Microclima: An r package for modelling meso- and microclimate

Abstract

Climate is of fundamental importance to the ecology and evolution of all organisms. However, studies of climate–organism interactions usually rely on climate variables interpolated from widely spaced measurements or modelled at coarse resolution, whereas the conditions experienced by many organisms vary over scales from millimetres to metres. To help bridge this mismatch in scale, we present models of the mechanistic processes that govern fine-scale variation in near-ground air temperature. The models are flexible (enabling application to a wide variety of locations and contexts), can be run using freely available data and are provided as an R package. We apply a mesoclimate model to the Lizard Peninsula in Cornwall to provide hourly estimates of air temperature at resolution of 100 m for the period Jan-Dec 2010. A microclimate model is then applied to a 1 km 2 region of the Lizard Peninsula, Caerthillean Valley (49.969°N, 5.215°W), to provide hourly estimates of near-ground air temperature at resolution of 1 m 2 during May 2010. Our models reveal substantial spatial variation in near-ground temperatures, driven principally by variation in topography and, at the microscale, by vegetation structure. At the meso-scale, hours of exposure to air temperatures at 1 m height in excess of 25°C ranged from 23 to 158 hr, despite this temperature never being recorded by the weather station within the study area during the study period. At the micro-scale, steep south-facing slopes with minimal vegetation cover experienced temperatures in excess of 40°C. The microclima package is flexible and efficient and provides an accurate means of modelling fine-scale variation in temperature. We also provide functions that facilitate users to obtain and process a variety of freely available datasets needed to drive the model.