makeFecObj: Function to build fecundity objects

In wpetry/IPMpack2: Builds and analyses Integral Projection Models (IPMs).

Function to build fecundity objects

Description

Allows a series of different glms to be fit all on the way to fecundity, e.g., probability of flowering, number of flower heads produced, etc; as well as fecundity into different discrete classes. Alternatively, the function creates fecundity objects following a specified list of formulas with specified corresponding list of coefficients.

Usage

makeFecObj(dataf, fecConstants=data.frame(NA),
		Formula=list(fec~size),Family="gaussian",
		Transform="none",meanOffspringSize=NA,
		sdOffspringSize=NA,offspringSplitter=data.frame(continuous=1),
		vitalRatesPerOffspringType=data.frame(NA),fecByDiscrete=data.frame(NA),
		offspringSizeExplanatoryVariables="1", coeff=NULL,doOffspring=TRUE, 
		reproductionType="sexual")

Arguments

dataf	a dataframe with columns "size", "sizeNext", "stage", "stageNext", and any additional columns with fecundity data. UIf fecundity data is transformed via log, etc, this MUST BE MADE CLEAR in the argument Transform since the fecundity object produced must generate total reproductive output.)
fecConstants	a list containing the value by which each of the fecundity rates will be multiplied in the order defined by the order in Formula. This data frame adjusts the probability of establishment of seeds or other stages in sexual reproduction that are not explicitly incorporated for each parent (e.g., 25% of seeds across all individuals germinate). The default is NA if no constants are used (equivalent to multiplying by 1).
Formula	a formulas describing the desired explanatory variables (interactions, etc) in classical R style, i.e. separated by ‘+’, ‘*’, ‘:’ and the response variables of choice. Possible covariates include ‘size’, 'size2' (size^2), ‘size3’ (size^3),‘logsize’ (log(size)), and ‘covariate’ (if this name is used, the assumption is made that this is a discrete covariate from which compound matrices may be constructed), and any other covariates available in dataf. Again, these must appear in the order defined by the Formula argument. See formula in base for more details.
Family	a character vector containing the names of the families to be used for the glms, e.g., binomial, poisson, etc. Again, these must appear in the order defined by Formula
Transform	a character vector containing the names of the transforms to be used for the response variables, e.g., log, sqrt, -1, etc. Again, these must appear in the order defined by Formula
meanOffspringSize	numeric vector, defining mean offspring size. Defaults to NA, in which case the function will use to data to assess the mean offspring size according to the relationship defined in offspringSizeExplanatoryVariables (which either simply fits a mean, or may fit more complex relationships linking maternal size to offspring size).
sdOffspringSize	numeric vector, defining standard deviation of offspring size. Defaults to NA, in which case the function will use the data to assess the standard deviation of offspring size; as described for meanOffspringSize
offspringSplitter	dataframe with values defining the number of offspring going into the indicated offspring category; will be re-scaled to sum to 1 within the function. This argument needs to be entered as a data.frame, and the names in the data.frame need to precisely match the used stage names in the data file.
vitalRatesPerOffspringType	dataframe defining which fecundity rates (both functions and constants) apply to which offspring category. This only needs to be specified when some fecundity rates do not apply to all offspring categories. The offspring categories in the column names of this dataframe should match those in the offspringSplitter exactly. The row names of the dataframe should match the fecundity column names in the data file and the supplied fecundity constants, in that order. In the dataframe, a '1' indicates that a fecundity rate applies to an offspring category, while a '0' indicates an omission. For instance, establishment and seedling survival rates may be applicable to seedlings, but not to seeds that go into a seedbank (depending on the life cycle and definition of vital rates).
fecByDiscrete	data.frame defining number of offspring produced by each discrete class ; defaults to 0. If specified, ALL discrete classes MUST appear in alphabetical order, so NO "continuous". e.g. fecByDiscrete=data.frame(dormant=0,seedAge1=4.2,seedOld=0)
offspringSizeExplanatoryVariables	a character defining the relationship defining offspring size; the default is "1", indicating simply fitting a mean and a variance; alternatives would including defining offspring size as a function of maternal size (i.e., offspringSizeExplanatoryVariables="size") or more complex polynomials of size (i.e., offspringSizeExplanatoryVariables="size+size2"). The corresponding relationship is fitted to the data contained in dataf, taking as the response variable the column "sizeNext" in dataf for rows where the column "offspringNext" is equal to "sexual" and the column "stageNext" is equal to "continuous".
coeff	list of numeric vector of required coefficients to be imposed if dataf is NULL for each element of the Formula list in order; must be compatible with Formula
doOffspring	argument defining whether you wish to fit an offspring regression as part of this function (TRUE), or do it separately (see makeOffspringObj); if this is FALSE, other argument relating to the offspring distribution will be ignored
reproductionType	argument specifying if "sexual" or "clonal" offspring should be considered; default is "sexual".

Details

This function fits a suite of subfunctions of fecundity towards creating a Fecundity transition projection model; e.g., the probability of flowering as a function of size, the number of seeds produced as a function of size, etc. Users can define the functional form of each relationship, as well as the distribution and any transforms. There is also a possibility of defining reproduction from discrete sizes into each of the subfunction outcomes; defined in the matrix fecByDiscrete.

Note that it is crucial that users appropriately set up the data to adequately reflect conditionality in the fertility kernel; for example, if there are two columns, with one reflecting the probability of flowering (0s and 1s) and the other reflecting seed output (integers) it is important that where the probability of flowering is 0, seed output is set to NA, as otherwise, meaningless 0s in the seed output column will bias the regression.

Finally, if the same variables are used in multiple equations, users should check for problems of endogeneity.

Value

an object of class fecObj

Author(s)

C. Jessica E. Metcalf, Sean M. McMahon, Roberto Salguero-Gomez, Eelke Jongejans & Cory Merow.

See Also

makeSurvObj, makeGrowthObj, makeOffspringObj

Examples

dff <- generateData(type="discrete")
fv1 <- makeFecObj(dff, 
	offspringSplitter = data.frame(continuous = 0.2, seedAge1 = 0.8, 
	seedOld = 0),
	fecByDiscrete = data.frame(dormant = 0, seedAge1 = 4.2, seedOld = 0))

#now without any data, imposing desired coefficients
fv1 <- makeFecObj(Formula=list(Fec~size),
				offspringSplitter = data.frame(continuous = 0.2, dormant=0,
				seedAge1 = 0.8),
				fecByDiscrete = data.frame(dormant = 0, seedAge1 = 4.2), 
				coeff=list(c(1,1)), meanOffspringSize=1,sdOffspringSize=1)
				

wpetry/IPMpack2 documentation built on Sept. 29, 2022, 9:41 a.m.