Phenological plasticity will not help all species adapt to climate change ANNE DUPUTI  E 1,2 *, ALEXIS RUTSCHMANN 2,3 *, OPH  ELIE RONCE 4 andISABELLE CHUINE 2 1 Laboratoire EEP, CNRS UMR 8198, Universit  e Lille1, F-59655 Villeneuve d’Ascq Cedex, France, 2 CEFE UMR 5175, CNRS, Universit  e de Montpellier, Universit  e Paul-Val  ery Montpellier, EPHE, 1919 route de Mende, 34293 Montpellier Cedex 05, France, 3 Station d’Ecologie Exp erimentale du CNRS  a Moulis, Unit  e de Service et Recherche 2936, 09200 Moulis, France, 4 Institut des Sciences de l’Evolution Universit  e Montpellier 2, CNRS, IRD, CC65, Place Eug ene Bataillon, 34095 Montpellier Cedex 5, France Abstract Concerns are rising about the capacity of species to adapt quickly enough to climate change. In long-lived organisms such as trees, genetic adaptation is slow, and how much phenotypic plasticity can help them cope with climate change remains largely unknown. Here, we assess whether, where and when phenological plasticity is and will be adaptive in three major European tree species. We use a process-based species distribution model, parameterized with extensive ecological data, and manipulate plasticity to suppress phenological variations due to interannual, geo- graphical and trend climate variability, under current and projected climatic conditions. We show that phenological plasticity is not always adaptive and mostly affects fitness at the margins of the species’ distribution and climatic niche. Under current climatic conditions, phenological plasticity constrains the northern range limit of oak and beech and the southern range limit of pine. Under future climatic conditions, phenological plasticity becomes strongly adaptive towards the trailing edges of beech and oak, but severely constrains the range and niche of pine. Our results call for caution when interpreting geographical variation in trait means as adaptive, and strongly point towards spe- cies distribution models explicitly taking phenotypic plasticity into account when forecasting species distribution under climate change scenarios. Keywords: climate change, climatic niche, European beech, phenology, Scots pine, sessile oak, species distribution model Received 8 December 2014 and accepted 27 January 2015 Introduction Concerns are rising about the capacity of species to adapt quickly enough to global warming (Burrows et al., 2011; Dawson et al., 2011; Hoffmann & Sgr o, 2011). These adaptations imply genetic changes as well as nongenetic changes in trait values (Hoffmann & Sgr o, 2011; Meril€ a & Hendry, 2014). In long-lived organisms such as trees, genetic adaptation is slow (Savolainen et al., 2004), and how much phenotypic plasticity, that is the production of several phenotypes from a single genotype in different environmental con- ditions, can help them cope with climate change remains largely unknown (Anderson et al., 2012; Franks et al., 2014; Meril€ a & Hendry, 2014). Phenotypic plasticity is adaptive when the phenotype changes in a direction favoured by selection in the new environment (Conover & Schultz, 1995); that is, the phenotypic change results in higher fitness than if there was no phenotypic change. For instance, Great Tits adjust their laying date according to spring tempera- ture, which allows matching the timing of high food demand with peaks of insect abundance, thus mitigat- ing the impact of climate change (Charmantier et al., 2008). Phenotypic plasticity can, however, result in imperfect adaptation in a changing climate, requiring further genetic changes to reduce maladaptation (Gienapp et al., 2013). As phenotypic plasticity determines fitness in spa- tially heterogeneous or changing environmental condi- tions, it is necessarily related to range size and climatic niche breadth, that is, respectively, the geographical and climatic spaces where fitness is not null. In the con- text of climate change, adaptive phenotypic plasticity is thus expected to mitigate fitness losses, resulting in broader range and climatic niche than in the absence of plasticity. Hence, adaptive reaction norms would result in fewer extinction rates at the trailing edge of the range and/or in wider colonizable areas at the leading edge. Theoretical models have indeed shown that adaptive phenotypic plasticity can limit range contraction under a changing climate (e.g. Valladares et al., 2014) and help further genetic adaptation to stressful environment by slowing down the population decline in those environ- ments (Chevin et al., 2010). However, phenotypic plas- ticity can also be maladaptive (Ghalambor et al., 2007), Correspondence: Anne Duputi e, tel. +33 3 20 43 49 91, fax +33 3 20 43 69 79, e-mail: anne.duputie@ens-lyon.org *Equal contribution. 3062 © 2015 John Wiley & Sons Ltd Global Change Biology (2015) 21, 3062–3073, doi: 10.1111/gcb.12914