Abstract Estimation of reliable design discharges under variable climate is a key challenge for today’s engineers. Therefore, researchers are intensively exploring different alternative approaches in order to improve standard methods for design discharge estimation. Paper investigates the performance of time-invariant, time trend- and precipitation-informed models based on generalized extreme value (GEV) distribution for 31 Slovenian discharge gauging stations with data availability from 1961 until 2015. Different rainfall durations are used as covariates in the case of precipitation-informed models. The selected catchments are located in different climate regions and characterized by five flow regimes. The results indicate that in most cases precipitation-informed models gave better fit to the measured data comparing to time-invariant and time trend-informed models. Relative differences in the design discharge estimations associated with 10- and 100-year return periods using time trend- and precipitation-informed models compared to time-invariant model were up to 60%. Additionally, the results indicate that identified best-fitting model of individual gauging station can to some extent be related to its flow regime.

中文翻译：

---

用于评估可变气候下设计排放量的时间趋势模型和降水模型比较

摘要 可变气候下可靠设计排放的估算是当今工程师面临的主要挑战。因此，研究人员正在深入探索不同的替代方法，以改进设计排放估算的标准方法。论文研究了基于广义极值 (GEV) 分布的时不变、时间趋势和降水知情模型的性能，这些模型具有 1961 年至 2015 年的数据可用性，适用于 31 个斯洛文尼亚流量监测站。在这种情况下，使用不同的降雨持续时间作为协变量降水信息模型。选定的集水区位于不同的气候区域，具有五种流态。结果表明，在大多数情况下，与时不变模型和时间趋势模型相比，降水模型更适合测量数据。使用时间趋势和降水知情模型与时不变模型相比，与 10 年和 100 年重现期相关的设计排放估计的相对差异高达 60%。此外，结果表明，确定的单个计量站的最佳拟合模型可能在一定程度上与其流态有关。