The Effect of Habitat Association and Edaphic Conditions on Tree Mortality during El Niño‐induced Drought in a Bornean Dipterocarp Forest

Abstract

The effects of El Niño-induced droughts on dipterocarp forests must be quantified to evaluate the implications of future global climatic changes for the tropical forests of Southeast Asia. We studied the mortality of trees ! 1 cm in diameter in a lowland dipterocarp forest in Borneo before, during, and after the 1997/1998 El Niño drought. The annual mortality rates were 1.30, 1.75, and 1.66 percent/yr for the pre-drought, drought, and post-drought periods, respectively. The effect of drought was tree size-dependent being greater for larger trees. Modified logistic regression analysis revealed a significant interaction effect between species' habitat association and edaphic condition on mortality rates in all periods. For species associated with wet habitat, drought effect was greater in dry conditions than in wet conditions, in both the drought and post-drought periods. The mortality rates of dry-habitat species were less affected by the drought both in dry and wet conditions. A similar pattern was also found in common Dipterocarpaceae species; mortality rates increased more in species associated with wet-habitat in the drought and post-drought periods. Species and families with higher mortality in the pre-drought period tended to experience greater mortality increases during the drought and post-drought periods. These results suggest that changes in drought regimes alter the species composition and spatial distribution of dipterocarp forests.

Crown damage and the mortality of tropical trees

“…There was also a direct negative effect of fertility on survival, possibly related to soil moisture rather than soil nutrient status. Previous studies have shown that trees growing in wetter habitats in aseasonal rainforests suffer more during occasional intense droughts (Itoh et al ., ; Zuleta et al ., ). Mortality in wetter habitats at Lambir was increased significantly following the 1997–8 ENSO (Table ; Nakagawa et al ., ; Itoh et al ., ).…”

“…Previous studies have shown that trees growing in wetter habitats in aseasonal rainforests suffer more during occasional intense droughts (Itoh et al ., ; Zuleta et al ., ). Mortality in wetter habitats at Lambir was increased significantly following the 1997–8 ENSO (Table ; Nakagawa et al ., ; Itoh et al ., ). The direct effect of fertility on mortality may reflect a lack of drought resistance in trees predominating in the wetter habitats.…”

Long-term effects of selective logging on dipterocarp populations in the Pasoh Forest Reserve, Malaysia

“…This drought increased community level mortality rate for lowland tropical rain forests (Aiba and Kitayama 2002;Potts 2003). Especially, Dipterocarpaceae were vulnerable to the drought event (Nakagawa et al 2000;Itoh et al 2012). Similar mortality rates might suggest a higher background mortality rate (mortality in the period without droughts) of trees in the logged forest than those in the primary forest.…”

Calculation of narrower confidence intervals for tree mortality rates when we know nothing but the location of the death/survival events

“…However, it is most common for ecologists to report confidence intervals for mortality rates as if they knew nothing but k and n, under the assumption of constant probability of survival either by using different analytical approximations to the binomial distribution (Condit et al, 2006(Condit et al, , 1995Condit, Hubbell, & Foster, 1993;Davies, 2001;Gilbert, Wright, Muller-Landau, Kitajima, & Hernández, 2006;Itoh et al, 2012;King, Davies, & Noor, 2006;Lewis et al, 2004;Nepstad, Tohver, Ray, Moutinho, & Cardinot, 2007;Queenborough, Burslem, Garwood, & Valencia, 2007;Welden, Hewett, Hubbell, & Foster, 1991) or by bootstrapping the binary observations directly (van Breugel et al, 2011;Thomas et al, 2013). However, it is most common for ecologists to report confidence intervals for mortality rates as if they knew nothing but k and n, under the assumption of constant probability of survival either by using different analytical approximations to the binomial distribution (Condit et al, 2006(Condit et al, , 1995Condit, Hubbell, & Foster, 1993;Davies, 2001;Gilbert, Wright, Muller-Landau, Kitajima, & Hernández, 2006;Itoh et al, 2012;King, Davies, & Noor, 2006;Lewis et al, 2004;Nepstad, Tohver, Ray, Moutinho, & Cardinot, 2007;Queenborough, Burslem, Garwood, & Valencia, 2007;Welden, Hewett, Hubbell, & Foster, 1991) or by bootstrapping the binary observations directly (van Breugel et al, 2011;Thomas et al, 2013).…”