Predicting climate-driven shift of the East Mediterranean endemic Cynara cornigera

Heba Bedair ^1,2 Yehia Abdulrahman Hazzazi ³ Asmaa Abo Hatab ⁴ Marwa Waseem A. Halmy ⁵ Mohammed A. Dakhil ⁶ Mubarak S. Algharani ⁷ Mari Sumayli ³ Dr. Ahmed El-Shabasy ^3* Mohamed El-Khalafy ⁴

• ¹ Department of Botany, Faculty of Science, Tanta University, Tanta, Gharbia, Egypt
• ² Department of Wildlife Sciences, Faculty of Forest Sciences and Forest Ecology, University of Goettingen, Göttingen, Lower Saxony, Germany
• ³ Department of Biology, College of Science, Jazan University, Jazan, Jizan, Saudi Arabia
• ⁴ Department of Botany and Microbiology, Faculty of Science, Kafrelsheikh University, Kafrelsheikh, Egypt
• ⁵ Department of Environmental Sciences, Faculty of Science, Alexandria University, Alexandria, Alexandria, Egypt
• ⁶ Department of Plant and Microbiology, Faculty of Science, Helwan University, Cairo, Beni Suef, Egypt
• ⁷ Libyan Authority for Scientific Research, Tripoli, Libya

Climate change poses significant challenges to the richness and distribution of endemics in the Mediterranean region. Assessing the impact of climate change on the distribution patterns of Mediterranean endemics is of critical importance for understanding the dynamics of these terrestrial ecosystems under the uncertainty of future changes. . This decline is linked to how ongoing climate change is affecting natural resources like water and the capacity of foraging sites. In fact, it is distributed in 3 fragmented locations in Egypt (Wadi Hashem (5 individuals), Wadi Um Rakham (20 individuals), Burg El-Arab (4 individuals)). In this study, we examined C. cornigera's response to predicted climate change over the next few decades (2020-2040 and 2061-2080) using species distribution models (SDMs). Our analysis involved inclusion of bioclimatic variables, in the SDM modeling process that incorporated five algorithms: generalized linear model (GLM), Random Forest (RF), Boosted Regression Trees (BRT), Support Vector Machines (SVM), and Generalized Additive Model (GAM). The ensemble model obtained high accuracy and performance model outcomes with a mean AUC of 0.95 and TSS of 0.85 for the overall model. Notably, RF and GLM algorithms outperformed the other algorithms, underscoring their efficacy in predicting the distribution of endemics in the Mediterranean region. Analysis of the relative importance of bioclimatic variables revealed Precipitation of wettest month (Bio13) (88.3%), Precipitation of warmest quarter (Bio18) (30%), and Precipitation of driest month (Bio14) (22%) as the primary drivers shaping the potential distribution of C. cornigera. The findings revealed spatial variations in habitat suitability, with the highest richness potential distribution observed in Egypt, (especially the Arishian sub sector), Palestine, Morocco, Northern Cyprus, and different islands in the Sea of Crete. Furthermore, our models predicted that the distribution range of C. cornigera would drop by more than 25 % during the next few decades. Surprisingly, the future potential distribution area of C. cornigera (SSP 126 scenario) for 2061 and 2080 showed that there is increase in the suitable habitats area. It showed high habitat suitability along the Mediterranean coastal strip of Spain,