A hybrid factorial stepwise-cluster analysis method for streamflow simulation – a case study in northwestern China

Abstract

In this study, a hybrid factorial stepwise-cluster analysis (HFSA) method is developed for modelling hydrological processes. The HFSA method employs a cluster tree to represent the complex nonlinear relationship between inputs (predictors) and outputs (predictands) in hydrological processes. A real case of streamflow simulation for the Kaidu River basin is applied to demonstrate the efficiency of the HFSA method. After training a total of 24 108 calibration samples, the cluster tree for daily streamflow is generated based on a stepwise-cluster analysis (SCA) approach and is then used to reproduce the daily streamflows for calibration (1995-2005) and validation (2008-2010) periods. The Nash-Sutcliffe coefficients for calibration and validation are 0.68 and 0.65, respectively, and the deviations of volume are 1.68% and 4.11%, respectively. Results show that: (i) the HFSA method can formulate a SCA-based hydrological modelling system for streamflow simulation with a satisfactory fitting; (ii) the variability and peak value of streamflow in the Kaidu River basin can be effectively captured by the SCA-based hydrological modelling system; (iii) results from 2 6 factorial experiments indicate that not only are minimum temperature and precipitation key drivers of system performance, but also the interaction between precipitation and minimum temperature significantly impacts on the streamflow. The findings are useful in indicating that the streamflow of the study basin is a mixture of snowmelt and rainfall water.

A Statistical Hydrological Model for Yangtze River Watershed Based on Stepwise Cluster Analysis

“…According to the five statistical coefficients, both the two schemes yielded acceptable simulation in all three stations. This result is consistent with previous studies (Fan et al, 2015;Li et al, 2015;Fan et al, 2016;Zhuang et al, 2016) which indicated that stepwise cluster analysis can provide reliable and efficient flow prediction. In the calibration period, measured and simulated monthly stream flows have a good match using the two schemes.…”

“…stepwise cluster analysis has received much attention for environmental issues such as air quality prediction (Huang, 1992), process control (Huang et al, 2006), climate projections (Wang et al, 2013), stream flow prediction (Cheng et al, 2016;Zhuang et al, 2016), groundwater simulation (Han et al, 2016), and ecosystem analysis and prediction (Sun et al, 2018). This previous researcher has indicated that the stepwise cluster analysis approach can characterize environmental processes with complicated nonlinear and dynamic relationships and provide satisfactory predictions.…”

A Stepwise-Clustered Simulation Approach for Projecting Future Heat Wave Over Guangdong Province

“…The framework of the developed approach is presented in Figure 1. SCA was introduced by Huang (1992) and has been widely used in pieces of environmental research, such as climate change, air pollution, process control, and hydrology prediction (Huang et al, 2006;Wang X. Q. et al, 2013;Li et al, 2015;Wang and Huang, 2015;Fan et al, 2016;Zhuang et al, 2016;Sun et al, 2019;Zhai et al, 2019;Duan et al, 2020;Ren et al, 2021b;Wang et al, 2021). These studies have indicated the SCA has satisfactory performance in projecting environmental processes with complex non-linear and dynamic relationships.…”

“…Previously, statistical downscaling methods mostly assume that the interested dependent variable is a function of independent variables (Wilby et al, 2002;Hessami et al, 2008;Gibson et al, 2017), while the improvements of the quality of downscaled simulation reproduced through such a function are likely to be limited relative to the raw GCMs' output owning to the complex climate system (Wilby et al, 2002;Wang X. Q. et al, 2013). However, the underlying complex relationships (including non-linear and discrete relationships) between dependent and independent variables can be revealed by the generated cluster tree without assuming the functional relationship compared to these downscaling methods (Wang X. Q. et al, 2013;Zhuang et al, 2016;Duan et al, 2020). Handling the complicated relationship among multiple dependent and independent variables at the same time is another significant advantage of SCA (Fan et al, 2015;Wang et al, 2021).…”

Ensemble Learning Enhanced Stepwise Cluster Analysis for River Ice Breakup Date Forecasting

“…Nevertheless, besides its calculation complexity, the performance of the SCA is sensitive to its inputs and internal parameters; the difference within leaf clusters of a SCA tree is usually not well described. The SCA has been applied to various water resources and environmental management problems, such as urban air quality prediction (Huang, 1992), lung cancer diagnosis (Ren et al, 1997), waste treatment process simulation (Sun et al, 2009;Sun et al, 2011), groundwater bioremediation optimization (Huang et al, 2006;Qin et al, 2007;He et al, 2008b;Wang et al, 2012;Zhao et al, 2017), open water forecasting (Fan et al, 2015;Li et al, 2015;Han et al, 2016;Zhuang et al, 2016b;Cheng et al, 2016;Fan et al, 2017;) and climate model downscaling (Wang et al, 2013;Zhuang et al, 2016a;Zhai et al, 2019). However, few applications of SCA to river ice forecasting are reported.…”