In this week’s episode, host Daniel Raimi talks with Michael Craig, an assistant professor at the University of Michigan who studies energy systems. Craig and Raimi discuss a recent study coauthored by Craig that explores how energy models can better incorporate variations in weather and climate and why an exchange of data between energy and climate modelers is crucial to helping keep the lights on. Craig outlines a research agenda that describes near-term and long-term steps to bridge the divide between energy and climate models; he also shares advice for interdisciplinary collaboration.
Listen to the Podcast
Notable Quotes
- The difficulties of incorporating climate data in energy models: “Typically, we also want very high spatial and temporal resolution in our data and energy-system models. I know where my wind plant is. I want weather data for that particular wind plant—not for a 100-by-100-kilometer grid cell that happens to include that wind plant and a bunch of others around.” (8:41)
- Informing decisions with improved models: “The idea that we’re trying to get at here is, let’s think about how the weather is changing. Let’s think of all the variability that introduces. And if we have that spectrum of variability, hopefully we can make better decisions than we can if we have a limited perspective on what weather can achieve.” (12:44)
- Account for the future consequences of climate change in today’s decarbonization efforts: “As we make these investments, we want to be thinking not just about how to mitigate climate change—we also want to be thinking about how these future investments fare under a changing climate. Are the power plants that I’m building going to be as useful in the future as I think they are now?” (14:22)
Top of the Stack
- “Overcoming the disconnect between energy system and climate modeling” by Michael T. Craig, Jan Wohland, Lauren P. Steep, Alexander Kies, Bryn Pickering, Hannah C. Bloomfield, Jethro Browell, Matteo De Felice, Chris J. Dent, Adrien Deroubaix, Felix Frischmuth, Paula L. M. Gonzalez, Aleksander Grochowicz, Katharina Gruber, Philipp Härtel, Martin Kittel, Leander Kotzur, Inga Labuhn, Julie K. Lundquist, Noah Pflugradt, Karin van der Wiel, Marianne Zeyringer, and David J. Brayshaw
- NextGenEC at the University of Reading
- Downscaling Techniques for High-Resolution Climate Projections: From Global Change to Local Impacts by Rao Kotamarthi, Katharine Hayhoe, Linda O. Mearns, Donald Wuebbles, Jennifer Jacobs, and Jennifer Jurado
- The Making of the Atomic Bomb by Richard Rhodes
The Full Transcript
Daniel Raimi: Hello, and welcome to Resources Radio, a weekly podcast from Resources for the Future. I'm your host, Daniel Raimi.
Today, we talk with Michael Craig, assistant professor in energy systems at the University of Michigan's School for Environment and Sustainability. Michael is an expert on building and using models of our energy system to inform policymaking and recently published a study with a large number of coauthors that addresses a really important topic: How can energy modelers better incorporate variations in weather and climate, particularly as climate change leads to more extreme weather events?
Michael will help us understand why answering this question is crucial to helping us keep the lights on and outline a research agenda that describes near-term and long-term steps to bridge the divide between energy and climate models. This episode will be of interest to all of our audience, but especially to those of you who work on developing, using, and interpreting energy- and climate-system models. Stay with us.
Michael Craig from the University of Michigan, welcome to Resources Radio.
Michael Craig: Thanks. It's great to be here.
Daniel Raimi: It's great to have you here. And here, I should just note for listeners that we are recording this episode in person in my house in Ann Arbor, Michigan, where we both live, and it's nice to be having a conversation in person. It's been over two years since we've done one of these podcasts in person. So, it's a pleasure to be with you in the flesh.
Michael Craig: It's an honor to be first in two years.
Daniel Raimi: So Michael, we ask all of our guests on the show how they got interested in energy and environmental issues, either as a kid or later in life. So, have you always been interested in this stuff?
