*2.3. New Analytics*

To make sense of this increasing amount of information, research and practice have made significant progress towards better analytics, including but not limited to those: (i) Capable of extracting valuable information from the data (from smart alerts to customized advice for water users); (ii) performing better stochastic simulations to improve the ability to produce longer timeseries (based on observations) for long-term scenario development and stress-testing; (iii) performing advanced optimisation to identify better solutions in this information richer environment; and (iv) providing novel ways of visualizing and understanding the decision tradeoffs within complex decision spaces. Examples of these new analytics, include AI/ML analytics for proactive managemen<sup>t</sup> of water distribution systems (including burst detection) demonstrated in UK case studies [40,41], asset deterioration assessment [42], as well as the use of deep learning techniques for defining novel control strategies that are more robust against cyber-attacks of water distribution systems [43]. Examples also include recent work on using smart meter readings to parametrise residential water demand models [44] as well as the methods and tools developed to investigate the properties of these timeseries at fine timescales [29]. Based on this growing body of work, we are now in a position to assess for the first time if smart meters are effective in water demand managemen<sup>t</sup> (see for example the review by Sønderlund et al. [27] based on 21 relevant reports and publications) or at least pinpoint the additional information needed to make this transition, including the information content, granularity, frequency and method of delivery etc.

However, getting better historical data is only part of the story. Additional work in stochastics is enabling hydroinformatics to develop simulated timeseries that explicitly represent each process of interest with any distribution model and hence conserve all of the characteristics of historical datasets (e.g., [45]). These longer timeseries can be used to drive hydroinformatic models of complex hydro-systems to better account for relevant uncertainties. However, this substantially increases the (time) burden for optimisation. Recent attention to 'optimisation on a budget' [46] shows how surrogate strategies can be employed to allow for less evaluations of expensive objective functions in evolutionary optimisation. Other authors have also focused on the challenging problem of optimal design under uncertainty and developed optimisation algorithms that exploit the concept of 'real options' [47], thus introducing flexibility into the long-term design for water systems [48]. Although an overview of the developments in optimisation is outside the scope of this paper, this is one of the most prolific fields in hydroinformatics to date. The interested reader is pointed towards an overview of this dynamic field, with a focus on water distribution networks, included in Mala-Jetmarova et al. [49] and in Maier et al. [50] for a more general overview of optimisation in water resources in general. Lastly, it is worth pointing out that developing new algorithms does not necessarily lead to better understanding or decision making. Recent attention to analytics for advanced visualisation of decision spaces sugges<sup>t</sup> that developing visual analytics to explore the decision space in multi-objective (e.g., [51]) or in multi-stakeholder problems [52] is both important and necessary.

### *2.4. New Whole Water Cycle Socio-Technical System Models*

The industry's interest in exploring new options for infrastructure provision (incl. new more distributed options discussed above) is driven in part by the process of aging infrastructure and the resulting investment gap [16]. The interest has also prompted the development and application of whole (socio-technical) system models [53] that attempt a more direct investigation of the interplay between centralized and distributed infrastructure solutions. Furthermore, the focus is also shifting towards the (often ignored) interplay between infrastructure and users (as also argued persuasively in the context of socio-hydrology by Sivapalan [54]). This integration is currently being delivered (mostly) around three axes:



It is important to note here that in support to these more integrative explorations, the hydroinformatics community has been developing and demonstrating: (a) Integrated modelling frameworks [73,74]; (b) models as services, often based on open source solutions [75]; and (c) cloud-based modelling systems [76,77], sometimes coupling both local model components and remote web services [78] in an effort to reduce the overhead required to create an integrated model in the first place and make their explorative power more accessible to the water research and practitioner communities.

### *2.5. New forms of Interactive and Immersive Decision Making*

The multi-faceted, multi-discipline and increasingly more inclusive multi-stakeholder nature of water managemen<sup>t</sup> considerations (and environmental managemen<sup>t</sup> in general [79]) have given rise to new ways of setting the questions, visualizing potential results and experiencing system performance under different stresses. These ways include Serious Games [80], augmented/virtual (or mixed) reality (AR, VR, MR) and their combinations that enable a different level of immersive, playful experience of problems, options and decisions that can be used in various contexts, including operational, strategic and stakeholder collaborative decision-making. The basic idea of these (relatively new) approaches is that practical water and environmental challenges (and options to address them) can be better understood through a more direct experiential approach. These game-based learning approaches improve critical thinking, creative problem solving and teamwork [79]. They also allow stakeholders to experiment with decisions and outcomes in a safe and fun environment.

Work by several authors is currently finding its way into practical applications, engaging water stakeholders in collaborative decision making for such diverse fields as urban flood managemen<sup>t</sup> [52], water resources managemen<sup>t</sup> [81,82] and integrated asset managemen<sup>t</sup> [83]. At the same time, augmented reality applications (including applications in handheld devices and smartphones) have begun to be actively used in infrastructure inspection and rehabilitations (see for example the Vidente application reported in Schall et al. [84]). The significant potential for this technology is especially evident in cases where infrastructure is underground as in the case of water distribution and sewerage networks. These applications typically superimpose data from GIS systems (such asset databases) or even data from simulations on real world views. This linking of spatial/georeferenced information directly on the real-world entities that they characterize, greatly facilitates the use of relevant data during field work (e.g., asset rehabilitation, water quality monitoring). As such, it ensures increased efficiency in maintenance activities, as well as increased understanding and learning in educational field trips and field-oriented stakeholder engagemen<sup>t</sup> processes (e.g., stakeholder visits in innovation demonstration case studies). An example of the latter is students participating in the EcoMOBILE project [85], who used an augmented reality application, as part of a field trip to an ecologically important lake. The virtual information was overlaid on the physical lake including hotspots—guiding students in collecting water quality measurements—but also increasing their understanding of underlying processes. It could be argued that such an increased (and more importantly shared) understanding between stakeholders, makes for a good basis for more inclusive, consensus-driven decision making.
