Assessment of Corrosion and Scale forming Potential of Groundwater Resources: Case Study of Dire Dawa City, Ethiopia

Almost 100% of the water supply of Dire Dawa City is from groundwater (including boreholes, dug wells, and springs). Recently, groundwater cause corrosion and scale problems to water distribution systems due to its content of dissolved ions that can cause public health and economic issues. The present paper investigates the corrosion and scale-forming potential of the groundwater in the city and visualize with mapping. Spectrophotometer, EDTA/Acid titration with calculation methods were used fors it water quality parameters analysis. GW Chart Calibration plot applied for the Piper diagram to categorizes the water types. Langelier saturation (LSI), Ryznar (RSI), aggressive (AI), Puckorius Scale (PSI), and Larson-Skold (LRI) indices were manipulated with Excel ® and visualized their spatial distribution using ArcGIS 10. The mean values of LSI, RSI, PSI, AI, and LRI obtained were 0.29 ± 0.28, 6.4 ± 0.5, 5.10 ± 0.48, 12.20 ± 0.24, and 1.4±1.57 respectively. LSI and RSI results indicate moderate to low scale-forming tendency of groundwater in most parts except the northeastern part with corrosive groundwater. Based on the AI value, the groundwater ranges low corrosion in almost all zones except the edge of the northeast and northwest region. PSI indicated the water tends to form salt-scale at a medium rate. The LSI results showed that chloride and sulphate are unlikely to interfere with the formation of protecting film except in northwestern and northeastern regions where localized corrosion might occur. In conclusion, in almost all distribution system of the city is affected by calcium carbonate scale formation. The groundwater in the northwest and northeast resulted in localized corrosion because of relatively high contents of chlorides and sulphates.

The purpose of the present study was to investigate the corrosion and scale forming potential of the groundwater source of Dire Dawa City using the water stability indices and visualize the geographical distribution on the studied area using GIS.

Geographical description and data collection
Dire Dawa City is located in the eastern part of Ethiopia taking the UTM coordinate system as a reference, the location of the city is between 804511m to 816913m west-east direction and 1059754m to 1067650m south-north direction as shown in Figure 1.

Water Quality Parameters
The samples were collected in pre-washed and rinsed polyethylene bottles.

Water Stability Indices
With the help of the groundwater chemical and physical parameter data, the corrosion and scale forming indices of LSI, RSI, PSI, AI and LRI were calculated Excel ® software for investigation of corrosion or water scale-forming. Summary of the most common indices interpretation and criteria for categorizing the stability of the water presented in table 1.

Langelier Index (LSI)
The Langelier index only shows the tendency of water to dissolve or form carbonate scale. it is only conducive to control the calcium carbonate content in systems (Langelier, 1936 T-temperature in ℃, TDS in mg/L, Ca 2+ and Total Alkalinity in mg/L as CaCO3, pH s is the pH in saturation state in calcite or calcium carbonate. The index could be efficient for systems at low water speed, TDS and temperature ranges (McNeill and Edwards, 2001;Singley et al., 1984).

Ryznar Stability Index (RSI)
While Ryznar index is experimental, it could only be used for (water inside the pipeline) with a velocity of about 0.6 m/sec. This index is often used in combination with the LSI to improve the accuracy of the prediction. Ryznar index is not suitable for the case of water saturation (Prisyazhniuk, 2007). RSI helps to monitor scale thickness.
RSI is calculated as follows.

Aggressive Index (AI)
The AI is a simplified form of the Langelier index and only approximates the solubility of CaCO3 and water acidity but the corrosion effect.
The index is calculated as A is Total Alkalinity and HC is Calcium hardness in mg/L as CaCO3 However, it can be a useful tool in selecting materials like asbestos-cement lining pipelines or treatment options for corrosion control (Von Huben, 1995).

Larson-Skold Index (LRI)
This index was developed by Larson and Skold (1958) for examining the extent of corrosion based upon the aggressiveness of chloride and sulphate in water and alkalinity as aggression reducer which is exposed to steel pipelines with light carbonic structure and cast-iron pipes (Singley et al., 1984;Singh and Chakradhar, 1998).
Larson-skold index is calculated as:

The Pockurius Scale forming Index (PSI)
PSI is based on buffer capacity. It is explanatory of the maximum quantity of scale that may bring in water. In this index, the use of equilibrium pH (pHeq) is more than actual pH.

