forecast: Caribou Population Forecasting

In psolymos/CaribouBC: Caribou Population Forecasting

Description

Functions to set demographic and cost parameter, to perform forecasting, and to inspect the results.

Usage

caribou_settings(pen.type = c("mat.pen",  "pred.excl",
    "moose.red", "wolf.red", "cons.breed"), herd = NULL, ...)

caribou_forecast(settings, tmax = 20, pop.start = 100,
    fpen.prop, fpen.inds)

caribou_breakeven(forecast, lambda=1,
    type=c("prop", "inds"), max=10^4, tol=0.01)

## S3 method for class 'caribou_settings'
print(x, ...)
## S3 method for class 'caribou_forecast'
print(x, ...)
## S3 method for class 'caribou_forecast'
plot(x, plot = TRUE, ...)
## S3 method for class 'caribou_forecast'
lines(x, pen = TRUE, plot = TRUE, ...)
## S3 method for class 'caribou_forecast'
summary(object, ...)
## S3 method for class 'summary.caribou_forecast'
print(x, ...)

Arguments

pen.type	character: maternity penning, predator exclusion, moose reduction, wolf reduction, or conservation breeding. It can also be a settings object.
herd	NULL (default, East Side Athabasca, Boreal ecotype) or character. The following herds (Southern Mountain ecotype) are available: "ColumbiaNorth", "ColumbiaSouth", "FrisbyQueest", "WellsGreySouth", "Groundhog", "Parsnip", "KennedySiding", "KlinsezaMoberly", "Quintette". Note: only "KennedySiding", "KlinsezaMoberly", or "Quintette" can be selected when pen.type = "wolf.red", beside a generic herd setting when herd is NULL that is specific to the 'wolf removal' treatment type (wild, i.e. 'no wolf removal' settings for these herds are available under the default pen.type = "mat.pen" treatment but no separate generic herd settings are provided for the 'no wolf removal' treatment).
settings	a settings object returned by caribou_settings.
tmax	positive integer, number of years to forecast after initial year 0.
pop.start	positive integer, initial population size in year 0.
fpen.prop, fpen.inds	fpen.prop is a single numeric value between 0 and 1 for the proportion of females penned; fpen.inds is a non-negative integer for the number of females penned, can be a vector giving number of females penned in each year to be accumulated. Only one of fpen.prop or fpen.inds can be provided at a time. Any of the two arguments can be NULL or missing, but not both at the same time.
forecast	a forecast object returned by caribou_forecast.
lambda	numeric, annual intrinsic growth rate, must be >0.
type	what to provide as output: proportion ("prop") or numbers ("inds") of females penned at breakeven point.
max	numeric (>0), maximum value for breakeven optimization when type = "inds".
tol	numeric, tolerance limit, i.e. the maximum acceptable deviation from lambda when solving for 'breakeven' proportion of females penned.
x, object	an object to print, summarize, plot, etc.
plot	logical, whether a plot is to be produced.
pen	logical, whether to display the pen or no-pen (baseline) results.
...	additional arguments to functions. For caribou_settings see Details.

Details

The following cost parameters can be set by caribou_settings:

pen.cap: how many individual Caribou can live in a single maternity pen.

pen.cost.setup: initial cost in thousands to set up pen.

pen.cost.proj: annual costs in thousands for project manager.

pen.cost.maint: annual cost in thousands for patrolling and repairing fence.

pen.cost.capt: annual cost in thousands to capture cows, monitor, survey, calf collar.

pen.cost.pred: annual cost in thousands for removing predators.

The following demographic parameters can be set by caribou_settings:

c.surv.wild: calf survival rate in the wild, annual.

c.surv.capt: calf survival rate when captive, annual.

f.surv.wild: adult female survival when wild, annual.

f.surv.capt: adult female survival when captive, annual.

f.preg.wild: pregnancy rate when wild.

f.preg.capt: pregnancy rate when captive.

Value

caribou_settings returns a settings object.

caribou_forecast returns a forecast object.

caribou_breakeven returns a numeric value representing the 'breakeven' proportion of females penned where lambda is within tolerance. It returns NA when proportion satisfying the lambda criterion cannot be found, alongside a warning.

The print method returns the input object x invisibly.

The summary method returns population and cost summaries for the forecast object.

The plot and lines methods return the plotted data invisibly (years, pen and no-pen population size for plot; years, pen or no-pen population size for lines). Both methods produce plots as a side effect.

See Also

caribou_matrix, caribou_breeding

Examples

## Predefined settings
(s1 <- caribou_settings("mat.pen"))
(s2 <- caribou_settings("pred.excl"))

## Modifying predefined settings
caribou_settings("mat.pen", c.surv.capt=0.65, pen.cap=30)
caribou_settings(s1, c.surv.capt=0.65, pen.cap=30)

## Forecast based on settings for 75% females penned
(f1 <- caribou_forecast(s1, fpen.prop = 0.75))
(f2 <- caribou_forecast(s2, fpen.prop = 0.75))

## Get population and cost summaries
summary(f1)
summary(f2)

## Plot the results
plot(f2)
lines(f1, col = 2)
legend("topleft", col = c(1,1,2), lty = c(2,1,1),
    legend = c("No pen", "Mat pen", "Pred excl"))

## Find 'breakeven' proportion of females penned where lambda=1
(b1 <- caribou_breakeven(f1, lambda = 1))
(b2 <- caribou_breakeven(f2, lambda = 1))
f3 <- caribou_forecast(s1, fpen.prop = b1)
f4 <- caribou_forecast(s2, fpen.prop = b2)
## See that lines are truly flat
op <- par(mfrow = c(1, 2))
plot(f3, main = "Mat pen")
plot(f4, main = "Pred excl")
par(op)

psolymos/CaribouBC documentation built on April 2, 2020, 3:54 a.m.