Groundwater resource assessment and safe yield analysis for a well field supplying Lomé (Togo)

by Manuel V. Arce ^1*, Álvaro Olave ² and Eva Molinero ²

¹  h2i S.L. manuel@hdosi.es

 ²   Quantum Solutions S.L ; aolabe@quantump.com, emolinero@quantump.com,

^* Corresponding author: manuel@hdosi.es

Abstract

During hydrologic year 2023-2024, there was a severe drought event affecting the Guadalope River catchment. The Ebro River Basin Organization set a pilot study in the area, which consisted of pumping groundwater from an existing drilled well into the Calanda reservoir at a flow of 250 l·s^-1 for 17 days. Results showed that, during the drought period, there was hydraulic connection among nearby aquifers but not between the Guadalope River and the contiguous aquifers. It can be concluded that it would be necessary to complete this pilot study to obtain new and more accurate hydrogeological data. It is also considered that those who will be benefited from the pumping should be the promoters of new pumping tests in the future.

Keywords

groundwater, drought, pumping, conjunctive water use, Ebro River basin.

Note:

Papers published in this special issue of the European Geologist journal have undergone a thorough peer-review process but have not been copy-edited. Authors bear full responsibility for the linguistic accuracy of their contributions.

This work is licensed under a Creative Commons Attribution 4.0 International License.

---

1. Introduction

The Greater Lomé area of Togo is experiencing rapid population growth, urban expansion and increasing water demand. Existing groundwater sources near Lomé, particularly the Continental Terminal (CT) aquifer of the Agoè plateau and the Paleocene aquifer, already show signs of stress, including overexploitation, declining heads and increased saline intrusion along the coast.

At the regional scale, the CT does not form a single continuous aquifer across the whole Maritime Region. It occurs in several plateau aquifers separated from each other by Eocene marly outcrops. The present study focuses exclusively on the Attitogon plateau, where the Sévagan, Vogan and Anfoin sectors are treated as hydraulically connected parts of the same plateau-scale CT aquifer.

Several regional studies since the 1970s have characterised the sedimentary coastal basin aquifers and highlighted their development potential and limitations. In the CT aquifer, the Sévagan, Vogan and Anfoin sectors were identified as productive zones underlain by an extensive recharge area on the Attitogon plateau. The Paleocene aquifer, while regionally important, currently exhibits long-term piezometric decline around Lomé, indicating that present exploitation is not sustainable. The Maastrichtian aquifer is highly heterogeneous, with sandy lenses embedded in low-hydraulic conductivity clays and poorly constrained geometry, making it a high-uncertainty target for large-scale water-supply development.

Against this background, the Togolese Ministry of Water and Sanitation has promoted the development of a new groundwater well field east of Lake Togo, to supply Lomé and secondary centres such as Hahotoé, Vogan and Tsévié. The study focuses on the CT aquifer and aims to: (i) quantify renewable groundwater resources and their temporal variability; (ii) define a conservative safe yield for a new CT well field; (iii) assess operational supply reliability under present climate and an RCP6.0 climate-change scenario; and (iv) derive design and operational constraints for a sustainable well-field layout [7].

The geographic scope of the study to identify groundwater-resource systems encompasses the entire maritime region in relation to the Paleocene and Maastrichtian aquifers. The area analysed for CT aquifer development is located in the most favourable zone, situated on the Attitogon plateau (Figure 1).

---

---

2. Materials and Methods

The study builds on five decades of regional hydrogeological [1] work in the coastal sedimentary basin. Key references include the 1975 UNDP groundwater prospection, which first defined the main aquifer systems and estimated recharge, and subsequent feasibility and planning studies by national and international agencies. These works provided conceptual models, preliminary recharge estimates and evidence of overexploitation in the CT and Paleocene aquifers around Lomé [1–7].

The work followed a multi‐step approach combining compilation of existing information, targeted field investigations and quantitative modelling of recharge, aquifer regulation and well-field behaviour. First, previous hydrogeological, geophysical and planning studies for the coastal sedimentary basin and the Continental Terminal (CT), Paleocene and Maastrichtian aquifers were reviewed in detail. This review was used to define the conceptual model, the limits of the study area, the aquifers to be considered, and realistic ranges for hydraulic parameters and boundary conditions.

Field investigations focused on the CT aquifer in the planned well-field study area and its recharge zone on the Attitogon plateau. An inventory of existing groundwater works (boreholes, piezometers and large-diameter dug wells) was carried out, compiling construction details, screened intervals, depths, static water levels, approximate yields and available water-quality data. A piezometric survey was then implemented to characterise regional groundwater gradients and seasonal fluctuations along representative flow paths from the recharge area towards the coastal discharge zone and Lake Togo. In parallel, in-situ electrical conductivity measurements and selected chemical analyses were used to identify salinity patterns and to delimit zones potentially affected by the saline wedge. Two exploratory CT boreholes with associated observation piezometers were drilled, logged and developed in the central part of the study area. Step-drawdown and constant-rate pumping tests were performed and interpreted to obtain transmissivity, storage and well-loss parameters suitable for subsequent analytical drawdown calculations and for checking the consistency of the regional conceptual model.

The lithological logs of these exploratory boreholes and their associated observation piezometers were synthesised together with previous regional information to define a representative hydrostratigraphic section and to constrain the spatial range of CT saturated thickness used in the conceptual model.

