Using multiple methods to assess detection probabilities of riparian-zone anurans: implications for monitoring

Abstract

Context Both manual call surveys (MCS) and visual encounter surveys (VES) are popular methods used to monitor anuran populations. Recent statistical developments, specifically the development of occupancy models that permit the use of data from various survey methods to assess method-specific detection probabilities, provide a rigorous framework for evaluating the effectiveness of field methods. Aim To compare species-specific detection probabilities generated by MCS and VES and to evaluate the effectiveness of these methods throughout the activity season of several riparian-zone anuran species. Methods During 2010 and 2011, we sampled 21 sites along the Broad and Pacolet Rivers, in South Carolina, USA, using MCS and VES. Anuran species were surveyed across three seasons (fall, spring and summer) each year. Key results For six species, MCS resulted in a higher mean probability of detection, whereas VES resulted in a higher mean probability of detection for four species. In addition, survey date was an important influence on detection probability of most anurans when using MCS, but largely unimportant when employing VES. Conclusions Our findings indicated that VES are as effective as MCS for detecting some species of anurans, and for others, VES represent a more effective method. Furthermore, when using VES outside the breeding window, some anurans can be reliably detected, and in some cases, detected more easily than by using MCS. Implications We suggest that VES is a complimentary technique to MCS and a potentially important tool for population monitoring of anurans. VES can provide more flexibility for anuran researchers, as robust estimates of detection and occupancy can be obtained outside a narrow breeding window.