1. SummaryThe Congo basin’s Groundwater Storage Anomaly (GWSA) datasets are generated by Benjamin M. Kitambo, Sly Wongchuig, Raphael M. Tshimanga, Adrien Paris, Alejandro Blazquez, Daniel Moreira, Frederic Frappart, Ayan Santos Fleischmann, Jean-Marie O. Kileshye , Mohammad J. Tourian, Fabrice Papa in the article entitled "Hydrogeological control of groundwater variations in the Congo River Basin revealed by GRACE water storage change decomposition", Water Resources Research (submitted).The dataset was generated using the decomposition method of Terrestrial Water Storage Anomaly (TWSA) which is the sum of the contributions of different hydrological storage compartments including changes in surface water storage anomaly (ΔSWSA), root zone soil moisture (ΔSMroot), snow storage (ΔSnow), and GWSA (ΔGWSA) (Frappart and Ramillien, 2018). The equation is as follows:                                                            ΔTWSA = ΔSWSA + ΔSMroot + ΔSnow + ΔGWSA  (1)ΔGWSA changes are obtained by subtracting ΔSWSA, ΔSMroot, ΔSnow from ΔTWSA on the same period of availability for all datasets, here 2002-2015. Since there is no contribution of snow storage in the Congo tropical region, this term is neglected in our study. Only GWSA datasets are provided from different Gravity Recovery and Climate Experiment (GRACE) solutions including Jet Propulsion Laboratory spherical harmonics (JPL-SH), JPL mass concentration blocks mascons (JPL-M), GeoforschungsZentrum Potsdam SH (GFZ-SH), CSR mascons (CSR-M), CNES-L3. To get ΔTWSA ,ΔSWSA, ΔSMroot datasets, please kindly refer to the aforementioned paper where all links are provided.2. Name Description●  Month_data_XX: Months of availability of GWSA data due to the missing values in GRACE’s TWSA XX (where XX stands for CNES-L3, Others which include JPL-SH, JPL-M, GFZ-SH, CSR-M).●  GWSA_XX: monthly groundwater storage anomaly variations from XX (where XX stands for JPL-SH, JPL-M, GFZ-SH, CSR-M, and CNES-L3 GRACE solutions).3. File Description The GWSA estimates from the decomposition method of TWSA provides the monthly variation of groundwater storage changes over 2002-2015, gridded on 0.25-degree of spatial resolution, each pixel covering almost 773 km². The files are organized in matrix:●  First column represents the latitude in degree.●  Second column represents the longitude in degree.●  From the third column: GWSA data in km³, there are 139 (respectively 147) columns representing each month of data availability of the period from 2002 to 2015 from CNES-L3 GRACE solution (respectively JPL-SH, JPL-M, GFZ-SH, CSR-M GRACE solutions).

1. SummaryThe Congo basin’s Groundwater Storage Anomaly (GWSA) datasets are generated by Benjamin M. Kitambo, Sly Wongchuig, Raphael M. Tshimanga, Adrien Paris, Alejandro Blazquez, Daniel Moreira, Frederic Frappart, Ayan Santos Fleischmann, Jean-Marie O. Kileshye , Mohammad J. Tourian, Fabrice Papa in the article entitled "Hydrogeological control of groundwater variations in the Congo River Basin revealed by GRACE water storage change decomposition", Water Resources Research (submitted).The dataset was generated using the decomposition method of Terrestrial Water Storage Anomaly (TWSA) which is the sum of the contributions of different hydrological storage compartments including changes in surface water storage anomaly (ΔSWSA), root zone soil moisture (ΔSMroot), snow storage (ΔSnow), and GWSA (ΔGWSA) (Frappart and Ramillien, 2018). The equation is as follows:                                                            ΔTWSA = ΔSWSA + ΔSMroot + ΔSnow + ΔGWSA  (1)ΔGWSA changes are obtained by subtracting ΔSWSA, ΔSMroot, ΔSnow from ΔTWSA on the same period of availability for all datasets, here 2002-2015. Since there is no contribution of snow storage in the Congo tropical region, this term is neglected in our study. Only GWSA datasets are provided from different Gravity Recovery and Climate Experiment (GRACE) solutions including Jet Propulsion Laboratory spherical harmonics (JPL-SH), JPL mass concentration blocks mascons (JPL-M), GeoforschungsZentrum Potsdam SH (GFZ-SH), CSR mascons (CSR-M), CNES-L3. To get ΔTWSA ,ΔSWSA, ΔSMroot datasets, please kindly refer to the aforementioned paper where all links are provided.2. Name Description●  Month_data_XX: Months of availability of GWSA data due to the missing values in GRACE’s TWSA XX (where XX stands for CNES-L3, Others which include JPL-SH, JPL-M, GFZ-SH, CSR-M).●  GWSA_XX: monthly groundwater storage anomaly variations from XX (where XX stands for JPL-SH, JPL-M, GFZ-SH, CSR-M, and CNES-L3 GRACE solutions).3. File Description The GWSA estimates from the decomposition method of TWSA provides the monthly variation of groundwater storage changes over 2002-2015, gridded on 0.25-degree of spatial resolution, each pixel covering almost 773 km². The files are organized in matrix:●  First column represents the latitude in degree.●  Second column represents the longitude in degree.●  From the third column: GWSA data in km³, there are 139 (respectively 147) columns representing each month of data availability of the period from 2002 to 2015 from CNES-L3 GRACE solution (respectively JPL-SH, JPL-M, GFZ-SH, CSR-M GRACE solutions).

