Assessing model performance via the most limiting environmental driver in two differently stressed pine stands

Area of Science:

• Ecology
• Forestry
• Climate Science

Background:

• Climate change poses risks to global forest productivity.
• Accurate forecasting requires robust process-based models.
• Traditional model validation methods may not reveal underlying flaws.

Purpose of the Study:

• To present and validate a novel methodology for assessing process-based forest models.
• To evaluate model performance by examining the reproduction of the most limiting environmental driver (MLED) for gross primary productivity (GPP).
• To identify potential improvements for forest simulation models.

Main Methods:

• Analyzed seasonal MLED for GPP in contrasting pine forests (Pinus halepensis and Pinus sylvestris) using eddy-covariance data.
• Simulated forest productivity using the LandscapeDNDC model over a three-year period.
• Assessed model accuracy by comparing simulated GPP and MLED seasonality against observational data.

Main Results:

• The LandscapeDNDC model reproduced GPP seasonality but showed slight overestimation without fine-tuning.
• The model correctly identified temperature (Finland) and soil water (Israel) as primary limitations.
• Model failed to capture high-temperature and vapor pressure limitations in the Mediterranean forest, suggesting issues with stomatal behavior representation.

Conclusions:

• The MLED seasonality approach is a valuable tool for evaluating process-based forest models.
• Model evaluation using MLED seasonality can reveal specific weaknesses, such as stomatal response.
• This methodology offers a pathway for improving forest simulation models for climate change impact assessments.