UCL Department of Geography


Description Photo Here

Personal tools
Log in
This is SunRain Plone Theme
UCL Home  /  Geography  /  Blog  /  Blog Entries  /  Gabon: quantifying forest structure

Gabon: quantifying forest structure

Posted by ucfaam0 at Oct 11, 2013 09:20 AM |

By Andrew Burt

Following on from Mat Disney’s previous post Shining a light on forest structure I have recently returned from a three and a half week field campaign in Gabon, West Africa to continue researching the potential of terrestrial laser scanning for quantifying forest structure.

Our main aim was to take a laser scanner into a tropical forest environment for the very first time and survey regularly monitored plots that form part of the Global Ecosystem Monitoring network. Here, a host of direct field measurements, including trunk diameter and height, are rigorously collected for each individual tree in a one hectare plot to identify forest state and dynamics. Destructive harvesting is used to understand how tree trunk diameter and height relate to tree biomass.

Research combining terrestrial laser scanning and 3D modelling work undertaken at UCL Geography (see here) and colleagues from Tampere University of Technology (see here) has allowed us to estimate forest biomass in a fashion that is completely independent of these direct field measurements. Laser scanning produces a full spatial representation of the scanned area in the form of a 3D point cloud as shown in the video and images below. By converting these point clouds into a topological description of tree structure we can derive volume and hence biomass. This methodology provides a wealth of further benefits as these 3D tree models can also be used to drive remote sensing models for instrument analysis and parameter retrieval algorithms. Collaboration with UCL Geography forest ecologist Simon Lewis is providing the opportunity for these models to answer fundamental ecological questions which are difficult to assess using traditional measurement techniques.

A fly-through of some of the 3D data from Lope National Park (video: Kim Calders)

An individual tree extracted from the point cloud and the 3D model of this isolated tree

An individual tree extracted from the point cloud (left) and the 3D model of this isolated tree (right). Data acquired at Brisbane Forest Park, Australia (image: Andrew Burt)

Applying this approach in Gabon at regularly monitored plots is a great way to compare the output of direct field measurements and remote sensing efforts. Through this, we can investigate their agreement and/or disagreement and ultimately improve how we go about measuring forest structure.

Accompanied by two colleagues, Kim Calders and Jose Gonzalez de Tanago Menaca from the Laboratory of Geo-information Science and Remote Sensing at Wageningen University and their RIEGL VZ-400 terrestrial laser scanner we arrived into the capital, Libreville, prior to heading out to the first site in Akandah National Park. In just seven days and battling with some extensive understory and a multitude of very large ants, we managed to survey two, one hectare plots.

Akandah National Park

Akandah National Park (image: Jose Gonzalez de Tanago Menaca)

Following this we journeyed to Lope National Park located in the centre of the country which is a wonderful mix of savannah through to old growth forest. Here we scanned one further plot and ten smaller monitored plots that covered the full range of forest types present at Lope.

The team in front of a 2.5m diameter tree in Lope National Park

The team in front of a 2.5m diameter tree in Lope National Park (image: Jose Gonzalez de Tanago Menaca)

It was here that we also came across a friendly elephant known to the locals as Billy who kindly allowed us to give him a quick scan!

Billy the elephant

Billy the elephant! (image: Kim Calders)

This short campaign was a huge success and has demonstrated the feasibility of using laser technology to monitor forest structure in harsh conditions. Work is now underway to analyse the 210GB of scanner data acquired and we hope for there to be more news on this soon!

Many thanks go to the NERC National Centre for Earth Observation for funding the campaign, Aida Cuni Sanchez (UCL Geography) and the Gabon National Park Agency for all their assistance.

Andrew Burt is a PhD candidate at UCL Geography. His research focuses on quantifying forest state and degradation using LiDAR and is funded through NERC in partnership with EADS Astrium.