Choice of baseline climate data impacts projected species' responses to climate change

Abstract

Climate data created from historic climate observations are integral to most assessments of potential climate change impacts, and frequently comprise the baseline period used to infer species‐climate relationships. They are often also central to downscaling coarse resolution climate simulations from General Circulation Models (GCMs) to project future climate scenarios at ecologically relevant spatial scales. Uncertainty in these baseline data can be large, particularly where weather observations are sparse and climate dynamics are complex (e.g. over mountainous or coastal regions). Yet, importantly, this uncertainty is almost universally overlooked when assessing potential responses of species to climate change. Here, we assessed the importance of historic baseline climate uncertainty for projections of species' responses to future climate change. We built species distribution models (SDMs) for 895 African bird species of conservation concern, using six different climate baselines. We projected these models to two future periods (2040–2069, 2070–2099), using downscaled climate projections, and calculated species turnover and changes in species‐specific climate suitability. We found that the choice of baseline climate data constituted an important source of uncertainty in projections of both species turnover and species‐specific climate suitability, often comparable with, or more important than, uncertainty arising from the choice of GCM. Importantly, the relative contribution of these factors to projection uncertainty varied spatially. Moreover, when projecting SDMs to sites of biodiversity importance (Important Bird and Biodiversity Areas), these uncertainties altered site‐level impacts, which could affect conservation prioritization. Our results highlight that projections of species' responses to climate change are sensitive to uncertainty in the baseline climatology. We recommend that this should be considered routinely in such analyses.

The pyrogeography of eastern boreal Canada from 1901 to 2012 simulated with the LPJ-LMfire model

“…Compared with the previous modelling efforts that had been conducted by Pfeiffer et al (2013) using the original LPJ-LMfire model, the results that are reported here show substantial improvement in the capacity of the DGVM to simulate fire ignition in the Canadian boreal forest. The use of a high-quality lightning strike data set instead of the low-resolution LIS/OTD global data set that was used by Pfeiffer et al (2013) allowed us to capture the spatial gradient of fire activity in a substantially better manner (Baker et al, 2016). The results confirmed that fire in the study area is strongly ignition limited, while most fire models have simply assumed that fire would always occur under appropriate weather and fuel conditions, e.g., SIMFIRE (Hantson et al, 2016).…”

Ozone effects on European forest growth—Towards an integrative approach

“…In case of limited measured data in terms of spatial and temporal extent, modelled data are valuable, but only if the resulting bias is quantified and considered in the interpretation of the results (e.g. Baker, Hartley, Butchart, & Willis, ).…”

Contributions of different bias‐correction methods and reference meteorological forcing data sets to uncertainty in projected temperature and precipitation extremes

“…Although bias correction is an unavoidable procedure for using GCM outputs as the inputs of impact models, it is an additional source of uncertainty in climate risk assessments because different bias-correction methods and reference daily weather data sets often lead to different impact outcomes Hawkins et al, 2013;Seaby et al, 2015;Baker et al, 2016]. Global retrospective meteorological forcing data sets, a hybrid of reanalysis data and gridded observations [e.g., Sheffield et al, 2006;Hanasaki et al, 2008;Hirabayashi et al, 2008a;Weedon et al, 2011Weedon et al, , 2014Iizumi et al, 2014], are used as the reference climatology in impact communities when bias correction is conducted at the global scale [e.g., Hempel et al, 2013].…”