2026-02-05_Water-Resources-Podcast_Dennis-Lettenmaier
Bridget Scanlon: [00:00:22] I am pleased to welcome Dennis Lettenmaier to the podcast. Dennis is a distinguished Emeritus Professor in the Department of Geography and Civil Engineering at UCLA and joined the faculty there in 2014. Prior to this time, Dennis was a Professor at University of Washington from 1976. So Dennis has covered a wide range of topics during his career, and so it's hard to focus on different things, but mostly hydrologic modeling and prediction, remote sensing, hydro climate and water management.
Bridget Scanlon: Today I think we're going to talk a bit about the Colorado River Basin, his work there, importance of snow in the Western US, flooding trends and then some general issues related to modeling and remote sensing. So we may not get to everything, but that's the idea. Anyway, thank you so much, Dennis, for joining me.
Dennis Lettenmaier: Well, thank you for having me.
Bridget Scanlon: So Dennis, I think it would be good to start off with the Colorado River Basin work because the regulations are being, it is being re-adjudicated or renegotiated in this year. And so there's a lot of interest in the declines in stores that we've seen over the past couple of decades.
You did some early work there and you talk about different droughts in the mid-1900s and also more recent droughts and emphasized the importance of temperature in the post 2000 droughts. And also your work you show, which sub-basins really contribute most of the flow in the upper basin.
So it's a very nice analysis. Maybe you could describe it just briefly.
Dennis Lettenmaier: Sure. So I think the work mostly in the Colorado River goes back to a 2004 paper that was part of something of a of a group of projects called Accelerated Climate Prediction Initiative, which the late Tim Barnett then at Scripps had convinced DOE to fund. And, so there were three major river basins across the west, the Columbia, Sacramento, San Joaquina, and the Colorado.
Interesting side note to that is that Tim got the money to do this, came around, wanted us to do some modeling, and we've been working on big model applications, especially in the Columbia. And he wanted done in three major river basins asked me for a budget, I put the budget together and he called me up and said, well, we don't have that much money.
You're going to have to cut something. So I said, okay, well, we'll cut out the Colorado River and we'll only do. So what ended up happening is a couple of months later Nicholas Christensen, who's a master's student came into my office and, and he needed a PhD topic. And I said, well, you know what? We already outlined what we would do in the Colorado.
Why don't you do that? So we ended up doing it anyway. And that first work is Christiansen, et all. 2004, I think it's climatic change. That's gotten a fair amount of attention. But Nicholas put together and we'd done this a number of other places, the VIC model, the model, the hydrology, but he also built a simple reservoir model of the system.
And we had a little bit of experience by then in dealing with reservoirs because it's one thing to go simulate the hydrology, but you know, the storage in Lakes Powell and Mead is four times the mean annual flow of the river. So everything's about water management, right. Just, just saying this is what the hydrology is going to be.
That that's half of it. The Bureau of Reclamation, of course, has a model of the system. It was very clunky and basically impossible to get your hands on, would've had to have been run on their computers and so on. And like many of these other things, there's certain parts of it are kind of hardwired to the historic climate.
So he put together a model that kind of produced the main aspects of the of the Bureau's model, but then was able to explore things like how power generation would, would fare water deliveries for irrigation, especially in water deliveries to Mexico. So we looked at both of those at that time.
My recollection is the climate projections, and that's really what the project was all about. How is climate going to affect all of this. We're coming from CMIP3 and it turned out that, and that was the third IPCC assessment and the climate models were run for that. There's a whole bunch of downscaling and other parts of this that go along with it.
But the bottom-line kind of was that group of models was quite dry across the sort of lower middle latitudes. And there was a lot of differences in how much the precipitation would decline. Our numbers, if I recall correctly, actually were towards the low end of the declines. In other words, less negative than what various other people. Some of this had to do with the downscaling and there was a whole bunch of discussions and meetings and so on, after that. Some of this got into Julie Vano wrote a BAMS paper on reconciling different projections for the Colorado River and why they were, and so on and so on. That came about, I think about 10 years later. But we did do that first study a little later along came the fourth IPCC, which was CMIP5, if I have this correct.
And the projections were for the drying to be less over the Colorado River basin. There was a second paper in 2007, kind of did an update. Now I may have the projections a little bit mixed up there. But basically, what happened was kind of interesting. I was on an academy panel whose primary charge was to try to reconcile different tree ring records because there had been a lot of work done in trying to reconstruct long-term flows, pre instrumental flows of the Colorado river. But there was a part of that doing with having to do with the instrumental record. The long and the short of it is I wrote a section for the Academy report and I'd had Nicholas, who was still around, and I asked him to go run the new climate projections.
We put the thing in there and I was told, well, you know, the problem, the thing is with academy reports, you can't have new work that hasn't been reviewed. This hasn't been reviewed, so we're going to have to take it all out. So, what we did, I don't think I'd ever published in HESS before, but Hydraulic Earth System Science, the EGU Journal at that time, they produced a pre-review copy, I think it was called discussions.
