Adams, K. D., R. M. Negrini, E. R. Cook, and S. Rajagopal (2015), Annually resolved late Holocene paleohydrology of the southern Sierra Nevada and Tulare Lake, California, Water Resour. Res., 51, 9708–9724, doi:10.1002/2015WR017850.
2.7 billion dollars of lost farming revenue. 18,000 jobs. Communities forced to cut water usage by 25%. This is the devastation caused by California’s drought, by far the worst in recent memory. But was this drought California’s worst ever? And could it have happened without climate change pushing weather to new extremes? This is difficult to answer without detailed historical records of water availability in the region. A new study creates a historical reconstruction of California drought conditions over the last 2000 years to put the current drought in perspective.
How we compare droughts
There are a few ways to reconstruct past drought conditions First we can look at historical data. The California Department of Water Resources has been measuring water flow in major rivers since 1894. It is clear from this dataset that the current drought is the worst this century—river flow in 2015 was less than ¼ of the historical average! But how can we look back further than the 120 years of the historical record?
One method is estimating past changes in lake levels. This is useful because it reflects modern water management: water levels in lakes (or dammed reservoirs) rise during wet years, then water is slowly distributed downstream over several dry years. Lake levels can be reconstructed directly from sediments accumulated on the lake bed using a variety of indicators, but this approach isn’t very precise because each layer of sediment represents several years of accumulation. This approach is great for spotting long-term trends but misses shorter events like the recent California drought.
Tree ring records provide a high resolution alternative for drought reconstructions. In dry regions like central California, tree growth is limited by water availability. In years with more rain and snow, trees grow quickly, and during drought years they might not grow at all. The width of tree rings increases with annual growth, so the tree ring record provides a record of water availability resolved at an annual time scale. Although the tree ring pattern could be affected by any number of factors in an individual tree (insect invasion, competition from a neighbor, etc.), these complications are smoothed out by averaging over a large number of trees. This is accomplished using the North American Drought Atlas, which compiles thousands of tree ring records collected by many scientists over decades of research. This study uses the Drought Atlas to simulate changes in lake levels over the last 2000 years, providing a long-term context for the severity of the 2012-2015 drought.
Constructing the model
The authors modeled the historical water levels of Tulare Lake in Central California (Figure 1). The lake itself no longer exists because the rivers feeding it have all been diverted for irrigation, but at the time of European settlement it was the largest lake east of the Mississippi River in terms of surface area. Tulare Lake is particularly interesting to model because it is fed by 4 rivers (the King, Kern, Tule, and Kaweah Rivers) that drain snowmelt from the Sierra Nevada mountains, which is a critical source of water for sustaining agriculture in the fertile but dry Central Valley.
The first step in developing the model was to establish a quantitative relationship between the tree ring dataset and the 100 years of measured water discharge for the 4 rivers (Figure 2). This allows the tree ring data (which dates back much further) to be used to estimate the water flow in each river for the time before the river flow data was available, creating a 2000 year estimate of discharge from the 4 rivers (Figure 3). Finally, the discharge from the rivers was combined to model the water level of Tulare Lake. This is offset somewhat from the river flows because the lake fills during wet years and takes several years to empty.
Results and significance
Precipitation and lake levels in California are quite variable, and very wet or very dry years occur regularly—look at all the peaks and valleys in figures 3 and 4! However, the 2000 year reconstruction shows that the recent drought was historically epic. 2015 was the driest year in the reconstruction for 3 of the 4 rivers (there were 25 drier years for Kings River). There were also droughts longer in duration: during the Medieval Warm Period (about 850 to 1300) there were 2 decades-long “megadroughts” covering the American West, during which the reconstructed water flow was only 65-70% of the historical average. During the 2012-2015 drought, water flow was only 37% of normal.
The reconstructed lake levels tell a similar story: although there have longer periods of sustained drought and single years where the water level was lower, the combination of severity and length of the recent drought was unprecedented. 2012-2015 was the driest 4-year period in central California in the last 2000 years, according to this reconstruction.
Putting the California drought in a historical context is important for efforts to detect the effects of climate change. Extreme climate events like droughts have occurred regularly (again, see all the spikes in the figures) without any help from fossil fuel emissions, making it hard to tell for sure if a disaster like a heatwave or a drought could have occurred naturally. Reconstructions like this help scientists assess the probability that something of this magnitude could happen without human impacts warming the planet. While we can’t be certain a drought this devastating wouldn’t have happened without human influences, we can say it hasn’t happened in at least 2000 years, so it isn’t very likely!