Modeling the impact of past and future land cover changes on a reservoir catchment hydrology in Semi-arid, Africa

doi:10.21203/rs.3.rs-1505575/v1

Download PDF

Research Article

Modeling the impact of past and future land cover changes on a reservoir catchment hydrology in Semi-arid, Africa

https://doi.org/10.21203/rs.3.rs-1505575/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Land-use change influences watershed hydrology and eventually, the sustainability of reservoirs. Understanding how reservoir catchment land use dynamic influences hydrology is imperative to devise effective land management and planning strategies for sustainable management of reservoir storage. This study evaluated the response of hydrological components of the Vea reservoir catchment (a sub-catchment of the White Volta Basin in West Africa) to past, present, and future land-use change scenarios (1985, 1996, 2006, 2020, 2030, 2040) using the Soil Water Assessment Tool (SWAT). The model performed satisfactorily with NSE values of 0.738 and 0.782 for calibration and validation respectively. The land cover change analysis revealed a continuous increase in farmland and built-up areas from 1986 to the projected 2040 resulting in increased Curved Numbers (CN). The model results revealed that an increase in CN from 81.26 to 86.06 between 1986 and projected 2040 resulted in a -7.8% and − 26.2% decrease in ET and base flow respectively, 46.9%, and 5.1% increase in surface runoff, and water yield respectively. Despite the general decline in base flow, the seasonal trend shows that base flow extends to January, indicating that the shallow aquifer store enough water that can be extracted for dry season gardening. While vegetation restoration is critical to reducing eroded sediment yield from the catchment to guarantee reservoir storage, constructing dams and dugouts within the catchment to harness the high surface runoff water for irrigation and livestock watering in the long dry season is also needful.

SWAT

Reservoir

Water balance

CA-Markov

future prediction

Vea

Water availability in an area is largely dependent on rainfall distribution, which is further distributed into various components including surface runoff into reservoirs, interflow, evapotranspiration, and percolation into groundwater (Kundu et al., 2017), and the proportions of these components are modified considerably by land-use changes in a watershed (Sajikumar and Remya, 2015). Meanwhile, Land use land cover changes (LULCC) have been tremendous over the past decade, the world over (Foley et al., 2005; Lambin and Meyfroidt, 2011). These changes come at a cost of reservoir sedimentation due to increased runoff, reduction in groundwater levels due to less percolation, and persistent alterations in geomorphological biodiversity and catchment energy balance (Gellis et al., 2006; Githui et al., 2009; Ristić et al., 2013; Paudel and Vetaas, 2014; Deng et al., 2015; Kundu et al., 2017; Zou et al., 2018). Therefore, the interconnection between land use and various components of a watershed system necessitates an understanding of how LULCC alters their spatial and temporal distribution (Chemura et al., 2020). Hence, a shift from the descriptive analysis of land-use change to a more quantitative analysis using diverse modeling tools is a new paradigm to land cover change studies. This is necessary to understand the interactions between variables that make up the complex watershed system (Sokolowski and Banks, 2011) to inform sustainable catchment planning and realistic management decisions (Deng et al., 2013).

Hydrological models enable us to present the real hydrological process in simplified systems (Tessema, 2011). Thus, serving as a mediator between the real world and theory (Babel and Karssenberg, 2013). Hydrological models can determine the watershed hydro-sedimentological reaction to the variability of land-use change through, the valuation of outcomes in surface streamflow and groundwater ensuing from temporal LULCC, and they are also capable of producing the spatial distribution of all water balance components (Magesh et al., 2012; Yan et al., 2013). Different hydrological models exist for quantifying water resources impact from LULCC. The Soil Water Assessment Tool (SWAT) is the most widely used due to its applicability to different regions of the globe as well as its interdisciplinary, comprehensive, and robust nature (Schilling et al., 2008; Deng et al., 2015; Kundu et al., 2017; Abbaspour et al., 2017). SWAT has been proven to have adequate potential to simulate the impact of land-use change on hydrological alteration the world over (Cornelissen et al., 2013). For example, Deng et al., (2015) examined the impact of LULCC on water and energy balance in Heihe river basin in China using the SWAT model and reported remarkable variations in evapotranspiration and runoff induced by LULCC. Júnior and Montenegro, (2019) also simulated the hydrological response of a repetitive watershed in the state of pernambuco, Brazil to reforestation and reported increased in ground water recharge and soil water storage to between 6% and 28%. In addition, Memarian et al., (2014) assessed the effect of land use change on hydrological components in hulu Langat watershed of Malasia and reported increase runoff, sediment load and ground water recharge.

In Africa, Woldesenbet et al., (2017) simulated the hydrological response of Lake Tanan and Beles watersheds in the Upper Blue Nile Basin, Ethiopia using the SWAT model. Their results revealed that expansion of farmlands alongside a decline in woody shrubs and woodlands led to an increase in surface runoff and a decline in actual evapotranspiration and groundwater components in both watersheds. Furthermore, Kidane et al., (2019) assessed the impact of climate change and LULCC on the Guder watershed, one of the headwaters of the Blue Nile basin, Ethiopia, and reported an 11.34% decline in streamflow in the dry season and 11.33–17.76% increase in rainy season flow and concluded that LULCC greatly influences watershed hydrology than Climate. De Jong Posthumus,( 2017) also simulated the water balance components of Tandjari reservoir in southeastern Burkinafaso and reported that SWAT was able to represent the reservoir water balance reasonably. Awotwi et al., (2019) likewise simulated water balance components' response to land-use change in the Pra River Basin of Ghana using the SWAT model and reported that a continuous increase in anthropogenic activities from 1986 to projected 2025 resulted in an increment of surface runoff and water yield and a reduction of base flow and evapotranspiration.

The findings of these studies confirm that adverse impacts are observed on hydrological regimes due to LUCC. Hence, evaluating the hydrological implications of LULCC is essential to inform practical policy planning and sustainable management of water and environmental resources (Napoli et al., 2017). Despite the significant impact of LULCC on runoff in tropical regions due to higher energy input and faster rate of change, very few studies have been conducted in that regard (de Paulo Rodrigues da Silva et al., 2018). Worse perhaps is that even though reservoir storage largely depends on runoff (Spence, 2007), reservoirs watersheds' hydrological response to LULCC has received far less research attention in West Africa. Against this backdrop, we aim to contribute to the limited literature by assessing the long-term past and future response of hydro-sedimentology to LULCC of a catchment with the reservoir downstream to it, greatly impacted by sedimentation. Specifically we, (i) Predict and evaluate the changes in LULCC up to 2040 in the reservoir watershed (ii) Examined the Water balance ratios in response to past and future LULCC scenarios, and (iii) Examined and critically analyzed the seasonal and spatiotemporal variations of hydrological components in response to LULCC scenarios in the Vea Catchment.

The Vea reservoir is located in the Upper East Region of Ghana between latitude 10.878133 and longitude − 0.838613. The reservoir provides water for municipal supply, and irrigation among others (Boateng et al., 2015). The study domain is a savanna ecological zone characterized by significant temperature variations and a unimodal, unpredictable rainfall pattern with three seasons; Rainy season (July to September), Dry season (February to May), and Hamatan season (November to January) (Barry, et al., 2005). The reservoir has a surface area of 385ha at full supply level (FSL) and a storage capacity of 17Mm³ (IDA, 1978). The reservoir is located in Ghana but its watershed extends to Burkinafaso with an area of 30500ha (Fig. 1). The main soils types found in the area include vertisols, luvisols, lithosols, and arenosoles and the average annual rainfall varies from 800mm to 1240 mm with the highest rainfall (more than 80%) experienced during the rainy season (July to September). The average monthly surface temperature ranges between 28°c and 39°c with the maximum temperature experienced in the dry season while the lowest is experienced in the Hamatan season. The Vea reservoir is challenged by a high sedimentation rate resulting in the chocking of flumes and a reduction in storage space (Adongo et al., 2019). Irrigation in the downstream end has also been affected by the chocked flumes and reduction in fresh water supplies.

The general workflow to execute the prediction in IDRISI-Terrset and analyze the impact of LULCC on water balance in the SWAT model is given in Fig. 2 below.

