Learning from a River’s History to Prepare for the Future

by |August 17, 2015
George_Caleb_Bingham_-_Boatmen_on_the_Missouri_-_Google_Art_Project

“George Caleb Bingham – Boatmen on the Missouri c. 1846 – Google Art Project” by George Caleb Bingham. Licensed under Public Domain via Wikimedia Commons

Ever since Lewis and Clark attempted to reach its headwaters more than 200 years ago, the Missouri River has played a major role in the history, economy and expansion of the American West, serving as a source of hydropower, a major inland waterway and a source of irrigation for some of the nation’s most productive agriculture. The longest river in North America, the Missouri drains a basin that covers nearly one-sixth of the continental United States before it empties into the Mississippi near St. Louis.

But for all its economic and ecological value, the climate conditions that affect how much water is flowing through the Missouri River system are not fully understood. High levels of variability in rainfall from year-to-year means that at any given moment water mangers, farmers and businesses that rely on the river must plan for the possibility of destructive floods, drought or anything in between. Longer-term natural climate variability complicates the situation, while climate change adds an entirely new level of uncertainty.

5804030823_00cc36e70e_o

Missouri River Flooding

Now a consortium of researchers from eight universities, including Columbia University, is using tree ring and glacier analysis to reconstruct the climate history of the Missouri River Basin in order to give policymakers and water managers better decision-making tools to manage the river. The team has received funding from the National Science Foundation to support its research, and held its first meeting in Bozeman, Mont., in June.

Naresh Devineni, an assistant professor of engineering at The City College of New York and an affiliated collaborator at the Columbia Water Center, is bringing his experience of reconstructing stream flow records on other rivers to the consortium.

Part of the problem, explains Devineni, is that there are very few observational records on river flows beyond 50–70 years. The operating rules of the reservoir systems are usually based on short historical records. These records are unable to adequately inform the risk of the persistent wet or dry periods, or transitions between them.

The mountain regions that collect moisture for the headwater streams of the Missouri river system get most of their moisture from the Pacific Ocean, which means the amount of moisture they receive is strongly influenced by semi-periodic cycles such as the El Niño/La Niña Southern Oscillation (ENSO) and the Pacific Decadal Oscillation. ENSO cycles last from 3–5 years, while the Pacific Decadal Oscillation takes place on a decade-long timeframe.

748px-Missouririvermap

Source: Wikimedia commons.

An example of the dramatic climate variability in the basin at different time scales could be seen in 2010, when the snowpack that feeds the headwaters of the river was 130 percent higher than normal, causing devastating flooding later in the year. Ironically, this happened during a decade when snowpack was one of the lowest it had been in 800 years.

“Climate has a long cycle,” Devineni says. “With a 50 or 60 year record, you only get what happens in that period. But if we can relate flows with tree ring data, we can infer the stream flow for the past 200 or 300 years.”

Trees typically add one ring of growth to their trunks every year. But the width of each ring is directly influenced by the amount of moisture that was available in that year. By gathering tree ring data from old trees, scientists can determine how wet or dry a particular year in the past was. This, in turn, can help them understand what are the decadal, multi-decadal or even century-long variations in climate that affect stream flow.

According to Devineni, understanding this history is critical in order to plan for the future. Consider navigation of waterways, he says: “In a dry decade, your operating policies could be modified to store more water in dams, or vice versa, you could release more water in wet periods.”

Since the 1970s, similar historical climate reconstructions of the Colorado River system showed that 20th century water allocations from the river were based on one of wettest periods of the past five centuries and that severe droughts had occurred in the past. A few years ago Devineni worked on a similar reconstruction of the Delaware River’s flow history in New York State, as part of a Consortium on Climate Risk in the Urban Northeast project designed to assess drought risk for New York City. That study found that while the extreme drought of 1961 to 1967 was unlikely to occur with much frequency, the historical record suggested the region was at risk of experiencing recurring droughts of shorter duration–a finding that could have major implications for how the city manages its reservoirs and future water supply.

“We finished the Delaware River study in 2012,” Devineni says. “The question then was, ‘Can we do this for a very large basin like the Missouri River where the reservoirs are in a network?’”

Until recently, the team was still in the data collection phase of its Missouri study, but is now ready to begin its analysis using Hierarchical Bayesian statistical models. They hope to complete this analysis by the summer of 2016.

What the experience of reconstructing the climate history of the Colorado River showed, says Devineni, is that failing to understand the full range of climate variability in a region overtime can lead to costly mistakes. After all, large-scale water infrastructure is expensive and lasts a long time, meaning there is a real risk of over- or under-designing, based on an incomplete understanding of how much water will actually be available.

But even once the infrastructure is built, having a better understanding of climate probabilities can affect operating procedures.


Leave a Reply

Your email address will not be published. Required fields are marked *