You’re out of free articles.
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
Sign In or Create an Account.
By continuing, you agree to the Terms of Service and acknowledge our Privacy Policy
Welcome to Heatmap
Thank you for registering with Heatmap. Climate change is one of the greatest challenges of our lives, a force reshaping our economy, our politics, and our culture. We hope to be your trusted, friendly, and insightful guide to that transformation. Please enjoy your free articles. You can check your profile here .
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Subscribe to get unlimited Access
Hey, you are out of free articles but you are only a few clicks away from full access. Subscribe below and take advantage of our introductory offer.
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Create Your Account
Please Enter Your Password
Forgot your password?
Please enter the email address you use for your account so we can send you a link to reset your password:
Wildfire evacuation notices are notoriously confusing, and the stakes are life or death. But how to make them better is far from obvious.

How many different ways are there to say “go”? In the emergency management world, it can seem at times like there are dozens.
Does a “level 2” alert during a wildfire, for example, mean it’s time to get out? How about a “level II” alert? Most people understand that an “evacuation order” means “you better leave now,” but how is an “evacuation warning” any different? And does a text warning that “these zones should EVACUATE NOW: SIS-5111, SIS-5108, SIS-5117…” even apply to you?
As someone who covers wildfires, I’ve been baffled not only by how difficult evacuation notices can be to parse, but also by the extent to which they vary in form and content across the United States. There is no centralized place to look up evacuation information, and even trying to follow how a single fire develops can require hopping among jargon-filled fire management websites, regional Facebook pages, and emergency department X accounts — with some anxious looking-out-the-window-at-the-approaching-pillar-of-smoke mixed in.
Google and Apple Maps don’t incorporate evacuation zone data. Third-party emergency alert programs have low subscriber rates, and official government-issued Wireless Emergency Alerts, or WEAs — messages that trigger a loud tone and vibration to all enabled phones in a specific geographic region — are often delayed, faulty, or contain bad information, none of which is ideal in a scenario where people are making life-or-death decisions. The difficulty in accessing reliable information during fast-moving disasters like wildfires is especially aggravating when you consider that nearly everyone in America owns a smartphone, i.e. a portal to all the information in the world.
So why is it still so hard to learn when and where specific evacuation notices are in place, or if they even apply to you? The answer comes down to the decentralized nature of emergency management in the United States.
A downed power line sparks a fire on a day with a Red Flag Warning. A family driving nearby notices the column of smoke and calls to report it to 911. The first responders on the scene realize that the winds are fanning the flames toward a neighborhood, and the sheriff decides to issue a wildfire warning, communicating to the residents that they should be ready to leave at a moment’s notice. She radios her office — which is now fielding multiple calls asking for information about the smoke column — and asks for the one person in the office that day with training on the alert system to compose the message.
Scenarios like these are all too common. “The people who are put in the position of issuing the messages are doing 20 other things at the same time,” Jeannette Sutton, a researcher at the University at Albany’s Emergency and Risk Communication Message Testing Lab, told me. “They might have limited training and may not have had the opportunity to think about what the messages might contain — and then they’re told by an incident commander, Send this, and they’re like, Oh my God, what do I do?”
The primary way of issuing wildfire alerts is through WEAs, with 78,000 messages sent since 2012. Although partnerships between local emergency management officials, the Federal Emergency Management Agency, the Federal Communications Commission, and cellular and internet providers facilitate the technology, it’s local departments that determine the actual content of the message. Messaging limits force some departments to condense the details of complicated and evolving fire events into 90 characters or fewer. Typos, confusing wording, and jargon inevitably abound.
Emergency management teams often prefer to err on the side of sending too few messages rather than too many for fear of inducing information overload. “We’re so attached to our devices, whether it’s Instagram or Facebook or text messages, that it’s hard to separate the wheat from the chaff, so to speak — to make sure that we are getting the right information out there,” John Rabin, the vice president of disaster management at the consulting firm ICF International and a former assistant administrator at the Federal Emergency Management Agency, told me. “One of the challenges for local and state governments is how to bring [pertinent information] up and out, so that when they send those really important notifications for evacuations, they really resonate.”
