Groundwater modeling and estimation of the safe pumping rate for the sustainable use of groundwater resources in Kombolcha town, upper Borkena watershed, Ethiopia Open Access

A major issue for over a billion people on the planet is a lack of clean drinking water (Jayawardena 2012). A large quantity of groundwater is used for home, industrial, agricultural, and aquaculture purposes, which results in over-pumping and dwindling groundwater levels. This can contribute to soil salinization and land subsidence (Lin et al. 2013). Groundwater is an important water source for domestic, industrial, and agricultural uses in Ethiopia. However, with the growth of industry, agricultural activities, and urbanization, the influence of humans on surface and groundwater is growing (Tsigie 2015). Intensive groundwater extraction alters the flow system and challenges future availability (Azeref & Bushira 2020). To avoid groundwater depletion, it is important to conduct a study on the impact of pumping in groundwater resources (Purwadi et al. 2023). To minimize this difficulty, numerous kinds of literature revealed that lowering pumping rates or balancing pumping with recharge rates is vital (Ruud et al. 2004; Zhou 2009; van der Gun & Lipponen 2010; de Graaf et al. 2019; Butler et al. 2020).

Previous studies have used water balance models to estimate the sustainable yield of groundwater pumping. However, there are controversies regarding groundwater safe yield and sustainable yield estimation, and the truth is that both the terms have the same meaning (Zhou 2009). To determine a safe pumping rate or yield, it needs groundwater level time series data, exact withdrawal of groundwater data, recharge data, and groundwater follow rate parameters. Currently measuring the groundwater withdrawal rate of the aquifer is difficult, especially in developing countries. To fill this gap, this study sets different pumping rate scenarios using the designed pumping rate of the existing borehole.

There are several techniques for estimating recharge, but choosing one is difficult (Lerner & Issar 1990; Scanlon et al. 2002). The management of recharge is essential for sustainable groundwater use. The groundwater flow and the transport model also include recharge as a crucial component (Tilahun & Merkel 2009). Each technique has various assumptions as well as restrictions. Thus, it is advised to integrate several methodologies to decrease ambiguity and enhance the conceptual clarity of recharging at a study site. Usually speaking, the choice of a strategy depends on the availability of data, which is frequently scarce in many areas. A less acceptable recharge estimating technique may be used as a result of a data shortage, and there may be no other methods available to confirm the findings (Walker et al. 2019).

Any computational technique that approximates a system of underground water is a groundwater model (Anderson & Woessner 1992). The effective management of groundwater resources and the prediction of future responses can be achieved through numerical modeling (Jayawardena 2012). Groundwater flow modeling is frequently done using the program MODFLOW, a finite-difference flow model created by the United States Geological Survey (USGS) and originally made available to the public in 1987. It assumes that a three-dimensional continuity equation describing fluid flow through porous media may be applied to model groundwater flow in an aquifer system (Mescher 2018). Groundwater models have proven to be useful tools over the years for addressing a variety of groundwater problems and assisting in the decision-making process, even though they are, by definition, a simplification of a more complex reality (Kumar 2019).

A model will be chosen based on its applicability and limitations, the data that is available, its suitability for a given hydrogeological condition, the establishment of a well, its stability, its general acceptability, researcher recommendations, etc. (Kumar 2019). A graphical user interface (GUI) for groundwater simulation tools 4.3 ModelMuse model was used to characterize the long-term impact of groundwater pumping on the Borkena aquifer around Kombolcha town. This model was chosen due to its ease of use, data requirements, and availability.

Currently, over 39 groundwater wells are actively supplying water to meet the municipal and industrial demands of Kombolcha town. The government and non-governmental organizations have significantly depleted the groundwater in the upper part of the Borkena River watershed. There are limited studies available on balancing groundwater recharge and pumping rates in the study area. Then, the study focuses on estimating the groundwater recharge rate, groundwater modeling, and estimating the safe groundwater pumping rate for the sustainable use of groundwater resources of the upper Borkena River watershed.

The annual average precipitation of the study area is 1,168.3 mm, and the average maximum and the minimum temperature of the study area are 24.35 and 10.56 °C, respectively. The catchment area exists between sub-humid and sub-tropical, within a zone identified as ‘Dega and Weyna Dega’, where the rainfall amount is suitable for two crop seasons.

The physiographic condition of the catchment area is known by rugged topography, mountainous, valley-forming landscape, and steep slope. At Tossa, the top of the ridge around the northwest water divide is about 3,500 m.a.s.l., while in the middle part of the watershed, the area gently slopes, and the average altitude is around 1,800 m.a.s.l. Land use is important in hydrological and groundwater studies because it is a prominent factor affecting the recharge. Eucalyptus, Acacia, and Juniper trees make up a minor fraction of the area's plant cover, whereas bushes and shrubs proportionally cover the majority of it. The primary soil types in the watershed consist of leptosols, heplic herosols, eutric cambisols, and chromic vertisols. The flat, gentle-moderate, and fairly steep slopes, respectively, are home to these various soil types (Tsigie 2015).

The average hydraulic conductivity of nine wells was utilized in this study; it was obtained from Amhara Drilling Well Enterprise's well completion report. Table 2 provides further details.

