Evaluation of recharge areas of Arusha Aquifer, Northern Tanzania: application of water isotope tracers

In Arusha urban, northern Tanzania, groundwater contributes about 80% of the water supply. However, elevated fluoride levels and evidence of anthropogenic pollution have been reported in the groundwater around Mount Meru which is a water source for Arusha urban. This study aims at understanding the recharge areas and flow pathways of groundwater in what has been a poorly monitored area. The study uses the isotopic ratio of oxygen and hydrogen to estimate the groundwater recharge area and flow pathway. The results show the recharge elevation of groundwater is between 1,800 and 3,500 m above mean sea level on the slopes of Mount Meru. The average fluoride contents in the study area are 5.3± 0.4 mg/L greater than the limits of 1.5 mg/L set by the World Health Organization (WHO) and Tanzania. The nitrate concentration of 83.9 mg/L at the lower elevation areas (<1,400 m above mean sea level) exceeds the 50 mg/L WHO limit. The relationship of F with δO and NO3 suggests the leaching of fluoride in high altitudes and dilution in lower altitudes.


INTRODUCTION
Shortage of adequate surface water sources, urban development, escalating per capita water consumption, and the influence of climate change have made groundwater the main source of water for domestic uses in many urban areas around the globe, including Tanzania (Li ; Foster Bousquet & Furey ). However, groundwater sources are often faced with various dynamics such as geochemistry, topography, geology, water-rock interaction, and anthropogenic activities (Mduma et al. ). In Arusha city, northern Tanzania, groundwater from wells and springs provides more than 80% of the freshwater used for both domestic and industrial purposes (Chacha et al. a). However, the greatest threats to maintaining freshwater supply in the city are the prevalence of fluoride contamination, a decrease in groundwater reserve, and degradation of water quality due to human activities in the potential recharge areas (Chacha et al. a). A comprehensive examination of the recharge areas and flow pathway of groundwater is a significant tool when considering sustainable groundwater resources management (Nakaya et al. ).
The Arusha urban water authority has recently devel- Several methods such as multivariate statistical analysis using chemical tracers, etc. (Bakari et al. ) can be used to delineate the recharge areas of groundwater. However, the stable isotopic ratio of oxygen and hydrogen (δ 18 O and δD) is the best method for tracing flow pathways and clarifying the groundwater origin and source areas (Nakaya et al. ) because these isotopes are naturally contained in the water molecule and cannot be modified by water-rock contacts (Kim ). Furthermore, the isotopic altitude effect is very useful for tracing and distinguishing groundwater recharged at high altitudes from that recharged at low altitudes (Li ; Nakaya et al. ). Various researchers have successfully used the isotopic altitude effect of the isotopic ratio of oxygen to determine the recharge areas and flow pathway of groundwater (e.g., Li ; Nakaya et al. Bouchaou et al. () reported that the variation in the isotopic values of groundwater is influenced by differences in the altitude of recharge areas. As the mean annual air temperature becomes low at high elevations, the composition in the stable isotope of water vapor in the atmosphere decreases by isotope fractionation (Farid et al. ).

).
Therefore, this study applied the stable isotope ratio method on groundwater and surface water samples from Mount Kilimanjaro watershed at an altitude >1,500 m.a.s.l.
The results of the study primarily provide critical information for the management of the groundwater resources in the study area. They might also help to indicate what could be happening in other areas and how it can best be studied.

THE STUDY AREA
The present study was conducted in urban areas of Arusha (including the Arusha district council) located on the south-western slopes of Mount Meru in the north-eastern part of Tanzania with an elevation of 1,400 m.a.s.l. (Figure 1).
According to the Tanzania population and housing census data of 2012, the approximated total population of the area is 739,640 (NBS ).
The area experiences a tropical climate with dry and wet seasons and the rainfall pattern is bimodal, with short rains between November and December and long rains between March and May or June with a total average annual precipitation of 842 mm (Chacha et al. b). The maximum temperature varies between 13 and 30 C with an annual mean value of 25 C and the area is characterized by a narrow variation of relative humidity (55-75%) with 924 mm annual potential evapotranspiration (Chacha et al. b).
The area is studded by volcanic deposits of variable ages and dumped alluvial residues (Ghiglieri et al. ). According to Chacha et al. (b), Mount Meru is the focal point of volcanic events in the area and the lava flow forms the main volcanic rocks such as basalts to phonolitic and nephelinitic tuff. These act as an aquitard, which restricts the infiltration of groundwater, and directs the surface run-off toward the lower slopes. Faults and fractures formed due to volcanic and tectonic activities act as groundwater conduits. Moreover, the study area is described by volcanic and sedimentary hydrogeological formation composed of rocks with minerals such as fluorapatite, natrite, halite, calcite, chabazite, nepheline, biotite, and illite. The geological properties in the area change with geologic time and the main groundwater aquifer is characterized by volcanic ash, pyroclastic deposits, weathered and fractured materials such as basalts, and phonolitic to nephelinitic materials (Ghiglieri et al. ; Chacha et al. b).