Michael Craig: Yeah, I've always been interested in the environment, well, at least as far as I can remember. My mom tells me that when I was one year old, I lived in San Diego and she'd take me to the Safari Park there every day as a kid. And so, I assume I just got attached to animals because I saw them as much as I did humans, basically, while I was growing up. And I really loved animals, the environment. I went to school wanting to work on them and I majored in environmental studies. So I was actually an ecology undergrad studying how animals behave, how they eat seeds in landscape-scale settings.
And then when I went to graduate, I thought, "How can I have a positive impact on the environment? Achieve the ends that I want to?" And I saw energy as a better means to an end for me and my skillset than working in ecology. And so, I got to work at a nonprofit in DC, Oceana, on ocean issues, energy issues, and that kind of set me down on the path to energy. And then, I had the great fortune of working with some great mentors through school, and that's where I really got hooked on energy, and that's where I've been working since.
Daniel Raimi: That's great. And for those listeners who have not been to the Safari Park near San Diego, it's really an extraordinary place.
Michael Craig: It is great. My parents brought me back when I was 24 years old, I think, and I think I probably loved it as much when I was 24 years old as I did when I was two years old.
Daniel Raimi: Yeah, I've only been once and I think I was around 24 years old, and it was fantastic.
Okay. So let's talk now about the work you do and a new paper that you've got out with a large number of colleagues. We're going to talk about that paper and how it focuses on the connection between the tools that model the energy system and the tools that are used to model the climate system and the future climate system. Before we get into details, can you give us kind of a real quick primer on what we mean when we use the terms "energy models" and "climate models" in today's conversation?
Michael Craig: Absolutely. So, you're right. This is a product, actually, of a very large initiative that's run by David Brayshaw at the University of Reading. It's called the NextGenEC Forum. And so, there's a bunch of different coauthors on this from the energy-system community and the climate-system community.
So, both these models are mathematical models, computational models, people writing code into a computer and running the programs. The energy-system models, the keyword there is on the system; what we're trying to do is not model individual power plants or individual pipelines, we're trying to model systems of energy.
And so you can think of this, for instance, in the context of a natural gas system, we might be modeling pipelines, compressors, producers of natural gas off-takers. And so, that we can describe via an energy-system model, or more relevant to this paper, I'd say, are power system-related models where we are not just, again, focusing on individual generators, but we have many different generators, transmission lines connecting large-scale generation or distributed generation to end-users or consumers of electricity.
Daniel Raimi: Great.
Michael Craig: So that's an energy-system model and we can use these for long-term planning, multiple-year short-term operations. And these are models that are used in many papers every year, but also utilities in the real world plan and operate power systems using these types of energy-system models that we're dealing with. And so, there's importance in this paper to researchers as well as to people who are actually planning and running these things in the real world.
Climate models also come in different flavors, but they are generally trying to represent and forecast climate and weather. That can be at a global scale. GCMs (global climate models or general circulation models) are looking across the earth. And so, when you see an Intergovernmental Panel on Climate Change headline that says, "The earth is on track to warm by X degree celsius by X year," somebody ran one or more likely many different types of global climate models to get some global average temperature change. And so, that's a global climate model, again, that's trying to represent climate and weather.
And then, we have some other types of climate models. Regional climate models, for instance (RCMs), that work on regional scales, maybe the southeastern United States or over an entire country, again, trying to do the same thing, to understand climate and weather processes to try to predict what it will look like in the future. Just the smaller spatial scale you have, the more spatial and temporal resolution you can get.
Daniel Raimi: Yeah, that makes sense. And so, when we're talking today about these models, we're referring to the whole swath of them and principles that can apply to any type of energy or climate model.
Michael Craig: Exactly.
Daniel Raimi: Great. So the paper itself, and we'll have a link to it in the show notes, is called “Overcoming the Disconnect between Energy System and Climate Modeling.” So that begs the question, What are some of the biggest disconnects that you and your authors are trying to identify and overcome?
