gibbonsecr-covariates: Model covariates
In dkidney/gibbonsecr: SECR Analysis for Acoustic Gibbon Surveys

Covariates can be used when constructing formulas for the various model parameters using the model argument in the gfit function. However, some parameters can only be modelled with certain types of covariates.

Covariates can be divided into the following types, according to the level at which they vary:

array-level: These are consistent within arrays. All listening posts and sampling occasions for a given array will have the same value for an array-level covariate, but values vary between arrays. For example, season is likely to be an array-level covariate, since it is highly unlikely that consecutive sampling occasions at a given array will overlap with more than one season.
trap-level: These are consistent within listening posts for a given array. All sampling occasions for a given listening post will have the same value for a trap-level covariate. For example, habitat and elevation are likely to be trap-level covariates since they may vary between listening posts but will be consistent across occasions.)
occasion-level: These are consistent within sampling occasions for a given array. All listening posts in an array will have the same value for an occasion-level covariate. For example, weather is likely to be an occasion-level covariate since it will probably be the same for all listening posts for a given occasion at a given array but may vary between occasions.
trap-occasion-level: These can vary between listening posts and between sampling occasions. For example, observer might be a trap-occasion-level covariate since observers will (always) vary between listening posts for a given occasion at a given array, and they may also vary between occasions for a given listening post at a given array.
mask-level: These are spatial covariates and can vary across mask points. Mask-level covariates from GIS polygon and point files can be imported and attached to mask objects using the import_shp and add_covariates functions.

The table below shows which covariates can be used to model which parameters.

Parameter	Covariate level
`D`	array, mask
`g0`	array, trap, occasion, trap-occasion
`sigma`	array, trap, occasion, trap-occasion
`z`	array, trap, occasion, trap-occasion
`bearings`	array, trap, occasion, trap-occasion
`distances`	array, trap, occasion, trap-occasion
`pcall`	array, occasion