Data from: Asymmetric species turnover underpins the latitudinal diversity gradient in the Western Ghats biodiversity hotspot

Latitudinal diversity gradients (LDGs), namely an increase in species
richness from higher to lower latitudes, are one of ecology’s best
documented patterns, yet how the fundamental components of compositional
mismatches (i.e., species gains and losses along a gradient) explain these
patterns remains poorly understood. As we argue here, one of two
alternative turnover scenarios must apply to a typical latitudinal
richness gradient. Such gradients can be the product of either a scenario
where both species gains and losses are correlated and increase with
geographical distance, or of uncorrelated change where species gains
increase with distance, but losses remain constant. We distinguish between
these alternatives using the freshwater fish communities of the Western
Ghats of India, a biodiversity hotspot with a strong latitudinal species
richness gradient. We find that the uncorrelated change scenario prevails
for both taxonomic and functional diversity, and on both the western and
eastern slopes of the mountain range, albeit with subtle differences. Our
results suggest that these uncorrelated changes result from high
nestedness among communities and reflect the dominance of dispersal
processes in the region. As the observed pattern arises due to an increase
in range-restricted species at lower latitudes, it is consistent with
Rapoport’s rule. Partitioning species turnover into its gain and loss
components thus helps explain the mechanisms underlying latitudinal
gradients of diversity, and informs our understanding of biodiversity
change.

# Data from: Asymmetric species turnover underpins the latitudinal
diversity gradient in the Western Ghats biodiversity hotspot Dataset DOI:
[10.5061/dryad.nk98sf86r](https://doi.org/10.5061/dryad.nk98sf86r) ##
Description of the data and file structure The folder contains R scripts
and data files used to investigate the latitudinal richness gradient by
measuring the underlying directional compositional changes in freshwater
fish communities in the Western Ghats. ### Main directory:
Beta_data_updated.zip #### **T_F_beta_diversity1.R** This script is used
to perform null model analysis, and calculate the directional measures of
taxonomic and functional beta-diversity. We calculated both the rough and
standardized effect size values of directional taxonomic and functional
beta-diversity. We also calculated standardized Jaccard replacement and
nestedness components by performing a null model analysis. Using this
script, we plotted the replacement and nestedness components across east-
and west-flowing basins. This script is also used to calculate geodesic
distance among basins using ‘*geodist*’ package. We used ‘adespatial’,
‘*mFD*’, and ‘*betapart*’ packages to perform the analysis. CSV files used
in the analysis: #### **beta_data_2_5_23.csv** This file contains basin by
species community data that is used to calculate the compositional
distances among communities. ##### Variables: * Basin: basins included in
the study * Column (2:326): Species presence (1) and absence (0) in basins
#### **env_15_7_24.csv** This file contains the basin-wise values for
environmental variables. ##### Variables: * Basin: 59 river basins
included in the study * flow.dir: The direction of the water flow- east
(1) & west (0) * position: Basins situated north of Palghat gap in
east (NPE) and west (NPW), and basins that are situated south of Palghat
gap in eastern (SPE) and western (SPW) side of the Escarpment. * flow:
Basins categorized based on whether a basin flows towards east or west *
gap: Basins categorized based on whether they are situated north (NP) or
south (SP) of Palghat gap                   * SpRich: Observed species
richness in each basin              * EndRich: Observed endemic species
richness in each basin   * area: Total area of a basin in km^2^ *
str.dens: Stream density        * soc: Soil organic carbon in
tonnes/hectare             * dis.pyr : Annual average water discharge
(cubic meters/second) * soil.div: Soil diversity (Shannon index) *
cover.div: Land cover diversity          * classarea: Percentage of basin
area above 1000-meter altitude             * alt.mean: Mean elevation
(meter)    * alt.min: Minimum elevation (meter) * alt.max: Maximum
elevation (meter)            * alt.range: elevation range  (meter)    *
HI.index: Hypsometric index              * slope.mean: Mean slope (in
degrees)          * rad.cv:  Solar radiation annual variability * rad.sum:
Overall annual sum of solar radiation (KJ/m^2^/day) * aet.sum: Overall
annual sum of actual evapotranspiration (mm)    * aet.cv: Overall annual
variability of actual evapotranspiration * prec.bio12: Precipitation
(annual mean; in millimetres)   * prec.cv.bio15: Precipitation annual
variability (seasonality)              * t.mean.bio1: Temperature (annual
mean; in degree celsius)       * t.range.bio7: Temperature (annual
range)   * varbio1: Quaternary climatic stability (annual mean
temperature; in degree censius)  * varbio7: Quaternary climatic stability
(temperature variability)       * varbio12: Quaternary climatic stability
(annual mean precipitation; in millimeters) * varbio15: Quaternary
climatic stability (precipitation variability)       * dor.pva: Degree of
regulation (in %) * ppd.sav: Population density (people per km^2^)        
* rdd.sav: Road density (meters per km^2^)       * urb.sse: Urban extent
(in %) * pac.sse: Protected area (in %)           * xcoord: Centroid
location of a basin (longitude in ˚E). * ycoord: Centroid location of a
basin (latitude in ˚N). * Perc_800: Percentage of basin area above
800-meter altitude                        * Column logArea to column
arcsinHI: Transformed variables (log or arcsine)       * Column logAreaS
to column slope.meanS: Standardized variables (z score; mean=0, standard
deviation=1) * clim1, clim2, energy1, energy2, iso1, iso2, history1,
history2: Variables derived from PCA analysis conducted on all climate,
all energy, all isolation and all Quaternary history variables.       
