WSIMOD (Water Systems Integrated Modelling)

WSIMOD is a Python modelling framework that integrates the movement of water as affected by physical and human processes.

Project website

WSIMOD documentation ❐

GitHub repository ❐

Details

The water cycle is highly interconnected; water fluxes in one part depend on physical and human processes in all other parts. For example, rivers are simultaneously:

• a water supply
• a receiver of waste water
• an aggregate of many hydrological, biological and chemical processes

Because of this, simulations of the water cycle that have highly constrained boundaries may exclude key interactions that could create unanticipated effects or unexpected opportunities. Integrated environmental models aim to resolve this, but they do have some important limitations; in particular, there is a significant need for a parsimonious, self-contained suite that is both accessible and easy to set up.

Traditional approaches to water system modelling broadly fall into two categories:

• highly numerical models that excel in representing individual subsystems
• systems dynamics models that create broad representations but lack a physical basis

Early attempts at a physical representation of the water cycle combined existing numerical models through an integration framework. While successful, this approach has an incredibly high user burden, because each subsystem model is so detailed and consequently difficult to customise. For example, SWAT is one of the most widespread models of the rural water cycle; SWMM is similarly widespread for the urban water cycle. It has been demonstrated that they can be coupled but integrated applications have been limited despite this apparently powerful combination of two almost ubiquitous models.

The WSIMOD Python package provides a parsimonious and self-contained suite for integrated water cycle modelling. It brings together a range of software developed over the course of three years during the CAMELLIA project. Urban water processes are based on those presented and validated in the CityWat model, while hydrological and agricultural processes come from the CatchWat model. WSIMOD also provides an interface for passing information between different model components, allowing all parts of the water cycle to interact with all other parts. The result is a simulation model that is easy to set up, highly flexible and ideal for representing water quality and quantity in 'non-textbook' water systems (which, in our experience, is nearly all of them).

The package provides a variety of tutorials and examples to help modellers:

• create nodes (representing subsystems within the water cycle)
• connect them together with arcs (representing the fluxes between subsystems)
• organise them into a model that creates simulations

Note that WSIMOD is not intended to be a substitute for sophisticated physical models or a system dynamics approach. In applications where detailed hydraulic or hydrological process representations are needed (for example, informing the design of specific pipes or cases where processes are hard to quantify, such as representing social drivers of population growth, etc.) there are likely better tools available.

Our case studies highlight that WSIMOD is most useful in situations where it is necessary to represent cross-system processes and interactions to answer questions. The parsimonious representations used are computationally fast and flexible in capturing a wide range of system interventions, giving another benefit to WSIMOD.

The WSIMOD has been used in multiple regional planning strategy projects for the Environment Agency and the Greater London Authority.

Informing London's subregional water strategy ❐