Forest successional processes following disturbance take decades to play out, even in tropical forests. Nonetheless, records of vegetation change in this ecosystem are scarce, increasing the importance of the chronosequence approach to study forest recovery. However, this approach requires accurate dating of secondary forests, which until now was a difficult and/or expensive task. Cecropia is a widespread and abundant pioneer tree genus of the Neotropics. Here we propose and validate a rapid and straightforward method to estimate the age of secondary forest patches based on morphological observations of Cecropia trees. We found that Cecropia-inferred ages were highly correlated with known ages of the forest. We also demonstrate that Cecropia can be used to accurately date disturbances and propose twenty-one species distributed all over the geographical range of the genus as potential secondary forest chronometer species. Our method is limited in applicability by the maximal longevity of Cecropia individuals. Although the oldest chronosequence used in this study was 20 years old, we argue that at least for the first four decades after disturbance, the method described in this study provides very accurate estimations of secondary forest ages. The age of pioneer trees provides not only information needed to calculate the recovery of carbon stocks that would help to improve forest management, but also provides information needed to characterize the initial floristic composition and the rates of species remigration into secondary forest. Our contribution shows how successional studies can be reliably and inexpensively extended without the need to obtain forest ages based on expensive or potentially inaccurate data across the Neotropics.
Citation: Zalamea P-C, Heuret P, Sarmiento C, Rodríguez M, Berthouly A, Guitet S, et al. (2012) The Genus Cecropia: A Biological Clock to Estimate the Age of Recently Disturbed Areas in the Neotropics. PLoS ONE 7(8): e42643. https://doi.org/10.1371/journal.pone.0042643
Editor: Ben Bond-Lamberty, DOE Pacific Northwest National Laboratory, United States of America
Received: May 15, 2012; Accepted: July 11, 2012; Published: August 10, 2012
Copyright: © Zalamea et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by an Ecos-Nord-Colciencias and Paris 13 University grant (C08A01), and the Institut de Recherche pour le Développement (France) doctoral grant to PCZ. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Neotropical forests are the greatest terrestrial reservoirs of biodiversity and carbon . Despite prolonged attention of researchers to habitat degradation, forest fragmentation and deforestation , currently tropical forests are the most threatened ecosystems worldwide . High rates of deforestation and selective logging in turn have resulted in a large increase in secondary forest areas . The resilience of secondary forests is poorly understood, although Norden et al.  showed that secondary forests could serve as habitat refugia and biodiversity reservoirs of tree species from mature tropical forests.
A common problem with studies of secondary forest ecology and succession, however, has been uncertainty over stand age. Tropical secondary forest succession is a process that takes several decades, and to our knowledge, no continuous data on vegetation change are available for secondary forests over periods longer than 30 years. Consequently, the chronosequence approach, where disturbance ages of secondary forests are known, has been frequently used to study secondary forest dynamics . Nevertheless, due to the difficulty of assessing the ages of secondary forest areas , there is a clear need for accurate and repeatable methodologies that allow us to date the age of disturbances over decadal time scales. More recently, attention has also been given to the need to monitoring and quantification of carbon sequestration processes to implement mechanisms for Reducing Emissions from Deforestation and Forest Degradation (REDD). However, one of the critical limitations in doing so is accurately estimating forests ages .
Currently, the most frequently used methods to date secondary forest are interviews of local people, use of remote-sensing techniques, probabilistic approaches based on tree growth rates, 14C dating, and tree ring analysis , –. All of these methods have limitations . For example, information from local people is not always available, and its accuracy is highly variable . Remote sensing or 14C dating are inaccurate and/or expensive techniques that may only work when suitable material is available.