RESULT AND DISCUSSION
According to Habteab and Jiri (2018)   The high concentration of bicarbonate, sulphate, and chloride causes the incrustation and corrosion action (Sawyer and McCarty, 1967;Aiman and Enab, 2007;Anon, 1983). The values of the groundwater physical and chemical parameters of sample points are presented in appendix 1, 2 and 3.

Total Dissolved Solids
Total dissolved solids (TDS) is mainly contents of inorganic salts and dissolved organic matter in groundwater (WHO, 2003). A TDS increases conductivity, as result, it facilitates corrosion electrochemical reaction. TDS may also affect the formation of protective films since it was related to the excess calcium-based bicarbonate, sulphate, and chloride. High TDS (>500) result in excessive scaling in water pipes, water heaters, boilers, and household appliances (WHO, 2003). The TDS higher than the desirable level (>500) were observed in boreholes and springs (Habteab and Jiri, 2018;Antonaropoulos, 2017 unpubl. data;Eyilachew, 2010;WWDSE, 2004).

Acidity (pH of water)
Water with pH > 9 tends to reduce corrosion and also tends to stabilize around pH = 6.5 (Health Canada, 2009). The mean values of wells and springs were found to be between 6.7 and 7.2 that were stable enough.

Alkalinity
Alkalinity serves to control the buffer intensity of most water systems. It provides a stable pH, as a result, prevents corrosion of lead, copper, and cement-based linings pipes. The mean values of alkalinity were between 248 to 440 mg/L. If the alkalinity level is higher than 100 mg/L, leads to scale-forming on a pipe wall. Also, the high alkalinity avoids point corrosion which is caused by rapid pH change of the water. The fact that bicarbonate more than 400 mg/l causes incrustation problems (Anon, 1983) that bicarbonate of the PW17, PW18, PW22, T-2, T-3 recorded more than the limit (<400 mg/l). A total of 66% of the samples, more than the limits (<400 mg/l), were all boreholes and dug wells. The groundwater of pH less than 8, had no concentration of carbonate (CO3 -2 ) so it was bicarbonate dominant (McDonald, 2006).

Hardness (Calcium and Magnesium)
Calcium and Magnesium are responsible for the hardness, originate from limestone and dolomite respectively. The values of calcium (157.6 to 568 mg/L) and magnesium ion (8.9 to 144 mg/L) were relatively high in both spring and wells that all samples were categorized as

Sulphate and Chloride
Sulphate is a strong corrosion catalyst implicated in the pitting corrosion of copper and lead (Health Canada 2009;Ferguson et al., 1996;Berghult et al., 1999). The mean sulphate values were between 68.8 to 1809.1 mg/l. The desirable sulphate ion is limited below 250 mg/L (Anon, 1983) to avoid incrustation in low pH. Only 35.4% of the samples were above the limit of calcium bicarbonate and sulphate with the highest value of borehole PW17, PW18, PW22, and Tome wells T-1 and T-3 in Police Meret and Legahare which are populated and oldest settlement of the city. Chloride is the most active and corrosive ion (Jones, 1996). 87.7% of the wells and springs samples were within the limit (200 mg/l). The aggressiveness of water containing high alkalinity, bicarbonate and sulphate resulted in lower corrosion since the larger percentage of the current carried by SO 4 2− or HCO 3 − competing with chloride. A chloride to sulphate mass ratio (CSMR) less than 0.7, is desirable in which the corrosion rate is low (EPA, 2016;Reiber et al., 1997;Nguyen et al., 2011). Value of CSMR of sample TW1, PW17, PW18, PW22, T-3 and T-2 were below 0.7 that the groundwater was less corrosive to galvanized metals. On the contrary, 76% of the well CSMR values were more the limit (> 0.7) that able galvanized corrosion.

Corrosion and Scale-forming Indices
Since a corrosive or scale-forming tendency of water is a complicated phenomenon, many attempts have been made to develop an index and several of the indices have been useful for prediction purposes (McNeill and Edwards, 2001). There is no single corrosion index applicable to all materials. Corrosion indices, particularly those related to calcium carbonate saturation, have given mixed results. The mean values of LSI, RSI, PSI, LRI and AI obtained as 0.29 ± 0.28, 6.4 ± 0.5, 5.10 ± 0.48, 2.06 ± 0.25 and 12.20 ± 0.24 respectively. The corrosion and scaling forming indices of the sample points were displayed in tables 2,3 and 4 above.
Mapping the spatial distribution of the corrosion indices in the study area was best interpolated with kriging method. To validate the map, the real indices values and the predicted ones was cross-checked.