Recharge to the CT aquifer was estimated using a daily soil–water balance model applied separately to the main recharge sectors. The model uses daily series of rainfall, temperature and reference evapotranspiration from nearby meteorological stations, and represents the partitioning of rainfall into actual evapotranspiration, changes in soil moisture, surface runoff and percolation below the root zone. Land use and soil characteristics were interpreted from existing cartography and satellite imagery to assign representative soil-water storage capacities and runoff thresholds. Model parameters were adjusted through sensitivity analysis and expert judgement to obtain a conservative representation of potential recharge. The resulting daily recharge series were aggregated to monthly and annual values for use in the groundwater-resource and reliability analyses.

To describe the regulation capacity of the CT aquifer and its discharge towards the sea and Lake Togo, the recharge time series were introduced into a lumped aquifer-reservoir model based on the classical Boussinesq–Maillet recession formulation. In this approach, the aquifer is treated as a linear reservoir draining towards a constant-head boundary. The model operates at a monthly time step and is controlled by a recession parameter that governs the decline of groundwater storage and discharge in the absence of recharge. Calibration was carried out against observed water-level time series in selected CT piezometers located both near the southern discharge boundary and further inland in the recharge area. Simulated changes in storage were converted into heads and adjusted to reproduce the amplitude and timing of the observed hydrographs, providing a consistent set of parameters and a time series of aquifer outflow.

On this basis, a probabilistic supply-reliability analysis was performed by confronting simulated monthly aquifer outflows with a set of hypothetical demand scenarios representing different intensities of well-field operation. For each scenario, “failure” months were defined according to simple shortfall criteria, and the frequency of failure and the complementary reliability were computed over the period covered by the climatic and recharge data. Years were grouped into dry, average and wet categories according to their total available resource, so that the performance of each operating scenario could be evaluated under contrasted hydrological conditions.

The potential influence of climate change on recharge, aquifer outflow and supply reliability was assessed using a regional climate scenario consistent with national projections for the Maritime Region. Projected changes in temperature and precipitation were applied as perturbations to the historical climatic series, and the full modelling chain (soil–water balance and aquifer-reservoir model) was rerun to generate climate-impacted recharge and outflow series. The same reliability framework was then used to compare the robustness of the different demand scenarios under current and projected climate.

Finally, a conceptual design of the future CT well field was developed and evaluated using analytical solutions for drawdown in aquifers. Hydraulic parameters derived from pumping tests and from previous regional studies were used to simulate several alternative layouts, varying the number, spacing and distribution of production wells between the main CT sectors. Drawdown superposition was used to estimate interference between wells and sectors, to check compliance with maximum allowable drawdown and desaturation criteria, and to verify that a sufficient hydraulic buffer is maintained against the landward migration of the saline wedge. Together, these methods provide a coherent basis to define conservative operational rules and a robust preliminary design for the planned CT well field.

3. Results

3.1. Hydrogeological context and aquifers considered

Around Lomé, the long-term behaviour of the deeper Paleocene and Maastrichtian aquifers already raises serious concerns for large-scale public supply. The Paleocene aquifer, although regionally important, shows a persistent decline in groundwater levels across much of the Maritime Region and clear signs of overexploitation in the Greater Lomé area. The Maastrichtian aquifer consists of discontinuous sandy lenses embedded in predominantly clayey formations, with highly variable productivity and poorly constrained geometry. This combination of depletion, heterogeneity and uncertainty led to a precautionary decision not to base the new well field on these two deeper systems.

In hydrogeological terms, the overexploited CT aquifer near Lomé corresponds to the Agoè plateau and should not be extrapolated to the Attitogon plateau. These two areas belong to different CT plateau aquifers, separated by Eocene marly outcrops. Within the Attitogon plateau, however, the CT behaves as a single aquifer system with three main exploitable sectors: Sévagan, Vogan and Anfoin.

Figure 1 includes a representative geological cross-section showing the general geometry of the CT aquifer on the Attitogon plateau.

Instead, the study focuses on the Continental Terminal (CT) aquifer west and north of Lake Togo, where conditions are more suitable for a new well field supplying Lomé and nearby towns. In this area, the CT is subdivided into three exploitable sectors (Sévagan, Vogan and Anfoin) aligned roughly west–east along the northern margin of Lake Togo and the coastal plain. These productive zones are underlain and sustained by a broad recharge area on the Attitogon plateau, bounded by the Haho and Mono rivers. The mapped recharge zone covers about 648 km², of which roughly 296 km² contribute to the Sévagan–Vogan sector and 352 km² to the Anfoin sector. On the plateau, CT formations crop out or subcrop beneath a thin soil and lateritic cover, under predominantly rural land use (grassland, mixed crops, scattered villages and patches of denser vegetation), which favours diffuse areal recharge.

On the Attitogon plateau, the CT aquifer consists of a detrital sandy sequence with sandy-clay interbeds and local lateritic horizons, resting on Eocene clays (Lama Formation) [12]. Hydrogeologically, it behaves as a predominantly unconfined aquifer. Its exploitable zone is a 5–14 km wide belt bounded to the north by the 0 m contour of the Eocene clay roof and to the south by the saline wedge, whereas areas farther north are only seasonally saturated or host small perched bodies. The mapped recharge area covers about 648 km² (296 km² for the Sévagan–Vogan sector and 352 km² for Anfoin) and is dominated by grassland and mixed crops, which favour diffuse rainfall infiltration. Piezometric monitoring from five CT piezometers indicates a generally shallow to moderate water table in the exploitable sectors, with small seasonal fluctuations in Sévagan and Vogan (<1 m) and larger amplitudes in the more inland Anfoin sector (about 2–4 m), but without a regional declining trend. Available study data indicate an initial saturated thickness of about 29 m in the northern marginal zone (Kpoguédé), while in the area of interest it reaches several tens of metres and up to about 65 m. Regional groundwater flow is directed from the Attitogon plateau towards Lake Togo, the lagoons and the coastal plain.