Here, we demonstrate that Cecropia, a widespread and abundant pioneer genus in the Neotropics, is an accurate chronometer for dating different types of disturbance. We base this demonstration on a causal chain of previous observations, i.e. the Cecropia developmental periodicity allows predicting the age of individuals of this genus –. In secondary forests, the age of individuals of pioneer tree species reflects age of the forest and thereby age of the disturbance, since these species are incapable of germination following canopy closure . In this study, we tested whether our estimation of Cecropia age based on two species and three different types of disturbance was reliable to the age of disturbance by studying the relationship between the estimated age of Cecropia and the real age of the disturbance. In addition, we propose twenty-one Cecropia species distributed all over the geographical range of the genus as potential chronometer species.
Materials and Methods
Procedures and permits for measuring living Cecropia trees for the slash-and-burn agricultural sites in Colombia and French Guiana were approved and obtained from private local owners on each locality. Procedures and permits for measuring and handling living Cecropia trees on the gold mining sites and the forestry road were approved and obtained from the French Guiana ONF (Office National des Forêts), which is in charge of building the logging road-network and controlling the environment on the mining sites.
In this study, we focused on two Cecropia species characterized by different geographic distributions. C. sciadophylla is a widespread species distributed throughout the Amazon basin, the Llanos region of Colombia and Venezuela, and the Guiana shield, while C. obtusa has a limited distribution on the Guiana shield and lower Amazon basin .
We gathered data for three different types of disturbance: i) slash and burn agriculture, ii) a forestry road, and iii) gold mining sites. For the slash-and-burn agricultural sites, stand ages were obtained through interviews with local people, while for the forestry road and the mining sites, ages were obtained via the French Guiana ONF databases, ensuring the accuracy of the disturbance age. For the slash-and-burn chronosequences, data were gathered in three different sites, two in Colombia and one in French Guiana. i) The first Colombian chronosequence was located ∼11 km from Leticia, in an agricultural matrix composed of mature forest, secondary forest, and native crop stands (Huitoto native community), located in the department of Amazonas (4°6′S, 69°57′W). The Leticia chronosequence was composed of 8 age classes between 2 and 18 yrs. old. ii) The second Colombian chronosequence was located at the municipality of La Primavera, a natural savanna located in the department of Vichada (5°24′N, 69°53′W); the landscape matrix was composed by savanna vegetation and gallery forest, where some years ago traditional slash-and-burn agricultural practices were developed. The stands established at La Primavera were composed of 5 age classes between 5 and 20 yrs. old. iii) The French Guiana chronosequence was located at Sparwine town in an agricultural matrix composed of mature forest, secondary forest, and native crop stands (near the border between French Guiana and Suriname, 5°16′N, 54°14′W). The Sparwine chronosequence was composed of 9 age classes between 1 and 14 yrs. old.
Data on the forestry road and mining disturbances were also gathered in French Guiana. The Counamama 60 km-long forestry road is located between the municipalities of Sinnamary and Iracoubo (5°18′N, 53°13′W). This road was constructed as an access way for a logging concession. Its construction started in 1989 and finished in 2006; during the construction process, different sections were opened successively generating a chronosequence composed of nine age classes that varied between 1 and 17 yrs. old. The Coralie gold mining sites were located between the municipalities of Roura and Regina in French Guiana (4°29′N, 52°27′W). Six sites corresponding to 6 different age classes that varied between two and eight years old were chosen. In order to avoid any bias generated by the prior knowledge of stand ages, the real stand ages were always obtained after the estimation of Cecropia age.
Plant material and measurements
In the French Guianan localities a total of 122 individuals of C. sciadophylla and 287 individuals of C. obtusa were measured, while in the Colombian localities a total of 80 individuals of C. sciadophylla were measured. Only straight trees without any evident trauma were selected. Although C. sciadophylla and C. obtusa are the most abundant Cecropia species in French Guiana, the proportion of the two species is not the same in all sites. The abundance of C. sciadophylla individuals relative to C. obtusa individuals is lower in mining and slash-and-burn agricultural sites.
