The Brief

Tens of millions of voters have had their citizenship status and other information checked using a revamped tool offered by the Trump administration, even as many states — led by both Democrats and Republicans — are refusing or hesitating to use it because of outstanding questions about the system.

U.S. Citizenship and Immigration Services (USCIS) says election officials have used the tool to check the information of more than 33 million voters — a striking portion of the American public, considering little information has been made public about the tool's accuracy or data security.

The latest update to the system, known as SAVE, took effect Aug. 15 and allows election officials to use just the last four digits of voters' Social Security numbers — along with names and dates of birth — to check if the voters are U.S. citizens, or if they have died.

The upgrade makes the tool far more accessible, since it now aligns with the information most states collect or have access to for most voters. But the Department of Homeland Security (DHS), which houses USCIS, has not responded to questions about the system from members of Congress, and numerous election officials NPR spoke with expressed concern about what else the Trump administration could do with the data it acquires from states.

"There's still uncertainty about what is happening, what happens to the data that are shared with USCIS," said Charles Stewart, a political science professor who directs the MIT Election Data and Science Lab. "I don't know if this means that the USCIS now has a depository of one-sixth of all [the country's] registered voters."

In recent months, several Republican-led states have brokered new agreements with USCIS to use SAVE, or announced the results of SAVE reviews. Ohio election officials will begin removing from their rolls thousands of inactive voters that SAVE identified as deceased. And Louisiana's secretary of state announced last week that officials identified 79 likely noncitizens who had voted in at least one election since the 1980s, after running nearly all of the state's 2.9 million registered voters through SAVE.

DHS is encouraging officials in other states to upload data to the system — even going so far as to make millions of dollars of grant money contingent on them using it.

But USCIS did not respond to NPR's questions about what happens to the data states upload and who has access to it.

And officials in other GOP-led states have expressed caution about using the system.

Last month, North Carolina's Republican-controlled state election board did not take up an offer by USCIS to participate in a "soft launch" of the upgraded tool. Spokesperson Patrick Gannon told NPR in a statement that state officials are pursuing "agreements to ensure that proper safeguards would be in place to protect and secure the data, if a decision is ultimately made to use the service."

Mississippi Secretary of State Michael Watson, a Republican, told NPR the upgraded SAVE seemed like a "fantastic tool," but he still has questions before he can run his voter list through it to ensure it is authorized under state law.

"Where's that data going? And at the end of the day, is it stored? What are they going to do with it? Who has access? Is it shared?" Watson told NPR last month. "I don't want to do something that I don't necessarily have the ability to do without legislative authority. So we just want to be very clear on that before we move forward."

SAVE concerns are compounded by other Trump moves on elections

The push for states to use the upgraded data system comes as the Trump administration is taking unprecedented steps to assert control over elections as well as collect and aggregate personal data on Americans — at times potentially risking the security of that data.

Numerous voting officials told NPR they felt the revamped SAVE tool could be useful for confirming citizenship status without encumbering voters, and many Democrats also acknowledged the efforts by Trump's USCIS to work in a bipartisan fashion.

But concerns about SAVE are compounded by other moves by President Donald Trump's administration. Trump continues to make baseless claims about widespread noncitizen voting, has attempted to change voter registration rules to include proof-of-citizenship requirements, and has directed the Justice Department to prioritize prosecuting noncitizens who register or vote. DHS also recently elevated a person who spread conspiracy theories about the 2020 election to a post on election integrity.

As part of the administration's stated aim to crack down on noncitizen voting, USCIS prioritized updating the data system known as Systematic Alien Verification for Entitlements, or SAVE, that state and federal agencies have for decades used to query DHS databases to determine if foreign-born individuals are eligible for various benefits.

The agency, working with the Department of Government Efficiency, or DOGE, made SAVE free to states and allowed bulk searches, as opposed to looking up one person at a time — changes that were celebrated by many voting officials.

"It's getting access to data that already exists and just making it so [election officials] can more easily compare things without making it hard for voters," said Idaho Secretary of State Phil McGrane, a Republican who has begun testing out SAVE's upgraded capabilities. "I'm trying to make it easier for voters while also doing these validations."

In May, USCIS linked data from the Social Security Administration, allowing election officials for the first time to check the citizenship of many U.S.-born citizens with the voter's name, date of birth and nine-digit Social Security number. NPR was the first news outlet to report on the change.

The integration with Social Security Administration data also means SAVE can show if someone appears on that agency's Death Master File.

