Map showing the bathymetry and location of the Moray Firth. The 50 m depth contour in the center of the firth demarks the Smith Bank.
Spatial extent of effort from four boat based surveys (a to d), and one aerial survey (e) carried out in the Moray Firth in (a) 2004 (August, September, October) and 2005 (April, May, June, July), (b) 2009 (June, August, September, October), (c) 2010 (April, May, June, July, August, September, October), (d) 2010 (April, May, June, July, August, September) and (e) 2010 (August, September).
Habitat variables (depth and sediment type) summarized over a 4 × 4 km grid. Sediment type is coded as the proportion of the area of the cell that was classified as sand or gravelly sand. © Crown Copyright/SeaZone Solutions Ltd. All Rights Reserved.
Location of harbor porpoise sightings from four boat based surveys (a to d) and one aerial survey (e) in the Moray Firth between 2004 and 2010.
Two dimensional smoother used in the porpoise habitat association model to describe the interaction between depth and the proportion of sediments that were sand or gravelly sand, and the relationship between these habitat variables and harbor porpoise relative abundance.
Maps of predicted relative abundance of porpoises (porpoise km−1) and the standard error of the prediction, from a GAMM with depth and the proportion of sand and gravelly sand sediments, given 1 km of effort in each cell. Predictions were not made in cells where depth was greater than 80 m because no survey data were available to inform these predictions (white areas in the figures).
(a) Spatial variation in median PPH detected on acoustic loggers deployed April to October in 2009 and 2010. Loggers less than 10 km from the coast are considered to be “coastal,” while those further from coast are considered to be “offshore.” (b) Comparison of detection rates in 2009 and 2010 at the 30 sites where data were available from both years. Median PPH for the sites in both years is shown, along with the line of best fit from a linear model for illustration.
Plots of the predicted porpoises km−1 from the habitat association model, against the median hours porpoises were detected acoustically within the same cell. (a) Data are included from all cells; (b) only cells where the standard error of the predicted number of porpoises was less than 1.4 are included. The line of best fit from a linear model is plotted for illustration.
Box plots of median number of detection positive hours from CPODs grouped by (a) predicted porpoise numbers for all cells in a habitat association model from visual survey data (N = 69) and (b) for cells with predictions with standard errors less than 1.4 (N = 54). The low group contains predicted porpoises km−1 values ranging from 0 to 0.039, the medium group contains values from 0.04 to 0.079 and the high group contains values from 0.08 to 0.13. Boxes represent the limits of the second and third quartile, while the bold central line represents the median. The range of the data (first and fourth quartiles) is shown by the dotted lines and outlying observations are represented by circles.
Details of five survey datasets of harbor porpoise, collected in the Moray Firth and used in habitat association modeling. Dataset a contains data previously published in Bailey and Thompson (2009) .
Results of a negative binomial GAMM used to analyze harbor porpoise counts, using a tensor smoother, with an interaction term between depth and the proportion of sand and gravelly sand.
Comparison of CPOD metrics used in correlation analyses with the porpoise habitat association model, from 69 CPODs deployed from April to October of 2009 and 2010. PPD is the proportion of porpoise positive days, PPH is median porpoise positive hours per day, PPM is median porpoise positive minutes per day and waiting time is the median number of minutes between successive porpoise detections. Correlations are Spearman's rank, on the full and reduced datasets. Coastal CPOD locations are within 10 km of land.
Article metrics loading...
Full text loading...