影响鱼类分布的生态水文因子建模及分析
Modeling and analysis of eco-hydrological factors influencing fish distribution
-
摘要: 人类活动干扰下河流自然水势改变,导致鱼类栖息地环境变化。为准确分析影响鱼类分布的生态水文环境,该文以美国阿肯色河流域的国家级濒危小型鱼类:产漂流性卵鲤鱼(Arkansas River Shiner,Notropis girardi,ARS)为例,根据其繁殖期特有的水文条件需求,选取流域40个水文站点的日径流数据,分别计算历史时期(1950-1962年)和当前时期(1990-2010年)的3个繁殖期生态水文因子:高流量,较高流量历时及断流历时。每个生态水文因子分别结合其他25个环境因子,基于最大熵理论(MaxEnt)构建物种分布模型。运用经典受试者工作曲线(receiver operating characteristic,ROC)及Parolo等提出的模型转移前后预测结果的Spearman’s rho相关系数,评价模型的自适应性(model fit)和转移能力(transferability)。3个模型在两个时期的ROC曲线与横坐标围成的面积均大于0.95,模型转移前后预测结果的Spearman’s rho相关系数均大于0.6。研究分析表明,模型具有较好的自适应性及转移性。3个生态水文因子中,高流量对模型的贡献率最大,对该种鱼类在两个时期的潜在分布影响最显著。不同的历史时期,物种的潜在适生概率随同一生态水文因子的变化呈现出了不一致的变化趋势。3个模型均预测出ARS的潜在生境有相同的空间变化趋势。该研究结果能够为中国保护河流生态系统与濒危鱼类提供借鉴经验。Abstract: The change of flow regime under the disturbance of human activity leads to alteration in fish distribution.To accurately analyze key eco-hydrological characteristics, Arkansas River federally endangered small-bodied fish: pelagic broadcast-spawning cyprinids reproductive guild(Arkansas River Shiner, Notropis girardi, “ARS” for short) was studied.According to the hydrological requirements for the spawning, daily flow data from 40 flow gages in the historic period(1950-1962) and the current period(1989-2010) were downloaded for the calculation of three important ecological hydrological fators: high flow, high flow duration and zero flow days.In combination with 25 other environmental factors, three species distribution model based on the theory of maximum entropy(MaxEnt) were established to analyze the contribution and influence of three flow metrics to the predicted distribution.The classical method(Receiver Operating Characteristic, ROC) and Parolo’s calculation of the Spearman's rho correlation coefficient before and after the transfer of the models were used to evaluate the model fit and transferability of the models.The area under the ROC curve(AUC) greater than 0.95 and the Spearman’s rho value greater than 0.6 for all the three models showed the good model fit and transferability of the models.Of the three flow metrics, high flow made the highest contribution to the model fit and had the most significant effect on the potential distribution of the fish.In different period, the species occurrence probability changed differently as the flow metric varied.Our study could provide valuable references for the protection of endangered fish in China.