This dataset shows sub-areas (sub-catchments) for West Africa and the estimated hydropower potential as well as several other interesting attributes for hydro power development. This dataset is not intended for local studies but only for regional comparison. The dataset was created using the following methodology: 1. Creation of the sub-catchment boundaries using the river network and the Hydrosheds 15s flow direction grid. Outlet points of sub-catchments were defined as the last cell point of river reaches, where the accumulated upstream catchment size exceeds 3000 km². In coastal areas the threshold was reduced to 1000 km². Outlet points were manually adjusted at large reservoirs. 2. Water balance and climate change variables were joined into GIS from the simulation outputs of a Fortran water balance model. Climate change scenario simulation results are based on the CORDEX Africa ensemble (RCP4.5 and RCP8.5). 3. Theoretical hydropower potential data were aggregated to sub-area values from detailed simulation results at the river network. 4. Attractive regions were identified using a classification system based on hydropower potential. 5. Hydropower plant types are based on a rough classification system using river network data, including mean annual discharge, seasonality in discharge and river slope. Local site studies may result in different suitable plant types. 6. Preferred machine types are based on a rough classification system using river slope and plant type. The following tools were used for creating this dataset: • ArcGIS 9.2: main GIS tool • ArcView 3.1: specific tasks with large attribute tables where more recent GIS versions fail • Fortran: main processing tool for various tasks o Pre-processing of GPCC and satellite precipitation data o Water balance modelling o Sub-catchment outlets definition • MS Excel: some data pre-processing and visualization • Libre Office: dbf file manipulation • CDO: Climate Data Operators for processing of CORDEX-Africa climate model data • Shell scripts: For automatic file processing of climate model data • Batch scripts: For automatic calls to Fortran programs For each sub-area the following attributes are available (units in brackets): • NB: ID number of sub-area • AREA: Local size (km²) of sub-area • PRECIP_Y: Mean annual precipitation (mm) in the period 1998-2014 • ETA_Y: Mean annual actual evapotranspiration (mm) simulated for the period 1998-2014 • RUNOFF_Y: Mean annual runoff (mm) simulated for the period 1998-2014 • TEMP_Y: Mean annual air temperature (°C) in the period 1998-2014 • P_2035_P25: Change in future mean annual precipitation in % (2026-2045 vs. 1998-2014) for the lower quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • P_2035_P50: Change in future mean annual precipitation in % (2026-2045 vs. 1998-2014) for the median projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • P_2035_P75: Change in future mean annual precipitation in % (2026-2045 vs. 1998-2014) for the upper quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • P_2055_P25: Change in future mean annual precipitation in % (2046-2065 vs. 1998-2014) for the lower quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • P_2055_P50: Change in future mean annual precipitation in % (2046-2065 vs. 1998-2014) for the median projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • P_2055_P75: Change in future mean annual precipitation in % (2046-2065 vs. 1998-2014) for the upper quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • E_2035_P25: Change in future mean annual actual evapotranspiration in % (2026-2045 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • E_2035_P50: Change in future mean annual actual evapotranspiration in % (2026-2045 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • E_2035_P75: Change in future mean annual actual evapotranspiration in % (2026-2045 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • E_2055_P25: Change in future mean annual actual evapotranspiration in % (2046-2065 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • E_2055_P50: Change in future mean annual actual evapotranspiration in % (2046-2065 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • E_2055_P75: Change in future mean annual actual evapotranspiration in % (2046-2065 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • R_2035_P25: Change in future mean annual runoff in % (2026-2045 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • R_2035_P50: Change in future mean annual runoff in % (2026-2045 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • R_2035_P75: Change in future mean annual runoff in % (2026-2045 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • R_2055_P25: Change in future mean annual runoff in % (2046-2065 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • R_2055_P50: Change in future mean annual runoff in % (2046-2065 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • R_2055_P75: Change in future mean annual runoff in % (2046-2065 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • T_2035_P25: Change in future mean annual air temperature in °C (2026-2045 vs. 1998-2014) for the lower quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • T_2035_P50: Change in future mean annual air temperature in °C (2026-2045 vs. 1998-2014) for the median projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • T_2035_P75: Change in future mean annual air temperature in °C (2026-2045 vs. 1998-2014) for the upper quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • T_2055_P25: Change in future mean annual air temperature in °C (2046-2065 vs. 1998-2014) for the lower quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • T_2055_P50: Change in future mean annual air temperature in °C (2046-2065 vs. 1998-2014) for