افزایش مدل های سیستم آب با یکپارچه سازی داده های بزرگ Enhancing water system models by integrating big data

1. Introduction The water resources community relies on computer models to conceptualize and reproduce behavior of systems, aiding in planning, design, and analysis. The use of computer models is growing due to the need for deeper insights into water systems and providing sustainable solutions for smart cities [1]. Models are formulated by developing a set of mathematical equations and rules, which mimic the real behavior of the system and decisions of stakeholders, and can be executed in an iterative fashion. These equations represent universal laws while parameters represent local systems. Parameters are typically characterized using averages, probability distributions to specify the likelihood of parameters at different states, and assumptions. Model parameters are updated to best reflect the actual system, often done manually when results deviate from field data. This fashion of updating models is time-consuming. Further, due to the speed at which some spatially heterogeneous variables (e.g. water demands and pre48 cipitation) change, it is nearly infeasible to manually update with fine resolution. Engineering advances in sensor and communication devices allow for the continuous monitoring of many systems including water systems. The purposes of these devices are to record and relay time series data with high frequency. Pertinent parameters measured by such devices include flow, quality, and stage; all of which are in situ. Technological advancements allow many sites to be monitored in near real-time with very little oversight. This type of measurement creates so-called big data, which relates to the collection in the data cycle – also including, storage, purification, and analysis of large-size data sets [2, 3].