Figure 2 summarizes the hydrogeological regime of the CT aquifer in the study area, including recharge zones, piezometric contours and groundwater-flow directions. Close to the southern discharge boundary in the Sévagan and Vogan sectors, groundwater levels show relatively small seasonal fluctuations, reflecting the buffering effect of the CT storage and the nearby constant-head boundaries. Further inland, particularly in the Anfoin sector and on the Attitogon plateau, seasonal water-level amplitudes are larger, in line with the role of this area as a recharge zone. Over the period covered by recent monitoring, no systematic regional decline has been observed in the CT, in clear contrast with the behaviour of the Paleocene aquifer. Taken together, these characteristics – extensive outcrop and recharge area, laterally continuous sandy formations, favourable geometry and more stable piezometric conditions – make the CT east and north of Lake Togo the most robust option for developing a new well field for Lomé.

---

---

3.2. Hydrodynamic parameters

Two zones with very different hydrodynamic characteristics can be identified within the aquifer. On the one hand, there are the so-called exploitable sectors, where the installation of the well field can be considered, and on the other hand, the large regions located to the north of these, whose hydrodynamic activity is highly seasonal and linked to recharge periods. For the exploitable sectors, prior information on their hydrodynamic parameters is available thanks to numerous pumping tests carried out as part of previous studies, both in the Attitogon plateau sectors and in other sectors of the CT aquifer in the maritime region. Much less hydrodynamic information is available on the recharge zones, although they play a key role in the aquifer’s recharge and regulation capacity.

Information on the hydrodynamic parameters of the CT aquifer is available, primarily compiled from previous works [2,3,5,8]. Sixteen pumping test data sets are available for the CT (Table 1).

---

Table 1: S and T values collected from previous studies.

N°	Sector	Location	Canton	S	T (m²/day)
1	Plateau Agoè	F52	Agoenyive		2 592
2	Plateau Agoè	CACAVELI F9	Agoenyive		173
3	Plateau Agoè	BRASSERIE B2	Agoenyive		864
4	Plateau Agoè	BRASSERIE B15	Agoenyive		38 880*
5	Plateau Agoè	AGOUEVE F33 BIS	Agoenyive	0.03	691
6	Plateau Agoè	AGOUEVE B6	Sagbado		320
7	Plateau Agoè	CACAVELI F304	Agoenyive		653
8	Plateau Agoè	CACAVELI F303	Agoenyive		864
9	Plateau Kpogame	B12 DEKPO	Abodo		518
10	Plateau Tabligbo	F50	Tokpli		40 608*
11	Plateau Agoè	WN3	Mission-Tove		29
12	Plateau Attitogon	F45	Aklakou		1 296
13	Plateau Attitogon	F46	Hompou		138
14	Plateau Attitogon	F47	Vogan	0.05	1 210
15	Plateau Attitogon	TdE – Vogan	Vogan		480
16	Plateau Attitogon	F4 (OTP)	Vogan	0.05	259
	(*)Outliers

---

Most of the tests come from the Agoè plateau area, associated with extractions around the city of Lomé. On the Attitogon plateau, five pumping tests are available, covering the three sectors analyzed: Sévagan, Vogan, and Anfoin.

Overall, the available transmissivity data indicate an average of 834 m²/day, and a storage coefficient value between 0.03 and 0.05. On the Attitogon plateau, which includes the sectors of Sévagan, Vogan and Anfoin, the measured values ​​of T vary between 140 and 1,300 m²/day. The highest values ​​are found in the central areas of these sectors, which explains their greater productivity. Production rates of up to 3,000 m³/day were achieved at Anfoin (F45), 2,000 m³/day at Vogan (F47), and 2,500 m³/day at Sévagan [2].

North of the so-called exploitable zones lies a vast area of ​​the aquifer whose functionality is largely limited to recharge periods (Figure 3), when it transfers its resources to the exploitable zones. During dry periods, these zones are largely desaturated, so their recharge role is significantly reduced. The boundary between the exploitable zones and these large recharge zones to the north was established by the UNDP (1975) [2].

---

---

During the dry season, these zones contain a residual volume that slowly flows towards the exploitable zones, with which they maintain a water connection. In order to assess the hydrodynamics of these zones, which are poorly understood to date, and to evaluate the effect and type of margin of the exploitable zones, two exploratory wells were drilled in the Sévagan region and north of the exploitable zone boundary (F1-Lycée de Sévagan and F2- Sévagan Kpoguédé). It should be noted that the productivity of these pumping tests is strongly influenced by their marginal position within the aquifer, although hydraulic conductivity assessments confirm that it is a highly permeable aquifer.

Table 2 summarizes the main findings of the analyses performed during the pumping tests. The most likely transmissivity values ​​are in the range of 1,100 to 1,200 m²/day, with hydraulic conductivity values ​​between 20 and 30 m/day. The values ​​obtained for the storage coefficient are 0.03 to 0.05, consistent with those reported in previous studies. In both tests, aquifer depletion effects were identified, attributable to its marginal location.

---

Table 2: Summary of the results of the analytical interpretations of the pumping tests.

