Moreau-Fournier Magali F, Daughney Christopher J
GNS Sciences, PO Box 30368, Lower Hutt 5040, New Zealand.
J Environ Monit. 2012 Dec;14(12):3129-36. doi: 10.1039/c2em30624g. Epub 2012 Oct 29.
Optimization of a water quality network through a change in sampling frequency is the only way to increase cost-efficiency without any reduction in the robustness of the data. Existing techniques define optimal sampling frequency based on analysis of historical data from the monitoring network under investigation. Their application to a large network comprised of many sites and many monitored parameters is both technical and challenging. This paper presents a simple non-parametric method for reviewing sampling frequency that is consistent with highly censored environmental data and oriented towards reduction of sampling frequency as a cost-saving measure. Based on simple descriptive statistics, the method is applicable to large networks with long time series and many monitored parameters. The method also provides metrics for interpretation of newly collected data, which enables identification of sites for which a future change in sampling frequency may be necessary, ensuring that the monitoring network is both current and adaptive. Application of this method to the New Zealand National Groundwater Monitoring Programme indicates that reduction of sampling frequency at any site would result in a significant loss of information. This paper also discusses the potential for reducing analysis frequency as an alternative to reduction of sampling frequency.
通过改变采样频率来优化水质监测网络，是在不降低数据稳健性的前提下提高成本效益的唯一途径。现有技术基于对所研究监测网络的历史数据进行分析来确定最佳采样频率。将这些技术应用于由众多站点和众多监测参数组成的大型网络，在技术上既具有挑战性。本文提出了一种简单的非参数方法来审查采样频率，该方法与高度删失的环境数据相一致，并旨在将降低采样频率作为一种节省成本的措施。基于简单的描述性统计，该方法适用于具有长时间序列和众多监测参数的大型网络。该方法还提供了用于解释新收集数据的指标，从而能够识别出未来可能需要改变采样频率的站点，确保监测网络与时俱进且具有适应性。将该方法应用于新西兰国家地下水监测计划表明，任何站点采样频率的降低都将导致信息的大量丢失。本文还讨论了降低分析频率作为降低采样频率替代方案的可能性。
