WAVELET ANALYSIS OF TIME SERIES AT RAINFALL AND FLOW STATIONS IN THE UPSTREAM TORTIYA WATERSHED (Northern Ivory Coast)

The objective of this study is to analyze the links between climate indices and the variability of precipitation and flow series. In order to better understand the non-stationarity of the different stations, the flow and rainfall data used concern the Tortiya station (1960-1996) and the Lafigué station (1977-1996). The climate indices coupled to these series are the NAO (North Atlantic Oscillation) and ENSO (El Niño - Southern Oscillation) over the same study period. The methodology of this work consisted in applying wavelet analysis and wavelet coherence on the different time series. These methods highlighted the different modes of variability occurring in the time series, namely : the sub-annual mode (< 1 year), the annual mode (1 year) and the interannual mode (1-2 years ; 2-4 years; 4-8 years). Firstly, the results of the present analysis show that the variability of the signal is explained at high frequencies (6 months to 1 year) in the different time series. At this frequency, it is the annual mode (1 year) that records all the signal variability between 30% and 70%. On the other hand, this work presents signals on other frequencies and periods but fairly, especially 2-year frequency, in the 1990s at the Lafigué station (8.2%). Wavelet analysis revealed that the dominant signal is very broadly significant  at the annual cycle level. In addition, the use of wavelet coherence between climate indices (ENSO, NAO) and precipitation, flow indicates a strong influence of NAO on rainfall and flow series.

Key words : Wavelet, coherence, periodicity, variability, upstream Tortiya watershed