Using Drought Forecasts to Improve Natural Resource Management

Video Transcript
Download Video
Right-click and save to download

Detailed Description

Natural resource managers face increasing challenges in dealing with drought. As competition for water increases between its various uses (water supply, energy demands, ecological services, recreation, and other environmental and ecological needs), our ability to forecast the onset and termination of drought becomes ever more important. This is particularly true given forecasts of the increasing frequency and intensity of drought suggested by climate change models. 

This talk presents a brief review of history of drought forecasting and its application in resource management. Drought indicators currently in practice are explored as well as emerging indicators. Additionally, the fundamentals of drought planning are reviewed, particularly the challenges of developing robust policies that work well when future conditions are difficult to predict. The challenges in applying these forecasts associated with the forecast accuracy are discussed, as will other limitations. Finally, the potential for improving our ability to predict regional drought in the near future are explored. 

This webinar was recorded on April 27, 2017 as part of the Climate Change Science and Management Webinar Series, held in partnership by the USGS National Climate Change and Wildlife Science Center and the USFWS National Conservation Training Center. 

Details

Date Taken:

Length: 00:52:24

Location Taken: MA, US

Video Credits

Speaker: Richard Palmer, University of Massachusetts Amherst, Northeast Climate Science Center
Hosts and Coordinators: John Ossanna, FWS National Climate Change and Wildlife Science Center; Elda Varela Minder, Shawn Carter, and Holly Padgett, USGS National Climate Change and Wildlife Science Center

Transcript

Shawn Carter:  It's my pleasure to introduce Dr. Richard Palmer. Rick is Professor and Department Head at the Civil and Environmental Engineering Department at UMass Amherst. He also serves as our university director at the Northeast Climate Science Center, which is based there. It's one of eight national centers funded by the USGS with Department of Interior.

His research focuses on the fields of climate impacts to water resources and urban infrastructure, and in the application of structured planning approaches to water and natural resource management.

Specifically, he looks at the impacts of climate change, drought planning, real‑time water resource management, and the application of decision support to civil engineering management problems.

Today, he's going to be talking about some of his work as it relates to drought. Without further ado, I'll turn it over to you, Rick. Welcome.

Richard Palmer:  Thanks Shawn, thank you very much. First Shawn, thank you, and others for inviting me to give this talk. It's a great honor. I want to thank everyone at the Northeast Climate Science Center here on the UMASS campus for all the support that they have provided us.

I'll just mention that our Climate Science Center also includes the University of Wisconsin, the University of Minnesota, the University of Missouri, the College of Menominee Nation, Woods Hole, and Columbia University as consortium members in our group.

I also want to mention my thanks to Mary Ratnaswamy, the federal director of our center. Even moving forward works. I wanted to have a couple of opening comments about my talk, because I think it's a little different than the talks that you've seen before in this series.

Previous talks ‑‑ one of the first talks ‑‑ Shawn and others introduced the concept of ecological drought, and two talks since then have really focused on soil moisture monitoring and exchange over time.

I just want to note that last week at our Northeast Climate Science Center, Doctor Keith Nislow, one of our PI's, gave a great presentation on the ecological impacts of drought in the Northeast. I really feel like a lot of important issues have already been covered.

What I want to do in this talk is take a slightly broader approach, and sort of talk about how difficult it is to bring drought forecasts into decision making.

Take it from my perspective, which is just a little different than some. For most of my career, I have been working in and out of looking at the impacts of droughts on large water supply systems, and on the river systems, and the environmental services that they provide.

Today, I'll be talking a little less about specific ecological droughts, but about droughts in general and how we try to cope with them.

In a perfect world, we would have forecasts of very high reliability, at spatial and temporal resolution that's perfect for decision making. They would go as far out into the future as we would care for them to go. Of course, that is not the case, in fact.

What I want to talk about today a little bit is a very structured planning approach to drought management, which I think is essential whether we're talking about ecological droughts, or droughts of any other sort.

