The authors have declared that no competing interests exist.
Designed the software used in analysis: SL. Reviewed and commented upon draft versions of the manuscript: SL ME. Conceived and designed the experiments: GD SL ME. Analyzed the data: GD. Contributed reagents/materials/analysis tools: SL. Wrote the paper: GD.
We quantify spatial turnover in communities of 1939 plant and 59 mammal species at 2.5 km resolution across a topographically heterogeneous region in south-eastern Australia to identify distributional breaks and low turnover zones where multiple species distributions overlap. Environmental turnover is measured to determine how climate, topography and geology influence biotic turnover differently across a variety of biogeographic breaks and overlaps. We identify the genera driving turnover and confirm the versatility of this approach across spatial scales and locations.
Directional moving window analyses, rotated through 360°, were used to measure spatial turnover variation in different directions between gridded cells containing georeferenced plant and mammal occurrences and environmental variables. Generalised linear models were used to compare taxic turnover results with equivalent analyses for geology, regolith weathering, elevation, slope, solar radiation, annual precipitation and annual mean temperature, both uniformly across the entire study area and by stratifying it into zones of high and low turnover. Identified breaks and transitions were compared to a conservation bioregionalisation framework widely used in Australia.
Detailed delineations of plant and mammal turnover zones with gradational boundaries denoted subtle variation in species assemblages. Turnover patterns often diverged from bioregion boundaries, though plant turnover adhered most closely. A prominent break zone contained either comparable or greater numbers of unique genera than adjacent overlaps, but these were concentrated in a small subsection relatively under-protected by conservation reserves. The environmental correlates of biotic turnover varied for different turnover zones in different subsections of the study area. Topography and temperature showed much stronger relationships with plant turnover in a topographically complex overlap, relative to a lowland overlap where weathering was most predictive. This method can quantify transitional turnover patterns from small to broad extents, at different resolutions for any location, and complements broad-scale bioregionalisation schemes in conservation planning.
Biogeographic breaks and transition zones have received considerable investigation at global and continental scales at coarse resolutions (e.g.
It is increasingly pertinent to examine biogeographic phenomena at finer resolutions and at regional to local scales. This tests the generality of concepts derived by investigations over broad extents, as patterns and processes may not exhibit scale invariance
Fine-scale quantification of biogeographic breaks also enables depiction of transitional zones in greater detail, facilitating their comparison with subtle variation in environmental gradients and potentially clarifying how different physical factors regulate species distributions across break transitions. This is useful for conservation planning because taxa occurring in transitional areas are often the most adaptive to a range of environmental conditions
Quantifying the rate of species turnover across a landscape has been used to identify transitional turnover patterns and biogeographic breaks for bees
This research maps breaks and transitions in multiple plant and mammal species distributions at 2.5 km resolution across the south-east coast of New South Wales (NSW) in Australia, which is a smaller subsection of the same region previously examined by Di Virgilio et al.
Secondly, Di Virgilio et al.
Previous studies of trees in south-eastern NSW have shown the importance of variables such as climate, lithology and topographic position in predicting the richness of several species groups, e.g.
Our results are interpreted in reference to the Interim Biogeographic Regionalisation for Australia version 6.1 (IBRA 6.1)
The South East Corner (SEC) bioregion defined in IBRA 6.1
Inset map shows the location of the South East Corner (SEC) Interim Biogeographic Regionalisation of Australia (IBRA) bioregion, which straddles New South Wales (NSW) and Victoria (VIC) in south-eastern Australia. Main map shows the part of SEC in coastal NSW that comprises the study area and its neighbouring IBRA bioregions in NSW (South-Eastern Highlands – SEH; Sydney Basin - SB). The white dashed line shows a 0.2° buffer which captured biotic and abiotic data around SEC in NSW for inclusion in moving window analyses of turnover.
