Chaotic Analysis of Hydro-meteorological Variables for Monsoon Season in the Savitri River Basin, India

Hydrometeorological variables are complex and inter-related, varying both spatially and temporally over a region. Identification of the complexity of these variables and their interactions is important for the assessment and management of water resources in a region. The present study examines the complexity of hydrometeorological variables in the Savitri River basin in India. Specifically, daily rainfall, temperature, evaporation, and relative humidity time series are analysed by employing the false nearest neighbour (FNN) algorithm, a nonlinear dynamic method. The FNN method is a dimension-based method and represents the complexity of the time series. In this study, the variables are analyzed independently, in a single-variable sense, and the phase space is reconstructed using the single-variable time series. Two different delay time values are considered for phase space reconstruction: (a) delay time value equal to 1; and (b) delay time value obtained from the the average mutual information (AMI) method. The results indicate that the FNN dimensions for the four series range from 4 to 15, suggesting different complexity levels for the four variables. The delay time value is found to have some significant influence on the FNN dimensions, with delay time equal to 1 generally yielding low dimensions (less than or equal to 7) and delay time from the AMI method generally resulting in high dimensions (equal to or more than 10).