This has been a really interesting field, because the use of climate states as indicators of future climate ‑‑ or drought in the short term ‑‑ has really been an exciting field for some time now.

I think if you look at the field, you'll see it's really rich with examples of how we've been able to get marginally improved forecasts. Of course, oftentimes we don't want slightly improved forecasts. We would love to have very largely improved forecasts.

I think if you look at the literature again, real demonstrations of how forecasts have changed the way we've managed, or have improved our management, are not very frequent. I'm going to talk about a little bit of the history, some examples of this, and I will also move to the idea of a drought monitor, and the drought forecast tools that are available to us.

As suggesting they do provide a good continental scale framework for thinking about drought managements. I'm going to end with a suggestion that I hope people will think about a little bit, and that is the value of practicing for droughts.

The notion here of a virtual drought is something I'm not familiar with in terms of natural resource management. I'm looking forward to the end of the talk, when we can make some discussion about the applicability of this.

My basic theme, I would say, is to ask the question, "Is there a strong reason for why ecological drought planning should be different than other forms of drought planning?" If you follow the general concepts of drought planning, you know there are variations.

They vary from utility‑level planning, whether it's a power utility or whether it's a water supply utility, to city planning ‑‑ almost all major cities in the US have drought plans ‑‑ to regional planning, such as in the area around Atlanta, when multiple water supplies have to coordinate their drought planning.

To state plans, which most states do now have drought plans. To what we might think of as more continental or landscape‑scale planning to deal with droughts. Regardless of whether you're looking at a utility level, at the smallest level, or a landscape‑scale planning, there are some general rules that disciplines follow in terms of drought planning.

What I have here on this slide is two examples that came from different areas. I think that you'll see they're quite similar. On the right‑hand side, I gave a list of 10 steps that the National Drought Mitigation Center, which is out of the University of Nebraska‑Lincoln, promotes in terms of what they consider good drought planning.

On the left‑hand side, I have listed a seven‑step process that came out of the US Army Corps of Engineers National Drought Study. If you look at the 7‑step process on the left, or the 10‑step process on the right, you'll see a number of things that are very consistent.

Basically, these consistencies are focused along the process of building teams to look at the problems that are associated with drought, really understanding our problem. Also, to developing clear objectives and ways to measure those objectives as being a very important step.

Then a description of what the status quo is. That is, what will happen if future droughts were to occur without any changes in management. Then, the formulation of alternatives. This process of formulating alternatives, and the next step of evaluating alternatives, typically are developed with the use of computer models.

Finally, once you've evaluated your alternatives and developed those that you feel are the most productive, the next step is basically to institutionalize that plan, to get it accepted by a region. Then the final step on both of these planning processes, is to either update, or evaluate, and revise drought plans as conditions change over time.

As I said, these two approaches came from different perspectives. One on the right being primarily associated with agricultural droughts and the one on the left, being associated more with urban drought situations.

Both reflect exactly the same multiple‑step process of engaging stakeholders, indicating what your true objectives are, and determining ways to measure them, looking at alternatives, and then finally selecting a plan and getting it implemented.

On both of these approaches, one of the things that I would say that you have to do to respond to droughts, is you need a series of things that we might think of as indicators. That is, how do you know that you're in a drought?

On that slide there, I've indicated stream flows might be an example of what an indicator of a drought would be. It could be a Palmer Drought Index. It could be a wide variety of things, but basically, you need a metric that indicates whether or not you're in a drought.

Then you need what we call "triggers," which is a value of that metric or a combination of metrics that indicates that you are either at an acceptable level or unacceptable level. My example there is that perhaps, if stream flows are less than 50 percent of average flows, it might be a trigger.

Then you need actions, things that you might do. Things like curtailing water allocations would be an example of an action.

The way I like to think about this, is that all drought planning focuses on how we'd make use of a forecast, if we can make a forecast, then how it would be used to take action. We'll come back to this idea of indicators, triggers, and actions in a few minutes in this talk.

This is an overview of the remainder of my talk. I'm going to present a series of papers from the past. Some of these are pretty old, and so it was pretty fun going back and looking at these.