Methods for estimating groundwater recharge that is often employed include chemical, isotopic, and water balance techniques. The water balance method was employed in this study to calculate groundwater recharge. However, other empirical formulae were used for the comparison and validation of water balance methods. The water balance approach is significant since recharge is often calculated using information that is already readily available data (rainfall, runoff, actual evaporation, and water levels).

The average rooting depth and the available water storage capacity were taken from Table 4 (1.25 m × 200 mm).

Aside from the water balance, the groundwater recharge rate was calculated using empirical methods. Utilizing several empirical correlations between recharge and rainfall that were created for various places with comparable climates, the evaluation of rainfall recharge was conducted (Gebrie et al. 2021).

Krishna Rao found the following correlations for various yearly rainfall amounts:

• ✓

  R = 0.20 (P − 400) for areas with annual normal rainfall (P) between 400 and 600 mm

• ✓

  R = 0.25 (P − 400) for areas with P between 600 and 1,000 mm

• ✓

  R = 0.35 (P − 600) for areas with P above 2,000 mm

Using the standards set by the Groundwater Estimation Committee, groundwater recharge from rainfall was also calculated. Ad hoc rules of rainfall infiltration were created by the Groundwater Estimation Committee in 1987 for calculating the recharge from rainfall.

• i.

  Alluvial areas: recharge could be taken as 20–25% of rainfall for sandy areas, and 10–20% for areas with high clay content (more than 40% clay).

• ii.

  Semi-consolidated sandstones: 10–15% of rainfall is considered a recharge.

• iii.

  Hard rock areas Granitic terrain: for weathered and fractured rocks, 10–15% of rainfall.

Unweather: 5–10% of rainfall

Basaltic terrain: 10–15% of rainfall

Weathered basalt: 4–10% of rainfall

Groundwater models provide additional information about the behavior of complicated systems and can aid in conceptual comprehension when appropriately developed. Once it has been established that they can be properly modeled, they may also be used to predict how future groundwater behavior will turn out, enhance decision-making, and enable the investigation of various management options (Kumar 2019). The GUI included in the ModelMuse software package can be used to create the flow and transport input files for PHAST as well as the input files for MODFLOW-2006. The temporal data in ModelMuse are independent of the stress periods, and the spatial data for the model are independent of the grid (Kumar 2019). The user can change the spatial and temporal discretization at any time because these data can be input independently.

The ModelMuse main window features a 3D view of the model that can be used to display model properties as well as top, front, and side views of the model that can be used for editing. The model grid can be created and edited using tools in ModelMuse (Winston & Goode 2017). It has a range of geographic functions and interpolation techniques that can be used to specify the model's spatial variability. The MODFLOW-2006 models' output can be viewed in ModelMuse, which can also be used to run PHAST and MODFLOW-2006. It can occasionally be challenging to provide a meaningful presentation of groundwater withdrawals due to the potential for symbol overlap for closely placed wells. Making a ‘footprint’ of the withdrawals is another way to depict groundwater withdrawals (Winston 2009). The ModelMuse software has different model types and flow packages. MODFLOW-NWT 5 is selected because it is more convenient for highly sloppy areas.

The following steps were used under the available data and the goals of this study:

• I.

  Fixing the mesh size or generating a grid: the mesh size controls the resulting quality. The smaller the mesh size, the higher the quality of the result, but when the mesh size becomes too small, it uses more space and the model does not simply converge. So, appropriate mesh size selection is important according to the size of the study area. This study used a 200 × 200 m mesh size.

• II.

  Specifying the number of layers or the number of aquifers: to specify the number of aquifers, understanding the subsurface geographical formation of the study area is a must. To identify the subsurface geographical formation of the Borkena watershed, 39 well borehole data were identified from the borehole completion report. The maximum depth of the boreholes in Borkena around Kombolcha is 333 m. This depth was used as the total depth of the aquifer. Based on well-log data, the aquifer is classified into four layers.

• III.

  Determining aquifer characteristics: determining the aquifer characteristics was the basic step in groundwater flow modeling. The most important parameters that determine groundwater movements are hydraulic conductivity, storage coefficient, and initial head under the dataset management of the ModelMuse. In the ModelMuse model, there are two ways of inserting hydraulic conductivity. The first one is inserting hydraulic conductivity through interpolation and the latter is directly inserting the average hydraulic conductivity. This study used both methods for comparisons. The initial head of the aquifer is the model top. The average hydraulic conductivity of the study area is 2.71 * 10⁻⁵ m/s.

• IV.

  Importing shape file: ModelMuse software is capable of importing GIS-executed shape files. Such shape files are either the boundary conditions or groundwater parameters specified under ModelMuse packages and program options. The Drainage shape file, Borkena boundary, shape file, recharge shape file, and evapotranspiration shape file were used for this study. This study used 0.384 m/year for the evaporation rate, 2.9 * 10⁻⁴m/s for the length that crosses the node of the cell used as a river conductance, considering the average width of the river as 2.7 m and the average depth of the river as 0.25 m in dry seasons.

• V.

