Understanding the relationship between groundwater chemistry and corrosion for rural water supplies

Corrosion of handpumps has been a persistent problem in sub-Saharan Africa and has been documented for at least forty years. Corrosion can impact hand-pump functionality and contaminate water supplies. The rate and severity of corrosion is controlled by groundwater chemistry as well as the type and quality of metal pump components.
A range of indicators has been used to assess the risk of corrosion of metallic components of handpumps installed in groundwater. The Larson-Skold index, the Langelier Saturation index and the Ryznar Stability index are the most commonly used indicators. These indices assess different aspects of the corrosivity of groundwater. The Langelier and Ryznar indices examine the effects of pH, temperature, calcium concentration, alkalinity and salinity; the Larson-Skold index accounts for the influence of the corrosive ions sulphate and chloride in relation to alkalinity.
The simplest indicator used to assess the risk of corrosion is pH. Observational evidence from sub-Saharan Africa by Otto Langenegger in the late 1980s and early 1990s was used to conclude that, where the pH of groundwater is less than 6.5, galvanised iron or mild steel should not be installed. However, for the next forty years, these materials continued to be installed and the problem of corrosion proliferated. To date, corrosion of hand-pumps has been documented in at least 20 countries across sub-Saharan Africa.
The aim of this study, undertaken as part of the ‘Stop the Rot’ initiative, was to characterise the corrosivity of the groundwater chemistry and to predict the probability of corrosion and leaching of corrosion by-products into the groundwater, using existing data from the Hidden Crisis project. The Hidden Crisis project took place between 2016 and 2019 in Ethiopia, Malawi and Uganda and was an interdisciplinary project designed to examine the physical, technical and social determinants of hand-pumped borehole functionality across two surveys. Water samples were collected and analysed across both surveys.
Interpretation of water-chemistry data and corrosion indices shows that:
• most groundwater from the sites sampled had a tendency towards corrosivity, as indicated by the corrosion indices
• pH values of less than 7 were common in the dataset, further indicating that groundwater at the study sites was corrosive
• elevated concentrations of iron, manganese and lead can be associated with corrosion but, in the Hidden Crisis dataset, these appear to be derived largely from the aquifer rather than the hand-pump
• chromium, cadmium and zinc may be associated with hand-pump corrosion
Statistical models were constructed to predict the likelihood of corrosion based on the water chemistry results and using the observations of corrosion made during the Hidden Crisis project. The statistical models showed that:
• the most effective predictive models of corrosion for galvanised iron rising mains and rods use the full water chemistry dataset; however, such models are challenging to interpret, especially as the water chemistry parameters that are important in these models vary significantly between component type
• pH below 6.5 is the most effective binary indicator of corrosion for galvanised iron components:
o for galvanised iron rising mains:
• the corrosion probability is 61 % if pH is less than 6.5
• the corrosion probability is 30 % if pH is greater than 6.5;
o for galvanised iron rods:
• the corrosion probability is 61 % if pH is less than 6.5
• the corrosion probability is 50 % if pH is greater than 6.5;
corrosion can still occur when pH is neutral or alkaline; in these cases, corrosion indices and more comprehensive assessments of water chemistry are required
Based on these results, we recommend that:
• comprehensive assessment of major-ion water chemistry should be conducted as a default where budgets allow and when galvanised iron is intended to be replaced by stainless steel
• where budgets are more limited, the use of pH thresholds can provide a useful indicator of the risk of corrosion
• pH must be measured accurately using well maintained and regularly calibrated pH sensors
• any galvanised metalwork used for groundwater installations should be in accordance with recommended manufacturing guidelines
• galvanised iron components should not be installed in groundwater with a pH of less than 6.5
• where pH is over 6.5, an assessment of the wider tendency of groundwater to be corrosive should be made using corrosion indices
• at a minimum, groundwater chemical analysis should investigate the following parameters:
o pH
o temperature
o major ions: calcium, magnesium, sodium, potassium, chloride, sulphate and nitrate
Better monitoring and preventative maintenance of rural water infrastructure can reduce the impact of corrosive groundwater on handpump condition and functionality:
• replacement of galvanised iron components with corrosion-resistant alternatives
• regular monitoring of condition of downhole components
• replacement of parts at regular intervals (as recommended in the hand-pump user manuals)
• adherence to standards as laid out in international and national specifications