Canopy wetness in the Eastern Amazon

Abstract

Canopy wetness is a common condition that influences photosynthesis, the leaching or uptake of solutes, the water status and energy balance of canopies, and the interpretation of eddy covariance and remote sensing data. While often treated as a binary variable, 'wet' or 'dry', forest canopies are often partially wet, requiring the use of a continuous description of wetness. Minor precipitation events such as dew, that wet a fraction of the canopy, have been found to contribute to dry season foliar water uptake in the Eastern Amazon, and are fundamentally important to the canopy energy balance. However, few studies have reported the spatial and temporal distribution of canopy wetness, or the relative contribution of dew to leaf wetness, for forest ecosystems.In this study, we use two canopy profiles of leaf wetness sensors, coupled with meteorological data, to address fundamental questions about spatial and temporal variation of leaf wetness in an Eastern Amazonian rainforest. We also investigate how well meteorological tower data can predict canopy wetness using two models, one empirical and one that is physically-based.The results show that the canopy is 100% dry only for 34% of the time, otherwise being between 5% and 100% wet. Dew accounts for 20% or 43% of total annual leaf wetness, and 36% or 50% of canopy wetness in dry season, excluding or including dew events that co-occur with rain, respectively. Wetness duration was higher at the top than bottom of the canopy, mainly because of rain events, whilst dew formation was strongly dependent on the local canopy structure and varied horizontally through the canopy. The best empirical model accounted for 55% of the variance in canopy wetness, while the physical model accounted for 48% of the variance. We discuss future modelling improvements of the physical model to increase its predictive capacity.

Leaf surface water, not plant water stress, drives diurnal variation in tropical forest canopy water content

“…Our simulated LW s at predawn (06:00 h) in top canopy leaves showed a consistent range at a similarly wet forest site and predicted that top canopy leaves are frequently wet at predawn (Fig. S2), which is consistent with a recent report at Caxiuanã using leaf wetness sensors (Binks et al, 2021). Altogether, these consistencies suggest the model predictions on LW s dynamics are realistic.…”

“…At the ecosystem scale, the contribution of LW s to VOD signals remains largely unknown despite LW s being an important component of the moisture budget, particularly in rainforest ecosystems where significant diurnal and seasonal variation in CWC occurs because of frequent rainfall interception and dew formation (Junqueira Junior et al, 2019;Binks et al, 2021) and where measurements of LW s beyond qualitative leaf wetness data (Binks et al 2019) do not exist. Therefore, ignoring the contribution of LW s to VOD can lead to overestimation of changes in leaf internal water (LW i ), which potentially biases the interpretation of VOD data as a measure of vegetation water stress.…”

Measuring the vertical profile of leaf wetness in a forest canopy

“…Leaf wetness sensors enable us to quantify the duration and spatial distribution of leaf wetness [1] , but manually analysing leaf wetness data is impractical and relies on subjective judgements that cans lead to inconsistency. Here we present the methods for: (i) setting up leaf wetness sensors (Phytos 31, Meter, Pullman, USA) in a forest canopy profile; and (ii) a method and R script for objectively distinguishing between rain and dew events (see co-publication [2] , Agr.For.Met.). The Supplementary Information includes an R script for detecting rain and dew events in leaf wetness sensor data (SI 1), and the program used for the data logger and multiplexer (SI 2); and the leaf wetness data on which this study was based (Caxiuana National park, Brazil, an Eastern Amazonia field site), has been uploaded to Mendeley Data ( http://dx.doi.org/10.17632/sbrbbn7skn.1 ).…”

“…We wanted to capture the transition of wetness conditions through the densest part of the upper canopy, and therefore we positioned sensors more closely together at the top of the rig than bottom. Sensor pairs (one sensor per rig) were located at heights of 4, 14, 19, 24, 28, 30, 32, 34, and 36 m from the ground, where the height of the top of the canopy was around 30 m with emergent trees attaining heights of 55 m (see [2] for canopy profile). Sensors were cleaned approximately every three months.…”

“…While the magnitude of the output from the leaf wetness sensors (millivolts) does not represent the level of saturation, the rate at which the signal increases can be used to distinguish between rain- or dew-related wetness ( Fig. 4 ., co-publication [2] ). Furthermore, over time the baseline value of a sensor (i.e., the sensor value when dry) can drift due to a build-up of detritus on the sensor, or possibly even a change of electrical resistance at the cable connections.…”