But while writing an emergency alert is a bit of an art, active prose alone doesn’t ensure an effective evacuation message.
California’s Cal Fire has found success with the “Ready, Set, Go” program, designed by the International Association of Fire Chiefs, which uses an intuitive traffic light framework — “ready” is the prep work of putting together a go-bag and waiting for more news if a fire is in the vicinity, escalating to the “go” of the actual evacuation order. Parts of Washington and Oregon use similar three-tiered systems of evacuation “levels” ranging from 1 to 3. Other places, like Montana, rely on two-step “evacuation warnings” and “evacuation orders.”
Watch Duty, a website and app that surged in popularity during the Los Angeles fires earlier this year, doesn’t worry about oversharing. Most information on Watch Duty comes from volunteers, who monitor radio scanners, check wildfire cameras, and review official law enforcement announcements, then funnel the information to the organization’s small staff, who vet it before posting. Though WatchDuty volunteers and staff — many of whom are former emergency managers or fire personnel themselves — actively review and curate the information on the app, the organization still publishes far more frequent and iterative updates than most people are used to seeing and interpreting. As a result, some users and emergency managers have criticized Watch Duty for having too much information available, as a result.
The fact that Watch Duty was downloaded more than 2 million times during the L.A. fires, though, would seem to testify to the fact that people really are hungry for information in one easy-to-locate place. The app is now available in 22 states, with more than 250 volunteers working around the clock to keep wildfire information on the app up to date. John Clarke Mills, the app’s CEO and co-founder, has said he created the app out of “spite” over the fact that the government doesn’t have a better system in place for keeping people informed on wildfires.
“I’ve not known too many situations where not having information makes it better,” Katlyn Cummings, the community manager at Watch Duty, told me. But while the app’s philosophy is “rooted in transparency and trust with our users,” Cummings stressed to me that the app’s volunteers only use official and public sources of information for their updates and never include hearsay, separating it from other crowd-sourced community apps that have proved to be less than reliable.
Still, it takes an army of a dozen full-time staff and over 200 part-time volunteers, plus an obsessively orchestrated Slack channel to centralize the wildfire and evacuation updates — which might suggest why a more official version doesn’t exist yet, either from the government or a major tech company. Google Maps currently uses AI to visualize the boundaries of wildfires, but stops short of showing users the borders of local evacuation zones (though it will route you around known road closures). A spokesperson for Google also pointed me toward a feature in Maps that shares news articles, information from local authorities, and emergency numbers when users are in “the immediate vicinity” of an actively unfolding natural disaster — a kind of do-it-yourself Watch Duty. The company declined to comment on the record about why Maps specifically excludes evacuation zones. Apple did not respond to a request for comment.
There is, of course, a major caveat to the usefulness of Watch Duty.
Users of the app tend to be a self-selecting group of hyper-plugged-in digital natives who are savvy enough to download it or otherwise know to visit the website during an unfolding emergency. As Rabin, the former FEMA official, pointed out, Watch Duty users aren’t the population that first responders are most concerned about — they’re like “Boy Scouts,” he said, because they’re “always prepared.” They’re the ones who already know what’s going on. “It’s reaching the folks that aren’t paying attention that is the big challenge,” he told me.
The older adult population is the most vulnerable in cases of wildfire. Death tolls often skew disproportionately toward the elderly; of the 30 people who died in the Los Angeles fires in January, for example, all but two were over 60 or disabled, with the average age of the deceased 77, the San Francisco Chronicle reported. Part of that is because adults 65 and older are more likely to have physical impairments that make quick or unplanned evacuations challenging. Social and technological isolation are also factors — yes, almost everyone in America has a smartphone, but that includes just 80% of those 65 and older, and only 26% of the older adult population feels “very confident” using computers or smartphones. According to an extensive 2024 report on how extreme weather impacts older adults by CNA, an independent, nonprofit research organization, “Evacuation information, including orders, is not uniformly communicated in ways and via media that are accessible to older adults or those with access and functional needs.”
