Astrid Leitner
Total score: 83/100
Title [[4/4 – nice and descriptive]]
Evaluation of Both Physical and Biological Habitat Associations for the Benthic Invertebrate Community in a Temperate Kelp Forest Ecosystem
Clarity [[12/14 – work on varying sentence length and omitting any extraneous words to give your writing more oomph.
Introduction [[15/20 – I especially like your ecological context and examples. You have the main elements, but it’s not flowing together terribly well and you didn’t address the novelty of this study]]
Understanding how diversity is maintained within an ecosystem has long been the a goal of ecologists. This is a complex and debated topic; however, ecologists have identified niche partitioning as an important process for maintaining high diversity. Niche partitioning maintains high diversity by reducing competition among species; instead of competing for the same resource the species use different parts of the same resource or niche, thus functionally dividing the resource between them (Schoener 1974) [[nice description]]. This allows for coexistence of previously competing speciesby reducing the size oftheir realized niches relative to their fundamental niches.One key way to study resource partitioning is through habitat association studies. By understanding each organism’s individual habitat, one can begin to piece together how different niches are partitioned throughout the coexisting community, and therefore how diversity is maintained [[nice]].
Species-habitat association studies have a long history in both terrestrial and marine ecosystems. For example, Robert Macarthur conducted a classic study highlighting how multiple species of warbler can persist by utilizing different sections of the same tree (Macarthur 1958). Resource partitioning has also been cited as a major cause of the present diversity of rockfishes within the Pacific coast kelp forest ecosystem (Hallacher and Roberts 1985). Species-species associations have also been previously documented; for example, in tropical reefs, associations between fish assemblages and corals as well as macroalgae are well documented (Syms 1995). However, in temperate waters there has been considerably less work done on species associations. Moreover, there are no studies that I am aware of that have attempted to link species to species as well as species to physical habitat structure and then attempted to evaluate the relative importance and strengths of these associations, especially not in temperate reefs. In this respect, this is a novel study. [[good examples with citations]]
A good approach to studying both types of associations (species-species and species-habitat) is to look at highly diverse communities such as tropical rain forests, coral reef systems, and kelp forests. In such systems resource partitioning maintains high levels of diversity, and so they are ideal for association studies. The kelp forest is also an ideal study area for this work because sampling done via SCUBA is fairly simple due to kelp’s shallow, near shore distribution. [[this is good, but you might want to reference Hopkins and its characteristics specifically right after the general description]]
Moreover, the kelp forest is an ecosystem that provides many important services to our society including provisioning services, regulating services, cultural services, and supporting services. For example, the kelp forest provides us with many products including kelp itself, whose canopy is often harvested, as well as food from the fisheries it supports. The kelp forest also attracts and supports many tourist activities that are important to local economies. [[this paragraph seems like a bit of a tangent]]
In order to effectively manage and protect this ecosystem and all the services the kelp forest system provides to humans, more information is needed about the intricacies of this complex system. This study aims to better understand the community by elucidating species-habitat and species-species associations by addressing several questions [[this is a bit awkward with the two “by”s]]: firstly, do species-habitat associations exist in the kelp forest; secondly, do these associations differ in relative strength for different species; thirdly, do species-species associations exist in the kelp forest, and do these associations differ in relative strength; finally, which association is more important in determining community composition, the physical or the biological?
We hypothesized that both species-habitat and species-species associations do exist in the kelp forest and that both these associations will vary in relative strengths. We also hypothesized that physical habitat associations will be more important in determining community composition because physical conditions are more often what determine a species fundamental niche where as the biological associations often shape the realized niche. [[good job introducing the questions and hypotheses]]
Methods [[16/18 – you’ve got all the essential bits, but I might reorder … notes below]]
In order to address these questions we conducted an observational field study at the Hopkins Marine Station in Pacific Grove, California (36°37'12.3"N 121°54'11.2"W) (Map Figure 1).
