Introduction

Nitrate (\({\text{NO}}_{3}^{-}\)) contamination in surface and groundwater is a global environmental issue of concern (Choi et al. 2007; Eppicha 2012). High inputs of nitrate to surface and groundwater have been attributed to intensive land use, application of N-containing organic and inorganic fertilizers, animal manure, and discharge of human untreated sewage as well as natural sources like soils, bedrock and atmospheric deposition (Xue et al. 2009; Elisante and Muzuka 2016). In Babati town, where this study was conducted, groundwater is a major source of drinking water. Previous studies conducted in areas of Manyara region, where Babati town is located, have indicated that nitrate pollution above the background levels of 10 mg/L is common, which threatens the water quality of the area (Hongoa 2014; Elisante and Muzuka 2017; Pantaleo et al. 2018).

In many communities, land use and land cover (LULC) change is inevitable, and it contributes towards pollution of water resources. The main anthropogenic drivers of LULC changes include deforestation, expansion of agriculture, conversion of wetland to pastureland and agriculture as well as urbanization (Miserendino et al. 2011). For example, in Europe, it has been observed that modifications in landscapes have been largely caused by political and socio-economic changes which took place during the first half of the 19th century (Bičík et al. 2015). Moreover, high and fast-growing population density in Africa has led to expansion of crop cultivation together with animal husbandry and consequent resettlement. These have been noticed as the major drivers of LULC changes (Tappan et al. 2004).

Integration of natural and human factors in the classification of LULC dynamics can be a complex mixture; apart from anthropogenic activities on LULC changes, natural factors such as climate change and variability have also contributed (Reid et al. 2000). A study conducted in the vicinity of Mount Meru, Tanzania, reported that many areas in Northern Tanzania have been converted to cultivated lands, grazing lands, human habitats, mining sites as well as industrial areas due to rapid increase in human population and consequent economic growth (Kitalika et al. 2017). Moreover, during the mid and late 1940s and in the late 1950s, many people migrated to Babati because of its fertile lands and easy access to the area’s natural resources. Huge land was cleared due to tsetse fly infestation, while most of the virgin land was converted into farmlands (Sandström 1995).

Land use and land cover changes are considered to be associated with changes in the quality of surface and groundwater which threatens human and environmental health (Choi et al. 2007; Namugize et al. 2018). Moreover, in most developing countries, sources of nitrate in surface and groundwater have not been well studied. Babati is an emerging town, with rapid population growth resulting into high demand for housing, crop farming, grazing lands as well as sanitation facilities; increasing the risk to water pollution (Sandström 1995; Hongoa 2014). Thus, it is imperative that sources and levels of aquatic pollutants be studied in such areas like Babati that will be the cities of tomorrow.

Intake of water with high nitrate concentration may cause several diseases including methemoglobinemia, gastric problems, thyroid gland retardation, even cancer (Tank and Chandel 2010; WHO 2011). Excessive loading of nitrate into lakes may cause eutrophication and hypoxia (Mitsch et al. 2001). According to WHO, the acceptable limit of \({\text{NO}}_{3}^{-}\) in drinking water is 50 mg/L. It is therefore important to determine the levels, distribution, and possible sources of nitrate in semi-urban aquatic environments in the developing countries of the world, especially in the rapidly growing urban areas.

The isotopic composition of N-nitrate (δ15N-nitrate and δ18O-nitrate) has been used to identify sources of nitrate because different sources of nitrate often have different isotopic signatures (Kendall and Aravena 2000; Chang et al. 2002; Mayer et al. 2002; Pardo et al. 2004). Synthetic fertilizers produced by fixation of atmospheric N2 such as urea, ammonium nitrate and potassium nitrate usually show δ15N values quite similar to that of atmospheric N2 i.e. levels between –4‰ and + 4‰ (Kendall and Aravena 2000). Organic fertilizers such as plant composts and animal waste have a wider range of compositions of δ15N from + 2‰ to + 30‰. Animal manure normally has high δ15N values in the range of + 10‰ to + 20‰. Widory et al. (2005) reported δ15N values for wastewater ranging from 10.3‰ to 23.5‰ and from 4.3‰ to 17.4‰. Nitrate derived from atmospheric deposition, has high positive δ18O-nitrate values ranging from + 50‰ to + 94‰ (Durka et al. 1994; Kendall et al. 2007). Nitrate derived from nitrification, for instance, in soils or derived from mineralization of aquatic phytoplankton and nitrogen fixing bacteria followed by ammonification and nitrification has δ18O-nitrate values between − 15‰ and + 15‰ that are dependent on the δ18O of water (Rock and Mayer 2004; Deutsch et al. 2006). Also, nitrate derived from manure and sewage have δ18O-nitrate values between − 15‰ and + 15‰ (Aravena et al. 1993; Wassenaar 1995). Since the values of δ15N-nitrate from different sources may overlap, then a combined use of δ15N and δ18O values of nitrate can offer a delineation tool for discriminating the sources of nitrate in surface and groundwater (Kendall et al. 2007).

