Carbon stocks in cool temperate forests with different stand structure due to deer overgrazing

Recently, an overpopulation of large herbivorous mammals has been observed in the Northern Hemisphere. Their overgrazing can degrade eatable understory vegetation, enhance establishment of unpalatable species, and increase overstory trees mortality. It is still unknown how these vegetation changes affect carbon (C) stock capacity in forest ecosystems. We aimed to evaluate the effects of forest vegetation changes on C stocks due to over 20-years grazing of Japanese sika deer in southern Kyushu, Japan. The study area was divided into less grazed forests (LG) and heavily grazed forests (HG). The HG was further divided into three treatments: forests with no understory vegetation (HG-nu), forests with dominance of unpalatable shrubs (HG-ud), and forests with gap areas created by overstory trees mortality (HG-gap). Four 100-400 m² survey plots were established for each treatment. We evaluated differences in vegetation structure (e.g., stem density) and C stocks, including overstory trees (height <2 m), understory vegetation, leaf litter, fine woody debris (FWD), and coarse woody debris (CWD) between the LG and HG treatments. We also separated overstory trees into eatable and unpalatable trees for sika deer diet. Stem density in LG, HG-nu, HG-ud, and HG-gap were 2548 ± 1813, 1544 ± 1145, 15619 ± 5326, and 63 ± 75 stems ha^-1, respectively. Lower stem density was found in HG-nu compared to that in LG resulting from low density of small-diameter eatable trees (i.e., individuals with the diameter of < 10 cm). Instead, higher stem density was found in HG-ud compared to that in LG resulting from high density of small-diameter unpalatable trees. C stocks of overstory trees in HG-nu (12526.1 ± 5367.0 g C m^-2) was comparable to that in LG (10771.4 ± 3351.3 g C m^-2). C stocks of overstory trees in HG-ud (5118.2 ± 5656.3 g C m^-2) and HG-gap (2028.9 ± 2343.4 g C m^-2) were 50% and 81% lower compared to that in LG, respectively. C stocks of overstory trees in these treatments was dominated by few large-diameter trees, and differences in the stem density driven by small-diameter trees of each treatment did not contribute to the difference of C stocks. Understory vegetation C stocks in three HG treatments (i.e., HG-nu, HG-ud, and HG-gap) were >96% lower compared to LG. C stocks of leaf litter and FWD in HGs was also 36-68% lower compared to LG. The understory vegetation biomass was positively correlated with the total amount of leaf litter and FWD, suggesting that understory contributes to litter production and holdings. C stocks of CWD in HG-nu and HG-ud were comparable to LG whereas HG-gap showed 6.8-fold higher C stocks of CWD than that of LG due to the death of large-diameter trees that occurred gap formation. Our results highlight that deer-induced vegetation changes decrease in C stocks due to the mortality of large diameter trees and the loss of understory vegetation, and changes contribution of their components.