Since most states only collect the last four digits of Social Security numbers from voters, the latest upgrade swung open the door for many more states to use SAVE. USCIS says almost 80% of the 33 million voters validated via SAVE were run through the system since the Aug. 15 change.

An unprecedented push for voter data

In mid-July, California Sen. Alex Padilla and two other Democratic U.S. senators formally requested information on the updated SAVE system from Homeland Security Secretary Kristi Noem, including the accuracy rate of the SAVE program, what data USCIS stores and who has access, and whether the agency is following protocols spelled out in federal privacy laws.

The senators asked for answers by July 29 but have not heard back.

"In light of the president's threats to issue unconstitutional executive orders to interfere with state elections, DHS owes Congress and the public some basic transparency about what they are doing with state voter rolls," Padilla told NPR.

A voting booth is seen during a Kentucky primary at a polling place in the city of Simpsonville on May 16, 2023.

(

Jon Cherry

/

Getty Images

)

SAVE pings a number of data sources, as opposed to being a database in and of itself. But under USCIS policy, all queries are saved for 10 years for audit purposes, so if a state runs its whole voter list through the tool, that data will remain with DHS for a decade.

Maine Secretary of State Shenna Bellows, a Democrat, told NPR that such a system sounded like the beginnings of a national voter registration list that would raise privacy and security concerns.

"They said they're going to keep our data for 10 years," Bellows said. "If [former Attorney General] Merrick Garland were asking for this or President Biden, I have to think that the red states would be calling for their heads."

The agreements that states sign to use SAVE include a clause that grants DHS permission to use information from states "for any purpose permitted by law, including, but not limited to, the prosecution of violations of Federal administrative or criminal law."

At a White House meeting about SAVE in late July, USCIS officials tried to reassure state voting officials about sharing sensitive voter data, according to Connecticut Secretary of State Stephanie Thomas, who attended the meeting.

"They said many times like, 'Look, we know the states are in charge of elections. We don't want your data. We don't want your lists,' " Thomas recalled.

But in recent weeks, the Department of Justice has been making controversial demands to numerous states to turn over such data, including in several instances voter rolls that include personal data like driver's license numbers and partial Social Security numbers.

Thomas, a Democrat, said that when a Republican secretary of state pointed out the DOJ's voter data demands to USCIS officials, the officials responded that they worked for a separate agency.

"The worst thing you can get is a false positive"

USCIS hasn't publicized detailed evidence about the accuracy of the tool or shared what testing went into the program before it was released to states, though it asserts SAVE's accuracy has markedly improved with the recent upgrades.

But voting rights groups and some election officials are voicing concerns that eligible voters could face barriers to casting ballots or be improperly removed from the rolls if states over-rely on incomplete information from SAVE.

USCIS acknowledges that certain categories of people who acquired U.S. citizenship, such as some foreign-born children of U.S. citizens, cannot be verified by SAVE.

Furthermore, data matching in elections is notoriously difficult and there are questions about the completeness of the Social Security Administration's citizenship data USCIS is relying on.

Wesley Wilcox, a Republican elections supervisor in Marion County, Fla., signed an agreement this summer giving his county access to SAVE. But he said he plans to double check any data he gets from the system, since he's encountered multiple instances in which SSA data indicated a person was dead when they weren't.

"It's like any new process that you put into effect … we're going to do that legwork," Wilcox said. "I want to be as accurate as humanly possible at all times."

For a portion of foreign-born individuals, SAVE prompts user agencies to submit more information, such as a person's naturalization certificate number or alien registration number, for their case to be manually reviewed. USCIS told NPR that of the 33 million voters submitted to the upgraded SAVE so far, less than 1% have required that manual review. The agency did not respond to NPR's question about the results of the manual review, how many noncitizens on voter rolls have been identified to date or what portion of the results so far were inconclusive.

While USCIS' materials say election officials are not supposed to reject voter registrations or remove voters from the rolls if the SAVE system asks for more voter information, it is not yet clear if there are consequences if states skip those steps.

Last year, thousands of U.S. citizens in states including Alabama, Virginia and Texas were removed from the rolls or deactivated after election officials relied on imperfect data to identify suspected noncitizens.

Conservative election integrity advocates have broadly celebrated SAVE's development, but among some, there is also an acknowledgement that states need to be cautious when removing people from voter rolls.

"I am hopeful that a great deal of care and a great deal of contemplation is going into this process because the worst thing you can get is a false positive," said J. Christian Adams, president of the conservative Public Interest Legal Foundation. "The worst thing you could do, as some states have done in the past, is remove a citizen from the voter rolls as a noncitizen. That should not happen."