The Congo River basin (CRB), the world's second-largest river system, is subject to extreme hydrological events that strongly impact its ecosystems and population. Here we present an unprecedented 38-year (1983-2020) hydrological reanalysis of CRB daily discharge that helped us to analyze the spatiotemporal dynamics of recent major CRB floods and droughts, and their teleconnection with El Niño Southern Oscillation (ENSO), the dominant driver of tropical precipitation.To develp this dataset, we employ a large-scale hydrologic-hydrodynamic model (MGB) with lake storage dynamics representation and a data assimilation (DA) technique using in-situ discharge and water level and remote sensing observations from radar altimetry.The performance of MGB is satisfactory, with KGE indices of 0.84 and 0.71 for calibration and validation, respectively. Incorporating lake representation significantly improves the simulations, from 0.3 to 0.63 for the Pearson correlation coefficient. Furthermore, DA shows a ~13% reduction in simulated discharge errors through cross-validation.This final version of the product uses all in-situ and radar altimetry observations during the simulation period, after satisfactory performance was verified in the DA cross-validation process. The original simulation period was from 1981 but the first two years were removed to avoid inconsistencies due to model initial conditions.  Please cite as:Wongchuig, S.; Kitambo, B.; Papa, F.; Paris, A.; Fleischmann, A.; Gal, L.; Boucharel, J.; Paiva, R.; Oliveira, R.; Tshimanga, R. M.; Calmant, S. 2023. Improved modeling of Congo's hydrology for floods and droughts analysis and ENSO teleconnections.

The Congo River basin (CRB), the world's second-largest river system, is subject to extreme hydrological events that strongly impact its ecosystems and population. Here we present an unprecedented 38-year (1983-2020) hydrological reanalysis of CRB daily discharge that helped us to analyze the spatiotemporal dynamics of recent major CRB floods and droughts, and their teleconnection with El Niño Southern Oscillation (ENSO), the dominant driver of tropical precipitation. To develp this dataset, we employ a large-scale hydrologic-hydrodynamic model (MGB) with lake storage dynamics representation and a data assimilation (DA) technique using in-situ discharge and water level and remote sensing observations from radar altimetry. The performance of MGB is satisfactory, with KGE indices of 0.84 and 0.71 for calibration and validation, respectively. Incorporating lake representation significantly improves the simulations, from 0.3 to 0.63 for the Pearson correlation coefficient. Furthermore, DA shows a ~13% reduction in simulated discharge errors through cross-validation. This final version of the product uses all in-situ and radar altimetry observations during the simulation period, after satisfactory performance was verified in the DA cross-validation process. The original simulation period was from 1981 but the first two years were removed to avoid inconsistencies due to model initial conditions.
Please cite as: Wongchuig, S.; Kitambo, B.; Papa, F.; Paris, A.; Fleischmann, A.; Gal, L.; Boucharel, J.; Paiva, R.; Oliveira, R.; Tshimanga, R. M.; Calmant, S. 2023. Improved modeling of Congo's hydrology for floods and droughts analysis and ENSO teleconnections.

The Congo River basin (CRB), the world's second-largest river system, is subject to extreme hydrological events that strongly impact its ecosystems and population. Here we present an unprecedented 38-year (1983-2020) hydrological reanalysis of CRB daily discharge that helped us to analyze the spatiotemporal dynamics of recent major CRB floods and droughts, and their teleconnection with El Niño Southern Oscillation (ENSO), the dominant driver of tropical precipitation. To develp this dataset, we employ a large-scale hydrologic-hydrodynamic model (MGB) with lake storage dynamics representation and a data assimilation (DA) technique using in-situ discharge and water level and remote sensing observations from radar altimetry. The performance of MGB is satisfactory, with KGE indices of 0.84 and 0.71 for calibration and validation, respectively. Incorporating lake representation significantly improves the simulations, from 0.3 to 0.63 for the Pearson correlation coefficient. Furthermore, DA shows a ~13% reduction in simulated discharge errors through cross-validation. This final version of the product uses all in-situ and radar altimetry observations during the simulation period, after satisfactory performance was verified in the DA cross-validation process. The original simulation period was from 1981 but the first two years were removed to avoid inconsistencies due to model initial conditions.

