The missing part of the past, current, and future distribution model of Quercus ilex L.: the eastern edge

Abstract

Ongoing climate change is anticipated to shift the geographical distribution range and impact local abundance of tree species by altering their ecological conditions. Given the lower resilience of populations at the species’ range edges, locally adapted range-edge populations are critical to the species’ survival under climate change. In this context, the distribution of holm oak (Quercus ilex L.) at the eastern border of its distribution range was assessed under current, past, and foreseeable future climate change scenarios, using species distribution models (SDMs). Current SDMs were developed using WorldClim 1.4 climate data as baseline at 30-second spatial resolution by using Generalized Boosted Regression Models (GBM) and showed moderate model performance. To compare temporal transferability and account for climate uncertainties of two versions of future climate data (CMIP5 and CMIP6), we used 4 Global Circulation Models (GCMs), 2 emission scenarios (moderate RCP45/SSP245 and pessimistic - RCP85/SSP585) for 2 different periods in the future (2040-2060 and 2060-2080). We also made predictions about the past (Mid-Holocene, about 6.000 years ago) using 4 CMIP5 GCMs. Most important variables of SDMs were distance to the sea, isothermality (BIO3), annual precipitation (BIO12), the mean temperature of driest quarter (BIO9), and the precipitation of driest month (BIO14). Our findings showed that the species’ potential distribution range probably used to be much wider in the mid-Holocene, which implies that the holm oak had a broader climatic niche during this period. The future projections indicate that its distribution area in the eastern border might increase particularly in the Black Sea region, while decreasing in the Aegean region resulting in a likely northward range shift in Turkey. However, other variables not included in our models such as land use changes might drive future shifts. Due to its high resistance to dry conditions and resilience, this species might continue to spread in southwestern Turkey in 2050s and 2070s. Finally, our study fills the gap in potential distribution predictions in context of climate change for the eastern boundary of the holm oak.

Keywords

• Species Distribution Model
• Global Circulation Models
• Holm Oak
• Turkey
• Range Edge
• Generalized Boosted Regression Models