Location	T (m2/day)	K (m/d)	S	Remarks
Kpoguédé EPP	1120	29
Recovery
Kpoguédé Long duration	1224	29		Aquifer depletion during pumping (aquifer edge effect)
Recovery
Kpoguédé Long duration	1148		0.03	Good adjustment of the piezometer data
Pumping (PZ)
Sévagan EPP	Not interpretable
Recovery
Sévagan Long duration	Not interpretable
Recovery
Sévagan Long duration	1122	22	0.05	Good adjustment of the piezometer data
Pumping (piezometer)
Sévagan Long duration	835	17		Good fit considering pressure losses. Aquifer depletion (aquifer edge effect)
Pumping (pumping well)

---

3.3. Climate and recharge

The study area lies in the Maritime Region of Togo, under a subequatorial climate with two rainy seasons (April–July and September–November) separated by drier intervals. Daily rainfall series for 2003–2023 from the main ANAMET stations on the Attitogon plateau indicate mean annual precipitation from about 800 mm/a at Akoumapé to more than 1,140 mm/a at Afagnan, with intermediate values at Tabligbo and Aneho (Figure 4).

Applying the daily soil–water balance model to these series yields an average diffuse recharge of 100 mm/year in the Sévagan–Vogan sector and 157 mm/year in the Anfoin sector. These values are of the same order of magnitude as those reported at the regional scale for the Togolese coastal sedimentary basin [13], where present-day recharge values range from 47 to 225 mm/year, with a mean of about 136 mm/year. The key outcome of the recharge calculation is its very strong interannual variability, with coefficients of variation in the annual series exceeding 100%, much higher than those of rainfall, highlighting the non-linear, threshold-controlled response of the system. There have been years with exceptional values ​​close to 500 mm (2009 and 2023), as well as years with 0 mm of recharge (2010, 2012, 2018).

Recharge is also highly seasonal: about three quarters of the annual total occurs during the first rainy season (April–July), once soil moisture deficits have been filled, with the remainder mainly during the second rainy season (September–December) (Figure 5).

An analysis of the evolution of soil water balance elements (Figure 6) shows that soil water deficit conditions (ETR < ETP) are dominant, with maximum values ​​between 100 mm and 150 mm accumulated monthly during the dry seasons. During the rainy seasons, the water deficit is reduced, and finally, when this deficit approaches 0 mm, aquifer recharge (infiltration) events occur. This makes recharge events episodic and highly irregular over time.

Aquifer recharge is therefore a very irregular and dynamic process, taking the form of episodic pulses of highly variable magnitude. The aquifer regulates these recharge events, which constitute the so-called renewable resource, and at a rate that depends on the hydrodynamic characteristics of the aquifer, it is discharged into the sea.

---

---

 

---

---

3.4. Groundwater quality

Groundwater-quality data indicate a clear contrast between the overexploited CT aquifer in the Lomé–Agoè area and the CT aquifer of the Attitogon plateau. In the Lomé sector, previous studies report highly mineralized Na–Cl groundwater affected by marine intrusion and nitrate pollution linked to urban pressures [8].

Five water samples were taken in the study area, three of them provided by the Société Nouvelle des Phosphates du Togo (SNPT). They all show a consistent Na–Cl facies with low to medium mineralization: total dissolved solids (TDS) 240–360 mg/L; electrical conductivity (EC) at 25 °C: 335–627 µS/cm; nitrate concentrations of 13.7–23.5 mg/L, and no exceedance of the analysed heavy metals. Charge-balance errors calculated from the reported major ions are within about ±3%, supporting the internal consistency of the chemical analyses (Table 3). Potential diffuse nitrate inputs from scattered settlements and agricultural land use in the recharge area cannot be excluded, but the available data do not indicate current potability problems in the sampled CT wells. The analyses of the Sévagan and Kpoguédé boreholes include heavy metals: in neither case were levels exceeding drinking water thresholds identified.

---

Table 3: Summary of chemical analyses of CT groundwater on the Attitogon plateau. Major ions and TDS in mg/L, EC in µS/cm. 

Location	Date	TDS	T°C	pH	EC 25°C	Ca	Mg	Na	K	HCO3	CO3	Cl	SO4	NO3	Bal. err. (%)
Sévagan	26/09/2024	360	29.4	6.9	627	30.4	11.5	78	1	61	0	152.2	7.2	23.5	0.5
Kpoguédé	16/09/2024	328	29.1	6.6	458	19.2	6.24	52	0.7	30.5	0	98.1	8.3	13.7	1.2
SNPT F15	20/09/2024	240	29	5.6	335	9.6	8.16	44	1.3	24.4	0	78.1	7	16.7	1.3
SNPT F14	20/09/2024	251	28.3	5.6	351	10.4	7.68	45	1.3	30.5	0	80.1	9.8	15.4	-1.1
SNPT F8	20/09/2024	298	29	5.6	416	13.6	6.72	60	1.4	30.5	0	98.1	7.2	16.5	2.6
Reference values (WHO, 2022)		600						200				250	500	50

---

The heavy metals for which regulations set quality limits—As, B, Cd, Cr, Cu, Hg, Ni, Pb, Sb, and Se—occur at concentrations ​​below the detection limit (<2 μg/L) or the parametric value in the case of Fe and Mn.

The available EC measurements and chemical data indicate that the seawater-intrusion zone is restricted to the coastal CT aquifer near the Lomé–Agoè sector, whereas the Attitogon plateau samples are characterized by lower mineralization and do not show evidence of present-day marine intrusion.