The Congo River basin (CRB), the world's second-largest river system, is subject to extreme hydrological events that strongly impact its ecosystems and population. Here we present an unprecedented 38-year (1983-2020) hydrological reanalysis of CRB daily discharge that helped us to analyze the spatiotemporal dynamics of recent major CRB floods and droughts, and their teleconnection with El Niño Southern Oscillation (ENSO), the dominant driver of tropical precipitation. To develp this dataset, we employ a large-scale hydrologic-hydrodynamic model (MGB) with lake storage dynamics representation and a data assimilation (DA) technique using in-situ discharge and water level and remote sensing observations from radar altimetry. The performance of MGB is satisfactory, with KGE indices of 0.84 and 0.71 for calibration and validation, respectively. Incorporating lake representation significantly improves the simulations, from 0.3 to 0.63 for the Pearson correlation coefficient. Furthermore, DA shows a ~13% reduction in simulated discharge errors through cross-validation. This final version of the product uses all in-situ and radar altimetry observations during the simulation period, after satisfactory performance was verified in the DA cross-validation process. The original simulation period was from 1981 but the first two years were removed to avoid inconsistencies due to model initial conditions.
Please cite as: Wongchuig, S.; Kitambo, B.; Papa, F.; Paris, A.; Fleischmann, A.; Gal, L.; Boucharel, J.; Paiva, R.; Oliveira, R.; Tshimanga, R. M.; Calmant, S. 2023. Improved modeling of Congo's hydrology for floods and droughts analysis and ENSO teleconnections.

The Congo River basin (CRB), the world's second-largest river system, is subject to extreme hydrological events that strongly impact its ecosystems and population. Here we present an unprecedented 38-year (1983-2020) hydrological reanalysis of CRB daily discharge that helped us to analyze the spatiotemporal dynamics of recent major CRB floods and droughts, and their teleconnection with El Niño Southern Oscillation (ENSO), the dominant driver of tropical precipitation.To develp this dataset, we employ a large-scale hydrologic-hydrodynamic model (MGB) with lake storage dynamics representation and a data assimilation (DA) technique using in-situ discharge and water level and remote sensing observations from radar altimetry.The performance of MGB is satisfactory, with KGE indices of 0.84 and 0.71 for calibration and validation, respectively. Incorporating lake representation significantly improves the simulations, from 0.3 to 0.63 for the Pearson correlation coefficient. Furthermore, DA shows a ~13% reduction in simulated discharge errors through cross-validation.This final version of the product uses all in-situ and radar altimetry observations during the simulation period, after satisfactory performance was verified in the DA cross-validation process. The original simulation period was from 1981 but the first two years were removed to avoid inconsistencies due to model initial conditions.  Please cite as:Wongchuig, S.; Kitambo, B.; Papa, F.; Paris, A.; Fleischmann, A.; Gal, L.; Boucharel, J.; Paiva, R.; Oliveira, R.; Tshimanga, R. M.; Calmant, S. 2023. Improved modeling of Congo's hydrology for floods and droughts analysis and ENSO teleconnections.