I'm sure you're familiar with this. You probably had papers in there as well. The discussions papers look exactly like a printed paper. So, and they do it, they, at that time, they did it in about two weeks. So this thing, we submitted it, it turned up as discussion. So then we cited as a discussions paper, and they put it in the academy report.
I think the date on that was 2007, the review comments came back and so on, which weren't that substantial. And, and that was the permanent version of the paper. So those two papers both looked at what the hydrology changing and, and even, you know, the one thing that kind of settled out of all of this is that the sensitivity in fractional change in runoff divided by fractional change in precipitation, which is what it is on the annual basis. That number is about two for the Colorado basin as a whole. So basically, if you don't know anything else, if you got a 6% decline in the precipitation, you're going to have a 12% decline in the runoff.
It's that sort of thing. And we'd looked at that. Also there was a lot of consternation in the change between CMIP3 and CMIP5. I never quite understood why there was no CMIP4. Okay. There was only CMIP3, 5, 6, and now they're working on on seven. But that change was fairly substantial and I think there was kind of a sigh of relief that, okay, things aren't really going to be as bad as we think, but that got into a lot more discussion of, you know, temperature sensitivities as well as precipitation sensitivities.
And that's carried on into some of the more recent work.
Bridget Scanlon: The Colorado is really fascinating somebody was suggesting to me recently. I've listened to a couple of books, I can't read books anymore, it takes too much time. But I was listening to a book by Kevin Fedarko called The Emerald Mile.
Dennis Lettenmaier: No, I know it. I've, I've read the book. Yeah. I actually read them. I don't like listening, but yeah. That is a fascinating book.
Bridget Scanlon: Yeah. Yeah. I always have a book that I'm reading and a book that I'm listening to. And he described Kenton Grua and going down the Colorado
Dennis Lettenmaier: Yeah. This was '83 right?
Bridget Scanlon: And so I guess there was so much flooding that they made releases to Lake Powell through the dam there and so really high flows, and they made it down in about 37 hours
Dennis Lettenmaier: It was some, yeah. Some remarkable time, but there's a lot of other stuff in the book. About the sort of spillway issues and cavitation, and,
Bridget Scanlon: Oh, yeah.
Dennis Lettenmaier: you know, reportedly there were chunks of concrete coming out of the spillway, the size of a Volkswagen and so on. I mean, depending on who you listen to, it's, oh yeah, it wasn't really that big a problem or, you know, they almost lost the dam.
But certainly they had not had, since Powell began to fill, and I can't remember, was that seventies maybe? Because it takes, the storage is double the flow of the river at that point takes a long, long time to, to fill it. They hadn't seen anything like that level of inflows.
And, and the other side story is very interesting is that. The forecasts really of the inflows were really a bust because what happened is there were some very substantial late seasons snow that, that is very abnormal. And the forecast models basically were all, typically they regress on April 1 snow pack.
And those regressions showed, yeah, I mean, it was a little bit above average, but it wasn't anything substantial. But then there was all this extra snow that the forecast models didn't know about. So that's all kind of the backdrop of the book, which is pretty fascinating.
Bridget Scanlon: And I guess 83, if you look the flow, at least very 83, that period was extremely high. Probably the highest on record, probably similar to, when the Compact was formed in the early 1920s. As high as those flows.
So anyway, if people are interested, that's and Kevin goes into a lot of detail about the dam and the Bureau of Reclamation and all of that, and you mentioned them earlier in their model and also VIC the variable infiltration capacity model that you guys developed. so that's excellent.
There have been a couple of papers out recently Chris Milly’s paper on the importance of temperature and increased evapotranspiration at the Colorado River Basin and the Clausius Clapeyron equation, with the increasing ET and what that would do to stream flow.
And then Marty Hoerling seems to take a kind of, maybe a different viewpoint in saying that maybe we will have some wet periods like we've seen in the past and that we will have high stream flow potentially sometimes. And his bounding estimates, minus 25 to plus 40% in the future.
And he mentioned that if he does his modeling and conditions it on the current drought from about 2000 to the present, that there's an increased likelihood of a wet period occurring in the future. So I'd really like to get your thoughts on those two viewpoints, importance of temperature and evapotranspiration versus likelihood of getting El Nino. And would that be swamped by the ET or everything is so dominated as you said by snow in the upper basin. How are those things going to pan out?
Dennis Lettenmaier: So they're Hoerling, let me go back a little bit because, there's a history backdrop to Marty Hoerling's work on the Colorado River, which perhaps isn't as well known. There was a paper of his early two thousands, sometime in a regional journal that doesn't exist anymore, but it was published down at University of Arizona. And in that paper they put together a very simple hydrologic model of the Colorado. This was the hydrology only no water management. And I don't remember which set of GCM scenarios they wereusing, but they concluded that the flow of the river the next century was likely to decline by 50%. A lot of people said, including myself, said, we don't believe that there were actually a series of meetings to try to figure out what was going on.