In conclusion, groundwater from the CT aquifer on the Attitogon plateau can be considered potable with respect to the analysed parameters and consistent with World Health Organization (WHO) drinking-water standards [10]. It is characterized by low to moderate mineralization, a dominant sodium chloride facies, and relatively low pH values (5–7). These low pH values are interpreted as reflecting natural geochemical processes in the aquifer, where groundwater mineralization is closely linked to silicate hydrolysis, as also reported in groundwater systems from Togo [11].

3.5. Aquifer regulation

Daily recharge series derived from the soil–water balance for the Sévagan–Vogan and Anfoin recharge areas were aggregated to monthly values and used as input to a lumped aquifer-reservoir model representing the CT aquifer draining towards the sea. The model conceptualises the CT as a single linear reservoir in which episodic recharge events increase groundwater storage, which then discharges to the sea following an exponential Boussinesq–Maillet recession law. The key calibration parameter is the recession index K (days), which controls the characteristic depletion time and therefore the regulation capacity of the system. K was calibrated against observed piezometric series at Vogan and Klologo by transforming simulated changes in storage into heads and matching the amplitude and timing of measured hydrographs.

Once the value of the recession index (k) was calibrated, the monthly groundwater outflow simulation was performed. The simulation results are presented in Figure 7. Over the 2003–2023 period, the calibrated model yields an interannual mean freshwater outflow to the sea of about 232,000 m³/d for the CT aquifer as a whole: roughly 81,000 m³/d from the Sévagan–Vogan sector and 151,000 m³/d from the Anfoin sector. The recession index obtained through calibration is around 110 days, indicating a relatively limited storage-regulation capacity. These groundwater outflow volumes are interpreted as an upper bound of the physically available renewable resource; sustainable operational abstractions must remain below this limit and respect the spatial distribution of discharge between sectors.

To determine the fraction of these resources that remains available for the planned well field, existing abstractions from the CT aquifer were explicitly accounted for. Current withdrawals, including Togolaise des Eaux supplies, industrial pumping and village-level schemes, are of the order of 8,400 m³/d in Sévagan, 1,400 m³/d in Vogan and 3,200 m³/d in Anfoin, i.e. approximately 13,000 m³/d in total—very small compared to the mean renewable groundwater outflow, but nonetheless deducted when assessing water volumes that can be reassigned to new production wells.

The 2003–2023 monthly groundwater outflow series shows a strongly seasonal behaviour. In the Anfoin sector (Figure 8), mean monthly groundwater outflow ranges from about 50,000–70,000 m³/d at the end of the dry season (March–April) to nearly 290,000 m³/d in June, with sustained values typically above 160,000–200,000 m³/d during most of the rainy season (May–November). Similar patterns, with lower absolute magnitudes, are obtained for the Sévagan–Vogan sector (Figure 9). This confirms that the CT aquifer substantially smooths the highly episodic recharge signal but still delivers a markedly seasonal discharge regime, which constitutes a key constraint for defining safe operational volumes and for the subsequent supply-reliability analysis.

---

---

---

---

3.6. Supply reliability

Supply reliability was assessed by confronting the monthly available CT resources, expressed as groundwater outflow from the aquifer-reservoir model, with hypothetical monthly demands corresponding to different maximum operational volumes of the future well field between 40,000 and 100,000 m³/d. For each demand level and shortfall threshold, a month is classified as a “failure” when the simulated available resource does not reach a given fraction of demand; the complementary probability gives the monthly reliability. The 2003–2023 simulation period was then stratified into dry, average and wet years according to the annual available resource, so that reliability could be analysed separately by year type. The results obtained are summarized in Table 4.

For relatively high operational volumes (80,000–100,000 m³/d), results show a marked contrast between year types. In average and wet years – which together represent about three quarters of the 2003–2023 period – monthly supply reliability typically ranges between roughly 73 and 81 %, depending on the exact operational volume and shortfall criterion. In these years, most failures are concentrated in the late dry season, when aquifer outflows are still recovering from the previous recession. In dry years, which account for the remaining quarter of the series, the system becomes much more fragile: for the same operational volumes, monthly reliability drops to about 28–37 %, reflecting the strong reduction in annual recharge and the limited capacity of the CT aquifer to buffer multi-annual deficits.

This systematic comparison across year types underlines that the CT aquifer can technically sustain high operational volumes during wet and average years, but only at the cost of frequent shortages in dry years if pumping were kept constant throughout the year. The analysis therefore highlights a strong mismatch between the highly seasonal and interannually variable recharge regime and the relatively flat demand profile typical of urban water supply. It also shows that acceptable reliability for high nominal volumes would require explicit operational measures, such as seasonal modulation of pumping, conjunctive use with other sources or additional storage, and that the definition of a conservative “safe yield” cannot be based solely on long-term mean resources but must incorporate the probability of failure by year type.

---

Table 4: Simulation of the monthly supply reliability according to the type of year.