#### **traitafterimpute_group1.csv** This file contains functional traits
for 325 species that are found in the Western Ghats. ##### Variables: *
Species: Species found in the Western Ghats region.       * Endemism:
Assigned endemic status to each species based on Dahanukar *et al*.
(2004). * genus: Genus of the species * family: Family of the species *
West: Number of basins a species occupies in west * East: Number of basins
a species occupies in east            * Presence: Presence of species in
east, west or in both sides of the escarpment        *
Presence_end_status: Presence of endemic (end) and non-endemic (ne)
species in east, west or in both sides of the escarpment.         *
body_form: Species are categorized according to their body forms (leaf
fish, carp, loach, carp-minnow, glassfish, snakehead, etc.).         *
body_form_major: Species are categorized according to their major body
forms (e.g., carp, carp-minnow, catfish, sucker catfish, snakehead, leaf
fish, loach, etc.)  * MBL: Maximum body length (in cm) * MBL_log: log
transformed maximum body length * BEL: Body elongation              * VEP:
Vertical eye position     * RES: Relative eye size             * OGP: Oral
gap position          * BLS: Body lateral shape * PFV: Pectoral fin
vertical position * PFS: Pectoral fin size * CPT: Caudal peduncle
throttling #### **nullresults_function.rds** This datafile contains the
basin-wise functional beta-diversity components (i.e. gains and losses)
that were generated by permuting the species×basin matrix for 99 times
using the ‘*independent swap’* algorithm. ####
**nullresults_taxonomic.rds** This datafile contains the basin-wise
taxonomic beta-diversity components (i.e. gains and losses) that were
generated by permuting the species×basin matrix for 99 times using the
‘*independent swap*’ algorithm. #### **nullresults_partition.rds** This
datafile contains the basin-wise Jaccard turnover and nestedness diversity
components that were generated by permuting the species×basin matrix for
999 times using the ‘*frequency*’ algorithm. #### **distance_decay1.R**
This script is used to perform the distance-decay modelling. We fitted
distance-gain/loss models along with distance-dissimilarity models using
the *decay.model* function from ‘betapart’ package. These models are
plotted using *plot.decay* function and two models were compared using the
*zdep* function of the ‘betapart’ package. This script was also used to
perform the Mantel tests. Packages ‘*Ternary*’ and ‘*ggplot2*’ are used to
plot the ternary plots. CSV files used in the analysis: ####
**dist_a.ee.csv** This square matrix (24 * 24) of geodesic distances (in
meters) among all east-flowing basins is used to fit distance-overlap,
distance-gains, distance-loss, and distance-β-diversity models. ####
**dist_a.ww.csv** This square matrix (35 * 35) of geodesic distances (in
meters) among all west-flowing basins is used to fit distance-overlap,
distance-gains, distance-loss and distance-β-diversity models. ####
**dist_a.e_w.csv** This square matrix (59 * 59) of geodesic distances (in
meters) among all east-to-west-flowing basins is used to fit
distance-overlap, distance-gains, distance-loss, and distance-β-diversity
models. #### **dist_a.w_e.csv** This square matrix (59 * 59) of geodesic
distances (in meters) among all west- to east-flowing basins is used to
fit distance-overlap, distance-gains, distance-loss and
distance-β-diversity models. #### **over.a.ee.csv** This square matrix (24
* 24) of species overlap among all east-flowing basins is used to fit
distance-overlap, and distance-β-diversity models. #### **over.a.ww.csv**
This square matrix (35 * 35) of species overlap among all west-flowing
basins is used to fit distance-overlap, and distance-β-diversity models.