In 2019, a court settlement required Adams to apologize to a group of Virginia voters his organization incorrectly claimed were noncitizens.

Louisiana's test case

The most in-depth data about how the updated SAVE is working came last week from Louisiana.

The state's Republican secretary of state, Nancy Landry, told reporters that by using SAVE, her office identified 390 people on the voter rolls who they believe are noncitizens. Seventy-nine of them were found to have voted. Landry said there was a review of each suspected noncitizen and her office worked with the FBI to investigate their citizenship status.

"They have been given notice we have reason to believe they are not a U.S. citizen," Landry said. "They have the opportunity to come in and provide documentation that they are in fact a U.S. citizen."

Anyone who does not respond or provide proof within 21 days is removed from the rolls, Landry said, and noncitizens who registered to vote — regardless of whether they were among the 79 who voted in recent decades — will be referred for criminal prosecution.

The number of suspected noncitizens Louisiana found to have voted amounts to less than 0.003% of the state's registered voters, a percentage that aligns with what many other reviews on the issue have found.

"The very small numbers suggest that the results are what we expected, and that they are not alarming," said Stewart of MIT.

He added that number might shrink further as some people may prove to be citizens.

Yet even as there are signs the SAVE program is generating helpful data, Stewart said it is troubling that USCIS is withholding answers to key questions, like what happens to the data that is uploaded or detailed breakdowns on SAVE results so far.

In Minnesota, state voting officials would need a change to state law to use SAVE for elections. But Secretary of State Steve Simon, a Democrat, said he won't push for such a change until he has more clarity on how it works.

"Certainly it's impressive what they've done in a short period of time," Simon said. "[But] I'd want to see a lot more by way of testing and assurances about the accuracy before we would be a willing and enthusiastic participant."
Copyright 2025 NPR

While living in L.A., I’d been hearing about this thing — “the desert.” Seemed a bit odd to someone who lived on the Westside next to the ocean, but OK, you can also see mountains from my house, so why not throw in an entire landscape food group?

A few years ago, I figured it was time to try it out, and my family and I headed to Joshua Tree for the weekend.

We began driving east and were soon in that vast no man’s land around the 10 Freeway. On one side, I could see a moving train, with all those boxcars. I was watching, entranced, waiting to see it go past. And it kept going. And going. It was like a vibrant desert serpent, wending its way through the landscape, each boxcar a different hue. It seemed so romantic, and then I thought “in each one, there’s a whole load of hanging car air fresheners, dental floss and Japanese waving cats” and suddenly the spell was broken. Ah well.

The eerie, alien Joshua Tree.

(

Sean Gallup/Getty Images,

/

Getty Images North America

)

As we got nearer, I said “Huh, what's that weird thing that looks like an alien?” and my husband said, “Yes, that’s a Joshua Tree.”

What? Wait, stop the car. There’s an actual Joshua Tree? First, who knew, and second, that is no tree. That, sir, is a weird misshapen cactus, with multiple crooked arms reaching into the air, each with its own spiky fur muff. Kinda like one of those waving inflatable guys if they were static. (Ah, I've been in L.A. too long).

Bougie smores

The vast desert sky.

(

Steve Holtzman

/

Courtesy Suzanne Levy

)

We finally arrived at our destination, a campsite — OK, a glamping campsite, a whole bunch of airstream trailers laid out with firepits. I knew it was bougie when I went to buy a smores kit from the shop and came back with 70% dark chocolate, artisanal graham crackers and single origin marshmallows. (I made that marshmallow one up, but it really is just a matter of time).

We made a fire, marveling at the gorgeousness of the sunset, the morphing colors and the vastness of the sky. I sat afterwards watching the flames, the cold air on my skin, under a blanket of stars. Tears pricked my eyes.

In the morning we headed to Joshua Tree National Park. When we got out of the car, the vista hit me, with that bright, bright sun, strange looking rocky outcrops, and nothing but Joshua trees as far as the eye could see. I thought: I have never seen anything like this. I couldn’t have even imagined something like this. In London, if you go on a day trip, you can go back in time, say to the thatched roofs of the Cotswolds — but not to a different planet.

What an extraordinary continent this is, I think. I am awestruck by its endless, varied landscapes — the plains, the Rockies, the oceans and the forests. And there’s something particular about the desert — its simplicity, its reduction to just three or four elements, which relaxes you, chills you out.