DRY YEAR					AVERAGE YEAR
Monthly supply reliability (%)					Monthly supply reliability (%)
operational vol.	N = 15%	N = 20%	N=30%		operational vol.	N = 15%	N = 20%	N=30%
(m3/day)					(m3/day)
100 000	28%	30%	37%		100 000	73%	75%	76%
90 000	30%	35%	37%		90 000	76%	76%	77%
80 000	37%	37%	38%		80 000	76%	77%	79%
70 000	38%	38%	45%		70 000	78%	79%	80%
60 000	45%	45%	46%		60 000	80%	81%	82%
50 000	46%	46%	55%		50 000	82%	82%	82%
40 000	55%	55%	55%		40 000	83%	83%	85%
WET YEAR
Monthly supply reliability (%)
operational vol.	N = 15%	N = 20%	N=30%
(m3/day)
100 000	78%	80%	81%
90 000	80%	81%	81%
80 000	81%	81%	83%
70 000	83%	83%	85%
60 000	85%	85%	86%
50 000	86%	86%	88%
40 000	88%	88%	88%

---

3.7. Climate change impacts

Climate change was analysed using the regional projections adopted by the Togolese Ministry of Environment in the Fourth National Communication on Climate Change [9], which provides scenario-based changes in temperature and precipitation for the Maritime Region up to 2080. Representative Concentration Pathway 6.0 (RCP6.0) was selected as the reference greenhouse-gas emission scenario because it is consistent with current policy trajectories and matches the 25-year durability horizon used in this study.

The Maritime Region reports an average temperature increase of about 2.75 °C by 2080 under RCP6.0 and a modest reduction in mean annual rainfall. These changes were applied to the historical daily climatic series (2003–2023) used in the soil–water balance: temperature was shifted by the projected increment and rainfall was reduced by the percentage change inferred from the national scenarios. The modified series were then used to rerun the daily soil–water balance, generating new recharge time series for the CT recharge areas, which were subsequently introduced into the aquifer-reservoir model to obtain future groundwater outflow series under climate change.

At the scale of the CT aquifer as a whole, the RCP6.0 perturbation leads to an overall reduction in diffuse recharge of roughly 15 % by 2080, which translates into a similar decrease in mean groundwater outflow to the sea. The shape of the seasonal signal remains largely unchanged: recharge and groundwater outflow are still concentrated in the main rainy season, but the volumes associated with both wet and average years are reduced. Dry years remain characterised by very low or even negligible recharge, so the system’s ability to recover from multi-year deficits does not improve under climate change.

When the climate-impacted groundwater outflow series are combined with the same demand scenarios used for present climate, supply reliability decreases but not uniformly across year types (Table 5). In wet years, the reduction in monthly reliability is modest because even reduced groundwater outflow volumes generally exceed demand for most of the year. In dry years, reliability deteriorates only slightly compared with the already low baseline, as these years are controlled more by the internal regulation capacity of the CT aquifer than by small relative changes in rainfall. In average years, however, monthly reliability decreases by about 16–18 %, closely tracking the reduction in recharge. This confirms that the CT aquifer remains a viable resource under RCP6.0, but only if the operational volume of the new well field is kept safely below the physical limit set by mean groundwater outflow and if operating rules explicitly account for inter-annual variability and climate uncertainty.

---

Table 5: Monthly supply reliability for the RCP6.0 scenario.

	DRY YEAR			DRY YEAR RCP6.0
Operational volume	Monthly supply reliability (%)			Monthly supply reliability (%)
	N = 15%	N = 20%	N=30%	N = 15%	N = 20%	N = 30%
(m3/day)
100 000	28%	30%	37%	21%	24%	26%
90 000	30%	35%	37%	25%	26%	28%
80 000	37%	37%	38%	28%	28%	33%
70 000	38%	38%	45%	29%	33%	38%
60 000	45%	45%	46%	35%	38%	40%
50 000	46%	46%	55%	40%	42%	36%
	AVERAGE YEAR			AVERAGE YEAR RCP6.0
Operational volume	Monthly supply reliability (%)			Monthly supply reliability (%)
	N = 15%	N = 20%	N=30%	N = 15%	N = 20%	N = 30%
(m3/day)
100 000	73%	75%	76%	52%	53%	59%
90 000	76%	76%	77%	55%	57%	59%
80 000	76%	77%	79%	59%	59%	62%
70 000	78%	79%	80%	61%	62%	64%
60 000	80%	81%	82%	64%	64%	66%
50 000	82%	82%	82%	66%	67%	69%
	WET YEAR			WET YEAR RCP6.0
Operational volume	Monthly supply reliability (%)			Monthly supply reliability (%)
	N = 15%	N = 20%	N=30%	N = 15%	N = 20%	N = 30%
(m3/day)
100 000	78%	80%	81%	73%	73%	77%
90 000	80%	81%	81%	73%	77%	78%
80 000	81%	81%	83%	78%	78%	78%
70 000	83%	83%	85%	78%	78%	82%
60 000	85%	85%	86%	80%	82%	82%
50 000	86%	86%	88%	82%	82%	85%

---

3.8. Well-field design concepts

Well-field design was guided by an operational definition of safe yield derived from the monthly groundwater outflow analysis, climate-scenario recharge projections and the reliability thresholds defined for urban supply. At aquifer scale, the mean renewable freshwater outflow is about 232,000 m³/d, of which roughly 81,000 m³/d corresponds to the Sévagan–Vogan sector and 151,000 m³/d to the Anfoin sector. After accounting for current abstractions, the design focused on the two sectors with the best balance between available resource, productivity and salinity constraints, namely Vogan and Anfoin. The reliability analysis tested total operational volumes between 40,000 and 100,000 m³/d, so the well-field layouts were dimensioned to remain within that range under conservative drawdown criteria.

