OR/15/047 Groundwater abstraction and groundwater levels

MacDonald A M, Bonsor H C, Taylor R, Shamsudduha M, Burgess W G, Ahmed K M, Mukherjee A, Zahid A, Lapworth D, Gopal K, Rao M S, Moench M, Bricker S H, Yadav S K, Satyal Y, Smith L, Dixit A, Bell R, van Steenbergen F, Basharat M, Gohar M S, Tucker J, Calow R C and Maurice L. 2015. Groundwater resources in the Indo‐Gangetic Basin: resilience to climate change and abstraction. British Geological Survey Internal Report, OR/15/047.

The IGB alluvium aquifer system is subject to considerable pressure, both from land use and groundwater abstraction, and changes to the hydrology from rainfall, river flow and irrigation. In this chapter we present datasets on the current groundwater abstraction across the IGB, and discuss how this is driven by agricultural activity. We also present data on groundwater level variations across the basin, interpreted from collating existing datasets across the IGB and a new statistical analysis of groundwater level data.

Groundwater abstraction

Figure 12 shows a map of the estimated groundwater abstraction across the IGB alluvial aquifer system and Table 2 gives the volume of groundwater abstracted from each typology. The methodology for compiling this dataset is discussed in Chapter 2. The total groundwater abstraction is estimated as 205 km³ per annum. Approximately 122 km³ is estimated from within India, 48 km³ from the IGB within Pakistan, 34 km³ from Bangladesh and less than 1 km³ from within Nepal

Table 2 Estimated abstraction from each typology.

Typology	Annual Abstraction In 2010 (km3)
T1 The Piedmont margin	8.9
T2 Upper Indus and Upper‐Mid Ganges	72.9
T3 Lower Ganges and Mid Brahmaputra	38.5
T4 Fluvial influenced deltaic area	12.7
T 5Middle Indus and Upper Ganges	58.5
T6 Lower Indus	4.9
T7 Marine influenced deltaic areas	3
T8 Minor typologies	5.2

The main crops produced are rice, wheat, cotton and sugar cane. Figure 13 shows the different distribution of these crops across the aquifer and the most recent available information on the production of each crop (2011/12 and 2012/13). Production from the IGB comprises >50% of the production of rice and sugar cane in Asia and two thirds of the region’s wheat production. Rice, cotton and sugarcane are mainly grown in the kharif, and wheat and pulses during the rabi season. Both are irrigated. The third crop between April and June can also be irrigated, and the ability to cultivate this third crop may be related to the availability of irrigation water. Research using remote sensing in the Indus basin has indicated that actual water use by cotton is 500–650 mm, rice 350–470 mm, wheat 320–400 mm and sugar cane 840–1100 mm (Bastiaanssen et al. 2002) — in all cases at least 25% lower than the water requirements published by the FAO (Doorenbos and Kassam 1979).

This practice of deficit irrigation over much of the basin has significant implications for the water resources, not only reducing the overall volume of groundwater used, but also limiting the salt flushed into the groundwater, and retaining the salt within the soil. Vegetables are grown in many parts of the basin and have higher water demands.

Groundwater abstraction for irrigation is seasonal. For much of the central basin, abstraction is highest during the kharif season (June to October) even though rainfall and surface water are most available at this time. Abstraction reduces during the rabi season, and reduces again during February and May. Groundwater abstraction is generally increasing annually across the basin as access to pumps and energy increases, and surface supplies become less reliable. Although reliable figures are difficult to come by, it is estimated that abstraction could be increasing across the basin annually at a rate of 2–5 km³ per year (Shah 2007, 2009, CGWB 2011, Quereshi et al. 2008, FAO 2013).

Groundwater forms an important part of municipal supply across the IGB alluvial aquifer system, with many cities reliant on groundwater for much of their supply, including Delhi, Dhaka, and Lahore. For example abstraction in Dhaka is estimated at approximately 0.8 km³ per year (DWASA 2012) and Lahore 1.1 km³ per year (Basharat and Rizvi 2011). Since abstraction is localised to within the urban areas it leads to local unsustainability and falling water tables (Chatterjee et al. 2009)

Figure 12 Map of estimated groundwater abstraction across the IGB aquifer system in 2010. The total for the basin is 205 km³. Sources discussed in the Methods Section.

Figure 13 Distribution of rice, wheat, sugar cane and cotton production across the IGB (Portmann et al. 2010). Rice: 122 000 tons (Pakistan 9400, India 57 000, Nepal 5100, Bangladesh 50 500); wheat 94 000 tons (Pakistan 23 500, India 67 700, Nepal 1800, Bangladesh 1000); sugar cane 229 000 tons (Pakistan 60 000, India 161 400, Nepal 2800, Bangladesh 4500); and cotton 3015 tons (Pakistan 2215, India 765, Bangladesh 35 (source FAO STAT 2015, GOI 2013).