The Study System
Our research was conducted within the Hopkins State Marine Reserve, which was previously known as the Hopkins Marine Life Refuge (Jones 1985). Since 1985 the taking of fish, plants, and marine invertebrates has been prohibited, making this an ideal place to study natural occurring habitat associations (Jones 1985). Moreover, this area has a high diversity of benthic species, making it an ideal place for species-species association work. Macrocystispyrifera, or giant kelp, dominates the kelp forest, and granite is the dominant rock type in this region (Watanabe 1984). The study area also has a variety of substrate types, including bedrock, boulder, cobble, and sand, as well as complex and varying habitat structure in the area ranging from shallow to nearly vertical relief. Therefore, the area is also ideal for habitat association studies. [[good]]
Methodology Overview
To answer our primary questions we used a combination of two different sampling methods: swath sampling and uniform point contact methodology (UPC). Both were conducted via SCUBA at the same transects off the Hopkins main transect cable. We used UPC to sample sessile benthic species and their physical habitat, substrate and relief. UPC primarily samples those species for which individuals are not identifiable (either because they occur in aggregations or because of their growth form) or those that are not countable due to their small size and sheer number. UPC gives a percent cover for those species, as well as for substrate type and relief. The swath surveying was done for countable species over a 30 m by 2 m area. This method gives a density for the species counted (individuals per area). The resulting data sets were then linked [[at what scale?]]so that we could determine if the physical and biological habitat characteristics (from the UPC) were associated with the distribution of countable species. [good level of detail]]
Swath Methodology [[I would move these full descriptions down until after you’ve explained WHY you did things this way (i.e. introduced the questions and specific methodology for each]]
We collected data from 20 transects running on and offshore off the permanent transect cable at Hopkins marine station, sampling from the 90 meter mark to the 135 meter mark. All swaths were 30 m in length and 2 m wide with each diver within two man buddy teams sampling 1 m oneither side of the transect tape. Each 30 m transect was split into six 5 meter long sections; therefore, the data could be easily combined with the UPC data. No organisms under 2.5 cm in diameter were counted so that easily undercounted, small species did not bias the data. Highly abundant individuals were subsampled. If over 15 individuals were encountered within one 5 m section, the meter mark at which the 15th individual was encountered was noted and the remaining individuals of that species within that section were not counted. 37 species were counted in these surveys including 9 seastars, 3 anemones, 2 species of urchin, 3 species of sea cucumber, 7 species of crustacean, 8 species of molluscs, 2 algae, 1 sponge, 1 solitary tunicate, 1 hydrocoral (see table 1). Note that Macrocystiswere only counted if the stipeswere greater than 1 m in height, and Cystoseira were only counted if they had a minimum diameter of 6 cm.
UPC Methodology
In UPC data is collected at uniformly spaced points along a meter tape. At each of these points the species of the primary substrate holder directly under the meter tape is identified and recorded. This makes it possible to survey invertebrates that act as primary substrate holders, which are often impossible to count as individuals (see table 2).
To collect the UPC data, divers in two man teams conducted two dives at each 5 m increment, the first heading offshore (90°heading) and the second one heading inshore (270°heading). Each transect had a length of 30 m, which was further subdivided into 5m increments. These 5 m increments were sampled at every half-meter with the first diver sampling the first half (points 0, .5, 1, 1.5, and 2 m) and the second diver sampling the second half of each increment (points 2.5,3,3.5,4,4.5, and 5 m). At each point divers identified the substrate as being either bedrock (rock greater than or equal to 1m in extent), boulder (between 10 cm and 1 m), cobble (less than 10 cm), or sand. Additionally, divers classified the relief at each point by noting the maximum elevation change occurring within a 1 m by .5 m rectangle (.5 m to either side of the transect line and .25 m ahead and behind the point). A four point relief scale was used to classify relief: F-flat for relief between 0 and 10 cm, S-shallow from 10 cm to 1 m, M-moderate from 1 to 2 m, and H-high for relief greater than 2 m. The presence of a superlayer of drift algae and juvenile Laminariales was noted as well. Finally the species that occurred directly under each point were identified and recorded (or the type of inanimate substrate if there was no biotic organism present). This methodology was repeated for all 5 m increments along each transect.
Table 2: UPC species list with codes used.
Cover / CodeInanimate / Bare rock / BARROC
Bare sand / BARSAN
Shell Debris / SHELL
Sediment/mud / SED/MUD
Dead Kelp Holdfast (any) / DEADHOLD
Red Algae / BRANCH-flat branching / BRANCH
LEAF- blade, unbranched / LEAFY
BUSHY-cylindrical branches / BUSHY
LACY-filamentous/dense / LACY
ENCRUSTING RED / ENCRED
TURF - red turf - < 2 cm / TURF
Coralline / Crustose coralline algae / CRUCOR
Articulated coralline algae / ARTCOR
Brown Algae / Cystoseiraosmundacea / CYSOSM
Dictyoneurumcalifornicum / DICCAL
Egregiamenziesii / EGRMEN
Desmarestia spp. / DESSPP
Macrocystis holdfast (Live) / MACHOLD
Laminariales Holdfast (Live) / LAMHOLD
Dictyotales (DictyotaDictyopteris) / DICTYOTALES
Tube Worms / Tubeworm - Other Solitary / TUBEWORM
Diopatraornata / Chaetopterus / DIOCHA
Phragmatopoma / PHRCAL
Dodecaceria spp. / DODFEW
Snails / Serpulorbissquamigerus / SERSQU
Petaloconchusmontereyensis / PETMON
Cnidarians / Corynactiscalifornica / CORCAL
Cup Corals / CUP
Other anemone / ANEM
Hydroids / HYDROID
Stylastercalif. (Calif Hydrocoral) / STYCAL
Tunicates / Colonial tunicate / COLTUN
Solitary tunicate / SOLTUN
Other / Scallop / SCALLOP
Embedded Cucumber / CUCSPP
Barnacle / BARN
Bryozoan / BRYO
Sponge / SPONGE
Mussel / MUSSEL
Linking Swath Data to UPC Data
In order to examine possible associations between the countable swath species and the UPC data, all transects were split into six 5 m long sections. For each section there was one set of species counts from the swath and 10 UPC points spaced half a meter apart (see Figure 2). For each section the following percentages were calculated: 1) percent of the substrate that is bedrock, boulder, cobble, or sand 2) percent of area that has flat, shallow, moderate, or high relief 3) percent of the section area made up by each UPC species sampled. These percentages were then linked to the swath species counts for each section. [[nice, but explain that they’re linked at a 10 sq m spatial scale]]
Habitat Associations
In order to determine if physical habitat characteristics, in this case substrate type and relief, determined community composition (for countable species) a dissimilarity matrix was calculated using Euclidean values for the different sections for each transect. This matrix was then linked to another dissimilarity matrix calculated using Bray-Curtis values for the swath species using a Spearman correlation. This was done to determine if swath species dissimilarity was related to habitat dissimilarity. From the results, the swath species dissimilarity could be plotted against the habitat dissimilarity to bring out any relationship between them.An analysis of variance table was then constructed between all swath species and substrate type and relief. A 95% confidence interval was used to evaluate the significance of the resulting relationship.
