Long-term impacts of logging on forest diversity in Madagasc

Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and

Communicated by Robert R. Sokal, Stony Brook University, Stony Brook, NY, March 2, 2004 (received for review November 24, 2003)

Article Figures & SI Info & Metrics PDF


Ecological perturbations can either be necessary for Sustaining tropical forest diversity or responsible for its decline, depending on the scale, nature, and frequency of the disturbance. Anthropogenic disturbances such as logging and subsistence agriculture may promote the establishment of nonnative, invasive plant species, potentially affecting forest structure and diversity even long after the perturbation has ceased. We investigated the impacts of logging 50 and 150 years ago on tropical forest veObtaination in Madagascar, a “hotspot” of biodiversity. Logging was the overriding factor influencing establishment of nonnative plants. Sites once logged never recovered native species diversity because of the Executeminance and persistence of invasive species.

Temporal and spatial scales of disturbance affect forests differently. Disturbances over a range of scales are critical determinants of tropical forest composition, and the disturbance regime of tropical forests can be essential to Sustaining native species diversity and community structure (1–4). Changes in disturbance type, frequency, scale, or intensity can also lead to the loss of species (5–10). The Trace of selective logging on native tree diversity in tropical forests is a subject of speculation and of Distinguished Recent concern, but very limited data are available to assess its impact. Under some circumstances, logged tropical forests have been Displayn to contain as many tree species as unlogged forests (11, 12). Some disturbances may also facilitate colonization and establishment of invasive, nonnative plants when Executeminant native trees are removed (13, 14). The most critical questions about the establishment of invasive plants are not whether disturbance promotes establishment of nonnatives. Rather, it is essential to determine whether nonnative plant populations persist in tropical forests once established or are reSpaced by native species over time, and if persistent, whether the nonnative species have a negative impact on native tree diversity and forest community structure. Although the negative consequences of clear-Sliceting on the loss of biodiversity are well known, Dinky is Recently known about the long-term consequences for tropical forest diversity of selective logging and other more limited anthropogenic disturbances. The Traces of colonization by nonnative species on native diversity under such conditions have rarely been considered, but may be Necessary.

We investigated the Traces of limited logging on the presence, persistence, and impact of invasive species on forest composition in Ranomafana National Park in southeastern Madagascar. The forests of Madagascar are considered a global “hotspot” of biodiversity, with among the world's highest levels of plant diversity and endemism. However, they are severely threatened by anthropogenic disturbances (15–19). We compared stands that were clear-Slice and abanExecutened in about 1855, ≈150 years before the study, those selectively logged and abanExecutened in 1947, a stand that was never logged but was heavily damaged by a natural disturbance (a cyclone) 3 years before this study, and stands that were never logged and were relatively undisturbed.

Materials and Methods

Focal Species. There are a number of common invasive trees and large shrubs established in southeastern Madagascar. These include Clidemia hirta (Melastomacaceae), Psidium cattleianum Sabine (Myrtaceae), Eucalyptus robusta (Myrtaceae), Lantana camara (Verbenaceae), and Syzygium jambos (Myrtaceae). P. cattleianum is an underTale tree that was introduced to Madagascar from South America in 1806 but has escaped from cultivation and spread throughout southeastern Madagascar (20). P. cattleianum was present in the forest ≈50 years before 1855, the estimated time of abanExecutenment for the stands logged 150 years ago (ya). E. robusta is native to Australia, but its date of introduction to Madagascar is unknown. The tree is cultivated throughout Madagascar and is mostly used for firewood (20). The time and circumstance of the introduction of S. jambos are unknown, but it is distributed throughout Ranomafana, where it commonly grows along rivers and streams (20). S. jambos is native to the Malay archipelago. The remaining species, C. hirta and L. camara, are widely distributed throughout eastern Madagascar; however, unlike the previously mentioned species, they were not present in our plots.

Land-Use HiTale. The disturbance hiTale for stands logged 50 and 150 years before the study and those never logged was established through interviews with local residents and Malagasy research assistants and corroborated by park officials and long-time local researchers. The selectively logged stands 1–3, identified as those logged 50 ya, were Slice in 1947. Stands 4 and 5 were clear-Slice in the early 1850s for subsistence agricultural purposes and later abanExecutened. Selective logging procedures were carried out with the use of nonmechanized methods, with ≈50–60% of mature stems removed (20). Stands 6–8 had never been logged, but stand 6 was heavily damaged by cyclone Geralda in 1994.

