OR/14/068 Discussion

Lapworth, D J, Gopal, K, Rao, M S, and MacDonald, A M. 2014. Intensive groundwater exploitation in the Punjab — an evaluation of resource and quality trends. British Geological Survey Internal Report, OR/14/068.

Groundwater flow and recharge processes: evidence from natrual and anthropogenic tracers

A range of natural and anthropogenic tracers have been used in this study to understand, groundwater flow processes, anthropogenic contamination, and the natural geochemical evolution of groundwater in the Bist-Doab. Stable isotope results confirm the largely diffuse meteoric sources of recharge across the catchment in both the shallow and deep groundwater, which have significant overlap in terms of δ¹⁸O vs δ²H values. However, for most sites there is a significant difference between sable isotope values for the paired deep and the shallow groundwater, with deeper sites showing isotopically depleted signatures relative to the shallow samples. This is consistent with different recharge zones and processes for the paired sites at any given location, with the deeper sites have a greater component of water that was recharged some distance up-gradient (i.e. towards the recharge zone at the foot of the Shiwalik range). This source has a relatively depleted isotope signature due to Raleigh distillation processes (see Figure 10). The values for the deeper sites (δ¹⁸O values of ca.-7 ± 0.5) are consistent with continental scale depletion in meteoric recharge as monsoon air masses track from the Bay of Bengal (δ¹⁸O values of ca.-4) in a NW direction, and the inland gradient of -2 per mil per 1000 km (Krishnamurthy and Bhattacharya 1991). There are also a small number of sites where the isotopic signatures for the shallow and deep sites are not significantly different, this sugg ests that there is a common local source or (mixtures of sources) for both depths and implies significant vertical leakage and mixing which may be enhanced due to pumping. This marks an important new finding and highlights the use of stable isotopes to delineate regional and local flow processes.

There are some shallow sites where there is evidence of recharge from fractionated sources of recharge, perhaps as a result of ponding, as well as surface water replenishment of aquifers adjacent to rivers (e.g. Amritpur, R Beas) as a result of rising river stage in the post monsoon period. This is consistent with recent results presented by Sharma et al. (2014) which showed between 40–70% surface water recharge in shallow aquifers adjacent to the R. Beas at Naushera Pattan and Amritpur based on δ²H values. Based on the results for shallow and deep sites, and comparing them to the published values for the Sutlej Canal (with a significantly depleted signature), there is no evidence of the Sutlej canal water being a significant component of recharge regionally — in fact the deep sites are more depleted than the shallow sites. However, this does not rule out significant recharge to shallow groundwater at locations close to canals or where canal leakage water is likely and canal water may be a significant component of irrigation water. The isotope values from the Beas River and the Kandi canal system, in the north of the Bist-Doab, overlap significantly with the groundwater isotope values, this means that delineating the influence of canal water in this region is not straightforward.

Trace elements were shown to be effective natural tracers of groundwater evolution as a result of mineral dissolution processes. For example, Sr, Mo, As, U, Mg/Ca all showed overall trends of increasing concentrations with increasing MRT. However, the trends showed a lot of scatter, suggesting perhaps that mixing processes, i.e. the convergence of groundwater with multiple residence times, is important in controlling trace element concentrations. Shallow sources of Se from soils rich in Se, observed across some in parts of the catchment (e.g. Dhillon and Dhillon 2003 and references therein), have been suggested as possible sources of high Se in shallow groundwater in these previous studies, it is also possible that high Se could also be associated with fertiliser use. This hypothesis was corroborated by evidence from this study which show enhanced Se concentrations in shallow groundwaters, although redox and pH controls are also likely to be locally important in mobilising trace elements and enhancing concentrations in shallow groundwater. Enhanced heavy metal (e.g. Pb, Cd, Cr) Cu, Ni and Zn in shallow groundwater compared to deeper aquifers, also suggests migration of anthropogenic sources of contamination in to the shallow aquifer system, most likely from waste water sources (Gopal et al., 2014).

Nitrate, Cl and SEC values are significantly higher in shallow groundwater due surface anthropogenic contamination from agriculture (e.g. fertiliser) and waste water sources. While there is evidence for denitrification in some sites (NO₃<0.03 mg/L) low concentrations were only detected in 18% of samples. Groundwater in this region is moderately oxygenated (mean DO of 1.2 mg/L) and are oxidising (mean Eh of +256 mV), suggesting that NO₃ is a useful tracer of modern recharge and contamination in this setting. Shallow groundwater NO₃, Cl and SEC values show an overall downward shift in post monsoon conditions suggesting that there is a rapid dilution due to meteoric monsoon recharge. Overall the deep sites show consistent median values pre and post monsoon.

