A nonlinear mixed-effects modeling approach for ecological data: Using temporal dynamics of vegetation moisture as an example

ODDI, FACUNDO J.; MIGUEZ, FERNANDO E.; GHERMANDI, LUCIANA; BIANCHI, LUCAS O.; GARIBALDI, LUCAS A.

Ecology and Evolution

wILEY

Año: 2019 vol. 9 p. 10225 - 10240

1. Increasingly,often ecologist collects data with nonlinear trends, heterogeneousvariances, temporal correlation, and hierarchical structure.Nonlinear mixed‐effects models offer a flexible approach to suchdata, but the estimation and interpretation of these models presentchallenges, partly associated with the lack of worked examples in theecological literature.2. We illustrate thenonlinear mixed‐effects modeling approach using temporal dynamicsof vegetation moisture with field data from northwestern Patagonia.This is a Mediterranean‐type climate region where modeling temporalchanges in live fuel moisture content are conceptually relevant(ecological theory) and have practical implications (firemanagement). We used this approach to answer whether moisturedynamics varies among functional groups and aridity conditions, andcompared it with other simpler statistical models. The modelingprocess is set out ?step‐by‐step?: We start translating theideas about the system dynamics to a statistical model, which is madeincreasingly complex in order to include different sources ofvariability and correlation structures. We provide guidelines and Rscripts (including a new self‐starting function) that make dataanalyses reproducible. We also explain how to extract the parameterestimates from the R output.3. Our modelingapproach suggests moisture dynamic to vary between grasses andshrubs, and between grasses facing different aridity conditions.Compared to more classical models, the nonlinear mixed‐effectsmodel showed greater goodness of fit and met statistical assumptions.While the mixed‐effects approach accounts for spatial nesting,temporal dependence, and variance heterogeneity; the nonlinearfunction allowed to model the seasonal pattern.4. Parameters of thenonlinear mixed‐effects model reflected relevant ecologicalprocesses. From an applied perspective, the model could forecast thetime when fuel moisture becomes critical to fire occurrence. Due tothe lack of worked examples for nonlinear mixed‐effects models inthe literature, our modeling approach could be useful to diverseecologists dealing with complex data.p { margin-bottom: 0.25cm; line-height: 120%; }