Variation in wood density across South American tropical forests

Abstract

Wood density is a critical control on tree biomass, so poor understanding of its spatial variation can lead to large and systematic errors in forest biomass estimates and carbon maps. The need to understand how and why wood density varies is especially critical in tropical America where forests have exceptional species diversity and spatial turnover in composition. As tree identity and forest composition are challenging to estimate remotely, ground surveys are essential to know the wood density of trees, whether measured directly or inferred from their identity. Here, we assemble an extensive dataset of variation in wood density across the most forested and tree-diverse continent, examine how it relates to spatial and environmental variables, and use these relationships to predict spatial variation in wood density over tropical and sub-tropical South America. Our analysis refines previously identified east-west Amazon gradients in wood density, improves them by revealing fine-scale variation, and extends predictions into Andean, dry, and Atlantic forests. The results halve biomass prediction errors compared to a naïve scenario with no knowledge of spatial variation in wood density. Our findings will help improve remote sensing-based estimates of aboveground biomass carbon stocks across tropical South America.

Fig. 1. Variation in wood density across South American tropical forests.

a Location of forest inventory plots where wood density was quantified. b Variation in basal-area weighted wood density between regions. Colours in a relate to regions in (b). N = 981 plots (Lowland-NW = 182 plots, Lowland-SW = 168, East-central Amazon = 205, Guiana Shield = 123, Brazilian Shield = 119, Atlantic forest = 69, Andes = 71, Dry forests = 44). Black points show mean values in each region estimated by a linear model with wood density as the response variable and region as the explanatory variable; lines show 95% confidence intervals from that model.

Fig. 2. Contribution of explanatory variables to models of spatial variation in wood density.

Global Shapley additive explanations (SHAP values) have been calculated for random forest (RF, blue) and generalised additive models (GAM, green) fitted with either just environmental or spatial variables (env/ spatial, open circles) or to both environmental and spatial variables (both, filled circles). Higher SHAP values indicate a greater contribution of a feature to model predictions. Variables are ordered on the x-axis based on their mean SHAP value across models.

Fig. 3. Predicted spatial variation in basal-area weighted wood density across tropical and sub-tropical South America at 1km resolution.

Ensemble predictions averaging the six input models are shown. Predictions for individual models are presented in Fig. S2, with inter-model uncertainty in Fig. S3. Forested areas outside the area of applicability of model predictions are shown in grey – see Fig. S2 for model predictions without this exclusion, and Fig. S7 for alternative definitions of areas of model applicability.