A Stand-Element Increment Modelling Framework for Forest Decision Support Systems

Forest decision support systems (DSS) increasingly require growth-modeling solutions that remain robust when forest stands are structurally complex. This abstract describes a machine-learning workflow that models annual increments in height and diameter at breast height (DBH) for stand-forming elements, using dendrometric data and forest site type as predictors.

The main focus is multi-structural representation and ease of deployment inside the DSS. The workflow supports use of separate models for stand elements and combining their predictions into stand-level outputs, covering stands with few elements as well as stands with many elements.

The workflow is suitable for both operational use and research. In a DSS, the prepared model system can be loaded, inputs can be read from a database, and stand-level outputs can be produced for decision support. The component can also be linked to a database and combined with other analytical models. Outputs can then be presented as decision-relevant tables and visualizations.

A Lithuanian forest inventory dataset was used for model development and validation, and an initial performance summary and a brief workflow check are reported. The framework allows accuracy improvements through model updates and provides a simple path for reusing updated models in a DSS.