Log-transform climate variables

This function performs logarithmic transformation on climr output to a user-specified base. This function currently only works with climr output in data.table format (i.e., not with raster output).

Usage

logVars(
  dat,
  elements = c("AHM", "DD", "Eref", "FFP", "NFFD", "PAS", "PPT", "SHM", "CMD"),
  base = exp(1),
  add.fields = FALSE,
  zero_adjust = TRUE
)

Arguments

dat: a data.table with columns of climate variables including the selected climate elements.
elements: character. Climate elements to log-transform.
base: numeric. Logarithmic base.
add.fields: logical. If TRUE, the new logged variables are added to dat (TRUE). Otherwise, original column values are replaced with the logs (FALSE).
zero_adjust: logical. If TRUE adjusts zeroes in raw data as: \(base^{\log_base{x_min} - 1}\). where \(x_min\) is the minimum non-zero, non-NA value.

Value

data.table

Details

In climate analysis it is often preferable to log-transform ratio variables, i.e., variables such as precipitation and degree-days that have a lower limit of zero and for which proportions are meaningful.

Climate elements include all related time variables, hence all column names that partially match strings in elements will be log-transformed. For example, specifying elements = "PPT" will #' result in log-transformation of c("PPT", "PPT_01", "PPT_02", ... , "PPT_at").

By default, zeroes are assigned a log value that is one order of magnitude less than the minimum positive value of the untransformed variable.