And some of the backdrop was this whole business of reconciling all of these different estimates, which showed up eventually in about 2010, 12 in this Julie Vano et al. paper. But it turned out that the major issue there was the very simple model they had was a very crude spatial scale model, which did not represent the effects of snow at high elevation and rather, obviously if you average snow pack over the entire basin and you take an average elevation and you figure out what's going to happen, the snow's all going to disappear because it's not cold enough anymore, then you're not representing the high elevation snow.
So Marty kind of disavowed that model and moved on to something else. The paper you mentioned, and I think they get sensitivities that are to us very low to temperature. They have numbers, if I remember correctly, are kind of like two, four, or something percent per degree C.
The Milly paper gets about 10% per degree C using a combination of satellite data for the radiation and some modeling for snow cover and so on.
We typically have been getting numbers and both Julie Vano’s work and more recent work by a student who wrote a paper or something on the causes of declining Colorado River Stream flows or something that might have been about 2018 WRR paper. We typically, with VIC modeled, been getting about 5% per degree C or a little more than half of what Chris got.
So there's kind of two things. One, the Milly paper, Milly and Dunne paper. The science one, 2020, I think is really a seminal work because it shows the reason for the declines on a physical basis, and it all distills down to something that's actually pretty simple, that most of it has to do with contrast in albedo at high elevations with warming for the period when in the warmer climate there's no snow.
And in the cooler climate there is that period is like a matter of weeks. But it happens at the time when solar radiation is the highest because you're at, at or near the vernal (spring) equinox. And that kind of extra net solar radiation gets converted into evaporation and by different society of absent precipitation changes.
hat difference comes out of the flow of the river, right? So that's where he gets the number. So what's kind of interesting, as an aside. My student wrote a paper looking at the so-called Millymechanism across the entire entire west to see where it was greatest and so on. And it shows, it mostly is maximized over the Colorado River area, probably in part at least because it's fairly far south in terms of where most hence the solar radiation is higher, but as an, and there's a bunch of stuff in that paper, but we kind of put in almost as an afterthought, buried in the papers, an analysis, trying to understand the reason why we get 5% and Chris gets 9% or so.
And there's a whole bunch of differences in models, methodology, and all the rest. But what about half of that difference distills down to is changes in net longwave radiation. And it's kind of interesting because, you know, in terms of net radiation that melts snow, you think solar is obviously going to be the big number. Solar isn't really affected by climate change, not directly that much. I mean, there's transmissivity of the atmosphere which changes and so on and so on. But in, in the overall scheme of things that's somewhat minor, what happens is in all the simulations we'd been doing for our simulation of long wave.
Basically, we just increase all the temperatures. And if you do that, you're taking differences of absolute temperature to the fourth power. So what happens if you increase that and you kind of got an effective temperature of the atmosphere and effective surface temperature? Well, a whole bunch of subtleties and, and details aside, if the atmospheric temperature and the surface temperature increased by the same amount, you increase them both the net or the downward component of the long wave actually goes down.
The effect of the difference, you get the differential long wave goes down, but in a warming world, you've got increased water vapor and increased CO2, which actually increased the emitted atmospheric emitted. When you accounted for that effect, it flips the sign of the change in longwave so that the net longwave is actually increasing instead of decreasing.
And I think by accounting for those simulations, and we were getting numbers more like 7% or so. It doesn't close all the differences and there's a bunch of methodological things, but it's kind of interesting. We since then have looked at this in the re-analysis, the ECMWF5 Reanalysis. And it appears that the trend for downward long wave in the re-analysis is actually negative. We've got to look at that a little bit more carefully. So do they not have the CO2 and water vapor effect in the re-analysis? I'm not quite sure there's something interesting to look at there.
Bridget Scanlon: I think you know it's healthy to have these different viewpoints because it forces us to think about what's going on and then to question different things.
Dennis Lettenmaier: Well, and also the basic physics. We never really thought that much because longwave is a fairly small part of net radiation in the springtime, to worry about that. And we realized, you know, the, that, warming signal in the atmosphere isn't in there. And he does in Chris's work, it's there because he's getting the long wave and the short wave out of some series satellite products.
So, arguably, it's observations and there it is kind of interesting that over 20 or 30 years that is enough to go. Cause you know, a few I believe 10 or so watts per square meter, which is not negligible.
Bridget Scanlon: So you mentioned earlier that the 83 flooding and, the discharges from Powell and stuff related to late season snow that they missed in the forecasting. Could we possibly get, some events like that again and if we got those events would we get high stream flow?
It seems to be a lot of uncertainty. I realize that temperature is much more robust parameter to predict in the future and all that sort of thing. But, there's a lot of uncertainty and and all sorts of things going on. It seemed like we could potentially get some wet years.
Dennis Lettenmaier: Well. Yes. I mean, it would not surprise me terribly if over the next 10 or 20 years we saw some kind of shift and you started getting some big precipitation years. I mean, that the climate models are notoriously uncertain in precipitation. The, you know, where everything lines up, the observations and, and the climate models is, is warming.
