Emissions of particulate matter (PM), reactive gases and greenhouse gases (GHGs) from industry, transport and domestic activities, degrade air quality in cities. In New Zealand, it is primarily emissions of PM from burning wood or fossil fuels that are of concern. PM remains suspended in the air where it can be inhaled, increasing the chances of developing heart-related illnesses and lung-related diseases.
In addition to PM, emissions of gases such as carbon monoxide (CO), nitrogen oxides (NO_x), and volatile organic compounds (VOCs), contribute to poor air quality. Furthermore, CO and VOCs react with NO_x to form tropospheric ozone, which is a GHG and an air pollutant. While most New Zealand towns and cities fortunately do not experience photochemical smog, for many megacities internationally, smog is a serious health hazard urgently requiring remediation.

Actions to mitigate PM and gaseous emissions require information on emissions sources, which have traditionally been identified using bottom-up accounting exercises. Through this project we are developing a new approach to inferring PM emission maps at high spatial and temporal resolution that capitalises on surface observations and a mesoscale atmospheric model enhanced to simulate aerosol microphysics. The system is called MAPM (Mapping Air Pollution eMissions).

Research Aims

There are a range of air pollution computer models that can take a prescribed emissions map for a city and simulate what the pollution levels around the city would be as a result of those emissions; the mesoscale model at the core of MAPM transforms a prescribed time varying two-dimensional surface PM emissions field into a time varying three-dimensional atmospheric PM concentration distribution. The model requires a description of the meteorology - primarily the winds that transport PM from their emissions sources to where they are measured.

The model also requires:
(i) a description of atmospheric temperature to capture the effects of inversion layers that trap PM close to the ground,
(ii) humidity, which affects aerosol microphysics, and
(iii) turbulence, which drives mixing of air masses.

But this model, on its own, doesn’t help us in the quest to identify where the emissions originated from. As well as measuring the pollution levels around a city (the output from the model), we want to figure out the emissions maps over the last 24 hours that would have produced the measured pollution levels (the input required for the model).

While knowing the level of pollution is useful, knowing where that pollution came from is far more valuable since city officials can then act to close down, or mitigate, those sources. We use an approach called inverse modelling that effectively ‘runs the model backwards’ so that it takes atmospheric measurements of pollution as input to infer what the pollution map must have been. The inverse modelling also generates uncertainties on those maps. While inverse modelling has been applied to similar problems elsewhere, it had not been applied to infer city-scale PM emissions fields from measured PM concentrations, operationally, to meet direct policy needs. Therefore, the design thinking for MAPM was conducted cognisant of stakeholder needs.

This project had some tough challenges. We needed to figure out what combination of measurements of atmospheric pollution levels leads to the best quality inferred emissions maps, where and what meteorological measurements around a city are required to constrain the model, and what model formulations will lead to the highest quality emissions maps. Small uncertainties in transport pathways of air parcels from their sources to where they are measured can result in large uncertainties in the inferred pollution emissions fields. Figuring out how to mitigate those so-called non-linear effects, was a huge challenge. The MAPM design criteria was tested through a field campaign in Christchurch in 2019.

The goal over the 2-year project was to develop the intellectual property (IP) that underpins the MAPM method and, once done, vest that IP in a new company that will then commercialise the capability as a service to city officials in offshore countries. Our project partners include the National Institute of Water and Atmospheric Research (NIWA), Environment Canterbury, University of Canterbury, University of Otago, Sigma Space, and Karlsruher Institute of Technology (KIT). The same technique can, and is, being applied to infer emissions of GHGs and gases participating in photochemical smog; combined retrieval of PM and these other trace gases has co-benefits. Therefore, while the primary focus of this project was on PM emissions maps retrievals, a stretch goal was to incorporate retrievals of CO, NO_x, and VOCs, and, through additional future funding, to extend the capability to include GHGs.

Our hope is that access to pollution source maps by city officials will reduce sources of pollution with subsequent improvements in urban air quality and improvements in human health. Developing and testing such a capability in New Zealand, and then deploying it as service also offshore, should provide a source of offshore revenue – a weightless export for New Zealand.