CFCs have been employed to trace recent recharge and estimate groundwater MRTs across the Bist-Doab. While there was evidence of contamination/degradation by CFCs at some sites, CFC-11 was much more affected by this than CFC-12. Where degradation or contamination were not occurring, there was generally good agreement between CFC-11 and CFC-12 MRTs and calculated modern fractions, providing important support for the interpretation of the CFC-12 data. The generally oxidising nature of the groundwater in this catchment also means that the use of CFC-12 as a conservative tracer is more reliable as CFC degradation can occur in sub-anoxic environments. Overall, shallow groundwater showed significantly younger MRTs compared to deep sites, as might be expected, although it is noteworthy that CFC-12 was detected in the majority of deep sites suggesting at least a component of modern recharge at depth. Even where concentrations were found to be greater than estimated modern recharge, indicating a potential source of local contamination, this data can still be used as a sensitive diagnostic tracer of connectivity between the shallow and deep aquifers and pathways for modern recharge.

Mean groundwater residence times in the shallow aquifers show a large range (0–>50 years) with average values of 29 years and 30 years under post-monsoon and pre-monsoon conditions respectively. Using Darcy’s law to calculate groundwater flow (Q) and literature values representative of alluvial sediments for hydraulic conductivity of 10–30 m/d (Rushton 1986); porosity of 0.2–0.3 (Todd 1959) and a regional gradient of 0.0004 (Bowden 1985) residence times of the order of ca.10³–10⁴ y in the deep aquifers (100–150 mbgl) 50 km from the recharge zone under natural flow regimes are estimated. Deep groundwater (>100 mbgl) have mean MRTs of 42 years, suggesting that the natural flow regimes for this aquifer system are highly perturbed by pumping, the young MRTs imply a significant component of recharge from vertical leakage induced by pumping from depth.

Implications for long-term groundwater security

There is clear evidence from historical groundwater level records that there has been a large decline in groundwater levels in shallow aquifers used for irrigation at a regional scale (ca.20–25% of the Bist-Doab) over the last 20 years. Natural flow regimes and recharge in the shallow groundwater system are highly perturbed by the sustained pumping for irrigation, as shown by the MRT and stable isotope results. Work by Datta and Goel (1977), using tritium techniques, also showed that areas with irrigation had significantly enhanced recharge compared to non-irrigated locations. In many cases pre-monsoon groundwater levels are now >20 mbgl and in some areas falling at ca.0.5 m per year, which has potential cost implications for long-term use of these shallow aquifers for irrigation. The area most affected by over-pumping for irrigation is the region SW of Adampur (i.e. zone 2 in Figure 3), the worst affected sites include Jalandhar, Phagwara, Nakodar and Shahkot, in the middle Doab and confluence, where pre-monsoon groundwater levels have declined by >20 m in the last 20 years.

There is evidence from the hydrographs that pumping and the declining pre-monsoon water levels are actually enhancing net recharge in the middle Doab and confluence region. However, in many of the sites in this region the post monsoon maximum are on a downward trajectory suggesting that net abstraction is outstripping the actual recharge.

A significant number of sites in the region of greatest groundwater decline (see Figure 7) show modern or over-modern CFC-12 recharge concentration suggests that the response in the shallow aquifers post-monsoon is due to rapid modern recharge. This has implications for the security of these shallow aquifers for i) sustained pumping towards the end of the pre-monsoon period, ii) the rapid migration of contaminants to depth within the aquifer, by-passing natural attenuation processes, and iii) inter-annual variability in rainfall and recharge during the monsoon. The higher groundwater levels (see Figure 7) and modern CFC values obtained from shallow and deep sites at Aima Mangat (most northern site) suggest that recharge in this region may be from shallow modern sources, perhaps due to the larger density of canals used for surface water irrigation in this part of the catchment or due to leakage from the Pong Dam. The similar enriched isotope values (δ¹⁸O values of ca.-5.4 to -5.7) for shallow and deep groundwater at this site suggests a common source of recharge at both depths and significant connectivity between the shallow and deep aquifers. The southern part of the Bist-doab also has a relatively high density of canals originating from the R Satluj, which could be enhancing overall recharge in this region. However, the stable isotope data suggests that this is probably insignificant compared to shallow groundwater irrigation and enhanced monsoon recharge from meteoric sources as the distinct depleted δ¹⁸O and δ²H signature (<-10 and <-70 respectively) of R Satluj (see Figure 9) and Rao et al., 2014) is not evident in any of the shallow groundwater in this region.

