Fighting Drought Will Be a Long-Term Battle, Says Study

Using new snowpack data collected by satellites, we now have a better view of California’s water deficit, and it is not a pretty picture. Steven Margulis of UCLA explains just how deep the problem is.

Written by Matt Weiser Published on Read time Approx. 8 minutes
Nearly barren ridges of the Sierra Nevada are seen near Sequoia National Park during an aerial survey performed by the California Department of Water Resources, April 28, 2015, a time when this area should have normally been buried in deep snow.Rich Pedroncelli, AP

The California drought is now in its fifth year. But what if we told you it could take four more years to get out of it?

That’s the alarming result of a study published June 21 in Geophysical Research Letters. The study analyzed California’s mountain snowpack to assess the severity of the current drought and compare it to past water shortages.

The study found that the current drought is, without question, the worst ever recorded in the state as measured by the “deficit” in the snowpack and the crucial freshwater it provides to the state. And largely because of its long duration, it will also likely take several years of winter storms to make up that deficit 4.4 years, to be exact.

That estimate was developed, first, by analyzing historical on-the-ground snowpack measurements together with a new resource: detailed satellite imagery of the mountain snowpack, gathered in recent decades by the federal government’s Landsat program. This new data provides a more comprehensive picture of the snowpack because it looks at all of it, not just location-specific data gathered by sensors on the ground.

The researchers, led by Steven Margulis, a professor of civil and environmental engineering at the University of California, Los Angeles (UCLA), then ran thousands of computer models using the data to estimate how much longer it will take to erase a drought of this magnitude.

Water Deeply recently spoke with Margulis to gain a better understanding of his findings and what they mean for California water management.

Water Deeply: Refresh us on this drought. How unique is it, based on your study?

Steven Margulis: One of the main novelties of our work was trying to embed this satellite-based data into a framework that allows us to tease out how much water is stored in the snowpack. By quantifying the amount of water stored, one of our specific results showed that 2015 was by far the lowest amount of snow water stored in the Sierra Nevada over the record we examined. The main period of our new dataset is 31 years, which corresponds to the remote sensing records.

And further extending it back to 65 years using in-situ [snow collecting] data, we found that 2015 was the driest year on record in terms of the Sierra, by far.

In our 31-year data set, the average amount of water stored in the Sierra was about 18.6 cubic kilometers (4.5 cubic miles), averaging over all years. And 2015 was only 2.9 cubic kilometers (0.7 cubic miles), so about 15 to 16 percent of the normal.

To put it another way, we used our dataset to calculate the return period, which tells you how often a part-event is expected to occur. What we found is that 2015 has a return period of over 600 years, meaning, on average, you wouldn’t expect it to occur but once every 600-plus years.

Last year – 2015 – just jumps out as a very extreme water year. Secondary to that, it happened on the tail-end of a multi-year drought. It was an extreme year which was compounding several other extreme years happening prior to it.

Water Deeply: You found that, historically, most California droughts ended after just one year of near-normal rainfall. But not this time. Why is that?

Margulis: We were able to develop a new metric, called a drought deficit. This dataset allows you to add up how much in deficit you are throughout the course of the drought. We were able to do that over the full 65-year record we have.

In California, with respect to snowpack, it’s typically a year-to-year thing. People are just waiting for the next above-average year to get things back to normal. What we found in computing this deficit is, in most cases, that’s fully justified. If you go back over the 65 years, in all other years but one, a [snowpack] deficit would go back to zero, meaning it would end, within one year. What stood out was that the deficit at the end of 2015 was almost twice as large as any other deficit on record. It was like digging this big hole.

The other subtle but important thing is that any time we have a drought year in snowpack, that’s water that is going to have to come from somewhere else, whether groundwater pumping, reservoir storage or conservation. What we’re saying is not that we’re not going to have above-average years going forward. There’s this hole that’s been dug and it’s going to take many years for the snowpack water to get back to normal. That effect that’s been propagated downstream is likely to be a multi-year recovery.

Specifically, what we found is the expected duration of recovery from this drought would be a little over four years.

Water Deeply: Four years, that’s a long recovery. How can this information be helpful to people who have already been through a drought for five years now?

Margulis: The method we developed was really motivated by trying to get better knowledge on how it varies year-to-year. In the depths of the drought, that’s useful to water managers. Most of the models they use are based on the historical data we have. If we can improve the historical knowledge, that’s helpful.

With respect to the drought specifically, it’s a cautionary note in that we as Californians have a short memory. What this research shows is that in these circumstances, the planning we need to do might be a little bit longer-term. One good year may not get everything back to normal.

Water Deeply: Does this mean we need four years of normal precipitation?