I'm going to give an example, a cautionary tale, about how drought forecasts were used, and perhaps misused. I'm going to take a look at how past efforts to use climate information were explored to improve management.

Then there is a paper which I hope some of you might be familiar with. It's called "Forecasts Are for Wimps." We'll look at this in detail in a few minutes. It's a paper that focuses on why forecasts perhaps are not used. I'm going to use that paper and a second paper to illustrate some of the challenges.

Then, I will go on to look at the National Drought Monitor and Forecasts, and then make this suggestion about using virtual droughts as a help in terms of natural resource management.

The first paper we're going to look at is a paper that was published many years ago by Marty Glantz. Marty published this in a common journal in my field, "Water Resources Research." It was published back in the early 1980s, but focused on a situation that occurred in 1977. It's nice to be able to talk about the 40‑year anniversary of this particular event.

In addition, Marty went on to publish quite a few other things. One of them was this book that I am showing here that Cambridge Press put out on looking at the impacts of El Niño and La Niña in management.

This paper has a very interesting title, I have always thought, "The Consequences and Responsibilities in Drought Forecasting." The thought here, of course, is that when you make a forecast, it does not stand by itself, but in fact, it has consequences. Those making those forecasts have certain responsibilities.

Just to give you a little background in this situation ‑‑ and I think this is fun to think about ‑‑ back in the early 1970s, again, more than 40 years ago, there was a lot of concern about climate change.

Part of this concern revolved around the behavior of Russia in their purchase of grain. What was feared then was that changes in climate were going to have a large impact on the USSR, former Soviet Union, in terms of being able to meet their grain needs.

About the same time, a little bit later, in the US and Canada, there was a major drought in the western portions of those countries. This is followed by one of the major droughts in California history, back in '76 and '77.

The story that's going to be told focuses on the Pacific Northwest when, in 1977, there was a really poor water supply outlook for this region. We're used to great graphics in terms of maps. If you think back to 1977, we didn't have a lot of those resources. If you look on the left, that was an official stream flow forecast put out by the Bureau of Reclamation.

What it looks like, what it suggests, is that all across the West, there were forecasts for extremely low stream flows, less than 30 percent of normal, for much of California, Nevada, Utah, Idaho, and Oregon, and then between 30 and 60 percent of normal in much of the rest of the West.

What we're going to focus on, or what this paper focuses on, is the Yakima River Basin, which is one of the tributaries to the Columbia River. It sits in a very dry part of the state of Washington, dry because it's in the rain shadow of the Cascades.

They tend to get about six inches of rainfall a year in much of that region, but it is a highly productive place because of the storage of water in the region. There are apple orchards, grape orchards, mint grown in the region, all very lucrative to that region.

Continuing, by February of 1977, the governor of the state of Washington has stepped in. He was very concerned about the situation. There's a little quote here, which you can imagine being said during the California drought or any other recent drought.

"Despite impressive gains in reducing the effect of drought, the trends of increasing population, increasing demands for food, scarcity of resources, and growing competition for water suggests that future droughts will continue to clearly, if not severely, impact society.

Within this portion of the Columbia River Basin, the Yakima Basin, there is about a half a million acre‑feet of water. That water is allocated to five major irrigation districts. They grow the types of crops that I mentioned before.

Now, this wasn't just hysteria. In the period of between October and January of 1977, they had what was the lowest precipitation ever recorded on record up to that point.

One of the things that the Bureau of Reclamation does each year for this region, and it's typically based on snowpack early in the year, is they create what is called the Total Water Supply Availability forecast.

This forecast is their best guess of how much water will be available to the region for the rest of the year. That forecast that came out in February said that, in general, the amount of water available is about half of a long‑term average for that river basin.

This caused a great deal of concern. Within the region, people began to look around for ways in which they were going to manage, given such a low water supply.

One of the things they looked at is what's called the 1945 Consent Decree. That consent decree basically prescribed how water would be allocated. This region has both riparian water rights, as long as water rights associated with first‑in‑use appropriated water rights.