the median projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • T_2055_P75: Change in future mean annual air temperature in °C (2046-2065 vs. 1998-2014) for the upper quartile projection of 30 climate model runs in the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • POWER: Theoretical hydropower potential (MW) for the period 1998-2014 (total of all river reaches located in the sub-area) • POW_MINI: Theoretical hydropower potential (MW) for pico/micro/mini hydropower plants (< 1 MW installed capacity) for the period 1998-2014 • POW_SMALL: Theoretical hydropower potential (MW) for small hydropower plants (1-30 MW installed capacity) for the period 1998-2014 • POW_MEDIUM: Theoretical hydropower potential (MW) for medium/large hydropower plants (>30 MW installed capacity) for the period 1998-2014 • ATT_MINI: Region with theoretical hydropower potential that is attractive (0: no, 1: yes) for pico/micro/mini hydropower plants (< 1 MW installed capacity) • ATT_SMALL: Region with theoretical hydropower potential that is attractive (0: no, 1: yes) for small hydropower plants (1-30 MW installed capacity) • ATT_MEDIUM: Region with theoretical hydropower potential that is attractive (0: no, 1: yes) for medium/large hydropower plants (> 30 MW installed capacity) • PLANT_TYP1: Region suitable (0: no, 1: yes) for hydropower plant type 1 (run-of-river without diversion) • PLANT_TYP2: Region suitable (0: no, 1: yes) for hydropower plant type 2 (run-of-river with diversion) • PLANT_TYP3: Region suitable (0: no, 1: yes) for hydropower plant type 3 (storage without diversion) • PLANT_TYP4: Region suitable (0: no, 1: yes) for hydropower plant type 4 (storage with diversion) • TURBINE: Preferred turbine type (text) • PT_2035_25: Change in future hydropower potential in % (2026-2045 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PT_2035_50: Change in future hydropower potential in % (2026-2045 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PT_2035_75: Change in future hydropower potential in % (2026-2045 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PT_2055_25: Change in future hydropower potential in % (2046-2065 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PT_2055_50: Change in future hydropower potential in % (2046-2065 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PT_2055_75: Change in future hydropower potential in % (2046-2065 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PL_2035_25: Change in future hydropower potential in % (2026-2045 vs. 1998-2014) of local rivers (having their source in the same sub-area) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PL_2035_50: Change in future hydropower potential in % (2026-2045 vs. 1998-2014) of local rivers (having their source in the same sub-area) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PL_2035_75: Change in future hydropower potential in % (2026-2045 vs. 1998-2014) of local rivers (having their source in the same sub-area) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PL_2055_25: Change in future hydropower potential in % (2046-2065 vs. 1998-2014) of local rivers (having their source in the same sub-area) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PL_2055_50: Change in future hydropower potential in % (2046-2065 vs. 1998-2014) of local rivers (having their source in the same sub-area) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PL_2055_75: Change in future hydropower potential in % (2046-2065 vs. 1998-2014) of local rivers (having their source in the same sub-area) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5)

This dataset shows the river network for West Africa and the estimated hydropower potential as well as several other interesting attributes. After detailed calculation, the attribute values were rounded to a reasonable accuracy. This dataset is not intended for local studies but only for regional comparison. The dataset was created using the following methodology: 1. The river network was derived from the Hydrosheds 15s dataset: a) A threshold of 10 cells in the flow accumulation grid was used to delineate river reaches. b) Reaches with zero long-term mean annual discharge (rounded at 0.000 m³/s) were removed. c) Some remaining reaches in desert areas were manually removed. d) Small tributary reaches in reservoir water bodies were manually removed. e) The topology between adjacent reaches was computed from nodes and the flow accumulation grid (to determine which of the adjacent reaches is located downstream). 2. Discharge was computed with an annual water balance model using precipitation (TRMM) and potential evapotranspiration (CRU) as inputs. The water balance model was calibrated with observed discharge data of about 400 gauges (Global Runoff Database). The last cellpoint at the river network before a confluence with a downstream reach at 15s Hydrosheds resolution was used to query discharge from the discharge grid. The monthly distribution of discharge was computed by superimposing a specific seasonality derived from observed discharge data. At some rivers (Niger, Bani, Yobe) diversions and evaporation losses were roughly considered. 3. Computation of slope for each river reach: a) Compute length of reaches. b) Aggregate Hydrosheds unconditioned DEM (based on SRTM DEM) from 3s resolution to 15s resolution. After extensive testing (also with ASTER DEM) the 10% percentile of the 3s elevation values within the 15s box was selected as aggregation value. c) Extract upstream and downstream elevation for each reach from the aggregated DEM (see previous step). d) Adjust upstream and downstream reach elevation values with the following method: i. Adjustment of elevation in downstream direction by carving through barriers. ii. Adjustment of elevation in upstream direction by filling sinks. iii. Compute average of previous two steps. e) Smoothing of elevation along the river network. Differentiate between main rivers and tributaries in the smoothing process. Several rounds of smoothing with different settings were applied. f) Consider artificial reservoirs (e.g. Lake Volta) by ensuring zero slope in reservoir water bodies. 