Data needed for modelling
Modellers talk about the information they need to input in their models to represent and simulate as datasets and they need various datasets to make models do different things. Many datasets of different types have been generated by previous scientific research and is provided via websites set up by datacentres. NERC (Nature and Environment Research Council) website has links to cntres hodling data used in environmental research here.
In order to simulate local environmental events, such as flooding, with a computer model it is important to have information about the physical features of the landscape, as well as about rivers and weather. One NERC centre with datasets useful for flood modelling is BADC. There are several organisations that generate data that can be useful, for example the Ordnance Service (OS) provides a range of data products, both freely accessble and licensed.
Further a model needs data about normal (non-flood) river flows and levels in the locality. The Environment Agency produces hydrological summaries and other statistical information about rivers in England and Wales. CEH (Centre for Ecology and Hydrology) provides an even wider range of data about hydrological and ecological processes in UK rivers.
When the model is set up to represent a specific locality in normal circumstances and the modeller had checked that it works as it should, the group can run scenarios to explore the issues of interest. Data for such scenarios can be synthetically generated by, for example, climate models. Scenarios can also be historical, using data from recorded past events such as floods or droughts. The IPCC (Intergovernmental Panel on Climate Change) has a datasite that people anywhere can access.
Scenarios can also be hypothetical in the sense of ‘what will the model output be if we increase water demand with an amount equivalent to 5000 households?’ Such specific questions become meaningful when they are based on in-depth knowledge about the locality modelled. This is where ECGs come in, making modellers step outside the confines of datasets recorded by scientific instruments, to also make use of locally collected information. In ECGs local information is brought to the group, discussed and compared with scientific datasets, which generates local knowledge.