dwpm {GenEst}

Fit density-weighted proportion (DWP) models.

Description

Carcass density is modeled as a function of distance from turbine. Format and usage parallel that of common R functions lm, glm, and gam and the GenEst functions pkm and cpm.

Usage

dwpm(data_DWP, type = "data", unitCol = NULL, dwpCols = NULL)

Arguments

data_DWP

data frame with structure depending on model type. In general, data_DWP would be a data frame if a model is to be fit or if point estimates only are provided as pre-simulated DWP data, and, if pre-simulated data with variation are provided, then a 2-d array (if one carcass class) or a list of 2-d arrays (if more than one carcass class). See "Details" for details.

type

model type may be rings, glm, TWL, or data. Currently, only the data type is supported.

unitCol

name of the column with the units, which must be non-numeric

dwpCols

name(s) of the columns with the DWP data

Details

The fraction of carcasses falling in the area searched at a turbine may be a function of carcass class (e.g., large or small) and/or direction from the turbine. Data may be provided for fitting a distance model(s) or, alternatively, simulated turbine-wise DWP data from custom-fitted models may be provided. If pre-fit, pre-simulated data are used, then glm returns a dwpm object with type = data.

To fit a model, data_DWP should be a data frame with a row for each carcass and columns giving (at a minimum) unique carcass IDs, turbine ID, distance from turbine, and fraction of area searched at the given distance at the given turbine. Optional columns may include carcass class, covariates that may influence detection probability (e.g., visibility class), and direction. If covariates are to be included in the model, then the fraction of area column would give the fraction of the area in the given covariate level at that distance. Alternatively, prefab data may be provided in a dataframe, with structure depending on data type. The simplest case would be that point estimates only are provided. In that case, if there are no distinctions among carcass classes (e.g., size), then data_DWP should be a dataframe with one column giving the unit (e.g., turbine) and one column with the DWP at each unit; if distinctions are made among carcass classes, then data_DWP would be a data frame with a unit column and a DWP column for each carcass class. If the DWP estimates incorporate uncertainties, then data_DWP should be an array with n_unit * nsim rows and with colunms for units and DWPs for each carcass class.

Value

an object of an object of class dwpm, which is a list with model type (currently only type = data is supported) and model, which gives the simulated DWP values as an array (if there's only a single carcass class) or a list of arrays (if there are more than one carcass classes).