Many studies have assessed the impact of irrigation on groundwater levels. Perhaps not surprisingly, they tend to show that increased pumping reduces the underground store of water, while increasing use of irrigation boosts evapotranspiration and makes more rainfall downwind. But none of these studies investigated the underlying behaviour of the aquifer. Does a depleted aquifer behave in the same way as a full aquifer? Could pumping groundwater change the nature of the system?

To investigate these questions Laura Condon and Reed Maxwell, both from the Colorado School of Mines in the US, used a hydrological model to simulate the impact of 20 years’ worth of groundwater fed irrigation, on the Little Washita Basin in Southwestern Oklahoma, US. This moderately sized basin – approximately 1600 square kilometres – is characterized by grasslands, rolling terrain and a semi-arid climate.

The researchers ran two simulations, one with irrigation and one without. In the simulation with irrigation they designated 385 cells to have farms, each with their own well. During the growing season the farms would pump water and irrigate crops if surface saturation was below a certain threshold, as real farms do. This differed from previous simulations, which have assumed constant irrigation demands.

Condon and Maxwell concentrated on water table depth and evapotranspiration – the two variables most likely to be affected by groundwater pumping and irrigation.

By comparing the irrigated scenario with the non-irrigated scenario they were able to show that pumping water and irrigating crops shifts the dominant timescales of variability and increases the temporal persistence within the system, making it less able to bounce back quickly from a change. The reason that irrigation has this effect is because it taps directly into the system, they believe.

"Previous research demonstrated that aquifers act as fractal filters, reducing the high frequency variability and increasing long-term persistence of noisy forcing variables such as precipitation," explains Condon, whose findings are published in Environmental Research Letters (ERL). "Here we show that seasonal groundwater pumping introduces additional high-frequency variability to the aquifer because variability in irrigation demand is translated directly from the surface to the aquifer, bypassing the natural infiltration processes."

Although the irrigation was localized (water was only pumped from the cells containing farms), the changes this caused impacted the entire basin. "These groundwater trends are induced far away from farms as a result of lateral groundwater flow, " Condon told environmentalresearchweb. And it wasn’t just the aquifer’s behaviour that was altered. Irrigation also affects the latent heat flux at the land surface, a proxy for the evapotranspiration from plants. Without irrigation the latent heat flux goes up and down fairly randomly, at periods of between one week and nine months, but when irrigation is applied the temporal persistence is increased. "This means that irrigation is increasing the memory in the surface system in irrigated cells," said Condon.

Until now we have rather taken for granted the role that groundwater plays as a buffer for natural variability. However, this study indicates that, in semi-arid regions at least, careful water management may be required in order to maintain that buffering capacity.

Related links

• Laura E Condon and Reed M Maxwell 2014 Environ. Res. Lett. 9 034009
• ERL
• Laura Condon
• Reed Maxwell

Related stories

• Transboundary aquifers feel the strain
• Groundwater study reveals climate-model variability
• African groundwater offers hope for climate variability 'buffer'