Spatial Distribution and Climate Warming Impact on Abies kawakamii Forest on a Subtropical Island

Abstract

Species distribution modeling (SDM) is currently the primary tool for predicting suitable habitats for species. In this study, we used Abies kawakamii, a species endemic to Taiwan. Being the only Abies species distributed in high mountains, it acts as an ecological indicator on the subtropical island. We analyzed a vegetation map derived from remote sensing and ground surveys using SDM. The actual distribution of A. kawakamii in Taiwan has a total area of 16,857 ha distributed at an altitude of 2700–3600 m, and it often forms a monodominant forest at 3100–3600 m with the higher altitude edge as a forest line. Exploring the potential distribution of A. kawakamii through MaxEnt showed that the suitable habitat was 73,151 ha under the current climate. Under the scenarios of temperature increases of 0.5, 1.0, 1.5, and 2.0 °C, suitable habitat for A. kawakamii will gradually decrease to 70.2%, 47.1%, 30.2%, and 10.0% of this area, respectively, indicating that A. kawakamii will greatly decline under these climate warming scenarios. Fire burning disturbance may be the most significant damage to A. kawakamii at present. Although A. kawakamii has been protected by conservation areas and its natural regeneration is in good condition, it rarely has the opportunity to migrate upwards during climate warming. We suggest that in the future, research on the natural regeneration and artificial restoration of A. kawakamii should be emphasized, especially in the forest line ecotone.

Modeling of the Spatial Distribution of Forest Carbon Storage in a Tropical/Subtropical Island with Multiple Ecozones

“…Among them, FIR-NF showed much lower C A than other natural coniferous types and a large variation in the uncertainty test, which might be not a misestimation because the DBH and H database used in the simulations are from permanent plots and can therefore be considered close to the real conditions. In fact, the distribution of fir forests is at an altitude range of 3100-3600 m, which is much higher than that of other pure coniferous forest types, and tall fir trees are easily broken due to strong mountain winds, resulting in a smaller wood stock [28,29]. Thus, the larger variation of FIR-NF was more likely due to its nature of wide variability of DBH and H. In addition, the C A of coniferous plantations in the simulation was much smaller than that of natural coniferous forests (57.87 t C ha −1 ).…”

Modeling of the Spatial Distribution of Forest Carbon Storage in a Tropical/Subtropical Island With Multiple Ecozones

“…Among them, FIR-NF showed much lower CA than other natural coniferous types and a large variation in the uncertainty test, it might be not a misestimation because the DBH and H database used in the simulations are from permanent plots and can therefore be considered close to the real condition. In fact, the distribution of fir forests is at an altitude range of 3100-3600m, which is much higher than that of other pure coniferous forest types, and tall fir trees are easily broken due to strong mountain winds, resulting in a smaller wood stock [35,36]. Thus, the larger variation of FIR-NF was more likely due to its nature of wide variability of DBH and H. In addition, CA of coniferous plantations in the simulation was much smaller than that of natural coniferous forests (57.87 t C ha -1 ).…”