Previous studies of two species, C. obtusa and C. sciadophylla, showed a high annual periodicity in reproductive and branching processes, as well as an annual alternation of long and short nodes –. Throughout plant ontogeny and over a wide geographic gradient, our previous results show that the production of ∼35 and ∼23 nodes per year is a remarkably stable trait for C. obtusa and C. sciadophylla, respectively. These results have shown that the ages of individual trees can be estimated through observations of external morphology. This occurs because the scars of fallen leaves, inflorescences and branches remain visible along the trunk in Cecropia species (Fig. 1) –. In this study we will use our previous results as a background to test the hypothesis that Cecropia could be used as a biological clock to estimate the age of recently disturbed areas.
Cecropia age estimation protocol consists in dividing the total number of nodes on the main axis by 23 for C. sciadophylla or 35 for C. obtusa. In the pictures (a), (b), and (c) leaf and stipule scars are shown at different heights in the tree. (d) Relationship between the number of nodes on the borne axes (An) and that of their bearing axis (An-1) above their point of insertion, for the same individual. The comparison of A2 in relation to A1 is represented by circles, A3 in relation to A2 by squares, A4 in relation to A3 by triangles and A5 in relation to A4 by “x” symbols. The dotted line in the panel (d) represents the 1∶1 line. The human silhouette represents a 1.8 m height scale.
The protocol that we propose for estimating the age of Cecropia individuals is thus based on morphological description of the main axis from the base to the apex (Fig. 1). The protocol consists in dividing the total number of nodes on the main axis by 23 for C. sciadophylla or 35 for C. obtusa (Fig. 1). When trees are higher than 2–3 m the use of binoculars is necessary to count the number of nodes. In order to determine the accuracy of the counting of nodes using binoculars, the total number of nodes was counted in 91 individuals (56 C. obtusa and 35 C. sciadophylla) at the Counamama forestry road using binoculars, and then those individuals were felled and the total number of nodes was manually re-counted. Regardless of the Cecropia species, we found that the total number of nodes counted using binoculars was highly correlated with the actual total number of nodes (regression slope = 0.962, intercept = 0.172, R2 = 0.983, P<0.001). In addition, we observed that for C. obtusa and C. sciadophylla individuals, the number of nodes per axis was positively correlated with the number of nodes of the bearing axis above its insertion point, suggesting that the emission of new nodes occurs at the same rate in all axes of a given individual (, ; see Fig. 1d). Thus, for dating purposes, if the main axis is dead or broken, it is possible to estimate the age by following the main axis and then counting nodes on a vigorous branch.
Given that annual flowering is one of the traits used to estimate the age of C. obtusa and C. sciadophylla individuals –, it is crucial to determine whether our chronometric approach is likely to be broadly applicable across Cecropia species. Zalamea et al.  used a herbarium dataset to characterize the reproductive phenology of 35 Cecropia species, and found that 21 of them are characterized by an annual flowering. Here we used this information, to generate the geographic distribution map of species presence. To do that, we divided the area comprised from 24°N to 31°S and from 104°W to 34°W into hexagons of 4° side, and then, based on 2244 herbarium vouchers, we mapped the presence of annually flowering species by hexagons using ArcMap .
The relationship between the estimated age of Cecropia and the real age of disturbances was assessed using linear-mixed effect models, including the locality of each tree as a random factor. We also included the following covariates as fixed effects: species, disturbance type, interaction between the real age of the disturbance and Cecropia species, and interaction between the real age of the disturbance and disturbance type. To determine a confidence interval for the correlation coefficient between the estimated Cecropia age and the real age of the disturbance, we performed a Pearson correlation analysis between paired samples, and used a bootstrap procedure with 10000 replicates. Cecropia estimated age and real age of the disturbance were both square-root transformed to meet normality and homoscedasticity assumptions. We used restricted maximum likelihood estimation in the mixed-effect model, and two-tailed tests with a significance level set to α = 0.05. All analyses were made in R software [version 2.13.1; 17] using the nlme [version 3.1; 18] package.