Sutton, the emergency warning researcher, also cautioned that more information isn’t always better. Similar to the way scary medical test results might appear in a health portal before a doctor has a chance to review them with you (and calm you down), wildfire information shared without context or interpretation from emergency management officials means the public is “making assumptions based upon what they see on Watch Duty without actually having those official messages coming from the public officials who are responsible for issuing those messages,” she said. One role of emergency managers is to translate the raw, on-the-ground information into actionable guidance. Absent that filter, panic is probable, which could lead to uncontrollable evacuation traffic or exacerbate alert fatigue. Alternatively, people might choose to opt out of future alerts or stop checking for updates.
Sutton, though she’s a strong advocate of creating standardized language for emergency alerts — “It would be wonderful if we had consistent language that was agreed upon” between departments, she told me — was ultimately skeptical of centralizing the emergency alert system under a large agency like FEMA. “The movement of wildfires is so fast, and it requires knowledge of the local communities and the local terrain as well as meteorological knowledge,” she said. “Alerts and warnings really should be local.”
The greater emphasis, Sutton stressed, should be on providing emergency managers with the training they need to communicate quickly, concisely, and effectively with the tools they already have.
The high wire act of emergency communications, though, is that while clear and regionally informed messages are critical during life-or-death situations, it also falls on residents in fire-risk areas to be ready to receive them. California first adopted the “Ready, Set, Go” framework in 2009, and it has spent an undisclosed amount of money over the years on a sustained messaging blitz to the public. (Cal Fire’s “land use planning and public education budget is estimated at $16 million, and funds things like the updated ad spots it released as recently as this August.) Still, there is evidence that even that has not been enough — and Cal Fire is the best-resourced firefighting agency in the country, setting the gold standard for an evacuation messaging campaign.
Drills and test messages are one way to bring residents up to speed, but participation is typically very low. Many communities and residents living in wildfire-risk areas continue to treat the threat with low urgency — something to get around to one day. But whether they’re coming from your local emergency management department or the White House itself, emergency notices are only as effective as the public is willing and able to heed them.
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
Plus more of the week’s top fights in data centers and clean energy.
1. Osage County, Kansas – A wind project years in the making is dead — finally.
2. Franklin County, Missouri – Hundreds of Franklin County residents showed up to a public meeting this week to hear about a $16 billion data center proposed in Pacific, Missouri, only for the city’s planning commission to announce that the issue had been tabled because the developer still hadn’t finalized its funding agreement.
3. Hood County, Texas – Officials in this Texas County voted for the second time this month to reject a moratorium on data centers, citing the risk of litigation.
4. Nantucket County, Massachusetts – On the bright side, one of the nation’s most beleaguered wind projects appears ready to be completed any day now.
Talking with Climate Power senior advisor Jesse Lee.
For this week's Q&A I hopped on the phone with Jesse Lee, a senior advisor at the strategic communications organization Climate Power. Last week, his team released new polling showing that while voters oppose the construction of data centers powered by fossil fuels by a 16-point margin, that flips to a 25-point margin of support when the hypothetical data centers are powered by renewable energy sources instead.
I was eager to speak with Lee because of Heatmap’s own polling on this issue, as well as President Trump’s State of the Union this week, in which he pitched Americans on his negotiations with tech companies to provide their own power for data centers. Our conversation has been lightly edited for length and clarity.
What does your research and polling show when it comes to the tension between data centers, renewable energy development, and affordability?
The huge spike in utility bills under Trump has shaken up how people perceive clean energy and data centers. But it’s gone in two separate directions. They see data centers as a cause of high utility prices, one that’s either already taken effect or is coming to town when a new data center is being built. At the same time, we’ve seen rising support for clean energy.
