Eigenvector-mapping methods such as Moran's eigenvector maps (MEM) are derived from a spatial weighting matrix (SWM) that describes the relations among a set of sampled sites. The specification of the SWM is a crucial step, but the SWM is generally chosen arbitrarily, regardless of the sampling design characteristics. Here, we compare the statistical performances of different types of SWMs (distance-based or graph-based) in contrasted realistic simulation scenarios. Then, we present an optimization method and evaluate its performances compared to the arbitrary choice of the most-widely used distance-based SWM. Results showed that the distance-based SWMs generally had lower power and accuracy than other specifications, and strongly underestimated spatial signals. The optimization method, using a correction procedure for multiple tests, had a correct type I error rate, and had higher power and accuracy than an arbitrary choice of the SWM. Nevertheless, the power decreased when too many SWMs were compared, resulting in a trade-off between the gain of accuracy and the loss of power. We advocate that future studies should optimize the choice of the SWM using a small set of appropriate candidates. R functions to implement the optimization are available in the adespatial package and are detailed in a tutorial.

Eigenvector mapping techniques are widely used by ecologists and evolutionary biologists to describe and control for spatial and/or phylogenetic patterns in their data. The selection of an appropriate subset of eigenvectors is a critical step (misspecification can lead to highly biased results and interpretations), and there is no consensus yet on how to proceed. We conducted a ten‐year review of the practices of eigenvector selection and highlighted three main procedures: selecting the subset of descriptors minimising the Akaike information criterion (AIC), using a forward selection with double stopping criterion after testing the global model significance (FWD), and selecting the subset minimising the autocorrelation in the model residuals (MIR). We compared the type I error rates, statistical power, and R² estimation accuracy of these methods using simulated data. Finally, a real dataset was analysed using variation partitioning analysis to illustrate to what extent the different selection approaches affected the ecological interpretation of the results. We show that, while the FWD and MIR approaches presented a correct type I error rate and were accurate, the AIC approach displayed extreme type I error rates (100%), and strongly overestimated the R². Moreover, the AIC approach resulted in wrong ecological interpretations, as it overestimated the pure spatial fraction (and the joint spatial‐environmental fraction to a lesser extent) of the variation partitioning. Both the FWD and MIR methods performed well at broad and medium scales but had a very low power to detect fine‐scale patterns. The FWD approach selected more eigenvectors than the MIR approach but also returned more accurate R² estimates. Hence, we discourage any future use of the AIC approach, and advocate choosing between the MIR and FWD approaches depending on the objective of the study: controlling for spatial or phylogenetic autocorrelation (MIR) or describing the patterns as accurately as possible (FWD).

No abstract

The canopy of many central African forests is dominated by light-demanding tree species that do not regenerate well under themselves. The prevalence of these species might result from ancient slash-and-burn agricultural activities that created large openings, while a decline of these activities since the colonial period could explain their deficit of regeneration. To verify this hypothesis, we compared soil charcoal abundance, used as a proxy for past slash-and-burn agriculture, and tree species composition assessed on 208 rainforest 0.2 ha plots located in three areas from Southern Cameroon. Species were classified in regeneration guilds (pioneer, non-pioneer light-demanding, shade-bearer) and characterized by their wood-specific gravity, assumed to reflect light requirement. We tested the correlation between soil charcoal abundance and: (i) the relative abundance of each guild, (ii) each species and family abundance and (iii) mean wood-specific gravity. Charcoal was found in 83% of the plots, indicating frequent past forest fires. Radiocarbon dating revealed two periods of fires: “recent” charcoal were on average 300 years old (up to 860 BP, n = 16) and occurred in the uppermost 20 cm soil layer, while “ancient” charcoal were on average 1900 years old (range: 1500 to 2800 BP, n = 43, excluding one sample dated 9400 BP), and found in all soil layers. While we expected a positive correlation between the relative abundance of light-demanding species and charcoal abundance in the upper soil layer, overall there was no evidence that the current heterogeneity in tree species composition can be explained by charcoal abundance in any soil layer. The absence of signal supporting our hypothesis might result from (i) a relatively uniform impact of past slash-and-burn activities, (ii) pedoturbation processes bringing ancient charcoal to the upper soil layer, blurring the signal of centuries-old Human disturbances, or (iii) the prevalence of other environmental factors on species composition.

