Application of uncertainty and sensitivity analysis to the air quality SHERPA modelling tool. (June 2018)

• Record Type: Journal Article
• Authors: Pisoni, E.; Albrecht, D.; Mara, T.A.; Rosati, R.; Tarantola, S.; Thunis, P.
• Is Part Of: Atmospheric environment. Volume 183(2018)
• Page Start: 84
• Page End: 93
• Publication Date: 2018-06
• Subjects: Uncertainty analysis -- Sensitivity analysis -- Air quality modelling -- Surrogate models -- Model quality assurance; Air -- Pollution -- Periodicals; Air -- Pollution -- Meteorological aspects -- Periodicals; 551.51
• DOI: 10.1016/j.atmosenv.2018.04.006
• Languages: English
• ISSNs: 1352-2310

Abstract

Abstract: Air quality has significantly improved in Europe over the past few decades. Nonetheless we still find high concentrations in measurements mainly in specific regions or cities. This dimensional shift, from EU-wide to hot-spot exceedances, calls for a novel approach to regional air quality management (to complement EU-wide existing policies). The SHERPA (Screening for High Emission Reduction Potentials on Air quality) modelling tool was developed in this context. It provides an additional tool to be used in support to regional/local decision makers responsible for the design of air quality plans. It is therefore important to evaluate the quality of the SHERPA model, and its behavior in the face of various kinds of uncertainty. Uncertainty and sensitivity analysis techniques can be used for this purpose. They both reveal the links between assumptions and forecasts, help in-model simplification and may highlight unexpected relationships between inputs and outputs. Thus, a policy steered SHERPA module - predicting air quality improvement linked to emission reduction scenarios - was evaluated by means of (1) uncertainty analysis (UA) to quantify uncertainty in the model output, and (2) by sensitivity analysis (SA) to identify the most influential input sources of this uncertainty. The results of this study provide relevant information about the key variables driving the SHERPA output uncertainty, and advise policy-makers and modellers where to place their efforts for an improved decision-making process. Highlights: SHERPA model can support decision makers in designing air quality plans. It is important to evaluate the quality of the SHERPA model with UA and SA. UA (Uncertainty analysis) can quantify uncertainty in the model output. SA (Sensitivity analysis) can reveal the most influential sources of this uncertainty. Key factor for uncertainty and sensitivity of SHERPA is the emission inventory.