Palm oil is considered one of the healthiest and most affordable vegetable oils. There’s a catch, though: its production comes at the expense of valuable ecosystems. Eco-labels are supposed to help make palm oil more sustainable. But do they deliver what they promise? Our visual investigation finds: only partially. Satellite images and data from Indonesia show that even on certified plantations, illegal clearing techniques and deforestation occur time and again.
The online article presents the investigation using maps, graphics, and text. It is accompanied by a short video. The story also appeared in print and on social media.
At the beginning of 2021, palm oil became a widely debated topic in Switzerland. The reason for this was the upcoming vote on the free trade agreement with Indonesia. Part of the agreement: Only palm oil with specific sustainability labels should benefit from lower import prices. The vote was approved on March 7.
The Neue Zürcher Zeitung (NZZ) was the only media organization that fact-checked the promises made by the sustainability labels in such a detailed analysis. In doing so, we made an essential contribution to the transparency of promises from democratically elected politicians.
The story was widely read and well-received by our subscribers. Many of them spent an above-average time on the article online. On social media, the story is still shared on a weekly basis in activist circles to point out the greenwashing issues of the palm oil industry. Internally at Neue Zürcher Zeitung, the article is considered a showcase for using satellite imagery as visual evidence and for investigative research with public data.
Our story focuses on one of the largest eco-labels for palm oil, the Roundtable on Sustainable Palm Oil (RPSO). We examine three of their most important rules:
1. Fires are not allowed to be used to clear land.
2. The clearing of virgin rainforest is prohibited.
3. New plantations may not be planted on peat.
Examining various publicly available data sources, we gained new insights into how their members adhere to these rules. First, we used data from Global Forest Watch to grasp the extent of palm oil plantations and the accompanying destruction of the tropical forest. We then overlaid fire hotspots measured by NASA with concession boundaries to reveal where fires occur on certified palm oil plantations. Next, we searched and analyzed publicly available infrared imagery from satellites to track where these fires start. Finally, we demonstrate how valuable tropical forests and peatlands are cleared for plantations using land cover data and true-color satellite imagery.
Documents from the palm oil industry also served as important sources. These include various reports from palm oil companies available to the NZZ and the rulebook of RSPO. Using the visual analysis process described above, we could either confirm or refute these documents.
The article presents the entire visual investigation in graphics, maps, and texts. Where high temporal and spatial resolution is available, we use map animations to show developments over time. Satellite images are annotated and presented in chronological order. Readers can also hide and show different data layers on top of the satellite images to view the evidence at their own pace. Finally, the text embeds the visual elements into the story, explains the complex world of sustainability labels to our readers, and guides them through the often opaque data.
What was the hardest part of this project?
The challenges of this story lay in the research and in presenting the complex data to our readers.
Most of the data and documents were public. But often, they were not easy to find and difficult to understand. Many of the documents were only found using Google search operators; other confidential documents were obtained through NGOs. The data on palm oil concessions was also incomplete and poorly documented. Some of RSPO’s data had to be downloaded from a hidden section of their website.
Satellite imagery posed further challenges: Dense cloud cover often hangs over Indonesia, making it challenging to find suitable images. In addition, we had to learn how to read the images with the help of experts, for example, to recognize typical traces of peat fire.
Once we had the data, we had to determine which of the hundreds of RSPO-certified palm oil plantations in Indonesia were breaking sustainability rules. To do this, we had to fully understand the RSPO rulebook, which proved complicated: the rulebook uses a very technical language, is full of sub-clauses, and keeps changing over the years. Many of the palm oil companies operate in the legal gray area. So it was also necessary to distinguish whether clearing and fires were actually breaches of the rules or whether they were environmentally harmful but nonetheless legal practices.
The final challenge was to convey our research to the reader: We processed satellite images so that events shown in them could be recognized even by laypersons. We visualized small-scale geographic developments in such a way that they were readable on desktop as well as on mobile screens. And in the text, we tell the story of the complex RSPO-rules and the opaque data in a compelling yet straightforward manner.
What can others learn from this project?
Consumerism in countries in the Western hemisphere often comes with an environmental and human cost for countries of the Global South. Political decisions on responsible consumerism often involve some form of greenwashing. As journalists in Switzerland, we can contribute to a more sustainable production cycle by fact-checking promises made by politicians and sustainability labels. We hope that our project can serve as an example for this kind of work. Furthermore, we hope that our article inspires journalists to make use not only of open, structured data, but of the wide variety of open-source material, such as satellite images, satellite data, and documents.