Data from: The importance of individual heterogeneity for interpreting faecal glucocorticoid metabolite levels in wildlife studies

1. Being a non-invasive and inexpensive method, the analysis of faecal corticosteroid metabolites (FCM) is increasingly being applied in wildlife research. Various environmental factors have been shown to influence FCM levels, but most studies did not account for inter-individual variance, which we hypothesized may substantially affect the results. 2. We combined FCM analysis with genetic analysis to identify the sex and individual of samples collected in three consecutive winters, with repeated samples per individual, across the entire range of an endangered population of capercaillie (Tetrao urogallus) in Southwestern Germany. Using generalized additive mixed models, we modelled FCM levels as a function of sex, season and environmental covariates at two spatial scales (sampling location and home range scales). We compared two models: one including information on the individual animal, and the other excluding this information (i.e. naïve model) to assess the influence of individual effects on the results obtained. 3. Most of the variance (44.0% and 45.1% at the sampling and home range scale, respectively) was explained by the inter-individual differences, and only very little (4.0% and 5.1%, respectively) by the environmental predictors. When ignoring individual effects, the model results changed considerably, with other, previously non-informative predictors, becoming significant. 4. In the full models, accounting for inter-individual variance, no effect was found of weather conditions, at either scale. FCM levels were negatively correlated with habitat quality, at the sampling location, whereas human recreation at the home range scale led to elevated FCM levels. In the naïve models, two additional predictors appeared significant: one weather variable at local scales and two at home range scale. In all models, seasonal FCM patterns differed significantly between males and females. 5. Synthesis and applications. Our results highlight the importance of considering the effects of individual heterogeneity when studying FCM in wildlife research, as ignoring information on the individual might lead to erroneous conclusions. Combining FCM analyses with genetic analyses can be an efficient approach to adequately address this issue.21-Feb-20181. 作为一种无创且成本低廉的检测方法，粪便糖皮质激素代谢物（faecal corticosteroid metabolites, FCM）分析在野生动物研究中的应用日益广泛。已有研究表明多种环境因素会影响FCM水平，但大多数研究未考虑个体间变异，我们推测该变异可能对研究结果产生显著影响。 2. 我们将FCM分析与基因分析相结合，以识别德国西南部濒危松鸡（Tetrao urogallus）种群整个分布范围内连续三个冬季采集的样本的性别与个体信息，且每个个体均配有重复采样样本。我们采用广义加性混合模型（generalized additive mixed models），以性别、季节以及两种空间尺度（采样点位尺度和家域尺度）下的环境协变量为自变量，对FCM水平进行建模。我们对比了两种模型：一种纳入个体动物信息，另一种则不纳入（即朴素模型），以此评估个体效应对所得研究结果的影响。 3. 大多数变异（在采样点位尺度和家域尺度下分别为44.0%和45.1%）可由个体间差异解释，而环境预测因子仅能解释极小部分变异（分别为4.0%和5.1%）。若忽略个体效应，模型结果会发生显著变化，原本无信息的其他预测因子会变得显著。 4. 在纳入个体差异的完整模型中，无论在何种空间尺度下，天气条件均未表现出显著影响。采样点位尺度下，FCM水平与栖息地质量呈负相关；而家域尺度下，人类休闲活动会导致FCM水平升高。在未纳入个体信息的朴素模型中，额外出现了两个显著预测因子：局部尺度下的一个天气变量，以及家域尺度下的两个天气变量。在所有模型中，雌雄个体的季节性FCM模式均存在显著差异。 5. 总结与应用。我们的研究结果强调，在野生动物FCM研究中考虑个体异质性效应的重要性——忽略个体信息可能会导致错误的研究结论。将FCM分析与基因分析相结合，可作为有效解决该问题的高效方法。2018年2月21日