Airborne Tree Crown Detection for Predicting Spatial Heterogeneity of Canopy Transpiration in a Tropical Rainforest

Abstract

Tropical rainforests comprise complex 3D structures and encompass heterogeneous site conditions; their transpiration contributes to climate regulation. The objectives of our study were to test the relationship between tree water use and crown metrics and to predict spatial variability of canopy transpiration across sites. In a lowland rainforest of Sumatra, we measured tree water use with sap flux techniques and simultaneously assessed crown metrics with drone-based photogrammetry. We observed a close linear relationship between individual tree water use and crown surface area (R2 = 0.76, n = 42 trees). Uncertainties in predicting stand-level canopy transpiration were much lower using tree crown metrics than the more conventionally used stem diameter. 3D canopy segmentation analyses in combination with the tree crown–water use relationship predict substantial spatial heterogeneity in canopy transpiration. Among our eight study plots, there was a more than two-fold difference, with lower transpiration at riparian than at upland sites. In conclusion, we regard drone-based canopy segmentation and crown metrics to be very useful tools for the scaling of transpiration from tree- to stand-level. Our results indicate substantial spatial variation in crown packing and thus canopy transpiration of tropical rainforests.

Predicting Tree Sap Flux and Stomatal Conductance from Drone-Recorded Surface Temperatures in a Mixed Agroforestry System—A Machine Learning Approach

“…We assume a non-linear relationship between canopy size and sap flux, which requires auxiliary variables to be predicted. Remotely sensed canopy size is an important factor for plant transpiration and was already found to be a suitable predictor in previous studies [6,80]. A general advantage of RF algorithms is their ability to produce accurate predictions even from small samples and with a large set of independent input variables [31]; in our study, they performed better than all other applied algorithms.…”

“…As such, previous ecohydrological assessments in the study region pointed to vast differences in water use between oil palms and trees including rubber trees of similar age, with substantially higher per-tree and stand transpiration rates of oil palms [78,79]. Drone-derived crown metrics of oil palms and adjacent trees further suggested that oil palms transpire two-times more water per unit of crown volume as agroforest trees [6] and about five-times more per unit of crown surface area than rainforest trees [80], which may well be the reason why the joined analysis of trees and palms was unsuccessful.…”

Complex canopy structures control tree transpiration: A study based on 3D modelling in a tropical rainforest

“…There is no pronounced dry season, with all months typically receiving more than 100 mm of precipitation (Drescher et al, 2016). The terrain is undulating, dividing the forest into upland and riparian sites (Ahongshangbam et al, 2020). The forest is rich in tree species (Rembold et al, 2017) but has a history of logging (Harrison & Swinfield, 2015).…”

“…As a baseline variable we used the crown volume (m 3 ) as derived from a convex hull algorithm (in the following referred to as 'crown volume'), which was reported to be a good predictor of per-tree transpiration in previous studies in the same region and partly on the same sample trees (Ahongshangbam et al, 2019(Ahongshangbam et al, , 2020; note: crown volume of one tree was corrected from the dataset of Ahongshangbam et al, 2020). As additional crown dimension variables, we calculated crown width (m, average of a North-South and an East-West measurement), the ground projection area (m 2 ) of the convex hull crown body into the 2D plane (in the following referred to as 'ground projection area', in other studies also called 'crown projection area'; note: ground projection area of several trees were corrected from the dataset of Ahongshangbam et al, 2020), the crown surface area (m 2 ) as derived from an alpha25-algorithm (in the following referred to as 'crown surface area'; note: these values differ from the convex hull surface areas as presented in Ahongshangbam et al, 2020) and the crown volume of the upper crown half (m 3 , convex hull) (Table S1). The total leaf area per tree, which was calculated by multiplying the ground projection area by the leafrelated variable leaf area index (see description in Section 2.3.4), is listed as a variable of crown dimension in our study due to its close correlation with other dimension-related metrics (R ≥ 0.77, Figures S3 and S4).…”

Evaluation of Crop Water Status and Vegetation Dynamics For Alternate Partial Root-Zone Drip Irrigation of Alfalfa: Observation With an UAV Thermal Infrared Imagery

“…Compared with satellite, UAV-based remote sensing is more suitable for field scale monitoring. Using UAV equipped with high-resolution TIR cameras, Ahongshangbam et al (2020) detected spatial variability of tropical rainforest transpiration, Hou et al (2021) estimated the soybean transpiration under different water conditions and verified with hydrogen-oxygen stable isotopes measurements, and Ortega-Farias et al ( 2021) evaluated the water flux of vineyard and verified with an eddy covariance system. These studies have confirmed the accuracy and applicability of UAV-based TIRs in transpiration estimation.…”