Qualitative Survey of the Hopkins Reef
Pike Spector (and the KFE Class of Fall 2011)
Title: 4 pts
Abstract [0, 0, 0, 0, 2, 0] 2 pts
The methodology and practice behind scientific sampling has a tendency to be readily involved in [I suspect you could say this with fewer word] usually requires the acquisition collection of raw quantitative data in the form of quantitative, numeric counts. We sought to characterize the distribution and abundance of common kelp forest species using a different approach: qualitative sampling. [Why???] Overall the question of abundance and distribution of common species by means of a qualitative survey yielded mixed results. Some species proved to be good candidates for a qualitative survey while others did not [based on what?]. However, through a combination of qualitative and quantitative survey species richness and abundance can be accurately, and precisely, sampled and recorded (Drury et al 2010). [not related to your work and don’t cite the literature in an Abstract]
Introduction [2, 4, 2, 2, 2, 2, 4, 4, 2, 4, 2] 30 pts
[You need to revisit and reduce sentences to achieve conciseness]
The methodology methods and practice behind scientific sampling has a tendency to be readily involved in the acquisition of raw data in the form of quantitative, numeric counts. [reduce this sentence to half its length] Here, instead, we assess effectiveness qualitative sampling (Mackey et al 1984). [why??] Through this method we want to determine the distribution and abundance of common kelp forest species. [why??]
Ecology as a whole is interested in examining interactions between organisms and their environment, and also between organisms of similar and different taxons taxa and species. By examining where species are and their abundance we can evaluate not only the overall health of the ecosystem but also the impacts that other factors and systems may have on the a system of interest (Scheiner et al 2011).
In order to asses the abundance and distribution of species we need a surveying tactic method that is reliable, accurate, feasible and, most importantly, unbiased (Eberhardt and Thomas 1991). The general question here is: can qualitative sampling via a qualitative means still yield reliable and applicable (i.e. precise and repeatable) data that can be accurately used? The most common approach to sampling in a kelp forest is by means of analytical quantitative surveying. This is the tradition because the school of thought behind such a survey is that it will yield reliable, unbiased results and allows for a repeatable statistical analysis (Eberhardt and Thomas 1991). However, this kind of surveying does not exclude bias indefinitely and employs a lot of time in the field to collect such data. By a qualitative survey data can be collected in a much more rapid fashion, but it also invites sampling bias and inaccuracy. Arguably quantitative sampling in conjunctionure with qualitative sampling could in fact yield an unbiased, precise and accurate description of a system overall (Drury et al 2010).
Even though the difference between qualitative and quantitative surveying have been exaggerated (Lund, 2005), in an analysis of cost between these two sampling techniques the most obvious discrepancy derives from an individual’s conception of the ecosystem. (what does this mean??) Doing an analytical fish count yields raw data, but assessing whether or not a diver personally thinks a particular fish species is abundant or not invites a lot of room for bias and a decline in precision. And what we see from our results indicates just that; buddy pairs did differ in their perception of species abundance. It does seem reasonable that a diver who has had extensive experience diving in the Monterey Bay would have a different idea of what to expect as opposed to a diver who has spent of their time diving in Southern California for example. Also, differences between diver’s experience overall may show in the results, especially if the diver has not had as much experience surveying in general. Through this survey we are able pose another question: can we acquire information regarding the abundance and distribution of kelp forest species yielding some indication of which species would be well suited to further qualitative sampling and which would not (Scheiner et al 2011).
All surveying was conducted in the field at the Hokpins Marine Life Refuge in the Monterey Bay, California. The primary ecosystem engineer in the kelp forests of the Monterey Bay is the giant kelp, Macrocysts pyrifera, which provides an extensive habitat for a multitude of invertebrates, algae and fish species. Using SCUBA we tested the effectiveness of qualitative sampling by the categorization of abundance of common species found in the kelp forest of three taxa: algae, invertebrates and fish.
