Estimation of real-time N load in surface water using dynamic data-driven application system

Agricultural, industrial, and urban activities are the major sources for eutrophication of surface water ecosystems. Currently, determination of nutrients in surface water is primarily accomplished by manually collecting samples for laboratory analysis, which requires at least 24 h. In other words, little to no effort has been devoted to monitoring real-time variations of nutrients in surface water ecosystems due to the lack of suitable and/or cost-effective wireless sensors. However, when considering human health or instantaneous outbreaks such as algal blooms, timely water-quality information is very critical. In this study, we developed a new paradigm of a dynamic data-driven application system (DDDAS) for estimating the real-time loads of nitrogen (N) in a surface water ecosystem. This DDDAS consisted of the following components: (1) a Visual Basic (VB) program for downloading US Geological Survey real-time chlorophyll and discharge data from the internet; (2) a STELLA model for evaluating real-time N loads based on the relationship between chlorophyll and N as well as on river discharge; (3) a batch file for linking the VB program and STELLA model; and (4) a Microsoft Windows Scheduled Task wizard for executing the model and displaying outputs on a computer screen at selected schedules. The DDDAS was validated using field measurements with a very good agreement prior to its applications. Results show that the real-time loads of TN (total N) and NO_x (nitrate and nitrite) varied from positive to negative with the maximums of 1727 kg/h TN and 118 kg/h NO_x and the minimums of −2483 kg/h TN and −168 kg/h NO_x at the selected site. The negative loads occurred because of the back flow of the river in the estuarine environment. Our study suggests that the DDDAS developed in this study was feasible for estimating the real-time variations of TN and NO_x in the surface water ecosystem.