基于哑变量的油松人工林和天然林生长模型

为准确掌握油松生长过程、改善油松经营管理模式，利用北京地区油松连续清查数据，在Richards模型基础上，考虑林分起源的差异，在模型中引入哑变量，建立北京地区不同林分起源相容性油松林分生长模型。结果表明：所建立的含哑变量的油松生长模型，对油松林分生长模型的拟合效果较好，决定系数高达0.9380和0.9918；油松蓄积量的拟合效果比断面积好，人工林的拟合效果高于天然林。用检验数据对模型进行适应性检验，林分断面积和蓄积量生长模型的预估精度均在90％以上。

Dummy variables models in Pinus tabulaeformis artificial forest and natural forest growth

Studying on Pinus tabulaeformis forest growth process and establishing its growth model can provide an important reference for grasping its growth process more accurately so as to improve forest management. Using the periodically inventory data of P. tabulaeformis in Beijing, on the basis of Richards model, compatible forest growth models of P. tabulaeformis with different stand origins were built through introducing the dummy variables in the model. The results showed that P. tabulaeformis growth models with the dummy variables were better simulation, R2 was as high as 0.938 0, 0.991 8, respectively. The fitting effect of stand volume for P. tabulaeformis was better than its stand basal area, the fitting effect of artificial forest was better than natural forest. With the independent test of established models using validation data, the forecasting precisions of stand basal and volume growth models were higher than 90%.