Climate change poses an imminent threat to global biodiversity, prompting urgent investigations into its profound impacts on ecosystems (IPBES, 2019). Moreover, a decline in biodiversity will deepen the climate crisis, with consequences including reduced species abundance, local extinctions, and the rapid degradation or loss of ecosystems such as mangroves, forests, peatlands, meadows, and seagrass (Macreadie et al., 2021). These changes not only affect the capacity of the planet to store carbon but also impede the ability of nature and humanity to adapt to or cope with evolving climate conditions (Pettorelli et al., 2021). Forests, vital agents in climate change mitigation, face unprecedented challenges due to the anticipated pace of environmental shifts (Hansen et al., 2001; Hamann and Wang, 2006; Warren et al., 2018). Despite their critical role, our understanding of how climate influences the spatial distribution (Canham and Murphy, 2017), abundance, and assembly of tree species remains limited (Chu et al., 2019). Future changes in tree assembly will significantly affect forest ecosystems and ecosystem services.

The capability of conservation and management practices to address climate change depends on the ability to forecast climate trends in the coming decades. This includes predicting ecological and biogeographic responses to climate change and anticipating the impacts of future land management practices (Colavito, 2017). A comprehensive understanding of how forest ecosystems react to climate change is crucial for reforestation initiatives, such as selecting species adapted to future climate conditions (Harger, 1993). In many temperate forest ecosystems, a shift from conifer-dominated to broadleaf-dominated ecosystems is anticipated (Hanewinkel et al., 2014; Thom et al., 2017). To forecast such changes, species distribution models (SDMs), also known as Ecological Niche Models, bio-envelope models, or species envelope models, have been widely used to determine the responses of species under different climate change scenarios (e.g., Guisan and Zimmermann, 2000; Elith and Leathwick, 2009; Tanaka et al., 2012; Nakao et al., 2013; Tang Y. et al., 2018; Ohashi et al., 2019).

Increasingly, research is examining long-term changes in vegetation in response to climate change (Anderson and Song, 2020) and the associated changes in the carbon balance (Mo et al., 2023). It seems unlikely that forests will be able to maintain their current stand structure and total carbon sequestration and storage potential with the changing climate (Thompson et al., 2009). Considering the likelihood that some species will decline under climate change, then which other tree species will be more resilient to the anticipated climate and may therefore be considered suitable alternatives? Instead of relying on models centered on individual species, it might be more beneficial to predict forthcoming changes in spatial distribution as a collective assembly (Zurell et al., 2020). It is important to predict potential habitat distributions at national, regional, and landscape scales under current and future climate conditions (Tanaka et al., 2005).

Temperate deciduous forests in the Northern Hemisphere, situated between evergreen and subarctic coniferous forests, experience distinct seasonal cycles and diverse compositions (Gilliam, 2016). These characteristics make them especially vulnerable to compositional changes driven by climate shifts. Despite their ecological significance, research on future species assemblages in East Asia, particularly in Japan, remains limited. Here the southern (lower) limit of distribution lies adjacent to the upper limit of evergreen broad-leaved forests, and if these forests are invaded, the light environment of the forest floor is likely to change significantly, leading to shifts in species composition and a potential decline in species diversity. The unique position of temperate deciduous forests makes them ideal systems for studying the impacts of climate change on forest ecosystems (Ohsawa, 1993). Moreover, temperate forests are home to numerous tertiary relict plants and endangered species, emphasizing their critical importance for biodiversity conservation and ecological research (Tang C. Q. et al., 2018).

Within this context, the Ogawa Forest Reserve (OFR) emerges as a unique example of an old-growth forest, situated in the southern Abukuma Mountains of Japan. Despite detailed studies on its forest dynamics and structure (Masaki et al., 1992; Nakashizuka et al., 1992; Shibata et al., 2010; Masaki et al., 2021; Wijenayake et al., 2023) there is little in-depth information on the future distribution of species in the vicinity of the Forest Reserve.

