Tracking the effects of China's autumn-winter air pollution action plan

2020-02-27 210浏览

  • 1.Is China winning the battle on air pollution? Tracking the effects of the autumn-winter air action plan
  • 2.Big picture – huge improvement in Beijing region in five years Winter 2012-13
  • 3.Big picture – huge improvement in Beijing region in five years Winter 2017-18
  • 4.Even as average levels fell, winter pollution peaks seemed to be getting worseSource:Greenpeace analysis of Ministry of Environmental Protection hourly air quality data
  • 5.Response:Winter action plan • Output restrictions on iron & steel and aluminum, closure of cement and coking plants – Up to 25% of steelmaking and 10% of cement capacity affected • New focusarea:2+26 cities • New environmental protection bureau for Jing-jin-ji • Massively ramped-up central governmentenforcement:tens of thousands of inspections • Replace coal heating&cooking in 5.5 million households • Ban on construction and restrictions on heavy trucks •Target:15% reduction in October-March average PM2.5, year-on-year
  • 6.Winter action plan created a new focus area (“2+26”), including most of the emission hotspots in Beijing region 2+26 cities
  • 7.Results:dramatic reduction in pollution levels, far exceeding targets
  • 8.YRD=Yangtze Delta PRD=Pearl River Delta Jingjinji=Beijing-Tianjin-Hebei
  • 9.Comparing 2017-18 winter with previous one • In 28cities:Major reduction in PM2.5, SO2 and CO levels, less impressive in NO2 levels • Increase in PM2.5 levels in Yangtze Delta, Jiangsu and much of the south. • Rest of the country saw negative progress on NO2. Part of this is related to relocation of industrial production and power demand away from 28 cities. • Lack of progress on NO2 is problematic for both PM2.5 (after falls in SO2 emissions NOx is now the main precursor to PM2.5 formation) and ozone (which has been getting worse)
  • 10.What was the role of policy vs. natural variation?
  • 11.Beijing monthly mean wind speeds (m/s) month • Beijing wind conditions were much more favorable than year before, especially in NovJan. • Simply comparing average pollution levels for this winter and past winter will give little useful information about the effectiveness of the action plan. year
  • 12.Wind speeds have a strong effect on average PM2.5 levels. Many other weather variables have an even stronger and more complicated impact.
  • 13.
  • 14.• While it’s clear that weather is the dominant factor in the variation of pollution levels on daily and monthly basis, the relationship between weather variables and PM2.5 levels is complex and non-linear. • Gradient boosting algorithms are generally the tool of choice for these kinds of problems. Comparison with linear and generalized additive models confirms the benefits of using the more complex model.
  • 15.R2:91.3%
  • 16.R2 for hourly values -trainingdata:'>data: