Deterministic chaotic dynamics in soil moisture across Nebraska

• Published in: Journal of hydrology (Amsterdam) Vol. 578; p. 124048
• Main Authors: Di, Chongli, Wang, Tiejun, Istanbulluoglu, Erkan, Jayawardena, A.W., Li, Siliang, Chen, Xi
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2019
• Subjects: Chaos; Correlation dimension; Deterministic; Local approximation prediction; Nebraska Mesonet; Soil moisture; Chaos; Local approximation prediction; Deterministic; Soil moisture; Correlation dimension; Nebraska Mesonet
• ISSN: 0022-1694; 1879-2707
• DOI: 10.1016/j.jhydrol.2019.124048

•The existence of chaotic behaviors of soil moisture was investigated.•The complexity of soil moisture systems generally decreased with increasing depth.•Dominant controls on soil moisture dynamics at different depths were different.•Soil moisture dynamics were more controlled by climate forcings at shallow depth.•Soil moisture dynamics were more controlled by soil texture at deep depth. Soil moisture systems generally exhibit complex spatiotemporal patterns due to their nonlinear interactions with surrounding environments. In this paper, the possible existence of chaotic behaviors in soil moisture time series was investigated using the chaos theory, specifically based on the techniques of phase space reconstruction, correlation dimension (CD), largest Lyapunov exponent (LLE), and local approximation prediction (LAP) methods. Daily soil moisture data for six years along with different influencing factors were obtained for 35 Nebraska Mesonet (NM) stations with varying hydroclimatic conditions and soil properties. The results of CDs and LLEs revealed the presence of chaos in soil moisture systems, implying that soil moisture dynamics was controlled by deterministic systems with a limited number of governing variables. The results obtained by the LAP method offered further evidence on the chaotic behavior of soil moisture systems. Specifically, Median of CD values of 4.51, 4.06, 3.23, and 1.40 were obtained for soil moisture time series at the depths of 10, 25, 50, and 100 cm, respectively, indicating that the complexity of soil moisture systems decreased with increasing depth. Moreover, the correlations of CDs with different influencing factors showed that the dominant controls on soil moisture dynamics switched at different depths. At shallower depths (e.g., 10 and 25 cm), CDs were more correlated with meteorological forcings, while stronger correlations emerged between CDs and soil texture in deeper soil layers (e.g., 50 and 100 cm). As the first attempt to analyze the complexity of soil moisture systems from the perspective of the chaos theory, the results of this study offer additional avenues for understanding chaotic behaviors of soil moisture dynamics.