Ecological Archives E091-097-A1

Delbert L. Smee, Matthew C. Ferner, and Marc J. Weissburg. 2010. Hydrodynamic sensory stressors produce nonlinear predation patterns. Ecology 91:1391–1400.

Appendix A. Map of study area, flow measurement parameters, predator abundances, and clam predation levels.

FigA1
 
   FIG. A1. Map of the study area.

 

Flow Measurement Methodology

We measured flow velocity and turbulence levels in the field using Acoustic Doppler Velocimeters (ADVs, NortekUSA™ Vector Model) and vendor supplied ExploreV™ software. We placed one ADV in each field site at 0.50 m above the mean lower low water line and measured flow velocity 0.50 m above the substrate. We selected this measurement height to maximize the duration of free-stream velocity measurements while reducing bias in turbulence between sites caused by localized differences in sediment topography. Flow measurement sites were adjacent to locations where we assessed clam predation (e.g., not in mid-channel or within oyster patch reefs or Spartina zones). Thus, our flow data provide a characterization of conditions relevant to our predation studies rather than an exhaustive description of all differences across the four regions encompassed by our investigation.

Flow velocity was sampled at a frequency of 16 Hz in 4-minute bursts every 15 minutes for 12 hours. After 12 hours, the ADVs were rotated between sites, and after 24–48 hours, the ADVs were retrieved to download data and recharge batteries. The ADVs were then redeployed so that each instrument was placed into a different site than one measured previously. This design insured that we sampled all sites equally during each 24 or 48 hour measurement period. We performed a total of 8, 12–h flow measurements per site and made 4 measurements during spring tide and 4 measurements during the week after spring tide to gauge the relative flow differences between sites as they vary with tidal exchange.

Flow Measurement Data Analysis

ADVs measure 3 dimensional flows, and we calculated the net flow velocity (U) using the formula where u, v, and w are the velocity components in the x, y, and z dimensions respectively. We determined the net flow velocity for each 4 minute measurement period and then averaged all of the measurement periods to determine the mean flow velocity for that 12-h ADV deployment.

Turbulence was calculated using the root mean square (RMS) of the velocity time series. As with flow velocity, we combined RMS in the x, y, and z dimensions for each 4 minute measurement period using the formula where these values represent the RMS levels in the x, y, and z dimensions respectively. We then averaged these RMS calculations from all measurement periods during the 12-h deployment to determine a mean turbulence level. Thus, we had n = 8, 12-h deployments per site for which we calculated mean velocities and RMS turbulence levels.

We compared mean flow velocities and mean RMS levels between sites from each 12-h ADV deployment using blocked one way ANOVAs with date as the blocking factor and site as the fixed factor (Sokal and Rohlf 1995). The blocking factor was significant and included in the ANOVA model. Pair wise comparisons between sites were performed using Fisher’s Protected Least Significant Difference (PLSD) post hoc tests (Sokal and Rohlf 1995). Data collected when the ADVs were out of water was not included in analysis nor were samples with low signal to noise ratios (<5) and when beam correlations were < 70. Within each 4 minute sampling burst, 3840 measurements were taken, and less than 200 measurements were removed from each of the bursts used for analysis.

FigA2
 
   FIG. A2. Relationship between number of clams eaten and the number of predators collected per site. Field sites are labeled. Blue crabs composed > 97% of the predators collected. ANOVA results comparing differences in # crabs caught per site did not reveal significant differences (P = 0.25, n = 20).


FigA3
 
   FIG. A3. Mean number (+ SE) of adult and juvenile clams consumed per site. Predation on juvenile clams accounted for >99% of total predation in the study.

[Back to E091-097]