If species' ranges are placed at random within a bounded gradient, they
overlap most in the middle, so richness peaks at mid-gradient even with
no environmental cause: the mid-domain effect (Colwell and Hurtt
1994; Colwell and Lees 2000). range_mde() builds this null model from
the observed ranges, in the spirit of RangeModel (Colwell 2008), and
compares it with the observed richness at each site.
Each species keeps its observed range size (the number of consecutive
sampled sites it spans, after range_through() interpolation), and its
position is randomised:
model = "uniform": every position where the range fits is equally likely (Colwell and Hurtt 1994). The expected richness is computed exactly; the interval comes fromnsimrandomisations.model = "spread": the range starts at a random site and spreads to random neighbouring sites (the "spreading dye" model of Jetz and Rahbek 2001, as in the rangemodelR package), which places more ranges near the ends than the uniform model.
Sites are treated as positions in order along the gradient, so the model assumes the sampled sites represent the domain evenly.
Arguments
- data
A data frame in long format, with one row per record.
- gradient
Column with the position of each site on the gradient (one value per site).
- site, species
<
tidy-select> The column holding site identifiers and the column holding species names.- model
"uniform"or"spread".- nsim
Number of randomisations.
- conf
Width of the simulation interval.
- interpolate
Apply
range_through()first (recommended).- .by
<
tidy-select> Optional grouping columns, kept in the output. Grouped data frames are also respected.- x
A
range_mde()result.- ...
Not used.
Value
A tibble of class td_mde with the grouping columns, the site
and gradient columns, observed richness, expected richness under
the null model, its sd, lower and upper simulation limits, and
outside ("above", "below" or "within"). Attributes r2 (the
variance in observed richness explained by the null model, one value
per group) and p_value (a global test: the proportion of
randomisations that deviate from the expectation at least as much as
the observed richness) summarise the fit; see glance().
References
Colwell, R. K. and Hurtt, G. C. (1994). Nonbiological gradients in species richness and a spurious Rapoport effect. American Naturalist, 144, 570-595. doi:10.1086/285695
Colwell, R. K. and Lees, D. C. (2000). The mid-domain effect: geometric constraints on the geography of species richness. Trends in Ecology & Evolution, 15, 70-76. doi:10.1016/S0169-5347(99)01767-X
Colwell, R. K. (2008). RangeModel: tools for exploring and assessing geometric constraints on species richness (the mid-domain effect) along transects. Ecography, 31, 4-7. doi:10.1111/j.2008.0906-7590.05347.x
Jetz, W. and Rahbek, C. (2001). Geometric constraints explain much of the species richness pattern in African birds. Proceedings of the National Academy of Sciences, 98, 5661-5666. doi:10.1073/pnas.091100998
Examples
set.seed(1)
mde <- range_mde(png_frogs, gradient = elevation, nsim = 199)
mde
#> # A tibble: 8 × 8
#> site elevation observed expected sd lower upper outside
#> <chr> <int> <dbl> <dbl> <dbl> <dbl> <dbl> <chr>
#> 1 MW0200 200 13 7.65 2.65 4 14 within
#> 2 MW0700 700 13 11.8 2.73 6.95 17 within
#> 3 MW1200 1200 17 13.7 3.13 8 21 within
#> 4 MW1700 1700 18 13.9 3.15 9 21.0 within
#> 5 MW2200 2200 14 13.9 3.11 8.95 21 within
#> 6 MW2700 2700 12 13.7 3.03 8 20 within
#> 7 MW3200 3200 5 11.8 2.76 6.95 17 below
#> 8 MW3700 3700 2 7.65 2.59 3 13 below
glance(mde)
#> # A tibble: 1 × 5
#> model n_sites n_species r2 p_value
#> <chr> <int> <int> <dbl> <dbl>
#> 1 uniform 8 55 0.354 0.055