So models have a social and political life. What does this mean for a modeller in practice?

How do you see your responsibility as modeller for the influence a model can have? What were you proud of, and what would you avoid in the future in relation to your modelling practice?

These were the questions we shared with speakers and participants of a conversation hosted at IHE Delft on 12th February. The topic especially attracted early-career modellers, which points towards a need to think deeper on the relations between models and society in education and in relation to using tools developed by others. To keep the conversation going, we share the outcomes of the session in this blog and conclude with thoughts on how we facilitated this session.

So why talk about politics and society when it comes to models?

Models validate and influence how people think about and interact with water, both in terms of the ideas behind them and the actual data and tools used. Many choices—whether obvious or hidden, and by modellers, funders, commissioners or users—affect how they are designed, what they can do, and who can engage with the process.

In the past years, a small but growing community of scholars has further explored how water models have a social and political life and with what effects [1][2][3][4][5]. Namely, the development and use of water models is influenced by the people, institutions, and technologies involved (or excluded), and in turn, water models and the modelling processes can influence how people and institutions perceive and understand water problems and conceive solutions. This asks for a modelling philosophy that concerns both the model, the modellers and society at large [6], and related to this, it brings up the question of what responsibility and agency modellers have.

So, how do we start up conversations about the social and political life of models?

Talking about responsibility, politics and modelling practices can be a sensitive topic. From our personal experiences, we noticed, for instance:

• The conversation can undermine deeply held beliefs in the neutrality and objectivity of models and science in general.
• Some may feel that disproportionate responsibility is put on the shoulders of the modeller in such conversations.
• Challenging another person’s practices can sometimes feel like the easiest form of criticism.
• Different disciplinary languages and ways of experiencing and navigating the world make it difficult to understand each other.

Therefore, for the session organised on the 12th of February at IHE Delft, we made the following choices:

• We set expectations with participants that the session would focus on sharing a diversity of experiences and opinions. We explicitly invited people to be curious and kind, and opted for small break-out groups.
• Participants shared a common goal: contributing to much-needed socio-environmental transformations, albeit in different ways.

We first asked Tobias Krueger (IRI THESys), Bich Tran, Claire Michailovsky and Jonatan Godinez Madrigal (IHE Delft) to share their experiences first, as input for conversations amongst the participants: As an experienced modeller, Tobias pointed out that sometimes modellers need to say no to modelling when it does not fit, but that it also could be a privilege to be able to say no. Bich’s research on uncertainties led her to question how to communicate uncertainties in a model, without undermining its usefulness. By working with social science disciplines, she is exploring how reflexivity and interdisciplinary collaboration can help her with this challenge. As a modeller, Claire experienced the challenge of developing models and communicating validity and usefulness to funders, commissioners and model users with high expectations and pressure to deliver. She is collaborating with communities in the Athi river basin, Kenya to explore the potential of using public-domain satellite data for local needs, and at the same time, working on improving the scientific rigor and technical feasibility of satellite-based models. Jonatan is an inter-disciplinary scientists who repurposed a model that was used to legitimise decisions that would harm rural communities living near the Zapotillo dam, Mexico. He emphasised that models need to align with the questions that are socially legitimate, as opposed to questions based on vested interests. The ethical questions are: which problem and whose interests do models serve? Which ones are left out? And most importantly, what can we do to challenge - and compensate for - these uneven playing fields?

These insights served as input for discussions in smaller groups in break-out sessions, from which we distilled the following key take-aways:

• Models as specific simplifications:

  • Contextualisation: Models are simplifications of reality, with specific decisions on boundaries and how to represent space and time. To support these design choices, always make sure to understand the context of what is modelled.
  • Constraints: Modelling is a way to explore the world, but the world’s complexity is difficult to capture in the constraints of a model.
  • Uncertainty: Uncertainty should be better understood, communicated, and managed. Model users should be engaged in setting expectations in a useful manner.
  • Society: How to include “the society” in models? There is no one-size-fits-all answer, and bringing up the question provides an opportunity for collaboration and ‘intra-action’ between disciplines.

• The modeller as mediator in practice

  • Political modelling choices: It is sometimes up to the modellers to recognise the politics in their modelling practice and model-supported decisions, and refer these back to other actors. For instance, optimisation criteria for water distribution are political decisions. It is the social responsibility of the modeler to make any implicit goals and decisions of a model explicit to all people who are impacted by such decisions.
  • Model uses & responsibility: It is the ethical responsibility of the modeller to prevent harmful uses of models and outcomes. The challenges is when the modeller has little influence on this.

• Modelling as a joint responsibility

  • Creating a shared responsibility: Making the social and political life of models explicit can be a way to create a shared responsibility of models development and use in a project.
  • Individual & institutional: The responsibility resides not only with individual modellers through ethical consciousness and training, but also with the institutions where these models are mobilized and used, necessitating regulatory frameworks.
  • Limits in agency: Novice modellers are often ‘driven’ to one or another model, often influenced by senior modellers. It takes time to understand what these choices mean, and often there is little space to choose other options.

These key ideas show that there are many aspects to take into account, from addressing the questions in education and training, individual responsibility, limitations in agency, institutional responsibility, to the responsibility of funders and users. These topics especially require dialectical conversions between different disciplines and actors. Each modelling case differs, but we think that bringing up the question of how to deliberately engage with the social and political life of models will help to better connect modelling to the situation and questions it aims to address. The open conversations during this session served as a safe space for participants to engage with this question, providing a practical example of how such discussions can be done.

This blog is based on a session held at IHE Delft on 12 February, making the most of Tobias Krueger’s (IRI THESys) visit to the Netherlands for the PhD defence of Janneke Remmers, on ‘Modellers as influencers? Analysing practices and standards in hydrodynamic decision-support modelling’. The session builds on the work done by Rossella Alba, Rozemarijn ter Horst, Bich Tran and Jonatan Godinez Madrigal on ‘Modelling practices in water governance’ https://cat-water-models.github.io/, and collaborations between IHE Delft, IRI THESys, Wageningen University and the GEAU - Montpellier on the same topic.

• [1] https://doi.org/10.1016/j.envsoft.2012.01.011
• [2] https://doi.org/10.3389/frwa.2022.1038322
• [3] https://doi.org/10.1016/j.jhydrol.2018.11.046
• [4] https://doi.org/10.1029/2018WR022958
• [5] https://doi.org/10.5194/hess-28-4157-2024
• [6] https://www.water-alternatives.org/index.php/blog/models

Twitter  facebook  Email