  Determining time steps: the two types of time steps in MODFLOW are steady-state and transient conditions. Steady-state simulation uses only hydraulic conductivity, whereas the transient condition needs storage coefficient. Then, due to a shortage of storage coefficient data, this study employed under 50-year steps in a steady-state condition from 2015 to 2065.

• VI.

  Selecting solver: choosing a solver: a suitable solver has to be chosen carefully. The NWT Newton solver was the best model solver because it was better suited for upstream weighting factors and high slope gradients.

• VII.

  Determining layer properties: Determining layer properties is an important step before executing the model. It is specific to which layer is simulated and which is not. The ModelMuse under model layer groups has three alternatives to specify the layer. These are non-simulated, confined, and convertible options. The convertible options were important to give the model itself the ability to determine the aquifer property, whether confined or unconfined.

• VIII.

  Executing the model: this is the final step in executing the model. Setting the working directory and saving it are required before launching the software.

Determining the pumping rate is crucial for the sustainable use of the aquifer without excessive pumping or groundwater mining because unsustainable pumping causes land subsidence and complete well failure. Therefore, it is necessary to determine the pumping rate to use groundwater resources sustainably. To reduce the pumping rate of the Borkena watershed, this study develops three pumping rate scenarios. As shown in Table 5, 39 wells were pumped at 4-, 8-, and 12-h intervals using their designed pumping rates. Generally the road map of the study is illustrated in Figure 3.

As shown in Table 6 and Figure 4, Bredenkamp overestimates recharge directly, whereas other methods estimate the average. As a result, the water balance method was used in this study because it is a small variance compared to other empirical recharge estimation methods. However, when access to river gauge data is limited, Groundwater Estimation Committee Norms were more capable than the other methods of predicting recharge.

According to Gobezie et al.’s (2023) study, the total Borkena watershed recharge was estimated to be approximately 122 mm/year. This variation was created due to the study area under consideration. This research, however, focuses on the upper portion of the Borkena watershed, which is a potential recharge zone. The region's geology is characterized by cracked or disintegrated basalt, creating secondary porosity. This formation increases the amount of water that enters the ground.

As a result, the water balance method estimates that 15% of the annual rainfall recharges the groundwater. Gebrie et al. (2021) estimated that the average coefficient of recharge in the Tana Basin is 0.18. Redda et al. (2024) estimated the groundwater recharge rate of the upper Awash Basin to be 185.9–280.5 mm/year. A study by Gebrie et al. (2021) compares empirical formulas for estimating groundwater recharge, and the results indicate that Groundwater Estimation Committee Norms have a recharge coefficient of 0.14, which is the same as the results of this study. This indicates that in the data-scarce area, the Groundwater Estimation Committee Norms are capable of estimating the groundwater recharge rate.

The ModelMuse model was employed to determine the long-term pumping impact on groundwater with three pumping rate scenarios. This study used a 43-year projection of groundwater use, with well pumps operating from 2015 to 2065. To assess the sustainability of the groundwater resource of the upper Borkena watershed, this study used a 335 m maximum borehole depth with a typical pumping rate of 4, 8, and 12 h for the upcoming 43 years. The relevant outcomes are displayed as follows.

The results indicate that the groundwater will be able to fulfill the needs, as shown in Figure 7, for the years between 2015 and 2039, when all wells were pumped for 12 h each day.

If we were to pump at a rate faster than this, it would have a variety of unfavorable effects, such as borehole failure, land subsidence, soil settlement, or sediment pumping.

To estimate the groundwater recharge rate and potential evapotranspiration of the study area, this study used 5 years of meteorological data and streamflow data (2015–2020). The ModelMuse's 4.5 groundwater flow model was used to determine the safe withdrawal of groundwater from the aquifer by comparing the inflow and outflow until the water balance report of the ModelMuse software reaches zero. To estimate the groundwater recharge rate of the study area, this study used four empirical formulas in addition to the water balance method. The study found that using the water balance method, the recharge coefficient is 0.14 of the annual rainfall. This result is considered more accurate than results obtained using other methods. The deviation of the water balance method from the average of the other formulas is insignificant. The pumping rates for 4, 8, and 12 h/day were applied over the past 7 years, from 2015 to 2022, and will continue for the next 43 years until 2065 using the design pumping rate for each of the 39 wells simultaneously. This was done to assess whether the amount of water being withdrawn equals the annual recharge rate, as determined by water balance methods. The result of the ModelMuse water balance report showed that there will be enough storage for 4 and 8 h, while there will be a negative water balance report for 12 h of pumping from 2039 to 2065. Based on 8-h pumping scenarios, the ModelMuse model result indicated that the safe pumping rate for the sustainable use of Borkena groundwater was 975 L/s. In all pumping scenarios, groundwater drawdown is observed and forms the same type of cone of depression. This means that while all wells are being pumped at the same time, there is an interference between the wells. This study generally suggests that scheduling groundwater pumping and designing groundwater monitoring wells are necessary to avoid interference between wells; artificial recharge methods are also necessary to balance the inflow and outflow of groundwater from the system.

To carry out the study, the authors would like to acknowledge the Director of the Wollo University, Kombolcha Institute of Technology Research Center for its support.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.