Canadian Forest Service Publications
Individual Tree Crown (ITC) Inventory and Analysis of the Petawawa Research Forest: Stand’s Species Composition and ITC-based Volume
Year: 2023
Issued by: Pacific Forestry Centre
Catalog ID: 41020
Language: English
Availability: PDF (download)
Abstract
Individual tree inventories promise precision advantages relative to conventional forest inventories and those enhanced by a LIDAR area-based technique. Although this promise has yet to be realized at scale in Canada, advances continue to bring individual tree inventory technologies closer to practical application. This report describes a practical demonstration of the Individual Tree Crown (ITC) approach on the Petawawa Research Forest in Ontario. A typical mosaic of multispectral aerial images (40 cm/pixel) was used to automatically delineate approximately 10 million ITCs, classify these into species and estimate their volumes. To mimic operational situations, the spectral signatures were generated from simple training areas produced using information from two previous forest inventories, with minor on-screen interpretation. Classification accuracy was ascertained for ten common species. Although species classification in this demonstration was generally poor due to some species confusion, the conventional forest inventory benchmark is also known to be poor for challenging situations. Conventional equations were used to establish volumes for 2,000 trees for which species, diameter at breast height (DBH) and height were measured in the field, and simple volume relationships were established for each species plotted as a function of height only (V = f (HT)) for application to the ITCs for which we only have species and height (i.e., no DBH). Summarizing at the stand level while compensating for the fact that some of the ITCs are actually tree clusters rather than single trees produced estimates consistent with standard volume assessment of the Petawawa Research Forest.
Plain Language Summary
Individual tree inventories promise precision advantages relative to conventional forest inventories. This report describes a practical demonstration of the Individual Tree Crown (ITC) approach on the Petawawa Research Forest in Ontario. A typical mosaic of multispectral aerial images (40 cm/pixel) was used to automatically delineate approximately 10 million ITCs, classify these into species and estimate their volumes. Classification accuracy was ascertained for ten common species. Stand species composition and volume were compared with existing forest inventories.
Conventional equations were used to establish volumes for 2,000 trees for which species, diameter at breast height (DBH) and height were measured in the field. Then, simple volume relationships were established for each species plotted as a function of height only (V = f (HT)) for application to the ITCs for which we only have species and height (i.e., no DBH). Summarizing at the stand level, while compensating for the fact that some of the ITCs are tree clusters rather than single trees, produced estimates consistent with standard volume assessment of the Petawawa Research Forest.
This is the first use of the CFS’s ITC-Suite to assess volume based on individual trees.
