COMPARISON OF MODEL-BASED FOREST STAND BIOMASS ESTIMATION: THE CASE OF SPRUCE PHYTOCOENOSES IN THE MURMANSK REGION

This study presents a comparative analysis of the effectiveness of various regression models for estimating forest stand phytomass stocks in spruce forests developed on Al-Fe-humic podzols at the northern tree line in the Murmansk region. Research was conducted on sample plots established across diverse landscape positions, including automorphic, transitional, and accumulative sites. The relevance of this work stems from the critical need for reliable data on phytomass stocks, as phytomass represents a key carbon reservoir within taiga phytocoenoses. Given the region’s heterogeneous and rugged terrain, accounting for topographic factors is essential to minimize errors in phytomass pool estimations. A comparison of four models revealed that the model developed by J. Repola et al. is the most applicable for assessing forest stand phytomass. Leveraging a large dataset from similar environmental conditions, this model demonstrates high predictability across all types of elementary landscapes and currently offers an optimal compromise for regional phytomass assessments, despite identified limitations in analyzing the internal fractional structure of phytomass. Conversely, the regional model is restricted to aboveground phytomass and is suitable primarily for automorphic conditions. Findings indicate that while V. A. Usoltsev’s model is appropriate for estimating spruce phytomass, it systematically underestimates birch stocks – a key associated species in northern taiga spruce forests. Furthermore, the model proposed in the guidelines of the Ministry of Natural Resources of the Russian Federation proved ineffective for evidence-based monitoring due to significant overestimations of birch phytomass. The lack of reliable allometric equations for belowground phytomass (roots) remains a substantial knowledge gap, alongside uncertainties in estimating the fractional structure of tree phytomass stocks. These results underscore the necessity of gathering experimental data to refine regional models and integrate landscape positions. Such refinements will facilitate the minimization of errors in remote sensing calibration and enhance the accuracy of quantifying the role of Murmansk region forests in the global carbon cycle.