There is new evidence from groundwater dating using CFCs that some of the deep aquifers are also being replenished by a significant component of more recent recharge. This is most pronounced in the region of groundwater decline in the SW of the catchment where MRTs shift by up to 20 years to younger values in the deep sites post-monsoon, at two sites in the catchment post monsoon recharge is effectively modern recharge. There are two obvious explanations for this, either the deep abstraction sites at these locations are poorly constructed and what we are seeing is by-pass flow along the casing and the MRTs are an artefact of the borehole completion. The detailed logs and construction details available at some municipal sites suggest that these are well constructed sites which case out all but the lower portion of the borehole. Alternatively, the thick horizons of lower permeability material are leaky and allow significant vertical movement of groundwater from shallow aquifers to depth due to pumping. The fact that there is significant overlap between the region of long-term significant groundwater decline and the observed shifts in MRT post-monsoon at shallow and deeper sites suggest that it is unlikely to be just a case of poor borehole construction and is in fact an anthropogenic signal of pumping induced recharge and vertical leakage. One explanation for this is that there is a high degree of lateral variation in vertical permeability in the confining layers between the shallower and deeper aquifers, or indeed they may be discontinuous over relatively short distances. This results in zones or windows of high permeability within the confining horizons allowing rapid vertical migration of younger recharge to depth (e.g. Rushton, 1986; Rushton and Tiwari, 1989). These variations in vertical permeability are not likely to be a significant control on vertical groundwater flow under natural flow regimes in this mid-plains aquifer setting, due to the low relief and low hydraulic gradients, however they may become significant if an aquifer system is stressed by sustained pumping in both the shallow and deep aquifers.

Natural and anthropogenic impact on groundwater quality

The shallow aquifers in this catchment are vulnerable to anthropogenic contamination from both agricultural (e.g. NO₃) urban sources (e.g. NO₃, heavy metals) and natural sources (e.g. Se in soils). In some shallow groundwater these contaminants are approaching or exceeding WHO guideline drinking water limits, e.g. for NO₃, Pb and Se. This shows the potential for contamination from surface sources are being flushed into the shallow aquifer, exceeding the capacity of the shallow aquifer to fully attenuate these contaminants through natural mechanisms such as sorption, dilution and denitrification. The enhanced pumping and resulting decline in water levels in some regions has meant that net recharge to the shallow aquifer has been enhanced, facilitating the rapid migration of recharge to depth within the shallow aquifer.

The leaky nature of the lower permeability horizons which separate the aquifer systems, demonstrated using a range of environmental tracers, means that the deeper groundwater is potentially vulnerable to vertical breakthrough of contaminants from shallow aquifers. There is evidence that this is already leading to nitrate contamination of some deep groundwater sources due to pumping enhanced vertical movement of groundwater. This has implications for the likely current levels and future trends for contaminants such as nitrate other anthropogenic contaminants such as pesticides. While these contaminants are not pressing concerns for groundwater security today they will need to be addressed in the long term.

In addition, some of the shallow groundwaters have SEC >1500 μS/cm, and have significantly higher SEC (p<0.05) compared to deep sites (see Figure 15), with potential implications for the use of this water for irrigation in the long-term due to the build-up of salts in the unsaturated zone. Elevated SEC in the shallow groundwater is likely due to the use of fertilisers and manure as well as evaporative effects due to irrigation in this semi-arid climate. The current levels of SEC are not prohibitive for irrigation, but trends in salinity build up in the shallow groundwater system need to be monitored. These results are comparable with trends in the deterioration of groundwater quality found in other parts of Indian Punjab, Pakistan Punjab and northern China due to irrigation in semi-arid environments (Kumar et al., 2007; Kijne 1995; Ò Dochartaigh et al., 2010).

Enhanced residence times and mineral dissolution within deep groundwater has resulted in more elevated trace element concentrations for elements such as As, Mo and U compared to shallow groundwaters. In Bist-Doab this results in median U concentrations >15 μg/L and as high as 70 μg/L in some instances (over twice the WHO provisional guideline value of 30 μg/L). There are potential implications for radon contamination, and radiological aspects of toxicity that warrant further investigation. Evidence from this study shows that As contamination is not a major groundwater quality issue within the Bist-Doab where all dissolved As concentrations were <10 μg/L, and median concentrations for both deep and shallow groundwater were <2 μg/L. Fluoride concentrations were below WHO drinking water limits of 1.5 μg/L in all samples.