In the previous tutorial, we learnt that one way to perform a source apportionment study in a city is the bottom-up Emissions modeling way. In this tutorial, we will learn how to build an emission inventory for a city.
Across this course we have dealt with data on pollutant concentrations, whose units are in µg/m3 or ppm. Concentration is thus the amount of pollutant present in unit volume of ambient air that we are breathing.
On the other hand, Emissions is the amount of pollutant directly emitted by a source. These sources could be vehicles, waste burning, biomass burning, etc., that we saw in the previous tutorial. The units of Emissions are thus in kgs or tonnes per day.
There is a simple equation that helps us calculate emissions from any source.
\[ Emissions = Activity * Emission Factor \]
For each source, \(Activity\) could be defined differently. 1. For Transport, it could be number of litres of fuel burnt, or number of kilometres travelled by vehicles.
\(Emission Factor\) is the amount of emissions released per unit activity. These factors vary for each source and activity. They are determined through lab-experiments and field studies. We can find emission factors of required pollutant from literature.
Each pollutant of interest would have an \(Emission Factor\) for each \(Activity\)
A city has 20,000 vehicles. Average speed of vehicles is 20 kmph. The Carbon Monoxide CO) Emission Factor is 2.7 g/km. What is the Emission Rate of CO in this city?
Here, we will define \(Activity\) as Kilometres travelled per hour. When we multiply that \(Activity\) with the given \(Emission Factor\), we would obtain Emission Rate in grams per hours (g/hr).
num_vehicles = 20000
avg_speed = 20 #kmph
activity = num_vehicles * avg_speed #km per hour
print(f"Vehicular activity: Vehicles in the city travelled {activity} km per hour")
co_ef = 2.7 #g/km
emission_rate = activity * co_ef #g per hour
print(f"Emission Rate: Vehicles in the city emit {emission_rate/1000} kilograms of CO per hour")Vehicular activity: Vehicles in the city travelled 400000 km per hour
Emission Rate: Vehicles in the city emit 1080.0 kilograms of CO per hourThus we have calculated the emission rate of Carbon Monoxide (CO) from all vehicles in the city. If we repeat this exercise for all other sources, we would build the CO Emission Inventory for the city and also identify all the sources of CO.
A few countries developed such Emission Inventories for various pollutants. The complexity of emission inventories depends upon the granularity of the emissions data. In the above example, we calculated CO emissions for the entire city. But with better datasets, we could calculate it for each 1 sq.km grid in the city. This is what the UK’s National Atmospheric Emissions Inventory (NAEI) achieved.
We can also download this gridded data from here: UK NAEI Gridded Emissions data download. I downloaded CO Emissions data for 2023.
import geopandas as gpd
co_point_sources = gpd.read_file('data/naei23_co_2023/point_sources_layer/point_sources_CO_2023.shp')
co_point_sources.sample(5)| Year | PollID | Poll_Name | PlantID | Site | Easting | Northing | Operator | SectorID | Sector | Emission | Unit | Country | Datatype | geometry | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1419 | 2023 | 4 | Carbon Monoxide | 11059 | Ferrybridge | 447940.0 | 425500.0 | Lafarge Plasterboard Ltd | 4 | Other mineral industries | 11.346540 | Tonnes | England | M | POINT (447940 425500) |
| 1272 | 2023 | 4 | Carbon Monoxide | 9647 | Grangemouth | 291649.0 | 681450.0 | Dalkia Plc | 10 | Chemical industry | 8.373685 | Tonnes | Scotland | M | POINT (291649 681450) |
| 2605 | 2023 | 4 | Carbon Monoxide | 43248 | Equus UK Topco Limited | 509265.0 | 431660.0 | Equus UK Topco Limited | 10 | Chemical industry | 3.091166 | Tonnes | England | M | POINT (509265 431660) |
