Greater effect of warming on community composition with increased precipitation and in moister landscape location

Questions We asked how plant community composition responded to experimentally produced warmer and drier climate conditions at the landscape scale with existing variation in local species composition and environmental conditions. We aimed to identify changes in community composition overall and the species with greatest response in abundance, and hypothesized that locally restricted species may be more sensitive to warming than more widespread species within the landscape based on the assumption that they have a narrower niche breadth with respect to environmental conditions. Location Semiarid, northern Mongolian steppe. Methods Open-top passive warming chambers (OTCs) elevated temperatures at two slope locations that differed in elevation, degree of slope, environmental conditions, and species composition. The OTC treatment was crossed with watering on the drier upper slope. Community composition differences among treatments were examined using canonical analysis of principal coordinates (CAP), which identified species contributing the most to differences. In response to warming, we also compared species locally restricted to one slope location with locally widespread species. Results Open-top passive warming chambers affected community composition more where soil moisture was greater, at the lower slope location and where warming was combined with supplemental watering on the drier upper slope. Locally restricted species responded negatively to the OTC while locally widespread species showed no overall response. Conclusions Community composition responses to warming differ even within the landscape over and above the initial differences that exist in community structure and abiotic factors. Our results suggest that a warmer and drier climate will impact community composition sooner under more mesic conditions, affect locally restricted species more strongly, and reduce variation in species composition across the landscape. To better predict community responses to future warming, we must consider combined and interactive effects with changes in precipitation and extant water availability.