The data for this project come from several sources, which were all reported in the manuscript by Kolker and Weathers.
The primary dataset of pollution maps is the Emissions Database for Global Atmospheric Research (EDGAR). This source is essentially a "big data" compilation of numerous data sources that calculates emissions for the entire globe for each year from 1970-2021 (for greenhouse gas) and 1970-2018 (for air pollutants).
EDGAR’s emission calculation method combines "emissions factors," with "activity data," - a common method used. An "emissions factor" is the amount of emissions released per unit of some action, and "activity data," is the total number of units of action. The two are multiplied to give the level of emissions from a given source.
To understand this method better, think about calculating the amount of carbon dioxide released from driving your car. We do not have working CO2 sensors on every car in the country, but we can calculate the amount of emissions using numbers we know well. We know how many molecules of CO2 are released by burning a gallon of gasoline since we know the chemical composition of gasoline. You know how many gallons of gas your car uses, because you see that number each time you fill up the tank. Multiply the number of gallons of gas you used by the number of CO2 molecules released by burning a gallon of gasoline, and you get the number of molecules of CO2 released by driving your car.
This method can be extrapolated from one car to the entire country using large data sets. It turns out that we know the number of gallons of gasoline the country uses relatively well, since the US Department of Energy tracks that. If we multiply the number of gallons of gasoline used each year by the number of molecules of CO2 released from burning each gallon of gasoline, we can get the total number of molecules of gasoline released by people driving cars.
This method is like the ways that other organizations calculate greenhouse gas emissions. The US Environmental Protection Agency's State Inventory Tool uses activity data and emissions factors to calculate greenhouse gas inventories for states. Those personal greenhouse gas calculators that you can find online also use a similar method. (For example, see this site https://www.epa.gov/energy/measure-impact-your-energy-use-environment.)
One big advantage of the EDGAR dataset is that it distributes these data spatially using
maps and demographic information that include where people live, where major roads are, and were major facilities are located. The output is a dataset for the entire world, with pixels that are 0.1o x 0.1o, which is about 6.5 miles long by 6.5 miles wide.
We determined the emissions for each parish/county using a digital geographic processing tool that added up all the emissions from each pixel lying within a county. If the pixel spanned multiple counties, the emission data was pro-rated proportionately. We did this for each county in Mississippi and each parish in Louisiana for each years’ data available.
While the EDGAR dataset is very good, like all datasets, it is not perfect. For example, some emissions factors and activity data are well known, like those for fuel consumption, others are less well known. For example, the amount of emissions from pipelines is based on the average size of a leak, and the number of leaks per mile for a pipeline. If pipelines in one state are more prone to leakage than another state, these numbers could vary with those differences.
We hope this project helps people talk about where our greenhouse gas and air pollution come from, so that we can come up with more effective strategies that protect people and the planet.
Acknowledgements: This project was led by Dr. Alex Kolker of the Louisiana Universities Marine Consortium, with important contributions from Dallon Weathers of Delta GeoMarine. The website was developed by Deep Fried. Support for this project was provided by the Rockefeller Family Foundation.