3.1 Future land cover map prediction

Historical Classified Landsat images of the Vea watershed for 1986, 1996, 2006 (Landsat 5 thematic mapper), and 2020 (Landsat 8 OLI/Tire) were employed for future predictions. Prediction and exploration of the land change dynamics were executed in TerrSet Geospatial Monitoring and Modeling system using the integrated cellular automata Markov (CA –Markov) model. The CA- Makov model analyzes land use classes of different periods and generates transitional probability matrixes, which denotes the number of pixels that are supposed to change from one class to another over a particular time unit. CA –Markov has been widely used in recent research to predict future land-use change and for other spatial events simulations (Sang et al., 2011; Wang et al., 2012; Mukhopadhyay et al., 2014; Al-sharif and Pradhan, 2014). The input files include historical LULC maps, suitability maps, and transition probability matrixes generated from the LULC maps (Eastman, 2012). The suitability maps used in this study include; Euclidian distance from the river, road, slope, and Elevation maps (Fig. 3). The high-quality land use prediction capabilities of the Markov chain integrated with the dynamic unambiguous spatial simulations of the CA model makes the CA-Markov efficient and effective when it comes to simulating spatiotemporal LULC patterns (Myint and Wang, 2006; He et al., 2008; Wang et al., 2012). Calibration and validation of predicted images is carried out by comparing the classified image of 2020 with simulated image of same year and future decadal predicted maps for 2030 and 2040 were generated as input maps in SWAT. The area and probability statistics are used to analyze the change in the future

3.2 Swat Model For Assessment Of Variation In Water Balance And Reservoir Components

The current hydrological condition and future LULCC scenarios for the Vea watersheds was simulated using SWAT version 2012. SWAT is a versatile watershed model developed by the US. Department of Agriculture (USDA), Agriculture research service (ARS)(Arnold and Fohrer, 2005), capable of simulating many processes including Agro-hydrologic water quality and plant growth among others (Gassman et al., 2007). It is a high-efficiency model and very appropriate for analyzing the impact of land-use changes (Kiros et al., 2015). During simulations, watersheds are partitioned into sub-watersheds, which are further delineated by hydrologic response units (HRU’s) that consist of homogeneous soil, land use, slope, and management characteristics (Arnold and Fohrer, 2005; Her et al., 2015). HRU is created by overlaying soil, land-use, and slope classes within each sub-basin. Water or pollutant routing is simulated in the model from the HRU’s to the watershed and then through the stream network to the entire watershed main outlet. The Vea was delineated into 15 sub-basins and further sub-divided into 156 HRU’s.

At each HRU, SWAT simulates hydrology using the water balance equation below (Sajikumar and Remya, 2015)

SW_{t =}SW_o +${\sum }_{t=i}^{t}{P}_{t}-{Q}_{sf}-{E}_{tc}-{Q}_{pe}-{Q}_{rt})----------(5)$

Where t is the time (day), SW_t is the amount of water content at day, SW₀ represents the initial amount of water content at day i, P_t is the amount of precipitation on day i, Q_sf gives the amount of surface runoff, E_t is evapotranspiration, Q_pe gives the percolation and Q_rt is the amount of return flow.

ArcSWAT uses two categories of data thus, spatial data inputs and temporal data. The spatial data includes DEM, Soil, and LULC and the temporal impute data sets consist of climate (temperature, rainfall, relative humidity, wind speed, and ETo), Hydrological Data (discharge), and reservoir data (reservoir sedimentation and storage). Details of the parameters used in the SWAT model are given in Table 1 below.

Table 1
Acquired SWAT model input data with an indication of the type of dataset, period of datataset and the Source of data set.
Dataset	Parameter	Year	Source
Elevation (Topography)	ASTER DEM	NA	United states geological survey (https://earthexplorer.usgs.gov/)
Land cover	Historical	1986, 1996, 2006 and 2020	United states geological survey (https://earthexplorer.usgs.gov/)
Land cover	Future	2030, 2040	Predicted using CA-Markov model
Gauge data	Water discharge	NA	West Africa Science Service Center on Climate Change And Adapted Land Use (WASCAL)
Soil	Soil map		Council for Scientific and Industrial Research (CSIR), Ghana
Climate	Rainfall, temperature relative humidity, and wind	2000–2018	Ghana Metrological Agency (GMET)

Table 2

Input parameters of Vea reservoir used in the SWAT model for Reservoir simulations.
Parameter	Description	Vea
MORES	Month in which reservoir becomes operational	2
IYRES	Year in which it becomes operational	1985
RES_EVOL(10^4m³)	Volume required to fill the reservoir to Emergency Spill Way	19.2
RES_ESA (ha)	Surface area for RES_EVOL	550
RES_PVOL(10^4m³)	Principal Spill Way Volume	17.3
RES_PSA (ha)	Surface area at RES_PVOL	405ha
RES_VOL(10^4m³)	Initial volume (volume on day reservoir becomes operational)	13.05
RES_SED (mg.l^− 1)	Initial sediment concentration	15.48904
RES_NSED(mg.l^− 1)	Equilibrium sediment concentration	15.48904
RES_D₅₀ (µm)	Median sediment Diameter	12.82
RES_K (mm.h^− 1)	Reservoir infiltration rate	0.25
EVRSV	Evaporation Co-efficient	0.6
IRESCO	Method of Reservoir Release	0
RES-RR (m³.s^− 1)	Maximum discharge of the principal spillway
OFLOWMN_FPS	Minimum discharge as a fraction of RES_PVOL	0
Source; Irrigation Company of the Upper Region (ICOUR), Ghana; 2020 bathymetric survey

With sediment grain size of clay, silt, and sand obtained from (Adongo et al., 2019), the median sediment particle diameter was computed from Eq. 6 (SWAT input-output document 2012)

$${d}_{50}=\text{e}\text{x}\text{p}\left(0.41\times \frac{{m}_{c}}{100}+2.71\times \frac{{m}_{silt}}{100} +5.7\times \frac{{m}_{s}}{100}\right) 6$$

Where d₅₀ is the median particle size of sediments (µ), in the surface soil layers m_c is the percent clay in the surface soil layer, m_silt is percent silt and m_s is percent sand.

For IRESCO the method of average annual release rate for uncontrolled reservoirs was chosen because it has the most appropriate to represent the open nature of the spillway in these dams. This method requires RES_RR, which is the maximum discharge release downstream when water exceeds the volume of the spillway. For both reservoirs, the mean of estimated values of water discharged between 2000 and 20118 was used. The default value of 0.6 was used for EV-RSV and an arbitrary value of 0.25mmd⁻, assigned for RES-K given the amplitude in the model from 0 to 1mm-day. Given the non-regulation of this dam and the absence of discharge for most months in the dry season (November to May), 0 was used for the OFLOW_MNFPS value. The monthly time series of maximum and minimum outflows, monthly targeted storage, and monthly consumptive water use were also obtained from ICOUR.

3.2. 1 Model Calibration, Validation, and Accuracy assessment

Though hydrological modeling of catchments with streamflow observation is generally challenged by uncertainties in every input data including LULC, hydro climatological data, and the DEM, Calibration significantly improves the performance of models (Merz and Blo¨schl, 2004). Calibration involves adjusting model parameter ranges until the best model performance is achieved (Jeffrey G. Arnold, 2012) while validation involves, using the range of calibrated parameters together with a separate period but the same kind of observed dataset in another alteration to authenticate the model. SWATCUP was used in the model calibration and validation. There are 5 calibration methods in SWAT CUP but the Sufi-2 method was used In this study because according to Wu and Chen, (2015) and Molina-navarro et al., (2016), the SUFI-2 method provides good accuracy, has the ability to represent the observed data with small uncertainties and require a lower number of runs to achieve a satisfactory result. The warm-up period was two years and the period of calibration and validation of discharge was 2013–2014 and 2015 respectively. Calibration and validation results are shown in Fig. 7.

Accuracy assessment evaluates the reliability of the simulated results to represent reality. The accuracy of the model to reproduce measured discharge was analyzed statistically by three error matrixes; RMSE root mean square error (Legates and McCabe, 1999; Singh et al., 2005)) Nash-Sutcliffe Coefficient ((Nash and Sutcliffe, 1970), and P-BIAS ( Gupta et al., 1999). The lower the RMSE the better the model (Moriasi et al., 2007). These matrixes are expressed by equations 1–4

NSE$=1-\frac{{\sum }_{k=1}^{n}{\left({Y}_{obs,k}-{Y}_{sim,k}\right)}^{2}}{{\sum }_{k=1}^{n}{\left({Y}_{obs,k}-{Ȳ}_{obs}\right)}^{2}} \left(1\right)$

PBIAS$=100\times \frac{{\sum }_{k=1}^{n}({Y}_{obs,k}-{Y}_{sim,k})}{{\sum }_{k=1}^{n}{Y}_{obs,i}} \left(2\right)$

RSR$=\frac{RMSE}{STDV}=\frac{{\sum }_{k=1}^{n}\sqrt{{\left({Y}_{obs,k}-{Y}_{sim,k}\right)}^{2}}}{{\sum }_{k=1}^{n}\sqrt{{\left({Y}_{obs,k}-{ȳ}_{sim,k}\right)}^{2}}} \left(3\right)$

Where n is the number of observed data, Y_obs,k is the observed streamflow on month k, Ȳ_obs is the average observed streamflow, Y_sim,k is the streamflow simulated on month k and Ȳ_sim is the average simulated streamflow.

NSE values range from -∞ to 1. 1 signifies a perfect agreement between predicted and observed while 0 indicates non-agreement. The model outputs are judged satisfactorily when NSE is higher than 60% and PBIAS is in the interval − 25% to + 25% over the calibration and validation period Moriasi et al., (2007). Also, the smaller the root mean square error, the better the model.

