We addressed Sustainable Development Goals (SDGs) and COVID-19 within the NASA Space Apps challenge by focussing on two SDGs: Reduced Inequality (#10) and Life on land (#15).
This project tackles two SDGs through combining Earth observation data with health data to create an interactive online tool, plotting the two together. We hope it will Reduce Inequality by using satellite data to shine a light on issues in low- and middle-income countries with least financial power to tackle COVID-19, and help Life on land, by serving as a tool that can be used by policy makers and NGOs to direct efforts protect both human and natural life.
The problem
Conservation International recently reported on their website that an increase in deforestation rates has been seen in countries such as Colombia and Brazil (Conservation International, 2020). Our first aim is to use satellite imagery to test the claims of Conservation International, and explore possible causations of forest change in Brazil and Colombia since the COVID-19 crisis began to affect those regions.
Possible explanations for why deforestation might increase while governments are forced to limit activity, if true, include the following:
None of these explanations are specifc to Brazil or Colombia, and so it is possible these trends if true, may be occurring in other low- and middle- income countries. This leads to the second aim of COVIDeforestAlert: explore whether forest changes have occurred in other regions beyond Brazil and Colombia that could be caused by COVID-19.While there has reductions in air and noise pollution have occurred as a result of COVID-19, these effects can be temporary and are all too easily reversed if transport and industrial activity does not change. Deforestation can have long-lasting impacts because carbon dioxide release from burning of wood has much slower but more persistent effects on climate, and it can take much longer timescales to restore soil carbon within regions affected. We will not even mention the complex issues surrounding inequality, politics, and ownership that might be an underlying driver of environmental damage in low and middle-income countries.
After identifying the problem and our approach to it, we emailed people with experience of forests in low- and middle-income countries, and received a reply from Mary Allen, of Practical Action, who advised on factors to take in account. We then looked for datasets of Earth observation data from NASA, and health data from John Hopkins university and EU CDIC. We opted for the EU dataset as it was less focussed on the US, and more on the regions in consideration.
For Earth observation data on deforestation we found that Global Forest Watch was ideal for our purposes, as it gives access to GLAD deforestation alerts updated daily. GLAD deforestation data provides high enough resolution to identify even relatively small clearings (30m x 30m), and has had extensive testing and coverage that would not be feasible within 48 hours. Establishing access to the API itself took a considerable proportion of the time, as the documentation was sparse.
We then wrote a Python script that downloads data on deforestation automatically from the Global Forest Watch API, as well as data on COVID cases from the European Centre for Disease Control. We then combined this into a fully-featured dashboard using Jupyter notebook, plotting the two variables against each other through a dynamic interactive graph using bokeh.
To tie it together, we created a website with documentation and a short story that motivates the use. It also serves as a link to the live code and GitHub repository.
Features of the live code we made:
+ Open-source, easy to read and can be edited
+ Hosted on Google Colaboratory and GitHub, so can be edited live without knowing or downloading Python
+ Attractive and responsive Interactive graphs using bokeh
+ Users can access data about almost any country within 30 degrees of the equator
Future work
We identified data for tho following indicators,
1. Proportion of GDP from remittances (those working abroad), an important source of security during hard times https://data.worldbank.org/indicator/BX.TRF.PWKR.DT.GD.ZS?locations=CO
2. Proportion of GDP from tourism. An activity strongly affected by the necessary travel restrictions, and which has had strong impacts on jobs and livelihoods for those reliant on it https://data.oecd.org/industry/tourism-gdp.htm
3. Unemployment rate. Estimates of unemployment rates for many countries in March https://tradingeconomics.com/country-list/unemployment-rate?continent=america
Unfortunately the tight scope of the competetion and limited resources meant we were not able to add these data to the graph, or explore correlations. However, this is an area that would be good to pursue
NASA ‘Landsat 8 « Landsat Science’. Accessed 31 May 2020. https://landsat.gsfc.nasa.gov/landsat-8/.
Hansen, M. C., P. V. Potapov, R. Moore, M. Hancher, S. A. Turubanova, A. Tyukavina, D. Thau, et al. ‘High-Resolution Global Maps of 21st-Century Forest Cover Change’. Science 342, no. 6160 (15 November 2013): 850–53. https://doi.org/10.1126/science.1244693.
Kiley Price. ‘Poaching, Deforestation Reportedly on the Rise since COVID-19 Lockdowns’, 30 April 2020. https://www.conservation.org/blog/poaching-deforestation-reportedly-on-the-rise-since-covid-19-lockdowns.
Watts, Jonathan. ‘Jair Bolsonaro Claims NGOs behind Amazon Forest Fire Surge – but Provides No Evidence’. The Guardian, 21 August 2019, sec. World news. https://www.theguardian.com/world/2019/aug/21/jair-bolsonaro-accuses-ngos-setting-fire-amazon-rainforest.