I remember heading back to L.A. after the weekend, happy and mellow, smiling as we drove west into the setting sun. It’s a cliche, but the desert does call you back. And I hope to get back there very soon.

We have another pip — the second egg in Big Bear’s famous bald eagle nest is showing signs of hatching.

The first pip, or crack, was confirmed in Jackie and Shadow’s egg no. 2 on Saturday morning, according to Friends of Big Bear Valley, the nonprofit that runs a popular YouTube livestream of the nest overlooking Big Bear Lake.

That’s about a day after the first egg started showing signs of hatching on Friday morning. The egg shell has continued to crack as the chick breaks through, revealing more of the eaglet’s fuzzy gray feathers as time goes on.

More than 38,000 people were watching the livestream shortly after the organization confirmed the second crack, compared with the more than 26,000 viewers who tuned in on Friday.

“The first egg is still in the process of hatching, it is not considered hatched until it is completely free of the egg shell. The chick has popped its head out of the shell to say a happy hatch day to mom and dad!” Friends of Big Bear Valley wrote on Facebook to more than a million followers on Saturday. “It also appears that the second egg has a pip. It is not well defined as of this morning, but we will likely see more progress throughout the day.”

Jackie and Shadow's usual incubation timeline is around 38 to 40 days, according to the nonprofit.

What’s next

With pips in place, it could take the chicks a day or two to complete the hatching process, as seen with last season’s trio.

Friends of Big Bear Valley won’t know for sure if any chicks are male or female, as the organization has said the only way to tell is with a blood test.

But once eaglets are around 9 or 10 weeks old, there should be signs that can help the nonprofit make an educated guess, including the chicks’ size, ankle thickness and vocal pitch.

Generally speaking, female bald eagles are larger than males. Female bald eagles also tend to have larger vocal organs — the syrinx — which leads to deeper, lower-pitched vocalizations, according to Friends of Big Bear Valley.

What do we call the chicks?

Historically, Jackie and Shadow’s chicks are given temporary nicknames initially, such as Chick 1 and Chick 2, or Bigger Chick and Smaller Chick (which some fans affectionately nicknamed Biggie and Smalls).

The final decision has then been left up to Big Bear Valley elementary school students. Previous chicks have been named Stormy, BBB (for Big Bear Baby), Simba and Cookie through that process.

Last year, Friends of Big Bear Valley crowdsourced more than 50,000 name choices in a week-long fundraiser, with the students voting from 30 finalists on official ballots delivered by the nonprofit.

The feathered parents’ eaglets were named Sunny and Gizmo last April.

One of last season’s three chicks didn’t survive a winter storm within weeks of hatching. Friends of Big Bear Valley named that chick “Misty” in honor of one of their late volunteers who is “still very missed,” the organization previously shared.

When Kekoa Gittens was 3, his preschool teacher told his mother he was a problem. He couldn’t sit still. He didn’t participate. When other kids learned the alphabet, he didn’t pay attention.

The next year, Kekoa’s classroom problems worsened. His mother, Sonia Gittens, took him to his pediatrician, who referred the boy to an eye doctor.

That doctor looked at the back of Kekoa’s eyes and diagnosed him with myopic degeneration, a dramatic form of nearsightedness.

“They are too little. They don’t know how to express themselves and say, ‘I cannot see it, teacher,’” said Sonia Gittens, who lives in the Marin County town of Corte Madera.

Today, Kekoa is a successful high schooler, but too many kids don’t get their eyes checked until they’re far behind in school.

Vision problems, particularly nearsightedness, have grown more common among American children. Roughly one in four school-age kids, or 25%, wear glasses or contacts, a proportion that increases as kids get older, according to 2019 federal survey data.

In California too few children on Medi-Cal like Kekoa are getting their eyes checked, and the problem is growing worse. Just 16% of school-age kids on Medi-Cal saw an eye doctor between 2022 and 2024 for first-time eye exams, continuing vision check ups or glasses, according to a report commissioned by the California Optometric Association. That’s down from 19% eight years earlier. The report, based on two years of Medi-Cal data, suggests that the state is moving in the wrong direction even as eye problems become more prevalent among kids.

Medi-Cal provides insurance for low-income Californians and those with disabilities.

“Every day when I see these children it is always a surprise to me that the kids are not getting the care they need,” said Ida Chung, a pediatric optometrist and an associate dean at Western University of Health Sciences in Pomona.