As we’ve seen in our own polling, nobody is coming out looking golden with the public amidst these utility bill hikes — not Republicans, not Democrats, and certainly not oil and gas executives or data center developers. But clean energy comes out positive; it’s viewed as part of the solution here. And we’ve seen that even in recent MAGA polls — Kellyanne Conway had one; Fabrizio, Lee & Associates had one; and both showed positive support for large-scale solar even among Republicans and MAGA voters. And it’s way high once it’s established that they’d be built here in America.
A year or two ago, if you went to a town hall about a new potential solar project along the highway, it was fertile ground for astroturf folks to come in and spread flies around. There wasn’t much on the other side — maybe there was some talk about local jobs, but unemployment was really low, so it didn’t feel super salient. Now there’s an energy affordability crisis; utility bills had been stable for 20 years, but suddenly they’re not. And I think if you go to the town hall and there’s one person spewing political talking points that they've been fed, and then there’s somebody who says, “Hey, man, my utility bills are out of control, and we have to do something about it,” that’s the person who’s going to win out.
The polling you’ve released shows that 52% of people oppose data center construction altogether, but that there’s more limited local awareness: Only 45% have heard about data center construction in their own communities. What’s happening here?
There’s been a fair amount of coverage of [data center construction] in the press, but it’s definitely been playing catch-up with the electric energy the story has on social media. I think many in the press are not even aware of the fiasco in Memphis over Elon Musk’s natural gas plant. But people have seen the visuals. I mean, imagine a little farmhouse that somebody bought, and there’s a giant, 5-mile-long building full of computers next to it. It’s got an almost dystopian feel to it. And then you hear that the building is using more electricity than New York City. This is very intimidating
The big takeaway of the poll for me is that coal and natural gas are an anchor on any data center project, and reinforce the worst fears about it. What you see is that when you attach clean energy [to a data center project], it actually brings them above the majority of support. It’s not just paranoia: We are seeing the effects on utility rates and on air pollution — there was a big study just two days ago on the effects of air pollution from data centers. This is something that people in rural, urban, or suburban communities are hearing about.
Do you see a difference in your polling between natural gas-powered and coal-powered data centers? In our own research, coal is incredibly unpopular, but voters seem more positive about natural gas. I wonder if that narrows the gap.
I think if you polled them individually, you would see some distinction there. But again, things like the Elon Musk fiasco in Memphis have circulated, and people are aware of the sheer volume of power being demanded. Coal is about the dirtiest possible way you can do it. But if it’s natural gas, and it’s next door all the time just to power these computers — that’s not going to be welcome to people.
I'm sure if you disentangle it, you’d see some distinction, but I also think it might not be that much. I’ll put it this way: If you look at the default opposition to data centers coming to town, it’s not actually that different from just the coal and gas numbers. Coal and gas reinforce the default opposition. The big difference is when you have clean energy — that bumps it up a lot. But if you say, “It’s a data center, but what if it were powered by natural gas?” I don’t think that would get anybody excited or change their opinion in a positive way.
Transparency with local communities is key when it comes to questions of renewable buildout, affordability, and powering data centers. What is the message you want to leave people with about Climate Power’s research in this area?
Contrary to this dystopian vision of power, people do have control over their own destinies here. If people speak out and demand that data centers be powered by clean energy, they can get those data centers to commit to it. In the end, there’s going to be a squeeze, and something is going to have to give in terms of Trump having his foot on the back of clean energy — I think something will give.
Demand transparency in terms of what kind of pollution to expect. Demand transparency in terms of what kind of power there’s going to be, and if it’s not going to be clean energy, people are understandably going to oppose it and make their voices heard.
This week is light on the funding, heavy on the deals.
This week’s Funding Friday is light on the funding but heavy on the deals. In the past few days, electric carmaker Rivian and virtual power plant platform EnergyHub teamed up to integrate EV charging into EnergyHub’s distributed energy management platform; the power company AES signed 20-year power purchase agreements with Google to bring a Texas data center online; and microgrid company Scale acquired Reload, a startup that helps get data centers — and the energy infrastructure they require — up and running as quickly as possible. Even with venture funding taking a backseat this week, there’s never a dull moment.
