• 关键词：预警  预警指数  贝叶斯推断  污染负荷  滇池流域

• 基金项目：国家水体污染控制与治理科技重大专项(No. 2010ZX07102-006,2013ZX07102-006)

• 刘岩
• 北京大学环境科学与工程学院, 北京 100871

• 周丰
• 北京大学城市与环境学院, 地表过程分析与模拟教育部重点实验室, 北京 100871

• 赵志杰
• 北京大学环境科学与工程学院, 北京 100871

• 摘要：以流域模型HSPF和贝叶斯递归回归树算法(BRRT)为计算模块,建立了滇池流域27条主要入湖河流和33个散流区的TN、TP入湖负荷预报预警系统.同时,构建了考虑历史排放规律和预测负荷计算预警指数(EWI)评价预警等级的两套预警体系.结果表明:BRRT替代模型在流域内以农业面源为主的柴河子流域校准和验证的可决系数R²均大于0.8,模拟结果相对可靠;根据预警时间选取预警体系;适用于6月份之前的预警体系一(EWS-1),利用现状排放量和历史排放量的关系计算预警指数.适用于6月份之后的预警体系二(EWS-2),主要考虑现状排放量、历史平均排放量、排放限值和预测排放量之间的关系计算3个预警指数,最终以最严格的作为综合预警指数EWI评价预警等级;根据柴河子流域"十二五"规划的TN和TP排放限值为130.2 t · a^-1、6.8 t · a^-1,应用此系统对2011年2、4、9和11月做出预警检验,各月份预警结果基本处于红色和黑色预警,该系统可为流域提供预警支持.

• Abstract：A forecasting and early-warning system was developed using Hydrological Simulation Program-Fortran (HSPF) and Bayesian Recursive Regression Tree (BRRT) as computing module for Lake Dianchi Watershed. It was designed for riverine TN and TP loadings of 27 major rivers and 33 bulk flow sub-watersheds. Early Warning Index (EWI) was used to evaluate the early-warning grade, considering historical emissions and predicted pollutants loadings. Two early-warning systems were then established using EWI. The coefficients of determination (R²) were greater than 0.8 for BRRT calibration and verification processes in the sub-watershed Chaihe, with agricultural non-point pollution as dominated source. The model results showed that (a) the calibrated BRRT model was reliable for further simulations; and (b) how to choose early-warning system depends on early-warning time. Early-warning system one (EWS-1) is applied for time before June, using the relationship between current pollutants loadings and historical emissions to calculate EWI. The early-warning system two (EWS-2) is applied for time after June. It considers current pollutants loadings, historical average emissions, emission limits and predicted pollutants loadings to calculate 3 EWI, using the most rigorous one to evaluate early-warning grade. TN and TP emission limits of sub-watershed Chaihe in the "12^th Five-Year Plan" was 130.2 t · a^-1 and 6.8 t · a^-1, respectively. Applying this early-warning system in sub-watershed Chaihe, it showed that the warning results of February, April, September and November of 2011 were all between red and black warning. This system can thereby provide early-warning support for Lake Dianchi Watershed.

• Key words：early-warning  Early Warning Index  Bayesian inference  pollution loadings  Lake Dianchi Watershed

• 
  摘要点击次数： 1271 全文下载次数： 2608