Pacific Gas and Electric, the beleaguered California utility at the center of the wildfire phenomenon, has turned to the Argonne National Laboratory for help. PG&E is looking for small-area weather and climate models that can help it make plans on a regional scale. Argonne’s Chief Scientist and head of its Department of Atmospheric Science and Climate Research, Dr. Rao Kotamarthi joined Federal Drive with Tom Temin to talk about how the arrangement works.
Insight by Red Hat: Learn how organizations are working to meet their missions in real-time by downloading this exclusive ebook.
Tom Temin: Dr. Kotamarthi, good to have you on.
Dr. Rao Kotamarthi: Thank you. Thanks for the invitation to talk to you.
Tom Temin: So I didn’t realize that a private entity like PG&E could come to Argonne and enter into an arrangement for research. Tell us how this works. And is it something that you do regularly?
Dr. Rao Kotamarthi: Yeah, it is fairly common for us to work with private sector partners. These kinds of collaborations range from like access to user facilities such as Advanced Photon Source at Argonne, high performance computing facilities to project like ours. So the laboratory is for public good, and supporting fundamental research into energy and supporting the energy infrastructure. So we do whatever we can to help. We are encouraged to work with private sector partners to share our knowledge and some of the inventions that we make at the laboratory so that it is made available for public and private industry and startups.
Tom Temin: And PG&E is paying for this work?
Dr. Rao Kotamarthi: Yes, it supports effort that my staff spends on it. Yes, it’s paid for that.
Tom Temin: Sure. And it sounds like that you will come up with a piece of intellectual property here, say, a way of directing climate models to smaller areas than simply whole nations or whole states. And will that belong to PG&E? It sounds like something that could be deployed throughout the world, really, for entities that have a interest in localized climate conditions.
Dr. Rao Kotamarthi: Yes, so this particular project is based entirely on open science research. We have developed the tools and datasets used for this research in peer reviewed journals and articles that we have done over the last several years. And hence, there is no IP issue for this particular project. This type of projects go through a screening process to identify any such issues and get cleared by the laboratory and DOE. In case there is an IP issue, things like development or testing of some new technology, there will be agreements in place for dealing with IP, before a contract is set up. But this particular research is, actually we are using the dataset we have developed previously for an application for trying to understand that, and help PG&E to develop better strategies for dealing with it as the climate is changing. We don’t have any IP issues in this project. It’s actually all based on open science.
Tom Temin: And what is it they’re trying to figure out exactly here? What is the essential problem you’re helping themselves?
Dr. Rao Kotamarthi: So we were asked to check six different indicators of wildfire that PG&E uses regularly to find out how these things will change into the future. So they have a protocol when this track is expired indicators. And whenever one of them exceeds or several of them exceed, they go into high alert mode and try to figure out how to combat that. So what they wanted to know is these six indicators that we are using right now, will they be changing into something, are there any projections into the future that we can use to better prepare or infrastructure for these changes? So these kind of changes would be like, how dry will the soil get? Will the wind speeds of direction of being changed in the future, condition in the atmosphere that drive particular types of weather patterns that are conducive to wildfire? And how they may change by mid century? So that’s pretty much, we’re asking is how this typical fire indicators that PG&E uses right now, change as we go into the future.
Tom Temin: And the models that you have then, these are developed from data that comes from the weather types of agencies?
Dr. Rao Kotamarthi: So the model itself is very high resolution in the sense that we can resolve not America “…” kilometers, but grid cells, which is climate model “…” at 100 kilometers. So it’s about 10 times higher resolution. It requires a lot of “…” computing. So the data for that comes from, for the current climate, we use current weather, of course, we do for the last 10 years, we take the weather conditions for the last 10 years to drive these models. When we do in the future, we actually use a climate models, there may be about “…” different climate models around the world. So they pick a few of them providing the conditions outside North America. So we can do a very high resolution simulation within North America. So they take into things like how would the greenhouse gases change in the future? There are different scenarios for that, uncertainties in physics that we don’t understand some other physics or how do we account for that and things like that. It gives you a projection into the future. Essentially, when you’re doing a climate simulation, you’re looking at scenarios of how this will evolve into the future. So it gives you an idea of how the scenarios will evolve. So essentially are projecting into the future. And then of course, when you do some kind of prediction, you want to understand what can fit it in this production system.
Tom Temin: We’re speaking with Dr. Rao Kotamarthi. He’s chief scientist and head of the Department of Atmospheric Science and Climate Research at Argonne National Laboratory. And are you able to develop any specific options for a place like PG&E? That is to say, when they understand what’s going to happen, say it’s going to get drier here or windier there, what do they do about it?
