Objective In this paper, taking the coniferous and broadleaved mixed forest at different development stages in Jiaohe, Jilin Province of northeastern China as research object, the interannual variation of kernel density estimation curve of seedling density was compared and analyzed. The relative effects of marginal changes of soil factors and distribution changes of soil factors on the dynamics of seedling density distribution were discussed.
Method Seedling investigation sample plots were systematically arranged in 420 m × 520 m half-matured forest (HF) sample plots and 500 m × 840 m mature forest (MF) sample plots. Quantile regression and counterfactual decomposition were used to test the marginal effects of environmental factors on seedling density at different quantile levels, and then to identify the leading factors resulting in the unequal change of seedling density.
Result The kernel density estimation curve of seedling density showed a positive skewed distribution with peak value shifting to the left and long tail extending to the right. At the high quantile of θ = 0.90, the change value of seedling density was −5.9 in HF sample plots and −2.6 in MF sample plots. The change value of seedling density in HF sample plot was 5.8 when θ = 0.75, and that in MF sample plot was 2 when θ = 0.50 and θ = 0.75. The greater change of seedling density at high quantile reflects the inequality of probability distribution of the right single tail. The relative effects of coefficient effect and covariant effect on seedling density distribution were different in varied estimated quantiles. The coefficient effect in HF sample plots had a high explanation on all estimated quantiles; in MF sample plots, the covariant effect had 89% explanation for the change of seedling density at θ = 0.50 quantile. In other estimated quantiles, the explanation of coefficient effect was higher. Therefore, the relative effect of distribution of soil factors on the distribution of seedling density was greater, and the marginal change of soil factors was the main factor leading to unequal change of seedling density. Most of the changes in the distribution of available nitrogen, available phosphorus and available potassium in soil explained less than 30% of the changes in seedling density at each estimated quantile, while soil water content and soil pH had a relatively greater effect on the change of seedling density.
Conclusion The interannual variation of seedling density is unequal in different quantiles, especially in the high quantile. The marginal change of soil factors and the distribution of soil factors determine the survival dynamics of seedlings.