Innovative framework to simulate the fate and transport of nonconservative constituents in urban combined sewer catchments

Abstract

We have developed a probabilistic model to simulate the fate and transport of nonconservative constituents in urban watersheds. The approach implemented here extends previous studies that rely on the geomorphological instantaneous unit hydrograph concept to include nonconservative constituents. This is implemented with a factor χ that affects the transfer functions and therefore accounts for the loss (gain) of mass associated with the constituent as it travels through the watershed. Using this framework, we developed an analytical solution for the dynamics of dissolved oxygen (DO) and biochemical oxygen demand (BOD) in urban networks based on the Streeter and Phelps model. This model breaks down the catchment into a discreet number of possible flow paths through the system, requiring less data and implementation effort than well‐established deterministic models. Application of the model to one sewer catchment in the Chicago area with available BOD information proved its ability to predict the BOD concentration observed in the measurements. In addition, comparison of the model with a calibrated Storm Water Management Model (SWMM) of another sewer catchment from the Chicago area showed that the model predicted the BOD concentration as well as the widely accepted SWMM. The developed model proved to be a suitable alternative to simulate the fate and transport of constituents in urban catchments with limited and uncertain input data.

Water quality modeling in sewer networks: Review and future research directions

“…In relative terms, biomarkers, CH4 and microorganisms were more likely to have a lower level of model accuracy, which is likely due to their greater level of complexity in the processes affecting these parameters. As shown in Figure 3, higher levels of model performance (R 2 or NSE greater than 0.9) were generally associated with good data availability, as was the case for empirical process-driven models for H2S (a type of sulfide), where continuously monitored data were available (Ganigue et al, 2018), or for smaller SNs, such as the kinetic process-driven model applied to a real SN with an area of 3.16 km 2 (Morales et al, 2016). Finally, it can be observed from Table 2 that no strong correlations existed between model performance levels and model types, and we deduced that the model accuracy level was mainly affected by the process complexities of the quality parameters (including their concentration levels), the data availability and the scales of the problems being considered.…”

“…In parallel to the development of data-driven models, process-driven models have also been used for sewer water quality modelling, benefitting from their capacity for representing the transformation processes involving water quality parameters in SNs explicitly (Morales et al 2016. Process-driven water quality models can be further divided into two main sub-categories based on their properties.…”

“…Understanding is often attained by using a process-driven modelling approach, as this enables the underlying temporal and spatial dynamics/evolutions of the water quality parameters within SNs to be determined as a function of varying external conditions (Sharma et al 2014. Control can be defined as the interventions adopted to influence the behaviour of water quality parameters, mainly through manipulating the factors that can affect their reaction processes (Morales et al 2016, Guo et al 2019). It should be noted that within the system controlling processes, the prediction of the status is often required for some specific control strategies, such as predictive control and feedforward control (Langeveld et al 2013, Liu et al 2016a).…”

A review of pollution-based real-time modelling and control for sewage systems

“…However, that leads to the effect that information regarding the concentration during storm events are lost. The solution proposed by [52] is to use a probabilistic model that is able to describe the dynamics of the process. This reduces uncertainties because it does not require a complex parametrisation of the complete sewer network and the model only needs an input value for each catchment.…”

“…11 , where the overall flow is the combination of a dry weather flow and an overland flow, which is separated as a pervious and impervious overland flow.

Figure 11 Representation of a catchment by [52] .

…”