An indirect method for assessing the abundance of introduced pest M yocastor coypus ( R odentia) in agricultural landscapes

Pest management requires the development of robust monitoring tools. In I taly, coypu M yocastor coypus (nutria) have been controlled since the early 1990s, but the effectiveness of these measures has never been tested. With the aim of developing a reliable and volunteer‐based method for the long‐term monitoring of coypu abundance in agricultural landscapes, we calibrated an index based on surveys for coypu paths against density estimates obtained through a standardized mark–recapture technique. Two trapping sessions were performed in winter for each of 12 1‐km long stretches of irrigation canals and watercourses using 15 baited cage traps. Trapping sessions lasted 7 days each, with a 10‐day break between sessions. Population size was assessed using three methods: P eterson– L incoln's formula, capwire estimators and accumulation curves. Active coypu paths and five habitat variables were recorded by walking on the edge of both banks. The variables were then related to population size ( y ) by means of multi‐regressive models, testing for the predictive power of the selected models by leave‐one‐out cross‐validation. Multi‐regressive models included only the number of coypu paths with the best performances achieved by the model based on P eterson– L incoln formula, supporting path count as an effective method to assess the abundance of the coypu in agricultural landscapes. Concurrently, to assess the field suitability of the indirect method, surveys for coypu paths were carried out on 122 randomly chosen 3‐km long stretches of irrigation canals and watercourses in the central part of the R iver P o valley ( c . 15 000 km 2 ; N I taly). The highest (>8/100 m) mean number of paths was recorded in the central part of the study area. According to the regression models, the overall number of coypu is predicted to range between 350 000 and 1 100 000, raising doubts about the effectiveness of current control measures.