Development of a screening tool for quick environmental assessment of mobility scenarios

Abstract

Large mobility projects, for example, construction of new highways, have attracted an increased interest from the public at large. Pressure groups and environmental activists often propose their own alternatives for the planned projects. The vast amount of scenarios that are proposed in this way lead to an enormous increase of work related to the environmental assessment. Therefore, a new screening tool is proposed that can serve as a first phase in the environmental assessment procedure for air pollution. The tool enables users to quickly estimate the impact of a scenario on local air quality. Only scenarios which show promising results in this screening tool are then to be discussed in the normal environmental assessment procedure which is much more time-consuming. There are several requirements in order to have a successful approach: the screening tool needs to be fast, it needs to provide results which are close to the detailed environmental assessment procedure and it needs to enable calculations for the major pollutants (NO2, PM10, EC/BC, …) including fast ozone chemistry. Such a screening tool is presented here. It determines annual average concentrations of the major pollutants (NO2, PM10, PM2.5, EC, C6H6) in a short calculation time of only some 10 min on 1 CPU, compared to calculations of 2 days on 24 CPUs for the detailed assessment procedure. The speed-up is obtained by using lookup tables of pre-simulated situations, which are then combined into the large-scale scenario. For instance, the effect of a 100 m road segment from north to south with a unit emission strength is calculated beforehand. If such a segment occurs in the input, this result is then used with the emissions scaled according to need. The tool does not calculate absolute concentrations (therefore, a standard model is still used) but only differences between scenarios. For some scenarios both the screening tool and the regular assessment procedure have been applied and results are compared. Comparing for both methods the screening tool with the full model yields small biases (−0.0022–0.0075 µg/m3), a small RMSE (0.02–0.21 µg/m3), a high R2 of (0.75–0.87) and a slope of the regression curve close to 1 (1.01–1.18), showing that for screening purposes the tool is well capable of making the cut between good and bad scenarios.