In the last couple of years deforestation caused by growing cocoa has leapt into the headlines. Efforts are under way to halt deforestation in protected areas, but how do you monitor vast tracts of land in remote areas in Côte d’Ivoire and Ghana?
The answer is, from the air, or rather, from space, using satellites, and artificial intelligence (AI), to provide accurate, customised data about where deforestation is taking place and provide farmers with equally accurate, customised information that they can use to grow more cocoa, but without encroaching on the forest.
In February 2019, Rainforest Alliance and partners issued a call to the remote-sensing and AI community to develop technology that can provide just the kind of data needed to preserve forests in West Africa and help cocoa farmers overcome the challenges they face.
Rainforest Alliance, Grameen Foundation, Touton, Satelligence, University of Ghana, and Waterwatch Projects set up the SAT4Farming programme to provide farmers with information and services to improve productivity and sustainability. The aim is to use digital technology and satellite imagery to create customised Farm Development Plans (FDPs) that can guide farmers and help them increase productivity. The aim is to increase their average productivity by up to 1,500kg per hectare – that would be a 300 per cent increase.
Together, Rainforest Alliance and its partners are calling on remote-sensing, hightech, and Earth-observation companies to tackle their ‘Ag-Tech Developer Challenge,’ a competition to build a remote-sensing data product that links to and adds value to the mobile-enabled FDPs. Two prizes will be awarded for the best ideas to be implemented, but all entries may be considered for future engagement with the Rainforest Alliance.
An FDP is a decision-making and planning tool for smallholder cocoa farmers and includes agricultural practices that they need to adopt as well as the financial investment needed to do so to increase their productivity. In addition, it can help field officers assist cocoa farmers through monitoring, training and financial planning. The key to the customised FDPs is that they provide recommendations that are specific to individual farmers and recommended interventions can be adjusted, based on the specific situation of a farmer and the household.
But the SAT4Farming project isn’t by any means the only way that satellite and remote sensing technology is being used to help farmers. Like most cocoa and chocolate companies, The Hershey Company is taking action to protect forests and restore forest cover in cocoa-growing regions in West Africa. One way that it is doing so is an action plan that is part of its commitments as a founding member of the Cocoa & Forests Initiative. Hershey has created specific goals, actions and targets for Côte d’Ivoire and Ghana, and plans to distribute hundreds of thousands of cocoa trees to farms there and use satellite mapping to closely monitor developments in its supply chain.
Using new data from the European Sentinel-1 and 2 satellites in combination with AI, it is already possible to accurately map cocoa growing areas and detect associated deforestation. Satelligence and partners Touton and SNV have been working on a project that saw them use improved resolution, scale and timeliness of map information based on satellite imagery to enable stakeholders to better assess deforestation risks and work with farmers to achieve viable livelihoods without encroaching into parks and reserves.
Until recently, it has been impossible to accurately and cost-effectively map cocoa production in entire countries at farm-level because cocoa trees show up as forest in conventional satellite images, but Satelligence developed a new approach using freely available Sentinel-1 and 2 satellite imagery and scientific concepts developed in Ghana in order to overcome this issue.
The technique developed by Satelligence (see elsewhere in this issue) combines the use of imagery at the 10m scale, advanced radar imagery that can capture the structure of the forest and of cocoa, and machine learning techniques that achieve higher accuracy than previously possible. The new technique has also overcome the problem of semi-persistent cloud and haze cover over West Africa. Leveraging the power of cloud computing, powered by the Google Compute Engine, it used more than 400 GB of Sentinel-1 and 2 data to generate a map of the entire cocoa landscape of Ghana at 10m detail.
The real excitement came when the company overlaid Touton’s ground-based farm boundary data on the satellite-derived cocoa land cover map. Although both datasets were generated independently, their alignment was spot-on. Having continuous cloud free observations from satellite-based radar is a ‘game-changer’ for early warning of deforestation and can help verify drivers of change in the forests of Côte d’Ivoire and Ghana. It can also help efforts to plan for rehabilitation of illegal cocoa farms in parks and reserves, providing local people with alternatives to cutting down trees to try to make a living income.