Sampling and analytical measurements
Groundwater and river water samples were collected throughout the study area ( Figure 1) in September 2018 for groundwater and from March to May 2019 for river water to document the spatial disparity in the isotopic signature of δ 18 O and δD. The groundwater samples (n ¼ 32), as presented in Table 2, were collected from springs as well as from public and privately owned deep and shallow wells close to the points of discharge to reduce the influence of evaporation and pollution from the atmosphere. The sampling points were as mapped in Figure 1 at altitudes ranging from 1,299 to 1,904 m.a.s.l. in Mount Meru watershed. The depth of the sampling wells from the ground surface ranged from 48 to 273 m for public wells and 11 to 26 m for privately owned wells. Most of the public wells were labeled with the wells' information, but the privately owned wells were not labeled and information was obtained from the owners. The groundwater levels were measured using a water level meter. The depth to groundwater level ranged from 2.7 to 93.2 m for public wells and 1.98 to 7.03 m for privately owned wells. One artesian well was also observed during the field survey.
The primary sources of groundwater recharge in the study area are precipitation and rivers. River water samples were collected from different points based on the ease of access within the study area to establish the isotopic composition of different recharge sources. To develop the local meteoric water line (LMWL) for Mount Meru watershed, rainwater samples were collected from five locations distributed in altitudes between 1,294 and 1,813 m.a.s.l., as presented in Figure 1. Also, the precipitation information of Dar es Salaam and Dodoma as presented in Table 2 were obtained from the International Atomic Energy Agency (IAEA) database. To trace the flow pathways and recharge areas of groundwater, water samples from (n ¼ 26) wells, (n ¼ 6) springs, (n ¼ 5) rivers, as well as rainwater samples

RESULTS AND DISCUSSION
Hydrogeochemistry Table 2 shows the physicochemical parameters and isotopic composition of δ 18 O and δD for groundwater, river water, and precipitation samples. The hydrogeochemistry analysis using the Piper diagram shows that the Na-HCO 3 water type dominates for both well, spring, and river waters except sample SP1 from spring which is the Na-Ca-HCO 3 water type ( Figure 2). The pH of the groundwater samples ranged between 6.18 and 8.56 with a mean value of 7.1 ± 0.09, indicating that the water was weakly acidic to alkaline, while pH values for the river water varied between 6.8 and The temperature of the groundwater samples varied between 17.2 and 25.5 C with a mean value of 21.1 ± 0.4 C, while that of river water samples varied between 13.4 and 18.2 C with a mean value of 15.9 ± 0.9 C. The EC of the groundwater samples varied widely, from 16.19 to 172.80 mS/m with a mean value of 70.6 ± 6.5 mS/m. However, the EC of river waters ranged within that of groundwater (16.19-54.50 mS/m), indicating stronger water-rock interaction in groundwater relative to river water. The concentrations of Cl À and Na þ suggest the dissolution of ions from rocks, which is also reflected in the EC (Table 1a), as also reported in other parts of the globe (Li The fluoride concentrations further showed correlations with Na þ (r ¼ 0.70, P < 0.05) and alkalinity (r ¼ 0.60, P < 0.05) as shown in Table 1a, suggesting that like Na þ , the fluoride contamination is natural through waterrock contact mainly from volcanic rocks. These findings are also in line with the suggestions by Kim & Kim () for the relationships of fluorides with Na þ and alkalinity in groundwater.
The results presented in Table 2 show a wide range of   It should be noted that the statistically significant positive relationship between well depth and elevation of sampling location (r ¼ 0.72, P < 0.05), as presented in Table 1b, indicates that most wells at higher altitudes are deep.

Stable isotopes of water
The stable isotopic values of the groundwater samples ranged from À6.0‰ to À4.0‰ for δ 18 O and from À31.3‰ to À17.7‰ for δD, whereas the isotopic composition of the river water samples varied from À6.3‰ to À4.9‰ for δ 18 O and from À34.7‰ to À23.0‰ for δD (Table 2). These results show that the river water samples had relatively narrow isotopic ranges and were more isotopically depleted than the groundwater samples, an indication of elevation effects since river water was sampled from high elevation areas (>1,700 m a.s.l) on slopes of Mount Meru (Table 2).
Similarly, the ratio between the isotopic values of δ 18 O and δD varied locally due to climatic and geographical differences in the area. Thus, they represent the LMWL.
This study estimated the LMWL for Tanzania (δD ¼ Dar es Salaam Rainfall 196104 55

Dar es Salaam Rainfall 196812 55
Dar es Salaam Rainfall 196902 55 Dar es Salaam Rainfall 196903 55 Dar es Salaam Rainfall 196904 55 Daar es Salaam Rainfall 196905 55 Dar es Salaam Rainfall 196906 55 Dar es Salaam Rainfall 196907 55 Dar es Salaam Rainfall 197001 55 to the slope of the African meteoric water line (¼7.48) and the LMWL for Tanzania (¼7.037). However, it is slightly less than the slope of the GMWL (¼8), an indication of the minimal influence of evaporative enrichment of precipitation during or prior to infiltration. Moreover, most of the groundwater data plotted close to the LMWL for Tanzania and the African meteoric water line, and according to Farid et al. (), this indicates that the groundwater in the study area is of meteoric origin (rain and/or snow).
The δ 18 O of groundwater showed a negative correlation (r ¼ À0.73, P < 0.0004) with well depth (Table 1b)   The concentration of NO 3 À at location 1 varied between 4.6 and 77.0 mg/L with a mean value of 22.4 ± 5.8 mg/L.
The groundwater at location 2 was observed to have relatively low pollution of NO 3 À , as the measured value ranged between 0.7 and 37.8 mg/L with a mean value of 11.9 ± 3.2 mg/L. A relatively wide range of NO 3 À contamination ranging between 1.9 and 83.9 mg/L with a mean value of 32.6 ± 10.6 mg/L was

CONCLUSIONS AND RECOMMENDATIONS
The altitude effect of the isotopic ratio of oxygen revealed that groundwater used in Arusha urban, northern Tanzania, comes from recharge areas located at altitudes between 1,800 and 3,500 m.a.s.l. The groundwater in the study area