A flexible modelling framework for model creation based on perceptual understanding in integrated human-water systems

Perceptual and conceptual modelling has been used historically by hydrologists to develop models rooted in a physical reality. As human activity increases and intertwines with the natural world, hydrological systems cannot be treated in isolation, particularly in urbanised areas. We argue that expanding our models and modelling approaches to consider interactions with water infrastructure can help us to identify the dominant processes and interactions within coupled human-water systems (CHWS) and guide our modelling processes towards models that produce results for the right reasons. We develop a three-level perceptual modelling approach that maps CHWS complexity in a systemic way. Perceptual models are representations of a system of interest based on stakeholder’s understandings and rooted in reality (e.g. visualised as a cross-section of the system with processes mapped on). Conceptual models are representations that break down the perceptual model to a component and state level (e.g. visualised as buckets and flows). From these definitions the framework was created to construct a computational model from an initial understanding of the region of interest. This framework prioritises engagement of different stakeholders at key junctions in the model making process, as well as providing a clear roadmap of modelling decisions. We applied this modelling approach for the Mogden Wastewater Catchment in North West London. The Mogden case study captures the interaction between surface water, groundwater and the sewer network, giving insight into the understudied field of sewer infiltration/exfiltration, highlighting the framework’s ability to better understand impact and behaviour of complicated flow paths. The case study highlights how this framework allows for the identification of interactions between human activity and the urban water system, producing models which are rooted in reality. The case study further revealed the benefit of flexible models, such as the implemented WSIMOD, for this framework, capturing diverse system perceptions and adaptability to include dominant processes.