Risk assessment for nonindigenous pests: 1. Mapping the outputs of phenology models to assess the likelihood of establishment

This paper demonstrates the use of phenology models mapped over the landscape as a tool in support of risk assessments for nonindigenous plant pests. Drawing on the relationship between pest development and temperature, the approach uses gridded sequential interpolated temperatures at a resolution of 1 km, linked with phenology models, to predict the potential for a pest to develop throughout the landscape. The potential for establishment of Colorado beetle (Leptinotarsa decemlineata) in England and Wales was used as an illustration. The likelihood of the pest completing a single generation during a 30-year period (1961–90) was computed. Summaries of phenology, based firstly upon point temperature series from weather stations and secondly upon temperatures interpolated across the landscape, were compared. The results revealed that the use of point data led to a 70% likelihood of over-estimating the area at risk from year to year. In the case of average long-term risk however, the point-based and landscape-wide distributions of establishment potential were similar. We demonstrate how the use of phenology models running on a daily time scale provides date based results, so allowing outputs to be tied in with periods in the cropping cycle. The application of daily data in computing the phenological results, unlike the main body of published work on pest risk assessment which uses averaged monthly data, reflects more fully the underlying variability and degrees of sensitivity of the pest to changes in weather.