Although there were lots of options on what people could do, what they ended up doing was to apply the consent decree. Those with senior water rights were originally promised 100 percent of the water that they would typically get. Next, those with prorated water rights, or second in line, were promised only six percent of their normal water allocation.

That created a lot of actions, obviously. The primary action that occurred for those who anticipated getting prorated water was to go out and start drilling wells. Even though surface water is managed, ground water, at that point in time, was not managed in the States.

If you could drill a well, you could take as much water as possible out of that well. In addition, people began leasing water rights from one another. There was a lot of discussion about pumping dead storage, or water which typically could not be taken from the reservoirs in that region.

The figure down at the bottom there is the Yakima Diversion Dam, from which the water is taken. There was even some discussion of a very large scale projects. Those couldn't be done quickly enough, but there was some actions taken to perhaps look into the future.

There were planes sent up to try to make it rain over the region. A very early attempt at a water bank was established to help move water from one user to another.

What happened? Did they really have enough water? As I said, a lot of this action took place in early February. That's when they suggested, as it turns out, a 90 percent delivery to the first‑in‑line, the non‑pro‑rated water users, and 6 percent to the pro‑rated water users.

As the year progressed, and as new estimates of how much water would be available became available, the allocation changed.

Basically between February, and April, and May, the amount of water that they thought would be available shifted to the pro‑rated group from about 6 percent to 50 percent. May is much too late for the farmers to take action. They were unable to take advantage of this excess water that appeared after the forecast.

What happened? It turned out that flows ended up being 83 percent of normal for that period. They probably could have provided all the water to farmers that everyone wanted with only the most modest cutbacks, if everyone took a very small cutback.

Another thing that was unfortunate that occurred was they discovered that the Bureau of Reclamation stream flow model that was being used was actually a flood forecasting model and not a low flow model, and it had not incorporated the irrigation return flows which during this summer represent a major contribution to the Yakima River.

Because of that simple modeling error, these forecasts were off dramatically, and it led as you would expect, to lots of disputes between the junior and the senior water rights holders. Litigation I know lasted at least 30 years for farmers trying to sue the federal government to get back the money they had invested to deal with the drought that did not occur.

Looking back, there is a nice quote in the paper that the Yakima superintendent suggests that, "Hey it's just a forecast, it's not a guarantee," whereas the farmers took it a little differently. They suggested that drought is when the government sends you a report telling you there is no water.

Droughts tend to be institutional sometimes, rather than actual droughts. This sort of setting continued in the Pacific Northwest, and right around 1999 or a little earlier, there is a lot of attempt to make use of the new, clear signals that people thought they were getting in terms of ENSO, La Niña, and El Niño, and the Pacific Decadal Oscillation.

Alan Hamlet and Dennis Lettenmaier of the University of Washington began to look at whether or not climate signals that could be interpreted several months in advance could be used to help operate the Columbia River system.

For those of you perhaps not real familiar with the Pacific Northwest water system, the Columbia River system has about 37 million acre‑feet of storage in the system. Principally from the dams created by the Corp of Engineers and by the Bureau of Reclamation.

Of course, mostly the purpose of the bureau's dams is irrigation, although there certainly is some flight control and some power production, and the Corps of Engineer's primary focus has historically been flight control, although of course their dams also produce hydropower production.

This 35 million acre‑feet of storage allows that region of the country to produce 40 percent of the hydropower throughout the United States, so it's an extremely large and productive system.

What was determined by these researchers, Hamlet and Lettenmaier, was basically that if you followed El Nino and PDO, you began to see some long term shifts over the historic record during those periods.

There were certain situations like cold PDO and a La Niña versus a warm PDO and an El Niño, where you could see average flows during those conditions perhaps being halved. That is to say during the warm PDO and El Niño flows on the Columbia might be half of what they would be during a cold PDO and La Niña.