On this basis, a set of 23 alternative well-field configurations was analysed with an analytically based drawdown model implemented numerically. In the Vogan sector, seven geometric configurations were evaluated using unit discharges of 1,500 and 1,800 m³/d per well, including two-line and staggered two- and three-line layouts, with tested spacing as low as 200 m. The final acceptable configuration in this sector was the T8 layout, with 13 wells at 1,800 m³/d, i.e. a sector abstraction rate of 23,400 m³/d. In the Anfoin sector, simulated layouts ranged from 16 to 40 wells and unit discharges from 2,200 to 2,500 m³/d. Sensitivity tests on linear “T” layouts showed that maintaining unit discharges of about 2,100–2,200 m³/d requires spacing of about 900 m or more, whereas spacing below 600 m reduces feasible unit discharge to less than about 1,500 m³/d. The final acceptable configuration in Anfoin was a concentric-ring layout (T4) with 35 wells at 2,200 m³/d, i.e. a sector abstraction rate of 77,000 m³/d.

Configurations satisfying the drawdown, desaturation and salinity-buffer criteria converge towards a combined scheme exploiting both sectors. The final conceptual layout therefore consists of 35 wells in Anfoin and 13 wells in Vogan, for a combined abstraction of 100,400 m³/d. This total is consistent with the upper end of the operational volumes tested in the reliability analysis, whereas the Anfoin sector alone (77,000 m³/d) closely matches the 80,000 m³/d scenario highlighted in the supply-reliability assessment. The Anfoin solution corresponds to a kilometre-scale concentric layout with inner and outer ring radii of about 1.0 and 1.5 km, respectively, while the Vogan sector is constrained to a smaller layout by the proximity of both the southern saline boundary and the northern negative barrier.

---

---

3.9. Methodological limitations

The methodology adopted in this study is appropriate for regional-scale resource assessment and preliminary well-field design, but it has some limitations that should be made explicit. Aquifer regulation was represented by a lumped linear-reservoir model at monthly time step, which is appropriate for estimating sector-scale renewable outflow and interannual variability, but does not reproduce short-term transient responses. Likewise, the well-field simulations were based on analytical drawdown superposition using simplified boundary conditions and spatially averaged hydraulic parameters. Therefore, the proposed layouts and operating volumes should be understood as a conservative planning framework rather than as a final design, and they should be refined during phased implementation using continued piezometric and salinity monitoring, additional pumping-test interpretation and operational feedback.

4. Conclusions

The results confirm that the Continental Terminal (CT) aquifer east and north of Lake Togo is the best solution that can reasonably support a new strategic well field for Lomé. While the Paleocene and Maastrichtian aquifers are regionally important, their current status makes them unsuitable as the main pillar of future supply: the Paleocene shows long-term piezometric declines of around 0.7–1.2 m/year in the Maritime Region, and the Maastrichtian is highly heterogeneous, with discontinuous sandy lenses and poorly constrained geometry. Against this background, the CT aquifer stands out because it combines a large recharge area (about 648 km² on the Attitogon plateau), significant saturated thickness and the absence of a regional depletion trend over the monitoring period.

The climate–recharge analysis highlights the key control on the whole system: recharge is not only modest in relative terms (around 130 mm/year on average, or roughly 15% of annual rainfall), but also extremely variable in time. Over 2003–2023, annual recharge ranges from virtually zero in dry years to more than 450–600 mm in the wettest years, with a coefficient of variation exceeding 100%. At the seasonal scale, approximately three quarters of annual recharge occur during the first rainy season (April–July), with a secondary contribution in the second rainy season. This strongly non-linear behaviour is typical of semi-humid tropical systems with pronounced soil-moisture deficits at the onset of the rains and has direct implications for how “safe yield” should be defined.

The lumped aquifer-reservoir model, fed by the monthly recharge series, provides a consistent link between this highly variable recharge and the smoother discharge behaviour of the CT aquifer. The calibrated recession index (around 110 days) indicates a moderate regulation capacity: the aquifer is able to store part of the recharge pulses and release them over several months, but it cannot fully compensate for multi-year deficits. The mean groundwater outflow to the sea over 2003–2023, about 232,000 m³/d (81,000 m³/d from Sévagan–Vogan and 151,000 m³/d from Anfoin), represents the long-term renewable resource that is currently lost to the ocean. Since existing abstractions amount to only ~13,000 m³/d in total, the CT system, in principle, offers a significant margin for new pumping. However, the variability of groundwater outflow – both seasonal and interannual – means that this margin cannot be exploited as if it were a constant, fully reliable flow.

The reliability analysis makes this point explicit. When monthly groundwater outflow (minus current uses) is compared with a range of constant operational volumes between 40,000 and 100,000 m³/d, the picture is very different depending on year type. In average and wet years (about 75% of the 2003–2023 sequence), monthly reliability for high operational volumes (80,000–100,000 m³/d) remains relatively high, typically around 73–81%, and most failures concentrate in the late dry season. In dry years, however, the same operational volumes lead to reliability values of only 28–37%. In other words, the system is able to deliver large volumes in most years, but becomes fragile and failure-prone in the driest quarter of the time series. This clearly shows that safe yield cannot be derived from mean annual groundwater outflow alone: it must incorporate an explicit tolerance to failure and a transparent view of how the system behaves in dry years.

