A knowledge-driven Multi-Criteria Decision Making- Analytical Hierarchy Process based geospatial modeling for the delineation of fluoride contamination zones in groundwater,

This study presents a framework to delineate the potential fluoride contamination zones within 15 the groundwater of the study area by employing GIS coupled with Multi-Criteria Decision 16 Making-Analytical Hierarchy Process approach (MCDM-AHP). In this context, various 17 groundwater contamination controlling hydrogeo-meteorological factors thematic layers were 18 prepared in the GIS environment and assigned with an appropriate rating and weights. All the 19 selected influencing factors were overlaid using a weighted overlaid index approach after 20 normalizing the weights and ratings of the respective layers using the AHP technique and then 21 computed the fluoride contamination zones (FCZs). The obtained results indicate that 61.50% 22 (1101.82 km 2 ) of the total study area was delineated as a relatively safe zone (F< 1.5 mg/L) and 23 the remaining 38.50% (689.18 km 2 ) was demarcated under the unsafe zone (F>1.5 mg/L). The 24 proposed FCZs model corroborates a significant agreement of 85% with the 61 observed 25 locations and thus testify to the model reliability. hornblende series minerals, muscovite, actinolite and fluor-apatite


29
Fluoride is one of the essential and helpful elements for human health as it is associated with 30 bone mineralization and dental enamel formation (Rahman et al., 2020;Gonçalveset al., 2020;31 Mondal et al., 2009). Moreover, some health issues are common, especially in children, such as 32 deficiency in bone mineralization, dental enamel formation, dental caries, when the amount of  (Shanker et al., 2003;Thapa et al., 2017;Su et al., 2019;Narsimha et al., 2019). 49 Geological materials consisting of fluoride-rich minerals such as fluorite, apatite, wohlerite, 50 tourmaline, herderite, sphene, hornblende series minerals, muscovite, actinolite and fluor-apatite 51 can trigger fluoride enrichment in aquifer solution through prolonged rock-water interaction 52 (Todd and Mays, 2005;Rao., 2009, Adimalla andVenkatayogi, 2017). Several clay minerals, 53 namely smectite, illite and chlorite also contribute significantly to fluoride concentration in 54 solution (Tossou et al., 2017). In recent times, several researchers have opinioned that climate, 55 evaporation, adsorption-desorption, ion exchange, ion competition, alkaline environmental 56 conditions, lithology, and geomorphology of the area are the significant factors that can produce 57 fluoride enriched groundwater (Mondal et al., 2009;Luo et al., 2018;Vithanage et al., 2014). 58 The higher the surface drainage density more is the surface runoff of rainfall and scanty is the 59 infiltration to groundwater storage. and other parts of the world (Ando et al., 2001;Pillai and Stanley., 2002;Madhnure et al., 2007).

71
The present study area comprises thick granitic litho-units and initially, the elevated fluoride 72 levels in groundwater were reported in some locations of the study area by the Central Ground  The proposed study area is a part of Jamui district in the state of Bihar, India and lies between 91 North latitude 24 0 40' to 25 0 10' and East longitude 85 0 50' to 86 0 35' with a total extent of 1791 92 km 2 . The geology of the study area is composed of rocky upland, pediplain and alluvial plains.

93
Alluvial plains in the study area were formed by continental Quaternary deposits and is a part of 94 the Jamui Formation and is generally termed as "older alluvium" in Indian geology. The major 95 rock types which cover the region are quartzite, quartz-mica schist, biotite-muscovite schist, 96 granites, composite gneisses, pegmatite, amphibolite and quartz veins. The most potential 97 aquifers are common in the alluvial formation. The thickness of the alluvium gradually increases 98 towards the north and finally merges with the Gangetic alluvium, south of the River Ganga. The 99 total thickness of the alluvium ranges from 90 m in the northern part and is finally reduces to less 100 than 12 m in the southern part. Some other landforms like, escarpment, inselbergs and valley fills 101 are also present in the area. Humid climate dominates the area and is the driving factor 102 responsible for the weathering of the overlying mantle. Elevated fluoride concentration in the 103 groundwater of the study area might have been generated due to weathering of this fluoride 104 bearing material from granite, granite-gneiss, amphibolites and mica-schists and deposition of 105 same over the parent rock as a weathered mantle. A tropical southwestern monsoon climate 106 governs the area with an average annual precipitation of 1042 mm. Infiltration from monsoon 107 rainfall contributes to the major part of the recharge of aquifers.     In addition, the AHP-based MCDM approach is also a suitable evaluation technique to evaluate 142 the consistency of final output; accordingly, it reduces the conflict involved in the decision-143 making process. Based on the above facts, in the current research work, a combined approach of 144 AHP and GIS technique is used to delineate the fluoride contaminated zones of the area under 145 study. The chosen ten thematic layers were supposed to be accountable for enhanced fluoride 146 concentration in the study region. Hence, the influencing factors were weighted according to 147 their contribution to groundwater contamination, especially with respect to fluoride, field study 148 and keeping in mind review of past studies. A layer with a higher weight illustrates a parameter  (Table 2).

