by Jay Lund
Every year is different for water management in California.
The 2012-2016 water years were among the driest and warmest on record. 2017 was the wettest year of record for much of California, with thousands of water managers struggling to store as much water as possible in reservoirs and aquifers.
So far for this 2018 water year (which began October 1), Northern California precipitation is about 67% of average for this time of year. Further south, the San Joaquin Basin precipitation is about 38% of average and Tulare basin is about 25% of average. Snowpack statewide is about 27% of average for this time of year.
December has had essentially no precipitation and a stationary ridge in the Pacific Ocean off California seems likely to block most moisture from the Pacific Ocean into January.
Fortunately, today California has 109% of average water storage in reservoirs for this time of year, 1.5 maf more than average reservoir storage. The wet 2017 water year substantially refilled Northern California’s less depleted aquifers. But only a small part of the additional drought groundwater withdrawals has been recharged for the more depleted aquifers of the southern Central Valley and southern California.
Here is some simple statistical analysis on three reasonable questions.
1) Will the 2018 water year be dry? Likely, but maybe not.
Let’s look at some historical statistics of monthly precipitation in Northern California.
For the last 97 years, when December precipitation was in the lowest 20% of all years, 79% of the overall water years were drier than the median precipitation. As the plot below summarizes, the dry December likely means a drier year, mostly because the annual total is deprived of December precipitation, but also because there is also some correlation that drags down precipitation in other months. From the regression equation slope, one inch loss of December precipitation averages 1.28 inches in lost annual Northern California precipitation. One inch less December precipitation tends to be accompanied by an additional 0.28 inches of less precipitation in other months. So there is a good chance that 2018 will be drier than the median and drier than average.
However, of the 19 historical years with the driest 20% of Decembers, four had total precipitation above the 97-year median. So there is roughly a 21% chance that 2017 will be an above-median water year.
2) What about floods? Does a dry December reduce the probability of flooding this year? Yes, but a flood could still occur.
A review of Northern California precipitation statistics shows that in 97 years, years when December has been in the lowest 20% of precipitation have never had flood levels of precipitation (more than 20 inches, corresponding to the 16 wettest months since 1921). Part of this is that a dry December means that one of the most flood-prone months did not have a flood. But a flood could still happen January through March. Viewed this way, the odds are low that 2018 will be a flood year – But water managers probably should not bet anyone’s flood safety on this statistic.
3) If 2018 is a dry year, is it the beginning of a new drought for California? Perhaps not.
With its long dry summers, every year California has a worse drought than most of the US has ever seen. This troubled early settlers, but today California’s city water supply and irrigation systems have enough reservoir and aquifer storage, supply interconnections, and institutional responsiveness so that most economic water uses are largely immune from one-year droughts. (Ecosystem impacts are a sadder story, however.)
California almost has to have at least two dry years together for there to be a noteworthy drought. Even if this year is dry, there is only about a 50% chance that the next year is drier than the median. Total water year precipitation has very little correlation between years, as seen in the two figures below.
Looking at the statistics for Sacramento Valley runoff (since 1906), by definition half of these 111 water years (55 years) are drier than the median. Thirty years are in the second or more year of successive dry years. 18 years are in the third or more consecutive dry years, 10 years are in the fourth or more consecutive dry year, 4 are in the fifth of more consecutive drier than median year, and 2 years are in the sixth consecutive drier year. No historical droughts exceed six consecutive drier years.
The frequency of longer drought years declines almost as if there were no correlation for dry years. This is shown more formally below, which plots annual Sacramento Valley runoff against the runoff in the previous year. There is a roughly 10% correlation in annual runoff, which explains about 1% of variance in annual runoff (some, but not much).
Bottom line – California’s 2018 water year will likely be drier than the median year, and still more likely to be drier than the average year. It could still be wet, but is less likely to have a major flood. However, floods are dangerous enough and still likely enough that it would be unwise to not prepare and operate for floods as well.
Although we would like to predict California’s hydrology for the coming season and years, we unlikely to have great skill in this – perhaps ever.
CDEC, The mighty California Data Exchange Center, http://cdec.water.ca.gov/index.html
Lund, Jay (2015), “The banality of California’s ‘1,200-year’ drought,” CaliforniaWatrerBlog.com, September 23.
Null, Jan (2017), “Dismal Beginning to SF Rainfall Season”, Golden Gate Weather Services Weather and Climate Blog, posted December 21, 2017
Swain, Daniel (2017), “Strikingly dry conditions persist; Thomas Fire now largest California wildfire,” December 24, 2017, http://weatherwest.com/archives/6030
Swain, Daniel (2017), “New insights into the Ridiculously Resilient Ridge & North American Winter Dipole,” December 4, 2017, http://weatherwest.com/archives/5982
Fun statistical fact (for geeks who read to the bottom):
Most years are drier than average. In the historical record of Sacramento Valley unimpaired flow, 57% of years are drier than average. Why? Averages are increased by very wet years, which do not affect the median flow (50% above and 50% below). Because precipitation and flow usually can’t be less than zero, extreme dry years cannot pull the average down as much as extreme wet years pull the average up. This skews their probability distributions away from zero, with more than 50% of years being drier than average. (But you already knew that water in California is not normal.)
Jay Lund is a professor of Civil and Environmental Engineering and Director of the UC Davis Center for Watershed Sciences.