Mixed forest specific calibration of the 3-PGmix model parameters from site observations to predict post-fire forest regrowth

Abstract

Fire disturbance has been one of the main reasons for the alteration of forest succession trajectory and forest composition in the Daxing’anling Mountain of Inner Mongolia. Therefore, predicting the dynamic process of post-fire regrowth is critical for understanding the specific forest succession trajectory of this region. The 3-PGmix model (i.e., Physiological Principles in Predicting Growth for mixed stands) has been reported as a powerful tool for predicting the growth of mixed forest species. However, simulating post-fire regrowth of mixed forest with 3-PGmix remains challenging due to the uncertainty of species-specific as well as site-specific model parameters. Based on the field measurements from a wide range of environmental conditions, the 3-PGmix model was calibrated by conducting the sensitivity analysis and optimization of species-specific parameters for mixed forest stands of larch (Larix gmelinii) and birch (Betula platyphylla) with the PEST model (Model-Independent Parameter Estimation), and by using our newly developed methods to accurately estimating site-specific parameters (e.g., fertility rating, stand density, climate factors). The calibrated 3-PGmix model was tested against the independent sites and predicted the characteristic of post-fire forest regrowth. The sensitivity analysis shows that the parameters describing the forest canopy are more sensitive to the objective function of the PEST model. The calibrated 3-PGmix model performed well on independent sites, resulting in agreement with the field-measured diameter at breast height (DBH), the biomass of components, and stand density with a bias of less than 15% on most sites. The 3-PGmix model prediction of post-fire forest recovery characteristics found that the highest recovery rate of 5.5% per year was predicted from the moderate burn severity scenarios. A higher fertility rating was predicted to accelerate the process of post-fire forest successional and shorten the duration time when the species proportion achieves balance. The mixed-forest specific calibration of the 3-PGmix model in this study enables the explicit prediction of post-fire forest succession, and more simulations over different spatial scales with a wide range of environmental conditions by coupling with remote sensing observations can be expected.