Michael Craig: Right. So, the big one is we want to capture future climate and weather inside of our energy-system models. And that is very hard to do and, almost always, it is not being done. Almost always, we are running energy-system models, whether it's in reality or in the research setting, using historic data. And so the main disconnect, again, we're trying to do is how do we get this future data that we know is more representative of the future than historic data into our energy-system modeling?
In the paper, we try to explain where the disconnects come from by walking through what an illustrative energy-system modeler does generally, what they're trying to do, and what an illustrative climate modeler generally does, what they're trying to do to explain not just how these things are not aligning, but why are they not aligning. And so through that process, we go from this big, overarching disconnect, which is, we want this data in our energy-system model. It's very hard to do. We break that down into a few smaller disconnects that all build up into having this disconnect happen.
So one of them is when we run energy-system models, we typically want a single annual time series. And that is true for most energy-system models. In research, and oftentimes in the real world as well, we're running energy-system models with one annual time series. So, a time series is just hourly or minute-to-minute data that lasts a whole year. You can think of in power systems, wind power is a function of wind speed. So, there's meteorological data. Solar power is a function of solar radiance demand. Electricity demand varies with temperature. And so, there is some meteorological relationship there.
And so when we run this with a single annual time series, we're missing a bunch of variability. Just totally forget climate change, there's multidecadal variability in the climate system. Anybody who has lived on the West Coast, for instance, is very familiar with El Niño, La Niña cycles. So, there's variability there. You have that interdecadal variability. Now, you layer on top of it non-stationary or changing conditions imposed by climate change. So, just using one annual time series is not enough to capture this variability. And in fact, climate modelers will never just give you one single annual time series because to them, that is totally anathema to their understanding of how the climate system works.
So we have energy-system modelers who want one annual time series; climate modelers typically give you 30 years. So, I'm an energy-system modeler. You give me 30 years of data—wow, that's a big change from what I expect to get, and so that can be challenging for all sorts of reasons. Typically, we also want very high spatial and temporal resolution in our data and energy-system models. I know where my wind plant is. I want weather data for that particular wind plant, not for a 100-by-100-kilometer grid cell that happens to include that wind plant and a bunch of others around.
And so, there's a mismatch there. And temporal, I mean, we worry about second-to-second, minute-to-minute, hour-to-hour variations in the energy system. Sometimes from climate models, you get 24-hourly data. So, they'll give you a single wind speed for an entire day. So, there's another disconnect. We want certain resolution or spatial and temporal resolution? Generally doesn't come from climate modelers. And there's good reasons why it doesn't. It's not like they look at us and say, "Forget you. We're not going to give it to you." Their models were built for very different things than what energy-system modelers were built for. And so, we have this disconnect there.
And then, there are two other small ones about how we want our data. We want in the energy system for a given hour, what are my wind speeds in my region, but what is my solar radiance and what is my temperature, so I can understand, as demand is increasing, what are solar and wind doing? Is solar dropping off when wind is dropping off, and now I have a problem of balancing supply and demand? Or is wind coming up with demand, and so I'm okay? And so we want to capture some covariability, how all these three things move together.
And generally, it's so hard to get that from a climate model because, in part, they don't give it to you on an hourly basis, but also because they're not really looking at and validating their output from their climate models on solar radiance and wind speed and temperature on an hourly basis—because who else besides from energy-system modelers want hourly wind speeds? Water planners typically don't want it. The agricultural community generally doesn't need that so much as they need precipitation and temperature. And so, there are new demands, I think, that are being placed on climate data from the energy-system community. And that means that we end up in the situation again where we want certain things that aren’t natively coming out of climate models, and so when we want to bring in future climate data into our energy-system models, it's very difficult.
Daniel Raimi: Yeah. That's great, and that all makes total sense. And it's really helpful for you to point out those three key pieces for the energy-system models, especially in the power sector here. We're talking about wind data, solar-radiation data, and then temperature data to help you get at those key energy-system elements.
