基于趋势及回归分析的珠三角城市群需水预测

对近10年珠三角城市群用水量进行了特征及结构分析,其中分析了总用水量和农业、工业、城镇公共、居民生活及生态环境用水量等与地区生产总值、年末人口数等参数的相关性,结果表明:与一般用水量随着经济与人口增长而增长的结论不同,近年来,珠三角地区城市总用水量与地区生产总值表现较强的负相关性,与年末人口数也呈一定程度负相关。基于相关性分析成果,构建基于经济、人口为自变量的二元回归模型及趋势模型,对城市群水资源需求进行了预测,并与常用的分项水量预测法进行对比,结果表明:趋势模型和二元回归模型预测精度良好,与实际值偏差在2%以内;趋势模型预测结果与实际值最为接近,二元回归模型预测结果偏小,而常用的分项水量预测法结果偏安全。相比分项水量预测法,趋势及二元回归预测方法具有较高精度和便利操作性,可供类似城市群需水预测研究参考。

regression analysis-based water demand prediction of Pearl River Delta urban agglomeration

The characteristics and structure of the water consumption of the Pearl River Delta urban agglomeration in recent 10 years are analyzed herein，in which the correlations between the total water consumption as well as the agricultural，industrial，urban public，resident living，eco-environmental water consumptions，etc. and the parameters such as gross regional product，population at year-end，etc. are analyzed. The results show that different from the conclusion，i.e. the water consumption generally increases along with the increases of both the economic and population，a stronger negative correlation presents between the total water consumption of the Pearl River Delta region and the gross regional product therein in recent years. Furthermore，it also has a negative correlation to the population at year-end. In accordance with the correlation analysis result，the dualistic regression model and trend model，which take the economics and population therein as the independent variables are established，and then the water demand of the urban agglomeration is predicted. The result is compared with that obtained from the conventional itemized water consumption prediction method，which shows that the prediction accuracies of both the trend model and the dualistic regression model are better with the offset within 2% of the actual value； i.e. the result from the trend model is most approached to the actual value， the result from the dualistic regression mode is a little bit too small， while the result from the conventional itemized water consumption prediction method is a bit too safe. If compared with the itemized water consumption prediction method，the prediction method given herein has a high accuracy and better operability，thus can be taken as a reference for the study made on the water demand prediction of the similar urban agglomeration.