Margulis: We did 10,000 random simulations [from the dataset], each of which sampled likely snowpack years from the historical record, but going forward. Some of those 10,000 realizations say if we have a huge snowpack next year, we could end the deficit. Others say we might not end the deficit for 10 years. If you take the average of all those realizations, that’s where the average of four years comes from.

It’s important to keep in mind, the drought could end sooner than that if we have a sequence of very wet years. Or it could take longer if we have more dry years.

This four years is not saying next year won’t be above-average. That’s certainly a possibility. It’s saying that an above-average year might not be enough to offset the several below-average years that we’ve had.

Water Deeply: It seems like lots of people get confused about this idea of a water supply deficit. Can you elaborate on that concept?

Margulis: In places like the Central Valley, where there’s lots of [groundwater] pumping going on, much of that pumping is because they’re not getting water from the snowpack. In thinking about drought, our study was really only looking at snowpack. But the system is really interconnected. If there’s a water deficit in one part of the system, that needs to be made up from another part of the system. So once you take that water out, now you have a deficit in groundwater.

We’re saying given what the snowpack may be, there’s this big hole the system is digging out of. That deficit is not just in the snowpack. Because it’s such a long-duration event that includes the biggest drought year on record, it’s going to take time for the system to recover because of its connectedness.

Water Deeply: Climate change predictions consistently warn that we’ll see less snow and more rain in the mountains. Did this happen in 2015?

Margulis: The characteristics that we saw are similar to that. The question is, what’s the mechanism? 2015 was an odd year in at least a couple ways. In terms of the amount of snowfall, it was the lowest on record. It was just a very, very dry year. But it also happened to be the warmest year on record.

Where temperature can play a role, there are two main ways. One is that precipitation will fall as rain instead of snow. If you just raise the temperature, where that freezing line is will be at a higher elevation, and lower elevations are going to experience rain instead of snow. The other way is that, in between storms there’s going to be more melt of the snow.

This study was not about climate change per se. It’s implicit, of course, in everything. But 2015, you could argue, maybe, that it’s a sign of things to come in the sense that it had most of the snowpack at the higher elevations. Temperature played a role, but it was just such a dry year.

There are implications with respect to the current [snowpack] monitoring system. All of these in-situ sensors, the snow courses and the snow pillows, they tend to be at middle elevations, largely for practical reasons. To get to those locations to maintain sensors is a difficult proposition in winter. So you can’t usually site them at the highest elevations of the range because they are just too hard to get to. So the whole system which has been set up to predict water resources in California is based primarily on these sensors at middle elevations.

So as the snow starts to recede upward, what’s sampled at these sites is less and less representative of what’s there. The sensors that we do have become problematic in that they’re not necessarily sampling where the snow is mainly located. That’s one of the benefits of our method is that [using satellite data] it provides estimates across the whole range, including at elevations where we don’t have sensors.

Water Deeply: Does that argue for deploying more sensors, or simply mastering the tools you’re working with?

Margulis: I think it’s both. Those kinds of in-situ estimates that do exist are very, very valuable for giving us on-the-ground information. There’s also definitely a push in the scientific community to develop these kinds of methods [using satellite and airborne tools] where you’re able to get pictures of what’s going on over the full [mountain] range. Because no matter where you site these things, they are very much point-scale estimates of something that can vary considerably over time. You’re getting an estimate at a given location when what you really want is the total amount of snow over the whole range.

Water Deeply: Your study correctly predicted the drought was very unlikely to end in 2016, estimating the likelihood at only 7 percent. What are the odds in 2017?

Margulis: I wouldn’t say we correctly predicted it. All indications are it hasn’t solved the deficit, but we haven’t confirmed that per se. Whether that’s proven, I’d be hesitant to say that matter-of-fact until we construct the snow water volumes for 2016.

We estimated what the likelihood will be out to five years. The 7 percent was the likelihood we would have expected in one year. But in our analysis, we have a probability for each year going forward. So in 2017, the probability of the deficit going to zero is about 25 percent – still rather low. Three years out, it’s about 45 percent. And then four years out, it’s a little over 60 percent. What we found is that once it’s more than 50 percent, you can think of that as, OK, it’s equally likely that it will end.

In many of the other droughts that have occurred over the years, when probabilities were greater than 50 percent that they would end in one year, they did. The only thing that’s different this year versus others is the starting point of the deficit. That’s the reason these numbers are low. It’s not projecting anything different going forward. The deficit is so large relative to anything seen over the last 65 years that it’s going to take longer to get out of that hole, in all probability.

Suggest your story or issue.

Send

Share Your Story.

Have a story idea? Interested in adding your voice to our growing community?

Learn more
× Dismiss
We have updated our Privacy Policy with a few important changes specific to General Data Protection Regulations (GDPR) and our use of cookies. If you continue to use this site, you consent to our use of cookies. Read our full Privacy Policy here.