This study aimed at identifying the contribution of land use practices towards nitrate contamination in surface and groundwater of a rapidly growing urban area in Tanzania. The study applied δ 15N-nitrate and δ 18O-nitrate to determine the sources of nitrate in surface and groundwater in Babati town. Samples collected from existing wells, springs, and Lake Babati were used to produce the needed information. Levels of major cations, anions and trace elements were also analyzed, and the results were compared against the WHO drinking water guidelines.

Materials and methods

Study area

Babati town is an administrative headquarter of the Manyara region, northern Tanzania. It lies between latitudes 3°–40S and longitudes 35°–36°E (Fig. 1). One of the most conspicuous landmarks of Babati town is Lake Babati. Lake Babati has diverse species of fish, water birds, and is known to be home to a sizeable population of hippos and other wild ungulates (Pantaleo et al. 2018). The lake is a source of water for various economic and domestic activities (Pantaleo et al. 2018). In the studied area, volcanic soils are dominant and range from sandy-loamy to clay and alluvial soils (Ghiglieri et al. 2012). However, alkaline soils may also be present in the lowlands such as those around Lake Babati (Nkotagu 1996).

Fig. 1
figure 1

Map of Tanzania (top left) indicating Manyara region (bottom left). Babati town and the respective sampling locations are shown on the right

Full size image

The climate of the study area is warm and relatively wet with a mean annual precipitation of 831 mm. The mean annual temperature is 20.2 °C. The area’s lithology was formed during the upper Mesozoic. Sedimentary rocks overlie a Karoo system and a Precambrian basement. These rocks comprise mainly of sandstone, limestone, shale, marl and local evaporates. Agriculture and livestock keeping are the primary livelihood activities in the study area. Smallholder farming includes maize, sesame, sunflower, horticultural crops and legumes. Furthermore, a recent study pointed out poor sanitation as one the major environmental and health challenges in the area (Pantaleo et al. 2018).

Water sampling

Surface water was collected from Lake Babati and a spring whereas groundwater was collected from boreholes and shallow wells. Sampling took place in the month of April 2018 and all water samples were collected during the same season. Each sampling location was geo-tagged using a handheld GPS unit and plotted on the Tanzania map using ArcGIS software. Twenty-one samples were collected from the lake, 16 from boreholes, four from shallow wells and one from a spring. All boreholes and wells were active therefore pre-pumping was not necessary during sampling. Lake samples were collected using a water sampler. Water samples for analysis of the levels of major ions and trace elements were filtered through a 0.45 µm filter paper and kept in prewashed 500 mL HDPE bottles. Samples for 15N-nitrate and δ18O-nitrate were collected in 30 mL HDPE bottles. All water samples were kept in cold condition and transported to the NM-AIST laboratory for analysis. Field measurements such as pH, temperature, EC, DO and TDS were carried out during sampling using an Eijkelkamp 18.52 SA Model (pH/mV/EC/T/sal/TDS/DO) multimeter.

Image acquisition and pre-processing

Three Landsat images were utilized in this study. Landsat 5 TM images captured on 02/01/1997, Landsat 7 ETM + images captured on 02/01/2008 and Landsat 8 OLI images captured on 17/2/2017. More information about satellite images used for the present study is shown in Table 1 below. Remote sensing imagery for the study area for 1997, 2008 and 2017 were downloaded from the USGS Earth Explorer for the Babati town. Landsat 5 data were obtained for 1997, Landsat 7 data were obtained for 2008 and Landsat 8 data were obtained for 2017 to examine land use changes over the past 20 years.

