Overview

All the slides shown on this page and other pages can also be found here:

Forest Foresight was created to prevent illegal deforestation, a significant challenge worldwide. It is a one-of-its-kind predictive early warning system that shows where deforestation takes place, before it takes place.

In the deforestation chain, interventions often happen too late as deforestation is only detected after the destruction is already underway, or due to long lead times to mobilize field teams. With predictive machine learning algorithms and near-real-time monitoring data, Forest Foresight aims to enable earlier interventions and prevent future deforestation before the destruction has set in.

Forest Foresight is more than a technological tool alone: it is a holistic approach from data to interventions, supported by technological tools (machine learning models and a user interface).

As a first step, the machine learning model predicts hot zones at risk of deforestation within the next 6 months, based on optical and radar remote sensing (satellite) data and other geospatial datasets. These predictions are clustered in groups of predictions that are close to each other.
After the machine learning model has prepared these prioritised alerts, the results are published in a user interface. This dashboard can be used to plan field investigations and keep track of follow-up activities, discussions and interventions for each alert.

The technical foundation to determine the deforestation predictions is key to this approach. Below infographic describes how the machine learning model determines these predictions. More information can be found in the Technical deep dive section.

Another essential aspect of this approach is the joint setup of an effective national governance mechanism for each region implementing Forest Foresight. WWF provides technical expertise and facilitates the exchange of learnings between the different regions implementing Forest Foresight. Regular monitoring of alerts and planning of field investigations and interventions, however, is driven by stakeholders in implementing regions. Typical stakeholders in implementing regions include national park agencies, various ministries governing a particular land use (mining, forestry, etc.), concession holders, communities, civil societies, etc.

Forest Foresight was first piloted in Kalimantan (Indonesia), where several field investigations were carried out based on our predictions. At the moment we are in different stages in different landscapes