UncertWeb is a research project funded by the European Commission that deals with the integration of resources, such as data and models, to construct complex models composed of chains of model and data components available as web services. When combining data of low quality, it is necessary to take into account uncertainties. This project develops mechanisms, standards, software tools and case studies for managing uncertainties in an interoperable model web context. Four demonstrators were conducted in different domains: biodiversity and climate change, land-use response to climatic and economic change, short-term uncertainty-enabled forecasts for local air quality, and Individual activity in the environment.
Firstly, one demonstrator for biodiversity and climate change is the e-Habitat model, which is being developed in the context of DOPA (Digital Observatory for Protected Areas), a biodiversity information system currently developed as a set of interoperable web services put in place by a number of international organizations active in the field of conservation. DOPA uses various datasets acquired from a wide range of biodiversity stakeholders worldwide and from remote sensing information. As the e- Habitat model did not deal with the various sources of uncertainty associated with the habitat data, an assessment was carried out.
Secondly, regarding land-use response to climatic and economic change, there are agent-based models of agricultural land-use which provide a tool for policy makers to investigate their impact on future land-use at local, regional, national and EU-wide scales. However, the inputs to land-use models are usually the uncertain products of other statistical or mathematical models (of varying resolution). The key drivers for land-use models are the climatic and economic conditions, but coupling land-use models to economic and climate models is difficult. Addressing this case was a challenge because complexity inflates computational costs and the feedbacks that exist between the economy and land-use are difficult to capture.
Thirdly, short-term uncertainty-enabled forecasts for local air quality are demonstrated by on-line and near-real time monitoring services for air quality, which are carried out by almost every country in Europe. Monitoring networks and on-line services are increasingly being complemented by modelling tools. Several air quality forecasting systems are running routinely in Europe. The MACC project represents the state of the art in regional air quality forecasting by bringing together these forecasts. In order to assess the uncertainty in those models, existing air quality and meteorological models have been applied to produce an ensemble forecast at the urban scale for the city of Oslo.
Fourthly, in order to demonstrate individual activity in the environment, Albatross is a model developed for travel patterns of individuals. It is uncertainty-enabled and allows quantifying errors due to model approximation, as well as propagating errors from uncertain inputs such as weather conditions. Assessments of traffic intensity and person density have been used to feed local air quality models.
These four case studies have been used to describe and quantify uncertainty following the methodology of the project and test the developed toolset.
UNCERTWEB, UNCERTWEB Website