How do we articulate and communicate whole system assessments and inherently uncertain predictions of the causes and effects of environmental change?
How do we reconcile conflicting interests with common necessity and purpose?
How do we encourage interdisciplinary working, stakeholder engagement and knowledge exchange?
Engagement and participation are difficult because of the lack of a common format, framework, language and story which simultaneously capture both the scientific knowledge (and ignorance) of these complex socio-environmental systems with practitioner and stakeholder perceptions of and expertise with them. Communication between interested parties and those with the techical knowledge - which is clearly necessary to define and pursue common visions - is difficult, and often avoided or short-circuited.
Sophisticated and complex computer models have been built of these systems, but have had limited impact in answering these questions. Apart from the technical difficulties with these models, there are simply too many different interests and understandings about these systems to reflect all, or even a representative set of opinions and perceptions. In short, those (few) who believe the model system are happy with the simplified representation, but hardly need the model. On the other hand, those who do not believe the model, for whatever reason, think that the model is irrelevant, however sophisticated. Nevertheless, the modelers themselves learn a lot about the environmental systems they model.
It ought to be possible to communicate our separate and different understandings about the ways in which these systems work without getting bogged down in technical detail or buried in sophisticated computer models. We believe we can identify the key systematic relationships between, for instance, land use, landscape appearance and environmental effects, and also identify the major differences of judgment and knowledge about the ways in which these key relationships work - what they mean for the management of the system. We will explore this approach with a number of stakeholders and practicing landscape managers, focusing on the Northumberland National Park in the first instance.
We expect to develop a set of mechanisms and procedures (a 'scope') to do this, with the primary purpose of helping articulate and communicate different perceptions and understandings of the major relationships and issues. We will illustrate the use of the 'scope' to systematically and coherently identify and communicate future options and scenarios.
We suggest that a Bayesian Belief Network (BBN) offers a real possibility for developing an intelligent framework for communicating and reflecting uncertainty and disputed knowledge about systematic relationships and potential outcomes, and about differences of judgement about underlying relationships and behaviours.
BBNs offer consistent semantics and mechanisms for representing uncertainty and an intuitive graphical representation of the interactions between various causes and effects. They are a very effective method of representing uncertain complexities that are believed to be essentially systematic, and of engaging differing perceptions and beliefs about such systems.
We will test this proposition by constructing a stylised representation of an upland region (the Northumberland National Park) as a BBN.
The first two research questions are:
We will then calibrate the resulting BBN system to more accurately represent a particular case, such as the Peak District, already the subject of a major RELU project. With this calibrated BBN of a known case, we should be able to replicate the major findings of this previous research, without substantial reference to additional input or interpretation from the specific expertise employed in the original project. If so, then this represents a substantial test of the methodological approach.
Subsequently, we will augment the BBN representation with the additional insights and beliefs of the actual stakeholders in the original project, to identify the critical gaps in the BBN representation, hence refining and tuning the model image to reflect the specific understandings and beliefs of the real world participants about their specific system. Sensitivity analysis of this refined model will then reveal the critical areas of uncertainty, mis-communication or conflicting beliefs, and indicate priorities for both future research and for communication and negotiation, as well as demonstrating proof of concept.