Modelowanie węgla w ekosystemach leśnych – przegląd narzędzi

The primary objective was to compare the structure, functionality, and applicability of five models commonly used in Europe – Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3), EFIMOD, European Forest Information SCENario Model (EFISCEN), CO2FIX, and CASMOFOR – under conditions relevant to Poland. The scope of the work included analysis of forest biomass dynamics, mortality, decomposition of dead organic matter, soil carbon pools, and simulation of natural and anthropogenic disturbances. The review was based on published literature and supplemented by a systematic comparison of model assumptions. The main features of the models are summarised in Tables 1 and 2, while differences in growth, mortality, decomposition, and disturbance handling are presented in Tables 3–6. The methodology involved a structured review of model architecture and input data requirements, emphasising their capacity to simulate forest growth and carbon balance. CBM-CFS3 was identified as the most advanced in handling disturbances and land-use change, offering a wide range of modules and disturbance matrices. EFIMOD integrates tree-level growth processes with soil dynamics via the ROMUL module, explicitly capturing C/N interactions. EFISCEN, based on matrix structures, allows large-scale simulations at landscape and European levels but is less precise for uneven-aged stands. CO2FIX and CASMOFOR use simpler cohort approaches, suitable for stand-level simulations, with CASMOFOR focusing on afforestation scenarios and short-term carbon balance. Figures and tables show that the models differ significantly in the representation of tree growth, mortality, and decomposition, with EFIMOD providing the highest spatial resolution and EFISCEN ensuring scalability to policy-level applications. The results highlight that each model offers distinct strengths. CBM-CFS3 provides detailed decomposition and disturbance modules, EFIMOD excels in linking tree and soil processes, while EFISCEN is most effective for large-scale scenario analysis. CO2FIX and CASMOFOR remain practical tools for stand-level management and afforestation monitoring. Despite these differences, no universal “best” model could be identified. Instead, model selection depends on research objectives, data availability, and spatial scale. A consistent conclusion is that the use of statistical measures such as RMSE, NSE, and R², combined with ecological validation (carbon balance, life cycles, response to disturbances), is essential for evaluating model efficiency. In summary, the comparative analysis (Tables 1-6) demonstrates that forest carbon cycle modelling is highly diversified, ranging from simple balance models to complex disturbance-oriented systems. Fragmentation of approaches creates challenges for harmonisation at European and national levels. Nevertheless, adopting hybrid evaluation frameworks can improve methodological consistency and enhance the role of carbon models in supporting climate policy and sustainable forest management.