【摘要】 为研究白洋淀水生生物栖息地适宜度,分别于2016年春、夏、秋3季进行了实地考察和采样,结合野外观测和实验室测量得到环境和生物信息。基于14个采样点的数据,通过构建广义可加模型(GAMs)模拟和预测底栖动物的分布。使用Margalef指数作为响应变量,采用向前逐步回归筛选解释环境因子,通过评价剩余偏差来判断模型的表现。水深、水温、底泥氨氮、底泥有机质和沉水植物生物量被用于建立最终优化的广义可加模型。Margalef指数对环境因子的响应曲线表明该指数与水深和底泥氨氮浓度成线性负相关,与水温和底泥有机质浓度成线性正相关,而与沉水植物生物量是单峰关系。结果表明,预测值与实测值呈现显著的强相关性(Pearson R~2=0.847,P<0.001),均方误差较小(MSE=0.013),模型性能表现良好。更多还原

【Abstract】 To explore the habitat suitability of aquatic communities in the Baiyangdian Lake, field investigations and samplings are done in the spring, summer, and autumn of 2016. The environmental and biological information are obtained combing in-situ measurements and in-lab testings. Based on data of 14 sampling sites, generalized additive models(GAMs)are established to simulate and predict the distribution of macroinvertebrates. Using the Margalef index as a response variable,explanatory environmental variables are selected by stepwise regression, and model performances are evaluated by residual deviations. Finally, water depth,water temperature, ammonium nitrogen in sediment, organic matter in sediment, and biomass of submerged macrophytes are included in the optimized GAM. The response curves of the Margalef index to those five environmental factors showes that, the index have negative linear relations with water depth and ammonium nitrogen in sediment, positive linear relations with water temperature and organic matter in sediment, while a unimodal relation with biomass of submerged macrophytes.The prediction showes that a significant high correlation exist between the predicted values and measured values(Pearson R~2=0.847, P<0.001), with a low mean square error(MSE=0.013). The results confirm a good predicting performance of the optimized model.更多还原