Plot climate anomaly time series — cd_plot

Creates a bar chart of anomalies over time with optional Theil-Sen trend lines. Positive and negative anomalies are colored differently.

Usage

cd_plot_timeseries(
  x,
  variable = NULL,
  period = "annual",
  trend = NULL,
  title = NULL,
  colors = c(pos = "#d73027", neg = "#4575b4")
)

Arguments

x: A tibble from cd_anomaly() with columns variable, period, year, anomaly, optionally anomaly_type, unit and long_name. Also works with cd_extract() output (uses value column). One row per year in the plotted series.
variable: Character. Which variable to plot. Default uses the first variable in x.
period: Character. Which period to plot. Default "annual".
trend: Optional tibble from cd_trend() to overlay trend lines. Only rows whose trend_on names the plotted column (anomaly or value) are drawn, so a table holding trends on both scales draws only the one that fits the bars, and warns when none does. Rows with no trend_on, or NA, are drawn whatever the plotted column (unlike cd_summary(), which reads them as anomaly trends). The earliest trend_start is drawn dashed, later ones solid.
title: Optional plot title.
colors: Named character vector of length 2 for positive/negative bar colors. Default c(pos = "#d73027", neg = "#4575b4").

Value

A ggplot2::ggplot object.

Details

The y-axis label is long_name and unit from x where present and not NA, otherwise from cd_variables() — resolved by the same rules as cd_summary() (see the input contract in cd_anomaly()) — else the plotted column's name. unit is the anomaly's unit, so on raw value input it is shown only where the anomaly type is absolute or pct_point_diff, whose anomaly unit is the unit of the values. Unlike cd_summary(), a long_name shared by several variables is not suffixed with the variable: the plot shows one variable, so name a station in title.

Examples

if (FALSE) { # \dontrun{
catalog <- cd_catalog()
aoi <- sf::st_read("my_aoi.gpkg")
ts <- cd_extract(catalog, aoi, variables = "tmean", periods = "annual")
bl <- cd_baseline(ts, baseline_years = 1951:1980)
ano <- cd_anomaly(ts, bl)
trn <- cd_trend(ano, trend_start = c(1951, 1981))
cd_plot_timeseries(ano, trend = trn)
} # }