The climate-change analysis under RCP6.0 reinforces this conclusion. A projected increase in temperature of about 2.75 °C and a modest decrease in rainfall by 2080 translate into an estimated reduction of diffuse recharge on the order of 15%, and a similar decrease in mean groundwater outflow. The seasonal pattern of recharge and discharge remains essentially unchanged – still concentrated in the main rainy season – but the volumes in average and wet years are reduced. When the climate-impacted groundwater outflow series are combined with the same demand scenarios, reliability decreases modestly in wet years and only slightly in already critical dry years, but it drops by roughly 16–18% in average years. Thus, climate change does not fundamentally alter the functioning of the CT aquifer, but it narrows the safety margin and makes any design based on optimistic assumptions riskier.

Within this context, the well-field design work plays a central role in translating hydrogeological understanding into operational rules. The 23 simulated configurations, combining different numbers of wells, individual discharges and layouts in the Anfoin and Vogan sectors, show that it is possible to approach the upper part of the operational range while keeping key constraints under control: maximum drawdown in production wells, residual saturated thickness in the recharge area and minimum head margins against the saline wedge. The most extensive configuration hosts about 35 wells in Anfoin and 13 in Vogan, arranged in circular or elliptical schemes.

However, the design is not a static “once and for all” solution. Its robustness depends on the implementation of a monitoring and control system capable of detecting and managing deviations from the expected behaviour. Long-term piezometric and electrical-conductivity monitoring in dedicated observation wells will be essential for tracking the position of the saline wedge and for verifying that heads remain above the control thresholds defined in the operational criteria. The proposed monitoring network and threshold-based rules are therefore an integral part of the safe-yield concept..

Overall, the study underlines that the CT aquifer west and north of Lake Togo offers a realistic and technically robust option to support the expansion of water supply to Lomé and its metropolitan area, but only if abstraction is framed within a dynamic concept of safe yield. This concept must combine: (i) a physical limit based on long-term mean groundwater outflow; (ii) probabilistic reliability targets that explicitly recognise the behaviour of dry, average and wet years; (iii) an allowance for climate-change impacts; and (iv) an operational design that integrates monitoring and adaptive control. Without these elements, there is a clear risk of repeating, in the CT system, the overexploitation patterns already observed in the CT aquifer in the Lomé area.

Acknowledgements: The authors wish to acknowledge the Togolese Ministry of Water and Sanitation and the regional services for their support and provision of data, as well as ANAMET for meteorological series used in this work. We would also like to express our gratitude to Mr. Koumayi Nathan Assoutom, project coordinator at the Ministry of Water and Sanitation of Togo, for his indispensable role in facilitating the project, and to Dr. Germain Bleza, of the Société Togolaise des Eaux, for his key contributions to the hydrogeological knowledge of the region.

Conflicts of Interest: The authors declare no conflict of interest.

---

References

[1] Fahy, J.C.; Honyiglo, L. 1974. Estimation of resources of the Continental Terminal aquifer in the Maritime Region of Togo.

[2] UNDP (United Nations Development Programme). 1975. Etude de Prospection des Eaux souterraines dans la zone cotiere du Togo (DP/UN/TOG-70-511-1). Technical report, Lomé, Togo.

[3] SAFEGE. 1991. Etude de l’etat des ressources en eau souterraine dans le bassin sedimentaire cotier du Togo. Technical report, Lomé, Togo.

[4] Cabinet Merlin. 2007. Etude de faisabilite pour l’amelioration de l’alimentation en eau potable de la ville de Lome. Technical report, Lomé, Togo.

[5] MEHV. 2013. Outils de gestion du bassin sedimentaire cotier: Rapport de modelisation des aquiferes du bassin sedimentaire cotier. Ministere de l’Eau, de l’Hydraulique Villageoise et de l’Assainissement, Togo.

[6] Gnazou, M.D.T.; Sabi, B.E.; Togbe, K.A.; da Costa, Y.D.; Agouda, K. 2015. Actualisation structurale de l’aquifère du Paléocène dans le bassin côtier du Togo. European Scientific Journal (ESJ), 11(27), 101-122.

[7] MEHV / SP-EAU. 2022. Etude du Plan Directeur de l’Alimentation en Eau Potable du Grand Lome (PDAEPGL): Etude des Ressources. Technical report, Togo.

[8] Bleza, A. 2023. Contribution a l’etude de l’aquifere du Continental Terminal du bassin sedimentaire cotier du Togo: approche multicritere. PhD thesis, Universite de Lome, Togo.

[9] MERF. 2022. Quatrieme Communication Nationale sur les Changements Climatiques. Ministere de l’Environnement et des Ressources Forestieres, Togo.

[10] World Health Organization. 2022. Guidelines for drinking-water quality: fourth edition incorporating the first and second addenda. Geneva: World Health Organization.

[11] Sodomon, A.K., Akpataku, K.V., Tampo, L., Mande, S.-L.A.-S., Benavente Herrera, J., Martín Rosales, W., Faye, S., 2025. Assessment of hydrogeochemical evolution of groundwater from the basement aquifer in the upper part of transboundary Mono River Basin, Togo. J. Hydrol. Reg. Stud. 58, 102200

[12] Gnazou, M.D.T., Bawa, L.M., Banton, O., Djanéyé-Boundjou, G., 2011. Hydrogeochemical characterization of the coastal Paleocene aquifer of Togo (West Africa). International Journal of Water Resources and Environmental Engineering 3(1), 10–29.

[13] Barry, R., Barbecot, F., Rodriguez, M., Mattéi, A., Djongon, A., 2024. Estimating Climate Change’s Impacts on the Recharge of an Ungauged Tropical Aquifer (Togolese Coastal Sedimentary Basin). Water 16(5), 731

---

This article has been published in European Geologist journal 61 – 5th IPGC Special Edition 2