No, you don’t need to check the prescription on your eyeglasses (or medications)—we really are talking groundwater depletion and sea level rise again. Just a few weeks ago, we covered a recent study on the topic published by researchers from Taiwan and the Netherlands and compared it to one from last year that was done by Leonard Konikow of the United States Geological Survey. There’s a good reason that we’re back at it again. But first, for those who didn’t take notes—what are we talking about?
In many places, the water table is dropping as groundwater is depleted. When groundwater is pumped up for use, whether for drinking water or irrigation, some portion of it fails to infiltrate back down into the ground. (In drier regions, the portion that infiltrates approaches nil). Instead, the water evaporates into the atmosphere or ends up in surface streams. In either case, most of it eventually makes its way to the ocean. In many places, the amount of precipitation that infiltrates into the ground is too small to make up for that loss. And as the volume of groundwater decreases, sea level must rise in turn. It’s an awfully big planet we’re on, though. Most of its surface is ocean, so you might not expect this to add up to much.
That’s where these studies become so interesting. They estimated that, currently, the volume of groundwater being depleted is equivalent to about 13 to 20 percent of the ocean volume change. This isn't the whole story, however. The construction of dams on rivers creates large reservoirs (or lakes) behind them, increasing the storage of water on land. As long as you keep building new dams, you continually counteract some portion of sea level rise.
The sea level breakdown in the 2007 IPCC report assumed that dams and groundwater depletion roughly cancelled each other out. Another study (that Konikow contributed to) found that the net result offset about 6 percent of sea level rise between 1972 and 2008. But the study out of Taiwan and the Netherlands used a larger estimate of groundwater depletion. It shows the net contribution for that same period could have been positive—adding about 6 percent to the total rate of sea level rise.
A study published by a group of Japanese researchers in Nature Geoscience a few days after our last story, however, came to an astonishingly different conclusion than either of those. Over the period of 1961 to 2003, they estimated the average contribution to be a whopping 42 percent of the total. (For a more direct comparison, that latest study we covered estimated the contribution to be slightly negative over that timespan).
The Japanese estimate was produced by a global model not completely unlike the one used by the group from Taiwan and the Netherlands. Available climatic and water management data is fed into these models, and they simulate a water balance. So how could they generate such different answers? It’s all about the types of data that are used and the assumptions that are relied upon.
The primary difference between the previous estimates and this most recent one is in the amount of groundwater depletion. Konikow authored last year’s study and told Ars the difference is due to “seriously flawed conceptual models.” These yielded answers that “are not based on or consistent with observed changes in [ground]water levels.” The Japanese model calculated a loss of around four times as much groundwater as Konikow’s study estimated.
Because good, global data on water use is hard to come by, they calculated total water demand based on a number of factors. If nearby surface water could not fulfill that demand, they assumed the remainder was withdrawn from groundwater and that nearly all of the groundwater that was pumped ended up in the ocean. That ignores surface water brought in from outside the immediate area, whether through canals or desalination of seawater. In addition, there’s no limit to the amount of groundwater that can be withdrawn in the model—even if there isn’t that much available in the real world.
It also runs afoul of a problem long known to haunt estimates based purely on simple budgets. As the actual water table drops, less groundwater will flow into streams and lakes, partially offsetting the impact of withdrawals. (A bit like how reducing your spending can soften the budgetary blow of a pay cut). This is why Konikow’s study relied so heavily on difficult-to-obtain measurements of groundwater level data and detailed groundwater basin models. It’s also a big reason why his estimates of groundwater depletion are the smallest.
Yoshihide Wada, a researcher from the Netherlands group, explained that their model used actual data on groundwater use rather than calculating it from total water demand. But they still ran into some of the same problems. When they compared estimates of groundwater depletion to more detailed regional numbers, they found that their model overestimated depletion in non-arid regions. To bring their numbers into line, they applied a correction factor. “Since our corrected groundwater depletion estimate [agrees well] with reported depletion estimate per region,” Wada wrote, “I would expect that their modeled groundwater depletion is likely overestimated.”
Other researchers in the field can spot dodgy methods in a paper like the mascot in a cereal box knock-off of Where’s Waldo?
The Japanese researchers point to the 2007 IPCC report in support of their higher number. In that report, estimates of the various contributions to sea level rise only amounted to about 60 percent of the observed trend. Since this estimate of the groundwater/reservoir contribution comes in at around 40 percent, it must have been the missing piece, right? Well, the error bars on those IPCC estimates were massive. More recent research has seemingly closed that gap without the need for stunning amounts of groundwater depletion.
So what happens when a sticks-out-like-a-sore-thumb study like this is published (in a high-visibility journal, no less)? Are scientists thrown into confusion? Is the whole subject automatically downgraded to “unsettled?” Not so much. Other researchers in the field can spot dodgy methods in a paper like the mascot in a cereal box knock-off of Where’s Waldo? Scientists know that every study is imperfect or incomplete in some way and are especially skeptical of results that contradict—rather than build upon—the existing science.
When lots of data has been published supporting one conclusion, and then a single data set points in a different direction, the most likely explanation is that something is wrong with that rogue data set. Of course, this isn’t always true, but as Carl Sagan was fond of saying, “Extraordinary claims require extraordinary evidence.” Groundwater depletion accounting for nearly half of sea level rise isn’t quite as far out there as faster-than-light neutrinos, but it would be fair to call it extraordinary.
Konikow told Ars he plans to submit a comment to Nature Geoscience laying out his objections to the study’s methods in detail. If it’s accepted (and it would be surprising if it wasn’t), the Japanese researchers will be given the chance to write a reply, and the two will be published together. The next IPCC report will probably cite all three of these studies, but will carefully take into account which were done most carefully and are best supported by the evidence. Unlike in the food industry, a scientific sausage is always judged by what went into it.
Nature Geoscience, 2012. DOI: 10.1038/ngeo1476 (About DOIs).
reader comments
144