Skip to contents

As the climate warms, species on mountains move upslope. tidydiversity 0.3.0 adds functions to describe where species sit on a gradient, to measure how far they moved between a historical survey and a resurvey, and to check whether a model predicted those moves.

Function What it does
range_edges() Lower and upper edges, midpoint, breadth and abundance-weighted optimum of each species, with flags for edges cut off by the ends of the transect
range_shift_resurvey() Shifts of each edge, midpoint and optimum between two surveys; species lost and gained
range_shift_test() Mean shift across species, bootstrap interval and exact sign test
range_shift_expected() Shift that would keep a species at the same temperature: warming / lapse rate
range_shift_validate() Scores predicted against observed shifts: bias, error, correlation, calibration and tracking

Ranges of frogs on a New Guinea mountain

png_frogs holds frog counts at eight elevations from 200 to 3,700 m on Mt Wilhelm (Dahl et al. 2024).

library(tidydiversity)
library(ggplot2)

edges <- range_edges(png_frogs, gradient = elevation, abundance = abundance)
edges
#> # A tibble: 55 × 10
#>    species lower upper midpoint breadth optimum n_sites abundance
#>    <chr>   <int> <int>    <dbl>   <int>   <dbl>   <int>     <int>
#>  1 FR01      200  1200      700    1000    296.       3       293
#>  2 FR02     1200  1700     1450     500   1330.       2        27
#>  3 FR03     2200  2200     2200       0   2200        1        32
#>  4 FR04     1700  1700     1700       0   1700        1       108
#>  5 FR05     2200  2200     2200       0   2200        1       137
#>  6 FR06     1700  2200     1950     500   1913.       2        47
#>  7 FR07     2700  2700     2700       0   2700        1        40
#>  8 FR08     1200  1200     1200       0   1200        1       133
#>  9 FR09     2700  2700     2700       0   2700        1        13
#> 10 FR10     2200  2700     2450     500   2690.       2       104
#> # ℹ 45 more rows
#> # ℹ 2 more variables: truncated_lower <lgl>, truncated_upper <lgl>
edges <- edges[order(edges$midpoint, edges$lower), ]
edges$species <- factor(edges$species, edges$species)
edges$cut_off <- edges$truncated_lower | edges$truncated_upper

ggplot(edges, aes(y = species)) +
  geom_segment(aes(x = lower, xend = upper, colour = cut_off), linewidth = 1.2) +
  geom_point(aes(x = optimum), size = 1.6) +
  scale_colour_manual(
    "Edge at the end\nof the transect",
    values = c(`FALSE` = "#2166AC", `TRUE` = "#B2182B")
  ) +
  labs(
    x = "Elevation (m)", y = NULL,
    title = "Elevational ranges of 55 frog species on Mt Wilhelm",
    subtitle = "Bars: lowest to highest record; dots: abundance-weighted optimum"
  ) +
  theme_minimal(base_size = 11) +
  theme(axis.text.y = element_text(size = 6), panel.grid.minor = element_blank())

Species whose edges lie at 200 m or 3,700 m may range beyond the transect. Their edges cannot show a shift, so range_shift_test() and range_shift_validate() leave them out by default.

How far should species move?

The shift that would keep a species at the same temperature is the warming divided by the lapse rate. With the 0.76 degrees C of warming measured on Mt Karimui between 1965 and 2012 and a lapse rate of 5.1 degrees C per km:

range_shift_expected(warming = 0.76, lapse_rate = 5.1)
#> [1] 149.0196

Did the models predict the climb?

With a historical survey and a resurvey, range_shift_resurvey() gives each species’ shift and range_shift_validate() scores predictions against it. The figure below summarises a hindcast of 136 bird species resurveyed on Mt Karimui, Papua New Guinea, in 1965 and 2012 (Freeman and Class Freeman 2014), using climate models from the companion package tidymapr. Tracking is the observed over the predicted mean shift: 1 means the birds kept pace with the warming.

With the regional warming estimate (0.39 degrees C) the birds seemed to outpace the climate by half again. With the warming measured on the mountain itself (0.76 degrees C) their upper limits tracked it: every interval based on local warming includes 1. The models predict the community’s mean shift well but each species’ shift only weakly.

The code pattern, for your own resurvey:

shifts <- range_shift_resurvey(survey_1965, survey_2012,
  gradient = elevation, abundance = count
)
range_shift_test(shifts, seed = 1)
range_shift_validate(shifts,
  data.frame(species = shifts$species, shift = range_shift_expected(0.76, 5.1)),
  seed = 1
)

References

Dahl, C., Richards, S. J., Basien, I., Mungkaje, A. J. and Novotny, V. (2024). Local and regional diversity of frog communities along an extensive rainforest elevation gradient in Papua New Guinea. Biotropica, 56, 90-97.

Freeman, B. G. and Class Freeman, A. M. (2014). Rapid upslope shifts in New Guinean birds illustrate strong distributional responses of tropical montane species to global warming. PNAS, 111, 4490-4494.