I applaud recent developments regarding one of the most devastating and widespread non-native tree-killing diseases, “Dutch” elm disease (DED). Brief descriptions of ecological importance of elms, the disease’s impact in North American and Europe, and difficulties in managing the rapidly evolving causal pathogens here. (See also a review of the ecological value of American elm here.)
Restoring America elm would be wonderful, so I rejoice at steps forward.
One task is to improve detection of the disease in forests. Currently detection is tardy because it relies on observation of visual symptoms followed by molecular confirmation. This process demands considerable time and labor; it is also error-prone. Earlier molecular detection methods also are labor intensive, costly, & have operational limitations.
A group of scientists led by Jian Jin and Songlin Fei are testing whether new spectral imaging & artificial intelligence can improve early detection. (See the publication by Wei et al.; full citation at the end of this blog.) Their goal is to detect subtle changes associated w/ disease developments before visual symptoms appear. The new technology — high-precision leaf spectral imagers — is already in use for agr crops. The devices needed are inexpensive and hand-held/portable. Can collect hyper- or multi-spectral images of a whole leaf in the field. This systems is also non-destructive & rapid.
To test applicability of this technology, the scientists inoculated the fungus responsible for DED into trees with known – and varying — disease susceptibilities. Then they collected spectral images of leaves from those trees to test accuracy of analyses conducted via both traditional machine learning & state-of-art deep learning models. These collections were made at three different times: 96 hours after inoculation / before visual symptoms; 4 weeks after inoculation / during visual symptom development; 15 weeks / foliar symptoms noticeable. They recorded the declining status of the using the traditional visual symptoms – wilting, yellowing, browning of leaves.
While detection accuracy varied by time of specimen collection and genetic heritage of the particular tree, machine-learning-based spectral & spatial analysis of high-resolution hyper & multi-spectral leaf images did detect DED symptoms. This advance would help detect pockets of disease in the forest and might be useful in screening elm genotypes for susceptibility to the pathogen. This latter ability would support resistance breeding programs.
However, further study is needed to determine whether light conditions, seasonal variations, or interactions with other pathogens might influence leaves’ spectral signature. Furthermore, scientists should test application of the process to additional elm genotypes. As Enrico Bonello and others have pointed out, however, the ideal would be to detect infection before even the start of symptoms – in other words, to detect even more subtle changes.
A second task is to breed American elms that can survive – even thrive – despite the continuing presence of the disease-causing pathogens. I rejoice here, too. So far, scientists have found varying levels of resistance in large “lingering” elms. This resistance appears to be heritable. Scientists are preparing reports of this progress for publication.
The USFS Northern Research Station is leading efforts of multiple partners to find and screen resistance of large elms across several regions. In New England, the principal partner is The Nature Conservancy; in the upper Midwest partners include the Army Corps of Engineers and Wisconsin Department of Natural Resources, In the lower Midwest the USFS is working with Metroparks Toledo, University of Illinois, Urbana Champaign, Appalachia Ohio Alliance and others. The Great Lakes Basin Forest Health Collaborative is helping to coordinate these efforts.
American elm has a huge range – covering much of the United States east of the Great Plains. Map Restoring the species to that range requires efforts throughout that range – so as to capture the genetic variability within the species and perpetuate its adaptations to the wide range of ecological conditions.
While restoring this magnificent and ecologically important tree species is worthwhile per se, a second motivation has emerged: using elms to restore riparian and wetland ecosystems now being harmed by loss of ash trees to the emerald ash borer.
Knight et al. (full citation at the end of this blog) note that these efforts’ success will depend not only on developing elms that can survive DED. It is also necessary to determine restoration strategies and silvicultural treatments that will promote the young trees’ ability to flourish despite challenges by storms, floods, competition with other plants, and wildlife feeding.
This team of USFS researchers describe ongoing tests of reintroduction strategies & silvicultural requirements in the Service’ Region 9. They note that reintroduction focuses on a single species. The goal of ecosystem restoration requires considering a broader range of factors. Both are important components for the success.
Testing Elm Reintroduction Factors
Research projects they describe include testing results of planting both bare-root seedlings and containerized stock. The latter approach is more labor-intensive but appears to provide better survival. When competing vegetation was removed & then controlled to prevent regrowth, large containerized trees had excellent survival & rapid growth. They also documented the value of caging trees to prevent deer browsing.
