Conservation effectiveness of protected areas for Hong Kong butterflies declines under climate change

On the Iberian endemism Eurylophella iberica Keffermuller and Da Terra 1978 (Ephemeroptera, Ephemerellidae): current and future potential distributions, and assessment of the effectiveness of the Natura 2000 network on its protection

“…Our results also emphasise the value of incorporating climate change information and species' future potential distributions into the designs of conservation programmes and priorities (e.g. Cheng and Bonebrake 2017). Moreover, few conservation planning studies have focused on insects, as compared to other taxa (Wilson and Maclean 2011;Sánchez-Fernández et al 2013).…”

Assessing the effectiveness of protected areas for conserving range‐restricted rain forest butterflies in Sabah, Borneo

“…However, some tropical butterflies have been found to decline during very severe droughts (Hill, ), whilst heavy rainfall may also be detrimental for some other forest taxa (e.g., see Ryan et al., ). Few studies have examined the distributions of tropical insects in relation to climate across Southeast Asia (e.g., see Cheng & Bonebrake, ; Klorvuttimontara et al., ), and so more research is needed to determine the causes and patterns in abiotic range limits across different insect taxa. Such information is vital for understanding the responses of insect species to climate change.…”

“…We modeled the distributions of 77 butterfly species across Borneo using R (R Core Team, ) with the biomod2 package (Thuiller, Georges, Engler, Georges, & Thuiller, ). We used an ensemble modeling approach to create a consensus of the predictions across three algorithms (see Cheng & Bonebrake, ; Marshall et al., ; Singh, McClean, Büker, Hartley, & Hill, ; Thuiller, Lafourcade, Engler, & Araújo, ), comprising: (a) A generalized linear model (GLM), with linear effects and stepwise selection based on Akaike information criteria (AIC); (b) a random forest (RF) model, using the default settings (no. trees = 501; node size = 5); and (c) maximum entropy modeling (MAXENT), including only linear and quadratic features (e.g., see Marshall et al., ).…”

“…trees = 501; node size = 5); and (c) maximum entropy modeling (MAXENT), including only linear and quadratic features (e.g., see Marshall et al., ). Maximum entropy modeling can perform well with few locality records (Phillips, Anderson, & Schapire, ; Platts et al., ; Wisz et al., ), whilst GLM and RF algorithms have also been used successfully in other insect studies (e.g., see Cheng & Bonebrake, ; Marshall et al., ).…”

“…It is important to focus conservation strategies on range‐restricted species in remaining areas of forest because these species can be particularly sensitive to disturbance (Bonebrake et al., ; Cleary & Mooers, ; Hill, Hamer, Tangah, & Dawood, ; Koh, Sodhi, & Brook, ) and because local extirpations could lead to extinction (Mittermeier et al., ). However, we currently know little about where range‐restricted insects occur in biodiverse tropical systems (Myers et al., ), and so understanding the climatic limits to species’ ranges may facilitate conservation planning in areas threatened by land‐use change (Cheng & Bonebrake, ; Klorvuttimontara et al., ). Knowledge about range‐restricted insect distributions in relation to aboveground forest carbon could also be important by linking species conservation to climate change mitigation strategies from avoided deforestation.…”