Tanaka et al. Projecting potential habitats of Pinus koraiensis in East Asia and assessing the impact of climate change. In: The East Asian Flora: and its role in the formation of the world’s vegetation

Reference # 6466333

TOPIC 1. HOLOCENE HISTORY OF VEGETATION IN EAST ASIA

Projecting potential habitats of Pinus koraiensis in East Asia and assessing the impact of climate change

Nobuyuki TANAKA¹, Katsuhiro NAKAO¹, Jong Hak YUN², Ikutaro TSUYAMA¹ & Motoki HIGA¹
¹ Forestry and Forest Products Research Institute (FFPRI), Tsukuba, Ibaraki, Japan
² National Institute of Biological Resource, Incheon, Republic of Korea

Statistical models such as GAM projecting potential habitats of plants are useful tools for assessment of plant geography as well as for assessment of climate change impact. Comparison of potential habitats with fragmented distributions for a plant species may suggest causes of limited distributions of the species. While spatial data on plant species distribution and environmental factors such as climatic ones are necessary for modeling, species distribution data with high accuracy are much more difficult to collect. In the case of a species occurring in more than one country, it is better to collect distribution data covering the whole range of the species to accurately project its potential habitats. In order to assess impact of climate change on natural population of Pinus koraiensis in East Asia, we made a statistical model, MARS (Multivariate Adaptive Regression Splines), in which we used the distribution data (DCL) made by digitizing the map of its whole range (Critchfield & Little 1966) asa response variable and four climatic indices prepared in Worldclim (Hijmans et al. 2005) as explanatory variables. We also used three independent datasets of distribution of the species in China (CDD), Japan (PRDB) and Korea (KAVeR) to validate the model. CDD was made by digitizing the distribution map of the species included in Higher Plant of China (Qingdao Publishing House, 1999-2005) while PRDB and KAVeR were our original databases of vegetation plots. Accuracy of the model was considered good because AUC values based on CDD, PRDB and KAVeR were 0.97, 0.92 and 0.81 respectively. The projection of potential habitats under current (1950-2000) and future (2080) climate conditions indicated that the area of habitat would become small in actual distribution areas of northeastern China, central Japan and Korea. This species occurring in unsuitable habitats after climate change would decline gradually. Potential habitats would shift northward and become larger in Russia. Suitable habitats in the areas around the boundary between China and Korea would be sustained under the future climate, suggesting future refugia for the species.