西安市气候变化人为影响度ARIMA测算评价

Prediction and evaluation of the human impact on climate change in Xi'an with ARIMA

利用时间序列ARIMA建模及灰色关联分析方法,分析了西安市气候变化人为影响度的变化特征。结果表明:气象资料通过了统计软件的预处理,证明了运用数理统计软件处理气象数据的可行性,为处理气象数据提供了另一种分析方法。经过比较多种模型,确定ARIMA(1,1,4)模型,因其BIC信息量最小。根据ARIMA(1,1,4)模型计算出的1950-1987年的气温值与原始计算值吻合很好,相对误差较小。进一步利用该模型迭代计算了西安市1988-2007年的气温值。通过对各指标的灰色关联度分析,最终确定西安市气温变化的人为影响度为43.45%,其中耕地面积的变化是主要因素,其次为工业发展及人口变动等。

Time - series model ARIMA and gray correlation were used to analyze the degree of human impact on climate change in Xi＇an City. Results indacated that the meteorological data was inspected successfully by pre - processing of the statistical software, which proved the feasibility of processing meteorological data by mathemati- cal statistical software ; meanwhile it provided another method to analyze meteorological data. The ARIMA （ 1, 1, 4） model was chosen to evaluat because its BIC value was the lowest. According to this model, the calculated result of temperature from 1950 - 1987 fitted well with the original value, and the error was in the rang of accep- ting. So the model was used to predict the temperature in Xi＇an City for 1988 -2007. Base on these, we can get the degree of human impact of 43.45% in Xi＇an City. The change of cultivated land area was the main factor and industrial development and population changes were followed.