Cite the Literature Correctly BEFORE the Period

Total= 4+18+38+27+

Lit Cited= 2

Clifton Herrmann

10/30/11

[4] Quantitative Comparison and Variance Analysis of Two California Coast Kelp Forest Ecosystems

Abstract:[4,2,3,3, 2,2, 2]= 18Comparative analysis of species abundance at two similar kelp forest locations creates an opportunity to explore sources of variability in data collection and analysis. Ecosystems were sampled for species abundance simultaneously on two separate days. Interaction between site and day variation is discussed, and understanding these values helps us pinpoint the sources of known variation. Our objective was to count individuals on the selected species list along two 30m by 2m transects at each site. Mean abundance data of all selected species were analyzed for their degree of difference between both site and day with respect to all three taxa, including comparative analysis among species within each taxon. Variance component analysis illustrates the consequences of inadequate sampling designs [???] and how they affect our interpretations of differences in abundance at our chosen spatial scale. Overall species abundance was found to be different at the two locations, and day and interaction effects were found to be significant in fish abundance data.

Introduction [4,3,4,4,3,3,4,2,3,4,4] = 38

[cite the literature correctly BEFORE the period]

Comparative analysis of species abundance at two similar kelp forest locations creates an opportunity to explore sources of variability in data collection and analysis. Variance in data can either show differences that exemplify the question(s) you had hoped to explore, or it can dilute reduce the integrity precision and accuracy of your results . (Morrisey et al. 1992). The latter form of variation is categorized into two groups: sources of variation that are known but still negatively affect the results, and those from unknown sources. Known variability sources can be statistically analyzed to understand exactly how they affect our data, whereas un-sourced variability is known as “error.” Two different kelp forest ecosystems were sampled for species abundance simultaneously on two separate days. (Providing four data sets.) Our goal in conducting this comparative study was to discuss [??] the three forms of statistical variability with respect to species abundance at our two sites. Mean species abundance with respect to day and site will show us potential differences, but it is important to include a statistical variance analysis to more accurately account for confounding factors.(Alkawri et al. 2009, Morrisey et al. 1992). In designing this study, it has been important to carefully consider the spatial and temporal scales chosen for sampling. Our two locations were close to one another along the California coast, having known differences that contribute to known or expected variability. Temporal considerations were fairly short term, (only two sample dates) but the assumption is that since they were during the same week, variation created by the day effect should be minimal. It is important to remember, however, that a two-day sample is not representative of the full range of temporal variation, but rather season and time-of-day specific. (Lewis, 1978)

Conducing a simple species abundance study at two sites over two days has allowed us to use multivariable multivariate statistical analysis in discussing identifying sources of variation from more than one source at once. Interaction between site and day variation is discussedassessed, and understanding these values helps us pinpoint the sources of known variation. Using this information, we can find out how the data patterns we identify are affected by unknown sources of variation, or error. Certain species were selected for this study to provide a clear overall understanding of the two kelp forest ecosystems, as well as provide a platform for comparative discussions. (Table 1)

Fish / Invertebrates / Algae
Damalichthys vacca
Embiotoca lateralis
Embiotoca jacksoni
Hexagrammos decagrammus
Oxylebius pictus
Sebastes atrovirens
Sebastes carnatus
Sebastes chrysomelas
Sebastes mystinus / Balanophyllia elegans
Callistoma ligatum
Haliotis rufescens
Loxorhynchus grandis
Pachycerianthus fimbriatus
Patiria miniata
Pisaster brevispinus
Pisaster giganteus
Pychnopodia helianthoides
Strongylocentrotus franciscanus
Tethya aurantia
Urticina lofotensis
Urticina piscivora / Chondracanthus corymbifera
Cystoseira osmundacea Dictyneurum californicum
Dictyneuropsis reticulata
Eisenia arborea
Macrocystis pyrifera
Pterygophora californica

Table 1: Selected species, sampled at both sites on both days, by taxa.

The specific questions we plan to address in this study revolve around the broader concept of comparative analysis between two similar, yet different, kelp forest locations, and how results may vary due to the influence from a myriad of factors. This comparison raised three main questions. 1) Was there a significant difference in species abundance between our two sites, and do these differences vary by taxa? 2) Was there a significant difference in species abundance by sample day, and do these differences vary by taxa? Does the number of sampling days chosen in your scientific design have an effect on your results, and can this vary by taxa? 3) Does the statistical analysis of both site and day create an interaction effect, and can this effect vary by taxa? We test hHypotheses that are based on these questions. Our null hypothesis states that there is no difference between the two sites based on site, day, or interaction effects. Significant values from our multidimensional statistical analysis will be used to reject this null hypothesis.

Methods: [2,0, 0, 3, 2, 2, 2, 0, 0, 0, 4, 1, 3, 3, 3, 2,] 27

General approach??