157
" Table 2 is about here" 158 Pair-wise comparison matrix (Table 2) constructed for the thematic layers is in following 159 accordance: .
. Table 2 and has been 163 exhibited by applying equation:
The normalized weight (wx) was calculated utilizing Eq. 3.
Then to examine the uncertainty, Satty (2004) proposed the principal eigenvalue and consistency 172 index and to examine the uniformity of judgment matrix, the consistency ratio.

173
Hence, for computing the consistency ratio (CR), the procedure adopted is as follows: (1)  In the above mentioned equation, n represents no. of thematic layers (n = 10 in this case) 178 whereas λmax represents the maximum eigenvalue of the pair-wise comparison matrix.

188
The normalized computed weights for all the thematic layers were finalized for further analysis.

189
Thereafter, the subclasses of thematic layers were re-classified in the GIS platform. The ratings 190 of subclasses of each thematic layer were allocated on a gradation of 1 to 9, according to their 191 relative significance with respect to the contribution to the groundwater fluoride contamination. respectively. The geomorphological pattern of the study area is shown in Fig. 2b.

234
Soil texture is one of the influencing factors considered for any groundwater-related studies as it 235 exhibits the infiltration rate of the soil media through which water flows and reaches the aquifer.

236
The coarser the soil media, the lesser will be the holding capacity and more will be the chances 237 of contaminants to reach the aquifer and vice versa. In the study area, coarse loamy soil extends 238 over an area of about 458 km 2 approximately, whereas fine loamy soil extends over an area of 239 about 1333 km 2 . Coarse loamy soil has been assigned higher weightage due to its high 240 infiltration rate. The soil texture map of the study area is shown in Fig. 2c.

241
The slope is another important factor in groundwater studies, that greatly influences the recharge 242 of groundwater, i.e., gentle to flat slopes will contribute to less runoff, thereby increases the 243 infiltration rate and groundwater contamination. On the contrary, segments of the study area, 244 with higher slope (steep slope) results in a higher runoff, lower infiltration rate and consequently 245 reduces the chances of groundwater contamination. According to the slope, the study area was 246 categorized into five classes such as 0-3%, 3-6%, 6-9%, 9-12% and more than 12% 247 encompassing over an area of approximately, 654 km 2 , 626 km 2 , 210 km 2 , 65 km 2 and 236.99 248 km 2 respectively. The slope map of the study area is shown in Fig. 2d. with 104 km 2 , 9-12 km/km 2 with 45 km 2 and more than 12 km/km 2 accounts for 9 km 2 264 approximately in the study area respectively. The drainage density map of the study region in the 265 Jamui district is shown in Fig. 2f. km 2 ) respectively. The rainfall distribution map of the study area is shown in Fig. 2g. promotes the infiltration of contaminants into groundwater. In the present study, lineament 296 density was classified into six subclasses i.e., 0-0.5 km/km 2 , which encompasses over an area of 297 about 1391 km 2 , 0.5-1 km/km 2 , 1-1.5 km/km 2 , 1.5-2.0 km/km 2 , 2-2.5 km/km 2 and more than 2.5 298 km/km 2 extends over an area around 157 km 2 , 61 km 2 , 51 km 2 , 33 km 2 and 91 km 2 respectively.

299
The prepared lineament density map is shown as Fig. 2j.

311
Based on the FCZ map (Fig. 3a), some parts of the study region belong to the safe category 312 covering an area of 1101.82 km 2 (61.52% of the total area). Groundwater from these zones is  (Table 3). Overall, 52 wells out of 61 wells with fluoride concentrations are having an agreement 343 concerning the respective fluoride contaminated zones. Therefore, it is prudent to conclude that 344 the map of FCZ zones predicts the occurrence of fluoride contamination with 85.24% accuracy.

345
" Table 3  as unsafe zones i.e., areas with more than 1.5 mg/L fluoride concentration in groundwater.

356
The main advantage of the current study was the integration of a number of the hydro-geo-357 meteorological thematic layers on a regional scale that can be successfully adapted to delineate  Xu, Y., Huang, H., Zeng, Q., Yu, C., Yao, M., Hong, F., Luo, P., Pan, X. and Zhang, A., 2017.

567
The effect of elemental content on the risk of dental fluorosis and the exposure of the 568 environment and population to fluoride produced by coal-burning. Environmental