Displacement Model: Foraging Seabirds

Technical Details of the Model

This simplified model includes three options for quantifying the magnitude of barrier effects. One option is to assume that they come from a normal distribution with mean 20 km and standard deviation 5 km; this corresponds to the way that barrier effects are specified in the full model. The remaining two options assume that the bird will incur an additional flight cost by flying around the wind farm rather than straight through it and quantify this additional cost by assuming that the wind farm is circular. The two options specify the radius of the circle in two ways:

1. To ensure that the set of bearings from the SPA that encounter the circle is the same that encounter the actual wind farm; or
2. To ensure that the area of the circle is the same as that of the actual wind farm.

Input Files

The model requires six input files; one file to provide species-specific information about the spatial distribution of the species within the region of interest, one file to provide parameter values for the model, and four files to provide meta-information.

The four meta-information files contain:

colonies.csv : locations of each colony (with columns ‘colony’, ‘longname’, ‘long’, ‘lat’)

colonysize.csv : size of each colony, in terms of number of pairs (with columns ‘SPA’, ‘Colony’, ‘Guillemot’, ‘Razorbill’, ‘Kittiwake’, ‘Puffin’, ‘Gannet’). Note that SPAs may contain more than one colony.

wfinfo.csv : locations, on the 0.5 x 0.5 km GEBCO grid, that are contained within the core and buffer regions for each wind farm. Columns are:  ‘WFcode’, ‘Longitude’, ‘Latitude’. Ten options are currently used for WFcode: ‘IC-footprint’ (Inchcape footprint), ‘IC-buffer’ (Inchcape buffer), and so on for NnG, R3A, R3B and R3AB.

wfstats.csv :  summaries of the relationship between each wind farm and each SPA.  Columns are: SPA, WindFarm, Dist, AngleMid, large.anglerange, large.radius, large.MAD, small.anglerange, small.radius, small.MAD. ‘Dist’ is the distance, in km, between the central colony of the SPA and the centre of the wind farm. The remaining columns refer to the sets of bearings from the SPA that would include the wind farm.

The parameter.csv file contains parameter values to use in running the displacement model – an example is provided. Information is required for all rows and all species. The ‘Version’ column is used when performing sensitivity analysis, and should be left blank otherwise.

Model Functions

Three functions are used to directly run the model and process the output; the actual calculations are performed by a number of additional functions, but these are not called directly by the user.

“vsdm.runmodel” runs the model for a single set of parameter values. The arguments are:

                dat : spatial distribution of the species

                species : species name

                meta : meta information about SPAs and wind farms

                pars.full : parameter values

                pars.version : which version of the parameters to use (default is to use pars.version=0, which corresponds to none of the alternative versions being used)

                ppop : which proportion of the population to use for simulation (default is to use ppop = 1, which corresponds to 100% of the population)

               min.prey.adjust : ratio to use for adjusting minimum prey values

                seedval.location : random seeds to use for simulating locations; set automatically if not specified

                seedval.behaviour : random seed to use for simulation foraging behaviour; set automatically if not specified

“vsdm.sensitivity” runs the model for all parameter sets listed in the ‘parameters.csv’ file. The arguments are identical to those for “vsdm.runmodel” except that ‘pars.version’ is no longer an option (since the model is now run for all versions).

“getsummary.zones” summarizes the output from “vsdm.sensitivity” in a format that can be used to create a CSV file (spreadsheet). The sole argument is ‘out’, the output from running "vsdm.sensitivity".

Example

The following contains the R code used to run the analysis for kittiwake.

setwd("~/Work/Project/CEH/Displacement/SimplifiedModel/CurrentModel/")

source("Simplified-Displacement-Model-Version17.R")

pars.full <- (read.csv("Parameters/parameters-gps.csv"))[,3:9]

meta <- list(wf = read.csv("Meta/wfinfo.csv"), colony = read.csv("Meta/colonies.csv"), colonysize = read.csv("Meta/colonysize.csv"), wfstats = read.csv("Meta/wfstats.csv"))

levels(meta$colony$colony) <- c("ForthIslands", "Boddam", "FAR", "Fowlsheugh", "ForthIslands", "StAbbsHead", "TPL")

levels(meta$wfstats$SPA)[3] <- "ForthIslands"

## Species-specific part :

dat <- read.csv("/tmpshare/adam/projects/seabirds/Displacement/out-sixth/kittiwake.csv")

levels(dat$Site)[3] <- "ForthIslands"

out.kittiwake <- vsdm.sensitivity(dat, species = "Kittiwake", meta = meta, pars.full = pars.full,  ppop = 1/3)

 write.csv(getsummary.zones(out.kittiwake), file="Output/sensitivity-kittiwake.csv", quote=FALSE, row.names=FALSE)