This study seeks to bridge this knowledge gap using species distribution models to predict the current and future geographical distribution of major tree species found in vicinity of the OFR under different climate change scenarios. More specifically, we address whether the projected changes in climate suitability are related to the structural stability of a plant community in terms of vertical and horizontal spatial organization and verify the generality of the findings based on the previous related studies on temperate forests. We hypothesized that even with the expected changes in climate, the structural stability of the species assembly will remain unchanged until the end of this century. This hypothesis is based on the observation that populations are already responding to climate trends in the same direction as expected in the future (Jump et al., 2009).

Our results provide detailed insights into the possible trends of the most abundant tree species and the impact of changes in the local tree species abundance and composition on the tree species assembly under future climate scenarios in the study area. They also provide information on potential forest ecosystem stability, which should be given special attention in the context of local species diversity. Overall, this work offers insights into the intricate relationships among species distribution, climate change, and the landscape surrounding OFR, offering valuable information for optimizing conservation and management strategies.

This study identified the climatic variables that seem most important for determining the habitat suitability of major tree species at present, and under two different future climate change scenarios. The majority of species, including Carpinus. cordata, Carpinus. laxiflora, A. amoenum, S. obassia, Carpinus. tschonoskii, F. crenata, Castanea. crenata, Cerasus. leveilleana, and Cornus. controversa, are affected by winter temperatures (BIO06 and BIO11), which is the most important factor affecting the future distribution of the species. Projected future changes for each species centered around the OFR under two SSP scenarios (a scenario implying low greenhouse gas forcing in the future [SSP2] and a fossil fuel development scenario [SSP5] for 2050 and 2,100) suggest that the distributions of tree species in and around OFR will change until the end of the 21st century. Under both climate scenarios, some structural changes are related to the species composition of the assembly of subcanopy species. However, the uncertainties and potential sources of error in model predictions may arise from complex biotic and abiotic interactions within the community, which are often challenging to capture accurately. Incorporating additional variables, such as terrain and soil characteristics, may only lead to modest improvements in model performance, as these factors might not fully account for the intricate ecological dynamics influencing species distributions.

The literature suggests that the structural complexity of the canopy in old-growth forests facilitates greater resource use efficiency while maintaining significant carbon storage potential (Hardiman et al., 2013; Obata et al., 2023). Near OFR, the subcanopy species C. cordata, A. rufinerve, and S. obassia will disappear completely by 2,100 under SSP5–8.5 while the remaining two subcanopy species (Carpinus japonica and Carpinus tschonoskii) will markedly decrease. In the 2,100 climate mitigation scenario, a similar trend is observed for two subcanopy species: S. obassia and C. cordata. The distribution of the canopy species F. crenata will shrink drastically by the end of this century. Another study of the possible future of F. crenata in Japan identified vulnerable habitats at low elevations on both the Sea of Japan and Pacific Ocean sides (Matsui et al., 2018). The distributions of the canopy species Betula grossa, A. amoenum, C. controversa, and K. septemlobus will shrink remarkably according to projections.

With the potential reductions in the abundance of subcanopy species, the forest structure is expected to change. However, prolonged exposure to unsuitable environmental conditions can lead to the physiological deterioration of trees, disrupting the rhythms of flowering, fruiting, and fertility (Selås et al., 2002; Wright et al., 2022). Consequently, the growth of the next generation of fruiting and juvenile trees may be hindered, resulting in a decline in population density (Niinemets, 2010). Concurrently, competition from species better adapted to warmer environments (e.g., evergreen broadleaved species) may induce changes in forest species composition and structure, accompanied by increased invasion and establishment from external sources, such as the growth of warm-temperate tree species (Lindner et al., 2010).

Stand structure has important implications for forest ecosystems, including the maintenance of wildlife habitat, sustaining biodiversity, nutrient cycling, regulating climate via carbon storage, and affecting forest regeneration (Brassard and Chen, 2006). In this sense, changing climate will likely affect the stability of the plant assembly and the habitats of various mammals and birds will also change. Accordingly, species diversity in the ecosystem surrounding OFR is expected to decrease. Based only on our findings, however, we cannot make a general statement regarding forest ecosystem stability throughout Japan. There may be places where climate change is working against Q. serrata and favoring subcanopy species.

