Mcera5: Driving microclimate models with ERA5 global gridded climate data

Abstract

Abstract Microclimate models predict temperature and other meteorological variables at scales relevant to individual organisms. The broad application of microclimate models requires gridded macroclimatic variables as input. However, the spatial and temporal resolution of such inputs can be a limiting factor on the accuracy of microclimate predictions. Due to its fine resolution and accuracy, the ERA5 reanalysis dataset is emerging as the favoured resource for global historical weather and climate data and has great potential for aiding microclimate modelling. Here we describe mcera5, an R language package that provides convenient access to, and wrangling of, the ERA5 climate datasets for use in microclimate models. Through this package, we provide functions to query ERA5 data for desired spatial and temporal extents, to correct for spatial biases and process outputs for easy interpretation by ecologists, thereby allowing faster and more accurate microclimate predictions. By validating with empirical observations from multiple biomes globally, we demonstrate that the use of ERA5 climate forcing via mcera5 improves the prediction accuracy of soil moisture, air temperature and relative humidity as compared to forcing with other globally available data and offers comparable performance when predicting soil temperatures. Through the provision of fine‐resolution ERA5 data, the mcera5 package fits into an ecosystem of tools for modelling microclimate in a spatio‐temporally explicit fashion, advancing our ability to efficiently predict microclimate for any place on Earth for the past, present or future. The package also provides convenient access to ERA5 datasets for a range of other applications.