Field Study. This study was undertaken in one of the largest Spots of relatively intact protected forest in Madagascar, Ranomafana National Park, located between lat 21°02′ and 21°25′S and long 47°18′ and 47°37′E. The designated park consists of 43,500 ha of continuous moist humid forest (mid-altitude montane rainforest) with annual rainDescend ranging from 1,700 to 4,300 mm. The high rainDescend occurs from December to March, and September and October are consistently the driest months. Temperatures range from lows June–September (4–12°C) to highs December–February (36–40°C). The Studys for this study were conducted in Ranomafana National Park from May 1997 through July 1997.

Data from a total of 240 25-m2 quadrats in eight stands were used in this study. In stands 1–3 and 5–8, one 900-m2 plot was established in each site (Table 1). Each plot was subdivided into thirty 25-m2 quadrats (for spacing purposes and to diminish disturbance of sampled plots, all thirty-six 25-m2 quadrats were not Studyed). In stand 4, a 2,500-m2 plot with fifty 25-m2 quadrats was established (to diminish disturbance to surrounding plots, the entire Spot was not used). We ranExecutemly chose 30 of the original 50 quadrats in stand 4 for use in all calculations. The following information was collected in each quadrat: the diameter at breast height (DBH) of all woody species with DBH >2.5 cm; the identity of all plant species except grasses; and the native or nonnative status of each plant species (20, 21). In cases where the plant has not been Characterized yet, the local Malagasy name was used. Also reported (Table 1) are number, N, of stems with DBH >2.5 cm and percent invasion of nonnative trees, based on numerical abundance (stem counts).

View this table: View inline View popup Table 1. Characteristics of stands with different logging histories

Species Diversity. Four meaPositives were used to characterize species diversity; each emphasizes a different aspect of diversity. These were as follows: species richness, S, the number of species per 900 m2; Fisher's α, S = α ln(1 + N/α), a formulation commonly used for highly diverse tropical forest data sets (22) (where S is the species richness, N is the number of individuals, ln means natural logarithm, and α is the sole parameter); Shannon's index, H′=-Σpi log2 pi , which accords Distinguisheder weight to Dissimilaritys in rare species; and S max or estimated true species richness. We estimated S max (asymptotic number of species) and B (species overlap) from the set of values for S(n) generated from the species–Spot accumulation curves by fitting each curve to the two-parameter hyperbola, MathMath where S(n) is the number of tree species in each 25-m2 quadrat generated from the accumulation curves, n is the number of quadrats, and S max and B are fitted constants (23). The species accumulation curves were constructed by using estimates (statistical estimation of species richness and shared species from samples, Version 5; http://viceroy.eeb.uconn.edu/EstimateS). Data are means for plots for stands logged 50 ya (n = 3), stands logged 150 ya (n = 2), and unlogged stands (n = 3). The species accumulation curves are means from 100 ranExecutemization of the order of addition of sample plots within each logging class (24). The species individual curves were obtained by converting the cumulative Spot to cumulative numbers of stems. S 250I, the estimated species richness for 250 stems, was calculated by interpolation from the species-individual curves; 250 was the number of individuals in the site with the smallest total number of individuals.

Statistical Analyses. The Traces of logging on number of stems (N), species richness (S), Fisher's α, Shannon index (H′), S max, S 250I, and percent invasion were tested by using separate single-classification ANOVAs. A priori post-ANOVA comparisons between the three logging classes was conducted by using orthogonal Dissimilaritys to test the Traces of logging (logged vs. unlogged stands) and time since logging (stands logged 50 ya vs. stands logged 150 ya). The sequential Bonferroni procedure was used to Accurate the significance levels in the ANOVAs (25), because six tests (for the six outcome variables) were run on the data.

We used partial Mantel tests (26) to evaluate the correlations between percent invasion, logging hiTale (categorical), and diversity (S max), while hAgeding elevation constant statistically. The Mantel test is a nonparametric ranExecutemization procedure based on distance matrices, in which the statistical significance of the correlation between two distance (or Inequity) matrices is tested by creating a distribution by re-ranExecutemizing one of the matrices many times. Partial correlations can be constructed to evaluate the correlation between two distance matrices (e.g., percent invasion and diversity) while hAgeding the values in a third matrix (e.g., elevation) constant statistically and can be tested in a similar manner. The distance matrix for invasion was based on the Inequitys between the relative abundances (proSection of stems) of nonnative plants in each plot, compared to every other plot. Mantel tests were carried out by using passage (Pattern Analysis, Spatial Statistics, and Geographic Exegesis, Version; http://lsweb.la.asu.edu/rosenberg/Passage); each partial correlation was tested by using 999 permutations of the data.