So, you've given us a great overview, introduction, to why this stuff matters and helping us understand the disconnects. Can you give us an example or two of—I want to say “real world” in air quotes, because we're talking about models here—but can you give us some examples either of historical events or events that could happen in the future where it would be particularly useful to have better integration between these two types of models?
Michael Craig: Sure. I'll let the suggestion that my models don't perfectly capture reality just slide. That's fine. Right. So one example would be, in California, the summer of 2020, they had an extreme heat storm. They called it a heat storm, and during that extreme heat storm, there were some outages happening across California. Rolling outages. People wanted electricity and they couldn't get it.
When California came after the fact and wrote a report asking, "How did this happen to us? How did we have these outages?"—which is the last thing an energy-system modeler or an energy-system planner wants to have, right?—they found that the extreme heat storm was partly exacerbated by climate change. So, we have some thumbprint of climate change on this event. Maybe it would've happened without climate change, but not as severe, perhaps. And because of that extreme heat storm that was at least worsened by climate change, they had several things happen in their power system that led to those rolling outages.
So this, to me, is one example where if we had a broader range of variability captured in the weather inputs that we were giving to our energy-system models, maybe we could have detected that sort of event and tried to plan to be able to withstand that sort of event. Now, it's always easy in retrospect to say, "Aha, this extreme event happened. You didn't do it well enough. You could have found it if you had this larger variability." I mean, people have talked about that with Texas in the winter from two years ago, as well. But the idea that we're trying to get at here is, let's think about how the weather is changing. Let's think of all the variability that introduces. And if we have that spectrum of variability, hopefully we can make better decisions than we can if we have a limited perspective on what weather can achieve.
And so, that's one event where—I'm not sure if it could have or not—but that is a type of event that we're talking about: something that a system operator or planner didn't exactly foresee led to these consequences like rolling outages.
Daniel Raimi: Is one way of thinking about this as a stress test? You're stress testing the models under different ranges of extremes?
Michael Craig: Yeah, exactly. We will talk later, I think, about the solutions that we have or the reforms that we think each community needs to adopt separately and then together. And one of them absolutely is stress testing your system to different future meteorology or different future weather. And you need to know what that future weather is going to be in order to do the stress testing appropriately.
Daniel Raimi: Great. Is there another example that comes to mind that you wanted to share?
Michael Craig: Yeah. I think just in general, the other example I'd give is that we are really changing how we consume and generate electricity across the United States and in many different parts of the world. We are trying to rapidly decarbonize to mitigate climate change. Those decarbonization activities, even if they go exceedingly well, past our wildest expectations, we still already have climate change happening. It will still get worse before it gets better.
So, as we're making these decarbonization decisions, we're building massive amounts of wind plants and solar plants. We're electrifying homes. We're making plans around all these different changes. We're still trying to make sure that we can balance supply and demand in the future as we make these investments.
And so, as we make these investments, we want to be thinking not just about how to mitigate climate change—we also want to be thinking about how these future investments fare under a changing climate. Are the power plants that I'm building going to be as useful in the future as I think they are now? And power plants will last 10, 20, 30, 40 years. Transmission lines will last even longer than that. And so, for instance, if I'm building massive amounts of transmission from the Midwest to the East Coast to bring a bunch of renewables, wind in particular, from the Midwest to the East Coast, which is what many decarbonization plans want to do, well, I have some expectation about the value of those transmission lines, what their capacity will be on a day-to-day and hour-to-hour basis. And if temperatures are getting warmer, that carrying capacity will generally be less, for instance. And so, thinking about our decarbonization decisions while accounting for changes in meteorology is another example I'd give that utilities across the nation are doing and, I think, are increasingly working on.
Daniel Raimi: Yeah, that makes sense. It's really interesting, especially in the context of transmission. I don't work with electricity models the way you do, day in and day out, and I sometimes forget about how transmission constraints can arise under different temperature conditions.
Michael Craig: Right.
Daniel Raimi: That's a great point.
