Statistical inference and uncertainty quantification for complex process-based models using multiple data sets

Making responsible decisions about landscapes is facilitated by the use of complex models able to represent multiple competing demands on land use. Decisions about land use require that trade-offs between competing demands be identified, and their consequences through time be characterised. Models consisting of stochastic computer simulations are increasingly used to make realistic predictions about real world processes from socio-ecological systems involving land use, to the effects of climate change. Because these models attempt to simulate all relevant aspects of a real physical system, they may involve many parameters, some of which will be difficult to set correctly. Because the final objective of these models is to assess the possible consequences of management decisions, such as the placement of wind turbines, it is crucially important that the uncertainty introduced by calibrating parameter values be understood. 

In order to make informed decisions, one needs to be able to consider the effects of a number of complex interacting temporal and spatial processes (e.g. hydrological, ecological, agricultural, economic, climate). We wish to predict the effect of making decisions on these competing demands, considering all information that we have available. 

With the above challenges in mind, the objectives of this project are: 

1.            Develop methods to perform calibration and uncertainty quantification for process-based models of the complexity required for the landscape decision problem, extending previous work to cope with the challenging cases encountered in practice, including the cases of Big Data and where the models are computationally expensive. 

2.            Introduce methods for accurate uncertainty quantification in coupled models that are robust to mis-specified models, when the constituent parts are simulation models. 

3.            Make available the methods developed in the project, through the release of software packages in R or python. 

4.            Work with investigators in the landscape decision-making programme to apply the methods developed above to their models, and to other existing models. 

PI Richard Everitt http://richardgeveritt.com/ 
Link to the project page on the UKRI Gateway to Research https://gtr.ukri.org/projects?ref=NE%2FT00973X%2F1