Autocorrelation as a critical factor of growth depensation of tropical trees in the Chocó biogeographic region

• Published in: Ecological modelling 2025-01, Vol.500, p.110949, Article 110949
• Main Authors: Martínez, Camilo E., Orrego, Sergio A., Giraldo, Jorge A., del Valle, Jorge I., Hernández-Barajas, Freddy, David, Diego A.
• Format: Article
• Language: English
• Subjects: autocorrelation; forest management; geographical distribution; Growth autocorrelation; Growth depensation; Growth modeling; growth rings; Mixed-effects models; population ecology; species; Tree rings; trees
• ISSN: 0304-3800
• DOI: 10.1016/j.ecolmodel.2024.110949

•Growth autocorrelation is a major contributor to intraspecific variability in diameter growth trajectories of tree populations.•We used a modeling approach that includes von Bertalanffy-type equations, random effects, and ARIMA specifications to account for autocorrelation.•Tree ring data facilitated the study of growth autocorrelation in tropical tree populations.•For the first time, we presented biologically meaningful ontogenic traits and silvicultural metrics relevant to the management of species from the ever-wet tropical forests. Growth depensation, the variation of size with age exhibited by populations, is attributed to biological, ecological, and environmental factors, as well as autocorrelation. Several studies have focused on the study of diameter growth of tropical trees to obtain information on ontogenic traits and silvicultural metrics of interest for ecology and forest management. However, few studies have rigorously and adequately considered autocorrelation as a primary factor contributing to growth depensation. The aim of this study was to investigate diameter growth in tree species from the Chocó biogeographic region. We used tree-ring data corresponding to 38 trees and 5 species. Our modeling approach included von Bertalanffy type equations to estimate diameter growth trajectories for each species using mixed effects models. ARIMA specifications were included in the residual terms to account for autocorrelation. The estimated parameters allowed us to calculate ontogenic traits and silvicultural metrics for each species. The results indicate that autocorrelation was a critical factor in growth depensation for all species studied, and was satisfactorily accounted for by the proposed modeling approach. Autocorrelation patterns on residuals showed a stochastic trend and were investigated by correlation structures of ARIMA(1,1,0) and ARIMA(2,1,0). Ontogenic traits and silvicultural metrics obtained for these species were biologically consistent, providing reliable and useful information to understand the population ecology of tropical trees and to inform management and conservation strategies of natural forests.