Diameter Estimation of Eucalyptus spp. Plantations in Southern Brazil Using Global Ecosystem Dynamics Investigation Data and Support Vector Regression

Forest plantations make up a large percentage of managed forest land globally. Assessing plantation productivity is vital from both commodity production and carbon management standpoints. Measuring the productivity of these areas is essential given their rapid growth and turnover. Transparent metrics to compare reported carbon storage with estimated values are required for internationally transferred mitigation outcomes under Article 6.2 of the Paris Agreement. Data from the Global Ecosystems Dynamics Investigation (GEDI) provide an excellent opportunity to measure plantation forests over large areas. We focused our efforts on Eucalyptus in southern Brazil and used data from an industrial partner to investigate plantation metrics (height, diameter, volume, stems per hectare, etc.) and to create a model of plantation diameter using Support Vector Regression (SVR). SVR enabled a robust model of tree diameter even given the heteroskedasticity and spatial auto correlation present in the GEDI data, which deleteriously impacted attempts at linear modeling. We could predict tree diameter in these plantations to within 1 cm using space-borne lidar, with broad implications for using space-borne lidars to monitor carbon accretion in secondary forest plantation.

Name thesis:degree_name

Master of Science

Level thesis:degree_level

masters

Discipline thesis:degree_discipline

Forestry

Department dc:contributor.department

Forest Resources and Environmental Conservation

Grantor dc:publisher

Virginia Tech

Year dc:date.issued

2022

Author dc:creator

• Miller, Benjamin Daniel

Chairs dc:contributor.committeechair

• Wynne, Randolph H.
• Thomas, Valerie A.

Committee member dc:contributor.committeemember

• Schons, Stella Z.

• Remote Sensing
• Eucalyptus
• lidar
• Machine Learning
• forest plantations

Statement dc:rights

• Creative Commons Attribution-ShareAlike 4.0 International

Licence dc:rights.uri

• http://creativecommons.org/licenses/by-sa/4.0/

Language dc:language.iso

en

Dc Identifier Other

vt_gsexam:35136

OAI identifier oai:identifier

oai:vtechworks.lib.vt.edu:10919/110918