| 1303 | 2023 | 4 | Carbon Monoxide | 9751 | East Kilbride | 261059.0 | 654957.0 | Coca-Cola Enterprises Ltd | 5 | Food, drink & tobacco industry | 7.703571 | Tonnes | Scotland | M | POINT (261059 654957) |
| 1562 | 2023 | 4 | Carbon Monoxide | 11918 | Hill of Tramaud Generating Station | 395290.0 | 813494.0 | Sita Holding UK Ltd | 22 | Construction | 6.847015 | Tonnes | Scotland | O | POINT (395290 813494) |
co_point_sources.Sector.unique()array(['Mechanical engineering', 'Non-ferrous metal industries',
'Waste collection, treatment & disposal', 'Chemical industry',
'Other industries', 'Electrical engineering',
'Paper, printing & publishing industries',
'Processing & distribution of petroleum products',
'Food, drink & tobacco industry', 'Vehicles', 'Miscellaneous',
'Water & sewerage', 'Lime', 'Cement', 'Other mineral industries',
'Textiles, clothing, leather & footwear', 'Commercial',
'Processing & distribution of natural gas',
'Oil & gas exploration and production', 'Iron & steel industries',
'Other fuel production', 'Major power producers',
'Minor power producers', 'Public administration',
'Agriculture, forestry & fishing', 'Construction'], dtype=object)We can see that it is a detailed information on how many tonnes of CO emitted by each point source in 2023. We can plot these points on a map with the size of the point representing the amount of emissions.
import matplotlib.pyplot as plt
uk = gpd.read_file('data/uk_regions.geojson')
fig, ax = plt.subplots(figsize=(8,8))
uk.plot(ax=ax, color="white", edgecolor="black", linewidth=0.5)
co_point_sources.to_crs(uk.crs).plot(ax=ax, markersize=co_point_sources['Emission']/50, color='red')
plt.title('Carbon Monoxide Point Sources in UK | 2023')
plt.show()
This point sources emission information is then distributed to each 1 sq.km grid in UK using specialised models. This is called Gridded Emissions Data and they are made available in TIFF format.
import rasterio
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.colors as mcolors
from rasterio.plot import show
import geopandas as gpd
# Load TIFF
tif_path = "./data/naei23_co_2023/GeoTIFF_layers/totalco23_2023.tif"
with rasterio.open(tif_path) as src:
co = src.read(1) # first band
co = np.where(co == src.nodata, np.nan, co) # mask nodata
extent = [src.bounds.left, src.bounds.right, src.bounds.bottom, src.bounds.top]
crs = src.crs
# Load admin boundaries — reproject to match raster CRS
uk = gpd.read_file('data/uk_regions.geojson').to_crs(crs)
# Plot
fig, ax = plt.subplots(figsize=(7, 7))
im = ax.imshow(
co,
extent=extent,
cmap='YlOrRd',
origin='upper',
vmin=np.nanpercentile(co, 2), # clip outliers
vmax=np.nanpercentile(co, 98),
)
# Admin boundary overlay
uk.boundary.plot(ax=ax, linewidth=0.5, color='black')
plt.colorbar(im, ax=ax, label='CO (tonnes)')
ax.set_title('CO Gridded Emissions - UK - 2023')
plt.axis('off')
plt.tight_layout()
#plt.savefig('co_griddedemissions.png', dpi=150, bbox_inches='tight')
plt.show()
This visualisation is also part of our Air Quality Visuals Course: Plotting Satellite images
The UK’s NAEI created a portal to explore such gridded emissions of each pollutant. You can check it out here: UK Emissions Interactive Map
In this tutorial, we learnt how to calculate amount of emissions of a pollutant by each source using the fundamental equation of \(Activity * Emission Factor\). We then explored the UK’s National Atmospheric Emissions Inventory (NAEI) and learnt about Gridded Emissions Datasets.
Other Emission Inventories: 1. National Emission Inventory of USA by US EPA
Emissions Database for Global Atmospheric Research (EDGAR) by European Commission: We can find global emission inventory here.