We found that Cecropia estimated age was highly correlated with the disturbance real age (F1,31 = 701.8; P<0.001; Pearson's correlation: r2 = 0.97, P<0.001; Fig. 2a), the slope is not different from 1 (slope = 1.04), and the intercept is different from zero (intercept = −0.26). This relationship was neither different between the two Cecropia species nor among the disturbance types (interaction P values≥0.152). Although Cecropia estimated age and real age of the disturbance were both square-root transformed to meet normality and homoscedasticity assumptions, we also present untransformed data in figure 2b. To determine a confidence interval for the correlation coefficient between estimated Cecropia age and disturbance real age we performed a non-parametric bootstrap. The size of the confidence interval was very small (95% CI: 0.965–0.976), indicating that the uncertainty in our correlation coefficient is small.
(a) Both ages were square-root-transformed to meet normality and homoscedasticity assumptions. The dotted line represents the equation adjusted by the linear mixed-effect model. For ease of interpretation by the reader we also present the untransformed data in the panel (b). Black diamonds represent C. sciadophylla and grey diamonds represent C. obtusa individuals. For the Colombian sites we described 52 C. sciadophylla individuals at Leticia, while 28 were described at La Primavera. For the French Guiana sites, we described 20 C. sciadophylla and 137 C. obtusa individuals at Sparwine, 84 C. sciadophylla and 82 C. obtusa individuals at Counamama, and 18 C. sciadophylla and 68 C. obtusa individuals at Coralie (see the methods section for a detailed study site description).
Based on the presence of 21 Cecropia species that show annual flowering , we generated a distribution map to highlight some of the potential Cecropia chronometers available across the Neotropics (Fig. 3 and Table 1).
The map illustrates the number of annually flowering Cecropia species identified by Zalamea et al. . The colors represent the number of species present in each hexagon of 4° side (see Table 1 for a list of species).
The correlation coefficient of the relationship between Cecropia estimated age and disturbance real age was very close to one. Moreover, in a previous study Zalamea  showed that the production of new nodes is extraordinarily stable throughout plant ontogeny and among populations of C. sciadophylla, both within and among years. This stability was also observed despite a high variation in annual precipitation and a wide geographic range among the study sites, suggesting a strong genetic control and a weak climatic or environmental influence on the observed periodicity. Together, both results show the high potential of the genus Cecropia as a proxy to estimate the age of secondary forests in the Neotropics.
In the relationship between Cecropia estimated age and disturbance real age (Fig. 2), the intercept was negative and close to zero, suggesting a small lag in recruitment of Cecropia individuals after perturbation. In anthropogenic perturbations, such as slash-and-burn agriculture, mining exploitation, and road construction, soil is highly disturbed, and seed banks could be affected or eliminated. For example, fires lasting five hours, such as those commonly applied when clearing land for agriculture, are sufficient to eliminate the C. sciadophylla seed bank up to a depth of 15 cm (Zalamea & Stevenson unpublished data). Cecropia is common on disturbed areas affected by fire and mining, indicating that recruits may come from seed rain rather than soil seed bank. Thus, the small lag in recruitment could be a response by timing of disturbance relative to seed rain, yet this lag is no longer than a year. In Barro Colorado Island, Panama, most C. insignis germination occurred in natural gaps within the first year after gap opening , and in central Amazonia, C. sciadophylla and C. purpurascens form even-aged stands in gaps resulting from different land uses , suggesting that Cecropia recruitment occurs in a single pulse, early in the regeneration process. Although several factors could delay recruitment, the intercept close to zero show that this delay in our database, is short in time, usually a few months.
Our method is limited in applicability by the maximal longevity of Cecropia individuals. Assessing individual longevity in tropical trees is challenging –, and in the case of Cecropia species, longevity data are scarce. In Cecropia, these estimations vary among 35 years for C. obtusifolia , 54 years for C. sciadophylla  and 96 years for C. insignis . Although the oldest chronosequence used in this study was 20 years old, in a recent study we have dated a 34 years old C. sciadophylla individual in French Guiana . Thus, at least for the first four decades after disturbance, the method described in this study provides very accurate estimations of secondary forest ages.
