Developing scientifically robust risk maps for priority plant pests
Project ID: 170607
Project Name: Developing scientifically robust risk maps for priority plant pests
CEBRA Project Leader: Assoc Prof Andrew Robinson and Dr James Camac
DA Sponsor: Dr Marion Healy, First Assistant Secretary
DA Division: Biosecurity Plant
DA Project Leaders:
- Dr Susie Collins, Director, Plant Health Surveillance and Diagnostics, Plant Health Policy Branch, Plant Division
- Dr Ranjith Subasinghe, Asst. Director, Plant Health Surveillance and Diagnostics, Plant Health Policy Branch, Plant Division
- Kylie Calhoun, Asst. Secretary, Plant Health Policy Branch, Plant Division
- Dr Simon Barry, CSIRO
- Dr Peter Caley, CSIRO
- Assoc Prof Jane Elith, University of Melbourne.
The Department plays a major role in surveying for the early detection of high impact invasive plant pests. Surveillance for early detection of invasive plant pests is labour intensive and costly to maintain. Efficient allocation of increasingly scarce surveillance resources across all risk areas presents a significant challenge for DAWR. Compounding the issue of prioritising which pest species to target in early detection surveillance, little to no information is available about where, when and how a new pest species is likely to arrive and establish in Australia.
In order to determine where surveillance resources should be allocated to maximise early detection or confidence in pest-freedom, it is imperative we have an understanding of how risk of pest establishment varies across space (Wintle et al 2012, Hauser et al 2009). Fundamentally, the risk that a pest arrives and establishes at a location is a function of three primary processes:
1. Its ability to arrive at a given location,
2. The environmental suitability of that location, and
3. The presence of hosts/vectors at that location.
Several approaches exist to estimate each of these processes (Dodd et al. 2016, Barry et al. 2016, Elith 2011, Václavík & Meentemeyer 2009, Work et al. 2005). However, previous CEBRA projects (e.g. 1402B - Barry et al. 2016; 1302A - Burgman et al. 2014) have highlighted that different methods can give very different results, likely as a consequence of making different assumptions and having differing data requirements (Guillera-Arroita et al. 2015). These studies have also highlighted that there is no single ‘best’ approach to estimating invasive species distributions (Barry et al. 2015). The combination of these uncertainties has made it difficult for decision makers to decide how best to estimate pest climate suitability, arrival rates, potential invasive pest distributions, and consequently, how to develop scientifically defensible maps of risk of establishment.
The primary objective of this three-year project is to develop a standardised approach for estimating risk maps that incorporate pest arrival rates, environmental suitability and the presence of hosts. Specifically, the project will develop practical guides (i.e. decision trees) for deciding how ‘best’ to estimate environmental suitability and arrival rates, in the face of varying data quantities/qualities, pest biology, and uncertainty about the most appropriate model fitting approach. These practical guides will then permit a standardised approach for the development of scientifically defensible maps of risk of pest establishment.
The continuation of the project into its second year is critical for developing scientifically robust risk maps that can be readily applied to most plant pests. The second stage will focus using on estimating the distribution of pathway-specific relative risk across Australia. Initial risk maps developed in stage 1 will then be refined by accounting for improved estimates of geographic risk and the incorporation of additional high risk pathways of pest arrival. Furthermore, underlying code used to create these risk maps will be updated and supplied to the department for incorporation into their IT systems.
It is imperative that stage 2 be conducted to ensure developed risk maps are fit for purpose –1) they meet scientific standards (the methods are transparent and reproducible); 2) Code developed can be readily implemented in the Departments IT systems; 3) risk maps can be created across the Nation using best available data; and 4) Risk maps can be used to inform optimal allocation of finite surveillance resources, pest-freedom and pest spread.