Integrating historical and recent data to measure long-term trends of endangered subterranean species

Abstract

Documenting population trends is pivotal to identify the underlying drivers of biodiversity changes and setting conservation priorities. Ascertaining trends often requires long-term, standardized, monitoring data that are not always available. Historical data provide important information on past species distribution, but their integration with recent data to obtain trend estimates is challenging. Here we show how site occupancy-detection models (SODMs) can allow combining data from recent monitoring with historical ones from the gray literature. Using data on the endangered cave salamander, Speleomantes strinatii, we tested whether SODMs can provide reliable trend estimates if i) historical data include repeated within-season surveys enabling the estimation of past detectability, or if ii) information on detection/non-detection is not available. We conducted repeated surveys across 40 caves covering the species range, for which historical (1940–1982) biospeleological data were available. We then developed Bayesian SODMs i) estimating species detectability from both recent and past surveys, and then assessing trends; ii) in absence of estimates of past detectability, assessing trends by comparing scenarios on the potential misdetection rate during historical surveys. Salamanders were widespread in the study sites. SODM estimated high detectability for both recent and historical surveys and suggested a growing occupancy. Changes in occupancy were unrelated to landscape modifications. Even without historical detection/non-detection data, realistic scenarios of past misdetection consistently suggested an increasing or stable trend. The application of tailored analytical approaches is fundamental to exploit the vault of information available in historical data, and can be linked to adaptive management to promote efficient conservation actions.