Corresponding author: Klaus Henle ( klaus.henle@ufz.de ) © Klaus Henle, Reinhard Klenke, Benjamin Barth, Annegret Grimm-Seyfarth, Diana Bowler. This is an open access preprint distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Citation:
Henle K, Klenke R, Barth B, Grimm-Seyfarth A, Bowler D (2024) Challenges and opportunities for assessing trends of amphibians with heterogeneous data - a call for better metadata reporting. ARPHA Preprints. https://doi.org/10.3897/arphapreprints.e137859 |
Abstract
Over the last decades, the worldwide decline of amphibian populations has become a major concern of researchers and conservationists. Studies have reported a diversity of trends, with some species declining seriously, others remaining stable and still others increasing. Only a few species have been monitored annually for a long period of time by specific monitoring programs. However, there are many heterogeneous datasets that contain observations of amphibians from professional surveys as well as diverse citizen science and other voluntary surveys. The use of these data brings a number of challenges, raising concerns about their validity and use in ecological research and applied conservation. We assessed to what extent such heterogeneous occurrence data can inform on the status and trends of amphibians by contrasting different approaches to overcoming challenges with the data, using the German state of Saxony as an example. We assessed the effects of data processing decisions to infer absences, the use of survey method information, and the statistical model (generalized linear mixed-effect occurrence model [GLMM] versus occupancy-detection model) and compared the trends with expert opinions (Red lists). The different data processing decisions mainly led to similar annual occupancy estimates, newts being an exception. Annual occupancy estimates were typically less certain when attempting to account for the effects of survey methods, which could be explained by many missing values on methods. Separate models for drift fence data reduced the uncertainty in the annual occurrence probability estimates of the GLMM models, but uncertainty remained high for occupancy-detection models. For both methods, strong peaks and troughs in the annual occupancy estimates occurred for several species, which were not biologically plausible. Some peaks align with periods of lower sampling effort and were probably caused by shifts in the sampling locations or target species among years. Only for three species (Bufotes viridis, Hyla arborea, and Pelophylax esculentus) were the trend results consistent among approaches and with expert opinions. For most other species, some inconsistencies appeared among models or approaches, indicating that trend assessments are sensitive to analytical choices. While heterogeneous data have proved useful for other taxa, our results highlight the complexity of using them for amphibians. We strongly recommend better harmonization of data collection and documentation, including explicit absence data, to enable more robust trend assessments in the future.