Table 1 Satellite data sources and detailed information for the analysis of LULC in the study area
Full size table

Satellite images required pre-processing to ensure that the primary object could be established into a more direct affiliation between acquired data and biophysical phenomena (Rawat and Kumar 2015). This pre-process was accomplished using ArcGIS Version 10.1 for georeferencing, mosaicking, and sub-setting of the image for the area of interest (AOI). Landsat 8 images underwent spatial sharpening using the panchromatic bands which resulted in images with a 30 m resolution. Further image processing analysis was carried out using ENVI 5.0. The image was displayed in natural color composite using a band combination of 3, 2, and 1 for Landsat 5 TM and 5, 4, 3, and 2 for Landsat 7 and Landsat 8. Maximum likelihood supervised classification was performed using several selected regions of interest (ROIs) based on defined classes which were agriculture, built-up area, water, bare lands, and forests area as shown below in Table 2.

Table 2 Classes defined according to the basis of supervised classification
Full size table

Accuracy assessment

This is a quantitative assessment of how effectively the pixels were sampled into the correct land cover classes. The accuracy assessment of an image classification was performed by creating the classification error matrix. In this confusion matrix, classification results were compared with ground truth data acquired during fieldwork. This was done by picking the menu item classifier > accuracy assessment, then importing the ground coordinates of the ground truth samples from an Excel file which was already saved in text format. According to Coppin and Bauer (1996), the accepted classification results should normally be > 70%. Table 3 shows the accuracy assessment results for the year 1997, 2008 and 2018 for image classification.

Table 3 Accuracy assessment results for Landsat images for years 1997, 2008, and 2017
Full size table

Analysis of water samples

Anionic concentration of \({\text{NO}}_{3}^{-}\) and \({\text{SO}}_{4}^{2-}\) was determined using a spectrophotometer while determination of the concentration of Cl−, \({\text{CO}}_{3}^{2-}\) and \({\text{HCO}}_{3}^{-}\) was done using titration method (APHA 2017). Determination of major cations (Na+, Mg2+, K+ and Ca2+) and trace elements (Fe2+, Pb2+, Zn2+, Cu2+, Mn2+, and Cd2+) was done using atomic absorption spectrophotometer. Trace elements and ionic concentration results were compared against the WHO guidelines. Water samples were randomly selected for detection of δ15N-nitrate and δ18O-nitrate from various water sources within the area. The removal of nitrite was done by using sulfamic acid (Granger and Sigman 2009) and kept in 30 mL HDPE bottles and shipped to the University of Waterloo, Ontario, Canada, for analysis. The δ15N–nitrate and δ18O-nitrate in 12 groundwater samples and 7 surface water samples were detected through conversion of nitrate to nitrite (Granger and Sigman 2009) using cadmium catalyst then chemically converted to N2O which was then analyzed on a Trace Gas–GVI IsoPrime-IRMS (TG-IRMS).

Statistical analyses

Paleontological Statistics (PAST, Version 3.08), MS Excel, and SigmaPlot software were used to perform all statistical analyses. Mean and standard errors were calculated for various parameters and significant relationships between various parameters were studied by using relevant correlations.

Results and discussion

Physico-chemical parameters

Understanding the physico-chemical quality of surface and groundwater is essential as it is an important factor in determining water suitability for human and environmental applications. Tables 4, 5 contain results of the physico-chemical laboratory and onsite analyses. Temperature of groundwater varied from 22.9 °C to 28.1 °C with mean value of 25.2 ± 0.3 °C whereby BH5 showed the highest temperature of 28.1 °C (elevation 1294 m.a.s.l.) and the lowest temperature in groundwater was 22.9 °C (elevation 1383 m.a.s.l.). This slight variation in water temperature may be attributable to the differences in altitude rather than aquifer type. The EC values in the present study indicated similar magnitude to findings of e.g. Yusuf (2007) (EC: 120–1860 µS/cm) who investigated the groundwater quality of Lagos city; Saana et al. (2016) (EC: 131–873 μS/cm) in borehole water samples from seven districts of Ghana; and those of Nkotagu (1996) (EC: 301–4740 µS/cm) in groundwater of a semi-arid region (Dodoma), Tanzania. In the present study, the relatively low EC values, EC: 701–2370 μS/cm for groundwater and 750–1068 μS/cm for surface water, may be attributed to low mineral content—thus water in the present study may be classified as freshwater (Saana et al. 2016).