3.2.2 Hydrological Dynamics Under Various Lulcc Scenarios

To understand the effect of different scenarios of land use land cover change on Tono and Vea reservoirs’ watershed hydrology, the calibrated SWAT model which used1986 LULC map as a baseline was subsequently run for three historical LULC (1996, 2006, 2020) and two projected (2030 and 2040) scenarios for the same climate period of 2000 to 2018 using the “fix-changing” method. This method involves varying LULC maps while keeping all other input parameters (Soil, DEM, Reservoir, and Weather data set) constant in the calibrated model. This method has been used widely in many countries such as the upper san Pedro watershed in Mexico (Nie et al., 2011); Upper Du watershed, china (Yan et al., 2013); Mula and Mutha rivers’ catchment, India (Wagner et al., 2013), Andasa watershed in the Blue Nile basin, Ethiopia (Gashaw et al., 2018); Dano catchment, Burkina Faso (Yira et al., 2016). The outcome of the simulations was then used to quantify the impact of LULCC on the reservoir and its watershed hydrology by comparing the mean annual basin values for each land use scenario. Pearson correlation was further used to establish the connection between LULC variables and hydrological components

4.1 Land use prediction and Land cover Assessment

4.1.1 Markov model validation to assess the accuracy of prediction

CA-Makov was used to simulate 2030 and 2040 future scenarios from classified land use maps. The transition area and probability matrix are created according to the1996 and 2016 classified land use maps based on the suitability atlas that has already been created (Fig. 3). The modeling results show a very strong similarity between simulated and classified images (Fig. 4). Furthermore, the predictive power was validated by comparing the results map for the 2020 simulation with the classified LULCC for the same year using kappa statistics. Generally, the validation results show a reasonable consistency between the observed and simulated with an overall spatial accuracy of 72%. Hence, this method can be used to predict the results for 2030, 2040 since kappa statistics between 0.6–0.88 and 0.81-1 means substantial and almost perfect agreement between simulated and actual images (Feinstein A.R. and Cicchetit V.D., 1990). The predicted 2030 and 2040 images were input in ArcSWAT to assess the response of hydrological variables to possible land-use changes in the future.

4.1.2 Historical And Future Land-use Change Dynamics

The conversion rate and trend of land use in the Vea catchments are depicted in Fig. 5. The statistics of the change map indicate a general decline in savannah forest and savannah woodlands and an increase in farmlands and built areas between 1986 and 2020. A marginal increase is observed for water within the same period. These changes are inevitable, anticipated to increase in the future as the drivers of change (population) are expected to grow (GSS, 2012)), resulting in a similar pattern in the future scenarios 2030 and 2040 with the most prevalent changes being derived by farmlands.

There is generally an increase in farmlands of about 6.0 km² within the first decade of the study period (1986–1996). The surge in agriculture development was remarkable between 2006 and 2020 (Fig. 6). This could partly be because of the wider gap of 14 years compared to the 10-year interval of the previous years. Projecting into the future, farmland is expected to increase by 34.89km² and 49.41km² between 2020–2030 and 2030–2040 respectively. This is validated by Ghana’s agriculture sector development policy, planting for food, and jobs program that resulted in a shift from subsistence farming to commercial agriculture (Martey et al., 2012). The program involves subsidizing farm inputs to encourage commercialization, to provide job opportunities for the youth and reduce rural-urban migration. In addition, built-up areas have seen a gradual increase of about 1.3km² between 1986 and 1996 and are projected to increase by 7.3km² between 2030 and 2040. The consequences of anthropogenic activities (agricultural and built area expansion) is glaring on natural resources including savannah forest, savannah woodlands, and water resources, which is projected to reduce considerably by -20.1 km², -14.2 km^2, and − 3.3 km² respectively between 2020–2030 as well as -21.3 km², -15.2 km² and − 4.3 km² respectively between 2030 and 2040. From the statistics, it is obvious that the reduction in the Vea catchment forest is largely attributable to the consistent increase in farmlands. A shift in land use toward farmlands and built areas at the disadvantage of forests are also reported in other studies; Gebremicael et al., (2013) in the Upper Blue Nile basin, Ethiopia, Ottinger et al., (2013) in the yellow river, China, Guan et al., (2011) in Saga, Japan, and Opoku et al.,( 2019) in southern Ghana. These changes have implications beyond just the area numbers and in the subsequent section, we discuss the implication of these changes in land use on water balance components and sedimentation in the basin.

4.3 Hydrologic Parameter Sensitivity Analysis

Sensitivity analysis determines the variables that when changed slightly can stimulate quick runoff generation. Sensitivity analysis is important since it decides whether suitable data can be attained to give realistic model output (Baker and Miller, 2013). The model parameters identified to be sensitive during the calibration process are presented in Table 3. Among them, the most sensitive parameters representing surface runoff, soil character, and groundwater are CN_mgt, SURLAG_hru, Soil AWC_sol, and ESCO_hru. Some of these parameters were also observed to be sensitive in the works of Guug et al., (2020) in the Sherigu Catchment of the White Volta basin, and Kundu et al., (2017) in a part of the Narmada river basin, India.

Table 3

Sensitive parameters adjusted for the calibration of the SWAT model: Note, r- Multiply default values of the parameter in ArcSWAT by (1 + bestfited value). V - Replace the default value of the parameter in ArcSWAT with the best-fitted value. The final values in the table are for the 1986 Scenario.
Parameter	Parameter description	Range	SWAT default	fitted Value	Method	Final value
Cn2.mgt	Scs runoff curve number	35–98	78.61	0.015	r	79.79
SOL_AWC.sol	Available water capacity of the soil layer	0–1	0.13	-0.19	r	0.105
GW_DELAY.gw	Ground water delay (days)	0-500	31	34.1	v	34.1
GW_REVAP.gw	Ground water revap to occur (mm)	0.02–0.2	0.02	0.0279	v	0.0279
REVAPMN.gw	Threshold depth of water in the shallow aquifer for revamp to occur (mm)	0-500	750	830	v	830
SURLAG.bsn	Surface runoff lag time	0.05-24	5.8	-0.907	r	0.54
GWQMN.gw	Threshold depth of water in the shallow aquifer required for return flow to occur	0-5000	1000	1205.3	v	1205.3
ESCO.hru	Soil evaporation compensation factor	0–1	0.95	-0.48	r	0.49
ALPHA BE.gw	Base flow alpha factor (days)	0–1	0.01	0.012	V	0.012

4.3 Swat Model Calibration And Validation

The performance of the model was verified by comparing the simulated streamflow with the measured streamflow. The sensitive parameters used for calibration in SWAT CUP are listed in table (3) together with their value ranges and fitted values. Though the model captures the rising and recession limbs of the hydrograph (Fig. 6), the peak flow was generally underestimated by 22.5% and 29.2% for calibration and validation periods respectively (Table 4). These deviations can be attributed to the extremely low rains in the Uppers East Region between 2014 and 2015 as well as the deficiency of SWAT in representing extreme flows occasionally encountered in the catchment process (Golmohammadi et al., 2014). Notwithstanding, the performance statistics indicate a good agreement between simulated and observed streamflow with NSE values of 0.738 and 0.782 for calibration and validation respectively (Table 4). The performance statistics of the model range from good to very good according to the performance criterion proposed by Moriasi et al., (2007). Hence, the SWAT model could be said to have performed satisfactorily and can be used for further analysis.

Table 4

SWAT Model performance statistics for monthly streamflow simulations

Calibration (2013–2014)

Validation (2015)

Simulated flows

Performance rating

Simulated flows

Performance rating

NSE PBIAS RSR

0. 738

-22.5

0.48

Very good

Good

0. 782

-29.2

0.45

Very good

Good

4.4 Hydrological Dynamics In Response To Lulc Change Scenarios

This section analyzes the hydrological and sedimentation effects of LULCC at the basin scale and sub-basin scale. At the basin scale, rainfall is partitioned into different water balance components and how these components vary at the annual and seasonal levels is examined.

4.4.1 Water balance ratios to mean annual rainfall under different LULC scenarios

The calibrated SWAT model was deployed to simulate the hydrological response of the Vea reservoir catchment to six LULC scenarios (1986, 1996, 2006, 2020, 2030, and 2040). The mean annual simulated water balance components for each scenario as a proportion of the catchment’s mean annual rainfall are given in Table 5. The annual average rainfall for the climate period used for model simulation and analysis (2000–2018) is 940mm. Precipitation is partitioned into various components in a watershed including baseflow, runoff, and evapotranspiration among others (Tang and Wang, 2017). The majority of rainfall in the Vea catchment is lost to evapotranspiration (ET) with the ratio of precipitation to ET ranging from 0.73 in 1986 and projected to decline marginally to 0.68 in 2040. Apart from water lost to ET, the remaining total amount of water available including surface runoff and groundwater is termed water yield (Ayivi and Jha, 2018). Water yield constitutes about 32% of the total rainfall (940mm) in 1986 and is projected to increase marginally to 35% by 2040, Out of which surface runoff accounts for about 7% and 16% in 1986 and projected 2040 while base flow consists of 16.9% and 10.2% of rainfall in same period respectively (Table 5). Percolation is simulated to account for the lowest proportion between 9% in1986 and 6% in 2040.