Groundwater levels

There has been much discussion about the change in groundwater levels across the basin with significant changes in groundwater level inferred from studies at different scales, using different data sources. Most notable are the estimates of groundwater‐level changes from terrestrial water storage based on large‐scale satellite measurements of gravity changes (e.g. Rodell et al. 2009; Tiwari et al. 2009; Shamsudduha et al. 2012; Murray 2013; Chen et al. 2014). The resolution of these satellite based gravity measurements are, however, too coarse a resolution (400 x 400 km) to capture the significant local variations in groundwater level changes present across the basin. These large‐scale remotely sensed studies are also often poorly constrained by local groundwater measurements, with some notable exceptions where good quality tubewell data are accessible (Shamsudduha et al. 2012). However, for much of the basin the quality of groundwater monitoring data is poor, and has not been easily accessible. There is also a paucity of deep groundwater‐level data (>100 m depth) across most of the basin region.

Within this current study we developed a map across the IGB alluvial aquifer of trends in groundwater levels using exclusively ground based measurements. This map has been developed by spatial interpretation of groundwater levels from various published national assessments as a basemap, which were reinterpreted by using a subset of monthly and quarterly water‐levels from >2300 shallow tubewells (0 – 100 m) across the basin to give the average depth of groundwater across the basin, and average seasonal variation in groundwater‐levels. The 2300 subset of monitoring points were a subset of higher quality water levels monitoring points that had undergone some QA screening. Using these data enabled narrower categories of groundwater level variation to be mapped. The results are presented in Figures 14 to 16.

Overall, groundwater‐levels are shallow across the basin, typically <5 m below ground level (bgl), with seasonal variation of a few metres. Against this, significant variations are evident, concurrent with areas of highest abstraction, and different recharge processes. In the Indian Punjab where abstraction is over double that in the rest of the basin, groundwater‐levels are significantly deeper — over 40 % of the borehole records indicating groundwater is 20–50 m bgl, with seasonal fluctuations significantly greater than 5 m. Linear long term trend analysis of monthly time series of groundwater‐level data, show this has equated to an average decline in groundwater level of up to 0.75 m/yr over the last 20 years — Figure 14. Downstream within the middle and lower Ganges, depth to groundwater systematically reduces to be <5 mbgl, with increasing rainfall recharge and lower abstraction — Figure 15. Long term trends of annual groundwater‐level change become increasingly weaker from ‐0.15 m/yr in the Middle Ganges, to ‐0.02 to +0.05 m/yr in the Lower Ganges. In the Indus, a similar systematic downstream change is observed, but driven by canal leakage effects in the arid climate — Figure 15. Long term trends of annual groundwater‐ level change from ‐0.45 m/yr in the Middle Indus, to >+0.10 m/yr in the Lower Indus, reflecting the much lower abstraction and water‐logging issues within the lower basin.

Greater detail to these trends can be seen in Figure 16, which presents 30 year time series rainfall and groundwater‐level data from typical individual borehole hydrographs for different parts of the basin. In addition to the general trends outlined above, clear seasonal groundwater changes relating to abstraction can be seen in the individual hydrographs, against the long‐term declining, stable or rising groundwater‐level trends in different parts of the basin. The time series data also indicate that the significant long‐term trend of groundwater‐level decline in the Punjab are unlikely to be driven by climatic change, since the long‐term average rainfall is consistent over the 30 year time series (see hydrographs A and B, Figure 16).

On a local scale, greater diversity is present within these large‐scale trends, highlighting the true complexity of the system, and the sensitivity of the system to differences in local abstraction and recharge. Within the Indian Punjab, for example, there are still some areas of rising long‐term trends in groundwater‐level (against the overall trend of significant depletion in the region) and most likely as a consequence of canal leakage. Within the Bangladesh and Bengal deltaic region where the deeper aquifer exists, the deeper groundwater is confined, and the piezometric head (groundwater‐level under pressure) cannot be simply be equated to increasing or decreasing recharge.

The recent observations of generally falling groundwater levels are part of a longer evolution of groundwater within the basin. Observation wells were first installed in irrigated parts of the basin in 1870 and the data for some wells are still available — Figure 17. These records demonstrate that for many decades, rising groundwater levels of up to 40 m was a major problem, due to the redistribution of river water from the major tributaries to the interfluve areas for irrigation. These rising groundwater levels caused water‐logging in many areas and institutions were developed in the early 20th century to help manage water‐logging and drainage. Figure 17 illustrates the current decline in groundwater levels now present in the Lower Bari Canal command area in Pakistan, due mainly to private abstraction for irrigation. In other parts of the IGB however, most notable in the Lower Indus, groundwater levels are still shallow and water logging remains an important concern.

Figure 14 Long term trends of groundwater‐level change from high resolution 25 year time series data sets.

Figure 15 Systematic changes in depth to groundwater and seasonal variance.

Figure 16 Thirty year time series rainfall and groundwater‐level data from typical individual borehole hydrographs for different parts of the basin. Monthly rainfall data is shown as deviation from the long‐term average.

Figure 17 Groundwater hydrographs from the Lower Bari Doab Canal irrigation area, Punjab, over 100 years. (Source: Basharat and Tariq (2013)).