Past and future hydroclimatic variability over West and Central Africa and their teleconnections

  • Moussa Sidibe

    Student thesis: Doctoral ThesisDoctor of Philosophy

    Abstract

    This thesis examined past hydroclimatic trends and variability across West and Central Africa from 1950 to 2005,and their interactions with both catchment properties and large-scale climate patterns. The understanding gained from this study was then used to assess the impact of near-term climate change on hydrological regimes. Two imputation methods were successfully applied to assemble a new reconstructed streamflow dataset and time series analyses revealed high non-stationarity in annual streamflow, with two step-changes occurring at the regional scale in 1970 and 1993 respectively. The results also indicated good agreement between precipitation and streamflow fluctuations from one decade to another. Moreover, detailed analysis of streamflow variability modes highlighted significant interannual to multi-decadal fluctuations which were found to be associated with large-scale climate variability and modulated by catchment physical properties. Building upon the detected streamflow-sea surface temperature teleconnections, a multi-timescale linear regression model was built and applied in combination with two hydrological models (GR2M and IHACRES) to examine the potential impacts of climate change on hydrological systems by the mid-21st century. The results highlighted a zonal contrast in future precipitation between western (dry) and eastern (wet) Sahel, and a robust signal in rising temperature, suggesting an increase in potential evapotranspiration, which are likely to induce a slight significant increase in discharge (~+5%) at the regional scale. More importantly, the findings indicated that uncertainties in streamflow predictions inherent to models and gridded observational datasets quality in Central Africa, could be narrowed by the teleconnections-based regression model.
    Date of AwardAug 2019
    Original languageEnglish
    Awarding Institution
    • Coventry University
    SupervisorBastien Dieppois (Supervisor), Damian Lawler (Supervisor) & Jonathan Eden (Supervisor)

    Keywords

    • data imputation
    • hydroclimatic trends and variability
    • teleconnections
    • hydrological regimes
    • climate change
    • bias-correction
    • West and central Africa

    Cite this

    '