科研之友
微信
新浪微博
Facebook
分享链接
摘要
Most studies on evaluating the potential in developing seasonal to interannual hydroclimatic forecasts have focused on associating low-frequency climatic conditions with basin-level precipitation/streamflow. The motivation of this study is to provide an understanding on how land surface characteristics modulate the low-frequency (interannual to decadal) variability in precipitation to develop low-frequency signal in streamflow. For this purpose, we consider basins with minimum anthropogenic impacts over southeastern United States and apply Singular Spectrum Analysis (SSA), a data-driven spectrum analysis tool, on annual precipitation and streamflow time series for detecting the dominant frequencies and for estimating the associated variability with them. Hypothesis test against an AR(1) process is carried out via Monte Carlo SSA for detecting significant (at 90% confidence level) low-frequency oscillations. Thus, the study investigates how the observed low-frequency oscillations in precipitation/streamflow vary over the southeastern United States and also their associations with climatic conditions. For most study basins, precipitation exhibits higher low-frequency oscillations than that of streamflow primarily due to reduction in variability by basin storage. Investigating this further, we found that the percentage variance accounted by low-frequency oscillations in streamflow being higher for larger basins which primarily indicates the increased role of climate and basin storage. To develop a fundamental understanding on how basin storage controls the low-frequency oscillations in streamflow, a simple annual hydrological model is employed to explore how the given low-frequency signal in precipitation being modified under different baseflow index conditions and groundwater residence time. Implications of these analyses relating to streamflow predictions and model calibration are also discussed.
- 出版日期2015-2