There's differences in how much, but it's pretty clear. It's been happening. It's been happening especially across the west and, and it's quite likely that'll continue that we have reasonable understanding of, but precipitation, pretty uncertain. A lot of decadal scale variability. You mentioned that wet years, in the early eighties, a series of them. And then, certainly back in the, you know, 1919, 1920 Colorado River Compact years is it certain we're never going to see that again? I certainly wouldn't go out on a limb and say that.
Bridget Scanlon: Dennis, you've done a lot of work on snow. And the snow is hugely important. The upper Colorado River Basin provides about 80% of the water for the entire Colorado basin, and snow is a primary driver. So you've looked at it in the Sierras we often hear in California, the governor going out April 1, what's the snow like and all of that sort of thing.
So you have described some of that in your work decreasing snow and losing that reservoir storage that we get from snow, and then importance of rain on snow. Maybe you can describe that a little bit, Dennis, and the importance of snow in the Western US for many of these basins.
Dennis Lettenmaier: Yeah, well we, as an aside, there's a paper written by Dongyue Li, maybe late 2017, 18 timeframe. Dongyue was a postdoc who worked for Steve Margulis in civil engineering and myself. And I had suggested to him that he might, this was when he was still a student at Ohio State that if he was looking for something to do, might see if he could develop some sort of algorithm, which gave us some real basis for understanding how much of the runoff in the Western US comes from snowpack.
And the side story on this is that, lots of people waved their arms and said how important it was. And there even were some numbers in some papers that were basically made up. And, and then those papers got cited as the source, right. And we were guilty of this ourselves because it's the kind of thing in the introduction to your paper, you, you say how important it is, right?
So Dongyue went in and for the big model put a tracker on it to go see in the model world where the water was coming from, how much originated as snow. And I mean, there's a map in that paper that's been fairly fairly widely used. So at least for that model, we now have some numbers.
Not surprisingly, aside from the coastal drainage areas where it doesn't amount to much for the Sierras, Cascades and so on, it's numbers that go up in the probably 60 plus percent range. And if you go to the interior of the west, even higher, up to 80% in some of the colder area. So, it's a big, big deal.
And for water management assuming the precipitation doesn't change you, you've still got the precipitation, but the water's going under the bridge earlier in the year and that, that's actually happening. There's a paper by Iris Stewart who was a PhD student of Dan Cayan in early 2000's or so. I think it's 2005. I've used a figure out of that paper in many, many talks because it basically just shows the change in timing and that that's been advancing earlier in the year over the last half century.
Bridget Scanlon: Yeah, so you're right. That would be critical for water management because the reservoir storage then that they have is not enough maybe to store it, to keep it for later in the summer than when you have dry periods in California and Central Valley and other regions like that.
Dennis Lettenmaier: That is the challenge. And it's interesting right now, we're in a, basically what people call snow drought across the west. The snow pack is way low. I mean, I was at a meeting last week in Salt Lake City and basically the, the Wasatch Range, there's hardly any snow on it.
Similarly in the Cascades. I occasionally read the comments, the articles in the Seattle Times, and there was a discussion. Well, I mean, had some very warm temperatures last week and, yeah, this is great, but I mean, we got no snow pack.
What's that going to mean? And somebody was, commenting, well, the Yakima River reservoirs are way above average. And, they had big floods up there in December and that took whatever early snowpack there was off, they got some cooler storms, but then there was another warm period.
And so there's not much snowpack. So the interesting question is, the water's in the reservoirs that should have been in the snowpack. Right. But what happens? Well, the problem is, and unless there's a big shift, which could happen, but they're kind of running out of time because by the time you get to about early March, you're not getting much snow accumulation anymore.
So realistically, there's about another month. There'd have to be a bunch of cold, very high precipitation storms, which it doesn't look like is going to happen. So what's going to happen is, is that, the reservoirs will be full or close to it early on, but they won't, you know, they begin to start irrigating in April. And generally, you've got inflows from snowpack continuing on into like June or something. Yakima River System is fairly low elevation. The higher elevation stuff in the Columbia, the peak flows are actually in July. Right. They're not going to get that they're not going to get that flow. And it's going to be a, it's a big water management problem.
It's better than the years where the drought is, you know, warm, dry drought so that you just didn't have the precipitation at all. But it's certainly way worse than a normal situation.
Bridget Scanlon: You've done some work on, so if snow is decreasing, then rain on snow and it's linkage to extreme flooding or whatever can you describe that a little bit?
Dennis Lettenmaier: Yeah, well, the rain on snow thing is kind of interesting because this is one of the climate community likes to think that, oh, flooding is going to be worse. It actually turns out, and the same postdoc Dongyue Li has a paper on rain on snow events across the US and their sensitivity to warming in coastal Mountain Maritime heading streams like those from the Sierras and the the Cascades.
