Environmental dependence of population dynamics and height growth of a subalpine conifer across its vertical distribution: an approach using high-resolution aerial photographs MASAYA SHIMAZAKI, TAKEHIRO SASAKI, KOUKI HIKOSAKA andTOHRU NAKASHIZUKA Graduate School of Life Sciences, Tohoku University, Aoba 6-3, Aramaki, Aoba-ku, Sendai, 980-8578, Japan Abstract Many studies have reported shifts in the altitudinal ranges of plant species in response to recent global warming. However, most studies of tree species have been conducted on a small scale and have focused on tree line ecotones by examining tree rings and age structure on account of the long life spans of the trees. To examine the impact of climate change on forest dynamics at a regional scale, we investigated differences in the population density and canopy height of a Japanese subalpine coniferous species, Abies mariesii, between 1967 and 2003 by analysis of high-resolution aerial photographs of the Hakkoda Mountains, Honshu, Japan. In 712 plots within the photographs we analyzed which environmental variables (including elevation, aspect, wetness, and distance from moorlands) account for these changes. The population density of A. mariesii decreased below 1000 m a.s.l. and increased above 1300 m a.s.l. It also increased around moorlands, which may provide refugia at low elevations. The rate of increase in canopy height was lowest on the southeastern slopes and on the periphery of the moorlands. The distinct changes in the population den- sity of A. mariesii at its distribution limits probably reflect the responses of the population to climatic changes during three decades. Areas surrounding the moorlands may offer refugia in spite of the poor growing conditions there. Keywords: aerial photographs, climate change, generalized additive mixed-effect models, global warming, Japan, population dynamics, refugia, upward shifts Received 25 January 2011; revised version received 17 June 2011 and accepted 22 June 2011 Introduction It is becoming clear that climate change in boreal, tem- perate, and tropical ecosystems is shifting the distribu- tion of many plant species polewards and upwards (Parmesan, 2006). Climate change affects vegetation structure through direct effects on demography and indirect effects on disturbance regimes (Overpeck et al., 1990). The resulting replacement of dominant plant species can drastically change the structure and loca- tion of biomes, alter the structure of ecosystems, and interfere with the provision of ecosystem services to human populations (Gonzalez et al., 2010). Although many studies have reported shifts in the altitudinal ranges of plant species in response to recent global warming, detailed measurements among woody species are rare (Pen ˜ uelas et al., 2007), because the life spans of trees are very long (Brubaker, 1986). As a result, most studies of tree species have focused on the tree line (i.e. the distribution limit of tree species at least 3 m in height; Ko ¨ rner & Paulsen, 2004; Harsch et al., 2009) by examining tree rings and a population’s age structure (Wang et al., 2006; Kullman, 2007; Pen ˜ uelas et al., 2007; Batllori & Gutie ´rrez, 2008). In this context, temporal changes in population structures at the tree line are good indicators of climate change (Wang et al., 2006), because alpine tree lines are climatically deter- mined species distribution limits (Ko ¨ rner, 1998; Ko ¨ rner & Paulsen, 2004) and are considered particularly sensi- tive to altered temperature regimes (Theurillat & Gui- san, 2001; Gehrig-Fasel et al., 2007). These species often respond to climatic warming by increasing recruitment (Kullman, 1986; Payette & Lavoie, 1994) or density (Camarero & Gutie ´rrez, 2004; Beckage et al., 2008) and by advancing uphill (Bradley & Jones, 1993; Gehrig- Fasel et al., 2007). However, it is difficult to investigate population dynamics from tree ring growth because tree rings can usually be obtained only from living trees. Furthermore, analysis of distributional extremes is sensitive to sampling effort (Shoo et al., 2006). Lenoir et al. (2008) investigated plant species’ responses in dis- tribution by measuring shifts in the altitudinal position of the species’ maximum probabilities of presence. They found a significant upward shift of species in a Correspondence: Takehiro Sasaki, tel. + 81 (0)22 795 5790, fax + 81 (0)22 795 5790, e-mail: sasa67123@gmail.com; sasa@m.tohoku.ac.jp © 2011 Blackwell Publishing Ltd 3431 Global Change Biology (2011) 17, 3431–3438, doi: 10.1111/j.1365-2486.2011.02491.x