Framework for developing hybrid process-driven, artificial neural network and regression models for salinity prediction in river systems

Abstract

Abstract. Salinity modelling in river systems is complicated by a number of processes, including in-stream salt transport and various mechanisms of saline accession that vary dynamically as a function of water level and flow, often at different temporal scales. Traditionally, salinity models in rivers have either been process- or data-driven. The primary problem with process-based models is that in many instances, not all of the underlying processes are fully understood or able to be represented mathematically. There are also often insufficient historical data to support model development. The major limitation of data-driven models, such as artificial neural networks (ANNs) in comparison, is that they provide limited system understanding and are generally not able to be used to inform management decisions targeting specific processes, as different processes are generally modelled implicitly. In order to overcome these limitations, a generic framework for developing hybrid process and data-driven models of salinity in river systems is introduced and applied in this paper. As part of the approach, the most suitable sub-models are developed for each sub-process affecting salinity at the location of interest based on consideration of model purpose, the degree of process understanding and data availability, which are then combined to form the hybrid model. The approach is applied to a 46 km reach of the Murray River in South Australia, which is affected by high levels of salinity. In this reach, the major processes affecting salinity include in-stream salt transport, accession of saline groundwater along the length of the reach and the flushing of three waterbodies in the floodplain during overbank flows of various magnitudes. Based on trade-offs between the degree of process understanding and data availability, a process-driven model is developed for in-stream salt transport, an ANN model is used to model saline groundwater accession and three linear regression models are used to account for the flushing of the different floodplain storages. The resulting hybrid model performs very well on approximately 3 years of daily validation data, with a Nash–Sutcliffe efficiency (NSE) of 0.89 and a root mean squared error (RMSE) of 12.62 mg L−1 (over a range from approximately 50 to 250 mg L−1). Each component of the hybrid model results in noticeable improvements in model performance corresponding to the range of flows for which they are developed. The predictive performance of the hybrid model is significantly better than that of a benchmark process-driven model (NSE = −0.14, RMSE = 41.10 mg L−1, Gbench index = 0.90) and slightly better than that of a benchmark data-driven (ANN) model (NSE = 0.83, RMSE = 15.93 mg L−1, Gbench index = 0.36). Apart from improved predictive performance, the hybrid model also has advantages over the ANN benchmark model in terms of increased capacity for improving system understanding and greater ability to support management decisions.

Introductory overview: Optimization using evolutionary algorithms and other metaheuristics

“…As mentioned in the Introduction, environmental simulation models are used extensively to support decision-making processes in a variety of application areas, such as: the development and evaluation of national and international environmental regulations (Giupponi, 2007;Laniak et al, 2013); land use management (Amato et al, 2018); natural hazard management (Newman et al, 2017); the operation and management of reservoir systems (Razavi et al, 2014); the assessment of environmental and human health (Morley and Gulliver, 2018;Reis et al, 2015); the management of river systems (He, 2003;Humphrey et al, 2016;Hunter et al, 2018;Ravalico et al, 2010) ; the management of drains (Humphrey et al, 2016); the management of air pollution (Baró et al, 2014;Borge et al, 2014); flood inundation assessment (Teng et al, 2017); groundwater management and remediation (Jakeman et al, 2016;Piscopo et al, 2015;Singh, 2014); the design of water distribution networks so as to minimize global climate impacts (Stokes et al, 2015a;Stokes et al, 2014b;Wu et al, 2010a); the prediction of and adaption to natural hazards such as floods or droughts (Basher, 2006); crop and livestock management (Moore et al, 2014;van Keulen and Asseng, 2018); the design of green infrastructure for stormwater management and urban renewal (Liu et al, 2014;Yigitcanlar and Teriman, 2015); and evaluating the effects of resource extraction by the petroleum (Fiori and Zalba, 2003), natural gas (McJeon et al, 2014), mining (Côte et al, 2010) and timber (Alavalapati and Adamowicz, 2000) industries. Environmental models are in such widespread use because they can be designed to effectively reproduce the dynamics of real-world systems under traditional management situations as well as alternative virtual realities, including different environmental conditions and management alter...…”

Process‐Guided Deep Learning Predictions of Lake Water Temperature

“…TGDS models use advanced empirical methods to extract pattern from data while also imposing structure or rules based on scientific theory. Because these hybrid models can be designed to remain true to accepted theory or physical laws while also learning very complex relationships when data are abundant, their predictions tend to be physically and biologically realistic, and more accurate than process-based models (see Fang et al, 2017;Humphrey et al, 2016;Hunter et al, 2018;Jia et al, 2019). Under the broad umbrella of TGDS models, Process-Guided Deep Learning (PGDL) pairs Earth systems process understanding with the most promising class of predictive tools.…”

A review of artificial neural network models for ambient air pollution prediction

“…The authors suggested that this requires further investigation in the future. A generic framework for developing both hybrid process and data-driven models of salinity in river systems, which is also applicable to other environmental engineering tasks, can be found in Hunter et al (2018). In the said proposed framework, the most suitable submodels are developed for each sub-process of a specific problem of interest based on consideration of model purpose, the degree of process understanding and data availability.…”