Michael Craig: And out west as well under wildfire season now, that is a whole new type of … well, maybe not a whole new type, but that coupling has become much stronger in recent years. And so, transmission of clean energy into California, say, from out of state is being threatened by increasing wildfires. Those wildfires in turn are being driven by climate change. And so, many different types of interaction here between what adaptation investments I need to make and what decarbonization investments I need to make.
Daniel Raimi: Yeah. And sometimes, wildfires are getting started by the transmission lines, but that's a whole other conversation.
Michael Craig: Exactly.
Daniel Raimi: So, another important part of the paper that you and your colleagues go into is about a research agenda. And because we have a lot of researchers who listen to our show, I thought it would be great if you could talk about that a little bit.
You've outlined a research agenda to overcome some of these disconnects. Can you give us some basics on what you and your colleagues think some of the near-term steps are and some of the longer-term steps might be to improve the kind of communication between these different types of models?
Michael Craig: Absolutely. So, we have two different sets of recommendations: one for the near-term things that we can start working on now and people already are starting to work on now, and another set for the long term. And then all these recommendations, near term and long term, is all geared toward getting this climate data into our energy-system models. In the near term, we want to take what's out there already, even if it's imperfect, and try to bring it into our energy-system models. In the long term, we're trying to bend the climate data to be better suited to energy-system models and in that way, have better data to incorporate into our energy-system models.
So, let's talk about the near term. We have two different recommendations here: one for the energy community, one for the climate community. For the climate community, our recommendation is to align those outputs from climate models, so that they can be useful to energy-system models. Oftentimes, climate models are being run. They're generating variables—things like wind speeds and solar radiance, for instance—and they're not really saving them, or if they're saving them, they're not making them publicly available, because there's no clear end user that has been defined for those variables so far.
So for instance, in conversations with climate scientists, sometimes you'll email them and say, "I really want to use the data from this GCM, global climate model, regional climate model, but I can't find this variable at this resolution. Do you have it?" And they'll say, "Absolutely, I have it. It's not on the public data portals, because that would cost me a lot of money to store and maintain, but I'm happy to send it to you." So, having that more publicly available would really just help uptake when I want to bring it into my energy-system model.
So, the first set of recommendations is for the climate-modeling community, aligning those outputs with energy-system modeling needs. And we have a specific list in our paper, for any climate modeler that's listening, where we try to itemize, "This is the resolution that we're looking for. These are the specific meteorological variables that we're looking for." And we also explain why we care about them.
The second one is for the energy-system modeling community and I, in case it has not been obvious so far, come from the energy-system modeling community. And so, these were very near and dear to me as well. And for the energy-system modeling, we are trying to make them better at handling uncertainty in weather. Like I said, we typically run with a single annual time series. Absent climate change, that is not a great idea to begin with, because you have this interdecadal variability. So we can get a lot of benefit just from making energy-system models better running on long-term time series of weather, not just one-year time series of weather. There's a lot of reasons why we don't do that. These models are very large. They're very hard to solve. You can't just throw more things into them and crank it, right? It's not going to solve that way.
And so, this has a lot of challenges we need to innovate in terms of how we run them, how we think about them. But the more we can handle uncertainty in weather, the better we're going to be able to handle long-term variability, absent climate change, and future variability. And the reality is if a climate modeler came to us tomorrow and said, "Here is everything you asked for. I have magically produced it overnight," we'd still have a really hard time incorporating it to our current type of energy-system model just because, again, it was not built for this 30-year data with multiple ensemble members and all sorts of uncertainty in it. So, that's our other near-term recommendation: get better at handling uncertainty in energy-system models.
Daniel Raimi: I know you want to speak to the long-term recommendations, too, but before that, I'm just curious, as you were speaking, whether the sort of improved computing power that we're seeing … What's Moore's Law? Double the computing power every two years or something like that. Is that helping you to be able to incorporate these additional data? Or is it more an issue of labor power and person-hours that go into rewriting the code and stuff like that?
