Fish Data Collection on the Canadian River 1995-2015

The use of streamflow simulations from the Vflo model and subsequent calculation of streamflow metrics to investigate flow-ecology relationships may be hindered by our inability to accurately model flow variability and extreme flows of the arid Great Plains. The Canadian River and other rivers in the Great Plains tend to have highly variable flows and harsh environmental conditions. The combination of these environmental conditions makes semi-arid and arid regions difficult to represent with a hydrologic model, especially extreme events. In some cases, overestimating flows may be acceptable to water managers (e.g., vulnerability of infrastructures), but could greatly affect estimates of fish species persistence. To address incidences where poor model performance affected metrics derived from Vflo simulations, we suggest three possible options. 1) Restrict flow-ecology relationships to the mainstem of the Canadian River below Lake Meredith, 2) Restrict assessments to streamflow data aggregated at a monthly time step (although typically, this does not match ecological processes well); 3) Focus on streamflow metrics with a high prediction accuracy (e.g., magnitude, timing and duration at some locations). To maximize the number of potential explanatory variables and survey locations available in the Canadian River basin for the development of flow-ecology response models and minimize bias and uncertainty, a combination of these approaches is likely warranted. To move forward on flow-ecology relationships with valid statistical power, the compiled fish data (see processing steps) is best combined with available gage data to improve the development of ecological relationships.

利用Vflo模型（Vflo model）的径流模拟结果及后续计算的径流指标开展径流-生态关系研究，可能会因我们无法精准模拟干旱大平原地区的径流变异性与极端径流过程而受到制约。加拿大河及大平原内其他河流往往兼具极强的径流变异性与严苛的环境条件。这类环境条件的叠加效应，使得半干旱与干旱区域难以通过水文模型（hydrologic model）实现精准表征，极端事件的模拟更是如此。在部分场景中，水资源管理者或许可接受径流被高估的情况（例如基础设施脆弱性评估），但这会严重干扰鱼类种群存续性的估算结果。

为解决模型性能不佳导致Vflo模型模拟衍生指标失真的问题，我们提出三项可行方案：1）将径流-生态关系限定于梅雷迪思湖下游的加拿大河干流范围；2）将评估限制在按月时间步长聚合的径流数据（尽管通常这类数据与生态过程的匹配度欠佳）；3）聚焦于预测精度较高的径流指标，例如部分点位的径流量大小、出现时刻与持续时长。

为最大化加拿大河流域内可用于构建径流-生态响应模型的潜在解释变量与调查点位数量，并最小化偏差与不确定性，综合运用上述三类方法或为合理选择。若要开展具备有效统计效力的径流-生态关系研究，最佳路径是将整理后的鱼类数据（详见处理流程）与现有水文站实测数据（gage data）相结合，以优化生态关系的构建工作。