#### **over_a.e_w.csv** This square matrix (59 * 59) of species overlap
among all east-to-west-flowing basins is used to fit distance-overlap, and
distance-β-diversity models. #### **over_a.w_e.csv** This square matrix
(59 * 59) of species overlap among all west- to east-flowing basins is
used to fit distance-overlap, and distance-β-diversity models. ####
**fe.over.a.ee.csv** This square matrix (24 * 24) of functional overlap
among all east-flowing basins is used to fit distance-overlap, and
distance-β-diversity models. #### **fe.over.a.ww.csv** This square matrix
(35 * 35) of functional overlap among all west-flowing basins is used to
fit distance-overlap, and distance-β-diversity models. ####
**fe.over.a.e_w.csv** This square matrix (59 * 59) of functional overlap
among all east to west-flowing basins is used to fit distance-overlap, and
distance-β-diversity models. #### **fe.over.a.w_e.csv** This square matrix
(59 * 59) of functional overlap among all west- to east-flowing basins is
used to fit distance-overlap, and distance-β-diversity models. ####
**gain.a.ee.csv** This square matrix (24 * 24) of species gains among all
east-flowing basins is used to fit distance-gains models. ####
**gain.a.ww.csv** This square matrix (35 * 35) of species gains among all
west-flowing basins is used to fit distance-gains models. ####
**gain.a.e_w.csv** This square matrix (59 * 59) of species gains among all
east-to-west-flowing basins is used to fit distance-gains models. ####
**gain.a.w_e.csv** This square matrix (59 * 59) of species gains among all
west- to east-flowing basins is used to fit distance-gains models. ####
#### **fe.gain.a.ee.csv** This square matrix (24 * 24) of functional gains
among all east-flowing basins is used to fit distance-gains models. ####
**fe.gain.a.ww.csv** This square matrix (35 * 35) of functional gains
among all west-flowing basins is used to fit distance-gains models. ####
**fe.gain.a.e_w.csv** This square matrix (59 * 59) of functional gains
among all east-to-west-flowing basins is used to fit distance-gains
models. #### **fe.gain.a.w_e.csv** This square matrix (59 * 59) of
functional gains among all west- to east-flowing basins is used to fit
distance-gains models. #### **loss.a.ee.csv** This square matrix (24 * 24)
of species losses among all east-flowing basins is used to fit
distance-loss models. #### **loss.a.ww.csv** This square matrix (35 * 35)
of species losses among all west-flowing basins is used to fit
distance-loss models. #### **loss.a.e_w.csv** This square matrix (59 * 59)
of species losses among all east-to-west-flowing basins is used to fit
distance-loss models. #### **loss.a.w_e.csv** This square matrix (59 * 59)
of species losses among all west-to-east-flowing basins is used to fit
distance-loss models. #### **fe.loss.a.ee.csv** This square matrix (24 *
24) of functional losses among all east-flowing basins is used to fit
distance-loss models. #### **fe.loss.a.ww.csv** This square matrix (35 *
35) of functional losses among all west-flowing basins is used to fit
distance-loss models. #### **fe.loss.a.e_w.csv** This square matrix (59 *
59) of functional losses among all east-to-west-flowing basins is used to
fit distance-loss models. #### **fe.loss.a.w_e.csv** This square matrix
(59 * 59) of functional losses among all west- to east-flowing basins is
used to fit distance-loss models. #### **taxo_SES_beta.ee._1_4_26.csv**
This square matrix (24 * 24) of standardised taxonomic dissimilarity among
all east-flowing basins is used to fit distance-dissimilarity models. ####
**taxo_SES_beta.ww._1_4_26.csv** This square matrix (35 * 35) of
standardised taxonomic dissimilarity among all west-flowing basins is used
to fit distance-dissimilarity models. #### **SES_beta.ee._1_4_26.csv**
This square matrix (24 * 24) of standardised functional dissimilarity
among all east-flowing basins is used to fit distance-dissimilarity models
#### **SES_beta.ww.1_4_26.csv** This square matrix (35 * 35) of
standardised functional dissimilarity among all west-flowing basins is
used to fit distance-dissmilarity models. #### **Range_analysis.R** This
script is used to perform range size analysis in freshwater fishes of the
Western Ghats. In first steps, latitudinal range sizes (latitude in ˚N)
were calculated for each species. Then, the decline in species range size
was assessed by constructing the 3 degrees bins from northern to southern
WG. ‘*dplyr*’, ‘*ggplot2*’, and ‘*ggbeeswarm*’ packages were used to
perform the range-size analysis and plotting. #### **functional_alpha.R**
This script is used to calculate the functional α-diversity and plot the
relationship between latitude vs taxonomic and functional richness.
‘*mFD*’ package was used to calculate the functional α-diversity.
‘*ggplot2*’ is used for plotting. CSV files used in the analysis: ####
**TDFD.csv** This file is used to run the linear models between latitude
and taxonomic and functional α-diversity. ##### Variables: * flow.dir: The
direction of the water flow- east (1) & west (0) * position: Basins
situated north of Palghat gap in east (NPE) and west (NPW), and basins
that are situated south of Palghat gap in eastern (SPE) and western (SPW)
side of the Escarpment. * flow: Basins categorized based on whether a
basin flows towards east or west * gap: Basins categorized based on
whether they are situated north (NP) or south (SP) of Palghat
gap                   * SpRich_new: Observed species richness in each
basin   * EndRich_new: Observed endemic species richness in each basin   *
xcoord: Centroid location of a basin (longitude in ˚E). * ycoord: Centroid
location of a basin (latitude in ˚N). * fun.richness: Observed functional
α-diversity * fun_rich_ses: Standardized functional alpha diversity ##
Code/software All analyses were done in R (R version 4.4.2)