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.

Significance: We develop a biogeographic approach to analyzing the presence of alternative stable states in tropical biomes. Whilst forest–savanna bistability has been widely hypothesized and modeled, empirical evidence has remained scarce and controversial, and here, applying our method to Africa, we provide large-scale evidence that there are alternative states in tree species composition of tropical vegetation. Furthermore, our results have produced more accurate maps of the forest and savanna distributions in Africa, which take into account differences in tree species composition, and a complex suite of determinants. This result is not only important for understanding the biogeography of the continent but also, to guide large-scaled tree planting and restoration efforts planned for the region.

A better understanding of how climate affects growth in tree species is essential for improved predictions of forest dynamics under climate change. Long‐term climate averages (mean climate) drive spatial variations in species’ baseline growth rates, whereas deviations from these averages over time (anomalies) can create growth variation around the local baseline. However, the rarity of long‐term tree census data spanning climatic gradients has so far limited our understanding of their respective role, especially in tropical systems. Furthermore, tree growth sensitivity to climate is likely to vary widely among species, and the ecological strategies underlying these differences remain poorly understood. Here, we utilize an exceptional dataset of 49 years of growth data for 509 tree species across 23 tropical rainforest plots along a climatic gradient to examine how multiannual tree growth responds to both climate means and anomalies, and how species’ functional traits mediate these growth responses to climate. We show that anomalous increases in atmospheric evaporative demand and solar radiation consistently reduced tree growth. Drier forests and fast‐growing species were more sensitive to water stress anomalies. In addition, species traits related to water use and photosynthesis partly explained differences in growth sensitivity to both climate means and anomalies. Our study demonstrates that both climate means and anomalies shape tree growth in tropical forests and that species traits can provide insights into understanding these demographic responses to climate change, offering a promising way forward to forecast tropical forest dynamics under different climate trajectories.