SEC has a mesothermal climate, with long mild summers and little variation in annual rainfall. Average monthly temperatures range from minima of −3.5–8.4°C to maxima of 19.2–28.8°C. Average monthly rainfall minima and maxima are 29–102 mm and 58–155 mm respectively. SEC’s terrain is heterogeneous, comprising coastal lowlands extending ∼15 km inland to the southern highlands to the west which rise to ∼1250 m.
A 0.2° (∼22 km) buffer was added around SEC (dashed white line in
The geographic locations of plants and mammals sampled within SEC plus the buffer were derived from the NSW Office of Environment and Heritage’s (OEH) Atlas of NSW Wildlife
The plant database contained 104,030 records, comprising 145 families, 595 genera and 1939 species and included a wide variety of different trees, shrubs, ferns, tussocks, herbs, grasses and other flowering plants. There were 41,392 mammal records (38 genera, 59 species). The mean positional accuracy of the unaltered Atlas mammal data is 1265 m with a standard deviation of 3391 m. Plant observations have a mean accuracy of 1137 m and standard deviation of 3451 m.
Elevation data was derived from the three arc-second Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM)
The spatial analyses adapted those described by Di Virgilio et al.
All biotic and abiotic data were aggregated to 0.0225°×0.0225° cells (∼2.5 km×2.5 km) by importing each data set into B
The Sørensen index was used to quantify the spatial turnover of biotic and lithological data:
To investigate which genera are exclusive to discrete areas of high or low turnover, we identified the plant and mammal genera unique to each sample set and common to both. Once gradational areas of high and low plant turnover were identified via the moving window analyses, the outline of a discrete, high turnover ‘break’ zone was assigned to the sample one set and the perimeter of one of two discrete low turnover ‘overlap’ zones was assigned to the sample two set. The plant or mammal genera common to both sets and those exclusive to either sample set one or sample set two were then identified by comparing the genera in the break with those in one of the overlap areas.
Since the topographic, climatic and weathering variables are interval scaled data, the moving window analyses used a numeric dissimilarity index (eq. 2), which compared the set of numeric values in each sample set by calculating the mean absolute difference between all pairs of values in samples 1 and 2
We also calculated species richness and sample redundancy (eq. 3) for plants and mammals for each moving window. Richness is an important component of species turnover
The mosaic of cells comprising a moving window were divided into one of two orthogonal, semi-elliptical neighbourhood sets (the sample one and two sets described above) and used the same geometric specification as the moving windows described by Di Virgilio et al.
The same window sizes were used for flora and mammals to aid comparison of turnover in each group. Since we aim to find the environmental correlates of species turnover, the same dimensions were used for moving window analyses of the environmental variables.
To gauge how biotic and abiotic turnover varies in different directions, moving window analyses were repeated in 25 increments of 15°; hence the window was rotated through 24 different angles from 0° to 360°.
We investigated how environmental variables influenced species turnover across the whole study area and also how their relative influence on break and overlap formation varied once the study area was stratified into three different areas of high or low turnover (see Results for details of the zones/stratification). Generalized linear models (GLM;
The correlograms revealed a range of ∼0.2° for flora, with a sill value of ∼0.92, where the range is the distance on the x-axis at which the median turnover values begin to plateau and is the maximum distance to which there is spatially structured turnover (
The species group correlograms for (A) flora and (B) mammal turnover which were used to calibrate the dimensions of moving window analyses of biotic and abiotic compositional dissimilarity. The correlograms are depicted as boxplots showing the distribution of turnover values between pairs of cells over geographic distance. Hence, for each boxplot, the black horizontal bar denotes the median turnover, the top and bottom of each box represent the 75th and 25th percentiles, respectively. The whiskers represent the minimum and maximum of all the data at each distance. Circles represent outliers. The distance on the x-axis at which the median turnover values begin to plateau is the maximum distance to which there is spatially structured turnover.
Plant richness was high across much of the SEC bioregion (
Plant (panels A–D) and mammal (panels E–H) richness and sample redundancy patterns in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia, for the 90° and 0° window orientation analyses. Panel A labels the main Interim Biogeographic Regionalisation of Australia (IBRA) subregions of SEC. The arrows (not drawn to scale) in the top left corner of each panel represent the overall orientation of the moving window used for that panel. Note that the arrow base shows the orientation of neighbour set 1, with the arrow indicating the direction of neighbour set 2.
