基于地理视角研究中国健康女性儿童流量-容积曲线上3项参考值的变化

目的：探索健康人肺功能与地理环境的关系，为用力依赖性肺功能参考值预计值方程的制定提供地理学基础。方法依照中国气候区划分区搜集并比较中国不同地区健康女性儿童 FEF25％、FEF50％和FEF75％参考值差异。同时获取经度、纬度、海拔高度、年日照时数、年平均气温、年平均相对湿度、年降水量、气温年较差和年平均风速等9项地理因素，并与3项参考值进行相关分析。运用支持向量机模型建立两者之间的关系，并插值出参考值的地理分布图。结果各气候区FEF25％、FEF50％和FEF75％参考值存在显著差异，并与地理因素中年日照时数和年平均相对湿度存在相关关系。依据两者关系建立的支持向量机模型具有较好的模拟精度，3项参考值空间分布存在显著的区域差异。结论地理因素与流量‐容积曲线上3项参考值存在显著关系，可以用来预测参考值范围。

Interpreting the variety of three reference values on MEFV of adolescent girls in China from a geographical perspective

Objective To explore the correlation between lung function of healthy people and geographical factors ,provide a geographical perspective to establish a predictive equation of force‐dependent lung function .Meth‐ods In accordance with China's climate zone districts ,the reference values of FEF25% ,FEF50% and FEF75% for healthy adolescent children in different parts of China were collected and compared .The normal reference value of forced expiatory flow was correlated to nine geographical factors ,including longitude (X1 ) ,latitude (X2 ) ,altitude (X3 ) ,annual sunshine duration (X4 ) ,annual mean air temperature (X5 ) ,annual mean relative humidity (X6 ) ,annual precipitation amount (X7 ) ,annual range of air temperature (X8 ) and annual mean wind speed (X9 ) .Support vector machines (SVR) was applied to establish predictive models for FEF25% ,FEF50% and FEF75% ,and the geographical distribution map of these three reference values were interpolated by using ArcGIS software .Results There were significant differences in the reference values of FEF 25% ,FEF50% and FEF75% for healthy adolescent children in dif‐ferent regions .Three reference values had significant correlation with annual sunshine duration and annual mean rel‐ative humidity .Based on SVR ,the established models had good simulation accuracy .There were notable regional differences in three spatial distribution maps .Conclusions Significance correlations exist between forced expiatory flow and geographical factors .Geographical factors can be used to predict three normal reference values .