Old-Growth Forest Mapping
(01)

Challenge
The Nova Scotia Department of Natural Resources sought to improve upon existing GIS tools for predicting the presence of old-growth forest. For forest stands to be conserved, their age must be verified in the field - a costly and time-intensive process. GIS tools help guide workers in the field towards land that is more likely to be old growth, and increase the amount of old-growth forest that is formally protected. Previous GIS tools had an accuracy rate of 25% for mapping old growth, leading to time and money spent in the field that did not result in more conserved forest. The goal of this project was to employ new remote sensing data and modern machine learning models to create a more accurate predictor tool.


(1) Prediction map for old growth - red cells indicate high old growth probability
(2) Prediction map for forest cover type - used as auxiliary data for old growth machine learning model
