New research on selective logging in Brazil

Matt Hethcoat’s Grantham Centre research used satellites to track changes in Brazilian forest cover caused by selective logging. Now Matt is building on this work to monitor selective logging more effectively.

Recently, Matthew and his team published research on mapping forests. You can read this research here: Mapping pervasive selective logging in the south-west Brazilian Amazon 2000–2019. Here, Matthew sets out how this new research fits in with wider goals to protect tropical forests. And he explains how an algorithm is helping him do it.

Losing tropical forests

a logging road used for selective logging
Narrow logging road cutting through the forest. Photo credit: Manoela Mollinari.

The value of tropical forests is incalculable. For instance, despite only covering about ten percent of Earth’s land surface tropical forests are estimated to host at least two-thirds of terrestrial biodiversity.

In addition, forest provide goods and services such as wood products, medicines, climate change mitigation and water filtration. As such, they are intimately linked to human livelihoods and the global economy.

However, a rapidly rising global population and continued loss and degradation of forested landscapes are putting incredible pressures on tropical forests.

In recognition of the challenges tropical forests face there has been a tremendous effort to improve our understanding of where, when, and how tropical forests are being lost.

Thankfully, this goal has largely been achieved. Because with advancements in satellite monitoring of forests, we now know (in almost real-time) when tropical forests are cut down.

Unfortunately – at the risk of downplaying a huge effort by many people – reaching this goal was the easy part.

What next?

The next big challenge involves understanding the where, when, and how of forest degradation.

Forest degradation is when a forest has been disturbed in some way, but remains a forest. So though the forest patch still exists, it has been affected by a fire, selective logging, small-scale mining, or some other human-induced disturbance.

Crucially, disturbance to forests lowers their ability to return to a pre-disturbance condition.

Selective logging – one type of forest degradation

My PhD research at the Grantham Centre focused on improving our understanding of one degradation type: selective logging.

Most tropical forests only contain a few species that are commercially valuable for wood products. Selective logging targets these, focusing on the most valuable. When this subset of trees is harvested the forest is left to regenerate for 30-50 years. Then there is another round of cutting.

Unfortunately, selective logging has been identified as a key driver of tropical forest loss, because many forest patches are eventually cleared when the selective logging frontier moves elsewhere.

A machine learning algorithm to map tropical selective logging

Last year my team published a paper outlining a methodology to detect and map selective logging. Ironically, logging records have been the key to our team’s success – the team used information from hundreds of thousands of tree harvest locations to build a machine learning algorithm.

Our algorithm understands what satellite imagery looks like when trees have been cut down. You can read the open-access paper about this research here.

Our new research on selective logging

the results of selective logging
Logs are measured and prepared on a log landing deck before being transported. Photo credit: Manoela Mollinari.

Our newest paper scales up the proof-of-concept previously published.

This new paper includes maps of logging across the Brazilian state of Rondônia (which is about the size of the UK) between 2000 – 2019. And these maps show widespread selective logging that improves upon existing forest disturbance monitoring tools.

If you’re interested, you can find this research here: Mapping pervasive selective logging in the south-west Brazilian Amazon 2000–2019

Overall, this new work is a key step in creating maps of logging and understanding the extent of disturbed tropical forests. Hopefully such maps will enable better accounting of selective logging’s impacts to biodiversity, carbon emissions, carbon sequestration potential, and more.

Ongoing work to track logging

This work has led to a collaboration with a government agency in Peru: El Organismo de Supervisión de los Recursos Forestales y Fauna Silvestre (OSINFOR). This organisation – the Organisation for the Supervision of Forest Resources and Wildlife in English – monitors forest resources.

Over the coming months the OSINFOR will be piloting a selective logging monitoring system. The hope is to improve current abilities to track logging activities across the country.

There will be more updates as this project develops – you can sign up to our newsletter to stay up to date.  


Written by Matt Hethcoat and Claire Moran.

Main image by Manoela Mollinari. Photo shows trucks loaded with tropical hardwood and ready for transportation.