The study system

Quantitative sampling of selected species was conducted at two sites on October 11th and 13th, 2011 at approximately 9:00am. Our class of 29 students was divided in half, and a group was sent to each site on the first day, switching locations on the second. Our objective was to count individuals on the selected species list along two 30m by 2m transects at each site. Hopkins Marine Station and Marine Life Refuge, owned and operated by Stanford University, is located in Monterey, California, at the southern horn of the Monterey Bay. This kelp forest location has been legally protected since 1931, making it the second oldest marine refuge in the state, and has been chosen as our fist site. This forest is geographically protected from high wave energy, which allows for the growth of a dense Macrocystis pyrifera canopy and subsequent associated fauna. A permanent transect cable has been placed at this site for the large amount of research taking place, which we used in our sampling. Our 30m transects were off and onshore using this cable as the focal point. Point Lobos State Natural Reserve is the location of our second site. Located 10 miles south of Hopkins along CA Highway1, Pt. Lobos is geographically exposed to higher wave energy, as it is not protected inside the Monterey Bay. Data were taken at Whaler’s Cove, just outside of The Pit. A temporary main transect was laid on each dive day, from which dive pairs measured two 30m transects. After decades of destruction from resource harvesting from abalone to coal, Pt. Lobos was designated a state park in 1933 in an attempt to secure and protect its natural beauty. [sexy descriptions!!!]]

Maps of Hopkins (left) including location of permanent transect, and Pt. Lobos (right) with Whalers Cove in the upper right corner.

Data collection:

Data were taken using SCUBA along a main transect for all three taxa at the same time. Fish were sampled while dive pairs reeled out their 30m line as to avoid bias from the effect of their presence. Invertebrates and algae were then counted either at once or one at a time depending on diver team preference before the meter tape was rolled in. This was completed for two transects or until dive pairs were low enough on air to turn around. Dives were completed on a single tank.

These simple sampling methods were designed to provide site-specific species abundance density data for comparative analysis between the two locations. The two locations were chosen based on their known similarities and differences, as they are common areas for research divers. We will used the known differences between sites to assess our accuracy in variance of abundance analysis.[how could you do that??] The two sites will bewere statistically compared based on species abundance to look for overall difference between sites, difference between sites by taxa, difference between sampling days, difference between days by taxa, and difference due to the interaction effect of the two variance sources both overall and by taxa.

[[Where are your hypotheses?????????? How did you use the data and analyses to test your hypotheses?????]]

Results:

Mean abundance data of all selected species were analyzed for their degree of similarity with respect to all three taxa, as well as among species within each taxon.To do this, we used multidimensional scaling (MDS) plots to visually represent the degree of similarity between points with respect to both day and site effects. Two-dimensional stress was below significance for each analysis, meaning that the flattening effect of our 3D model does not significantly affect our ability to analyze the points within each figure. Multidimensional PERMANOVA analysis was used to gauge the strength of effects from site, day, and the interaction effect between the two. (Table 2) This analysis was conducted for overall species abundance, as well as for each taxon. A P-value less than 0.05 denotes significant effect from a given source. Variance component analysis was then conducted to verify and reinforce conclusions of variance sources by site, day, and the interaction between the two. Finally, we will test the power index of each taxa with respect to numbers of transects completed. This type of analysis will shed light on whether or not our sampling was thorough enough to accurately create a fair and level comparative platform for species abundance. [ALL of this belongs up in the Methods section!!! ALSO… it is all PAST tense…. Not “we will do this or that”.]]

Depth Effect

The effect of depth on our species abundance data can be excluded, as there is no significant variation between sites. (Figure 1) Upper and lower ranges are nearly identical, with only slight difference in the shallow portion of the median quartile. Depth is hereby excluded as a possible source of variation.

Overall Species Abundance

Multidimensional scaling (MDS) of all species shows an overall site effect seen by major spatial distinction between overall species abundances at Pt. Lobos and Hopkins. (Figure 2) Multivariable analysis reveals a significant site effect (PERMANOVA, P-value=0.001) Day and interaction effects do not exhibit significant correlation.

By Taxa: Algae

Figure 3: MDS plot (left) for algae data shows level of similarity as a function of distance between points. Numbers 1 and 2 denote sample day, and colors represent location. Green is Hopkins, and blue represents Lobos data

Figure 4: Species abundance of algae (right) at each site for day 1 and 2. Bars show mean abundance from all algae data.

Algal MDS analysis shows a strong site effect on mean species abundance, represented by distinct regions denoting site. (Figure 3) Mean algal abundance by species shows similarities at sites for each day. (Figure 4) We see high numbers of Chondracanthus corymbifera and Cystoseira osmundacea at Hopkins with slight variation by day, but relatively low levels of these species at Pt. Lobos. Abundances of Pterygophora californica and Eisenia arboreta are high for both sampling days at Pt. Lobos, while nearly nonexistent at Hopkins. Multivariable analysis reveals significant site effect for algal abundance. (PERMANOVA, P-value=0.001) There is no significant correlation for day or interaction effects in algal abundance data.

By Taxa: Invertebrates

Figure 5: MDS plot (left) for invertebrate data shows level of similarity as a function of distance between points. Numbers 1 and 2 denote sample day, and colors represent location. Green is Hopkins, and blue represents Lobos data

Figure 6: Species abundance of invertebrates (right) at each site for day 1 and 2. Bars show mean abundance from all invertebrate data.

