Instituto Hidrográfico, Portugal.
Centro de Química Estrutural - Faculdade de Ciências da Universidade de Lisboa, Portugal.
Mar Pollut Bull. 2020 Sep;158:111371. doi: 10.1016/j.marpolbul.2020.111371. Epub 2020 Jun 18.
The assessment of long-term trends in river water composition is hampered by river composition heterogeneity, and sampling and sample analysis uncertainty. This work describes a novel methodology for the reliable detection of small river composition trends by taking all relevant uncertainty components into account. The methodology was applied to study the variation of nutrients concentration of Tagus river estuary in the extremely dry 2017 year. Mean nutrient concentrations were determined with an uncertainty that combines sampling and sample analysis uncertainty by the Monte Carlo Method. The nutrient concentration variation observed in two occasions is meaningful if the difference of mean concentrations is metrologically different from zero for a 95% confidence level. The observed meaningful NO increase, and SiO and NO variations is justified by dissolved oxygen reduction, decreased freshwater input and algal productivity variations. The developed tool can be applied to the assessment of other composition trends in rivers.
长期以来，河流水质成分的变化趋势一直受到河流成分异质性、采样和样品分析不确定性的阻碍。本研究描述了一种新的方法，通过考虑所有相关的不确定性成分，可以可靠地检测小河流水质成分的小趋势。该方法应用于研究在极其干旱的 2017 年塔霍河口营养物浓度的变化。通过蒙特卡罗方法确定了均值营养物浓度，该浓度不确定性结合了采样和样品分析的不确定性。如果两次测量的均值浓度差异在 95%置信水平下在计量上与零不同，那么所观察到的有意义的氮增加以及硅和氮的变化就可以得到解释，这是由于溶解氧减少、淡水输入减少和藻类生产力变化所致。所开发的工具可用于评估其他河流的成分变化趋势。