The results obtained from this study are affirmed by previous studies within the Volta Basin. For instance, using a spreadsheet-based soil water balance model (WaSIM-ETH), Martin, (2006) simulated the water budget of Atankwidi, which is a sub-catchment of the white volt basin adjacent to the Vea watershed. With an annual average rainfall of 990 mm for the 1961–2001 simulation period, he reported that 11–20% of annual rainfall goes into the surface runoff, 2–13% goes into groundwater recharge and 63–82% of the annual rainfall is lost to ET. Similarly, Ibrahim et al., (2015) determined the water balance for the Vea catchment from water budget modeling using GR2M MODEL and found about 74.6% of precipitation going into ET processes with surface runoff and recharge being 11.9% and 12.9% of the annual rainfall. Also, Larbi et al., (2020) simulated water balance components of the Vea catchment from 1993 to 2017 and found that 74.3% of mean annual rainfall (954mm) is lost to ET while 8.6% is lost to surface runoff. Guug et al., (2020) also reported that 70.02–73.9% of annual average rainfall (844.1mm) in the Sherigu catchment of the white volta basin which covers Northern Ghana and Southern Burkina Faso is lost to evapotranspiration while base-flow contributes 13.7%-11% and surface runoff consist of 25%-9.3%.

Table 5

Ratio of precipitation to hydrological components in Vea catchment. ET-Evapotranspiration, PERC-Percolation, GWQ-Baseflow, SURF-Surface runoff, WYLD-Water yield
WBC	1986	1996	2006	2020	2030	2040
CN	79.79	80.24	80.96	81.26	83.83	86.06
ET	0.731	0.728	0.718	0.707	0.698	0.681
PERC	0.096	0.093	0.089	0.081	0.070	0.058
GWQ	0.168	0.160	0.151	0.138	0.124	0.102
SURF	0.072	0.081	0.090	0.108	0.130	0.159
WYLD	0.323	0.327	0.331	0.338	0.343	0.351

4.4.2 Mean Annual And Percentage Change In Hydrological Components

The mean annual and percentage of the basin’s parameters response tested (surface runoff, water yield, percolation, base flow, and ET) are given in Fig. 7 and table 6. The impact of land-use change indicates a general increase in weighted CN values (Table 5) and an associated increase in water yield, and surface runoff as well as a consequent decline in Evapotranspiration (Fig. 7, table 6).

The average weighted CN values for the Vea catchment increased rapidly from 79.79(1986) to 86.06 (2040), attributable to the significant forest degradation due to the increasing urbanization and agricultural land expansion. Linking surface runoff variation in Vea catchment with their respective LULC scenarios, the relatively higher curve numbers corresponded to a decrease in the ability of the soil to retain rainfall, and hence produce much more runoff. A remarkable increase in surface runoff is expected from122.23mm in 2030 to 149.75mm in 2040 representing 22.5% while the least increase in runoff from 76.7mm to 85.3mm (11%) was observed between 1996 and 2006 (Fig. 7VA and Table 6). The overall projected increase in surface runoff from 2020 to 2040 is 46.9%. The excess runoff in the catchment is attributable to the increase in imperviousness associated with settlement expansion to the disadvantage of the Savanah forest. This is confirmed by Cerdà et al., (2017) findings that, the expansion of cultivated lands and the reduction of natural vegetation would be considered, triggers for the generation of substantial runoff and soil erosion. Gashaw et al., (2018) also reported increasing surface runoff resulting from an increase in agriculture and built areas as well as a decrease in forested areas. Many studies show that impervious surfaces associated with urban expansion alongside loose soils from farmland expansion, lead to more water flowing through runoff networks and increased eroded sediment yield. Thereby, resulting in a higher sedimentation impact on reservoirs downstream (Kondolf et al., 2014). Hence, the high sedimentation rate in the Vea reservoir and other reservoirs within the catchment, widely reported in the literature (Adwubi et al., 2009; Adongo et al., 2016; Adongo et al., 2019) may likely be linked to the excessive runoff in the catchment.

The increase in surface runoff tends to reduce percolation and for that matter contributes to the decline in groundwater as base flow and percolation in the Vea catchment declined from − 5% and − 3.6% between 1986 and 1996 to a projected 16.9% and 17.5% respectively between 2030 and 2040 (Table 6). Notwithstanding, the impact of decreasing percolation may not be felt on groundwater supply because of the existing high groundwater storage in Northern Ghana coupled with the general insignificant abstraction rate due to the prevalent Fluoride concentration underlying the geological terrains of the Bongo area located in the Vea catchment (Agyekum and Asare, 2016). It is worth noting that, because water yield comprises surface runoff, vertical flows, and groundwater (Ayivi and Jha, 2018), the increasing effect of surface runoff is neutralized by the decreasing effect of base flow and percolation, resulting in a rather marginal increase in water yield from 1.4% in 1986 to a projected 2.8% in 2040(table 6). The lowest increase in water yield is observed from 303.1 (1986) to 307.4 (1996) and the highest increase is expected to be from 324.3mm (2030) to 333.5mm (2040) (Fig. 7VB). In any case, surface runoff, which forms a significant part of the water yield, could be harnessed through the construction of dams and dugouts for use in the long dry season for irrigation and livestock watering among others. In the long term, dealing with surface runoff through institutional collaboration is critical to dealing with sedimentation in dams and dugouts within the catchment and particularly the Vea dam. Regarding ET in the catchment, the highest decline in actual ET is expected to be -2.4% between 2030 (656.0mm) and 2040 (640.3mm) with the lowest decline of -1.3% observed between 1996(684.6mm) and 2006(675.5mm) (Fig. 7VF). The major factors influencing ET are evaporation from water bodies and transpiration and photosynthesis processes from plants and trees (Bhatt and Hossain, 2019; Alhassan and Jin, 2020). Hence, the reduction in transpiration due to excessive vegetation cover removal is augmented by the increase in evaporation from water bodies due to the construction of mini dams and dugouts within the Vea catchment between 1985 and 2006 for dry season irrigation and livestock watering (GIDA, 2010). Thus, resulting in a rather minimal decline in ET than expected.

Table 6 Annual percentage of change between scenarios in the Vea catchment; SURQ-Surface runoff (mm), Water yield(mm), - WYLD, Percolation(mm) –PERC, Baseflow(mm)- GW_Q, Evapotranspiration(mm)- ET
Scenario	SURQ	WYLD	PERC	GWQ	ET
1986–1996	12.65747	1.420632	-3.5985	-5.06664	-1.37966
1996–2006	11.22276	1.307844	-4.80502	-5.35851	-1.32431
2006–2020	19.50973	1.907786	-9.14556	-8.65718	-1.4875
2020–2030	19.90668	2.199062	-13.0118	-10.5339	-1.42235
2030–2040	22.515	2.847457	-16.8851	-17.4894	-2.40159
2020–2040	46.90367	5.109136	-27.6999	-26.181	-7.76638

4.4.3 Monthly And Seasonal Response Of Hydrological Variables To Lulcc

The mean monthly distribution of simulated water balance components' response to LULCC in the Vea catchment is given in Fig. 8. The catchment is located in the savannah ecological zone of Africa, characterized by long dry periods (Namara et al., 2014; Berhane et al., 2016). Rainfall begins in May and surface runoff and water yield peak in August because the early rains tend to build up the soil moisture level towards the end of July. The presence of vegetation, as well as the initial soil moisture in the subsequent month (August), tend to allow rainfall to satisfy their soil moisture deficit quickly and then more water runoffs the catchment from upstream into the reservoirs, hence producing more runoff in late August (Fig. 8V1). Percolation is water that infiltrates into the deeper layers of the soil to form part of the groundwater. Since percolation takes some days after rainfall or irrigation incident (Schreiner-McGraw and Vivoni, 2017), the Vea catchment percolation hydrograph peaks in August-September (Fig. 8V3). Baseflow is the sub-surface flow that is fed to streams by delayed paths, mostly derived from the precipitation of previous months. While rainfall and surface runoff hydrographs peak in August, the base flow ascending limb peaks in September and the recession limb extends to February (Fig. 8V4). Larbi et al., (2021) also observed a similar trend in the Vea catchment where surface runoff peaks in August and base flow extends to February. Evapotranspiration is highest in August and September due to the abundance of water from rain and water content in vegetation. In response to land-use change, the peak ET in August decreases marginally from 125.2(mm) in 1986 to a projected 122.6(mm) in 2040 (Fig. 8V2). Surface runoff increased all through the rainy season in response to agriculture expansion and is projected to increase from 26.7mm (1986) to 48.9mm (2040) in August (Fig. 8V1). On the other hand, base flow is expected to reduce from 41.9mm (1986) to 24.7mm (2040) in September. Similarly, Gebremicael et al., (2013) reported that the rising trend in surface runoff and water yield in the wet seasons spanning from March to September, and the downward trend in the same variables in the dry season was mostly linked to Agriculture expansion and Urbanization in El Diem station of the Blue Nile basin. Generally, the seasonal responses of LULCC to parameters show minimal variations in the dry season (December-April).

