杉木不同造林密度生长预测模型的研究

A Study on the Growth Prediction Model of Chinese Fir in Different Forestation Densities

通过设置５种不同的造林密度试验来研究密度对杉木生长的影响，并运用人工神经网络方法来建立在不同造林密度条件下杉木生长模拟预测模型。结果表明：运用人工神经网络所建立的胸径生长模型、树高生长模型和材积生长模型的模拟精度都很高，分别达９６．６８％，９７．９２％和９８．５３％，因此，利用已建立的３个生长模型就可以对杉木的生长动态进行预测，这可为杉木的合理经营提供科学的依据和手段。

This paper deals with the effect of five kinds of densities (1875, 2805,3750,4500 and 6000 trees per hm 2) on growth in 29-year-old Chinese fir plantation, and the growth predictive models were built by the mean of artificial neural network. The results showed that the accuracy of D. B. H model, height model and individual volume model are all high, which are 96.68%、97.92% and 98.53% respectively. With the three models, the growth dynamic of Chinese fir can be predicted, and it also provided scientific method for the reasonable management of Chinese fir.