Ahead of the Rivian R2’s launch later this year, the EV-maker has partnered with EnergyHub, a company that aggregates distributed energy resources into virtual power plants, to give drivers the opportunity to participate in utility-managed charging programs. These programs coordinate the timing and rate of EV charging to match local grid conditions, enabling drivers to charge when prices are low and clean energy is abundant while avoiding periods of peak demand that would stress the distribution grid.
As Seth Frader-Thompson, EnergyHub’s president, said in a statement, “Every new EV on the road is a win for drivers and the environment, and by managing charging effectively, we ensure this growth remains a benefit for the grid as well.”
The partnership will fold Rivian into EnergyHub’s VPP ecosystem, giving the more than 150 utilities on its platform the ability to control when and how participating Rivian drivers charge. This managed approach helps alleviate grid stress, thus deferring the need for costly upgrades to grid infrastructure such as substations or transformers. Extending the lifespan of existing grid assets means lower electricity costs for ratepayers and more capacity to interconnect new large loads — such as data centers.
Google seems to be leaning hard into the “bring-your-own-power” model of data center development as it looks to gain an edge in the AI race.
The latest evidence came on Tuesday, when the power company and utility operator AES announced a partnership with the hyperscaler to provide on-site power for a new data center in Texas. signing 20-year power purchase agreements. AES will develop, own, and operate the generation assets, as well as all necessary electricity infrastructure, having already secured the land and interconnection agreements to bring this new power online. The data center is set to begin operations in 2027.
As of yet, neither company has disclosed the exact type of energy infrastructure that AES will be building, although Amanda Peterson Corio, Google’s head of data center energy, said in a press release that it will be “clean.”
“In partnership with AES, we are bringing new clean generation online directly alongside the data center to minimize local grid impact and protect energy affordability,” she said.
This announcement came the same day the hyperscaler touted a separate agreement with the utility Xcel Energy to power another data center in Minnesota with 1.6 gigawatts of solar and wind generation and 300 megawatts of long-duration energy storage from the iron-air battery startup Form Energy.
The microgrid developer Scale has acquired Reload, a “powered land” startup founded in 2024, for an undisclosed sum. What is “powered land”? Essentially, it’s land that Reload has secured and prepared for large data centers customers, obtaining permits and planning for onsite energy infrastructure such that sites can be energized immediately. This approach helps developers circumvent the years-long utility interconnection queue and builds on Scale’s growing focus on off-grid data center projects, as the company aims to deliver gigawatts of power for hyperscalers in the coming years powered by a diverse mix of sources, from solar and battery storage to natural gas and fuel cells.
Early last year, the Swedish infrastructure investor EQT acquired Scale. The goal, EQT said, was to enable the company “to own and operate billions of dollars in distributed generation assets.” At the time of the acquisition, Scale had 2.5 gigawatts of projects in its pipeline. In its latest press release the company announced it has secured a multi-hundred-megawatt contract with a leading hyperscaler, though it did not name names.
As Jan Vesely, a partner at EQT said in a statement, “By bringing together Reload’s campus development capabilities, Scale’s proven islanded power operating platform, and EQT’s deep expertise across energy, digital infrastructure and technology, we are supporting a more integrated approach to delivering power for next-generation digital infrastructure today.”
Not to say there’s been no funding news to speak of!
As my colleague Alexander C. Kaufman reported in an exclusive on Thursday, fusion company Shine Technologies raised $240 million in a Series E round, the majority of which came from biotech billionaire Patrick Soon-Shiong. Unlike most of its peers, Shine isn’t gunning to build electricity-generating reactors anytime soon. Instead, its initial focus is producing valuable medical isotopes — currently made at high cost via fission — which it can sell to customers such as hospitals, healthcare organizations, or biopharmaceutical companies. The next step, Shine says, is to scale into recycling radioactive waste from spent fission fuel.
“The basic premise of our business is fusion is expensive today, so we’re starting by selling it to the highest-paying customers first,” the company’s CEO, Greg Piefer told Kaufman, calling electricity customers the “lowest-paying customer of significance for fusion today.”