A greater proportion of SEC showed high mammal species richness (
Richness patterns for both biotic groups changed slightly according to the orientation of the moving window because different combinations of cells are considered by different window orientations. For example, the shape of the northern low plant and mammal richness zones changed from the 90° to 0° windows, but their overall form was still consistent across orientations.
Floristic sampling redundancy (
Plant and mammal species turnover maps in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia, for the Sørensen moving window analyses at bearing 90°. The plant turnover map labels discrete zones of high and low turnover. The mammal turnover map also shows the Interim Biogeographic Regionalisation of Australia (IBRA) subregions comprising this part of SEC and some of the principal towns.
There was continuous variation in species turnover patterns with several breaks, overlaps and transition zones shown in greater detail than the 5 km resolution analyses of Di Virgilio et al.
A western low turnover area to the south-west of the break overlies high, topographically complex terrain separated from the break by a small valley (
The outlines of the northern break zone, the hilly western overlap and the coastal eastern overlap are sketched in
Approximate extents of a discrete high plant turnover zone (‘Northern high turnover zone’, black dotted line) and low plant turnover zones (‘Eastern low turnover zone’, black solid line and ‘Western low turnover zone’, grey dashed line) identified via moving window analyses of plant turnover in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia.
Mammal turnover was lower across SEC and less sharply defined than plant turnover (
Biotic turnover was not as smooth as the corresponding richness and redundancy patterns. Breaks in mammal distributions appeared in both high and low redundancy areas and across extremes of richness. In contrast, the main portion of the floristic high turnover zone corresponds to areas of low to medium sampling redundancy and richness.
The northern break zone and the smaller eastern overlap contained similar numbers of plant species and genera, although eight times as many observations were sampled in the overlap (
Turnover Zone | No. Observations | No. Genera | No. Species |
|
|||
Plants | 2385 | 348 | 740 |
Mammals | 3961 | 34 | 47 |
|
|||
Plants | 18610 | 334 | 701 |
Mammals | 3874 | 33 | 46 |
|
|||
Plants | 5803 | 265 | 521 |
Mammals | 349 | 26 | 33 |
Species Group | No. Genera Common to BothRegions (Both Neighbour Sets:List A Genera) | No. Genera Unique to Break (Neighbour Set 1: List B Genera) | No. Genera Unique to Overlap (Neighbour Set 2: List C Genera) |
|
|||
Plants | 270 | 79 | 79 |
Mammals | 32 | 2 | 1 |
|
|||
Plants | 214 | 135 | 57 |
Mammals | 26 | 8 | 0 |
A greater variety of plant genera and species were sampled in the northern break relative to the western overlap (
Similar numbers of mammal observations, genera and species were sampled in the break and eastern overlap (
Generally, environmental turnover did not correspond to the prominent biotic turnover zones (
Plant and abiotic turnover maps in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia, for the Sørensen and numeric dissimilarity moving window analyses at bearings showing the strongest correlative relationship between environmental and plant turnover. Turnover maps show Interim Biogeographic Regionalisation of Australia (IBRA) bioregion boundaries.
