Return to

Newsletter

Sign up and stay in-the-know about The Crowd & The Cloud and the world of citizen science.
* indicates required

#CrowdCloudLIVE After each episode's WORLD premiere in April, show host, producer, and people seen on the show participated in post-premiere roundtable discussions. Viewers like you listened in, asked questions, and were able to dive deeper into the power of Citizen Science.

Watch the recorded Facebook Live events now. Discover more about how Citizen Science is revolutionizing the ways we gather, analyze, and utilize the data that fuels scientific research, discovery, and community action.

Find your local station and showtimes here, or go to:
Aptonline.org

CoCoRaHS: Volunteers vs Extreme Weather
Real-time Weather Reporting
Q&A with Chad Gimmestad

My name is Chad Gimmestad and I work for the National Weather Service as a Senior Forecaster.

​Tell​ ​us​ ​a​ ​little​ ​bit​ ​about​ ​how you got involved in CoCoRaHS?

I was actually working the night of the flood in Fort Collins, in 1997. I got there the next morning, looking at the damage, so I have that personal connection to how the project got started. I was interested right away when Nolan Doesken started doing this, and became an observer shortly afterward. I'd asked Nolan, there's already a couple other people close to me who are doing it. He said, "Yeah, we need all the data we can get," and it turns out the observers kind of come and go. But there's always somebody from our neighborhood that's reporting. We got interested pretty quickly. We were getting these hail reports from people, and what do we do with them? How can we make them useful in real-time, rather than just being in some climatologist’s drawer somewhere? And so we figured out a way to get that information in the hands of forecasters quickly.

The NOAA building where Chad works.

How did the flood of 2013 differ from the flood of 1997? What difference did CoCoRaHS make?

For one thing, the 1997 storm was a lot smaller, really localized. In some ways, that's harder to deal with because you're less likely to have somebody under there who's observing it, whereas the 2013 flood was a really large-scale event. But there was a difference in the information that we had. In the Fort Collins flood in 1997, we had some professional weather observers that we were talking to on the east side of Fort Collins who had a few inches of rain. And they were telling us, "Yeah, we're getting heavy rain, but it's two or three inches in a couple of hours, so it's causing some localized street flooding." Meanwhile, on the other side of town, there's this huge deluge coming in, where there's 10 or 12 inches of rain in that same amount of time.

(In 2013) our awareness increased almost as quickly as the flooding itself increased, instead of lagging behind. The whole goal is to shrink that lead time so that we're aware of what's happening as quickly as possible. We got great documentation after the fact, too, people who were going back and studying what happened and trying to improve river models and engineering questions about how big should the floodplain be and how big should our bridges be. They're going back in and finding 50 or 100 CoCoRaHS reports that really help with that.

What the early warning does is it gets the emergency responders moving faster. In the Fort Collins flood, the fire station was right across the street. We didn't know about the situation there, where the people were most threatened, but the emergency responders did. They knew what the problem was going to be, when we told them there's going to be a flood on this creek. Here in Boulder, during the 2013 flood, we worked with the emergency managers to help them understand how serious it was. That got them moving faster. Using the radar data and the CoCoRaHS data, we were able to help them pinpoint, "Here's where it's the worst, and they can go evacuate those areas first." Helping them get to the right place and get there sooner is what saves lives. In those cases, sometimes a few minutes does matter.

How did you go from being there at ground zero, to creating something that the National Weather Service incorporates in its assessments, models, warning systems today?

We're always looking for every piece of data we can get, especially, when we are making important decisions about people's lives. Nothing replaces somebody on the ground looking at something, so we were very hungry for this information. We figured out a way to get the computers to talk to each other, for CoCoRaHS to forward reports to the computer system in Boulder (at NOAA’s National Weather Service offices.) I run a program that would quickly turn around and send that report to whatever Weather Service office needed to get that information. At the time CoCoRaHS was just in northern Colorado. As it grew, we expanded the list. As they grew across the country, we decided to include the whole weather system in the process.

Chad shows the system to Waleed.

Can you walk me through, from measurement to outcome, describing the steps involved?

Let's say, somebody out in Fort Morgan, Colorado, puts in a hail report on the CoCoRaHS website. They're looking out the window, they are hearing the hail on their roofs, they run out right after the storm is over, they measure, and they've got one and quarter inch diameter hail. They enter it on the CoCoRaHS website, and that report's going to come to me about one minute later. I'm going to be sitting here, monitoring storms, looking at radar data, and we're constantly evaluating, do I need to issue a warning for this storm or not? If we think the storm is going to become severe, we want to get a warning out as quickly as possible.

I'm sitting here, trying to make a decision, and the radar data is not always conclusive. We are making the best estimate that we can, with the data that we have. A lot of times, there are borderline cases where the same radar picture could be half inch diameter hail, it could be one and a half inch diameter hail depending on the environment, and we don't know. Now an alarm goes off. I click on this alarm, and I see the report; this person is getting 1.25 inch diameter hail, and they are getting minor damage from the hail. If I didn’t have a warning out at this point I’d need to issue a warning for that storm. I wasn't sure whether it was producing large hail, now I know that it is, so I want to issue a warning as quickly as I can.

How does data generated from CoCoRaHS volunteers complement your radar data?

