Population forecasting of Australian Plague Locusts using machine learning
Project summary
Outbreaks of the Australian Plague Locust pose a serious threat to agriculture in eastern Australia, causing up to $30 million in damage annually. Large swarms of locusts can cover vast distances and consume enormous amounts of vegetation in their path, impacting crop production and grazing areas.
The Australian Plague Locust Commission, in Agriculture, Fisheries and Forestry, has developed more advanced forecasting tools to improve early detection and responses to locust outbreaks.
The project has been delivered collaboratively by the APLC and the University of Melbourne’s Centre of Excellence for Biosecurity Risk Analysis. It uses open-source tools and is hosted within the national Biosecurity Commons platform.
Australian Plague Locust (Chortoicetes terminifera) nymph.
Image: Dr Gordon Berg
How AI contributes to the project
The project has delivered a type of machine learning AI algorithm, called Random Forest Classification, that more accurately predicts short-term forecasts of locust outbreaks up to 2 weeks ahead. The model incorporates over 37 years of APLC locust data records with high-resolution climate, vegetation, soil, land use and land cover information. This historical harmonised dataset will be used for future analysis and model refinement.
The AI-generated forecasts support the commission’s in-house simulation model, Life System Simulator. This assists medium-term strategic planning by predicting seasonal forecasts for plague locust population dynamics across their entire range, including the development, survival and behaviour of locusts across all life stages.
The Australian Plague Locust Commission is integrating this AI-driven forecasting system into a new interactive data dashboard, called WebMapper v2. The dashboard will deliver the forecasting outputs in a secure system for visualising monitoring data, forecast maps and environmental layers such as vegetation indices and climate anomalies.
Outcomes and next steps
The new forecasting system has strengthened the Australian Plague Locust Commission’s capacity to prioritise field surveys and control operations effectively, leading to better planning and resource allocation. Weekly likelihood maps of nymph presence, stage and density improve early detection and support proactive management.
Next steps include retraining the AI models with new monitoring data and refining the predictor layers. This will be achieved using updated Bureau of Meteorology climate reanalysis products, as well as Pasture Growth Index surfaces, land use and cover data. Automated pipelines will be developed to deliver regular, operationally ready forecasts.
Staff training will ensure ongoing in-house capability, supporting DAFF’s biosecurity goals and the National Biosecurity Strategy’s emphasis on evidence-based, data-driven risk management.
Find out more
Baumgartner, J.B., Spessa, A., Deveson, E., and Camac, J.S. (2024) Short-term population forecasting of the Australian Plague Locust (Chortoicetes terminifera). CEBRA project 23A. Technical Report for the Australian Plague Locust Commission, accessed 5 August 2025.
The University of Melbourne (2025) Biosecurity Commons, Biosecurity Commons website, accessed 24 July 2025.
Department of Agriculture, Fisheries and Forestry (2024) Strengthening Locust Management with Advanced Forecasting Models, DAFF website, accessed 24 July 2025.