Langelier and Ryznar Indices
The Langelier index values in all zones ranged from -0.37 to 0.98. According to LSI, all the samples were above zero with a minimum value related to point PW 18. Besides, majority of water sources (95%) were supersaturated which could form scale at moderate rate and consequently damaged drinking and other water-consuming and distribution systems. In this regard, as per figures 5 and 6, the groundwater that tendency to form a scale located in most   More than half of the boreholes and dug wells (59.7%) had a scaling tendency on pipeline and water operating equipment and the other 20.8% of groundwater source points were balanced. However, the other 19.5% of the groundwater points were corrosive to metals and deteriorated effects on cement-lined and plastic pipes. The Ryznar index, in figure 6, showed relative scale-forming and saturated north part. However, the edge south and northwest part groundwater had shown corrosion tendency. Langelier index, on one hand, is preferably used for still waters and is used for control of bicarbonate scale formation. In this case, TDS was calcium bicarbonate dominant groundwater. On the other hand, the Ryznar index may be used with moderate to hard water but does not apply to soft or saline waters.

Aggressive Index
The aggressive index mean values in all zones ranged from 11.59 to 12.93. Almost all (87%) data were nonaggressive (AI greater or equal to 12) groundwater. The other 13% exhibited a corrosion rate ranged low to moderate. The wells with minimum AI values were located in the northeastern edge and the maximum values in central and northwestern. AI does not incorporate temperature or TDS effects. The groundwater ranged from low to medium corrosion and its effect is only applicable to asbestos-cement pipelines.

Puckorius Scale forming Index
The mean values of the index in all zones ranged from 4.23 to 5.87. Based on the results, the water tended to be scale-forming at moderate in most of the regions especially northwestern © CNCS, Mekelle University 123 ISSN: 2220-184X and central parts. The groundwater has high calcium concentration and high alkalinity consequently possess high buffering capacity. The calcium precipitated from the oversaturation over the less of pH change. The high values PSI in the southern and northeast edges indicated that groundwater was from less scale forming and tended to dissolve any existing scales.

Larson-Skold Index
The groundwater that has high calcium and alkalinity (bicarbonate), possesses high buffering capacity (Sajil Kumar, 2019). So the groundwater aggressiveness containing sufficient buffering capacity and alkalinity is due to the increase in chloride and sulphate (Larson and Skold, 1958). northwest is populated and has poor municipal northwest is populated and has poor municipal waste management. Animal and human waste are also the main source of sulphate and chlorine (Eyilachew, 2010).
In summation, the calcium and magnesium, bicarbonate and sulphate dominant salt in groundwater caused a moderate scale as indicated by LSI, RSI and PSI. On the other hand, the groundwater was also the source of low corrosion problems as indicated by AI and LRI mainly due to high TDS and warm temperature. In northeastern and northwestern, the groundwater contained significant amount of chloride and sulphate that dangerous pitting corrosion might exist which is fast and localize according to Larson and Skold (1958). Geographically, the corrosion tendency of groundwater in Dire Dawa City increases from south to north direction. The scaling problem was more aggressive at the center of the city, where most of the groundwater sources exist.

CONCLUSIONS
Assessment of the corrosion and scale forming tendency of the groundwater in Dire Dawa City was performed using five water stability indices. These indices prove very helpful to the city public drinking water distribution systems and other industrial water utility supervisors, field engineers, and policymakers in monitoring and manage water corrosion and scale-forming inhabited and old settlements that the groundwater had localized corrosion tendency due to relatively high content of chloride and sulphate. A recent over-pumping of the drinking water also caused a sudden drop in the groundwater level. Consequently, this increased salinity of the water exceeds limits and also led to carbonates and sulphate precipitation in pipes and water supply systems. This increasing trend requires more protection of the water resources and implementation of water treatment systems.

ACKNOWLEDGMENTS
The authors would like to acknowledge Dire Dawa Water Supply Authority for their support on laboratory utilization, fieldwork and documents. Dire Dawa University, Engineering Unit (Water Sanitation Facility) laboratory utilization and Dire Dawa Institute of Technology for financial support. The authors wish to thank anonymous reviewers for their valuable suggestions and comments.

CONFLICT OF INTEREST
There is no conflict of interests.