What I'm measuring with the remote sensing isn't actually what I want to know, because what I really want to know is, how big is the hail? How many hailstones are there, what impact is it having? Or rain or flooding, and so on. What the radar's telling me is there's this much energy return from the radar, and we go out a long ways on trying to figure out what that means, but there is nothing like having somebody there with eyeballs and a ruler to actually tell me. The human can tell me exactly what I want to know, and we're using a lot of computer power and other fancy science to try to estimate that based on what this radar is hearing from 100 miles away.

CoCoRaHS data can feed back into computer models, when we are talking about rainfall amounts. Our river forecast centers are able to bring in CoCoRaHS rainfall data into their hydrologic model. They're starting with radar data, which is a rough estimate of how much rain fell, but then they add rain gauge reports, which are fairly infrequent. The official rain gauges tend to be one or two per county, which gives ground to the radar estimate. Say the rain gauges on the ground are saying it rained half as much as the radar estimate. They'll adjust that whole radar estimate down 50 percent to match with the rain gauges said.

Well, the CoCoRaHS program is giving them 10 or 100 times as many of those rain gauges on the ground, (so) they're going to have a much better estimate. They can put in real data at all of those points, instead of just going with the radar estimate. It's kind of an honesty test for the radar data. It's turning the radar data from an estimate into something that's much more precise.

The CoCoRaHS data has also been very useful for climate studies, because again, we've got an official network of stations that's fairly sparse. For large-scale applications, it's good, but if you want to go in and study, what is the snowfall pattern in Boulder County, Colorado, where you have plains, and mountains, and all of these complicated patterns, then, you need much more detailed data.The CoCoRaHS data lets you come in and see there's 10 years worth of this really detailed data. That's enough for climatologists to see the finer patterns that we've been missing before.

What's different about measurements of rain, snow, and hail? (C&C: remember, CoCoRaHS stands for the Community Collaborative Rain, Hail and Snow network)

For the rainfall, it's how much rain was in their gauge. That's important to me because if it's a lot it's going to cause flooding. The greatest value of that, day in and day out, is to climatologists, people who are studying water supply, soil moisture, the river forecasters. The needs vary a lot from one part of the country to another. But there's a ton of users for that information everywhere. Drought monitoring, for instance, a big application that maybe wasn't thought of to begin with, and in that case, getting a quarter to a half inch rain makes a big difference, versus nothing. And so, the drought monitoring has gone from a state scale down to a fraction of the county scale because of CoCoRaHS. Because now, we have consistent, daily data from a half dozen points in the county.

People don't realize that snowfall varies a lot. We tend to stay home when it snows and we see what happens at our house. A lot of weather systems, there are small bands of showers that make it snow twice as much in one place than it does 5 or 10 miles away. Of course, when you get into situations where there's mountains or coastlines, even little hills, it can have big effects on where the heavier snow is versus the lighter snow. So, there's a ton of variability. We get data from a few points, but what's happening in between? We'll have a 10 mile wide band of heavy snow on our radar that's between our official stations. Chances are there's a couple of CoCoRaHS observers in there. They'll let us know what happened. Most places got 2 inches of snow, but this little spot got 10 or 12 inches of snow. We can pass that on and let people know.
It's taking me from a place of, as a weather forecaster, where instead of making a scientific estimate, I've actually got real data.

Hail measurement is really interesting because it's not a widespread thing like rain and snow, and not everybody has it all the time. There isn't a lot of data on hail in the United States, because it's overall a relatively rare thing. From my perspective as a National Weather Service person, my main concern is getting the reports quickly, so I can issue warnings. Hail does cause a lot of damage. There are some things people can do, like moving your car into a garage, running out and covering the garden. If I give you 15 or 30 minutes notice that a hail storm is coming, you may prevent some damage. Certainly, you can get inside and get to a safe place, which is our number one priority, because big hail can injure people and occasionally kill people.

So we're protecting people, and to some extent, protecting property with the hail warnings. But there's also an interesting thing with CoCoRaHS. It's really the first time we've had, over a large area, this hail data being routinely reported. If you think of thousands of people who are reporting every time it hails, all of a sudden you've got great information on how often it hails and how often it hails with large hail sizes, all across the country, and that hasn't existed before.

What do you hope for the future of CoCoRaHS?

One big thing is we'd have much better hydrologic modeling. If we go 10 or 15 years into the future, our computer researchers are getting to that point, but being able to integrate all this stuff into modeling a flash flood would be tremendously helpful. It would keep me from issuing warnings when something's not going to happen, and it would keep me from missing things that I miss now, being able to feed that real time information in. We're going to be doing it with the radar data, but oftentimes, it's off by quite a bit. As we move into the world where we're doing that kind of modeling, to have ground truth to fix that would be very, very valuable.

Then, having a better understanding of what happens for all the different kinds of science that are going on. If you look at how agriculture has modernized in the last decade, even, with computer technology and remote sensing. Now, you're putting sensors in cows and in the ground, and you got people flying drones over fields sensing soil moisture. This data feeds right into that, of being able to specify the rainfall and snowfall. It's going to have a modest economic value all over the place.

We're also moving into a world where I'm carrying around a phone that's giving me real-time traffic information from other people who are carrying their phone around. The weather world's getting to be like that too, where we're all helping each out more, we're all sharing way more information than we used to. (CROWD & CLOUD: see the “Cowboy CoCoRaHS” video to see how rancher Skyler Flake uses archived precipitation data to help manage his herds of cattle. He’s not just giving data to the CoCoRaHS network, but getting valuable information back.)

One of the things I love about my job is that we are that bridge, that we are connected to the people. I'm getting the data from people, and I'm also helping people every day. I love being in that place.

randomness