Qualitative Sampling Methods and Their Effectiveness In

Clifton Herrmann

10/11/11

UCSC Kelp Forest Ecology

Qualitative Sampling Methods and their Effectiveness in

Describing Distribution and Abundance of Selected Species Within

A California Cold Water Kelp Forest Ecosystem

Introduction:

Species abundance and distribution within a given ecosystem are important components to understand in order to gain insight on its general biogeochemical interactions and how they may influence future system structure. Constant competition for coexistence creates a delicate balance of relative abundance and distribution, driven by the strength of interactions between species that rely on various shared resources for survival. (Sala et al., 2002) The main focus of this study is to determine relative abundance and distribution of selected species within a California kelp forest ecosystem. Determining site-specific descriptions of the distribution and abundance of species allows ecologists to draw conclusions about how changes in resource availability may influence current ecosystem structures. Tracking these changes may shed light on which resources have the strongest effect on the ecosystem as a whole, and allows ecologists to further understand trophic food web health and structure. (Graham, 2004) Spatial and temporal scales chosen within a study also have an influence on the way in which a given ecosystem appears to be changing. Understanding the various levels of influence that ranging disturbances can have on your ecosystem helps paint a more concise picture of the biogeochemical interactions taking place, and how these disturbances affect the functional ecosystem unit. (Edwards, 2004)

Sampling methods can be specifically tailored to best depict desired data, but are more broadly classified as either quantitative or qualitative descriptions. There are costs and benefits associated with the use of each of these two descriptive methods. Quantitative data are generally collected when scientists and ecologists want to use hard numbers to support predicted outcomes. Data are unbiased and generally consistent, as they follow strict rules that govern how, when and where they are collected. Statistical analysis tends to be straightforward, and quantitative data can generally be easily graphically depicted. Collection of data can be outsourced to research volunteers more easily due to the instructional nature of quantitative studies, and numerical data retain their relevance even over long periods of time. Quantitative studies are also less site-specific as they generally work to describe relationships within a larger system, rather than the system itself. This allows for the logical application of quantitative results to other similar ecosystems. This can become a problem when your study requires a thorough understanding of the system as a whole. Qualitative data were collected for this study in order to gain that complete understanding of this kelp forest ecosystem. However, the most prominent cost to using qualitative methods of data collection is the high level of relativity. Personal opinion plays a major role in perceived abundance or absence. Definition of “abundant” varies with relative expected quantities of a given species. Seeing fiftyBalanophyllia elegans (orange cup coral) could seem extremely abundant to one, whereas another person may perceive that as only a small population since they were expecting to see many. For this reason, qualitative data collection requires a very strong understanding of the expected distribution and abundance of species within the ecosystem before conducting the research. Using consistent buddy pairs on dives and repeating the sample would allow ecologists to accurately track ecosystem changes based on abundance and distribution shifts.

Statistical analysis of variance components were plotted graphically in order to provide an understanding of how accurate our qualitative study describes this specific kelp forest habitat. Variance components were modeled against the source of the data. These three sources were depth, either the deep or shallow side of the permanent transect, buddy, data variance between dive partners, and meter mark along the transect. These variance sources provide insight to high percentages of difference in data collected between individuals within buddy pairs. Influencing factors could include diver familiarity with the specific habitat, and the relative visibility of the selected species. Divers must have a clear understanding and knowledge of local selected species, and be able to identify with ease in order to accurately describe the habitat. Variance may also be enhanced by the fact that some species are much more visible than others. Macrocystis pyrifera, for example, is much more visible and easily noted than Strongylocentrotus franciscanushiding in small cracks and crevices. Other species that proved to be highly visible include Patiria miniata, Sebastes chrysomelas, and Urticina piscivora. These species showed low levels of variance due to their easily identifiable features. Species showing high variance include many of the selected fish species, and there may have been some confusion between the structural similarities of Dictyoneurum sp. vs. Dictyoneuropsis sp. With continued sampling and consistent dive partners, qualitative methods have great potential to accurately track changes in site-specific ecosystems.