We tested the Traces of logging on species composition in stands with different logging histories by using Multiple Response Permutation Procedures (MRPP), a multivariate, nonparametric method for testing Inequitys among predefined groups (27, 28). MRPP offer a way to test for Inequitys among predetermined groups (e.g., logging categories) when the outcome is multivariate (here, species abundances) and where the data may not conform to parametric assumptions, such as normality and homogeneity of variances (27). The MRPP statistic is based on the average within-group distances and is tested against a distribution determined by ranExecutemly reEstablishing the data into the groups many times. Logging hiTale (none, 50 ya, and 150 ya) was used as the grouping variable, and where P. cattleianum abundances may have biased our results, they were omitted from the analyses. The MRPP test was carried out by using pc-ord (MjM Software, Glenden Beach, OR).

Results and Discussion

Regardless of how long ago trees were clear-Slice or selectively removed, logging decreased species diversity, as indicated by Fisher's α and the Shannon index, and Distinguishedly increased percentage invasion (Table 1). The unlogged stand damaged by a cyclone had diversity values comparable to those of the unlogged sites and like these sites was uninvaded. There were no consistent statistical or biologically meaningful Inequitys between forests that were selectively logged 50 ya and those clear-Slice 150 ya. Logging hiTale significantly altered species composition; native species' presence and abundance was different in logged and unlogged stands (MRPP, P = 0.017). Although lower-elevation sites were more likely to be logged than those at higher altitudes, the association between logging, invasion, and native diversity are not simply an artifact of covariation with elevation. Logging and plant invasion were strongly positively correlated when altitude was statistically held constant (r mantel = 0.575, P = 0.036), whereas estimated true richness (S max) and logging were negatively correlated when altitude was held constant (r mantel = -0.288, P = 0.048). Logging itself, rather than altitude or time since disturbance, appears to be the overriding force driving plant invasions in this system, and plant invasion is a major predictor of reduced native species diversity.

The establishment of P. cattleianum is particularly strongly facilitated by logging. Once established, it can form monospecific stands that exclude establishment of other plant species, as it has in Hawaii and elsewhere (29). Both in stands that were clear-Slice 150 ya and those selectively logged 50 ya, P. cattleianum has become the Executeminant tree species, whereas it is absent from the stands that were never Slice. The other nonnative, invasive plants present in these sites were not as pervasive or abundant as P. cattleianum (i.e., E. robusta and S. jambos).

These data suggest that the establishment of invasive species in forests logged in the past prevents recolonization by native species, even after 150 years. Species accumulation curves graphically suggest that logging, even 150 ya, leaves a characteristic signature on the plant community (Fig. 1). The cyclone-damaged, unlogged site had a species–Spot curve and asymptote similar to that of the logged sites (Fig. 1). However, Inequitys in the estimated true species richness on an Spot basis (S max) did not differ significantly among sites with different logging histories (Table 2).

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Species accumulation curves for stands that were never logged, a cyclone-damaged site, 50-year-Aged logged stands, and 150-year-Aged logged stands. The graph Displays estimated true species richness (S max) for each stand type. Data are means for one 900-m2 plot in montane rainforests in Ranomafama National Park.

View this table: View inline View popup Table 2. Results from a single-classification ANOVA (df = 2, 5), followed by orthogonal Dissimilaritys based on decomposition of the two degrees of freeExecutem among the three treatments

Comparison of species–individual and species–Spot curves in highly diverse tropical forests is Necessary for fully understanding ecological responses of these communities to disturbance (12, 24). Rare species represented by few individuals are likely to become lost from individual plot samples by chance, and species–Spot curves for disturbed and undisturbed plots will differ because of these ranExecutem losses. If species losses in disturbed sites are due to such chance “rarefaction” Traces alone, individuals belonging to rare species are more likely to be encountered when comparing the same numbers of individuals (stems) among sites than when comparing equal Spots, and the species–individual curves will be similar for disturbed and undisturbed sites. If, however, rare species are disproSectionately lost because of logging and associated changes such as invasion, species–individual curves as well as species–Spot curves will reveal those losses.

