Tree-equality?
Trees regulate extreme temperatures and manage flash flooding. They also absorb carbon dioxide from the atmosphere, helping to offset greenhouse gas emissions.
Given these benefits, it’s clear that green spaces have an implicit value in our warming world. In this article, we explore tree coverage inequality within cities and how this perpetuates existing imbalances between the rich and the poor.
We propose a KPI to assess the equality of tree coverage and the resultant health benefits in any city.
Our KPI:
Our KPI allows cities around the world to find and systematically evaluate regions with low tree coverage. This information can and should be used to drive socially equitable public investment in green spaces.
We have created an easy and generic interface for calculating and visualizing this KPI for any city. All you need is the name of a city and our software organizes the rest.
How it works
- Take a city name as input
- Determine city boundaries from satellite images using a custom ML model
- Collect thousands of ‘zoomed-in’ images within the city bounds and asses green coverage
- Combine results to form a Lorenz curve and calculate the Gini coefficient
- Visualize areas with the least tree coverage on a heat map
Examples:
The following notebook shows our application being applied to Sydney and Los Angeles. We use static images below to explain these results but please don't hesitate to check out the link for dynamic content.
As shown above, our heat maps clearly visualize areas with low green coverage and therefore guide effective equitable public investment.
The chart above shows the Lorentz curve for the two cities. Although seemingly complex, this maths can be distilled down to one number: the Gini coefficient. Often used for evaluating income inequality, a Gini coefficient of 0 indicates the equal distribution of resources whereas 1 indicates inequality.
Applied in this context, we can clearly see that Sydney has done a much better job at distributing their green areas equally throughout the city as opposed to Los Angeles. This serves as a useful tool for comparing cities and evaluating progress.
Importance:
Our work provides a systemic measurement that can be applied to any city. For many years the public has been told that their local governments are tackling climate change and inequality, but how can they know?
Our tool gives information to everyone and allows cities to be held accountable as we tackle two very large issues.
Limitations:
Although the above results are promising there are still areas of improvement:
- The model used to classify greenery needs to be improved (it can still be tricked by some particularly green water)
- Population density should be taken into account. In the included examples, multiple airports appear highlighted as vulnerable regions. Although they contain few green spaces, the lack of residents removes many of the public health risks.
Conclusion:
Please let us know what you think of our work so far. Engagement is our biggest motivator and will definitely push us to keep improving the product.
For an in-depth breakdown of V1 visit this notebook.
Stay healthy!