Table 4 Physical, chemical, and isotopic composition parameters measured in the groundwater samples collected from the studied area
Full size table
Table 5 Physical, chemical, and isotopic composition parameters measured in the surface water samples collected from the studied area
Full size table

Furthermore, when raw data in Tables 4, 5 were used to plot correlations between EC and TDS, the R2 values were high—0.96 and 0.93, respectively (Fig. 2). This means that, generally, a high linear correlation existed between water conductivity and total ions’ concentrations. The Excel file containing raw datasets has been included as Supplementary Materials.

Fig. 2
figure 2

Linear correlation plots between TDS and EC values for (a) groundwater samples and (b) surface water samples collected at the study site

Full size image

The concentration of major cations in groundwater and surface water was in the following order Ca2+ > Mg2+ > Na+ > K+ (Tables 4, 5). Figure 3 is a Piper diagram that shows major ion composition in surface and groundwater of Babati. Both surface and groundwater indicated a calcium–magnesium bicarbonate water type. The chemical properties of groundwater showed hardness in all samples and were dominated by alkaline earth elements (Ca2+ and Mg2+), which exceeded the alkalis (Na+ and K+). During infiltration, groundwater may dissolve the CaCO3 and MgCO3 found in water rocks to add up to the concentration of Ca2+ and Mg2+ in water (Dehnavi et al. 2011). A significant negative correlation between Ca2+ and \({\text{HCO}}_{3}^{-}\) in surface water (Table 7; r = − 0.51, p = 0.018) may indicate that Ca2+ was constantly being replaced by Mg2+ meaning that MgCO3 or Mg(HCO3)2 were more dominant species rather than CaCO3 or Ca(HCO3)2 (Jahnen-Dechent and Ketteler 2012). This phenomenon is further proved by a significant negative correlation between Mg2+ and Ca2+ in surface water; with r = − 0.52 and p = 0.016 (Table 7). It further follows that the solubility of MgCO3 probably decreased with increasing surface water temperature and that is probably why Mg2+ was negatively correlated with temperature (r = − 0.58, p = 0.047; Table 7) (Dong et al. 2008).

Fig. 3
figure 3

Piper diagram for hydro-chemical composition for surface and groundwater samples collected in the study area. The bottom left ternary shows the cations while bottom right ternary shows the anions. The top diamond-shaped chart represents both anions and cations distribution in the water samples

Full size image

Usually, the Piper diagram (Fig. 3) reveals the analogies, dissimilarities and different types of waters (Utom et al. 2013). It shows the variation or domination of cations and anion concentration in water. It is made up with three parts, the lower left triangle showing the cations, the lower right triangle that shows anions and the diamond plot at the middle that shows the combination of the two. Results from the present study indicated that alkaline earth type of water (Ca2+, Mg2+) exceeded alkalis (Na+, K+) whereas weak acid (\({\text{HCO}}_{3}^{-}\)) exceeded the strong acids \({\text{SO}}_{4}^{2-}\) and Cl− which may be indicative of slight hardness in all groundwater samples. This was further corroborated by the Pearson correlation values in Tables 6, 7.

Table 6 Pearson correlation matrix for groundwater samples
Full size table
Table 7 Pearson correlation matrix for surface water samples
Full size table

Surface water temperature varied from 22.9 °C to 23.8 °C with a mean value of 23.4 ± 0.1 °C (Table 5). The pH of groundwater ranged from 7.1 to 7.8 with a mean value of 7.4 ± 0.1 (Table 4) while for surface water the pH varied from 7.3 to 7.7 with mean value of 7.5 ± 0.1 (Table 5). These pH values fell within the normal pH of the natural aquatic environments. The EC of groundwater varied from 657 to 2370 µS/cm, with a mean value of 1301.18 ± 119.74 µS/cm (Table 4). For surface water, the EC ranged from 750 to 1063 µS/cm with a mean value of 1027.43 ± 15.13 µS/cm (Table 5). The values of TDS and EC correlated well in both surface and groundwater (Fig. 2). For surface water, the correlation coefficient valued indicated that EC significantly correlated with Ca2+ (r = 0.71, p = 0.009) indicating that calcium was probably the dominant major cation (Table 7). This finding is further corroborated by the Piper diagram (Fig. 3). Moreover, in surface water, the EC was negatively but significantly correlated with \({\text{SO}}_{4}^{2-}\) (r = − 0.84, p = 0.0007); this would probably mean that \({\text{SO}}_{4}^{2-}\) was not the major anion and decreased with increasing dissolved solids in surface water (Table 7).