1. Summary The Congo basin’s Surface Water Storage (SWS) datasets are generated by Benjamin Kitambo, Fabrice Papa, Adrien Paris, Raphael M. Tshimanga, Frederic Frappart, Stephane Calmant, Omid Elmi, Ayan Santos Fleischmann, Melanie Becker, Mohammad J. Tourian, Rômulo A. Jucá Oliveira, Sly Wongchuig in the article entitled "A long-term monthly surface water storage dataset for the Congo basin from 1992 to 2015", Earth System Science Data (submitted). The dataset was generated using two methods, one based on a multi-satellite approach and one on a hypsometric curve approach. The multi-satellite approach consists of the combination of surface water extent (SWE) from the Global Inundation Extent from Multi-satellite (GIEMS-2) and satellite-derived surface water height (SWH) from radar altimetry (long-term series ERS-2_ENV_SRL) on the same period of availability for the two datasets, here 1995-2015. The hypsometric curve approach consists of the combination of SWE from GIEMS-2 dataset and hypsometric curves obtained from various digital elevation models (DEMs) (i.e., ASTER, ALOS, MERIT, and FABDEM). Both methods estimate monthly spatio-temporal variations of SWS changes across the entire Congo River basin. 2. Name Description HYPSO_XX: hypsometric curve providing the surface water extent area-elevation relationship from XX (where XX stands for ASTER, ALOS, MERIT, and FABDEM DEMs). HYPSO_CORR_XX: corrected hypsometric curve providing the surface water extent area-elevation relationship from XX (where XX stands for ASTER, ALOS, MERIT, and FABDEM DEMs). AREA_STOR_XX: hypsometric curve providing the surface water extent area-storage relationship from XX (where XX stands for ASTER, ALOS, MERIT, and FABDEM DEMs). SWS_XX: monthly surface water storage variations from XX (where XX stands for ASTER, ALOS, MERIT, FABDEM DEMs, and Multi-satellite approach). 3. File Description The SWS estimates from the multi-satellite approach (1995-2015), as well as the hypsometric curves providing the surface water extent area-elevation relationship from the four DEMs (before and after the corrections), the surface water extent area-storage relationship, along with the four SWS estimates (1992-2005). The dataset is gridded on equal-area of 0.25° spatial resolution at the equator, each pixel covers almost 773 km². The reference point for calculating the volume variation is the minimum of surface water extent for each pixel. The files are organized in matrix: First column represents the latitude in degree. Second column represents the longitude in degree. From the third column: data. In case of hypsometric curve, the data represents elevation in meter on 101 columns representing the increment of 1% flooding in each 773 km² pixel from GIEMS-2. In case of SWS data (in km³), there are 288 (respectively 252) columns representing each month of the period over 1992-2015 (respectively 1995-2015) from hypsometric curve approach (respectively multi-satellite approach).

1. Summary The Congo basin’s Surface Water Storage (SWS) datasets are generated by Benjamin Kitambo, Fabrice Papa, Adrien Paris, Raphael M. Tshimanga, Frederic Frappart, Stephane Calmant, Omid Elmi, Ayan Santos Fleischmann, Melanie Becker, Mohammad J. Tourian, Rômulo A. Jucá Oliveira, Sly Wongchuig in the article entitled "A long-term monthly surface water storage dataset for the Congo basin from 1992 to 2015", Earth System Science Data (submitted). The dataset was generated using two methods, one based on a multi-satellite approach and one on a hypsometric curve approach. The multi-satellite approach consists of the combination of surface water extent (SWE) from the Global Inundation Extent from Multi-satellite (GIEMS-2) and satellite-derived surface water height (SWH) from radar altimetry (long-term series ERS-2_ENV_SRL) on the same period of availability for the two datasets, here 1995-2015. The hypsometric curve approach consists of the combination of SWE from GIEMS-2 dataset and hypsometric curves obtained from various digital elevation models (DEMs) (i.e., ASTER, ALOS, MERIT, and FABDEM). Both methods estimate monthly spatio-temporal variations of SWS changes across the entire Congo River basin. 2. Name Description HYPSO_XX: hypsometric curve providing the surface water extent area-elevation relationship from XX (where XX stands for ASTER, ALOS, MERIT, and FABDEM DEMs). HYPSO_CORR_XX: corrected hypsometric curve providing the surface water extent area-elevation relationship from XX (where XX stands for ASTER, ALOS, MERIT, and FABDEM DEMs). AREA_STOR_XX: hypsometric curve providing the surface water extent area-storage relationship from XX (where XX stands for ASTER, ALOS, MERIT, and FABDEM DEMs). SWS_XX: monthly surface water storage variations from XX (where XX stands for ASTER, ALOS, MERIT, FABDEM DEMs, and Multi-satellite approach). 3. File Description The SWS estimates from the multi-satellite approach (1995-2015), as well as the hypsometric curves providing the surface water extent area-elevation relationship from the four DEMs (before and after the corrections), the surface water extent area-storage relationship, along with the four SWS estimates (1992-2005). The dataset is gridded on equal-area of 0.25° spatial resolution at the equator, each pixel covers almost 773 km². The reference point for calculating the volume variation is the minimum of surface water extent for each pixel. The files are organized in matrix: First column represents the latitude in degree. Second column represents the longitude in degree. From the third column: data. In case of hypsometric curve, the data represents elevation in meter on 101 columns representing the increment of 1% flooding in each 773 km² pixel from GIEMS-2. In case of SWS data (in km³), there are 288 (respectively 252) columns representing each month of the period over 1992-2015 (respectively 1995-2015) from hypsometric curve approach (respectively multi-satellite approach).