Introduction

Eddy covariance towers provide continuous measurements of ecosystem exchanges of carbon, water, and energy between the ecosystem and the atmosphere (Figure 1). These fluxes can be used as a representation of the ecosystem, for example the Gross Primary Productivity (GPP), Net Ecosystem Exchange (NEE) and the Evapotranspiration (ET), being the amount of carbon captured by photosynthesis, the net carbon accumulation or loss, and the sum of water evaporation from soil and transpiration from plants. However, the interpretation of these measurements as a representation of the ecosystem requires assumptions such as that the tower samples come from relatively consistent and homogeneous source over time.

In the field, eddy covariance measurements come from a changing footprint that varies with wind speed, wind direction, and atmospheric stability. As a result, the area of contribution can shift substantially over time. If the area surrounding the tower is heterogeneous, the variation in source area may influence the observed fluxes independently of ecosystem processes. Knowing if the flux measurements remain consistent across the area, considering wind speed and direction, can help decide if the data are representative of the ecosystem dynamics.

Another challenge is to understand how environmental conditions are associated with variability in ecosystem fluxes. Tropical dry forests are characterized by a marked seasonal variation in water availability, leading to pronounced changes in atmospheric demand, temperature, and vegetation (Abdaki et al., 2024). Variables such as Vapor Pressure Deficit, Air temperature, radiation, and vegetation indices (Normalized Difference Vegetation Index and Enhanced Vegetation Index) may influence ecosystem fluxes (Medlyn et al., 2011). However, these variables often covary, making it difficult to know which factors are most associated with changes in carbon uptake and water exchange rates.

This study first evaluates whether eddy covariance measurements from a dry tropical forest in Costa Rica (Kakubari tower) are representative of the area, by analyzing the fluxes when varying the distance of measurement, wind speed, and wind direction. Then, examine how environmental variables are associated with variability in ecosystem fluxes, applying multivariate statistical techniques to account for the covariance between predictor and response variables.

This study was guided by two main objectives:

1. Evaluate whether eddy covariance measurements at the Kakubari tower reliably represent the ecosystem's carbon and water exchanges, and whether variations in wind conditions influences the observed fluxes.

2. Identify which environmental variables are associated with the variability in ecosystem carbon and water fluxes at the Kakubari tropical dry forest site.