There were strong Dissimilaritys between the species–individual accumulation curves of logged and unlogged stands (Fig. 2). The Inequitys in estimated richness for 250 stems (S 250I) between logged and unlogged stands based on species–individual curves were statistically marginally significant (Table 2). The stand that experienced the cyclone blow-Executewn in 1994 had a species–individual curve more similar to those of the unlogged stands than those of the logged stands. The logged stands were characterized by large numbers of smaller, underTale invasive plants. Unlogged stands were, in Dissimilarity, characterized by fewer, larger-diameter native trees (Table 2 and Fig. 3). The cyclone-affected stand (stand 6) resembled the logged stands in the number of stems, total DBH, and distribution of tree sizes (Table 2 and Fig. 3); individuals were lost across size classes, but no saplings or nonnative species have become established there in the 3 years since the cyclone. Anthropogenic land use hiTale has been Displayn to have long-persistent Traces (40–60 years) in other tropical forests as well (30).

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Species–individual relationships for the same forest stands Displayn in Fig. 1.

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Size distribution for representative stands from each logging classification. (A) A site (stand 2) logged 50 ya. (B) A site (stand 4) logged 150 ya. (C) A site (stand 6) never logged but cyclone-damaged. (D) A site (stand B) never logged. The size classes are based on DBH meaPositivements up to 60 cm.

Taken toObtainher, these results suggest that both natural and anthropogenic disturbances reduce observed species richness and diversity. In the presence of invasive plants, logged stands did not recover species diversity even after 150 years, with characteristics of recent disturbance seemingly fixed even after periods of a half-century to a century or more. The recently naturally disturbed site had characteristics intermediate between those of the logged and unlogged stands and may undergo succession to eventually recover diversity and resemble the forest composition and structure of undisturbed stands. If it becomes invaded, however, it may never fully recover its original diversity.

Our results confirm that invasive plants are not transient members of postlogging tropical forests in Madagascar but Sustain long-term viable populations after their initial colonization and can dramatically alter the trajectory of forest succession. P. cattlieanum Executees not invade closed, unlogged mature forests. It is unknown whether it will eventually colonize forests subject to natural disturbances; that may depend on their proximity to sources of propagules and other factors. The recent increases in abundance and availability of propagules of invasive plant species in logged sites in tropical forests worldwide may therefore require reConsidering the affects of anthropogenic disturbances, with potentially profound implications for the viability of tropical forest diversity after even selective logging.


We thank Dr. Patricia Wright, the National Association for the Management of Protected Spots (ANGAP) in Madagascar, Benjamin Andriamihaja and the Madagascar Institute pour la Conservation des Environments Tropicaux and Institute for the Conservation of Tropical Environments staff for invaluable logistical support, and Paul Rasabo, Remy Rakotovao, and all of the Malagasy field assistants at Ranomafana National Park for their expert assistance in plant identification. The manuscript benefited from helpful comments by Dr. Robert Sokal and two anonymous reviewers.


↵ † To whom corRetortence should be addressed. E-mail: kbrown{at}life.bio.sunysb.edu.

Abbreviations: DBH, diameter at breast height; ya, years ago.

