Abstract: In West Africa, more than 80% of the original forest cover has disappeared due to the exponential growth of human populations in a recurrent search for new agricultural land. Once the fertility of the land is exhausted, these areas are abandoned and left to be reforested through natural succession. Despite the widespread presence of secondary forests of various ages in West African landscapes, little is known about the trajectories of recovery and the environmental factors that influence recovery rates. We set up 96 0.2 ha forest plots, along a chronosequence of 1 to 40 years and including 7 controls, on which all trees larger than 2.5 cm in diameter at breast height were inventoried. We modelled the recovery trajectories of four complementary dimensions of biodiversity (richness, diversity, composition, indicators of old-growth forest) in a Bayesian framework. Our results show that the four dimensions of biodiversity recover at different rates, with composition recovering much faster than floristic diversity. Among the local, landscape, and historical factors studied, the number of remnants and proximity to old-growth forests have a positive impact on recovery rates, with, under good environmental conditions, the composition, richness, and diversity being almost completely recovered in less than 25 years. Our results demonstrate the very high resilience of the composition of the semi-deciduous forests of West Africa, but also suggest that the management of these post-forest areas must be differentiated according to the landscape context and the presence of isolated trees, which are the last vestiges of the former forest. In unfavourable conditions, natural dynamics should be assisted by agroforestry practices and local tree planting to allow for a rapid restoration of forest goods and services to local populations.

Abstract: Automatic tree crown segmentation from remote sensing data is especially challenging in dense, diverse, and multilayered tropical forest canopies, and tracking mortality by this approach is even more difficult. Here, we examine the potential for combining airborne laser scanning (ALS) with multispectral and hyperspectral data to improve the accuracy of tree crown segmentation at a study site in French Guiana. We combined an ALS point cloud clustering method with a spectral deep learning model to achieve 83% accuracy at recognizing manually segmented reference crowns (with congruence >0.5). This method outperformed a two-step process that involved clustering the ALS point cloud and then using the logistic regression of hyperspectral distances to correct oversegmentation. We used this approach to map tree mortality from repeat surveys and show that the number of crowns identified in the first that intersected with height loss clusters was a good estimator of the number of dead trees in these areas. Our results demonstrate that multisensor data fusion improves the automatic segmentation of individual tree crowns and presents a promising avenue to study forest demography with repeated remote sensing acquisitions.

Abstract: Biological nitrogen fixation (BNF) is a fundamental part of nitrogen cycling in tropical forests, yet little is known about the contribution made by free-living nitrogen fixers inhabiting the often-extensive forest canopy. We used the acetylene reduction assay, calibrated with 15N2, to measure free-living BNF on forest canopy leaves, vascular epiphytes, bryophytes and canopy soil, as well as on the forest floor in leaf litter and soil. We used a combination of calculated and published component densities to upscale free-living BNF rates to the forest level. We found that bryophytes and leaves situated in the canopy in particular displayed high mass-based rates of free-living BNF. Additionally, we calculated that nearly 2 kg of nitrogen enters the forest ecosystem through free-living BNF every year, 40% of which was fixed by the various canopy components. Our results reveal that in the studied tropical lowland forest a large part of the nitrogen input through free-living BNF stems from the canopy, but also that the total nitrogen inputs by free-living BNF are lower than previously thought and comparable to the inputs of reactive nitrogen by atmospheric deposition.

Abstract: Tree stems and soils can act as sources and sinks for the greenhouse gases (GHG) carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). Since both uptake and emission capacities can be large, especially in tropical rainforests, accurate assessments of the magnitudes and temporal variations of stem and soil GHG fluxes are required. We designed a new flexible stem chamber system for continuously measuring GHG fluxes in a French Guianese rainforest. Here, we describe this new system, which is connected to an automated soil GHG flux system, and discuss measurement uncertainty and potential error sources. In line with findings for soil GHG flux estimates, we demonstrated that lengthening the stem chamber closure time was required for accurate estimates of tree stem CH4 and N2O flux but not tree stem CO2 flux. The instrumented stem was a net source of CO2 and CH4 and a weak sink of N2O. Our experimental setup operated successfully in situ and provided continuous tree and soil GHG measurements at a high temporal resolution over an 11-month period. This automated system is a major step forward in the measurement of GHG fluxes in stems and the atmosphere concurrently with soil GHG fluxes in tropical forest ecosystems.

Abstract: Drought-induced xylem embolism is considered to be one of the main factors driving mortality in woody plants worldwide. Although several structure–functional mechanisms have been tested to understand the anatomical determinants of embolism resistance, there is a need to study this topic by integrating anatomical data for many species. We combined optical, laser, and transmission electron microscopy to investigate vessel diameter, vessel grouping, and pit membrane ultrastructure for 26 tropical rainforest tree species across three major clades (magnoliids, rosiids, and asteriids). We then related these anatomical observations to previously published data on drought-induced embolism resistance, with phylogenetic analyses. Vessel diameter, vessel grouping, and pit membrane ultrastructure were all predictive of xylem embolism resistance, but with weak predictive power. While pit membrane thickness was a predictive trait when vestured pits were taken into account, the pit membrane diameter-to-thickness ratio suggests a strong importance of the deflection resistance of the pit membrane. However, phylogenetic analyses weakly support adaptive coevolution. Our results emphasize the functional significance of pit membranes for air-seeding in tropical rainforest trees, highlighting also the need to study their mechanical properties due to the link between embolism resistance and pit membrane diameter-to-thickness ratio. Finding support for adaptive coevolution also remains challenging. © 2020 The Authors New Phytologist © 2020 New Phytologist Foundation