Overall, Q. serrata is projected to remain stable and even expand its distribution, even under extreme climate conditions in 2100. Notably, it is the only one of the 17 species that expands its distribution area under both the climate adaptation and mitigation scenarios (Figure 4). These results suggest a simplified species composition of the canopy layer and a reduced abundance of subcanopy species in OFR in the future. Q. serrata occurs in the intermediate temperate forest zone and is widely distributed in lower and warmer areas, particularly in less snowy environments (Matsui et al., 2018). A demographic study of the life history of OFR indicated that Quercus species may continue to dominate the future landscape, primarily due to the higher survivorship at diameters at breast height of 10–40 cm (Wijenayake et al., 2023). The canopy species A. pictum and C. crenata will also remain stable under SSP5 in 2100.

Taking proactive steps to understand how trees respond to climate, through well-planned tests in specific areas, could bring tangible advantages in the future (Rehfeldt et al., 2014). The findings in Appendix 2 can help identify suitable conservation strategies for the tree species examined here. For example, for the subcanopy species that are expected to undergo considerable loss in the future, efforts could focus on facilitating range expansion to the northwest of OFR, which may enhance their stability facing the consequences of rapid climate change. Furthermore, the northern range limits of evergreen broadleaved tree species are projected to expand to higher elevations and northwards into areas dominated by Quercus (Nakao et al., 2011; Matsui et al., 2018). Our study also suggests that the Fagaceae will show resilience and expand their distributions, particularly Q. serrata, and especially under extreme climate conditions in 2100.

The species C. controversa and C. crenata had lower AUC values, which is a study limitation. C. controversa is a fast-growing, shade-intolerant pioneer species that may have lower correlations with climate variables. C. crenata is a valuable food resource related to agroforestry and it would be affected by past socioeconomic activities rather than bioclimate variables (Freitas et al., 2022). Therefore, it is important to note that the distribution changes are not only affected by climate factors but also by socioeconomic factors such as anthropogenic disturbances (Koyama et al., 2021), and land use history, which could have a lasting impact on forest resilience. For species such as C. controversa with potentially long seed dispersal distance (Wijenayake et al., 2024), the incorporation of spatially dynamic processes such as seed dispersal models should also be important to improve model prediction from landscape perspective. Long-distance dispersal of seeds significantly influences plant population dynamics via enabling vulnerable species to colonize more suitable distant habitats; however, our understanding of its implications at a landscape scale remains limited. For this reason, we excluded seed dispersal factors to minimize complexity. Climate change will increase disturbances in forests, but these disturbances should not be seen only as a threat to stability. They have the potential to enhance tree species diversity, with positive impacts on productivity and other ecosystem functions (Silva Pedro et al., 2016), while leading to changes in habitat quality (Hovenden and Williams, 2010). These vulnerabilities are not necessarily a realistic picture of the area around OFR. Note, however, that the future changes in this area may not be as severe due to topographic features. The current dominant tree species will remain similar, although their relative dominance will change with the climate.

To verify the generality of findings from previous related studies, we figured out sharply contrasting outcomes. Hansen et al. (2001) examined the potential impacts of human-induced climate change on forest biodiversity in the eastern U.S., revealing that habitat suitability for tree species varies based on their individual responses to environmental changes. Their study predicted a northward shift in habitat for many species, ranging from 100 to 530 km, with some species extending their optimal habitats beyond the U.S. border. This contrasts with our findings, where a high-resolution ecosystem-based climate envelope model predicts climate change impacts on forests in British Columbia, emphasizing climate as the primary driver of plant distribution (Hamann and Wang, 2006). By 2050, a significant proportion of present-day beech and sessile oak forests in zonal positions at low altitudes in Central Europe may fall outside their current bioclimatic niches. However, in our case the current dominant tree species are expected to remain, their relative dominance is likely to shift with changing climatic conditions. We believe that our study expanded the generality of knowledge of this field.

This study demonstrated the utility of species distribution models for addressing knowledge gaps regarding changes in major tree species under varying climate change scenarios. We project that the canopy species Q. serrata will become more dominant than other canopy species by 2,100 under the baseline climate scenario, potentially leading to shifts in the species composition of the OFR. Consequently, the stability of forest ecosystems and local species diversity may be compromised under future climate change scenarios. This information is crucial for developing vegetation monitoring, ecosystem management, and climate change adaptation pathways. Further refinements should consider factors such as topography and land use changes in the surrounding landscape.