PNNL

Abstract

We would like to extend our sincere gratitude to the discusser for his insightful and positive feedback on the next-generation intensity-duration-frequency (NG-IDF) technique, which is designed to improve upon the traditional precipitation-based IDF methods, especially in snow-dominated regions (Yan et al., 2018, 2019, 2023a; Sun et al., 2022). The discusser's emphasis on seasonal analysis underscores the importance of capturing peak floods occurring in different seasons, driven by distinct hydrologic processes. As the discusser noted in Yan et al. (2024), the American River Basin (ARB) experiences peak floods in distinct seasons. Winter floods often result from rain-on-snow (ROS) events, where precipitation exceeds snow accumulation, typically driven by atmospheric rivers (ARs). In contrast, spring floods are predominantly caused by snowmelt, with minimal precipitation involvement. The discusser also pointed out that using seasonal analysis—such as deriving NG-IDF curves based on water available for runoff (W) during the warm season—can yield larger values than those based on annual maximum data, providing more conservative (i.e., higher) flood estimates for hydrologic design. While we agree that seasonal analysis can be valuable in regions with distinct seasonal flood peaks, such as the ARB, it falls outside the scope of Yan et al. (2024). Our paper focuses on validating the NG-IDF technique at the basin scale using observational data and compares results with standard methods, such as the United States Geological Survey (USGS) Bulletin 17C (England et al., 2018), which rely on annual maximum flood data. However, we acknowledge the importance of further research on the NG-IDF technique and recognize that several approaches could enhance its application in basins like the ARB. One such approach, as the discusser suggests, is seasonal analysis. Another possibility is employing physics-based hydrologic models to explore the mechanisms of flood generation or W events, followed by mixed-population frequency analysis. For instance, recent studies have explored these topics: Yan et al. (2023b) classified W events into rainfall, ROS, and snowmelt mechanisms, developing separate hyetographs for each; Sun et al. (2024) classified peak floods into four mechanisms and analyzed the impacts of climate change on both flood magnitudes and their underlying mechanisms; Barth et al. (2019) conducted a process-driven flood frequency analysis using a weighted mixed-population approach, distinguishing between AR-driven and non-AR floods. We believe the real challenge, however, is not solely about enhancing NG-IDF curves through seasonal or mixed-population approaches. The real challenge lies in bridging the gap between advances in hydrological science and actionable information that can be effectively used by the engineering community. The NG-IDF method attempts to provide a potential solution by offering a framework to incorporate snowmelt into traditional IDF design with minimal additional training required for practitioners. Could the method be further improved? Absolutely. Is there an alternative approach? Certainly. We fully support the use of distributed physics-based hydrological models to facilitate process-driven flood frequency analysis, especially in the context of climate change. However, there remains significant technology transition challenges such as lack of guidance in local regulations and high-computational and labor costs to implement. We believe addressing these challenges requires fostering stronger collaboration between the climate, hydrology, and engineering communities.