Just moving ahead here, they looked at all these potential six conditions, and if you look at this graph you can see there is a great deal of variability between these six climate conditions, again two forms of PDO and three of ENSO. Basically there are some indications of being able to anticipate either higher flow periods or lower flow periods if you know the climate conditions.

They looked at this over the entire record and this final figure shows shifts and changes. These are functionally around the PDO and not so much the ENSO. Let me say that given this, they came back with a structured approach at operations.

What they were suggesting in their research is that if you can calculate the ENSO and the PDO, then you can do your analysis with basically a censored set of data that is reflective of that climate condition, and you really shouldn't look at all possible futures but just those that match your ENSO and PDO condition.

The more sad part of this story is that this was never actually used inside of the operating system in the Pacific Northwest. At best we can say it helped inform the system. I want to comment now on another paper that has this provocative title. I don't know how well you can see, but I'll make this a little larger and we can highlight it.

"Weather Forecasts are for Wimps. Why Water Resource Managers Do Not Use Climate Forecasts." What we would hope in a very structured process is that you would make use of any good data that would help you better manage into the future.

Rayner was at Oxford at the time and other authors were at Oregon State University. They published this in "Climatic Change," basically to report the results of the survey of what world water managers thought about making use of forecasts. They took a wide variety of settings and interviewed operators.

I would say in summary, the paper really addressed two concerns. Do we have the ability to make good forecasts with high reliability, and what are the range of challenges associated with actually taking those forecasts and putting them into action?

The basic theme I think is one that, if you read this type of literature, you see often. It is that integrating the use of uncertain climate information into complex corporate operations is extremely difficult for managers. This is a fairly long paper, but I'm going to summarize it in a very short number of slides.

What they noted, and this was done a little over 10 years ago, they felt that the skill and our ability basically to improve weather forecasts or predict variations in weather had really slowed from perhaps the previous decade.

They asked water managers basically, "Do you use probabilistic forecasts? What are the institutional factors that affect your decision making? If you're not using them, would using them perhaps help, and what institutional changes would be required?"

I'm going to summarize their numerous pages of responses into just this simple slide. I'd say that this is a bit sobering, although in a moment, we'll see not too surprising. What they found in terms of their interviews was that all of the managers felt that their institutional setting was very unique and very complex, and that really made it challenging to use forecasts.

They admitted that managers tended to be very conservative, and conservative can mean a lot of things. Conservative can mean not taking advantage of new ideas, but the way they explained it was principally that the managers really looked at longer‑term horizons in terms of operations than the forecasts could provide still.

They felt like the probabilistic forecasts were unreliable, and that's certainly a barrier to incorporating them. They almost always felt that the temporal and spatial scale for which they were provided forecasts somehow didn't meet their needs, which is again another large barrier.

Many of them pointed to the fact that within their agency, innovations were difficult to implement. Another very interesting result, I think, is that they noted that the managers needed flexibility and once you began to incorporate forecasts in a very structured way, it decreased the flexibility the managers had.

Maybe closely related to that, they noted that their standard practices, which were in place and accepted by their agencies, did not use forecasts, and so they became very hard to put into the system.

You would like to feel that that kind of result is ‑‑ I'm not sure ‑‑ old fashioned, perhaps not accurate. If you take a look at Nate Silver's book which is, I'm not sure, two or three years old, "The Signal and the Noise," there is a really nice chapter on weather and climate forecasting.

Nate Silver in probably a much different way came to very much the same conclusions, that basically people believed that weather forecasts longer than eight days really have very little statistical power. He writes very eloquently about how forecasts really must do better than simply persistence.

Which is to say, "We think tomorrow, temperature and precipitation will be the same as today's," and they have to do better than climatology, that is looking at the long‑term average for that particular time of year.

In his book, he goes on to note that the National Weather Service forecasts are extremely well calibrated, which is to say, "If historically, the forecast is for a certain percentage of rainfall, the models do tend to reproduce that." He notes that commercial forecasts are not well calibrated.

His explanation of this is particularly interesting. He notes that, really, what commercial forecasters care about are the perception of their users, or purchasers.

