Abstract: Modelling the local density of tropical saplings can provide insights into the ecological processes that drive species regeneration and thereby help predict population recovery after disturbance. Yet, few studies have addressed the challenging issues in autocorrelation and zero-inflation of local density. This paper presents Hierarchical Bayesian Modelling (HBM) of sapling density that includes these two features. Special attention is devoted to variable selection, model estimation and comparison. We developed a Zero-Inflated Poisson (ZIP) model with a latent correlated spatial structure and compared it with non-spatial ZIP and Poisson models that were either autocorrelated (Spatial Generalized Linear Mixed, SGLM) or not (generalized linear models, GLM). In our spatial models, local density autocorrelation was modeled by a Conditional Auto-Regressive (CAR) process. 13 explicative variables described ecological conditions with respect to topography, disturbance, stand structure and intraspecific processes. Models were applied to six tropical tree species with differing biological attributes: Oxandra asbeckii, Eperua falcata, Eperua grandiflora, Dicorynia guianensis, Qualea rosea, and Tachigali melinonii. We built species-specific models using a simple method of variable selection based on a latent binary indicator. Our spatial models showed a close correlation between observed and estimated densities with site spatial structure being correctly reproduced. By contrast, the non-spatial models showed poor fits. Variable selection highlighted species-specific requirements and susceptibility to local conditions. Model comparison overall showed that the SGLM was the most accurate explanatory and predictive model. Surprisingly, zero-inflated models performed less well. Although the SZIP model was relevant with respect to data distribution, and more flexible with respect to response curves, its model complexity caused marked variability in parameter estimates. In the SUM, the spatial process alone accounted for zero-inflation in the data. A refinement of the hypotheses employed at the process level could compensate for distribution flaws at the data level. This study emphasized the importance of the HBM framework in improving the modelling of density-environment relationships. (C) 2008 Elsevier B.V. All rights reserved.

Abstract: Tropical forests are predicted to harbor most of the insect diversity on earth, but few studies have been conducted to characterize insect communities in tropical forests. One major limitation is the lack of consensus on methods for insect collection. Deciding which insect trap to use is an important consideration for ecologists and entomologists, yet to date few study has presented a quantitative comparison of the results generated by standardized methods in tropical insect communities. Here, we investigate the relative performance of two flight interception traps, the windowpane trap, and the more widely used malaise trap, across a broad gradient of lowland forest types in French Guiana. The windowpane trap consistently collected significantly more Coleoptera and Blattaria than the malaise trap, which proved most effective for Diptera, Hymenoptera, and Hemiptera. Orthoptera and Lepidoptera were not well represented using either trap, suggesting the need for additional methods such as bait traps and light traps. Our results of contrasting trap performance among insect orders underscore the need for complementary trapping strategies using multiple methods for community surveys in tropical forests.

Abstract: A major challenge for forest managers is to define the optimal cutting cycle to ensure that the resource is sustained in the long term. Matrix models of forest dynamics allow time-projection of diameter-class distributions and thus assessment of the time needed, after logging, to recover a given part of the exploitable stock. They are easy to build and they only require, as input variables, the diameter structure of the population(s) under scope. However, such models are based on a coarse description of tree population dynamics and must be used with caution. In particular, as trees are only described from a diameter threshold (usually 10 cm dbh), recruitment of a new tree cannot be linked with the preceding generation since too much time elapsed between seed dispersal and the installation of a 10-cm recruit. This causes predictions of matrix models to be highly questionable in the long term when ingrowth to larger dbh classes greatly depends on the way recruitment has been modelled. We used a case study from French Guiana to test whether or not a simple matrix model is reliable enough to help forest managers choose between management alternatives. We focused on the major timber species Dicorynia guianensis Amshoff (Caesalpiniaceae) harvested under a selective cutting regime. We compared predictions of D. guianensis stock recovery in the short and long term provided by two models: StoMat, a non-regulated matrix model, and SELVA, a single-tree distance dependent model explicitly simulating the entire species life cycle. Both models were independently calibrated on data from Paracou permanent sample plots. We showed that: (i) the short-term recovery of the exploitable stock predicted by StoMat is reliable for a large range of disturbance conditions; (ii) recruitment implementation in StoMat does not influence projections until the third felling cycle; (iii) for shared initial stand conditions SELVA and StoMat give consistent mid- and long-term predictions: the simple recruitment model used into StoMat could efficiently summarise the regeneration processes of the species under low felling intensity. Our results indicate that the current felling regime used in French Guiana may not be sustainable on a long-term basis. In any case, no more than 60% of the initial stock would be recovered after logging. We conclude that simple models can provide as reliable predictions as more complicated ones. They may be sufficient to assess the recovery of a species' exploitable stock even in the long term, or at least assess the (un)sustainability of particular harvesting regimes. (c) 2005 Elsevier B.V. All rights reserved.

Abstract: In order to progress in the understanding of mechanical stress generation, the mesoporosity of the cell wall and its changes during maturation of poplar (Populus deltoides × P. nigra) tension wood (TW) and opposite wood (OW) were measured by nitrogen adsorption-desorption. Variations in the thickness of the gelatinous layer (G-layer) were also measured to clarify whether the mesoporosity change simultaneously with the deposition of the G-layer in TW. Results show that mesoporous structures of TW and OW were very similar in early development stages before the deposition of G-layers. With the formation of the S2 layer in OW and the G-layer in TW, the mesopore volume decreased steeply before lignification. However, in TW only, the decrease in mesopore volume occurred together with the pore shape change and a progressive increase in pore size. The different patterns observed in TW revealed that pores from G-layers appear with a different shape compared to those of the compound middle lamella, and their size increases during the maturation process until stabilising in mature wood. This observation strongly supports the hypothesis of the swelling of the G-layer matrix during maturation as the origin of maturation stress in poplar tension wood.