4. Slope was computed from length and elevation difference. 5. Gross theoretical hydropower potential (attribute POWER) was computed from mean annual discharge (Q_YEAR) and elevation difference in reach (ELEV_DIFF). Specific hydropower potential (POWER_SPEC) was computed from POWER and reach length. 6. Preferred hydropower plant size is based on a classification scheme using specific hydropower potential (attribute POWER_SPEC) and mean annual discharge (attribute Q_YEAR). The classification scheme is based on an analysis of existing hydropower plants in West Africa and some general considerations for hydropower plant design. 7. River names were manually assigned by comparison with the following data sources: a) GRDC gauges b) Nigeria JICA gauges c) Sierem GIS dataset d) Google maps e) Open Streetmap f) Michelin map (online version) g) Travelmag map (analog version) h) Various other sources (Wikipedia, detailed local maps, etc.) 8. For rivers forming international borders the country entries in the attribute table were manually assigned. The manual editing was required because the GIS datasets (river network and countries) do not align fully accurate at country borders. 9. Future change in annual discharge was computed by running the water balance model using the delta-change method for future precipitation and air temperature. The climate change projections were extracted from 30 Regional Climate Model (RCM) runs included in the CORDEX-Africa ensemble for the representative concentration pathways RCP4.5 and RCP8.5. Future change in potential evapotranspiration was estimated using an empirical relationship between air temperature and potential evapotranspiration. The following tools were used for creating this dataset: • ArcGIS 9.2: main GIS tool • ArcGIS 10.0: Python scripting and model building • ArcView 3.1: specific tasks with large attribute tables where more recent GIS versions fail • GDAL: automatic merging of DEM tiles • Fortran: main processing tool for various tasks o Pre-processing of GPCC and satellite precipitation data o Water balance modelling o Spatial aggregation of DEMs from 3s to 15s resolution o Adjustment of reach elevation (fill sinks, carve through barriers, smoothing) • MS Excel: some data pre-processing and visualization • Libre Office: dbf file manipulation • CDO: Climate Data Operators for processing of CORDEX-Africa climate model data • Shell scripts: For automatic file processing of climate model data • Batch scripts: For automatic calls to Fortran programs For each reach of the river network the following attributes are available (units in brackets): • ARCID: ID number of reach • TOARCID: ID number of next downstream reach • FROMARCID: ID number of dominant upstream reach (largest inflow) • NB: ID number of sub-area • RIVER: River name (English) • RIVER_FREN: River name (French) • COUNTRY_1: Country (ISO code) • COUNTRY_2: Second country (ISO code) if reach forms international border • AREA_Km2: Total upstream catchment area (km²) of reach • LENGTH_Km: Length (km) of reach • EXRIVER: Flag indicating external river originating from another sub-area (0: local river, 1: external river) • ELEV_US_m: Elevation (m) at upstream end of reach • ELEV_DS_m: Elevation (m) at downstream end of reach • ELEV_DIFF: Elevation difference (m) in reach • SLOPE: Slope (m/m) of reach • PWR_MW: Theoretical hydropower potential (MW) for the period 1998-2014 • PWR_MW_KM: Specific hydropower potential (MW/km) for the period 1998-2014 • QYEAR_m3_s: Mean annual discharge (m³/s) simulated for the period 1998-2014 • QJAN_m3_s: Mean monthly discharge (m³/s) 1998-2014 in January • QFEB_m3_s: Mean monthly discharge (m³/s) 1998-2014 in February • QMAR_m3_s: Mean monthly discharge (m³/s) 1998-2014 in March • QAPR_m3_s: Mean monthly discharge (m³/s) 1998-2014 in April • QMAY_m3_s: Mean monthly discharge (m³/s) 1998-2014 in May • QJUN_m3_s: Mean monthly discharge (m³/s) 1998-2014 in June • QJUL_m3_s: Mean monthly discharge (m³/s) 1998-2014 in July • QAUG_m3_s: Mean monthly discharge (m³/s) 1998-2014 in August • QSEP_m3_s: Mean monthly discharge (m³/s) 1998-2014 in September • QOCT_m3_s: Mean monthly discharge (m³/s) 1998-2014 in October • QNOV_m3_s: Mean monthly discharge (m³/s) 1998-2014 in November • QDEC_m3_s: Mean monthly discharge (m³/s) 1998-2014 in December • Q_2035_P25: Change in future mean annual discharge in % (2026-2045 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • Q_2035_P50: Change in future mean annual discharge in % (2026-2045 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • Q_2035_P75: Change in future mean annual discharge in % (2026-2045 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • Q_2055_P25: Change in future mean annual discharge in % (2046-2065 vs. 1998-2014) for the lower quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • Q_2055_P55: Change in future mean annual discharge in % (2046-2065 vs. 1998-2014) for the median simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • Q_2055_P75: Change in future mean annual discharge in % (2046-2065 vs. 1998-2014) for the upper quartile simulation using 30 climate model runs of the CORDEX-Africa ensemble (RCP4.5 and RCP8.5) • PLANT_SIZE: Preferred hydropower plant size (0: none, 1: <1MW, 2: 1-30MW, 3: >30MW installed capacity) • LAT: Latitude (decimal degrees North) at end of reach • LON: Longitude (decimal degrees East) at end of reach