Wind speed data for shore and forest on an island in French Polynesia WaldropLindsay 2014 <p>Wind data were collected on three days during the second and third weeks of June in 2010 on Motu Tiahura when the weather was clear or partly cloudy and calm. The two sites for collection were the "shore" on open sand approximately 3 meters inland from the high-tide mark and at least 3 meters from the beginning of the tree line and "forest" on the leaf-litter under tree canopy at least 10 meters from the tree line. Hermit crabs were collected from both sites where wind data were collected. Wind speed measurements were collected using a hot-wire anemometer (Kurz 443M, Kurz Instruments, Inc, Monterey, CA) in a water-proof casing. Data was routed through a DAQ card (National Instruments, Inc., Austin, TX) to a Matlab software acquisition GUI on an IBM Thinkpad laptop. The anemometer was calibrated in a wind tunnel across a wide range of wind speeds prior to being taken into the field. Data were collected at 12 heights above the substratum from 5 mm to 0.68 m, covering the height range which hermit crabs would likely capture odors. Wind speeds were sampled at data were taken for three-minute intervals and sampled at 100 Hz. Measurements were repeated three times (on three different days) and averaged for each position above the substratum. Turbulence intensity was calcuated by dividing the standard deviation of wind speed in the sample by the mean speed of the sample. Statistics were performed with <em>R</em> and values are reported with standard error.</p> <p>Mean_vel.pdf reports wind speeds (in meters per second) graphed against height above substratum (in meters). Red dots indicate the "forest" location and blue dots indicate "shore" location. In a linear regression model (speed~height*location), there is a significant correlation between speed and height above substratum (slope = 0.18 ±0.04, t-value = 4.477, dof = 20, p = 2e-4) but only a slight correlation between speed and location (coef = 0.0506 ±0.02, t-value = 2.252, dof = 20, p = 0.03). </p> <p>turbint_regression.pdf reports turbulence intensities graphed against height above substratum (in meters). In a linear regression model (turbulence intensity~height*location), turbulence intensity has a significant correlation with height above substratum (slope = 0.0845 ±0.01, t-value = 5.712, dof = 20, p = 1e-5) but no relationship with location (coef = -0.00851 ±0.008, t-value = -1.07, dof = 20, p = 0.30).</p>