phenModel (version 1.0)

sensitivity: Sensitivity Analysis for Evaluated Phenology Model

Description

Runs a sensitivity analysis on the evaluated phenology model, by simulating random changes to fitted parameters and/or specifying a range of base temperatures to determine a chill day.

Usage

sensitivity(Tmax, Tmin, Tbase, lat, day,
            bud_pars, weib_pars, lactin_pars, max_day_length, 
            pop_quantile = .5, data,
            study_type = c("Tbase","lactin","weibull","bud_days","bud_parms","cdl"),
            percent_variation = .1, resolution = 1000)
summary_sensitivity(obj)
boxplot_sensitivity(obj)

Arguments

obj

an output from a call to the sensitivity function

Tmax

a vector of daily maximum temperatures

Tmin

a vector of daily minimum temperatures

Tbase

the base temperature

lat

the latitude

day

day of the year (vector)

bud_pars

vector of parameters for the budburst model

weib_pars

matrix of parameters for the Weibull model for each developmental stage

lactin_pars

matrix of parameters for the lactin model for each developmental stage

max_day_length

critical day length threshold

pop_quantile

quantile of the population to be evaluated by the phenology model

data

a data.frame

study_type

a character string indicating what type of sensitivity study to be performed. "Tbase" runs a sensitivity study for the temperature threshold that classifies a chill day; "lactin" runs a sensitivity study on variations of the values for the parameters estimated by the lactin model; "weibull" does the same as the latter, but for the Weibull model; "bud_days" changes the budburst day while "bud_parms" randomly changes the parameter estimates for the exponential decay function associated with budburst prediction; "cdl" changes the critical day length threshold

percent_variation

percentage of variation of the values to be changed for the sensitivity analysis

resolution

number of model evaluations to be performed

Value

A list containing the sensitivity analysis results as the first component, and the original model evaluation results as the second component.

See Also

phen_model

Examples

# NOT RUN {
data(casement9596)

casement1995 <- data.frame(casement9596, 
                           with(casement9596, 
                                get_total_info(Tmax, Tmin, Tbase = 7, Lat, day)))

data(pvulg_budburst_parameters)
data(pvulg_weibull_parameters)
data(pvulg_lactin_parameters)

sens_casement1995 <- sensitivity(Tmax = casement9596$Tmax,
                                 Tmin = casement9596$Tmin,
                                 Tbase = 7,
                                 lat = casement9596$Lat,
                                 day = casement9596$day,
                                 bud_pars = pvulg_budburst_parameters,
                                 weib_pars = pvulg_weibull_parameters,
                                 lactin_pars = pvulg_lactin_parameters,
                                 max_day_length = 14.92, 
                                 pop_quantile = .5,
                                 data = casement9596,
                                 study_type = "lactin",
                                 percent_variation = .1,
                                 resolution = 100)
                                 
summary_sensitivity(sens_casement1995)

boxplot_sensitivity(sens_casement1995)
# }