Biodiversity responses to forest cover loss: taxonomy and metrics matter

The actions required for biodiversity conservation depend on species responses to habitat loss, which may be either neutral, linear, or non-linear. Here, we tested how taxonomic, functional, and phylogenetic diversity of aquatic insects, dragonflies, frogs, and terrestrial mammals, as well as their species composition respond to forest cover loss. We hypothesized that taxonomic, functional, and phylogenetic diversity would respond nonlinearly (thresholds) to forest cover loss. Our findings do not support the current idea that a single threshold value of forest cover loss is applicable across tropical regions, or that some biodiversity facets are consistently more sensitive than others across different taxa. Species compositional responses to forest cover loss showed general patterns with thresholds between 30-50%. These results highlight the importance to consider multiple biodiversity facets when assessing the effects of forest cover loss on biological communities.


42
Human activities have substantially altered landscapes around the world, increasing species     119 We compiled data from studies conducted on the Bodoquena plateau, Mato Grosso do 120 Sul, central Brazil (Fig. 1). The vegetation consists mostly of deciduous and semideciduous 121 forests, woodland, and arboreal savannas (cerradão and cerrado stricto sensu, respectively),   163 We used a dataset that included 21 sampling sites in 15 forest fragments. Anurans 164 were sampled using 1 h in 2 Km 2 plots in each site. Sampling was performed by one observer 165 at night (5 pm until midnight) using visual encounter (Crump and Scott Junior, 1994) and 166 auditory survey (Zimmerman, 1994). We quantified forest cover within a 500-m buffer.  For medium to large (> 500 g body mass) terrestrial mammals, we recorded animals 175 by actively searching along six 1-km transects in each of 15 previously described landscapes.

176
One trained observer walked continuously along each transect from dawn to 11h. Species 177 sightings were also recorded through opportunistic encounters. We also used camera-traps to

258
The responses of biodiversity facets to the gradient of forest cover loss were highly 259 variable among taxonomic groups. A linear model provided the best fit for medium to large 260 terrestrial mammal taxonomic diversity and the relationship between taxonomic diversity and 261 percentage of forest cover was positive (Fig. 2). For anuran taxonomic diversity, the 262 piecewise model provided the best fit and it detected a threshold at around 59% of forest 263 cover. Intermediate levels of forest cover had the lowest anuran taxonomic diversity ( Fig. 2;   264 Supplementary material, Table S2). For aquatic insects and odonates, the null model provided 265 the best fit (Supplementary material, Table S2).

267
The piecewise model provided the best fit for anuran functional diversity, and it 268 showed a threshold at 65% forest cover. Anurans had their lowest functional diversity 269 between 55-70% of forest cover ( Fig. 2; Supplementary material, Table S2). The piecewise 270 model provided the best fit for anuran phylogenetic diversity and had a threshold of around 271 62% forest cover. The anuran phylogenetic diversity pattern was quite like that of anuran 272 taxonomic and functional diversity (Fig. 2). The highest values of phylogenetic diversity 273 were observed at high or low percentages of forest cover ( Fig. 2; Supplementary  Species composition showed thresholds between 30 and 60% of forest cover for three 277 of the four taxonomic groups (Fig. 3). Aquatic insects showed a community composition 278 threshold around 40% of forest cover, while odonates had a threshold close to 45% (Fig. 3). 279 For medium to large terrestrial mammals, we detected thresholds around 35% of forest cover 280 (Fig. 3). Results for anurans and small mammals did not show species with adequate 281 reliability index values, so we were unable to obtain meaningful results from species  Figure S1, Table S3). forest cover loss. However, this does not mean that the idea of using ecological thresholds to 302 orient conservation and restoration strategies should be discarded. Instead, our findings 303 suggest that these efforts will be most effective when selecting sensitive metrics and best fit 304 models to conservation problems that require knowledge of multi-taxa taxonomic, functional, 305 phylogenetic and community compositional responses to vegetation loss. Adapting these 306 models for specific taxonomic groups and regions will be a critical step in selecting sensitive 307 bioindicators of the effects of forest loss, and based on our results, non-linear responses 308 should be considered during the process of model development.

318
Only when we focused on community compositional changes did a more general pattern 319 emerge, showing thresholds between 30-60% of forest cover for some taxonomic groups.  al., 2020). However, it will be important to consider regional contexts and the unique 393 responses of multiple taxa and metrics, rather than relying on a general threshold value that is 394 appropriate for a specific group, but potentially harmful to other components of biodiversity.