Michael Craig: Yeah. Labor hours are of course constraints, but ultimately it is still the computational constraint. And despite all the progress we've had in processing power … I started running these sorts of energy system models in my master's. That was, wow, three plus two plus three plus two—what is that? 10 years ago. And 10 years ago, I was running a power-system model at very roughly the level of complexity that I am now. I mean, we have better computational resources. Supercomputers help, parallelization helps. So, you can run more now than you could 10 years ago, let alone earlier than that. But the problems are just way too large in order to run it across a 30-year time horizon, especially if I have 30 years of hourly data—but now I've got multiple 30-year sets because of all the uncertainty. So, it's just not something that you can throw it all into the model.
Daniel Raimi: Okay, great. So yeah, tell us more about your long-term research agenda.
Michael Craig: In the long term, what we're trying to do is not just work within each of these disciplines, but instead have them come together to try to think about a new approach. Not just energy-system modeling, not just the climate modeling, but together as a whole. And so, this is our transdisciplinary area approach, where if I am a climate modeler trying to think about how I can better help energy-system modelers, one way would be to ask them what the outputs in it are and just give them those outputs. That would be something like an interdisciplinary approach. Transdisciplinary would be, "Let me think about how I validate my climate model. Let me think about how I assess the outputs that are coming from it." And so, what we're trying to do here is the same end, again—get these climate-system outputs that are better suited to energy-system models, but we want to do it in a new way and a longer-term way.
And the two strategies that we have for that are, have some energy-system modelers go into the climate-modeling community and inform them about what we need. And let's do that assessment together, the assessment of the climate-model outputs. And then once we have those climate-model outputs, have the climate modelers come into the energy-system community and help us better understand how we can leverage those data sets. And so, that is a long-term vision. We think this will require much more time than just within a community recommendation that we have, but ultimately, what we'll get out of this is more progress towards that end that we want, having future weather data enter the energy models in a sophisticated and effective way.
Daniel Raimi: I think that our listeners will have a sense of this before I ask you the question based on our conversation over the last 20 minutes or so, but we do have a lot of researchers who listen to the podcast, as I said, and I'm sure some of them are interested in working on these topics. I know some of them are. So I'm just wondering, What types of expertise or tools or skills are you looking for to complement this work? Are you still looking for folks who have certain types of capabilities to work with you and the climate modeling community to try to do this work? And if so, what might they be and how do they get in touch with you?
Michael Craig: Yeah, absolutely. It's hard to think of a discipline where there is not some value being added here. And so, I think if anybody is listening to this and they think, "Wow, I think I could add to this," I think there almost certainly would be room for them. And so, we'd love to hear from you. You have my email from my school webpage. We have this NextGenEC community that's run out of the University of Reading, David Brayshaw. We have an annual workshop coming up in two months (it's virtual), and we have a series of presentations, a couple tutorials. So, even if you're new to this field and you want to think about how you can incorporate this into your research, we've got tutorials to train you on how to work on these topics.
So, some specific disciplines that come to mind. Of course, energy-system modelers, climate modelers, whether you're working on downscaling techniques in the climate community, whether you're working on global climate models, whether you have some experience before guiding agricultural decisions or water-resource decisions, all those disciplines, of course, would be wonderful to have here.
And our author list from this paper, in fact, includes people from all these different areas. We have a lot of data computation challenges within energy-system models, within climate models, within getting those high-resolution outputs from climate models that we care about. And so, people who are trained in computer science in general, machine learning, statistics, mathematics, all those, there are a lot of hard computational challenges that we need to overcome here.