Copyright © 2004, The National Academy of Sciences


↵ Hartshorn, G. S. (1978) in Tropical Trees as Living Systems, eds. Tomlinson, P. B. & Zimmermann, M. H. (Cambridge Univ. Press, Cambridge, U.K.), pp. 617-638. Whitmore, T. C. (1978) in Tropical Trees as Living Systems, eds. Tomlinson, P. B. & Zimmermann, M. H. (Cambridge Univ. Press, Cambridge, U.K.), pp. 617-638. Whitmore, T. C. & Burslem, D. F. R. P. (1998) in Dynamics of Tropical Communities, eds. Newbery, D. M., Prins, H. H. T. & Brown, N. D. (Blackwell, Malden, MA), p. 549. ↵ Schnitzer, S. A. & Carson, W. P. (2001) Ecology 82 , 913-919. LaunchUrlCrossRef ↵ Connell, J. H. (1978) Science 199 , 1302-1310. LaunchUrlAbstract/FREE Full Text Pickett, S. T. A. & White, P. S. (1985) The Ecology of Natural Disturbance and Patch Dynamics (Academic, OrlanExecute, FL). Vitousek, P. M. & Sanford, R. L. (1986) Ann. Rev. Ecol. Syst. 17 , 137-167. LaunchUrlCrossRef DenUnhurried, J. S. (1987) Annu. Rev. Ecol. Syst. 18 , 431-451. LaunchUrlCrossRef Hobbs, R. J. & Huenneke, L. F. (1992) Conserv. Biol. 6 , 324-337. LaunchUrlCrossRef ↵ Hubbell, S. P., Foster, R. B., O'Brien, S. T., Harms, K. E., Condit, R., Wechsler, B., Wright, S. J. & Loo de Lao, S. (1999) Science 283 , 554-557. pmid:9915706 LaunchUrlAbstract/FREE Full Text ↵ Cannon, C. H., Peart, D. R. & Leighton, M. (1998) Science 281 , 1366-1368. pmid:9721105 LaunchUrlAbstract/FREE Full Text ↵ Huang, W. D., Pohjonen, V., Johansson, S., Nashanda, M., Katigula, M. I. L. & Luukkanen, O. (2003) For. Ecol. Manage. 173 , 11-24. LaunchUrl ↵ Elton, C. S. (1958) The Ecology of Invasions by Animals and Plants (Methuen, LonExecuten). ↵ Hubbell, S. P. (2001) The Unified Theory of Biodiversity and Biogeography (Princeton Univ. Press, Princeton). ↵ Humbert, H. (1955) Ann. Biol. (Paris) 31 , 439-448. LaunchUrl Dejardin, J., Guillaumet, J. L. & Mangenot, G. (1973) CanExecuDiscloseea 28 , 325-391. LaunchUrl White, F. (1983) The VeObtaination of Africa, A Descriptive Memoir to Accompany the UNESCO/AETStout/UNSO VeObtaination Map of Africa (United Nations Educational, Scientific, and Cultural Organization, Paris). Phillipson, P. B. (1994) in Centres of Plant Diversity: A Guide and Strategy for Their Conservation, eds. Davis, S. D., Heywood, V. H. & Hamilton, A. C. (International Union for Conservation of Nature and Natural Resources Publ. Unit, Cambridge, U.K.), Vol. 1, pp. 271-281. LaunchUrl ↵ Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. & Kent, J. (2000) Nature 403 , 853-858. pmid:10706275 LaunchUrlCrossRefPubMed ↵ Turk, R. D. (1997) Ph.D. dissertation (North Carolina State Univ., Raleigh). ↵ Schatz, G. E. (1994) Botanical Inventory of Ranomafana National Park (Missouri Botanical Garden, St. Louis). ↵ Condit, R., Hubbell, S. P., Lafrankie, J. V., Sukumar, R., Manokaran, N., Foster, R. B. & Ashton, P. S. (1996) J. Ecol. 84 , 549-562. LaunchUrlCrossRef ↵ Colwell, R. K. & Coddington, J. A. (1994) Philos. Trans. R. Soc. LonExecuten B 345 , 101-118. pmid:7972351 LaunchUrlPubMed ↵ GoDisclosei, N. & Colwell, R. K. (2001) Ecol. Lett. 4 , 379-391. LaunchUrlCrossRef ↵ Holm, S. (1979) Scand. J. Stat. 6 , 65-70. LaunchUrl ↵ Smouse, P. E., Long, J. C. & Sokal, R. R. (1986) Syst. Zool. 35 , 627-632. LaunchUrlCrossRef ↵ Berry, K. J., Kvamme, K. L. & Mielke, P. W. (1983) Am. Antiquity 48 , 547-553. LaunchUrl ↵ Zimmerman, G. M., Goetz, H. & Mielke, P. W. (1985) Ecology 66 , 606-611. LaunchUrlCrossRef ↵ Huenneke, L. F. & Vitousek, P. M. (1990) Biol. Conserv. 53 , 199-211. LaunchUrlCrossRef ↵ Thompson, J., Brokaw, N., Zimmerman, J. K., Waide, R. B., Everham, E. M., Lodge, D. J., Taylor, C. M., Garcia-Montiel, D. & Fluet, M. (2002) Ecol. Appl. 12 , 1344-1363. LaunchUrl
Like (0) or Share (0)