Generally, rain on snow floods go down with warming. And the reason is less antecedent snow. Okay? In order to have rain on snow, you have to have snow, but if it warms up, you end up a lot of that lower elevation snow pack there. There's no snow pack there in the warmer scenario. The places where you get more, is the colder regions of the interior of the West. And there's a figure in his paper that that shows that pretty clearly the places where they show some increases are high elevations in the interior in observations.
We wrote a paper. It actually started as a class project in my graduate class.
Huilin Huang is the first author, maybe it's a 2022 or 2023 paper in GRL. Looks at changes in flooding across the Western US and they do a classification into different categories and they split things up. In the interior of the West, you tend to get snow melt floods, so the peak floods mostly come from warm temperatures in the spring. This would be streams draining the Rockies and so on.
In the coastal west, most of them are associated with atmospheric rivers. That's one of the partitionings. But you also, they partition out rain on snow events as separate from the others. And it's based on how much antecedent snowpack, and so on.
And then there's a monsoon category down in the southwest and so on. Turns out that the only statistically significant changes in any of the categories is the rain on snow events, which have been decreasing and they're predominant in the coastal river basins and it's for exactly the reason I just described.
We surmise that less snow is led to a general warming. There's a little bit of an increase. There's not very many stations where there's dominant mechanism, but in the interior where convective storms, where they tend to get summertime convection can lead to at least some of the major floods. That's actually seems to have gone up a little bit. Everything else, there's nothing statistically significant going on.
Bridget Scanlon: I really appreciate that you guys look at the data in detail. Oftentimes we kind of make these logic jumps, oh, this is happening and this is happening, ergo this is going to happen. And then people forget to look at the data, what has been happening? And so it's very important.
Dennis Lettenmaier: A lot of this stuff is more complicated than it seems at first, and even the snow stuff. When we started on this back of this EPA report to Congress in the late eighties the details have all changed, but, the big picture was warming less snow change in the seasonality.
And there have been a gazillion other studies done since then. And in the West, I mean, that's basically the result you get. But there's a whole lot of fine points that kind of get buried in that. And there's a really interesting paper by Keith Musselman at NCAR a few years back, which is the title is something like Slower Melt in a Warmer Climate.
And the first thing is slower melt and warmer climate. What do you mean? Well, the reason is, is that as things warm up, snow packs move earlier in the year and you've got less available radiation to melt. So the melt rates tend to go down. And I mean, he shows this from a very detailed analysis of a whole bunch of observations at Snotel sites and so on.
So there there's other things that kind of go on in a warming climate. The other one that we wrote Zhaoxin Ban last dissertation paper is something like decelerating rate of change in Western US stream flow or something like that. And it's like decelerating rate of change.
Well, it's fairly obvious when you think about it because as the snow slowly goes away, as you warm up, there's less and less of it there. So there's less to be impacted by warming. And so the rate of change tends to go down and ET, runoff and so on and so on. So it is a little more complicated than is often made to seem. I think in the realm of flooding, it's kind of analogous to the temperature. What do we know and what don't we know? Well, you expect precipitation extremes would change. There's a lot of complexity there, but just because water holding capacity of the atmosphere going up that more water in the atmosphere and if it gets tripped off, you could get higher extremes.
And there's some evidence of this happening in the observational record, okay. We'd done an analysis that was never published, but of a thousand stations or something across the CONUS. And it was something like at 5% significance level, there were significant trends in about 10% of the stations in the extremes.
The annual maximum precipitation, I can't remember what duration. And about two thirds of those were upward. We also did some work looking at major urban areas a few years ago, and there seemed to be increasing changes, but in the floods we're not picking it up. And so interesting is why not?
Bridget Scanlon: Before we move to the flooding, I would like to just mention a little bit the, I really liked your paper on the 2023 Snow Deluge in California, and you did
Dennis Lettenmaier: Well, first, first off, I, I have to correct you. That was Adrian Marshall's paper and she kindly invited me on board, basically as, as a co-author because there's some VIC simulations and I can't, I think they were some ones that we had archived or something.
Okay. She did all the heavy lifting on that paper was Adrian's.
But I do think it's a very interesting paper
Bridget Scanlon: And I really nice I love one of the graphs in there, where you talk about wetter and cooler and look at the different periods and, so much emphasis on drought periods. I've memorized droughts in different regions, but we don't talk as much about the wet periods.
And I think that was a really nice analysis. And, 2023 snow deluge, is that linked, do you think, to, I presume that's linked to the atmospheric rivers too, at that period.
Dennis Lettenmaier: Well, there were certainly a bunch of very, very strong atmospheric rivers that contributed to that. What's interesting is because you mentioned that paper, I went back to look at it because I remember exactly what was there and there is an interesting figure in there where she flags I think a total of five of these years in record. And, the most recent one prior to 2023 is, I guess it was in 83, I think it's 83, was the one also the same year as the Colorado River event that you'd mentioned. So it's been 40 years, but the earlier ones are all separated by more or less 10 years or something.