Nitrate concentration in groundwater ranged from 1.7 to 54.5 mg/L, with an average value of 11.5 ± 3.1 mg/L (Table 4). Only one groundwater sample (BH3) exceeded the WHO drinking water guideline of 50 mg/L (WHO 2011). Higher levels of nitrate in drinking water systems are known to cause health and aesthetic problems. Furthermore, nitrate did not have strong correlation with other anions or cations for both groundwater and surface water (Tables 6, 7). This may suggest that nitrate in the analyzed samples did not share the source with other ions. Surface water showed relatively low concentrations of nitrate ranging from 1.2 to 4.4 mg/L with an average value of 2.5 ± 0.2 mg/L (Table 5). The main nitrogen species in surface water are ammoniacal nitrogen and organic nitrogen. The concentration of sulfate in both surface and groundwater samples was low compared to the WHO guideline (250 mg/L). The highest concentration of sulfate detected in groundwater was 46 mg/L and the lowest was 1.0 mg/L with the mean value of 18.0 ± 4.1 mg/L (Table 4), whereby in surface water the mean value was 1.7 ± 0.2 mg/L; highest concentration was 4.0 mg/L and the lowest was 1.0 mg/L (Table 5). Chloride concentration was higher in groundwater samples than in surface water samples whereby the maximum chloride value in groundwater was 1706.79 mg/L, and the lowest was ~ 7 mg/L with a mean value of 447.54 mg/L (Table 4). The origin of the observed chloride in groundwater was probably linked to the application of inorganic fertilizers in the study area (Kelly et al. 2012). In surface water, chloride ranged from 0.5 to 34 mg/L with an average value of ~ 7 mg/L (Table 5). Bicarbonate in groundwater ranged from 648 to 1927 mg/L with an average value of 1171.57 ± 376.58 mg/L (Table 4) whereas in surface water it ranged from 939.6 to 1479.1 mg/L with an average value of 1257.2 ± 135.9 mg/L (Table 5).

In groundwater, calcium values varied from 7.0 to 314.2 mg/L with a mean value of 146.3 ± 90.9 mg/L (Table 4) while in surface water, calcium varied from 10.7 to 89.0 mg/L with a mean value of 50.1 ± 15.7 mg/L (Table 5). High levels of Ca in groundwater systems may be attributed to the dissolution of carbonate and gypsum rocks (Chenini et al. 2010). The desirable limit of calcium in drinking water as specified by WHO is 75 mg/L; it was observed that in five groundwater samples Ca levels had exceeded the limit set by WHO. Magnesium concentration in groundwater varied from 2.6 to 203.4 mg/L with a mean value 67.1 ± 65.2 mg/L while in surface water it varied from 19.0 to 80.0 mg/L with mean value of 50.8 ± 17.2 mg/L. According to WHO, the desirable limit of Mg is 50 mg/L meaning that eight groundwater samples, equivalent to 47%, exceeded the WHO limit. High levels of Ca and Mg are indicative of water hardness and imply that some water sources in the studied area may not be potable. Similar to Ca, high levels of Mg in groundwater may also be linked the dissolution of carbonate minerals and gypsum rocks and, most probably, the dissolution of nesquehonite (MgCO3 • 3H2O) (Dong et al. 2008; Chenini et al. 2010). Sodium concentration in groundwater varied from 5.4 to 21.9 mg/L with a mean value of 10.5 ± 4.2 mg/L, while in surface water it varied from 23.4 to 38. 4 mg/L with a mean value of 27.7 ± 4.2 mg/L. Potassium concentration in groundwater varied from 0.06 to 1.3 mg/L with a mean value of 0.6 ± 0.4 mg/L while in surface water it varied from 0.3 to 1.0 mg/L with a mean value of 0.7 ± 0.2 mg/L. A similar study was conducted in Chad Basin in Nigeria, whereby Ca–Mg–HCO3 was traced to water recharge and its sources were connected to rainfall and dissolution of silicate minerals (Nur et al. 2012). Thus, high levels of Ca2+ and Mg2+ in the present study could be linked to silicate minerals leached by weathering due to the action of weak carbonic acid (Nur et al. 2012). Interestingly, in groundwater samples of the present study, Mg was significantly correlated to bicarbonate (r = 0.58, p = 0.014) whereas Ca had a good correlation with bicarbonate (r = 0.47, p = 0.05) (Table 6). This is probably an indication that the groundwater of Babati and its suburbs are of Ca–Mg–HCO3 type. This could also be the reason why water samples collected from Lake Babati indicated a significant correlation between Mg and bicarbonate (r = 0.54, p = 0.011) (Table 7).

