CICEET Progress Report for the period 9/15/04 Through 3/15/05

Project Title: Application of an in situ infrared camera system for evaluating icthyofaunal utilization of restored and degraded mangrove habitats: developing a set of reference conditions from a NERRS site.
Principal Investigator(s): Susan S. Bell, William L. Ellis

Figures

Figure 1

Figure 2

Figure 3*

Figure 4*

Figure 5

Figure 6 *Figures 3 and 4 are Windows Media movies that require Windows Media Player

B) Compare the activity and species composition of a fish community at a NERR mangrove restoration site to that of a natural mangrove fish community.

C) Compare the activity and species composition of a fish community along an obviously modified shoreline to that of an unmodified mangrove shoreline.
Both tasks B and C are intended to explore the potential of video sampling as a means of evaluating habitat quality.

Progress on Tasks
A) To illustrate within-site spatial variability, we simultaneously recorded the activity of intertidal fishes at four unevenly spaced (2, 4, and 8 m apart) points along a single 15 m segment of the mangrove/open water ecotone in Rookery Bay. The species composition of the fish community seen by each of the four cameras over a 5.5 h recording period was compared using Bray-Curtis similarity measures.

B) We made several unsuccessful attempts to record the intertidal movements of fishes in the Rookery Bay mangrove restoration sites (see Difficulties Encountered).

C) Fish utilization of seawall fortified shorelines and nearby mangrove shorelines was compared at three sites in Tampa Bay. We have just completed the videotaping for a companion study, conducted at several sites in Rookery Bay, in which we manipulated food availability (absent vs. highly abundant) within the field of view of cameras position along the mangrove/open water ecotone. By doing this we have documented a resource dependent behavioral response by a common mangrove inhabitant, the mojarra (Eucinostomus spp.).

Difficulties Encountered
We encountered several difficulties during this last reporting period. Most significant of these, has been our inability to record useful video in the intertidal zone of the Rookery Bay mangrove restoration site. Our two major impediments have been the weather and boat traffic. On several occasions during the periods of planned recording, thunderstorms have struck southwest Florida. Since any attempt to deploy the recording gear during this sort of weather would have likely resulted in severe damage to the digital video recorders, we had no choice but to cancel the planned activities. More troubling is the high turbidity of the water at the restoration site. The restoration site is located on a narrow and shallow portion of Henderson Creek that is highly trafficked by boaters. Except during high tide, boats must be on a plane to navigate these waters. We have found that the wake produced by the fast-moving boats ensures that fine sediment is continually suspended in the water column. As a result, all of our recordings in the restoration site have been under near “zero visibility” conditions. We are hoping to get some video footage from the restoration site after the winter tourist season is over.

In our previous report, we described our attempts to create an automated measure of video quality. Our attempts to compensate for the interference of site specific features in this measurement have not met with success. Instead, we have identified yet more sources of interference but no practical way to effectively address them all. So, as of now, our automated scheme will only work under a narrow range of conditions. We will continue seek new solutions this problem.

Anticipated Success in Meeting Project Objectives
We anticipate meeting the major project objectives in the scheduled project period.

Preliminary Data
A) Spatial Variability-
We anticipated that the degree of similarity of the fish community seen in the field of view of each camera would be related to the distance between them. Contrary to that expectation, no strong relationship was found between camera spacing and Bray-Curtis similarity (r2= 0.501, p=.116; See Figure 1). Instead, similarity in the species composition at the four different positions along the mangrove shoreline is apparently a function of microhabitat features (e.g. oyster substrate vs. soft mud) that do not necessarily vary consistently from one end of the series of cameras to the other. Some portion of the recorded community dissimilarity between the cameras is also likely a result of visibility issues and subsampling artifacts. Surprisingly, visibility was not consistent among cameras. Periodic, spatially uneven increases in turbidity likely concealed some species resulting in an underestimate of their abundance at some but not all cameras. Also, several of the species identified in the video moved laterally along the mangrove/open water ecotone and were “missed” by subsampling as they passed by one or more of the cameras. Indeed, a separate accounting of species presence/absence over the entire recording period indicated that Sphyraena barracuda, Mugil cephalus, and Harengula sp. were omitted from species lists when they were, in fact, present. The variability in the similarity measures from these sites within a superficially homogenous mangrove habitat may be a call for caution in the interpretation of multivariate comparisons of fish community structure in mangrove habitats.