The trend indicated in the report is alarming, Chung said. In her clinic, where about half of children are on Medi-Cal, it’s common for kids with congenital vision problems to visit for the first time when they’re in first grade or later. That indicates to Chung that many kids don’t have enough access to eye care.

Though kids might be getting basic vision screenings at school or from a pediatrician, some eye problems are still overlooked. “It’s something the child had before they were born,” Chung said.

Eye exams decrease statewide

Colusa County, a rural farming region north of Sacramento, saw the sharpest drop in kids’ eye doctor appointments in the state from 20% between 2015-16 to just under 2% between 2022-24.

Nearly all counties — 47 out of 58 — performed worse on vision care than they did in the past, the report shows, with some, like Colusa, declining significantly.

Most of the severe declines happened in rural areas, although urban counties like San Francisco and Los Angeles also saw decreases. Only seven counties improved the rate of children receiving eye exams or glasses. Four counties were excluded for comparison in the report because the numbers were too small.

“The decline in performance here is so widespread that something really needs to happen,” said David Maxwell-Jolly, a health care consultant who authored the report and the former director of the Department of Health Care Services, which oversees Medi-Cal. “These numbers are way lower than what you would expect to be seeing if we’re doing a good job of detecting kids with treatable conditions.”

A spokesperson for the Department of Health Care Services said in an email the state could not confirm the accuracy of an external report, noting that vision services can be difficult to track because “not all encounters are captured in a single, comprehensive dataset.”

For example, many initial vision screenings take place in the pediatrician’s office during well-child visits, which include eye and hearing screenings as well as immunizations and developmental checks. State data shows about half of kids with Medi-Cal receive well-child visits.

Still, experts say the low numbers tell a real story: if children were reliably getting follow-up care from initial screenings, the share who get comprehensive eye exams and glasses would be closer to 25-30% — in line with the known prevalence of vision problems among kids — rather than the 16% found in the optometric association’s report.

Maxwell-Jolly said the analysis he conducted replicated an internal, unpublished department report tracking vision services between 2015 and 2016. His analysis, based on data obtained through a public records act request, updated the results for more recent years.

The state’s most recent Preventive Services Report, which measures how well Medi-Cal delivers preventive care to children, shows the rate of comprehensive eye exams for children and young adults ages 6-21 is similar to the optometric association’s analysis at 17%.

Contra Costa County experienced the third largest decline in children’s eye care in the state. A spokesperson for Contra Costa Health Plan said Medi-Cal health plans are not required by the state to track vision benefits and that it would take time to understand the data. The state, however, does track vision services internally, according to the health care services department.

A bill sponsored by the optometric association and authored by Assemblymember Patrick Ahrens, a Democrat from Cupertino, aims to require the state to establish vision benefit quality measures and report performance data publicly. The goal of the legislation is to track where kids do not have enough access to vision services and to ensure that Medi-Cal providers are improving services.

Rural challenges

Amy Turnipseed, chief strategy and government affairs officer for Partnership HealthPlan of California, said rural parts of the state struggle to find enough providers. The nonprofit health insurer provides Medi-Cal for 24 northern counties, including Colusa and Modoc.

In Modoc County, which borders Oregon and Nevada, one optometrist serves a 90-mile radius. Partnership has worked closely with that optometrist to ensure they continue accepting Medi-Cal patients, Turnipseed said.

“In rural counties with lower populations, losing even one provider can exponentially impact the access to services to families,” Turnipseed said. “In the past few years we’ve seen vision providers reduce or limit their Medi-Cal, which makes it harder for families to see providers.”

An assortment of glasses at Vision to Learn mobile optometry clinic at Esther Lindstrom Elementary School in Lakewood on March 20, 2026. Photo by Ariana Drehsler for CalMatters Modoc is one of just seven counties where more children have received vision care in recent years, according to the report.

Providers frequently cite low reimbursement rates from the state as a reason for not accepting Medi-Cal patients. The California Optometric Association estimates only about 10% of its members accept Medi-Cal. The reimbursement rate for a comprehensive eye exam is about $47, said Kristine Shultz, association executive director.

“Our reimbursement rates haven’t increased in 25 years. Imagine getting paid what you were paid 25 years ago,” Shultz said.

Schools check kids’ vision, but follow-up is spotty

State law requires schools to periodically check kids’ vision starting in kindergarten. Those screenings are a good bellwether for if a child is struggling to see in class, said Chung with Western University. The problem is getting the kids who fail the screening to an eye doctor.