Methods:

Qualitative data were taken by fourteen dive buddy pairs using SCUBA. Descriptive data of species distribution and abundance will be used to provide an overall understanding of the habitat’s systems and structures. Each pair sampled 30m transects both on and offshore from the permanent transect cable at Hopkins Marine Research Sanctuary in September of 2011. Data were taken for each “leg” of the transect, meaning that four different sets were recorded. Selected species were rated on a 1-5 scale, 1 denoting a complete absence, and 5 reflecting high abundance. Species recorded are as follows.

Results:

Variance component analysis (VCA) was run[jf1] on the data. We wanted to show both species abundance and distribution, as well as provide a thorough analysis of the data set in terms of error and relativity when using qualitative sampling methods. Six visual interpretations were generated: variance component analysis in general and by taxa, mean abundance for all species, relative difference between buddies, percent disagreement between buddies using presence/absence data, and relative difference between buddies as a function of mean abundance. These figures present a thorough analysis of both the habitat and the effectiveness of our sampling method[jf2].

VCA shows percentage of variance contributed by each source as a portion of the total (100%.) The significance of the first figurelies in the large percentage of variance found within buddy pairs- nearly 40% of the total variance. This suggests potential issues regarding the quality and consistency of our qualitative data. The assumption would be that buddies on the same transect would be seeing similar

the validity of high meter source variation. For fish, however, low buddy variation suggests that certain fish were found heavily along some meter marks and scarcely along others. This could be caused by territoriality, sampling error due to the presence of 30 divers changing fish behavior, or even resource availability gradients that focus higher abundances in certain areas of the forest. The invertebrates portion of the figure tells a very different story. With 0% of the variation coming from distance along the permanent transect, it is logical to conclude that invertebrate assemblage is much more depth dependent than anything else. We do experience strong buddy variance, however. This could be attributed to the very hidden nature of many invertebrates. The use of qualitative sampling methods runs the risk of sampling imprecision due to the divers’ level of attention to detail. Qualitatively describing fish and invertebrates simultaneously is difficult due to the divers general position in the water column being different when looking for these two types of biota. For this specific data set, we see an overall high level of variation within buddy pairs. For future qualitative studies, the data would be more consistent if all dive pairs had a thorough briefing before the dive on what they will define as abundant for each species. Setting guidelines and being thoroughly associated with the particular ecosystem being studied are important parts of collecting qualitative data.

The mean abundance figure relates[jf3] directly to the data rather than the degree of difference between points within the data. This figure is separated by taxa and describes the overall abundance of the selected species. From this information, we are able to draw conclusions about species interactions and coexistence as well as resource partitioning and use.

We see that within the fish taxon, abundance is fairly even between the different species. These data show that fish abundance and distribution is probably quite stable. Resources are being evenly distributed, and each fish species is keeping the others in check. Had we seen an absence of a particular species, we may have also seen a resulting peak in another. This would be because a dominant species was killed off and the suppressed species moved in to the previously occupied area and allowed to proliferate. (Holbrook, 1995) Algal abundance shows a higher level of dominance. Macrocystis pyrifera and Cystoseira osmundacea are key species within the kelp forest. M. pyrifera provides the canopy and vertical substrate while C. osmundacea is a commonly abundant as an understory kelp. A possible explanation for the higher abundance of C. osmundaceais seasonality. During the fall, C. osmundacea produces reproductive structures in its macroscopic stage that extend upward with large pneumatocysts. The higher abundance may be a result of its enhanced visibility during this stage of its life history. Other understory kelp species appear consistent, and the lower level of Phyllospadix spp. are most likely due to their distribution within the upper subtidal, making it more common on the onshore side than offshore. Invertebrate species appear to vary in abundance based on their specific life history strategies and levels of resource consumption. Patiria miniata (bat star) was the most abundant invertebrate species found. Bat stars are highly common in the sandy subtidal environment and are active detritovores. They are commonly and heavily found at the lowest level possible, as that is where drift material tends to settle. They are highly reproductive and resilient. Haliotis rufescens (red abalone), for example, was the least abundant invertebrate. While the red abalone was recorded the least, its diet and lifestyle are very similar to that of a bat star. One main difference is where they reside in the benthos. H. rufescens generally wedge deep within cracks and crevices for better protection. They were probably present in lesser quantities, but there are likely many red abalone that were not counted.

