Predicting climate-driven changes in reservoir inflows and hydropower in Côte d'Ivoire using machine learning modeling

Salomon Obahoundje, Arona Diedhiou, Komlavi Akpoti, Kouakou Lazare Kouassi, Eric Antwi Ofosu and Didier Guy Marcel Kouame

Energy, 2024, vol. 302, issue C

Abstract: This study investigates the impact of climate change and variability on reservoir inflow and hydropower generation at three key hydropower plants in Côte d'Ivoire including Buyo, Kossou, and Taboo. To simulate inflow to reservoir and energy generation, the Random Forest (RF), a machine-learning algorithm allowing fewer input variables was applied. In three-step, RF k-fold cross validation (with k = 5) was used; (i) 12 and 6 multiple lags of precipitation and temperature at monthly increments were used as predictors, respectively; (ii) the five most important variables were used in addition to the current month's precipitation and temperature; and (iii) a residual RF was built. The bias-adjusted ensemble mean of eleven climate models output of the COordinated Regional Downscaling Experiment was used for the representative concentration pathways (RCP4.5 and RCP8.5). The model output was highly correlated with the observations, with Pearson correlations >0.90 for inflow and >0.85 for energy for the three hydropower plants. The temperature in the selected sub-catchments may increase significantly from 0.9 to 3 °C in the near (2040–2069) and from 1.7 to 4.2 °C in far (2070–2099) future periods relative to the reference period (1981–2010). A time series of precipitation showed a change in range −7 and 15 % in the near and −8 to 20 % in the far future and more years are with increasing change. Depending on the sub-catchment, the magnitude of temperature and precipitation changes will increase as greenhouse gas emissions (GHG)(greater in RCP8.5 than RCP4.5) rise. At all time scales (monthly, seasonal, and annual), the simulated inflow and energy changes were related to climate variables such as temperature and precipitation. At the annual time scale, the inflow is projected to change between −10 and 37 % and variability may depend on the reservoir. However, the energy change is promised to change between −10 and 25 %, −30 to 15 %, and 5–40 % relative to the historical (1981–2010) period for Taabo, Kossou, and Buyo dams, respectively at an annual scale. The changes may vary according to the year, the RCPs, and the dam. Consequently, decision-makers are recommended to take into consideration an energy mix plan to meet the energy demand in these seasons.

Keywords: Time series change; Climate variability and change; Hydropower generation; Machine learning; Random forest; Côte d’Ivoire (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544224016220
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:302:y:2024:i:c:s0360544224016220

DOI: 10.1016/j.energy.2024.131849

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().