Things like rain, he feels is much better to over‑predict the possibility of rain rather than under‑predict it, because people do not like to be rained on when they don't expect it. They're not quite as angry when it doesn't rain when the forecast suggests that it is raining.

Even more recently, a group of us at the University of Massachusetts did a similar survey on water managers and the Connecticut River, which is shown on the right there. We were looking at the use of seasonal, three‑month forecasts in terms of helping them manage.

Our result, I would say, somewhat sadly, were extremely similar to those many years before that basically, operators felt that the relative advantage of the forecast is really dependent upon the particular system that we're operating and how objectively accurate the forecasts are.

The group we interviewed had very little confidence and longer term forecast, only minimal confidence and forecasts longer than 10 days, and virtually no confidence that a six‑month forecast was useful for them and their operations.

That's a sad story, I would say, up to this point. Let me say a couple more things. We've sort of summarized how a poorly managed drought forecast impacted the Yakima River Basin. We looked at how researchers really see the value of incorporating seasonal forecasts, but we're unable to get them implemented in a meaningful way within a large operating system.

We have numerous examples now of water managers not trusting forecasts, both from over a decade ago to more recent studies. Let me suggest a couple of things right now. Most of us really would like to be able to make use of forecasts in the work that we do.

I think this is particularly true of something like an ecological drought, where the actions that you take may not be undone. You would like some confidence to what the future might hold.

This is just a picture on the right. This is a picture of our local river forecast system and the dotted purple lines in the right are the forecast of the stream flows, things that someone might be able to use.

There are tools out there. We'll talk about their ability to be useful. We hope that we can use to communicate with our stakeholders. Two other things that I'll cover very quickly here in the next few minutes is the US Drought Monitor and the drought outlook. I hope that most of you are familiar with these, particularly those working in the area of drought.

The US Drought Monitor, again, is a multi‑agency tool that comes out on a weekly time step. This is the one that was just posted for April 25th. I live up in Massachusetts, so you can see a very light yellow, anomalously dry area up here by us in New Hampshire and in Connecticut.

This tool has been around a little while now. I think the first ones came out in 1999. There are 11 different groups of individuals who are engaged in producing the drought monitor. It's really a production model in the sense that you can count on it coming out every week.

Basically, the attempt is to capture the current situation using a variety of metrics that are indicative of dry settings. I note here on the bottom here on the right, David Miskus provided a couple of the slides that you're about to see in preparing this talk.

The drought monitor just looks at what is, and the drought outlook ‑‑ seasonal drought outlook ‑‑ is a three‑month forecast into the future. This is, again, I think the most recent drought monitor. Maybe one came out this afternoon, but it's a very recent one.

What you can see is some drought persisting in the Alabama, Georgia, South Carolina area right now. We all know that much of California was extremely dry up until last year, and only portions of that remain in Arizona and Southern California.

Then you can see another example of droughts developing, but in general, the state of this map suggests that much of the US is not forecasted to be in drought.

A question that we want to ask about this...This is a picture of basically some of the inputs that go into the drought outlook. It begins with the drought monitor. They do an extended, two‑week soil moisture forecast. Then they use a longer‑term, three‑month precipitation and temperature outlook to attempt to update the drought outlook each week.

I want to move on to this. This figure is different than what we were just looking at. Let me take a step back. That's an example of the outlook.

This figure here is an example of the verification of how well the drought outlook is working. What these two figures suggest ‑‑ and it was great for the agency to share this figure with me ‑‑ it illustrates two examples of a point in time when the drought outlook is skillful and when it is not skillful.

On the left‑hand side, let me just say, this is a time in which there was either a developing La Niña or El Niño. What it suggests ‑‑ these scores that are in red or purple, they're on the left‑hand side ‑‑ that is if you take away the persistence measurement ‑‑ that is things just remaining as they are ‑‑ the skill score on this is about 23.5 percent. It's providing valuable information that you would not have without a forecast.