No abstract

A comprehensive understanding of Ca cycling in an ecosystem is desirable because of the role of this element in tree mineral nutrition and its status as a major base cation on the soil exchange complex. The determination of the origin of Ca in forests is particularly indicated in regard of important changes linked to acid inputs and intensive logging. Natural strontium isotopes are increasingly used as tracers of Ca in forest ecosystems for qualitative and quantitative assessments. Nevertheless this method is limited to relatively simple systems with two sources of nutrients. Some recent studies coupled Sr/Ca or Sr/Ba ratios to Sr isotopic measurements in order to solve more complex systems. Such method has however associated with it some uncertainties: this approach assumed that Ca, Sr and Ba behave similarly throughout the ecosystem and does not take into account the Ca biopurification processes occurring in some tree's organs which can alter element ratio. The present work focuses on two deciduous species covering large areas in Europe: European beech (Fagus sylvatica L.) and pedunculate oak (Quercus robur L.). In order to test the similarity of behaviour between Ca, Sr and Ba, their concentrations were measured extensively in the major compartments of two forest ecosystems. In parallel, the discrimination process inside tree organs was studied in 23 stands for beech and 10 stands for oak. We found that Sr and Ca behave similarly in all soil and tree compartments. By contrast, Ba and Ca appear to have contrasting behaviours, especially in streams, soil solution and soil exchange complex (no correlations between element concentrations). Sr/Ba and Ba/Ca ratios must therefore be used with care as tracer of Ca. The Ca biopurification is absent in roots and slight in bole wood but is large in bark, twigs and leaves. The discrimination factors (DF) between wood and leaves are characteristic of the two species studied and do not change significantly as a function of the soil Ca status (acidic or calcareous soils). Therefore, strontium-calcium DF can be used as a correction factor of the Sr/Ca ratio of leaves when this ratio is used in connection with Sr isotopic ratios. This correction allows to solve systems of tree nutrition with more than two sources of Ca.

Questions Soil properties have been shown to partially explain tree species distribution in tropical forests. Locally, species turnover across space can result not only from edaphic heterogeneities but also from limited seed dispersal. To characterize the contribution of each process, contact areas between contrasted soil types offer ideal settings. In the present study, we aimed to test species and species assemblage responses to a sharp edaphic discontinuity in a tropical forest tree community. Location Yoko forest reserve (6975 ha), Democratic Republic of the Congo. Methods We set up four 500–600‐m long parallel transects crossing two contrasted edaphic habitats, one lying on clayey soil and the other on sandy soil. The canopy and subcanopy trees were identified and geo‐referenced along the transects over a width of 50 m and 5 m, respectively, and soil samples were collected every 50 m to characterize each habitat. Results Correspondence analyses indicated a clear differentiation of tree communities between sandy and clayey soils. Using a torus‐translation method combined with Chi‐squared non‐parametric tests, we observed that ca. 40% and 18% of the species represented by at least 12 individuals displayed significant density differences according to habitat in the canopy and subcanopy, respectively, although very few species displayed significant differences in their relative abundance. Nevertheless, whole community tests of differentiation (in species relative abundances) between soil types were significant in both strata, even after removing individual species or families displaying a significant habitat preference. Conclusion While only a minority of species displayed a clear habitat preference, we still observed a community‐wide impact of the edaphic discontinuity on species assemblages at a local scale. Our results provide further evidence for the major contribution of environmental heterogeneity in maintaining biodiversity in tropical forests.

A B S T R A C TStable zinc (Zn) isotope fractionation between soil and plant has been used to suggest the mechanisms affecting Zn uptake under toxic conditions. Here, changes in Zn isotope composition in soil, soil solution, root and shoot were studied for ryegrass (Lolium multiflorum L.) and rape (Brassica napus L.) grown on three distinct metal-contaminated soils collected near Zn smelters (total Zn 0.7-7.5%, pH 4.8-7.3). The Zn concentrations in plants reflected a toxic Zn supply. The Zn isotopic fingerprint of total soil Zn varied from À0.05% to +0.26 AE 0.02% (d 66 Zn values relative to the JMC 3-0749L standard) among soils, but the soil solution Zn was depleted in 66 Zn, with a constant Zn isotope fractionation of about À0.1% d 66 Zn unit compared to the bulk soil. Roots were enriched with 66 Zn relative to soil solution (d 66 Zn root À d 66 Zn soil solution = D 66 Zn root-soil solution = +0.05 to +0.2 %) and shoots were strongly depleted in 66 Zn relative to roots (D 66 Zn shoot-root = À0.40 to À0.04 %). The overall d 66 Zn values in shoots reflected that of the bulk soil, but were lowered by 0.1-0.3 % units as compared to the latter. The isotope fractionation between root and shoot exhibited a markedly strong negative correlation (R 2 = 0.83) with transpiration per unit of plant weight. Thus, the enrichment with light Zn isotopes in shoot progressed with increasing water flux per unit plant biomass dry weight, showing a passive mode of Zn transport by transpiration. Besides, the light isotope enrichment in shoots compared to roots was larger for rape than for rye grass, which may be related to the higher Zn retention in rape roots. This in turn may be related to the higher cation exchange capacity of rape roots. Our finding can be of use to trace the biogeochemical cycles of Zn and evidence the tolerance strategies developed by plants in Znexcess conditions.