So that's not a statistical analysis, but it does kind of suggest they're not happening as often. And she does an analysis with future climate runs and so on, and shows that you'd expect the probabilities of return periods of these things go up in a warmer climate, but you're not going to see as many of them, but it doesn't mean you're not going to see any of them.
Bridget Scanlon: Another paper that I really liked was If Precipitation Extremes are Increasing Why Aren’t Floods
Dennis Lettenmaier: The backdrop for that, my main contribution to that paper is that I made up the title. The lead author is the Ashish Sharma. The second author is Conrad Wasco. I was the external examiner on his committee. I think it's University of New South Wales in Australia. And he did, I think the best job of addressing this. There's a whole series of papers now about almost 10 years ago from his dissertation, but basically positing different reasons why, you know, the climate community waved their arms and say, because of increased water holding capacity, the atmosphere, more of extreme precipitation because of more extreme precipitation, more floods.
But you know, those of us in hydrology know and I have actually have a visual in some other talk, what causes a flood? Alright, well, extreme precipitation or melt if it's a snow melt driven right? Obviously increased extreme water supply. But there's a bunch of other things. A big, big one is antecedent soil moisture, right?
I mean, everybody in hydrology sort of knows that. I have a figure also, and I can't remember the author of the paper. I came across it somewhere where they had conditioned, I think they used model soil moisture from the VIC model, but observed flood extremes by regions, I think across the CONUS, and then they just plotted the fraction of the upper one percentile, one percentile or something precipitation that led to upper something or other percentile floods, and they categorized them by wet and dry soils.
It was just a bar plot. In fact, I think I re plotted as a bar plot. They had it some other way, but you know, not surprisingly, the bars are way, way higher for the wet antecedent condition. So what Conrad said was that, well, in a warming climate, what, you know, you're going to get enhanced evaporation. What's that going to do to antecedent soil moisture?
Well, you expect it's going to make it drier. So wouldn't that lead to, a decreasing effects? One of these things, like many of these things there, there's more than just one aspect of it. He also hypothesized for some Australian conditions, warm season precipitation, which is not what we have in Western US, but that the scale of storms well the intensity, the maximum intensity appeared to be going up.
And he has some observations to show this, that the spatial scale is going down. And so now, depending on how big your river basin is. You may have, some parts of it have got in enhanced precipitation, but much of it may have reduced. So, and, and there's some other things potentially go on in there that we don't, we haven't entirely sorted out.
Bridget Scanlon: But I think it's great that you ground it with data and you look at the data, and, it's very easy to make the logic steps, increase precipitation extremes, increasing flooding, that's easy to think of, but to look at the data then and really examine it and then evaluate the processes that feed into it.
But, most of us, we, we just can't think of more than one thing.
Dennis Lettenmaier: The other thing that goes against this, and I kinda rant and rail against this because, you know, I can afford to, I don't have to publish anymore than I want in any place I don't want to. But, you know, so-called high impact journals, want to see big splash type of stuff. So, you know, a paper that shows extreme floods increasing is going to get a lot more attention than ones that argues that they aren't.
And so you, we get a certain amount of this, this stuff that, that is a little bit of selective processing.
Bridget Scanlon: People have to sensationalize stuff to get them in those high impact journals, it seems like. And I guess, one was the precipitation extremes. The other was increases in temperature do not translate to increased flooding. That was another analysis.
And I guess your study was saying or, some study that you were involved with a global stream flow extreme more likely to be decreasing than increasing with climate change and flash flooding, more likely related to snow melt processes.
Dennis Lettenmaier: Well, there, there's a bunch of aspects of that and again, that some of this kind of comes into what we know and what we don't know. I mean, I don't want to seem like the climate skeptic saying, no, there's no evidence of increased flooding and so on. You know, there are a number of potential reasons for this.
But, but amongst various others, we analyze what's going on our data at the places where we have it. Where do we have it? We have it where there are stream gauges. Where do they put stream gauges? Well, in general, they don't put very many stream gauges and certainly not ones with long-term records in small urban basins.
But what would you expect in an urban basin with high impervious area? Well, you'd probably expect a stronger connectivity of changes in flooding, the changes in precipitation than you would in larger basin with a bunch of these other things going on. And certainly, you know, a parking lot doesn't, there's no antecedent soil moisture in a parking lot, right?
So, but how many gauges do we have that represent those highly urbanized basins? Not very many. The other thing, may well be just floods are extreme events. We're looking at extreme precipitation. We're out in the tails of the distribution. The signal may well not have emerged from the noise. Doesn't mean that there isn't something going on there that when people look at this 50 years from now, it may be more obvious.
We don't know that yet.
Bridget Scanlon: Dennis, much of your career, you've been involved in various aspects of modeling and more recently you've been looking at the National Water Model, that's promoted by NOAA and comparing that to what the National Weather Service Forecasts are doing.
And you've been doing that in California and other regions. Can you maybe briefly describe the National Water Model and you know how you feel like it compares with the River Forecast Centers.