4.4.4 Hydrological Components Response To Lulcc At Subwatershed Scale

The sub-watershed-wise spatial variation in hydrological variables from 1985 to 2040 is given in Fig. 9. The spatial map shows a significant increase in surface runoff, occurring principally in the Upper northern and southeastern part of the catchment corresponding to the spatial distribution of settlements and farmlands dominant areas (Fig. 4) and potential future expansion. The impermeability of topsoil associated with built areas usually results in high surface runoff leading to a reduction in groundwater recharge rate demonstrated by the spatial pattern of percolation and base flow (Fig. 9 VP and VB). Thus, though the general trend of base flow and percolation is that of a decreasing one, the rate of decline is lowest in areas where there are still patches of tree reserves, particularly in the Northern and southeastern parts of the catchment. This is substantiated by Feddema, (1998) assertion that degraded vegetation in an environment usually results in a decrease in soil water retention and for that matter a decrease in the recharge of the aquifers, generating an increase in surface runoff. The Hotspots of decreases in ET are localized in the mid-portion where the Vea reservoir is located and the northern part where there is still a substantial amount of forest cover (Fig. 9VS). Therefore, the reduction of natural vegetation is the trigger of the declining ET in the catchment. Conforming to the present study, Sajikumar and Remya, (2015) reported an increase in peak runoff because of land-use change, and Niu and Sivakumar, (2013) reported increased flow due to deforestation and decreased flow due to afforestation. Additionally, the increased discharge has been long recognized to be associated with urbanization and farmlands (Du et al., 2012).

4.6 Disparities In Watershed Water Balance Modeling

Land, climate, and water resources are integral components of the watershed ecosystem (Gessesse et al., 2015). Different land uses have different influences on hydrology as they influence the curve number differently (Kundu et al., 2017). The land cover transition from tree cover to tree cover removal tends to decrease the CN resulting in an increase in surface runoff and the accompanying reduction in groundwater recharge rate due to a decline in sub-surface flow. Since ET depends on evaporation and transpiration, tree cover removal and reduction in water resources consequently will result in ET reduction. At the same time, IPCC, (2001) pointed out that even with high precipitation in sub-Saharan Africa, surface runoff may decrease due to increasing ET driven by increasing temperatures. In any case, a reduction in tree vegetation means a decline in ET while the increased temperature will increase ET. Implying therefore that inevitable reduction in future tree cover removal and increase in future temperature will affect water balance components in different directions.

Assessing the impact of land cover change on hydrology in Africa, ( Li et al., 2007) indicated a decrease in ET and an increase in surface runoff as well as streamflow in response to deforestation in the Niger and lake chad basins. In lake Tanan and Guder watershed Blue Nile basin, Ethiopia, expansion of farmlands and decline in forests, shrubs, and woodlands was linked to the increase in surface runoff and decline in ET and base flow (Woldesenbet et al., 2017; Kidane et al., 2019). An Increase in surface runoff and water yield due to the continuous increase in cropland and decrease in forest cover was also reported in an East African watershed (Baker and Miller, 2013) and Pra basin in Ghana ( Awotwi et al., 2019). In other parts of the world, similar trends were reported ( Warburton et al., 2012; Wagner et al., 2013; Kundu et al., 2017).

On the other hand, climate change's impact on watershed hydrology has been widely studied and an opposite trend is observed. For instance, Kankam-Yeboah et al., (2013) projected a decrease in mean annual rainfall by 19% and a 1.9^oc increase in temperature for the 2050s resulting in a decrease in streamflow in the white Volta basin. Gemechu et al., (2021) assumed a constant future land use while changing climate variables under RCP Scenarios in the Upper Abbay Basin (Ethiopia) and reported a decrease in water yield from 0.6–9% under RCP 4.5 and RCP 8.5 scenarios, which was attributed to precipitation decline and increased temperature. Despite a 20.1mm increase in rainfall by 2050 in the Densu river basin of Ghana, Justice Ankomah-Baffoe et al., (2021) reported a 61% increase in ET and a 23mm decline in water yield due to a 2^o increase in temperature.

The current study with a 2000–2018 simulation period and projecting to 2020–2040 for the Vea catchment found a -7.8% decrease in ET, 46.9%, and 5.1% decrease in surface runoff and water yield due to an increase in CN from 81.26 to 86.06. Contrarily with mean monthly rainfall showing relatively no change in the same catchment, Larbi et al., (2021) reported about 8.3% increase in ET and − 42.8% and − 38.7% decrease in surface runoff and water yield due to a 1.3^oc increase in mean annual temperature under RCP 4.5 Scenario, with 1981–2005 simulation period and projecting to 2020–2045. Climate was held constant with the current study while varying land cover change varied whereas Larbi et al., (2021)’s SWAT model held constant the land use and varied climate. Therefore though uncertainties regarding model calibration can not be ruled out(Parajuli et al., 2009), the different outcomes can mainly be attributed to the difference in the direction of influence of past and future climate and land-use changes on the water the balance components. Since the variability in surface runoff and water yield depends on several factors including LULCC, soil, surface rise, climate change, and soil water recharge among others, treating variation in climate and land-use change as stand-alone may not afford a holistic view of future water balance variations in a watershed. In other to understand and predict water resources realistically, taking into account the climate and land-use changes is imperative to present the reality of balance between the input of water from precipitation and outflow by evapotranspiration, groundwater recharge, and streamflow.

4.7 Uncertainties

Uncertainty refers to a quantity or process associated with the measurement or analysis that is not known. Though it will be ideal to have data and modeling processes free from uncertainties so that models can represent the real-world system perfectly, but this is not possible. The uncertainties associated with the current study include; uncertainties linked to the SWAT input data, and the SWAT modeling process.

I. Uncertainties connected to the SWAT input data;

Uncertainties associated with fixed and dynamic input data maybe liked to hydro-climatic input data such as the measured discharge data used for calibration of streamflow, temperature, precipitation, and DEM. Additionally, the SWAT hydrological model inherits uncertainties from LULC maps derived from the classified satellite images.

II. SWAT modeling process;

Every modeling process comprises some uncertainties. In the case of this study, the uncertainty is associated with the model assumption that land cover change and climate variations are stand-alone so that climate is held static while varying land cover change. This may not afford a holistic view of water balance variations in a watershed since the climate is not static and the fact that, land cover changes influences climate and vice versa and both influence watershed hydrology (Zhang et al., 2009). Also, because of the lack of data to calibrate sediment, the outcome of the model only provides the foundation for sediment management. It must be stated that the uncertainties connected to the image projection are more than those linked to the hydrological model.

The current study modeled the hydrological response of the Vea reservoir catchment in the White Volta basin to past, present, and future LULCC using the SWAT model. Model calibration and validation indicated satisfactory performance with NSE values of 0.738 and 0.782 for calibration and validation respectively. LULCC scenarios reveal a continuous increase in anthropogenic activities (farmland and built areas expansion) to the detriment of savannah woodlands and forest cover leading to an increasing CN from 1986 to the projected 2040. The model results reveal an increased surface runoff and reduced base flow and percolation in response to increased CN. Somewhat surprisingly, water yield did not alter much due to the counteracting effect of the increasing surface runoff and decreasing base flow and percolation. ET also reduced minimally from − 1.3% (1986) to the projected − 2.4%(2040). This is so because some reservoirs were constructed in the catchment between 1986 and 2006 to provide water for dry season irrigation and livestock watering (GIDA, 2010), hence the decreasing transpiration from trees is somewhat neutralized by the increased evaporation from these reservoirs. Seasonal analysis of hydrological components in response to land-use change reveals that variations are predominant in the rainy season with surface runoff increasing from 26.7mm (1986) to a projected 48.9mm (2040) in August while the base flow is expected to reduce from 41.9mm to 24.7mm in the same period. The high increase in surface runoff in the catchment could likely be responsible for the high sedimentation rate in the Vea reservoir. Hence, devising management options to reduce runoff by protecting the soil from raindrop impact and increasing surface roughness to reduce the velocity of flow and soil erosion is necessary. These can include but are not limited to afforestation, agroforestry systems, and contour plowing in farmlands. The implementation of these sustainable land reforms would require effective institutional collaboration. Alternatively, by constructing dams and dugouts, the abundant surface runoff in the rainy season could be harnessed for use in the long dry seasons. Also, the seasonal variation in hydrological components reveals that base flow peaks in September and extend to early march though at a declining rate, indicating therefore that the shallow aquifer store enough water that can be extracted at a minimal cost for dry season gardening.

