Skip to contents

Interpolate environmental data onto a finer time step

Source: R/retime.R
downscale_time.Rd

Puts a coarse series onto a finer one, so a monthly product can be matched to observations on their actual dates rather than by month.

Usage

downscale_time(
  env_dat,
  to = c("day", "month"),
  vars = NULL,
  method = "step",
  extrapolate = TRUE
)

Arguments

env_dat

an sf POINT object from accessEnvDat()

to

the target step: "day" or "month"

vars

columns to aggregate; NULL uses all covariate columns

method

one of "step", "linear", "spline"; length one for all variables, or a named vector per variable

extrapolate

hold the first and last values constant beyond the outermost period midpoints. With FALSE those half-periods are NA, since there is no second point to interpolate between.

Value

an sf POINT object with one row per cell per target step

What this does and does not do

As with downscale_grid(), this adds time steps rather than information. A monthly mean rendered daily still resolves nothing within the month.

There is a further trap specific to the time axis, which is why step is the default:

linear and spline do not preserve the period mean. Interpolate twelve monthly means to daily values and average those days back up, and you will not recover the months you started from. The interpolated series is a plausible smooth curve through the monthly values, not a disaggregation of them, and any budget or total computed from it will be off. step does preserve the mean, because every day in the month carries the month's own value.

  • step — every fine step takes its containing period's value. Preserves the period mean; discontinuous at period boundaries. Required for categorical columns, and applied to them automatically.

  • linear — straight lines between period midpoints. Continuous, and the usual choice when a series is going into a model as a smooth covariate.

  • spline — a natural cubic spline through the midpoints. Smoother than linear, and can overshoot past the source range between points, which for a bounded quantity like chlorophyll can produce negatives.

Where a period's value sits

linear and spline need each source value placed at a point in time, and a monthly mean is placed at the middle of its month rather than the first. Placing it at day 1 would shift the whole interpolated series half a month early, which is a systematic bias rather than a rounding difference.

See also

upscale_time() for the other direction, downscale_grid()

Examples

if (FALSE) { # \dontrun{
monthly <- accessEnvDat(vars = "SST", years = 2010, months = 1:12, bounding_box = bb)

# Daily steps, each carrying its month's value
daily <- downscale_time(monthly, to = "day")

# A smooth seasonal cycle instead, for a covariate going into a model
smooth <- downscale_time(monthly, to = "day", method = "spline")
} # }