On the right‑hand side of this graph, it's for a different time period. This just happens to be June of 2016. What this one shows is that it's during an ENSO neutral period. It happens to also be during the summer where snowpack is not as important, and precipitation is more important.

The skill score on this is providing almost two percent more information than just looking at persistence. That is assuming that things will continue as they were at the beginning of the period.

I think the usefulness of this is to say that there are times a year, or certain conditions, in which the drought outlook has value. And other times for which it provides almost no more useful information than knowing the current state of the system.

Finishing this up, I just want to go back to this idea that drought planning is a structured approach. It really involves engaging stakeholders to understand what's important in a region. It involves establishing what the conditions are now, and what they're likely to be in the future, and looking at wide range of alternatives.

I'll mention again, these things typically can only be done with the aid of some sort of computer simulation model. Continuing with this, I'll just note that I would encourage managers...I would love to hear people comment on whether or not these are things that have been done with resource managers.

I would encourage you to consider the process of the virtual ecological drought. This idea is just one of practice. How does one deal with unusual circumstances? In general, droughts don't occur all the time. Those who are responsible for managing them can benefit by having experience practicing.

Imagine a gathering of stakeholders, managers, scientists, and elected officials who could ask questions like the following. How would we plan for and manage the worst ecodrought on record if it happened today, with today's population, and the environmental services that are expected today?

The worst drought on record for a region might be in the 30s or it might be in the 60s or it might be in the 80s. That circumstance is extremely different, could be different, than it was when the drought actually occurred.

The logical question, when we think about climate change, of course, is not only planning for the worst drought that we've seen, but for planning for even more severe droughts if they occurred with today's demands on the environmental services and other issues.

Along those same lines ‑‑ because droughts are quite different than floods, because they start slowly and are hard to identify ‑‑ you have to ask yourself in a virtual ecodrought exercise, how would you recognize a drought?

How far along would it have to go before you would declare it the drought? What are the indicators? What were the triggers? What actions would you take if you found yourself in a drought?

Basically, I think one of the great reasons for having these virtual exercises is to ask yourself, "How do multiple agencies, how can they be best coordinated to maximize their effectiveness?" Those are things, unless you practice them, perhaps don't occur quite as naturally as we might hope.

I'm going to basically finish up with this final thought. Imagine this exercise, that you have a highly interactive and extremely flexible, probably object‑oriented, computer simulation model, that helps you walk through their multi‑year drought.

Maybe at a monthly time step, that you could replicate the past or generate new future drought scenarios that you have to manage, that you would encourage all the participants to develop their own indicators, and triggers, and actions.

Then you would test those to see which ones are the most appropriate. You would test the effectiveness of the approaches that you took. Then, as I said before, maybe equally important, determine the institutional arrangements. Who has to talk to who? What actions have to be taken to guarantee success?

Any time you do this type of exercise, you have the opportunity to test the resiliency of your approach. How well does your approach work under a wide variety of different drought scenarios?

With that, I'll stop. I'm very happy to take questions. I certainly appreciate all of you taking your time on a busy Thursday to listen in.

John Ossanna:  Thank you, Richard, for your presentation. Real quick, while we wait for more questions, I have a few questions for you, Richard, if that's OK. First, have you used the virtual drought in real‑time situations? What was the impact?

Richard:  Thanks, John. This idea of virtual droughts came out of work that we did back in the '80s. I've been involved in these workshops in the Pacific Northwest in particular. This sort of work goes all the way back, for me, in the early '80s.

The Potomac River Basin, the water supply for Washington, D. C., Northern Virginia, and Southern Maryland, there's an agency there called the Interstate Commission on the Potomac River Basin. They ran these virtual droughts for many, many years. Way back in the '70s, there were some significant droughts.

Also, in the '80s, along the East Coast. Not to mention, the largest drought of record, the '60s. There was an annual event on the Potomac River where people used computer models to help simulate what could happen during a drought.

What that resulted in was that when they did have major droughts, and they have a unique system. They have what's called a Water Master.