It is hoped that, the outcomes of this research will offer an essential foundation for stakeholders planning toward sustainable future water resources and sediment management in the Vea catchment. Non the less, there were two main uncertainties associated with the study including uncertainties linked to the SWAT input data and the swat modeling process. This implies that additional research is needed to enhance the results, particularly regarding catchment sedimentation by collecting field data to properly calibrate the sediment transport and yield to understand the reality of the sedimentation process in the catchment, given the fact that sedimentation is a very big problem in the Vea reservoir. Hence, further assessment of the sedimentation process using measured data to calibrate the model is necessary. Also, the impact of interconnected climate and landcover change influence of watershed water balance is necessary. This therefore will be the focus of the forthcoming research.

Acknowledgment

The authors duly acknowledge the funding support from the regional water and environmental sanitation center, Kumasi (RWESCK), an African Center Of Excellence under the auspices of the World Bank and Government of Ghana. Irrigation Company of the upper region (ICOUR) is much appreciated for the Data support.

Authors contributions.

Conceptualization, Data collection, and Preprocessing were done by JAA and AJ. JAA also contributed to Literature review and modeling while writing the first draft and editing was performed by JAA and SA. Supervision, Further Editing, Reviewing, and Validation were done by AAK, EDO, and OSN. All authors read and approved the final manuscript.

Data availability statement

The data that supports the findings of this study are available from the corresponding author (JAA), upon reasonable request.

