REA Shiny

Session 1 Session 2 Session 3 Session 4 Session 5 Session 6 Session 7 Session 8 Session 9 Session 10 Session 11 Session 12
Min Likely Max
Min Likely Max


Rapid Eradication Assessment (REA) is a software interface to allow eradication managers to estimate the probability of successful eradication of a pest animal species from a defined area (such as an island) using a combination of monitoring tools (i.e. absence following repeated non-detection). REA was designed by a collaboration of the University of Auckland, Landcare Research, and Conservacion de Islas. The original description of the method can be found in the paper by Samaniego-Herrera et al. published 2013 in the Journal of Applied Ecology.

There are two ways to use REA. The first is to estimate the probability of eradication success following an actual eradication program or incursion response using an uploaded existing monitoring regime. This is only done when grid-spacing is set to zero. The second use presents what would be the probability of eradication success of a hypothetical monitoring regime specified by the user. This is done by simulating random grids of specified uniform spacing when grid-spacing is set to >0. This is useful for eradication managers planning eradication monitoring to a given level of confidence before placing devices in the field.

Confirming eradication is challenging as absence of evidence is not evidence of absence, i.e. just because no pest are detected does not mean none are present. REA works through repeated simulations of low density eradication survivors or incursions being monitored under a specified scenario of data and parameters. The resulting estimates of the probability of successful eradication are presented as a histogram, with a red line in the graph (posterior 2.5% quantile) with the median estimate as the blue line (posterior 50% quantile). The user chooses an acceptable lower bound over which they would be confident of pest absence (e.g. Target = 90%). The REA model provides a measure (the credible interval value) of how often this level of confidence would be met. Managers should aim for their acceptable minimum level of confidence to be met nearly 100% of the time.

Data Input

Users must first enter the target island's geography as a GIS shape file. All components of the shapefile (SHP, SHX, DBF etc) must be combined into a single ZIP file, before being uploaded. If spacing is set to zero to estimate the probability of eradication success using an existing grid of static devices, the users must also upload a CSV file with x (easting) and y (northing) locations of the devices, and a third column with device type number (1, 2, .). If mobile tracks (e.g. detection dogs) are used the users can also upload a CSV file with x (easting) and y (northing) tracks of the tracks, and a third column with track number (1, 2, .). If pest individuals are from a known location (e.g. incursion response) the users can also upload a CSV file with x (easting) and y (northing) of the single known location.

Monitoring data can be uploaded as a CSV file or entered manually for up to 12 sessions of eradication follow-up monitoring. These data include the uniform grid spacing for simulation of hypothetical grids (or set spacing to zero for uploaded grids), the number of nights of monitoring, the number of iterations to be undertaken (e.g. 2,000), the target minimum acceptable threshold probability for confirming absence (e.g. 0.90), and the number of years since eradication (e.g. 0.5 is 6 months) up to two years.

Static device parameter g0 (the probability of detection at the home range centre) can be uploaded as a CSV file or entered manually. These include minimum, most likely and maximum values for g0 (probability a single device would detect an animal if it encountered it on a single night). Currently only up to two different g0 can be used.

Biological parameters can also be uploaded as a CSV file or entered manually. These include minimum, most likely and maximum values for sigma (the standard deviation of a half-normal home-range kernel), the prior probability of the site having been made pest-free (e.g. 0.80 for global rat eradication attempts on islands), the probability of animals reinvading from elsewhere (e.g. 0.01 for remote islands), the annual growth rate of the target pest species if left uncontrolled (e.g. 5), the average dispersal distance for a juvenile from its parent's home-range centre (e.g. 125 m for a rat), and the average dispersal distance for a newly colonising individual (e.g. 625 m for a rat).

Example: Great Mercury Island

Click here to download an example containing Great Mercury data.

The zip contains 7 files: the Static Device Locations (GMIStaticDeviceLocations.csv) the Mobile Track Locations (GMIMobileTrackLocations.csv) the Shapefile ( the Incursion Location (GMIIncursionLocation.csv) the Monitoring Data (GMIMonitoringData.csv) the Detection Probability Parameter (GMIG0Parameter.csv). and the Biological Parameters (GMIParameters.csv).

Or automatically populate the user interface with the button below:


Kim, J. H. K., Corson, P., & Russell, J. C. (2020). Rapid eradication assessment (REA): a tool for pest absence confirmation. Wildlife Research, 47(2), 128-136.

Russell, J. C., Binnie, H. R., Oh, J., Anderson, D. P., Samaniego-Herrera, A. (2016). Optimizing confirmation of invasive species eradication with rapid eradication assessment. Journal of Applied Ecology, 54(1), 160-169.

Samaniego-Herrera, A., Anderson, D. P., Parkes, J. P., & Aguirre-Muoz, A. (2013). Rapid assessment of rat eradication after aerial baiting. Journal of Applied Ecology, 50(6), 1415-1421.

REA core R script

REA simulation R script

For any questions, problems or feedback please e-mail