It allowed them to test different operating rules and operating policies in that region so when a drought occurred, they were really set to do this. I've done this in a variety of different locations. Portland, in Seattle, in Dallas, in the Kanawha River Basin.

Again, the goal is really to give people an opportunity to practice unusual events that don't occur every year, but which do require careful coordination.

John:  My other question. Do you think we will be seeing better weather or climate forecasts in the near future?

Richard:  I'd love to hear someone else opinion on this. No. [laughs] Answer to that question is, I'm not certain that we will get better short‑term weather forecasts in the future.

There is this problem of basically model ‑‑ I'm sure there are people more familiar with this than me ‑‑ in terms of forecasts longer than 8, to 9, to 10 days, there has been very little indication that these weather models/climate models are going to be able to improve in the short term.

Part of it is just the feedback within the models where the actual modeling process begins to dominate the results of the models. It becomes more about the model that you're using rather than the physics of weather. Again, I'm not a climatologist. If anyone out there has a differing opinion, I'd love to hear it.

John:  Thank you for at least giving us your opinion. [laughs] I see a few people typing away. I do want to give them the time to do that. Are there any final thoughts that you have, Richard, while these people are typing some questions away, possibly?

Hold on one second. Do you have comments about scientific papers about risk, or likelihood of multi‑decadal drought? That's from Nancy Green. Thank you, Nancy.

Richard:  Thank you, Nancy. Where I work at the University of the Massachusetts, paleoclimate is an extremely a hot topic. On our staff, we have Ray Bradley, who was the co‑author of the paper, which is typically called the hockey stick paper, where they recreated long‑term climate records back for thousands of years.

If you look at those records, what you see on those records are exactly what Nancy is asking about, periods in the past where we had much longer‑term decadal and sometimes multi‑decadal drought periods in the past.

The work I've done on the Colorado River, suggests that those things had a very large impact in the Colorado River, hundreds, if not, a thousand years ago.

I think there's every reason to believe that we will face much more severe droughts than we've seen on our typical 100 to 150‑year hydrologic records that we're very familiar working with. The question becomes, how do we develop robust systems that can deal with these longer decadal shortages in water supply?

John:  Great. John looks like he has a bit of a comment about where he is in the Pacific Northwest, watching the ENSO updates from NOAA and the Potential Service Water Availability. Great.

Richard:  Thank you. I appreciate those comments. That work that I was looking at is back with Lettenmaier and Hamler back in about 2000. That scenario in which basically the Pineapple Express, which is impacted by El Niño, La Niña, really has a large impact on annual precipitation. Should be a strong signal. It doesn't always work, but it should be a strong signal.

John:  Excellent. Do you feel there is a data gap in empirical continuous monitoring stream flow data to inform drought forecast?

Richard:  Great, Adrian. Thank you. Let's see. Some of us remember the USGS Stream Flow Network prior to Ronald Reagan. In fact, a portion of that network was dismantled during those years with the feeling that we have been collecting data for almost a hundred years, and why would we need more?

I probably do not fall into that camp. One of the things most of concern in ecological droughts is our ability to forecast stream flows, and smaller streams, and even streams that are ephemeral.

As the network gets smaller, and we focus on only the larger streams and streams that somehow are viewed to have clear economic value, but sometimes miss the smaller upper watershed streams which are extremely important to many of our environmental concerns.

I do believe we're developing data gaps where we could learn much more about how smaller streams behave during extended droughts.

John:  OK. Well, once again thank you Richard. I'd like to extend an invite as well for our next webinar in this series. Which would take place on May 18th at 3:00PM. That presentation is on hydrologic research and assessment from local to regional scales. Our presenter will be Jacob LaFontaine, who's with USGS South Atlantic Water Science center.

Be on the lookout for that, and also be on the lookout for the information of this recording will be available online shortly. I look forward to all of your future presentations, or join us for the future presentations. Thank you to everyone who joined us today. Thank you again Richard and Elda, for the cooperation with USGS. Thank you.

Richard: Thank you very much.