Frequency and magnitude variability of Yalu River flooding: Numerical analyses for the last 1000 years

Accurate determination of past flooding characteristics is necessary to effectively predict future flood disaster risk and the dominant controls. However, understanding the role of environmental forcing on past flooding frequency and magnitude is difficult due to the deficiency of observations and too short measurement time series. Here, a numerical model HydroTrend, that generates synthetic time series of daily water discharge at a river outlet, is applied to Yalu River to: (1) reconstruct annual peak discharges over the past 1000 years and estimate flood annual exceedance probabilities; (2) identify and quantify the impacts of climate change and human activity (runoff yield induced by deforestation and dam retention) on the flooding frequency and magnitude. Climate data obtained from meteorological stations and ECHO-G climate model output, morphological characteristics (hypsometry, drainage area, River length, slope and Lapse rate) and hydrological properties (groundwater properties, canopy interception effects, cascade reservoirs retention effect and saturated hydraulic conductivity) are form the significant reliable model inputs. Monitored for decades and some proxies on ancient floods allow for accurate calibration and validation of numerical modeling. Simulations match well present-day monitored data (1958–2012) and historical flood events literature records (1000–1958). They indicate that flood frequencies of Yalu River increased during AD 1000–1940, followed by a decrease until the present day. Frequency trends were strongly modulated by climate variability, particularly by intensity and frequency of rainfall events. The magnitudes of larger floods, events with a return period of 50 to 100 years, increased by 19.1 and 13.9 %, respectively, due to climate variability over the last millennium. Anthropogenic processes were found to either enhance or reduce flooding, depending on the type of the human activities. Deforestation increased the magnitude of larger floods by 19.2–20.3 %, but the construction of cascade reservoirs in AD 1940 significantly reduced their magnitude by 36.7 to 41.7 %. We conclude that under intensified climate change and human activity in the future, effective river engineering should be considered, particularly for small and medium-sized mountainous river systems, which are at higher risk of flood disasters due to their relatively poor capacity for hydrological regulation.