Conflict of interest.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper

Abbaspour KC, Vaghefi SA, Srinivasan R (2017) A guideline for successful calibration and uncertainty analysis for soil and water assessment: A review of papers from the 2016 international SWAT conference. Water (Switzerland) 10. https://doi.org/10.3390/w10010006
Adongo TA, Kugbe JX, Gbedzi VD (2016) Siltation of the Reservoir of Vea Irrigation Dam in the Bongo District of the Siltation of the Reservoir of Vea Irrigation Dam in the Bongo District of the Upper East Region, Ghana. Int. J. Sci. Technol
Adongo TA, Kyei-Baffour N, Abagale FK, Agyare WA (2019) Assessment of reservoir sedimentation of irrigation dams in northern Ghana. Lake Reserv Manag 36:87–105. https://doi.org/10.1080/10402381.2019.1659461
Adwubi A, Amegashie BK, Agyare WA, Tamene L, Odai SN, Quansah C, Vlek P (2009) Assessing sediment inputs to small reservoirs in Upper. Lakes Reserv. Res. Manag. 2009 14 279–287 279–287. https://doi.org/10.1111/j.1440-1770.2009.00410.x
Agyekum WA, Asare EB (2016) Challenges Associated With Groundwater. Ghana J Sci 56:39–51
Al-sharif AAA, Pradhan B (2014) Monitoring and predicting land use change in Tripoli Metropolitan City using an integrated Markov chain and cellular automata models in GIS. Arab J Geosci 7:4291–4301. https://doi.org/10.1007/s12517-013-1119-7
Alhassan A, Jin M (2020) Evapotranspiration in the Tono reservoir catchment in upper east region of Ghana estimated by a novel TSEB approach from ASTER imagery. Remote Sens 12. https://doi.org/10.3390/rs12030569
Arnold JG, Fohrer N (2005) SWAT2000: Current capabilities and research opportunities in applied watershed modelling. Hydrol. Process. 19, 563–572. https://doi.org/10.1002/hyp.5611
Awotwi A, Anornu GK, Quaye-Ballard JA, Annor T, Forkuo EK, Harris E, Agyekum J, Terlabie JL (2019) Water balance responses to land-use/land-cover changes in the Pra River Basin of Ghana, 1986–2025. CATENA 182:104129. https://doi.org/10.1016/j.catena.2019.104129
Ayivi F, Jha MK (2018) Estimation of water balance and water yield in the Reedy Fork-Buffalo Creek Watershed in North Carolina using SWAT. Int Soil Water Conserv Res 6:203–213. https://doi.org/10.1016/j.iswcr.2018.03.007
Babel, Karssenberg, and D (2013) Hydrological models are mediating models. Hydrol Earth Syst Sci Discuss 10:10535–10563. https://doi.org/10.5194/hessd-10-10535-2013
Baker TJ, Miller SN (2013) Using the Soil and Water Assessment Tool (SWAT) to assess land use impact on water resources in an East African watershed. J Hydrol 486:100–111. https://doi.org/10.1016/j.jhydrol.2013.01.041
Barry B, Obuobie E, Andreini M, Andah W, Pluquet M (2005) The Volta River basin, Assessment, Comprehensive Management, Water.Water1–187
Berhane G, Gebreyohannes T, Martens K, Walraevens K (2016) Overview of micro-dam reservoirs (MDR) in Tigray (northern Ethiopia): Challenges and benefits. J Afr Earth Sci 123:210–222. https://doi.org/10.1016/j.jafrearsci.2016.07.022
Bhatt R, Hossain A (2019) Concept and Consequence of Evapotranspiration for Sustainable Crop Production in the Era of Climate Change. Adv Evapotranspiration Methods Appl. https://doi.org/10.5772/intechopen.83707
Boateng A, Musah AkuruguBismarkA, Zango,Abanyie S, Kojo S, Ampofo S (2015) Assessing the Impact of a Dam on the Livelihood of Surrounding Communities A Case Study of Vea Dam in the Upper East Region of Ghana. J Environ Earth Sci ISSN 5:2224–3216 ISSN 2225 – 0948
Cerdà A, Keesstra SD, Rodrigo-Comino J, Novara A, Pereira P, Brevik E, Giménez-Morera A, Fernández-Raga M, Pulido M, di Prima S, Jordán A (2017) Runoff initiation, soil detachment and connectivity are enhanced as a consequence of vineyards plantations. J Environ Manage 202:268–275. https://doi.org/10.1016/j.jenvman.2017.07.036
Chemura A, Rwasoka D, Mutanga O, Dube T, Mushore T (2020) The impact of land-use/land cover changes on water balance of the heterogeneous Buzi sub-catchment, Zimbabwe. Remote Sens Appl Soc Environ 18:100292. https://doi.org/10.1016/j.rsase.2020.100292
Cornelissen T, Diekkrüger B, Giertz S (2013) A Comparison of Hydrological Models for Assessing the Impact of Land Use and Climate Change on Discharge in a Tropical Catchment A comparison of hydrological models for assessing the impact of land use and climate change on discharge in a tropical catchme. J Hydrol 498:221–236. https://doi.org/10.1016/j.jhydrol.2013.06.016
De Jong Posthumus EC (2017) Water Balance Modelling of the Tandjari Reservoir in Burkina Faso Master of Science Thesis Ewoud de Jong Posthumus 73
de Silva PR, Silva V, Singh MT, de Souza VP, Braga EP, de Holanda CC, Almeida RM, Sousa F, Braga ACR (2018) Simulation of stream flow and hydrological response to land-cover changes in a tropical river basin. CATENA 162:166–176. https://doi.org/10.1016/j.catena.2017.11.024
Deng X, Shi Q, Zhang Q, Shi C, Yin F (2015) Impacts of land use and land cover changes on surface energy and water balance in the Heihe River Basin of China, 2000–2010. Phys Chem Earth 79–82:2–10. https://doi.org/10.1016/j.pce.2015.01.002
Deng X, Zhao C, Yan H (2013) Systematic modeling of impacts of land use and land cover changes on regional climate: A review. Adv. Meteorol. 2013. https://doi.org/10.1155/2013/317678
Du J, Qian L, Rui H, Zuo T, Zheng D, Xu Y, Xu C (2012) Assessing the effects of urbanization on annual runoff and flood events using an integrated hydrological modeling system for Qinhuai River basin. China 465:127–139. https://doi.org/10.1016/j.jhydrol.2012.06.057
Feddema JJ (1998) Estimated impacts of soil degradation on the African water balance and climate. Clim Res 10:127–141. https://doi.org/10.3354/cr010127
Feinstein AR, Cicchetit VD (1990) HIGH AGREEMENT BUT LOW KAPPA: I. THE PROBLEMS OF TWO PARADOXES * 43, 543–549
Foley JA, DeFries R, Asner GP, Barford C, Bonan G, Carpenter SR, Chapin FS, Coe MT, Daily GC, Gibbs HK, Helkowski JH, Holloway T, Howard EA, Kucharik CJ, Monfreda C, Patz JA, Prentice IC, Ramankutty N, Snyder PK (2005) Global consequences of land use. Sci (80-) 309570–574. https://doi.org/10.1126/science.1111772
Gashaw T, Tulu T, Argaw M, Worqlul AW (2018) Modeling the hydrological impacts of land use/land cover changes in the Andassa watershed, Blue Nile Basin, Ethiopia. Sci Total Environ 619–620. https://doi.org/10.1016/j.scitotenv.2017.11.191
Gassman PW, Reyes MR, Green CH, Arnold JG (2007) The soil and water assessment tool: Historical development, applications, and future research directions. Trans ASABE 50:1211–1250. https://doi.org/10.13031/2013.23637
Gebremicael TG, Mohamed YA, Betrie GD, van der Zaag P, Teferi E (2013) Trend analysis of runoff and sediment fluxes in the Upper Blue Nile basin: A combined analysis of statistical tests, physically-based models and landuse maps. J Hydrol 482:57–68. https://doi.org/10.1016/j.jhydrol.2012.12.023
Gemechu TM, Zhao H, Bao S, Yangzong C, Liu Y, Li F, Li H (2021) Estimation of hydrological components under current and future climate scenarios in guder catchment, upper Abbay Basin, Ethiopia, using the swat. Sustain 13. https://doi.org/10.3390/su13179689
Gessesse B, Bewket W, Bräuning A (2015) Model-Based Characterization and Monitoring of Runoff and Soil Erosion in Response to Land Use/land Cover Changes in the Modjo Watershed. Ethiopia L Degrad Dev 26:711–724. https://doi.org/10.1002/ldr.2276
Ghana Irrigation Development, (GIDA) (2010) Final Report of the Ghana Social Opportunity Project (GSOP)
Ghana Statistical Service (GSS) (2012) Final Summary Report Of 2010 Population & Housing Census
Githui F, Mutua F, Bauwens W (2009) Estimating the impacts of land-cover change on runoff using the soil and water assessment tool (SWAT): Case study of Nzoia catchment, Kenya. Hydrol Sci J 54:899–908. https://doi.org/10.1623/hysj.54.5.899
Golmohammadi G, Prasher S, Madani A, Rudra R (2014) Evaluating three hydrological distributed watershed models: MIKE-SHE, APEX, SWAT. Hydrology 1, 20–39. https://doi.org/10.3390/hydrology1010020
Guan D, Li H, Inohae T, Su W, Nagaie T, Hokao K (2011) Modeling urban land use change by the integration of cellular automaton and Markov model. Ecol Modell 222:3761–3772. https://doi.org/10.1016/j.ecolmodel.2011.09.009
Guug SS, Abdul-Ganiyu S, Kasei RA (2020) Application of SWAT hydrological model for assessing water availability at the Sherigu catchment of Ghana and Southern Burkina Faso. HydroResearch 3, 124–133. https://doi.org/10.1016/j.hydres.2020.10.002
He J, Tang Z, Chen HH, Wang S (2008) An accurate Markov model for slotted CSMA/CA algorithm in IEEE 802.15.4 networks. IEEE Commun Lett 12:420–422. https://doi.org/10.1109/LCOMM.2008.071460
Her Y, Frankenberger J, Chaubey I, Srinivasan R (2015) Threshold effects in HRU definition of the soil and water assessment tool. Trans ASABE 58:367–378. https://doi.org/10.13031/trans.58.10805
Gupta HV, Sorooshian S, STATUS OF AUTOMATIC, CALIBRATION FOR HYDROLOGIC MODELS: COMPARISON WITH MULTILEVEL EXPERT CALIBRATION (1999)J. Hydrol. Eng.25,135–143. https://doi.org/10.1002/fut.20174
Ibrahim B, Wisser D, Barry B, Fowe T, Aduna A (2015) Hydrological predictions for small ungauged watersheds in the Sudanian zone of the Volta basin in West Africa. J Hydrol Reg Stud 4:386–397. https://doi.org/10.1016/j.ejrh.2015.07.007
Intergovernmental Panel on Climate Change (2001) Climate Change 2001: Synthesis Report
Irrigation Development Authority (IDA) (1978)Tono Irrigation project; Abrief description
Arnold JG (2012) Model use, calibration, and validation. Biol Syst Eng 55:1249–1260. https://doi.org/10.13031/2013.42256
Júnior RF, Montenegro A, IMPACT OF LAND USE CHANGE ON, THE WATER BALANCE IN A REPRESENTATIVE WATERSHED IN THE SEMIARID OF THE STATE OF PERNAMBUCO USING THE SWAT MODEL the (2019) main springs and streams are generally intermittent, and both native vegetation and the watershed.The aim o4430,110–117
Yawson JAnkomah-Baffoe, Okae-Anti D (2021) Assessment of the Impacts of Climate Change on Some Hydrological Processes of The Densu River Basin. Ghana 13:147–166
Li KY, Coe MT, N.R. and R.D.J (2007) Modeling the hydrological impact of land-use change in West Africa 258–268. https://doi.org/10.1016/j.jhydrol.2007.01.038
Kankam-Yeboah K, Obuobie E, Amisigo B, Opoku-Ankomah Y (2013) Impact du changement climatique sur les débits de plusieurs bassins au Ghana. Hydrol Sci J 58:773–788. https://doi.org/10.1080/02626667.2013.782101
Kidane M, Bezie A, Kesete N, Tolessa T (2019) The impact of land use and land cover (LULC) dynamics on soil erosion and sediment yield in Ethiopia. Heliyon 5:e02981. https://doi.org/10.1016/j.heliyon.2019.e02981
Kiros G, Shetty A, Nandagiri L (2015) Performance Evaluation of SWAT Model for Land Use and Land Cover Changes under different Climatic Conditions: A Review. J Waste Water Treat Anal 06. https://doi.org/10.4172/2157-7587.1000216
Kondolf GM, Gao Y, Annandale GW, Morris GL, Jiang E, Zhang J, Cao Y, Carling P, Fu K, Guo Q, Hotchkiss R, Peteuil C, Sumi T, Wang H, Wang Z, Wei Z, Wu B, Wu C, Yang CT (2014) Sustainable sediment management in reservoirs and regulated rivers: Experiences from five continents. Earth’s Futur 2:256–280. https://doi.org/10.1002/2013ef000184
Kundu S, Khare D, Mondal A (2017) Past, present and future land use changes and their impact on water balance. J Environ Manage 197:582–596. https://doi.org/10.1016/j.jenvman.2017.04.018
Lambin EF, Meyfroidt P (2011) Global land use change, economic globalization, and the looming land scarcity 108. https://doi.org/10.1073/pnas.1100480108
Larbi I, Hountondji FCC, Dotse S-Q, Mama D, Nyamekye C, Adeyeri OE, Djan’na Koubodana H, Odoom PRE, Asare YM (2021) Local climate change projections and impact on the surface hydrology in the Vea catchment, West Africa. Hydrol Res 00:1–15. https://doi.org/10.2166/nh.2021.096
Larbi I, Obuobie E, Verhoef A, Julich S, Feger KH, Bossa AY, Macdonald D (2020) Water balance components estimation under scenarios of land cover change in the Vea catchment, West Africa. Hydrol Sci J 65:2196–2209. https://doi.org/10.1080/02626667.2020.1802467
Legates DR, McCabe GJ (1999) Evaluating the use of “goodness-of-fit” measures in hydrologic and hydroclimatic model validation. Water Resour Res 35:233–241. https://doi.org/10.1029/1998WR900018
Lopes JWB, de Araújo JC (2019) Simplified method for the assessment of siltation in semiarid reservoirs using satellite imagery. Water (Switzerland) 11:1–18. https://doi.org/10.3390/w11050998
Warburton ML, Schulze RE, G.P.W.J.S (2012) Hydrological impacts of land use change in three diverse South African catchments.J. Hydrol.414–415
Magesh NS, Chandrasekar N, Soundranayagam JP (2012) Delineation of groundwater potential zones in Theni district, Tamil Nadu, using remote sensing, GIS and MIF techniques. Geosci Front 3:189–196. https://doi.org/10.1016/j.gsf.2011.10.007
Martey E, Al-hassan RM, Kuwornu JKM (2012) Commercialization of smallholder agriculture in Ghana: A Tobit regression analysis. 7:2131–2141. https://doi.org/10.5897/AJAR11.1743
Martin N (2006) Development of a water balance for the Atankwidi catchment, West Africa – A case study of groundwater recharge in a semi-arid climate. Ecol Dev Ser 49:176
Memarian H, Balasundram SK, Abbaspour KC, Talib JB, Sung B, Sood CT, A.M (2014) Modélisation hydrologique de scénarios d’utilisation des sols en zone tropicale à l’aide du modèle SWAT. Hydrol Sci J 59:1808–1829. https://doi.org/10.1080/02626667.2014.892598
Merz R, Blo¨schl G (2004) Regionalisation of catchment model parameters 287. 95–123. https://doi.org/10.1016/j.jhydrol.2003.09.028
Molina-navarro E, Hallack-alegría M, Martínez-pérez S, Ramírez-hernández J, Mungaray-moctezuma A, Sastre-merlín A (2016) Hydrological modeling and climate change impacts in an agricultural semiarid region. Case study : Guadalupe River basin, Mexico. Agric Water Manag 175:29–42. https://doi.org/10.1016/j.agwat.2015.10.029
Moriasi DN, Arnold JG, Liew MW, Van, Bingner RL, Harmel RD, Veith TL 2007. MODEL EVALUATION GUIDELINES FOR SYSTEMATIC QUANTIFICATION OF ACCURACY IN WATERSHED SIMULATIONS D. Colomb. Med. https://doi.org/10.1234/590
Mukhopadhyay A, Mondal A, Mukherjee S, Khatua D, Ghosh S, Mitra D, Ghosh T (2014) Forest cover change prediction using hybrid methodology of geoinformatics and markov chain model: A case study on sub-Himalayan town Gangtok, India. J Earth Syst Sci 123:1349–1360. https://doi.org/10.1007/s12040-014-0476-2
Myint SW, Wang L (2006) Multicriteria decision approach for land use land cover change using Markov chain analysis and a cellular automata approach. Can J Remote Sens 32:390–404. https://doi.org/10.5589/m06-032
Namara RE, Hope L, Sarpong EO, De Fraiture C, Owusu D (2014) Adoption patterns and constraints pertaining to small-scale water lifting technologies in Ghana. Agric Water Manag 131:194–203. https://doi.org/10.1016/j.agwat.2013.08.023
Napoli M, Massetti L, Orlandini S (2017) Hydrological response to land use and climate changes in a rural hilly basin in Italy. CATENA 157:1–11. https://doi.org/10.1016/j.catena.2017.05.002
Nash JE, Sutcliffe JV, RIVER FLOW FORECASTING THROLIGH CONCEPTUAL MODELS PART I - (1970) A DISCLISSION OF PRINCIPLES.J. Hydrol. Eng.10,282–290. https://doi.org/10.1080/00750770109555783
Nie W, Yuan Y, Kepner W, Nash MS, Jackson M, Erickson C (2011) Assessing impacts of Landuse and Landcover changes on hydrology for the upper San Pedro watershed. J Hydrol 407:105–114. https://doi.org/10.1016/j.jhydrol.2011.07.012
Niu J, Sivakumar B (2013) Study of runoff response to land use change in the East River basin in South China. https://doi.org/10.1007/s00477-013-0690-5
Opoku E, Macgregor CJ, Sloan S, Sayer J (2019) Deforestation is driven by agricultural expansion in Ghana ’ s forest reserves 5. https://doi.org/10.1016/j.sciaf.2019.e00146
Ottinger M, Kuenzer C, Liu G, Wang S, Dech S (2013) Monitoring land cover dynamics in the Yellow River Delta from 1995 to 2010 based on Landsat 5 TM. Appl Geogr 44:53–68. https://doi.org/10.1016/j.apgeog.2013.07.003
Parajuli PB, Nelson NO, Frees LD, Mankin KR (2009) Comparison of AnnAGNPS and SWAT model simulation results in USDA-CEAP agricultural watersheds in south-central Kansas. 763:748–763. https://doi.org/10.1002/hyp
Paudel S, Vetaas OR (2014) Effects of topography and land use on woody plant species composition and beta diversity in an arid Trans-Himalayan landscape. Nepal J Mt Sci 11:1112–1122. https://doi.org/10.1007/s11629-013-2858-3
Ristić R, Ljujić M, Despotović J, Aleksić V, Radić B, Nikić Z, Milčanović V, Malušević I, Radonjić J (2013) Reservoir sedimentation and hydrological effects of land use changes-case study of the experimental dičina river watershed. Carpathian J Earth Environ Sci 8:91–98
Sajikumar N, Remya RS (2015) Impact of land cover and land use change on runoff characteristics. J Environ Manage 161:460–468. https://doi.org/10.1016/j.jenvman.2014.12.041
Sang L, Zhang C, Yang J, Zhu D, Yun W (2011) Simulation of land use spatial pattern of towns and villages based on CA-Markov model. Math Comput Model 54:938–943. https://doi.org/10.1016/j.mcm.2010.11.019
Schilling KE, Jha MK, Zhang YK, Gassman PW, Wolter CF (2008) Impact of land use and land cover change on the water balance of a large agricultural watershed: Historical effects and future directions. Water Resour Res 45:1–12. https://doi.org/10.1029/2007WR006644
Schreiner-McGraw AP, Vivoni ER (2017) Percolation observations in an arid piedmont watershed and linkages to historical conditions in the Chihuahuan Desert. Ecosphere 8. https://doi.org/10.1002/ecs2.2000
Singh J, Knapp HV, Arnold JG, Demissie M (2005) Hydrological modeling of the Iroquois River watershed using HSPF and SWAT. J Am Water Resour Assoc 41:343–360. https://doi.org/10.1111/j.1752-1688.2005.tb03740.x
Spence C (2007) On the relation between dynamic storage and runoff: A discussion on thresholds, efficiency, and function. Water Resour Res 43:1–11. https://doi.org/10.1029/2006WR005645
Tang Y, Wang D (2017) Evaluating the role of watershed properties in long-term water balance through a Budyko equation based on two-stage partitioning of precipitation. Water Resour Res 53:4142–4157. https://doi.org/10.1002/2016WR019920
Tessema SM (2011) Hydrological modeling as a tool for sustainable water resources management: a case study of the Awash River Basin, Trita-LWR. LIC NV – 2056
Wagner PD, Kumar S, Schneider K (2013) An assessment of land use change impacts on the water resources of the Mula and Mutha Rivers catchment upstream of Pune, India. Hydrol Earth Syst Sci 17:2233–2246. https://doi.org/10.5194/hess-17-2233-2013
Wang SQ, Zheng XQ, Zang XB (2012) Accuracy assessments of land use change simulation based on Markov-cellular automata model. Procedia Environ Sci 13:1238–1245. https://doi.org/10.1016/j.proenv.2012.01.117
Woldesenbet TA, Elagib NA, Ribbe L, Heinrich J (2017) Hydrological responses to land use/cover changes in the source region of the Upper Blue Nile Basin, Ethiopia. Sci Total Environ 575:724–741. https://doi.org/10.1016/j.scitotenv.2016.09.124
Wu H, Chen B (2015) Evaluating uncertainty estimates in distributed hydrological modeling for the Wenjing River watershed in China by GLUE, SUFI-2, and ParaSol methods. Ecol Eng 76:110–121. https://doi.org/10.1016/j.ecoleng.2014.05.014
Yan B, Fang NF, Zhang PC, Shi ZH (2013) Impacts of land use change on watershed streamflow and sediment yield: An assessment using hydrologic modelling and partial least squares regression. J Hydrol 484:26–37. https://doi.org/10.1016/j.jhydrol.2013.01.008
Yira Y, Diekkrüger B, Steup G, Bossa AY (2016) Modeling land use change impacts on water resources in a tropical West African catchment (Dano, Burkina Faso. J Hydrology Elsevier B V. https://doi.org/10.1016/j.jhydrol.2016.03.052
Zou J, Bui KTT, Xiao Y, Doan C, Van (2018) Geo-Spatial Inf Sci 21:149–157. https://doi.org/10.1080/10095020.2017.1386848. Dam deformation analysis based on BPNN merging models