C) Comparison of Seawall to Mangrove Shorelines-
With this in mind, we began our investigation into the differences in fish utilization of mangrove and seawall shorelines. At each of three locations in Tampa Bay, we selected a pair of mangrove and seawall sites that were within 16 m or less of one another so that fish had equal access to both habitat types. During the deployment of the recording gear, it was immediately apparent that the simple classification of sites as either mangrove or seawall would likely be inadequate given noticeably inconsistent physical characteristics among members of each class (i.e., mangrove or seawall). Two of the mangrove sites (Weedon Island and Coquina Key) had similar slopes, elevation, substrate, and mangrove vegetation whereas the third (Tierra Verde) was noticeable different from the first two in all of these features. The seawall sites were more dissimilar from one another: The Weedon Island seawall site was characterized by a higher elevation than the other two sites. The seawall was made of wood and had virtually no epibionts attached to it and was faced by a featureless sand bottom. In contrast, the Coquina Key seawall site was an estimated 15 cm deeper than the Weedon Island site, and was littered with concrete rubble. The Tierra Verde Seawall site was an estimated 45 to 50 cm deeper still. Its seawall was covered in tunicates, sponges, and barnacles. The substrate was a silt sand mixture and was sparsely covered with seagrass and sea whips. Given these physical differences, we predicted that the Weedon Island seawall site and the Tierra Verde mangrove site, would support fish assemblages that were similar to each other yet distinct from the remaining four sites. Likewise, the Tierra Verde site was so completely different from the other sites, we anticipated it to support a fish assemblage that was distinct from the others.

An MDS plot (Bray-Curtis similarity measure applied to square root transformed abundance data; See Figure 2) of abundance data did not agree with these expectations. However, an MDS plot that examined the similarities in bottom picking behavior (# of picks per minute per individual present in that minute; See Figure 3) and aggression (proportion of total minutes during which a species was noticeable aggressive; See Figure 4) by the five numerically dominant species agreed perfectly with our predictions (See Figure 5). This suggests that studies of even simple behaviors may improve our ability to distinguish between sites of differing habitat quality.

We are also in the midst of investigating a somewhat more complex measure of behavior that may prove to be more useful as a distinguishing feature between “high” and “low quality” mangrove sites. Based on our initial in situ video observations and some laboratory work performed by Burrows and Gibson (1995), we propose that the swim burst duration of the ubiquitous mojarra may serve as a reliable indicator of food availability. We have just completed the field work for a rudimentary test of this hypothesis. In this study, we placed sand-filled trays along the waterward edge of the mangroves in Rookery Bay and recorded mojarra behavior around them. Frozen brine shrimp were added to half of the trays and covered with a light layer of sand before the experiment began. No shrimp were added to the other trays. We predict that swim bursts will be shorter over the resource rich trays. If this prediction is verified, we will re-examine earlier video footage for differences in mojarra swimming behavior that we will attempt to link to site-specific food availability.

As somewhat of an aside, we have found that the use of video sampling over the full tidal cycle may aid in avoiding some of the pitfalls associated with sampling (e.g. with a seine net) at a single point in time. For example, consider the case of an ecologist collecting a fish sample from the mangroves at high tide. Typically, within the same study, all other samples will also be collected at high tide for consistency sake. What is often missing from such designs is the acknowledgement that the habitat may have been used quite differently by a species at a different point of the tide cycle. For example, mojarra were absent at Coquina Key mangrove site at high tide but were highly abundant there at mid-flood (See Figure 6). The only alternative to the researcher that employs nets would be to repeatedly sample the same site over time and risk inserting a sampling disturbance artifact into the sampling design.

Tasks and Activities for the Next Reporting Period

Tasks for the next reporting period
We will complete the video analysis of footage from Rookery Bay and Tampa Bay. Following that, we will submit our final report. If possible, we will acquire and analyze footage from the Rookery Bay mangrove restoration site.

Work plan to accomplish tasks
We will continue video analysis until it is complete. We anticipate field activity at the restoration site in early May.

Concerns or difficulties
It is possible that the visibility in the waters of the restoration site will not be suitable for meaningful video recording.

Expenditures
Expenditures were in the range anticipated for the work accomplished to date.