This figure shows the relative difference between buddies broken into data by taxa. The main significance of these figures is to visually show that there was some level of sampling bias taking place during the qualitative study. Taking algae for example, higher variation was seen between buddies for the Dictyoneurum sp. vs the Dictyoneuropsis sp. This correlates with their very similar morphological features, making them harder to decipher, resulting in higher variance between buddies. M. pyrifera, however, represents the lowest level of variation between partners due to its extremely easily identifiable features. Having a very site-specifically trained research team would reduce the variance seen in this figure.

By[jf4] removing the factor of relative quantity, we are able to view the data with less bias created by difficulty with identification of morphological characteristics. The data are more accurate when you just look at whether or not a specific species was present at all. This presence/absence analysis can be helpful in painting the general picture of what species are out there and visible. The following figure shows this presence/absence data as a function of percent difference between buddies. Unfortunately, this type of analysis still shows the high level of variation between dive partners experiencing similar locations. It still does, however, exemplify previous statements that the visibility and ease of identification of certain species have large effects on buddy agreement and whether or not the presence of a species is actually noted.

Again, looking at the algae figure, it is obvious that all buddy pairs agreed on the presence or absence of M. pyrifera. Dictyoneuropsissp. and Dictyoneurum sp. show large degrees of disagreement, most likely due to their easily confused morphology.

The final figure plots relative difference between buddies as a function of mean abundance. This[jf5] figure is particularly valuable because it shows how our huge

specific experience, presence/absence data collection reduces the influence of those factors. Despite the high variation between buddies in abundance and distribution data, mean abundance figures were still valuable in providing a clear description of the general structure within this specific kelp forest ecosystem. With consistent practice and continual data collection, qualitative sampling methods can provide viable descriptions of any given site. Using VCA is very important, allowing ecologists to tell exactly how accurate your data are, and how they should account for any possible biases. Qualitative methods seem to be most useful when creating a description of species distribution and abundance over temporal scales rather than having just one sampling period. Consistently repeating the sample allows ecologists to create a comparative and accurate description of trends in abundance over time. Tracking these changes is ecologically important because better understanding relationships, processes and systems within many different ecosystems can provide a strong overall understanding of the many influencing factors that play a role in species and ecosystem structure that can extend beyond the system itself.

References:

Enric Sala, Michael H. Graham: Community-wide Distribution of preditor-prey interaction strength in Kelp ForestsPNAS, March 19, 2002. Vol 99. No. 6

Michael H. Graham: Effects of Local Deforestation on the diversity and Structure of southern California Giant Kelp Forest Food Webs Ecosystems, 2004. Vol 7 no. 4

Matthew S. Edwards: Estimating scale-dependency in disturbance impacts: El Niños and giant kelp forests in the northeast PacificEcosystem Ecology, 2004

Sally J. Holbrook, Russel J. Schmitt: Compensation in resource use by foragers released from interspecific competition Journal of Experimental Marine Biology and Ecology, 1995 Vol. 185 no. 2

General Notes: Figures always require legends. Figures are standalone and a reader should have all the information required to interpret them. Refer to guidelines.

Good job, you have obviously put a lot of though into the discussion. The key now is to practice making it concise. Figure out your main points and develop them. Leave out extraneous oinformation that does not them.

[jf1]Avoid passive tense

[jf2]No need to restate the methods, this whole paragraph is redundant. Start with the findings of your study

[jf3]Do not refer to the figures this way, in the discussion you rarelyrefer to figures at all. You have already presented them in the results.

[jf4]Again a lot of this is for the methods and results and does not need to be repeated in the discussion

[jf5]Also, use logical linkages…why did you analyze it this way? Why were you interested in that question? That is how the paragraph should start.