Dennis Lettenmaier: Well, so that, that's an interesting question. We, you know, developed the VIC model 30 years of effort into that and, so on. Done a lot of studies with it. In the more recent past, since I've been at UCLA, we've done several studies where we've tried to go look at different models so that we can avoid potential criticism that, well, is this all model dependent?
So in my mind, I mean the hydrologic core of the national water model as, as it were, is NOAH-MP Right? And we've run that in a number of cases. It, it's sort of interesting. The Noah-MP comes about from the old Noah model, which was the land scheme that the National Weather Service used as nearly as I can tell.
The only thing in common between Noah and Noah-MP is the name. Okay? They're completely different models. And the old NOAH model was not what I would call hydrologically credible. It focused much more on kind of land atmospheric exchanges than things like runoff production, which it, it didn't do very well.
Noah-MP is in my mind at least, is hydrologically credible. That's what I would say and that's one reason that we started doing some work on it. The other reason is that we've been working closely with Marty Ralph’s Center down at Scripps Center for Western Weather and Water Extremes. And a guy named Rob Hartman, who's retired now from the Weather Service, but he was a hydrologist in charge at the Sacramento River Forecast Center for a number of years. So he's been doing a bunch of work with them and we'd had some discussions and I'd asked about National Water Model slash NOAH-MP and why the Weather Service doesn't use that.
And Rob is kind of a crusty guy. He just said, well, because it doesn't work. Okay. It's like, well, what do you mean? So that is some of the work that you by Lu SU, a PhD student of mine who finished a couple of years ago I charged her with going off and looking at performance of Noah-MP and Rob through his good graces because he does have long-term connections with the RFC, was able to get from the Sacramento RFC and the Northwest one, access to archive forecast for actual floods. Okay, so we had the forecast. And for Sacramento, we went back 20 years, and we looked at a bunch of these coastal basins up and down California, and then up in the Pacific Northwest as well.
And there's, there's a couple of papers around on this. There's a more recent one by my current, soon to finish student on this as well that with proper calibration and so on, NOAH-MP seems to be comparable and in some cases there's even a little better, I think that there's a lot of not invented here stuff going on. Now, as an aside, at this meeting I was at, in Salt Lake City, which was like Colorado River Water Users group or something like that. A lot of bright people from universities, different agencies and and so on, I was unaware that the whole hydrologic forecasting aspect of the National Weather Service has been reorganized, and somehow the RFCs are now directly connected to this water center down in Alabama. I don't know if they are going to make the shift or not. It will be interesting. It'll be interesting to see.
Bridget Scanlon: So the last thing I would like to just mention, I mean, Dennis, you use all the data that you can access and you've been integrating remote sensing into your analysis a lot over your career. And I guess the most recent thing is the SWOT mission and what it does. So maybe you could briefly describe, how you think- and SWOT is surface water and ocean, topography- how that helps with the river discharge and reservoir storage and all of that sort of thing.
And your analysis, you've done some analysis and reservoirs and stuff in the US. How you think that will advance our understanding.
Dennis Lettenmaier: Yeah, so the, connection with SWOT, is a long term one. I spent a year at NASA headquarters in 97, 98. There were a couple of workshops on post 2002 mission planning. Well, it was post 2002 it was because the EOS, earth observing satellites, had been launched by then and they were looking at a so-called budget wedge and what are we going to do next, basically.
So there's a lot of sort of blue sky thinking, and I remember one of those that there were two major workshops and, you know, kind of what are we going to do about water? SWOT was one of 'em. So the soil moisture people were there. That was a pretty strong community there that had been in place for a while.
It took,them, eventually turned into SMAP. Took a couple of tries before that of missions that got canceled and so on and so on. But that was clearly on track. The snow community, frankly, was better organized than the surface water people. But for whatever reason couldn't quite coalesce on a direction to go.
And then there was a whole idea of altimetry, could we do something with altimetry? So the oceanographers had altimetry since Topex Poseidon. I think a launch in that was 92 or something like that. But that's tracks with very large separation. So there's some work out there that has been used on very large reservoirs, Amazon River and so on.
But in order to get anything out of it, you really needed track lengths of like 10 kilometers or so. So it was very big stuff and the discussion ended up evolving into doing SWOT altimetry so that rather than having tracks, you would basically have maps. And I won't go into the technology on that, but it turned out as this all evolved and it took decades, right?
I mean, those workshops were 97 and 98. And the launch of SWOT was what? Late 2022. So however, many decades that ends up being a bunch. But there was a series of academy reports, the first decadal review. I was on that in 2000, and the report came out in 2007, and that was the first time Earth Science had come together to say, okay, what should the priorities be?
And there's a whole lot of fine points there. We won't go into, but. It was clear by then that the oceanographers were interested in SWOT Altimetry because some of the stuff they were doing well, they got a lot of good science out of the sort of Topex, Jason and other altimeters, they couldn't get close to the shore and there were certain other things, small scale eddies and so on.