Although Cecropia is a widespread genus on the Neotropics, the method proposed here to estimate the age of disturbed areas could not be used yet for all the species of the genus. This is because we do not know how many nodes are produced per year for each Cecropia species. However, as flowering is one of the traits we used to estimate the age of Cecropia individuals in our previous studies –, annual flowering periodicity suggests high potential for Cecropia species to become secondary forest chronometers, as it has been proposed for C. obtusa  and C. sciadophylla . The distribution map of the annually flowering Cecropia species (Fig. 3) implies that the Cecropia dating method could be used across a wide geographical range (i.e. all the Neotropics). Preliminary results showed an annual growth periodicity for C. angustifolia, C. ficifolia, C. insignis and C. distachya (P.C. Zalamea and P. Heuret, unpublished data). These results reinforce the idea that Cecropia could be used as a chronometer to estimate the time since forest disturbance. Although this study was based in anthropogenic disturbances, this methodology could also be used to date natural gaps. However, the dense canopy occurring in natural gaps compared to anthropogenic disturbances complicates the Cecropia age estimation using binoculars (i.e. it is difficult to count all nodes along the main axis). In such cases, an accurate estimation could be made using a destructive method.
The age of pioneer trees not only provides information needed to calculate the recovery of carbon stocks that would help to improve forest management , but also provides information needed to characterize the initial floristic composition and the rates of species remigration into secondary forest , . In addition, this information could be useful to establish differences in substrate or soil fertility among successional stands and to model secondary forest population dynamics , . In the current Neotropical scenario of forest fragmentation and degradation, little is known about the variability of responses from different primary forest to degradation effects , . Thus, our contribution shows how successional studies can be reliably and inexpensively extended without the need to obtain forest ages based on remote sensing data or potentially inaccurate interview data across the Neotropics. In addition, growth-monitoring studies in secondary forests in the Neotropics are scarce and recent, suggesting that integrative approaches that use chronosequences and permanent plots can provide the valuable knowledge urgently needed in the current portfolio of tools for secondary forests management.
We thank F. Bompy, A. Oddi, J. Migeot, E. Bahamon, M. Peuch, T. Wiesinger, J.B. Gratecap, B. Larcher, P. Nlungu, L. Roux, and J. S. González for their assistance in the field. We are grateful to F. Munoz for enlightening discussions during data analyses and to C. Pedraza for his help in elaborating the map shown in Fig. 3. We thank J. Dalling, C.E.T. Paine, D. McKey, H. Beeckman, and J.P. Gómez for comments on an earlier version of this manuscript. We thank B. Bond-Lamberty and two anonymous reviewers. We also thank the Office National des Forêts in French Guiana and the local people at Leticia, La Primavera, and Sparwine who provided us the real disturbance ages from the studied sites.
Conceived and designed the experiments: PCZ PH PRS DB. Performed the experiments: PCZ MR CS SG CD EN. Analyzed the data: PCZ AB. Contributed reagents/materials/analysis tools: PCZ PH PRS SG. Wrote the paper: PCZ. Intellectual support and editorial input: PH CS PRS.
- 1. Malhi Y, Baker T, Phillips O, Almeida S, Alvarez E, et al. (2004) The above-ground coarse wood productivity of 104 Neotropical forest plots. Global Change Biol 10: 563–591.
- 2. Bawa KS, Dayanandan S (1997) Socioeconomic factors and tropical deforestation. Nature 386: 562–563.
- 3. Laurance WF (2007) Have we overstated the tropical biodiversity crisis? Trends Ecol Evol 22: 65–70.
- 4. Wright J (2005) Tropical Forests in a Changing Environment. Trends Ecol Evol 20: 553–560.
- 5. Norden N, Chazdon RL, Chao A, Jiang Y-H, Vílchez-Alvarado B (2009) Resilience of tropical rain forests: Tree community reassembly in secondary forests. Ecol Lett 12: 385–394.
- 6. Chazdon RL, Letcher SG, van Breugel M, Martínez-Ramos M, Bongers F, et al. (2007) Rates of change in tree communities of secondary Neotropical forests following major disturbances. Phil Trans R Soc B 362: 273–289.