Of course, economists thinking about the trade-offs that we have once we get our energy-system models running, that will really help us make our decisions. And I'd say, the risk-analysis community as well is a community I've been trying to learn from more and more recently, who I've been dealing with, for instance, thinking about decisionmaking under deep uncertainty. I will never know exactly what the future is going to look like. I don't even know, maybe, what next year is going to look like with 100 percent precision. Given that, how do I make decisions that I think will protect me despite the range of outcomes that I might see? And so, the risk-analysis community, I think, would be wonderful to have join us in this endeavor.
Daniel Raimi: That's great. Yeah, and especially, I mean, in the last, what? Two or three years, especially in the last several months, we've all been reminded about the importance of planning for uncertain futures and having resilience against those uncertain futures.
Michael Craig: Exactly. I think the Texas event really underscores that point where the Texas outages during the winter of 2021 were surprising to many people, and it was surprising to Texas, in part because that system planner had planned for a worst-case scenario that was based on 2012, 2013, somewhere around there. So, looking back 10 years, for instance. Since that event has happened, there's been two or three papers that have come out that said, "Well, if you just took the historic record and you look back 40 years, here are some even worse events that you could have seen in the past." Even absent climate change, you look backwards, you capture that long-term variability, there's an immense amount of value in that in energy-system models. And we have a lot of work to do in our own energy-system modeling community at getting better in incorporating that historic uncertainty.
Daniel Raimi: Yeah. So last question, Michael, before we go to our top of the stack segment, which is related to this issue of interdisciplinarity and transdisciplinarity. Anyone who's ever worked across disciplines knows that it can be incredibly rewarding, and sometimes incredibly challenging, because people have different training, different backgrounds, different cultures, different communities. So, what have you learned so far about building bridges between different disciplines, and what are some of the challenges that you've encountered as you've tried to do so?
Michael Craig: Great. So, I think you've already hit a couple. Different backgrounds and different incentive structures. So, understanding why somebody thinks the way they do, how they were trained, where they're coming from, and also what they need to do to succeed professionally, I think can help you understand, here is a shared pathway forward where both sides benefit. I think language is another very important one. I have had meetings where we use the same exact word, but we mean radically different things. And so, making sure that you have a common syntax and understanding. "When I say this word, and you say that word, are we talking about the same thing?" would be great.
I first started working with climate scientists in my PhD under Paulina Jaramillo at Carnegie Mellon, and the most value I had through that exchange was I just went and sat with them for a week. And sitting in a group of climate scientists and hydrologists for a week and talking to them, that regular back and forth, I found infinitely valuable, and having that face time again, because that's where you really surface these, "I think the data you’ve given me is this. Let me make sure it is. Oh, it's actually not at all what I thought it was. Okay, let me change course and account for that." And so that language, understanding why and where they're coming from, all that can really help.
And the last thing I'd say is just the compromise aspect of it. I want to run energy-system models the way I have always run energy-system models. Many people want to run their models and that's not because we're stubborn. Well, I might be a little stubborn, but it's mostly just because we have designed the models to do a certain thing because that's what we think is important. Energy-system models are often run on an hourly or sub-hourly basis because we have to constantly balance supply and demand. Maybe, at this point, it's just a little too much to ask for hourly outputs, for instance. And so having that compromise, understanding how far I can bring you towards me but also how far I need to go towards you, can also help find common ground and get better than what you would have if you had not done this in the first place.
Daniel Raimi: Yeah, really interesting. Well, Michael, this has been a fascinating conversation. It's a fascinating stream of work and I hope folks will reach out to you and find ways to collaborate to pursue it, because it's I think fairly obvious that it's really important, and I really enjoyed talking to you about it.
So now, let's go to our last question that we ask all of our guests, which is asking you to recommend something that you've read or watched or heard lately that you've enjoyed. And I have the pleasure of seeing you fairly regularly. So, feel free to re-recommend something to me if you want to talk about something that we've already talked about. But what would you recommend to our listeners as something that's on the top of your literal or your metaphorical reading stack?
Michael Craig: All right. So, I have two books: one very in line with this conversation and one somewhat adjacent, but I still think very relevant to thinking about environmental challenges in general.