MDS analysis of invertebrate species abundance shows a slightly weaker visual distinction between data from the two sites than seen for algae, with some mixing of the two colors. (Figure 5) Regions are still relatively distinct, however, and exhibit a statistically significant site effect. (PERMANOVA, P-value=0.001) We still see no significant day or interaction effect in our multivariable analysis for this taxon. Mean invertebrate abundance data by species exhibit similarities at sites for each day. (Figure 6) Hopkins had high numbers of Balanophyllia elegans and Patiria miniata. While these species are also the most abundant species at Pt. Lobos, their mean values are lower than those at Hopkins. Pachycerianthus fimbriatus and Callistoma ligatum were found to be relatively abundant at Hopkins, while very few C. ligatum and absolutely no P. fimbriatus were noted at Pt. Lobos.

By Taxa: Fish

Figure 7: MDS plot (left) for fish data shows level of similarity as a function of distance between points. Numbers 1 and 2 denote sample day, and colors represent location. Green is Hopkins, and blue represents Lobos data

Figure 8: Species abundance of fish (right) at each site for day 1 and 2. Bars show mean abundance from all fish data.

For fish, site was not seen as the main source of influence on species abundance. The MDS analysis shows no clear distinction between data points from the same site, but rather a combination of relationships between day and site effects. (Figure 7) PERMANOVA analysis reveals no significant site effect, but effect from day, (PERMANOVA, P-value=0.043) and the interaction effect of the two sources (PERMANOVA, Si*Sa, P-value=0.015) were both significant. Actual mean species abundance data show a generally higher abundance trend for most species on day 1 than 2, (Figure 8) enhanced by the lack of consistency from site effect. The only fish species remaining relatively consistent at sites per day were Damalichthys vacca and Sebasted chrysomelas at Hopkins.

Variance Component Analysis: All Taxa

Figure 9: Variance component analysis showing variance from either day or site sources as a percentage of the whole (total variation) separated by taxa.

Variance component analysis of abundance data by taxa with respect to site and day reinforces our multidimensional statistical analysis. (Figure 9) Here we see that variance in algal and invertebrate abundances are completely controlled by the effect of site. Source of variance for fish data, however, rests more heavily on the effect of sample day than site.

Power Indices

Power index analysis shows us the strength of our data based on the number of transects used. As the line horizontally asymptotes, you have reached the highest possible power, yielding diminishing return per unit number of transects after that point. Here we see that Algae has reached a strong power and leveled out in relatively few transects. Invertebrate data power is fairly level as well, but still increasing slightly. Power of our fish data has yet to reach any type of asymptote. Any statistical power greater than two is generally considered credible.

Table 2: PERMANOVA P-values as a result of multivariate analysis. P-value of less than 0.05 denotes statistically significant difference based on the given source

All Species / P-values
Site
Day
Interaction / 0.001
0.382
0.375
Algae
Site
Day
Interaction / 0.001
0.724
0.926
Invertebrates
Site
Day
Interaction / 0.001
0.505
0.569
Fish
Site
Day
Interaction / 0.319
0.043
0.015

Discussion:

The three main questions presented in the introduction of this paper laid a general framework for the actual hypotheses we set out to answer. Each hypothesis is broken into two pieces, and all six are as follows:

H1: There is a difference in species composition between Hopkins and Point Lobos.

H1a: Difference in species composition between Hopkins and Pt Lobos varies by taxa.

H2: There is a difference in species composition between days.

H2a: Difference in species composition between days varies by taxa.

H3: Both Site and sampling day affect species composition (interaction effect.)

H3a: the interaction between site and sampling day varies by taxa

H0: There is no difference in species abundance between Hopkins and Pt. Lobos.

Given that our overall MDS plot of all species shows a significant difference due to site effect, and not day or interaction effects, we can discuss potential reasons for the difference in species composition between Hopkins and Lobos. Wave energy due to level of protection from rocks in a given kelp forest has a strong effect on algal species composition, and subsequently the composition of associated fauna. Forests subject to high wave energy are likely to have lower levels of Macrocystis pyrifera, since it experiences a high level of drag from its many blades, and will be torn from its holdfast in high energy hydrodynamic environments. This morphological strategy, however, maximizes its ability to photosynthesize and grow very quickly. For this reason, M. pyrifera is quickly replenished after major disturbance events such as storms or destructive human influence (i.e. harvesting.) Much lower levels of M. pyrifera were noted at Lobos than Hopkins due to its lower level of protection from wave energy. Lobos, instead of M. pyrifera, exhibited high concentrations of understory kelp species such as P. californica and E. arborea. These deeper, stronger species of macroalgae can more easily survive in areas with higher wave energy such as Lobos. This fundamental difference of kelp forest structure creates the statistically significant difference we see between the two sites with respect to both algae and invertebrate species composition. For fish, however, there is no significant difference between the two sites. Even though the two sites are statistically different overall, this result is not consistent for all taxa.