Effects of soil properties, stand structure, and species diversity on forest ecosystem carbon stocks in a cold- temperate watershed of northeastern China

Abstract Cold-temperate forests store large amounts of carbon, but the distribution of carbon among vegetation, litter, and soil pools remains poorly quantified at watershed scales. We assessed forest ecosystem carbon stocks in the Ergong River Basin of the Xiaoxing'an Mountains, northeastern China, using field measurements and forest inventory data. Carbon stocks were estimated for tree, shrub, herb, litter, and soil layers across four dominant forest types. Mantel tests, principal component analysis, and piecewise structural equation model-ling were used to identify the soil, stand, and diversity factors associated with carbon storage. Soil was the dominant carbon pool, accounting for 61.57–77.68% of total ecosystem carbon stocks, whereas tree biomass contributed 88.15% of vegetation carbon. Total ecosystem carbon stocks were highest in mixed conifer-broadleaf forest (231.42 Mg ha -1 ), followed by larch forest (227.06 Mg ha -1 ), birch forest (204.24 Mg ha -1 ), and mixed broadleaf forest (187.54 Mg ha -1 ). High total carbon stocks were concentrated mainly in low-elevation ri-parian zones, whereas high vegetation carbon densities occurred more often at higher eleva-tions with weaker disturbance. Soil total carbon, nitrogen, organic carbon, phosphorus, and MnO were strongly associated with ecosystem carbon stocks. Mean diameter at breast height was the main correlate of vegetation carbon. The structural equation model indicated that soil total carbon had the strongest direct association with ecosystem carbon stocks, with addi-tional associations linked to stand structure and species dominance. These findings indicate that carbon accumulation in cold-temperate watershed forests is closely associated with soil nutrient status and stand structural attributes. The results provide field-based evidence for forest carbon assessment and management in northeastern China.