David Smith
Forest Pathologist/Entomologist | ArborCarbon
Tuesday | 9:30am | ARB Stream
TALK TITLE
The use of tree databases, remote sensing and artificial intelligence to assist in urban forest biosecurity preparedness and response
BIO
David has worked as Senior Forest Biosecurity Officer with Agriculture Victoria within the
Chief Plant Health Officer Unit and as Honorary Research Fellow with the University of Melbourne.
David has experience in investigating the biology and control or amelioration of biotic and abiotic agents that threaten the productive capacity of softwood and hardwood plantations and nurseries, the productivity and ecology of native forests, and the control and monitoring of endemic and exotic pests and pathogens in farm forestry and urban forest landscapes.
He has designed, established and delivered of forest health monitoring and surveillance programs for plantation companies and state and local governments. He was involved in the eradication program for an exotic pine nematode, and in ongoing eradication or containment/transition to management programs for chestnut blight, palm fusarium wilt, myrtle rust and Giant pine scale in Victoria. He has provided technical advice to local, state and federal governments, along with National Forest Health and Biosecurity
Committee (which includes Urban Forests), and the Victorian Committee for
Amenity Tree Health, on endemic and exotic pests and pathogens of trees and
wood products.
ABSTRACT
There are many threats to Australian Urban forests from exotic and endemic insects and pathogens. Biosecurity surveillance programs focus on mitigating the risks posed by these biotic agents. Understanding pest’s pathways and their hosts is essential for preparedness and surveillance. Urban forests around high-risk entry pathways like ports can be used as sentinels for early detection, but assessing large numbers of trees can be time-consuming.
In Melbourne, Australia, many local governments maintain accurate tree databases that allow for precise host mapping within public land. However, this is not the case for all other cities in Australia and remains a challenge within private land. Host mapping during a response following the detection of a pest is time-consuming and resource intensive; generally conducted in vehicles and foot patrols. Recent work in Australia has shown that using tree databases, ArborCamTM airborne remote sensing camera system mounted to a fixed-wing aircraft and the use of artificial intelligence allows for automation and detection of host trees with damage within the crowns.
The use and analysis of ArborCamTM multi-spectral data provides information on tree health, size, and shape, detailed information on vertical and horizontal structure, and spectral and thermal information for object identification and measurement. The benefit of the system is that it is focused on detecting spatial and temporal changes in vegetation and particularly tree health across all land tenures enabling early response to impacting factors or evaluation of control activities. Remote sensing can be valuable in remote communities where on ground surveillance is difficult due to limited access and their vast areas. Tree databases and remote sensing can aid in identifying areas for targeted surveillance, speed up responses and increase efficiency of on-ground surveillance.