Research papersInvestigating anthropically induced effects in streamflow dynamics by using permutation entropy and statistical complexity analysis: A case study

- We investigated the dynamics of São Francisco streamflow, Brazil.
- We applied permutation entropy and statistical complexity.
- Original and deseasonalized series show different dynamical patterns.
- Streamflow dynamical change could be attributed to Sobradinho dam.
- ENSO co-induced effects were detected in the streamflow dynamics change.

In this paper we investigated the influence of the construction of Sobradinho dam on daily streamflow of São Francisco river, Brazil, using permutation entropy method. We analyzed a long daily streamflow time series recorded during the period 1929-2010 encompassing the construction of Sobradinho dam between 1973 and 1979. We found that the original and deseasonalized streamflow time series are characterized by clear different complexity and entropy patterns before the construction of the dam; while, after it, their degree of randomness and complexity are nearly identical. Furthermore, investigating the oscillatory behavior of the entropy and complexity time variation, the periodicity of 3.67 years was identified, identical to one of the main periodicities revealed in the Multivariate ENSO Index (MEI). Such finding confirms the close relationship between streamflow dynamics and ENSO phenomenon. After the construction of the dam, the time variation of entropy and complexity changes almost abruptly toward stochastic regime characterized by higher entropy and lower complexity. Although the dam operations could be considered responsible for such abrupt dynamical change in the streamflow, we cannot exclude the presence of a co-induced ENSO effect; in fact, the analysis of MEI shows a strikingly similar and concomitant change in the long-term trend, identified by using the singular spectrum analysis.