So they had an interest in it as well. So it became clear. That if something was going to happen with surface water is going to be done jointly with oceanography. So there's another case where my main contribution to the mission was the name I chaired the part of the academy committee. There, there were seven different subpanels and the one on water I chaired, and we were a little bit paranoid if we got involved with the oceanographers.
They had such a long history that they would push us out of the way. So we wanted a name where the hydrology basically came first. So that's where SWOT, where SWOT came from. So going forward in time, the technology began to solidify. Various other things happened that there were three tiers of missions, and SWOT was tier two, which probably ordinarily wouldn't have gone, but there were some other missions where various things didn't work out and they kind of got pushed off to the side and there weren't any very major deal breakers on SWOT plus the French put up, I think it was a couple hundred million euros, which was a big, big thing.
The, that I don't think SWOT would've happened. Even, although about 80% of what ultimately was spent was US. That 20% was really, really important. So to me, the mission having launched is in, in terms of global hydrology is, is a huge, huge breakthrough. And we're just starting to get, you mentioned, we've done looked at some reservoirs in California where, where we have good storage estimates from in situ and can compare and so on.
The thinking is that you probably for surface water bodies with surface areas greater than about one square kilometer. So you can get pretty good data out of it. Then this is just starting to come. I mean, it will, there's a lot of details on processing and so on and evolution of the processing and it will only get better with time.
You know, lots of discussions with that ongoing now. But it is pretty interesting stuff. So for the storage, reservoir lake storage, and if you try to figure out how much water is stored globally and what are the dynamics of that, what do you have to work with? You basically, you got little pieces of things you can get here and there, but you can't get systematic information.
Now we have a chance to have systematic information for river discharge. Lots of effort going into that. It's more complicated than the storage because you basically have to derive you, you get the slope gets observed directly and river width. So typically river widths, they're thinking a hundred meters possibly 50 m.
It's in that range. You won't get quality of data or quality of information on river discharge that you would get out of say, a USGS gauge. But this will also all evolve. Probably it's going to be done with data assimilation into hydrodynamic models, which are already our efforts to go run global scale hydrodynamics and so on.
It's only going to get better in time, but it's a huge, huge resource and trips off a whole areas of science that we haven't really been able to do too much with in the
Bridget Scanlon: It must be nice and satisfying to see something that you were involved in a couple of decades ago to come to fruition because most of the stuff we're involved with goes nowhere.
Dennis Lettenmaier: Yeah. Well, it is interesting because, you know, after the 97 - 98 workshops and before the 2007 decadal review, there were some working groups and there was a working group for, you know surface water stuff, and you could fit everybody in that working group around a fairly small conference table, okay.
That, that's what there was in the early days. So to fast forward to some of the more recent. SWOT meetings where you got a whole big room filled with, you know, 200 people and so on and a bunch of people online. It is just a whole different, I mean the interest and and so on is, has exploded and there's a lot of good work that will come out.
It also, I think, somewhat reshapes the whole field. I mean, if you look at other fields which have had kind of global scale, remote sensing data, atmospheric sciences and oceanography in particular, you know, hydrology didn't have very much. And, and now we've got, you know GRACE, which you've been very involved in all kinds of work has come out of GRACE, SMAP, SMOS soil moisture missions and, and now SWOT and, and then some others that we get other things out of the way.
The, the way work in the field is done, certainly on the science, has completely changed.
Bridget Scanlon: I think we're running out of time, Dennis, and we've only touched a little bit of all of the things that you've been doing, but I really appreciate, it's great that you, the backdrop on some a lot of these discussions and how things evolved over time, because when you read the end result, you don't know how
That mostly just comes from having been around a long time.
Bridget Scanlon: But I think you know, the other thing that you and I have in common, both soft money researchers and that keeps you on your toes.
Dennis Lettenmaier: That does change your outlook. Well. You've been longer than I am, but it was 39 years on soft money at Univ. of Washington, so.
Bridget Scanlon: But also, you were very involved with the agencies. You spent a stint with the USGS. you spent a time at NASA and everything. You were critically involved in all of those things, and so you had a good knowledge, of what the agencies were thinking and could feed back into them then to talk about how things should move forward,
Dennis Lettenmaier: Well, it does broaden your thinking and for sort of younger scientists who have that opportunity, you know, I was very happy. I was on something called Intergovernmental Personnel Act loan on those two different occasions, USGS and, NASA, and that played a big part in my career.
Bridget Scanlon: And I know at your retirement party recently, there was an army of people around you. so many of your students and everything so you have a huge following and you've influenced so many people. Thank you so much for doing the podcast. Dennis Lettenmaier is a Distinguished Emeritus professor at the Department of Geography and Civil Engineering at UCLA.
And thanks for all your work Dennis. I really appreciate what you do.
Dennis Lettenmaier: Okay. Well thank you Bridget, for taking the effort to put these together. It's a big effort. Thank you.