Methods [2, 0, 0, 4, 0, 0, 2, 0, 0, 0, 0, 0, 4, 0, 2, 2, 0] 16 pts
[See writing guidelines for missing pieces and use of subheadings]
Survey SiteSystem
[write in first person!] The We conducted the survey was conducted at Stanford University’s the Hopkins Marine Station (affiliated with Stanford University, Palo Alto, California) located in on the Monterey Bay, California. [where?] The substrate at southern end of the Monterey Bay is primary composed of granitic rocky reef with sandy patches and a perennial kelp forest ecosystem that goes through annual cycles of impact, decline and re-growth. This site was chosen because of its ease of access for the divers and the abundant kelp forest ecosystem under the protectorate of the Hopkins Marine Life Refuge.
[now describe the species we sampled and why we chose them]
Sampling
Two thirty 30 meter transects where run off of the main transect running, from North to South (0º and 180º respectively), in approximately 40 feet of water; all surveys were conducted using SCUBA. Fourteen buddy pairs ran first a “deep” transect (heading into the bay at approximately 270º) and a “shallow” transect running towards shore (approximately 90º). [why these different depths?] In order to assess the differences between buddy pairs, and to test the validity of qualitative sampling, no prior discussion of sampling methods were to be discussed between the buddy pairs. Divers were given a list of common species one would expect to see in the kelp forest and told to indicate on their slate the relative abundance ranked from 1-5 (one being absent, five being abundant); species were grouped intoincluded three taxa:how many Algaealgae, Fish fish and Invertebratesinvertebrates. Divers were instructed to sample in, on and under rocks as well as to include any and all fish seen given by their visibility in the water column.
Analyses????
Results
Variance of Components
In order to test the effectiveness of qualitative sampling we are more interested in looking at, specifically, the differences between buddy pairs. Therefore we can look at the results of the Variance of Components Analysis (figure 1) to see where most of the explained variation comes from. Surprisingly, about 42% of the described variance comes from the length of main transect along which buddy pairs ran their perpendicular thirty meter transects. However, about 38% of the variance comes from the buddy pairs themselves; depth of the transect accounts for only about 20% of the total variance.
Differences Between Divers
We predicted that there would be a lot of dissimilarity between species indicated by divers (and there relative abundance) and as indicated by figure 2 the variance was spread across all species. Some species, such as the red abalone (Haliotis rufescens) related very little discrepancy because almost none were seen. However, for the fish count, the stripped surfperch (Embiotoca lateralis) contributed to most of the inconsistency.
Most of the disagreement came not from the distance between 3 and 4 (present to common) but rather between 1 and 2 (absent to rare). The relative present difference between buddy pairs as a function of mean abundance of species (figure 3) indicates either preconceived notions of species abundance individuals had or an unfamiliarity divers had with the species they were seeing.
Variance by Taxa
Because the main goal of the survey was to qualitatively assess the abundance of common species expected at the Hopkins Reef we compiled the data and calculated a mean abundance for each species (figure 4). We then were able to evaluate the difference across taxa as a function of depth, distance along the main transect (“meter”) and buddy pair. Here all species were grouped into their three taxa: Fish, Algae and Invertebrates (figure 5). For Algae the main variance came from distance along the meter, depth accounted for only about 10%. Again with Fish, meter accounted for the majority of the discrepancy, but buddy pairs accounted for the next variance (with a difference of about 10%). For Invertebrates the meter distance explains almost no variance whereas depth and buddy pairs account for almost the entirety of the discrepancy.
Figure 4: Mean abundance for all species
Discussion
By running this relatively simple survey we are able to look at both the broad question of abundance and distribution of species as well as the intricacies of the proposed results of qualitative sampling. From the analysis of the data the results show much more interesting trends in diver psychology and practice then they do in the proposed distribution of species. Interaction strengths, community composition and abundance in kelp forest ecosystems have already been well documented (Dayton et al 1992, Sala and Graham 2011,) so truly here we are examining the overall effectiveness of our qualitative sampling done at the Hopkins Reef.