Habitat Association Strengths
To test the hypothesis that differences exist in the strengths and signs (positive or negative) we performed a series of correlations between swath species and their substrates and relief types. All correlations greater than 0.1 and less than -0.1 were significant using a 95% confidence interval and were therefore considered strong associations. Correlations less than .1 and greater than -.1 were considered weak associations and not used in the analyses.
Four pairs of species were selected to be examined in detail: Macrocystispyrifera and Cystoseiraosmundacea, Balanusnubilus and Styelamontereyensis, Cryptochitonstelleri and Lithopomagibberosa, and Patiriaminiata and Pisastergiganteus. These species were selected because they were abundant and showed interesting and related habitat associations.
Species-Species Associations
In order to test the hypothesis that biological habitat characteristics, in this case UPC species, influenced the community composition for countable species, a Bray-Curtis dissimilarity matrix was calculated for UPC species and compared to the swath species dissimilarity matrix again using the Spearman rank correlation. Again the results were used to plot the swath species dissimilarity against the UPC species dissimilarity. A slope of 0 would show that there is no relationship between the two. An analysis of variance table was constructed to determine whether the associations were significant or not. 95% confidence intervals were used to determine significance.
Species-Species Association Strengths
To test the hypothesis that species-species associations varied in relative strength and sign, another series of correlations were done this time between UPC species and swath species.95% confidence intervals determined that all correlations greater than 0.1 and less than -0.1 were significant. Only significant correlations were included in the results.
Relative Importance of Physical versus Biological Associations
In order to determine whether physical or biological attributes had a larger impact on community composition a variance component analysis was run on the data. This analysis provided us with the relative contribution of each attribute to the resulting community composition. [[I envisioned this going further up, before the detailed methods and including the field sampling methods that enabled you to do these analyses]]
Results [[12/16 – you’ve got a lot more details than you need here, and it’s drowning out the important points you’re making. Good job being specific about whether the results support or reject the hypotheses. Also, I’d prefer if you put the figures all in one section rather than interspersing in the text as it makes it difficult to follow]]
The kelp forest community at the Hopkins Marine Station was highly diverse in both physical habitats and biological species. Based on this observational field study, we determined that specific physical habitat associations are most important in determining community composition. Nevertheless, species-species associations do also exist with a wide variety of strengths, and they are still important to the community’s structure, especially for understanding resource partitioning that maintains the diversity of communities such as this kelp forest.
Habitat Associations
The null hypothesis for habitat associations was rejected (P-value <.0.000001) (see table 3). Habitat associations were found to exist in the community, and the physical attributes of substrate and relief did influence community composition (see figure 3).
Habitat Association Strengths
For the eight species that were examined in more detail, habitat associations varied in strength and sign with some species showing positive correlations to some substrates and negative correlations to others (see figure 4).
Figure 4: For eight selected species the correlations to each substrate type (bedrock: rock greater than or equal to 1m in extent, boulder: between 10 cm and 1 m, cobble: less than 10 cm, or sand) and to each relief type (flat for relief between 0 and 10 cm, shallow from 10 cm to 1 m, moderate from 1 to 2 m, and high for relief greater than 2 m). Correlations greater than 0.0 are positive, meaning the species is found on that physical habitat more than would be expected by chance; correlations less than 0.0 are negative, meaning the species is found on that physical habitat less frequently than would be expected by chance. The dashed lines represent significant positive and negative correlation values (with 95% confidence intervals).
Both algal species examined here, Cystoseiraosmundaceaand Macrocystispyrifera, showed significantly positive correlations with bedrock and lower relief; both showed significantly negative correlations with small clast size substrates. Cystoseira showed negative associations with sand and moderate relief and had positive associations with shallow relief and bedrock. Macrocystis showed positive correlations to bedrock and shallow relief as well and was negatively correlated with cobble and flat relief.