Model predictions of key revegetation species response to climate change differ between correlative and mechanistic models: implications for adapting restoration

Using native species to revegetate and restore areas of the landscape is important to redress habitat degradation and fragmentation resulting from land‐use change. However, under a changing climate, the restoration of previously cleared land in urban and agricultural areas may be increasingly difficult as life history traits of plants, such as seed dormancy, stratification, germination, and seedling survival, could be impacted. Using 10 native species commonly used in revegetation projects in southeastern (SE) Australia and the Port Phillip and Westernport (PPWP) region, we compared correlative and mechanistic model predictions of how climate change could impact the climatic suitability of each of these species. We compared climatic suitability under projected future climate conditions for 2090 relative to current climatic suitability. The correlative models were substantial to excellent in predicting species occurrence and showed that in SE Australia and the PPWP region, all 10 species were at a moderate to extreme risk under the 2090 projection. Substantial divergence between the correlative and mechanistic modeling outputs was also evident, with species such as Allocasuarina verticillata , Olearia lirata , and Bursaria spinosa showing quite different responses for correlative versus mechanistic models. This was due to the species' response to water stress during germination and establishment. Differences between the correlative and mechanistic models highlight the need to consider both approaches when modeling a species' response to climate change.