Buddy pairing was, at best, random prior to the survey so no indication of similarity between divers could be assed; all divers went through similar training procedures prior to the dive but at different times of the year and under different conditions. Coincidently there is a large amount of varying disagreement between buddy pairs (figure 6) that is spread across all taxa, but varies by species. Not surprisingly there was no disagreement in presence of M. pyrifera as it is the dominant canopy forming species, as it is impossible to miss at Hopkins. Similarly the percent disagreement with two morphologically similar species of kelp Dictyoneurum californicum and Dictyonueropsis reticulata was 40% and 48% respectively which implies that one diver at least saw one of the two (because both are present throughout Monterey and central California) but only recorded seeing one or vice versa (Spalding, Foster and Heine 2003). And for fish species the percent disagreement between buddies for the striped surfperch, Embiotoca lateralis, was about 44%. This species is a mid-water fish that, when disturbed, will readily take flight. It is easy to assume that one diver saw a fish (specifically only one) before it swam away and before the other diver could take notice. Most surprisingly is the 39% disagreement regarding the orange cup coral, Balanophyllia elegans. In most cases all a diver has to do is look down over an exposed rock to see the vibrant cnidarians but, according the data, not everyone took notice. One would think that B. elegans would make for a good species for use in a qualitative survey. Likewise, a cryptic invertebrate, such as the red abalone, Haliotis rufescens, does not make for a good candidate because they tend to hide in cracks and crevasses. Moreover, the red alga, Chondrachanthus corymbiferus, is quite abundant and abundance can be easily quantified.
In consistency with figure 6 is the Variance Components Analysis (figure 1) which also indicates the highest variance following a discrepancy between an absence or presence of species noted by buddy pairs. This seems to coincide with an argument posed in the introduction; does experience matter? I would assert that more time spent in the water does yield higher precision and more consistent results. A diver can look both in cracks, over and on the substrate and in the water column while still recording accurate results only if they are comfortable with the study site and being comfortable in the water. This explains the difference between a record of presence and absence of a species; if there were only one cryptic black and yellow rockfish, Sebastes chrysomelas, an inexperienced diver would probably not see it wedged under a rock and record an absence of this vital species. A more experienced diver should know to look in large cracks where this species tends to seeks refuge.
But diver experience alone does not explain the variance between the other components of the data; conceptions of a present, common or abundant species (ranked 3, 4 and 5 respectively) as indicated (figure 4). One possible explanation for this variance is derived from the main argument of experience. That is, if an individual has spent a lot of time in the kelp forest and knows what to expect and what they should be seeing (and in what abundance) then certain factors might influence the results. For example, the blue rockfish, S. mystinus, is almost always abundant in the kelp forests of onterey. However, if a diver saw only 12 individuals, for example, then they might record that as being only present or even rare. Whereas said diver’s buddy might be from Southern California where seeing 12 blue rockfish in a single dive would be beyond abundant causing this diver to mark the abundance as such.
Overall the question of abundance and distribution of common species by means of a qualitative survey yielded mixed results. Some species proved to be good candidates for a qualitative survey while others did not. The afore mentioned algal species, D. reticulata and D. californicum, do not make for good examples because it would appear as though most divers cannot distinguish between the two (figure 6). Conversely, the bat star, Astarina miniata, would be a perfect candidate for qualitative surveying (figure 6). Because of its general abundance and easily recognizable morphology a diver of any experience level can note its distribution and profusion in a, hopefully, precise fashion.
There is, however, much to be said for the accuracy of this kind of surveying. As already stated there is much room for error. In the methods I stated how divers were told specifically to not discuss their conceptions or tactics prior to the dive. Had divers agreed on expectations per species before the diver I am convinced we would see more constant results. This, however, does not give any indication of greater precision. Perhaps if all divers had been screened for number of dives in a kelp forest and paired accordingly the data could be compiled to yield more exact results.