Trace elements

Iron (Fe) concentration in groundwater varied from 0.01 to 1.93 mg/L with a mean value of 0.28 ± 0.52 mg/L whereby BH2 and BH5 showed higher concentration of Fe than the aesthetic value set by WHO (Table 4). In surface water, Fe varied from 0.06 to 0.16 mg/L with a mean value of 0.095 ± 0.03 mg/L (Table 5). All the analyzed surface water samples had Fe concentration within the aesthetic value set by WHO—0.3 mg/L (Table 5). The concentration of Cu in groundwater ranged from 0.01 to 0.62 mg/L with a mean value of 0.17 ± 0.2 mg/L, while that of surface water varied from 0.001 to 0.06 mg/L with a mean value of 0.028 ± 0.016 mg/L. All the water samples were within the recommended WHO standard for Cu levels (Tables 4, 5).

For Zn, the concentration in groundwater ranged from no detection to 0.12 mg/L with a mean value of 0.047 ± 0.037 mg/L, while that of surface water ranged from no detection to 0.12 mg/L with a mean value of 0.028 ± 0.028 mg/L (Tables 4, 5). The concentration of lead (Pb) in groundwater samples ranged from no detection to 0.12 mg/L with mean value of 0.028 ± 0.033 mg/L while that of surface water ranged from no detection to 0.040 mg/L with a mean value of 0.017 ± 0.009 mg/L (Tables 4, 5). Because Pb is a known neurotoxin, especially to children aged < 5 y, the levels of Pb in some of the samples seem to be alarming. Manganese concentration in groundwater varied from no detection to 0.02 mg/L with a mean value of 0.01 ± 0.003 mg/L, and that of surface water varied from 0.01 to 0.2 mg/L with a mean value of 0.04 ± 0.05 mg/L (Tables 4, 5). All the water samples were within the WHO indicative aesthetic value for Mn. The concentration of Cd in groundwater ranged from no detection to 0.05 mg/L with a mean value of 0.02 ± 0.003 mg/L while that of surface water varied from no detection to 0.030 mg/L with a mean value of 0.02 ± 0.006 mg/L (Tables 4, 5).

In groundwater and surface water systems, in most cases, trace elements originate from the parent rock and sediments but may, sometimes, originate from anthropogenic loadings e.g. mining, industry, transport and agriculture (Ramesh et al. 1995; Molina et al. 2009). The levels of trace elements found in the present study were lower than the levels found in the drinking water from wells in small scale mining in Ghana (Cobbina et al. 2015) and were comparable to the levels reported by Ilyas and Sarwar (2003) in groundwater from Peshawar in Pakistan.

Isotopic studies

The values of ẟ15N-nitrate in groundwater ranged from 2.4‰ to 20.1‰ with the mean value of 11.8 ‰ ± 1.8 ‰ while in surface water ẟ15N-nitrate ranged from 2.1‰ to 12.6 ‰ with the mean value of 4.9‰ ± 1.4‰. High concentration of nitrate (> 10 mg/L) in groundwater BH3, BH5, SW2, and SW3 may be indicative of enriched values of ẟ15N-nitrate, which varied between 2.3‰ and 19.4‰ (Fig. 4). These values bear isotopic signatures of synthetic fertilizer (− 4‰ to + 4‰), manure (+ 10‰ to 20 ‰) and sewage (+ 9‰ to 19‰) (Aravena and Mayer 2009). For the present study, intensive agricultural activities were probably the major source of nitrate. In surface water, there was low concentrations of nitrate (Fig. 4) as well as low ẟ15N-nitrate and ẟ18O-nitrate values. This may be indicative of nitrate derived from soil N (Kendall 1998).