- 7. Brienen RJW, Lebrija-Trejos E, van Breugel M, Pérez-García EA, Bongers F, et al. (2009) The potential of tree rings for the study of forest succession in southern Mexico. Biotropica 41: 186–195.
- 8. Sanchez-Azofeifa GA, Castro-Esau KL, Kurz WA, Joyce A (2009) Monitoring carbon stocks in the tropics and the remote sensing operational limitations: from local to regional projects. Ecol Appl 19: 480–494.
- 9. Martínez-Ramos M, Alvarez-Buylla ER (1998) How old are tropical rain forest trees? Trends Plant Sci 3: 400–405.
- 10. Neeff T, Lucas RM, dos Santos JR, Brondizio ES, Freitas CC (2006) Area and age of secondary forests in Brazilian Amazonia 1978–2002: an empirical estimate. Ecosystems 9: 609–623.
- 11. Heuret P, Barthélémy D, Guédon Y, Coulmier X, Tancre J (2002) Synchronization of growth, branching and flowering processes in the South American tropical tree Cecropia obtusa (Cecropiaceae). Am J Bot 89: 1180–1187.
- 12. Zalamea P-C, Stevenson PR, Madriñán S, Aubert P-M, Heuret P (2008) Growth pattern and age determination for Cecropia sciadophylla (Urticaceae). Am J Bot 95: 263–275.
- 13. Berg CC, Franco P (2005) Flora Neotropica Monograph 94. Cecropia. Organization for Flora Neotropica. Bronx, NY: The New York Botanical Garden Press.
- 14. Zalamea P-C (2010) Cecropia growth pattern periodicity: could a Neotropical genus be a good biological clock to estimate the age of disturbed areas? Ph.D. thesis, Université Montpellier 2, France.
- 15. Zalamea P-C, Munoz F, Stevenson PR, Paine CET, Sarmiento C, et al. (2011) Continental-scale patterns of Cecropia reproductive phenology: Evidence from herbarium specimens. Proc R Soc B 278: 2437–2445.
- 16. ESRI (2008) ArcMap 9.3. Redlands, CA: Environmental Systems Research Institute, Inc.
- 17. R Development Core Team (2011) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Available: http://www.Rproject.org. Accessed 2012 Jul 23.
- 18. Pinheiro J, Bates D, DebRoy S, Sarkar D (2009) the R Core Team (2009) nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1-92
- 19. Brokaw N (1987) Gap-phase regeneration of three pioneer tree species in a tropical forest. J Ecol 75: 9–19.
- 20. Mesquita RCG, Ickes K, Ganade G, Williamson GB (2001) Alternative successional pathways following deforestation in the Amazon Basin. J Ecol 89: 528–537.
- 21. Terborgh J, Flores C, Mueller P, Davenport L (1997) Estimating the ages of successional stands of tropical trees from growth increments. J Trop Ecol 14: 833–856.
- 22. Chambers J, Higuchi N, Schimel J (1998) Ancient trees in Amazonia. Nature 391: 135–136.
- 23. Martínez-Ramos M, Alvarez-Buylla ER (1999) Reply to Chambers and Trumbore… Tropical rain forest tree life-history diversity calls for more than one aging method. Trends Plant Sci 4: 386.
- 24. Laurance WF, Nascimento HEM, Laurance SG, Condit R, D'Angelo S, et al. (2004) Inferred longevity of Amazonian rainforest trees based on a long-term demographic study. Forest Ecol Manag 190: 131–143.
- 25. Korning J, Balslev H (1994) Growth rates and mortality patterns of tropical lowland tree species and the relation to forest structure in Amazonian Ecuador. J Trop Ecol 10: 151–166.
- 26. Condit R, Hubbell S, Foster R (1993) Identifying fast-growing native trees from the Neotropics using data from a large, permanent census plot. Forest Ecol Manag 62: 123–143.
- 27. Lugo AE (2009) The emerging era of novel tropical forests. Biotropica 41: 589–591.