Kaitlyn E. Horisk ^1* Sarah Ivory ^1,2 Katherine Freeman ¹ Allison Baczynski ^1,3 Joy McCorriston ⁴ Andrew Anderson ⁵ R Scott Anderson ⁶ Ali Al Kathiri ⁷

• ¹ Department of Geosciences, College of Earth and Mineral Sciences, The Pennsylvania State University, University Park, Pennsylvania, United States
• ² Earth and Environmental Systems Institute, College of Agricultural Sciences, The Pennsylvania State University, University Park, Pennsylvania, United States
• ³ Laboratory for Isotopes and Metals in the Environment, Earth and Environmental Systems Institute, College of Agricultural Sciences, The Pennsylvania State University, University Park, Pennsylvania, United States
• ⁴ Department of Anthropology, College of Arts and Sciences, The Ohio State University, Columbus, Ohio, United States
• ⁵ Royal Botanic Garden Edinburgh, Edinburgh, Scotland, United Kingdom
• ⁶ School of Earth and Sustainability, College of the Environment, Forestry, and Natural Sciences, Northern Arizona University, Flagstaff, Arizona, United States
• ⁷ Ministry of Heritage and Tourism, Muscat, Oman

Over 1/3 of the Earth's human population relies on dryland ecosystems for food and water resources. While these ecosystems are highly sensitive to changes in climate, we lack observational data as to how changes in hydrology influences plant communities. Paleoecological data for southern Arabia show woodland communities transitioned to more dry-adapted herbaceous plants, which suggests rainfall decreased across the Holocene. To assess relationships between hydrology and ecology, we employed leaf wax n-alkane distributions, δ 13 Cwax, and δDwax records from rock hyrax (Procavia capensis) middens in Dhofar, Oman. The biomarker properties allowed reconstruction of changes in C3/C4 vegetation and local moisture availability, in tandem with community changes represented by a published pollen record. To constrain interpretations, n-alkane analyses were conducted on herbarium specimens of leaves collected in Dhofar. For the modern specimens, xeric plants typically contained longer homologues than mesic plants. Across the fossil middens (4038-109 cal yrs BP), the proportions of plant-wax homologues do not show major changes, and thus do not suggest a shift between xeric versus mesic plants. Similarly, δ 13 Cwax values indicate little or no change in the distributions of C3 and C4 vegetation. Limited δDwax data from the middens confirm overall drying occurred into the late Holocene, punctuated by a wetter pulse at ~1.6 ka. Taken together, plant wax distributions and isotope data indicate changes in moisture availability across the late Holocene did not alter the structural composition of the plant communities and that the proportion of C3/C4 vegetation remained stable. We infer vegetation changes associated with late Holocene drying involved reshuffling of community composition and not major changes in vegetation structure.