College of Life Sciences, Capital Normal University, Beijing, 100048, China.
Environ Monit Assess. 2010 Jan;160(1-4):465-78. doi: 10.1007/s10661-008-0711-x.
This study used geographic information system techniques and geostatistics methods to evaluate the effectiveness of routine water quality monitoring in the western segment of the Miyun reservoir in Beijing. Methodologies as well as the sampling design are evaluated. The single-layer evaluation and three integrated evaluation methods including principal component analysis (PCA), ordinary kriging (OK)_Mean, and Mean_Layers were used to validate the effectiveness of evaluation methods, and the effectiveness of each sampling design was validated by comparing their errors. Results indicated that, while a single-layer evaluation only shows the trophic state of water at a specific level, an integrated evaluation synthetically analyzes and evaluates the trophic state of the entire water body. Furthermore, results of the integrated analysis show that a PCA method is more accurate and can represent the trophic state of the entire water body. The OK_Mean and Mean_Layers methods are only able to represent the mean level for trophic state of the entire water body but cannot reflect local trophic state and distribution details. Although methods used in the routine monitoring of Miyun reservoir have some similarities to the OK_Mean and Mean_Layers methods, their range of errors and uncertainty are greater because of a lack of detailed spatial continuous information. The analysis on the number of sampling points shows that, within a certain range of error, minor changes of sampling points will have no obvious impact on the monitoring results. For the routine monitoring of western Miyun reservoir, using only three to five sampling points for monitoring is inadequate. According to our analysis, it is more appropriate to use at least ten sampling points for monitoring these areas.
本研究运用地理信息系统技术和地统计学方法，评估了北京市密云水库西部常规水质监测的效果。评估了方法和采样设计。采用单因子评价和三种综合评价方法，包括主成分分析（PCA）、普通克里金（OK）均值和均值层，验证了评价方法的有效性，并通过比较其误差来验证了每种采样设计的有效性。结果表明，单因子评价仅能显示特定层次的水体富营养状态，而综合评价则综合分析和评价整个水体的富营养状态。此外，综合分析的结果表明，PCA 方法更准确，能够代表整个水体的富营养状态。OK 均值和均值层方法仅能代表整个水体富营养状态的平均水平，而不能反映局部富营养状态和分布细节。尽管密云水库常规监测中使用的方法与 OK 均值和均值层方法有些相似，但由于缺乏详细的空间连续信息，其误差和不确定性范围较大。对采样点数目的分析表明，在一定误差范围内，采样点的微小变化对监测结果不会产生明显影响。对于密云水库西部的常规监测，仅使用三到五个采样点进行监测是不够的。根据我们的分析，在这些区域使用至少十个采样点进行监测更为合适。