Fig. 4
figure 4

ẟ15N-nitrate values against nitrate concentration in three water populations

Full size image

The range of signatures of ẟ18O-nitrate in ground and surface water was low, indicating that probably nitrate was derived from nitrification (− 10‰ to + 10‰), fertilizers, manure, and soil N (0.8–5.8‰) (Fig. 5). However, mixing of water of different concentration and sources of nitrate affect the nitrate concentration as well as changes the isotopic composition of nitrate (Kendall et al. 2007; Aravena and Mayer 2009). Surface water can receive groundwater inflow, recharge groundwater or both. Such movements of water between groundwater and surface water systems may lead to mixing of water qualities. Hence nitrate and other dissolved nutrients in surface water can be transferred to the connected groundwater system and vice versa.

Fig. 5
figure 5

ẟ15N-nitrate values against ẟ18O-nitrate values for three water populations

Full size image

Land use and cover change vs. nitrate enrichment

The LULC changes data show that the land use type in 1997 for built-up area was 65 km2 (14.4%), agriculture was 78 km2 (17.3%), water was 8 km2 (1.8%), forest was 73 km2 (16.2%) and bare land was 227 km2 (50.3%) (Figs. 6, 7). In 2008, the built-up area had doubled to 130.4 km2 (28.9%) whereby agriculture land also doubled up to 182.9 km2 (40.6%), respectively. However, the area covered by water had reduced 7.5 km2 (1.7%); forest cover also reduced to 57.9 km2 (13.2%) and bare land reduced by a factor of three to 70.2 km2 (15.6%). Data from 2017, indicate that the built-up area and agriculture land maintained the highest proportion, occupying 228.8 and 98.7 km2 (50.7% and 21.9%), respectively. The area occupied by water bodies has been reduced to 7.4 km2 (1.6%) whereas the forest area seems to have slightly increased to 63.1 (14%) km2 probably due to recent forest conservation efforts run by both the government and private sector. On the other hand, bare land has shrunk by a factor of four from its 228.8 km2 of 1997 to a mere 53.2 km2 (11.8%) probably due to increasing population and demand for settlement areas. These changes may have negative impacts on the quality of water and water sources.

Fig. 6
figure 6

The land use and landcover change for three sampled years of 1997, 2008 and 2017 in the study area

Full size image
Fig. 7
figure 7

LULC change detection for 1997 (bottom), 2008 (mild) and 2017 (top) of Babati town

Full size image

The presence of nitrate and other nutrients at high levels in natural waters usually indicates that there is anthropogenic influence on the water quality (Salvado et al. 2006). In the studied area, agriculture is a major economic activity followed by animal keeping. These activities may have contributed to the pollution risk on water resources in the area because of the possibility of pollutants being released from both point and nonpoint sources (Camara et al. 2019). Elevated levels of nitrate were detected in some boreholes and shallow wells located near agricultural and piggery project areas. Similar cases of groundwater pollution with high levels of nitrate have been reported in Babati and other parts of Tanzania (Nkotagu 1996; Elisante and Muzuka 2017; Pantaleo et al. 2018). The sampling point BH5 with a nitrate concentration of 54.5 mg/L was located in a lowland area surrounded by maize farms where the application of fertilizers was intensive. However, water samples from Lake Babati revealed levels of nitrate close to background values, indicating less influence of anthropogenic activities on the lake’s water quality.

Conclusions

Generally, for the present study, the analyzed ground and surface water samples revealed relatively low levels of nitrate and trace elements. Five groundwater samples had nitrate concentration higher than the background levels and one sample exceed the WHO nitrate limit of 50 mg/L. Such levels may be associated with human activities such as application of synthetic fertilizers and animal manure in farms around the study area. Chemical composition of ground and surface water from the study area is largely influenced by the aquifer lithology than anthropogenic activities. Interpretation of hydro-chemical results showed that calcium carbonate dissolution and ion exchange processes controls the major cations chemistry of Ca2+, Mg2+, Na+, and K+ as well as the anionic chemistry of \({\text{HCO}}_{3}^{-}\), Cl−, \({\text{NO}}_{3}^{-}\) and \({\text{SO}}_{4}^{2-}\).

The present study has observed land use change in the study area whereby huge proportion of the land has been converted into human settlement and a shrinkage in the agricultural lands due to rapid urbanization and population growth. Farming activities and animal husbandry seem to be a lead cause in nitrate enrichment in ground and surface water of Babati town whereby enriched values of δ 15N-nitrate were mostly found at points close to the agricultural fields and in areas characterized by livestock keeping.