Abstract
Citizen scientists regularly collect monitoring data for threatened species to improve the spatial and temporal resolution of sampling. Such programs should adopt robust data assurance measures and statistical approaches to reduce observer bias and better inform uncertainty estimates while supporting management decisions. In this study, we estimated trends and drivers of malleefowl (Leipoa ocellata) breeding activity within a Bayesian hierarchical modelling framework using 1823 site x years of nest count data collected by volunteers in Australia. Our modelling suggests malleefowl breeding activity decreased 4.8% annually in South Australia ( - 0.050; 95%CIs - 0.062, - 0.037), decreased 2.1% annually in Western Australia ( - 0.022; 95%CI - 0.040, - 0.004), was stable in Victoria ( - 0.001; 95%CI - 0.010, 0.009) and increased 4.8% annually in New South Wales (0.047; 95%CI 0.009, 0.086). We found strong evidence for positive associations between winter rainfall (0.084; 95%CI 0.004, 0.165), time since fire (0.288; 95%CI 0.179, 0.399) and an interaction between time since fire and the proportion of a site burnt (0.292; 95%CI 0.173, 0.410). Malleefowl breeding activity was negatively associated with patch size ( - 0.255; 95% CI - 0.642, 0.020) and the proportion of a site burnt ( - 0.191; 95% CI - 0.363, - 0.030), suggesting small reserves are important for conservation and the extent and frequency of fire should be managed cautiously. While our index of fox abundance decreased as baiting effort increased ( - 0.484; 95%CI - 0.640, - 0.317), there was little evidence for this benefiting malleefowl. This study demonstrates how volunteers can play a vital role understanding population trends and informing conservation of a threatened species at a national scale.