Topographic and temperature turnover was substantially more predictive of variation in mammal turnover than for plants, as shown by the maximal
Correlates | ||||
Abiotic | Biotic | R2 | β1 |
|
Geology | Flora | 0.015 | 0.123 | 5.386 |
Mammals | 0.015 | 0.133 | 5.348 | |
Weathering | Flora | 0.029 | −0.276 | −7.584 |
Mammals | 0.023 | −0.312 | −6.931 | |
Elevation | Flora | 0.035 | 0.260 | 8.635 |
Mammals | 0.249 | 0.698 | 25.613 | |
Slope | Flora | 0.055 | 0.346 | 10.936 |
Mammals | 0.246 | 0.771 | 25.769 | |
Solar Radiation | Flora | 0.016 | 0.161 | 5.808 |
Mammals | 0.247 | 0.622 | 25.467 | |
Annual Precipitation | Flora | 0.028 | 0.322 | 7.766 |
Mammals | 0.068 | 0.469 | 12.023 | |
Annual Mean Temperature | Flora | 0.039 | 0.236 | 8.779 |
Mammals | 0.247 | 0.698 | 25.446 |
GLM regressions between biotic and environmental turnover in the stratified study area produced the maximal
Correlates | Northern Break Zone | Eastern Overlap | Western Overlap | |||||||
Abiotic | Biotic | R2 | β1 |
|
R2 | β1 |
|
R2 | β1 |
|
Geology | Flora | 0.084 | 0.070 | 6.467 | 0.041 | 0.268 | 2.762 | 0.099 | 0.504 | 3.848 |
Mammals | 0.093 | 0.438 | 6.875 | 0.077 | 0.267 | 3.851 | 0.070 | 0.297 | 2.706 | |
Weathering | Flora | 0.033 | 0.083 | 4.122 | 0.112 | 0.354 | 4.770 | 0.156 | 0.487 | 4.662 |
Mammals | 0.048 | 0.263 | 4.744 | 0.088 | 0.337 | 4.177 | 0.089 | − |
−2.884 | |
Elevation | Flora | 0.014 | 0.069 | 2.673 |
|
− |
− |
0.183 | − |
− |
Mammals | 0.288 | 0.826 | 13.667 | 0.288 | 0.800 | 8.510 | 0.076 | 0.399 | 3.194 | |
Slope | Flora | 0.073 | 0.107 | 6.333 | 0.066 | − |
− |
0.189 | 0.519 |
|
Mammals | 0.245 | 0.696 | 12.238 | 0.280 | 0.662 | 8.333 | 0.175 | 0.709 | 4.557 | |
Solar Radiation | Flora | 0.027 | 0.059 | 3.743 | 0.096 | − |
− |
0.106 | − |
− |
Mammals | 0.261 | 0.647 | 12.753 | 0.223 | 0.610 | 7.185 | 0.141 | 0.483 | 4.519 | |
Precipitation | Flora | 0.034 | 0.092 | 4.184 | 0.080 | 0.435 | 3.947 | 0.076 | 0.405 | 3.347 |
Mammals | 0.196 | 0.633 | 10.422 | 0.160 | 0.653 | 5.833 | 0.128 | 0.576 | 4.276 | |
Temperature | Flora |
|
− |
− |
0.039 | − |
− |
0.168 | − |
− |
Mammals | 0.273 | 0.816 | 13.114 | 0.285 | 0.739 | 8.445 | 0.076 | 0.324 | 3.036 |
Several differences between the biotic-environmental relationships emerged within the different turnover zones. Topographic and climatic turnover were more strongly predictive of mammal turnover in the northern break and eastern overlap compared to the western overlap, whilst geology and weathering turnover showed comparatively weaker associations with mammal turnover across all three zones. In comparison to mammals, environmental turnover had limited predictive capacity for variation in plant turnover, except for in the western overlap where weathering, elevation, slope and temperature turnover were substantially more predictive. Weathering turnover was also much more predictive of plant turnover than the other variables in the eastern overlap.
Most regression slopes were positive, but there were some negative relationships between biotic and environmental turnover. These occurred mainly in the two overlap zones, with only one negative slope in the northern break, and applied most consistently to relationships between plant, elevation, solar radiation and temperature turnover.
This research aimed to identify high resolution biogeographic breaks and transitions in multi-taxon distributions, the specific genera driving break formation and the environmental role in generating these patterns using a study area in south-eastern Australia as an example. Biotic and environmental turnover patterns in this region were depicted in substantially greater detail than was achieved by Di Virgilio et al.
The apparent breaks and overlaps do not appear closely linked to differing land-use patterns, except for an area of cleared/disturbed land which corresponds with the location of the southern extension of high plant turnover (
Plant species turnover maps overlaid with the locations of cleared/disturbed land (left panel) and national parks, reserves and New South Wales (NSW) state forests (right panel).