Daniel Raimi: Great.
Michael Craig: The first one I have to recommend is a book that came out last year called Downscaling Techniques for High-Resolution Climate Projections. The reason I think this book is very nice is it's written for a non-climate-science audience, despite the name of the book.
Daniel Raimi: The name does not inspire generality.
Michael Craig: Absolutely, but they have several chapters in there that are written for practitioners, for city managers, city planners for instance, trying to inform people who are not familiar with climate models, "Here is how to think about climate model outputs. Here is where they're good. Here is where they're not good." So, it's great for that audience, and I think it's also great for an audience like myself who have some familiarity with climate system modeling. I've worked with a lot of colleagues in the area, but I could also always brush up on understanding downscaling, for instance, which is generating high-resolution outputs. And so, if you want to work with climate data, I think this book is one way that is a great starting point in that journey.
Daniel Raimi: Excellent. What else?
Michael Craig: The other book that I recommend, I read six months ago, and I can't stop thinking about it, and that's The Making of the Atomic Bomb by Richard Rhodes. And I think it is an incredible book that has really stuck with me, because it's a beautiful story about science, and it's very inspiring. And I find myself thinking about just the characters in the book and what they accomplished and how they accomplished it, while I'm walking to work, for instance. And I also think he does a great job talking about how this was an incredible process with feats of heroic scientific achievement—and it also produced something that has harmed a lot of people.
I think it's impossible to read the end of that book without becoming emotional, and I think it is just a great depiction of how science can be done in different ways, and the end result can be sometimes good, sometimes not so good, and from whose perspective you can also argue that. And when we think about environmental problems, I think often, there are some very clear-cut environmental problems. I think oftentimes, though, it is not, "This one thing is good and this one thing is bad. And just get rid of the bad thing and more of the good thing." It's much more complex than that. And so I think, to me, that book is a very stark encapsulation of how you can have good and bad under the one roof.
Daniel Raimi: Yeah, that sounds fascinating. And his previous book was about the energy system, right? I think his previous book was called Energy.
Michael Craig: Yes, that's his brand new book that came out three or four years ago. Yes.
Daniel Raimi: Yeah. And the atomic bomb book is older?
Michael Craig: No, it came out a long time ago. It came out–
Daniel Raimi: Oh, it's been out for a long time.
Michael Craig: Right. It came out 20 or 30 years ago, I would think. It came out right as the records were being declassified.
Daniel Raimi: Okay, great.
Michael Craig: Right.
Daniel Raimi: Great. Okay. So the most recent book is Energy, but an oldie and a goodie is the one Michael is recommending.
Michael Craig: Absolutely. I am just somebody new to the book. It has been out for a long time. So, many people listening to this are probably thinking, "Yeah, of course, it's great. It won the National Book Award and every award under the sun when it came out, however long ago." But some of us are too young to have read it when it first came out.
Daniel Raimi: Fair enough. Great. Well, Michael Craig from the University of Michigan, thank you so much for joining us today on Resources Radio. It's been a fascinating conversation.
Michael Craig: Thanks so much for having me.
Daniel Raimi: You've been listening to Resources Radio, a podcast from Resources for the Future (RFF). If you have a minute, we'd really appreciate you leaving us a rating or a comment on your podcast platform of choice. Also, feel free to send us your suggestions for future episodes.
This podcast is made possible with the generous financial support of our listeners. You can help us continue producing these kinds of discussions on the topics that you care about by making a donation to Resources for the Future online at rff.org/donate.
RFF is an independent, nonprofit research institution in Washington, DC. Our mission is to improve environmental, energy, and natural resource decisions through impartial economic research and policy engagement. The views expressed on this podcast are solely those of the podcast guests and may differ from those of RFF experts, its officers, or its directors. RFF does not take positions on specific legislative proposals.
Resources Radio is produced by Elizabeth Wason, with music by me, Daniel Raimi. Join us next week for another episode.