Dr. Rao Kotamarthi: So it’s an interesting problem, right now most of the private industry in the U.S. is trying to figure out how do they account for climate change in their plant? So this is probably the initial phase, people are trying to figure out, is the data itself useful for me to make a decision, right? If the data says that the incidence will change by two or three years, it is 10 now, it could be 11, or 12. And the uncertainty on that is 30%. But how much of a credence should I give to that, and how much money should I be spending? So they’re trying to understand the center projections and uncertainties in them and how it affects their business plans. For example, for PG&E, they are really interested in something called a Diablo wind. I think Diablo is a mountain in California. And the winds come from Northeast. So that is one of the biggest indicators of whenever they see this Diablo wind over some threshold, they do have a good idea that this is gonna lead to wildfires, especially in the North and Central parts of California. And we have been looking at the incident of these fires, of these winds, how the intensity is changing, how the frequency is changing, how the duration of this is changing into the future. The idea being that if you can develop some statistics of how these are changing into the future, maybe there could be some of the dataset that will be helpful in planning. Let’s say, right now, the Diablo winds are mostly around the coastal part of Northern California, maybe they move a little closer to the mountains, what kind of action should we take now so that we can do better planning for the future? So these are the kinds of questions they’re asking. And at this level, at this time, the idea is to just be aware of these things and start planning. I’m not yet sure how the industry will actually implement this into their future activities, how it will affect investment decisions. So this is all part of the idea that we have to adapt to changing climate. Even if you do mitigation, there is this big need for adapting to changing climate. How do we go about doing that? This is a challenge, and also a need. And we’re trying different things. And this is like maybe getting closer to generating the kind of data maybe the industry can actually start using in their spreadsheets and stuff like that. So that they can look at it in their, in terms of cost, right? It’s a whole idea is that essentially at some point, you have to figure out how much it costs and how much action they take can take.
Tom Temin: Yeah, my question then is, what do they do over time? Once you are done with the research, and you come up with a refreshed model, say for them to use, this has to be deployed, I would think in the models run on an ongoing basis. And so will they be able to have the ability to take what they get from Argonne, and perhaps use it themselves to keep an ongoing prediction of what they need to know?
Want to stay up to date with the latest federal news and information from all your devices? Download the revamped Federal News Network app
Dr. Rao Kotamarthi: Yeah, I think the whole idea is that they mostly look at current weather. So if you can build this trends of how this will change in the future, they can build it into their models so that they are looking at, let’s say 20 years from they’re putting some of the wires underground or something like that, where do they should prioritize, maybe that will help them decide those things. So these are the decisions that will be made by industry, and they are trying to figure out those pathways. And I think the kind of data we are developing will help them do that.
Tom Temin: And do you have a separate set of data and algorithms and models and so forth for not so much wildfire, but for say flooding, which might affect utilities elsewhere.
Dr. Rao Kotamarthi: So what we have done at Argonne, and several other people have done around the world too so I don’t want to take all the credit for this, but what we have done is that we have developed a really high spatial resolution climate data for North America. It’s a process called downscaling. And we did a few years ago. To do this, so we have simulations for the current decade, like let’s say recent past, mid century and end of the century, we do different greenhouse gas emission scenarios, like I said before. So this in total is about 300 individual “…” and about five petabyte of data for performing analysis. So there’s a large dataset from that we can extract, almost all kinds of meteorological variables, I mean precipitation, and other things, or temperatures from which we can calculate things like a fire index. So the fire index is one derived product from the climate model simulations we did, we do additional calculations to do that. So similarly, we have to do flooding, we have done flooding too at a really high resolution. So because we have precipitation and other things at every three hours, we run a separate flood model, at very high resolutions for calculating flooding intensities, for example, both coastal and inland flooding. So that will be helpful for infrastructure that is affected by flood. So this particular, PG&E for example, is very interested in fires, obviously. So the whole idea of a fire is that you develop an index, and the index tells you is the number is very high, you can have the potential for fires, where you see sometimes in the West, “…” fire potential index is high, it’s yellow, red and things like that. Essentially “…” end up indices. So you want to calculate those indices for current and future climate and see how they are varying, which parts of North America, for example, may become more fire prone in the future and things like that. So, risk analysis in some sense, once you’re done all these calculations, you’re doing risk of flooding fire and things like that. And based on that, then industry or government can take action on considering the risk and uncertainty calculating the risk.
Tom Temin: Yeah, so the “…” we have now is that the power stays on at Argonne, so that you can help everybody else.
Dr. Rao Kotamarthi: [Laughing], and the computer keeps running.
Tom Temin: Alright.
Dr. Rao Kotamarthi: They’ll buy a bigger computer. That’s all.
Tom Temin: Dr. Rao Kotamarthi is chief scientist and head of the Department of Atmospheric Science and Climate Research at Argonne National Laboratory. Thanks so much for joining me.
Dr. Rao Kotamarthi: Thank you. You’re welcome.