INSIGHTS Wanted: new allometric equations for large lianas and African lianas Zewei Miao* 1 , Sally E. Koerner* 1 , Vincent P. Medjibe* 1,2 , and John R. Poulsen* 1,2,3 1 Nicholas School of the Environment, Duke University, P.O Box 90328, Durham, NC 27708, USA 2 Agence Nationale des Parcs Nationaux, Batterie IV, BP. 20379 Libreville, Gabon ABSTRACT Liana abundance appears to be increasing, possibly to the detriment of trees, but methods for measuring liana biomass are undepend- able. We show that five commonly used allometric equations produce disparate results and discuss two large information gaps—Central African lianas and large lianas—that currently preclude accurate liana biomass estimation. Abstract in French is available with online material. Key words : aboveground biomass; Afrotropics; carbon sequestration; Gabon; tropical forest; tropics. LIANAS (WOODY VINES) HAVE LARGE INDIRECT NEGATIVE EFFECTS ON GLOBAL CARBON STORAGE DUE TO INTENSE COMPETITION WITH TREES (Schnitzer & Bongers 2011, Tobin et al. 2012). Although lianas have historically accounted for just 4–5 percent of the total aboveground biomass (AGB) in moist tropical forests (Putz 1983, DeWalt & Chave 2004), lianas are increasing in abundance (Phillips et al. 2002, Wright et al. 2004) in response to various glo- bal change factors. Phillips et al. (2002) documented a doubling of liana abundance over a 20-yr period in Amazonia, Northwest South America, and Central America, with increases occurring in both small and large lianas. Numerous studies from the Neotrop- ics (e.g., Ingwell et al. (2010), Laurance et al. (2014)) now support this increasing liana hypothesis (ILH), with forest fragmentation, logging, changes in temperature and seasonality, and increased atmospheric CO 2 often invoked as causal factors (Phillips et al. 2002, Schnitzer 2015). Since most of the tropics are severely threatened by at least one, if not all, of these factors (Malhi et al. 2013), the liana contribution to total AGB is likely to grow over the coming decades. Lianas substantially decrease tree diversity, recruitment, growth, and survival (van der Heijden & Phillips 2009, Ingwell et al. 2010, Schnitzer & Carson 2010), and thus increasing liana abun- dance may not only restructure tropical forests but also drastically modify carbon dynamics (Duran & Gianoli 2013, Schnitzer et al. 2014). Dependable methods for quantifying liana AGB are needed to accurately estimate both their direct and indirect effects on total AGB in tropical forests (Schnitzer et al. 2006), yet these methods are largely lacking. Current approaches rely on allometric equations to predict liana AGB from liana diameter. While similar equations are used extensively in forestry to measure trees, the same allomet- ric relations are not applicable to lianas (Schnitzer et al. 2006). Lianas have a much greater canopy biomass relative to stem bio- mass than trees because their stems are non-structural (Putz 1983, Kazda et al. 2009). In addition, liana growth is more erratic, often looping around the understory, unlike the relatively linear growth of trees (Schnitzer et al. 2006). Yet, although allometric measure- ments for lianas may be more complicated than trees, the only alter- native is destructive sampling and weighing of the liana. As a result of the paucity of destructive sampling studies, most studies of liana biomass rely on allometric equations developed at the regional or global level (Addo-Fordjour & Rahmad 2013a). Numerous allo- metric equations have been put forward, yet rarely are they com- pared (but see Addo-Fordjour & Rahmad 2013a). Here, we compared the performance of five previously pub- lished liana allometric equations (Table 1) in estimating large liana (≥10 cm diameter at breast height [dbh]) AGB in Gabon, Africa. If a given region, like Central Africa, does not have a local allo- metric equation for lianas, will pantropical and other regional allo- metric equations produce significantly different estimates? In 2012, Gabon initiated the establishment of a national resource inventory (NRI), a network of 104 permanent plots (100 9 100 m) each located randomly within a cell created by a 50 9 50 km grid laid out across the country. The locations span a wide precipitation gradient and are located in old growth, sec- ondary, and logged forest. Across the 104 plots, we recorded 1354 large lianas for which we measured the diameter at 130 cm from the roots (D 130 ). These measurements were used to calculate plot- level liana AGB using five different allometric equations (Table 1). While we used our specific dataset, liana measurements from any site or study could be used to examine the variability in liana AGB estimates produced by various allometric equations. The most commonly used pantropical equation (PT1) for liana AGB was developed by Schnitzer et al. (2006) using data from five independent data sets in four countries (Brazil [x2], French Gui- ana, Cambodia, and Venezuela). The equation, which estimates liana AGB from stem diameter at 130 cm from the roots, Received 15 December 2015; revision accepted 25 April 2016. *Authors contributed equally. 3 Corresponding author; e-mail: john.poulsen@duke.edu ª 2016 The Association for Tropical Biology and Conservation 561 BIOTROPICA 48(5): 561–564 2016 10.1111/btp.12353