Trees play a crucial role in the water, carbon and nitrogen cycle on local, regional and global scales. Understanding the exchange of momentum, heat, water, and CO 2 between trees and the atmosphere is important to assess the impact of drought, deforestation and climate change. Unfortunately, ground measurements of tree properties such as mass and canopy interception of precipitation are often expensive or difficult due to challenging environments. This paper aims to demonstrate the concept of using robust and affordable accelerometers to measure tree properties and responses. Tree sway is dependent on mass, canopy structure, drag coefficient, and wind forcing. By measuring tree acceleration, we can relate the tree motion to external forcing (e.g., wind, precipitation and related canopy interception) and tree physical properties (e.g., mass, elasticity). Using five months of acceleration data of 19 trees in the Brazilian Amazon, we show that the frequency spectrum of tree sway is related to mass, canopy interception of precipitation, and canopy–atmosphere turbulent exchange.

The mass of trees is influenced by physiological processes within the tree (e.g. transpiration and root water uptake), as well as external loads (e.g. intercepted precipitation). Recent studies have found diurnal variations in radar backscatter over vegetated areas, which might be attributed to mass changes of the vegetation layer. Field measurements are required to study the driving processes. This study aims to use measured three-dimensional displacement and acceleration of trees, to detect and quantify their diurnal (bio)mass variations. Accelerometers and dendrometers were installed on seven different tree species in the Amazon rainforest. Trees were selected to cover a broad range of wood density. Using spectral analysis, the governing frequencies in the acceleration time series were found. The governing frequencies showed a diurnal pattern, as well as a change during precipitation events. Our results suggest that we can separate and potentially quantify tree mass changes due to (1) internal water redistribution and (2) intercepted precipitation. This will allow further investigation of the effect of precipitation and water stress on tree dynamics in forest canopies.