported.

DISCUSSION & RECOMMENDATIONS
There was no detectable effect of kelp harvesting, as currently practiced off Monterey and Pacific Grove, on the density or spatial extent of the fall maximum kelp canopy. However, it must be emphasized that statistical power to detect such an effect was very low given the small sample size (n = 2) and high natural variability of kelp canopies. Therefore, these results do not necessarily indicate that there wasn't a harvesting effect, only that such an effect was undetectable given the available data. Because of the small sample size and high natural variability, we could have only detected a very large harvesting effect, so there could indeed have been an impact ranging from the limit of what we could detect to no impact at all. Increasing sample size (i.e. more yearly aerial kelp canopy surveys in the future) is the only way to further refine these results. Assuming future winter harvesting is maintained at present (or greater) levels within the study area and no persistent environmental factors are confounding, four or more annual aerial kelp surveys are necessary to meaningfully assess harvesting effects. The inherent limitation to greater resolution of our question is a fixed sample size (n = 6 for the period of " low harvesting" from 1986 - 1995). Statistically, it would be ideal to compare periods with equal sample sizes, in which case a one-tailed, two-sample t-test could be used to compare the " low harvest" period with the " high harvest" period.

Since we could not make rigorous, meaningful comparisons with data from the " high harvest" period because of insufficient sample size, comparisons between the periods of " moderate harvest" and " low harvest" might suggest patterns to expect within the Ed Ricketts Park region in the future. BACI testing yielded statistically non significant differences between the " moderate harvest" and " low harvest" periods at the Cannery Row and Lovers Point sites, but a significant difference in Canopy Area at the San Carlos site. However, this result may have been confounded by short-term harvesting effects (and we suspect it was) if kelp harvesting was localized in this area year-round during the l 970s and early 1980s. One of the assumptions we made in the analyses was that short-term canopy effects (i.e. missing canopy due to recent harvesting) were negligible during the fall period of maximum canopy. This assumption was probably violated during the "moderate harvest" period of 1976 - 1985, however, because Monterey Abalone Farm most likely concentrated their kelp harvesting activities in the San Carlos area (since San Carlos was the closest kelp forest to the harbor mouth, it was thereby the easiest and most economical canopy to harvest). If these results were not confounded, they suggest that a harvesting effect on the abundance of kelp canopy may occur if extant kelp harvesting activities are focused in a localized area, particularly in the vicinity of Lovers Point and Otter Point during the winter.

As discussed in the Background section above, previous research has suggested that routine kelp harvesting operations have unimportant effects on canopy yield and plant mortality, with the caveat that individual plants were not harvested more than three times per year. The concentration of kelp harvesting effort at Lovers Point and Otter Point during the winter (i.e. 414 tons taken from bed 220 during the winters of 96/97 and 97/98) increase the likelihood that an individual kelp plant may be repeatedly harvested and thus result in negative effects. The recent designation of a no-harvest area between Lovers Point and the Monterey Breakwater will provide a useful second Control area, but also may exacerbate the possibility of multiple pruning since harvesting would be further concentrated in other areas.

To put the amount of kelp currently harvested into perspective, Harrold et al. (1998) estimated that the kelp forests occurring from the Monterey Breakwater to Cypress Point (Bed 220) generate more than 200,000 tons of drift kelp per year (wet weight production estimate from Gerard, 1976). The study area is approximately 1/6 of the size of Bed 220, and therefore a rough estimate of drift kelp production from this area is 33,000 tons per year. The current annual kelp harvest in the study area by local aquaculture firms is approximately 600 tons per year, or 2% of the estimated amount of drift kelp from this same area. Therefore, further reasoned debate over the effects of kelp harvesting and its management might be sharpened by a consideration of scale.

ACKNOWLEDGMENTS
The authors would like to thank the many individuals and organizations who were integral to successful completion of this report, particularly R.F. VanWagenen of Ecoscan Resource Data for doing the important 1998 aerial kelp survey and generously providing access to archived kelp canopy surveys - this study would not have been possible without his assistance. We also graciously thank Chris Harrold and the Monterey Bay Aquarium for providing an invaluable data set of aerial kelp surveys, the California Department of Fish and Game and Rob Collins for providing access to kelp harvest records and computer workspace, Steve Webster and the Monterey Bay National Marine Sanctuary for facilitating last-minute funding for the 1998 aerial kelp survey (contract order # 40ABNC801827), and the cities of Monterey (contract order # 95110021612) and Pacific Grove (contract order # 9899-308) for major funding. Dave Ebert of U.S. Abalone, Art Seavey of the Monterey Abalone Company, Bill Williamson of the California Kelp Company, and Gary Russell of the Pacific Abalone Farm were all very helpful in providing information on harvesting activities and unselfishly providing access to their kelp harvest records. We appreciate the valuable historical information provided by Earl Ebert and George Lockwood, and regulatory input from Jerry Spratt at CDFG. Thanks also to the reviewers of this report, including Steve Webster and Chris Harrold of the Monterey Bay Aquarium, Greg Cailliet of Moss Landing Marine Laboratories, and especially Kari Boylan of Larkin Elementary School. Helpful software consultation was provided by Jennifer Fisher and Marty Gingras of CDFG.

PERSONAL COMMUNICATION
Ebert. David. Proprietor, U.S. Abalone Inc. Davenport, CA

Ebert, Earl. Research biologist (retired). Granite Canyon Abalone Laboratory, California Dept. of Fish and Game. Monterey, CA

Lockwood, George. Proprietor, Monterey Abalone Farm (retired). Monterey, CA

Seavey, Arthur. President, Monterey Abalone Company. Monterey, CA.

Spratt, Jerry. Associate Biologist. California Dept. of Fish and Game. Monterey, CA (831) 649-2870

VanWagenen, Robert. Proprietor, Ecoscan Resource Data. P.O. Box 1046, Freedom, CA 95019 email: ecoscan@aol.com

Williamson, Bill. Proprietor, California Kelp Company. Monterey, CA.

FIGURES AND TABLES

Figure 1

Figure 2

Figure 3a

Figure 3b

Figure 3c

Figure 3d

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Table 1

Table 2

Table 3

Table 4a

Table 4b

Table 5

Table 6

 

   
National Marine Sanctuaries | National Ocean Service | National Oceanic and Atmospheric Administration
U.S. Department of Commerce | NOAA Library

click here to go to the NOAA home page For Website comments/questions, contact the MBNMS Webmaster.
For programmatic comments/question, contact the appropriate MBNMS Staff.
MBNMS Privacy Statement
This page last modified on: 01/14/02
click here to go to the national marine sanctuaries home page

URL: http://montereybay.noaa.gov/research/techreports/kelpcover/kelpcover.html