基于哑变量的云南松蓄积生长模型

The volume growth model of Pinus yunnanensis based on dummy variables

为了解云南松林分蓄积生长规律,为更好的经营云南松林分提供理论依据,笔者基于云南省云南松一类清查数据,以Schumacher和Richards模型为基础模型,选取不同的林分密度指标,来拟合云南松林分蓄积生长模型,并在最优基础模型上引入哑变量,将不同林分类型合并,建立同时适用于间伐林分和未间伐林分的蓄积生长模型。结果表明:4个传统生长模型的模拟效果都较好,R^2和预估精度最高分别为0.971 6、96.79%,不同的林分密度指标对模型的预测效果有较大影响,且以选用林分断面积作为林分密度指标的Richards模型的预估效果最好,其预估精度为96.79%。在最优传统生长模型中引入哑变量后,模型的R^2和预估精度都比传统的生长模型稍高,分别达到0.973 0、96.84%。通过对模型的适应性检验,所建的哑变量模型的预估精度超过98%,且对间伐林分的拟合效果更佳,可以用来描述不同措施下云南省云南松林分的蓄积生长规律,也解决了不同类型林分合并建模不相容的问题。

In order to understand the laws of volume growth of Pinus yunnanensis Franch and provide a basis for a better management,by using the inventory data of P. yunnanensis in Yunnan Province, based on the Richards and the Schumacher models,through choosing different density indexes,the volume growth models were simulated. On the basis of the best conventional model,the dummy variables were introduced to combine the different type of stands trying to build a volume growth model which is applicable for both thinned stands and un-thinned stands. The results showed that all the four conventional models were fitting well,the maximum R^2 and forecasting precision reached 0.971 6,96.79% respectively. Besides,different density index has a significant impact on the forecast effect of the model,and the one based on the Richards as well as taking the stand basal area as the index proved to be the best with its forecasting precision as high as 96.79%. The R^2 and forecasting accuracy were slightly improved after introducing dummy variables into the best traditional model,the value of R^2 was 0.973 0 and the forecasting accuracy was 96.84%.Independent test of the models indicated that the forecasting precisions of the dummy variables model was higher than 98% and the fitting effect of thinned stands was better. The dummy variable model can be used to describe the volume growth rule of P. yunnanensis with different management measures in Yunnan Province,and can be compatible with unified modeling for different types stands.