Study of temporal streamflow dynamics with complex networks: network construction and clustering
- PDF / 3,796,284 Bytes
- 17 Pages / 595.276 x 790.866 pts Page_size
- 74 Downloads / 189 Views
ORIGINAL PAPER
Study of temporal streamflow dynamics with complex networks: network construction and clustering Nazly Yasmin2 • Bellie Sivakumar1,2 Accepted: 7 November 2020 Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Applications of the concepts of complex networks for studying streamflow dynamics are gaining momentum at the current time. The present study applies a coupled phase space reconstruction–network construction method to examine the clustering property of the temporal dynamics of streamflow. The clustering of the temporal streamflow network is determined using clustering coefficient, which quantifies the tendency of a network to cluster (a measure of local density). Monthly streamflow time series observed from each of 639 stations (i.e. 639 networks) in the United States are studied. The presence of links between nodes (i.e. phase space reconstructed vectors) in each streamflow network (i.e. station) is identified using the Euclidean distance. Different distance thresholds are used to examine the influence of threshold on the clustering coefficient results and to identify the critical threshold. The results indicate that the distance threshold has significant influence on the clustering coefficient values of the temporal streamflow networks. With the critical distance threshold values, the clustering coefficients for the 639 stations are found to be between 0.15 and 0.81, suggesting very different types of network connections and dynamics. The clustering coefficient values are found to provide useful information on the influence of a given month (i.e. timestep) of the year on the temporal dynamics. Reliable interpretations of the clustering coefficient values in terms of catchment characteristics and flow properties are also possible. Keywords Streamflow Temporal dynamics Complex networks Nonlinear dynamics Clustering coefficient Distance threshold
1 Introduction Adequate understanding of the temporal (and spatial) dynamics of streamflow is important for a wide range of purposes in hydrology and water resources, including for forecasting of floods and droughts and estimation of soil erosion. The past decades have witnessed the development and applications of numerous scientific concepts and
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00477-020-01931-9) contains supplementary material, which is available to authorized users. & Bellie Sivakumar [email protected] 1
Department of Civil Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India
2
UNSW Water Research Centre, School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052, Australia
mathematical techniques to study the temporal dynamics of streamflow. In recent years, applications of the concepts of complex networks (A network is a set of points, called nodes, connected by a set of lines, called links) to study the temporal dynamics of streamflow have been gaining momentum (e.g., Braga et al
Data Loading...