We also quantified sample redundancy across SEC to reconcile the turnover patterns with the effects of poor sampling. Plant sampling redundancy ranges from low to medium levels in the north of the study area (
Despite some fine-scale similarities, overall the environmental turnover maps lacked consistent parallels with biotic turnover. For instance, high and low environmental turnover corresponded to separate areas of low biotic turnover. This appears to support the contention that greater environmental diversity may not always represent increased species richness
Generalised linear model (GLM) regressions between biotic and abiotic turnover across SEC were much more predictive of mammal turnover than for plants. The positive relationships between most physical variables and biotic turnover suggest an intrinsic link between higher rates of environmental turnover and higher biotic turnover across SEC, although the opposite trend applied to weathering intensity. Relationships between plant and environmental turnover were weak across SEC, and varied only slightly. This contrasts with strong relationships between topography, temperature and mammal turnover, suggesting variation in these variables is influencing mammal turnover across SEC.
However, the
Negative relationships between environmental and biotic turnover were more common to plants and were exclusive to the two low turnover zones, with the exception of plant-temperature turnover in the northern break, which was only slightly negative (β1 = −0.021). This suggests that greater environmental variability in the break influences higher biotic turnover and potentially lower overall beta diversity and species richness. In contrast, increased variability in elevation, slope, solar radiation and temperature within the low turnover zones is associated with lower plant turnover and hence higher species richness because the distributions of multiple species overlap in these regions. Complex, heterogeneous environments have often been associated with higher species richness because of the greater variety of habitats and more opportunities for resource exploitation that they provide
The northern break zone and the adjacent, smaller eastern overlap both contained similar genera and species numbers, despite many more observations being sampled in the latter. This overlap contains several cells of particularly low turnover, plus a band of low turnover extending ∼25 km southwest and connecting with a separate low turnover zone at −37°. These low turnover cells indicate the presence of comparatively stable communities with high richness and diversity. They also coincide with areas of relatively low topographic and climatic variability. The connecting strip is 4 to 8 km in width and mostly comprises reserves and NSW state forests, though there is some incursion of cleared/disturbed land at its narrowest point. Nonetheless, this connecting strip of low turnover warrants investigation as a potential corridor linking the two overlaps, especially since corridors comprising natural landscape features may increase species movement in fragmented landscapes
The eastern overlap may also be viable for additional conservation reserve siting. A key goal of conservation is to maximise biodiversity persistence
The northern break zone contained comparatively high proportions of unique plant and mammal genera and these same genera may play a significant role in driving higher rates of turnover.
Relative distribution densities (red = high density; blue = low density) of plant and mammal genera point observations common to the northern break and eastern overlap zones (panels A and E), unique to the eastern overlap zone (panels B and F), all plant and mammal genera in the northern break zone (panels C and G) and plant and mammal genera unique to the break zone (panels D and H) in the South East Corner (SEC) region of south-eastern New South Wales, Australia.
The distribution of mammal genera in the break and eastern overlap zones shows the same general pattern (
We have mapped gradational biotic and environmental turnover patterns in significantly greater detail and for a smaller region than Di Virgilio et al.
Plant species turnover maps in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia, for Sørensen moving window analyses rotated through 360° in 15° increments. Plant species turnover maps, moving window orientations 90° to 315°.
(TIF)
Plant species turnover maps in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia, for Sørensen moving window analyses rotated through 360° in 15° increments. Plant species turnover maps, moving window orientations 300° to 165°.
(TIF)
Plant species turnover maps in the South East Corner (SEC) bioregion of south-eastern New South Wales, Australia, for Sørensen moving window analyses rotated through 360° in 15° increments. Plant species turnover maps, moving window orientations 150° to 90°.
(TIF)
Comparison of the plant and mammal genera unique to northern break and eastern overlap zones, unique to the same break and a western overlap zone and common to the intersections of each pair.
(DOC)