Chung runs an academic optometry clinic that works with local schools in Pomona. Each year up to 35% of students fail the screening, meaning they likely have a vision problem. But based on conversations with school nurses, Chung said only about 7% of those children then go to an eye doctor and come back to school with glasses.

Chung, who chairs the children’s vision committee for the California Optometric Association, said colleagues who work with school districts around the state report similar experiences.

“If a high number of those children are not getting the follow up care, we may just be fooling ourselves and checking a box,” Chung said. “We’re in compliance with the law in California but are we really helping the children?”

For some families, the answer is no. That’s what happened to Kekoa when he was 3. The school checked his eyes and said he might have vision problems, but his mother, Gittens, waited. Her son was still learning his numbers and letters. How would he be able to read an eye chart, she reasoned. It wasn’t until his problems got worse that Gittens took Kekoa to an eye doctor.

Now, at 15, Kekoa wears contacts and likes athletics. He needs to see to compete in capoeira martial arts competitions and surf on the weekends, his mother said.

First: Dr. Kiyana Kavoussi shows letters on a monitor during Noah Mattison’s, 11, visual acuity test. Last: Optician Maya Ortega looks at Italia Martin’s, 6, eyes before she chooses new glasses inside the Vision to Learn mobile optometry clinic at Esther Lindstrom Elementary School in Lakewood on March 20, 2026. Photos by Ariana Drehsler for CalMatters Many parents lack the resources to take their kids to the doctor, or simply wait. Notes from school nurses flagging that a child failed a vision screening may also get lost in a backpack on the way home, educators say. The California Department of Education does not track the results of school vision screenings.

Vision To Learn, a nonprofit, created a mobile eye clinic to help bridge the gap between kids failing school vision screenings and getting glasses. The group brings an optometrist to campus, meaning kids that need an eye exam can get one the same day and go home having gotten a prescription and ordered glasses.

Damian Carroll, chief of staff and national director, said Vision to Learn’s numbers tell a similar story to Chung’s. About one-third of students screened are unable to read the eye chart, but very few of those kids have adequate glasses.

In the California schools where the program operates, around 70% of kids who have been prescribed glasses did not own a pair. Another 20% had glasses with outdated prescriptions, according to internal data, Carroll said.

And that gap can drastically affect learning outcomes or behavior in school.

“First and second graders who try on glasses the first time are blown away because they just thought that’s how the world looked,” Carroll said. “They can see the leaves on the trees and the math on the board, and it’s shocking to them.”

For the record: This story has been updated to reflect that Maxwell-Jolly’s study replicated the methodology of an earlier one by the Department of Health Care Services, but did not republish department findings.

Supported by the California Health Care Foundation (CHCF), which works to ensure that people have access to the care they need, when they need it, at a price they can afford. Visit www.chcf.org to learn more.

This article was originally published on CalMatters and was republished under the Creative Commons Attribution-NonCommercial-NoDerivatives license.

For more than 15 years, botanist Naomi Fraga of the California Botanic Garden has been trying to collect seeds from the rare Death Valley sage, for safekeeping in a vault of native California seeds. Each time, she's come home empty handed. But this year, with the desert in the midst of a big bloom, she's trying again.

"It's a little bit of a gamble," she says. "But, you know, the plant's having a really good year. I feel hopeful."

The plant has silvery-green pointy leaves, fuzzy buds and striking deep purple flowers. But it is challenging to study and to sample. Fraga says she often has to hike or scramble up mountainsides, or drive on backroads to find it. Very little is known about the plant's pollinator. And in exceptionally dry years, the Death Valley sage doesn't flower at all – meaning no seeds either.

The sage's habitat is mostly protected, within the boundaries of Death Valley National Park. But climate change doesn't respect park boundaries – and could push these plants that are already living on the brink into even more existential peril.

Naomi Fraga says for the first time since 2009, she found the Death Valley sage seeds. Soon, she says, she'll return with a team to make the first big harvest.

(

Krystal Ramirez

/

NPR

)

"You can imagine that if conditions were to get more difficult with a changing climate, it's going to be harder and harder to collect seed," Fraga says.

In late March, Fraga headed into the foothills of the Nopah Range, near an abandoned mine, to check on one of the largest populations she knows of. And for the first time since 2009, she found the seeds. Soon, she says, she'll return with a team to attempt the first big harvest of Death Valley sage seeds.

A bee pollinates a Death Valley Sage in the Nopah Range near Death Valley.

(

Krystal Ramirez

/

NPR

)