Download PDF

Version 1

posted

You are reading this latest preprint version

Modeling the impact of past and future land cover changes on a reservoir catchment hydrology in Semi-arid, Africa

Status:

Version 1

Abstract

Figures

1.0 Introduction

2.0 Study Area

3.0 Data And Methodology

3.1 Future land cover map prediction

3.2 Swat Model For Assessment Of Variation In Water Balance And Reservoir Components

NSE\(=1-\frac{{\sum }_{k=1}^{n}{\left({Y}_{obs,k}-{Y}_{sim,k}\right)}^{2}}{{\sum }_{k=1}^{n}{\left({Y}_{obs,k}-{Ȳ}_{obs}\right)}^{2}} \left(1\right)\)

PBIAS\(=100\times \frac{{\sum }_{k=1}^{n}({Y}_{obs,k}-{Y}_{sim,k})}{{\sum }_{k=1}^{n}{Y}_{obs,i}} \left(2\right)\)

RSR\(=\frac{RMSE}{STDV}=\frac{{\sum }_{k=1}^{n}\sqrt{{\left({Y}_{obs,k}-{Y}_{sim,k}\right)}^{2}}}{{\sum }_{k=1}^{n}\sqrt{{\left({Y}_{obs,k}-{ȳ}_{sim,k}\right)}^{2}}} \left(3\right)\)

3.2.2 Hydrological Dynamics Under Various Lulcc Scenarios

4.0 Results

4.1 Land use prediction and Land cover Assessment

4.1.1 Markov model validation to assess the accuracy of prediction

4.1.2 Historical And Future Land-use Change Dynamics

4.3 Hydrologic Parameter Sensitivity Analysis

4.3 Swat Model Calibration And Validation

4.4 Hydrological Dynamics In Response To Lulc Change Scenarios

4.4.1 Water balance ratios to mean annual rainfall under different LULC scenarios

4.4.2 Mean Annual And Percentage Change In Hydrological Components

4.4.3 Monthly And Seasonal Response Of Hydrological Variables To Lulcc

4.4.4 Hydrological Components Response To Lulcc At Subwatershed Scale

4.6 Disparities In Watershed Water Balance Modeling

4.7 Uncertainties

5.0 Conclusion

Declarations

References

Status:

Version 1