Skip to contents

Summarise what the observations actually contain

Source: R/summarise_obs.R
summarise_obs.Rd

Reports, for every observed variable: how many observations there are, when they span, how they are distributed across years and seasons, and whether they are depth profiles or single-depth samples.

The profile question is the one that usually decides whether a calibration is possible. A variable sampled hundreds of times can still be useless for constraining stratification if every one of those samples is a surface grab, and observation counts alone hide that completely.

Each variable is classified as one of

"profile"

at least half its sampling dates carry min_depths or more distinct depths - a depth-resolved record.

"discrete"

depths are recorded, but most visits carry fewer than min_depths of them - surface grabs, or surface/bottom pairs.

"scalar"

no depth information at all, e.g. water level.

The summary also carries the date range of every forcing series - met, inflows, outflows - because those, not the observations, are usually what limits how early a simulation can start. See suggest_sim_period(), which turns this summary into a runnable period.

Usage

summarise_obs(aeme, vars_sim = NULL, min_depths = 3L)

Arguments

aeme

Aeme object.

vars_sim

Character. Variables (var_aeme values) to summarise. Default NULL summarises every observed variable.

min_depths

Integer. Distinct depths a visit needs before it counts as a profile. Default 3L.

Value

An object of class aeme_obs_summary: a list with

variables

one row per variable - n_obs, n_dates, first, last, n_years, n_months (distinct calendar months, i.e. seasonal coverage), n_profiles, profile_frac, median_depths, depth_min, depth_max and kind.

years

one row per variable per calendar year - n_obs, n_dates, n_profiles.

forcing

one row per forcing series - source, first, last, n.

window

the date range over which every forcing series has data, i.e. the widest period the model can be run over.

See also

suggest_sim_period(), set_sim_period(), get_obs()

Examples

if (FALSE) { # \dontrun{
aeme <- readRDS(system.file("extdata/aeme.rds", package = "AEME"))
s <- summarise_obs(aeme)
s
s$variables
s$years[s$years$var_aeme == "HYD_temp", ]
} # }