Binomial mixture models (BMMs) have been increasingly applied to account for imperfect detection and to estimate abundance from count data, but their performance has not been thoroughly evaluated. Here, I conducted simulation experiments to examine parameter estimates in BMMs under various situations. I generated data by assuming that abundance followed a Poisson distribution with an expected value λ and that the number of detected individuals followed a binomial distribution with an individual detection probability p. In simple simulations without covariates for λ and p, when the number of sampling sites (n) was between 20 and 160, BMMs could recover λ and p under the following conditions: 0.1≤λ≤160 and p≥0.1. However, within these ranges of λ and p, the estimates were variable under lower values of λ and p, although the situation improved as n increased. When λ and p are expected to exceed these ranges and the sample size is small, the results suggest that sampling and/or modeling designs should be reconsidered. I then conducted simulation experiments with covariates. I assumed that λ increased with a covariate (x) across 20 sampling sites. I varied p, number of visits (v), and their dependency on a covariate. To compare BMMs with analyses that did not accommodate imperfect detection, I fitted ordinary Poisson generalized linear models to mean and maximum counts (GLMmean and GLMmax). The results showed that GLMmax was superior to GLMmean because GLMmean underestimated λ when p was small. GLMmax underestimated a coefficient of the covariate (slope) when v was negatively correlated with x. BMMs successfully recovered true values of the intercepts, slopes, and λ in most cases. However, when p and v were small, and when p and λ were highly negatively correlated due to their inverse dependency on x, estimates from BMMs were more variable.
The range of Eurasian Coot Fulica atra in Japan has expanded gradually since the 1980s, and the numbers of coot wintering at several sites in Japan have increased. In this paper, we summarize the trends in coot numbers wintering in Japan and mainland East Asia. In Japan, wintering numbers of coot have increased in most regions, except in northern Chiba Prefecture. The increase began in the late 1990s or around 2000 at several water bodies in the Kanto, Chubu, Kinki and Chugoku regions of Honshu. It also increased in the mid-2000s at several water bodies in the Tohoku, Kanto, Chubu, and Kinki regions of Honshu, and perhaps also on the island of Kyushu. In China, huge numbers of coot overwintered at several sites in the lower Yangtze floodplain (e.g., Poyang Lake, Dongting Lake) and along the Jiangsu Coast near the Yangtze River mouth (e.g., Yancheng Nature Reserve, Hongze Hu Lake) during the 1990s. However, during the 2000s, no sites in China held more than 10,000 wintering coot. The number of coot wintering in coastal south China has also decreased. In South Korea, the number of wintering coot has increased especially since the mid-2000s. The increases in the populations of wintering coot in central and western Japan and in South Korea seem to be mainly as a result of changes in coot populations and/or habitats outside Japan. We suggest two possible causes: the breeding population of coot on the East Asian mainland may have increased, or some coot that used to overwinter in China may have begun wintering in Japan and South Korea.
We studied the effects of climatic conditions on the number of juvenile Whooper Swans wintering in Aomori Prefecture, northern Japan, from 1983 to 2009. We analyzed the relationship between the abundance of wintering juveniles and eight climatic variables using generalized linear models; (May and June temperatures recorded in Cokurdah, Yakutia, Russia within their breeding range, the temperature in Nikolayevsk-na-Amure within their migratory stopover range, snowfall in Obihiro, Hokkaido, a stopover site, and temperature and snowfall of the previous and present years in Nobechi, Aomori Prefecture, where they winter. The models with good fit showed that the mean highest daily temperatures in May on the breeding grounds had a positive effect on the abundance of wintering juveniles, while the total snowfall of both previous and present years in the wintering range had a negative effect on juvenile abundance.
For species conservation, identifying factors that affect the distribution of the target species is essential. To provide basic information for the conservation of Greater White-fronted Geese in Japan, this study investigated factors associated with the distribution of the species by using data from the nationwide waterfowl survey. The geese were observed in 88 grid cells (about 10 km×10 km) in 2008–2012. The maximum entropy approach was adopted to model the distribution using presence-only data. Explanatory variables were the proportion of rice field area, urban area, and lake area; distance to lakes; and maximum snow depth. Spatial filters were also included in the model to account for spatial autocorrelation. Habitat suitability for the geese was high in areas with a high proportion of rice fields, having a certain proportion of urban area and lake area and less snow. Consequently, the suitable habitats tended to be distributed on the plains along the Sea of Japan and Pacific coasts. This study represents the first step towards habitat management and restoration of Greater White-fronted Geese in Japan.
In Japan, Rice Oryza sativa fields are regarded as important sites for wintering waterbirds. Rice fields harbor considerable amounts of post-harvest waste rice, which is an important food resource for dabbling ducks in winter. The availability of rice fields as duck foraging sites is affected by agricultural management methods, harvest methods or post-harvest treatment of residues. In the present study, we examined the effects of two agricultural practices after harvesting, flooding with shallow water and plowing the surface of fields, on foraging site selection by ducks in rice fields near Katano-kamoike, Ishikawa Prefecture, Japan. To investigate the distribution of duck flocks in rice fields, night censuses were conducted during winters 2004–2005 to 2006–2007. Field conditions (flooded or plowed) in the census area were also recorded. By GLMM model selection, it was shown that water cover had a positive effect on the number of ducks, but post-harvest plowing had a negative effect. Furthermore, the effect of plowing was tested by a field experiment. The initial grain density on the ground was higher at unplowed plots. Unplowed rice fields are considered to be more suitable for ducks as foraging sites, with plowing reducing the attractiveness of rice fields as foraging sites by burying residual rice grains.