Skip to contents

Plot temporal trends in plankton data with fitted lines

Source: R/plot_timeseries.R
pr_plot_Trends.Rd

Create time series plots with fitted trend lines to examine long-term changes, seasonal patterns, or interannual variability in plankton indices. The function can plot raw data over time, monthly climatologies, or annual means.

Usage

pr_plot_Trends(df, Trend = "Raw", method = "lm", trans = "identity")

Arguments

df

A dataframe from pr_get_Indices() containing timeseries data

Trend

The temporal scale for trend analysis:

  • "Raw" - Plot all data points over time with a trend line through the full timeseries

  • "Month" - Monthly climatology showing seasonal patterns (averaged across years)

  • "Year" - Annual means showing interannual variability

method

Smoothing method for the trend line. Options include:

  • "lm" - Linear regression (default, good for detecting long-term trends)

  • "loess" - Local polynomial regression (good for non-linear trends)

  • "gam" - Generalised additive model (requires mgcv package)

  • Any other method accepted by ggplot2::geom_smooth()

trans

Transformation for the y-axis scale:

  • "identity" - No transformation (default)

  • "log10" - Log base 10 transformation (useful for abundance data)

  • "sqrt" - Square root transformation

  • Any other transformation accepted by ggplot2::scale_y_continuous()

Value

A ggplot2 object showing the timeseries with fitted trend line(s)

Details

This function is designed to help identify temporal trends in plankton data:

  • Raw trends: Shows all observations over time with a smoothed trend line. Useful for detecting long-term changes (e.g., increasing or decreasing abundance).

  • Monthly trends: Aggregates data by month across all years to show seasonal patterns (e.g., winter blooms, summer stratification effects).

  • Annual trends: Shows year-to-year variability, useful for detecting regime shifts or responses to climate oscillations (e.g., ENSO).

The shaded ribbon around trend lines represents the 95% confidence interval. For NRS data, values are averaged across depths if present.

See also

pr_plot_TimeSeries() for plots without trend lines, pr_plot_Climatology() for alternative climatology visualisations, pr_get_model() for extracting model coefficients

Examples

# Examine long-term trends in zooplankton biomass
df <- pr_get_Indices("NRS", "Zooplankton") %>%
  dplyr::filter(Parameters == "Biomass_mgm3") %>%
  pr_model_data()
pr_plot_Trends(df, Trend = "Raw", method = "lm")
#> Warning: Removed 61 rows containing non-finite outside the scale range
#> (`stat_smooth()`).
#> Warning: Removed 63 rows containing missing values or values outside the scale range
#> (`geom_point()`).


# Examine seasonal patterns
pr_plot_Trends(df, Trend = "Month", method = "loess")
#> Warning: Removed 17 rows containing missing values or values outside the scale range
#> (`geom_point()`).


# Examine interannual variability
pr_plot_Trends(df, Trend = "Year", method = "lm")
#> Warning: Removed 8 rows containing non-finite outside the scale range (`stat_smooth()`).
#> Warning: Removed 9 rows containing missing values or values outside the scale range
#> (`geom_point()`).


# Use log transformation for abundance data
df <- pr_get_Indices("CPR", "Phytoplankton", near_dist_km = 250) %>%
  dplyr::filter(Parameters == "PhytoAbundance_Cellsm3",
                BioRegion %in% c("South-east", "Temperate East"))
pr_plot_Trends(df, Trend = "Raw", method = "lm", trans = "log10")
#> Warning: Removed 15180 rows containing non-finite outside the scale range
#> (`stat_smooth()`).
#> Warning: Removed 15180 rows containing missing values or values outside the scale range
#> (`geom_point()`).