广东省森林覆盖率的影响因素分析和模型预测 ——基于灰色关联分析和GM(1,1) 模型

基于2005—2015 年广东省森林覆盖率的原始数据，选取年平均气温、降水量、年日照时数、有 林地面积、活立木蓄积量、林业产值、造林面积及育苗面积8 个指标对森林覆盖率的影响要素进行灰色关 联度分析。分析结果显示，关联度最大的3 项因素依次是：有林地面积0.993 4、年平均气温0.992 1、年 降水量0.973 2。为验证森林覆盖率灰色系统预测模型的适用性性，根据2005—2010 年的森林覆盖率数据， 分别建立GM(1,1) 模型、多项式回归模型和Logistic 回归模型，对2011—2016 年的森林覆盖率进行预测， 并与实际值进行比较，3 个模型的预测相对误差平均值依次是0.69%、1.08%、1.28%。结果表明在获取的 年份数据较少时，采用灰色系统森林覆盖率预测模型精度高于多项式回归模型和Logistic 回归模型，预测 适用性更优。

Analysis of Influencing Factors and Model Prediction of Forest Coverage in Guangdong Province: Based on Grey Correlation Analysis and GM (1,1) Model

Based on the original data of the forest coverage in Guangdong province from 2005 to 2015, eight indicators were selected to analysis the influencing factors on forest coverage through the gray correlation analysis. The indicators were annual average temperature, amount of precipitation, annual sunshine duration, forest land area, standing forest stock, output value of forestry, afforestation area and seedling area. The results showed that the three factors with the highest degree of correlation were: the forest land area (0.993 4), the annual average temperature (0.992 1) and the annual precipitation (0.973 2). In order to verify the adaptability of the gray system forecast model of forest coverage, three models (GM (1,1) model, polynomial regression model, and Logistic regression model) were established based on the original data of forest coverage from 2005 to 2015. The prediction of forest coverage from 2011 to 2016 was compared with the actual value, and the average relative error of the three models was 0.69%,1.08% and 1.28%, respectively. The results showed that, when the annual data was less, the accuracy of the gray system forecast model was higher than that of the polynomial regression model and the Logistic regression model, and its adaptability is more excellent.