EO-ALLert: Early-Warning to the Impacts of Alluvial Mining on Sensitive Areas Using Earth Observation

EO-ALLert “Early-Warning to the Impacts of Alluvial Mining on Sensitive Areas Using Earth Observation” aims to provide the means for development agencies and authorities to identify priority areas to be preserved from the impacts of alluvial small-scale mining. It considers the case of alluvial gold mining in Colombia and works on integrating information from satellite data and stakeholder priorities. This results in scientifically and geo-spatially based understanding of mining activities and their impacts.

Why small-scale mining?

Small-scale and artisanal mining exist in many parts of the world and produce the majority of worldwide sapphire and about 20% of gold and diamond. While this type of mining provides livelihood to many families in rural areas, it impacts the landscape, degrades the land, and can contaminate the food chain with heavy metals. Thus, the applications of this work are not limited to Colombia. It contributes to the development of new approaches to monitor and assess small-scale mining worldwide. Furthermore, these extractive activities have major implications on several of the United Nations Agenda 2030 for Sustainable Development Goals (SDGs) including the following ones:

Alluvial small-scale gold mining and the benefits of satellite data

There are several types of small-scale mining in Colombia including alluvial mining and underground vein mining. EO-ALLert considers alluvial mining of gold particles that is taking place on land, typically along riverbanks. This extractive activity of gold utilizes heavy excavation equipment and leaves a large footprint on land
Studying small-scale mining has several challenges including its remote setting in rural areas, large footprint spread into wide zones along rivers, and is typically caught in armed conflict. Earth observation through satellites can allow us to have a better understanding of the story. The Copernicus program has been providing free satellite data every 6 days and with 10m resolution, leading to new opportunities in analyzing such a dynamic activity.

Study site in Colombia

In order to interpret satellite data and obtain useful information, machine learning techniques need to be used. To train these techniques, we need knowledge of aspects we are searching for. Thus, it is important to have study areas which get to study and learn about very well and use to train and validate our machine learning models. For EO-Allert, the study area is a mining region close to the town of El-Bagre and is around the border between the municipalities of El-Bagre and Zaragoza in the Department of Antioquia (see image on the left).

Our approach

The diagram on the right describes the EO-ALLert methodology and workflow.

Project Profile

EO-ALLert project is carried out by the University of Liège (GeMMe group) in collaboration with with United Nations Environment Programme (UNEP) and UNEP-GRID-Geneva. It is funded by RawMatCop 2018-2020 Programme which aims to develop skills, expertise and applications of Copernicus data to the raw materials sector. It is the continuation of the preceding RawMatCop project CopX.
Ongoing 2019-2020.
Contact gemme@uliege.be