Creating a Heat Map for Fossil Fuel Operation Emissions

In this blogpost I'll detail the motivation and process behind creating the above heatmap.

The map you see above was created using climateTrace data. Please check out this incredible initiative at www.climatetrace.com. I created this map because recently I have been working on adding different layers to the offsetmap. Up until now, I have added a variety of nature-related layers, so this was my first attempt at exploring industrial emissions data. The data you see above is a combination of three data sets, coal mining emissions, oil and gas production and transport emissions, and oil and gas refining emissions. To be clear, this data is related to the emissions embedded in the upstream of the hydrocarbon production process. This data does not account for energy production, transport, or manufacturing (all of which would significantly change this map). This was just a starting point to look at some of the total industrial emission process. 

I chose to use 2021 because it is the last year with data available for all three data sets. The image you see is actually just a single frame from an animation showing the dynamic from 2015 to 2021. I chose to use a static representation of the data because climateTrace does not have accessible data for coal mining related emissions prior to 2021. 

Working with this data-set was relatively straightforward. It was in an csv format, so it was easy to filter and process the data in python. What you see is actually the co2e_20yr emission data for each of those three categories. Taken together, the processes involved in extracting, transporting, and refining oil and gas all over the world account for nearly one in six tons of human-made greenhouse gases (GHGs) when accounting using a 20-year global warming potential (GWP). In 2021, these sectors totaled more than 11 billion tons of CO2e, more than the annual emissions of the United States. Climate TRACE’s latest global inventory of greenhouse gas emissions brings new data, covering every country and all emitting sectors of our global economy, plus over 70,000 of the highest-emitting facilities around the world.[1] 

The data set contained the point sources, their coordinates, and a time-series of related emissions - all of which were used to generate the map you see above. Geopandas was used to generate the map outlines you see, scipy was used for the smoothing (gaussian filter) of the heat map data, and matplotlib was used for the heatmap itself. None of the data was excluded in the map above. 

In the future I hope to include the other industrial emissions included in the climateTrace database. Part of the challenge will be the sheer size of some of those data sets. To put it into perspective, this relatively small data set (54000 rows, 1.5mb) is dwarfed by some of the large data sets (i.e., Transport - 18GB). Thankfully, the climateTrace datasets are well organized and homogenized across all data reports, so the additional data sets should not require too much pre-processing. 

Check out the super cool initiative at their website www.climatetrace.org