Other research projects explore elm seedlings’ ability to tolerate cold, floods, and shade. Scientists in New England and Wisconsin are observing how well progeny from various crosses between DED-tolerant American elms & local survivor trees are enduring the regions’ winters. One test is deploying progeny from paternal lines that are from different plant cold hardiness zones. It will be important to identify and plant trees that are adapted to local environmental conditions on top of being resistant to the DED pathogen.
Another group of tests investigates flood tolerance. Even minor dips or rises on floodplains lead to very different flooding intensities. Some of these experiments also consider shade tolerance. This is because managers hope can establish understory trees poised to grow rapidly by planting elm seedlings before harvest or mortality of canopy trees (e.g., ash). In one experiment in floodplain forests in Ohio, so far many elm seedlings have survived extensive spring & fall flooding. The seedlings are thriving across a range of microsite light environments. Even competition from invasive herbaceous plants does not appear to have impeded the elms’ survival.
DED has two methods of infecting nearby elms: that pathogen is either vectored by beetles that burrow below the tree’s bark, or through direct fungal contact via grafting of roots. Scientist do not yet know whether trees that tolerate DED infections caused by beetle attacks can withstand infection via root grafts. An experiment using paired elms was initiated in 2011. At the time of their writing, the trees had not yet grown sufficiently large to form root grafts – necessary before scientist could begin the experimental inoculations.
Finally, these many plantings have revealed some “unknown unknowns” — factors not previously identified. Knight et al. describe two studies:
1) Under the National Elm Trials, scientists are studying growth, stress and pest resistance, and horticultural performance of DED-tolerant American elm cultivars & other elm species and hybrids in 16 states. (See details here.)
2) A system of sentinel restoration sites has been established. Multiple DED-tolerant American elm selections have been planted in eight locations in four states to be an “early warning” system to identify additional pathogens of concern. Knight cites detection of a wood wasp at one site in Ohio and competition of thick grass and feeding by rodent on their roots in Minnesota.
Testing Restoration Strategies
As Knight et al. remind us, Eastern forests experience many forms of disturbance, including non-native pests and plants, increases in deer populations, land clearing, grazing, & climate change. Foresters want to know whether DED-resistant American elms might be used in restoration plantings in response to these natural and anthropogenic disturbance? They value elm for its ability to thrive in a wide variety of conditions. Furthermore, the species supports a diverse array of insect herbivores, which then support higher trophic levels, e.g., birds (Tallamy 2009). Another factor, not mentioned by Knight et al., is that even vulnerable elms can grow to some size before they are killed by DED. Knight et al. say multiple studies are testing use of American elm as one of several native species to be plant in ash ecosystems devastated by EAB. In northern Minnesota, the experiment is occurring in wet forest ecosystems formerly dominated by black ash. In Ohio researchers are observing elms planted in riparian systems where green ash forests used to be found. They report that early data indicate good initial survival of American elm in both studies.
The Nature Conservancy’s Connecticut River Program planted over 1900 disease-tolerant American elm cultivars at 76 sites in four New England states over the decade 2010 – 2021. Several DED-tolerant selections and their progeny were planted. Survival has varied considerably, they think depending on site factors, e.g., ice flows, height and density of competing vegetation, climate, damage from voles, deer browsing, others.
More recently, the partners have moved away from crossing survivor elms with cultivars because that results in too many related progeny, insufficient genetic diversity. In addition, the trees would not be adapted to the planting site because one parent was not local).
The Nature Conservancy’s participation has been funded by a grant of ~$2.4 million from a private foundation. TNC is helping to identify “lingering” or “survivor” American elm and restore them to floodplains and urban forests across New England. TNC has also funded groundbreaking research at the USFS to accelerate the breeding program and develop best practices for American elm reintroduction.
The Vermont chapter has been particularly active. Since 2014 it has been managing experimental elm trees plantings at 10 TNC natural areas and 26 partner-owned sites across the state. This effort has yielded ~7,000 trees that represent 142 novel crosses between 23 survivor elms identified by TNC in New England & several varieties identified by USFS from other parts of the country. Scientists plan to inoculate these trees in spring 2026. The trees’ vulnerability to the pathogen will then be evaluated over two years.
Knight et al. expect that in a decade or less these and other research projects will contribute needed understanding of various American elm propagules’ cold tolerance, flood tolerance, shade tolerance, response to competing vegetation, & root grafting. This information will allow managers to maximize survival of planted elm trees. It will also demonstrate how to usefully employ elms in ecosystem restoration. They caution that guidelines will probably vary to fit specific situations & site characteristics e.g., forest type, competing species, local hydrology, etc.
Knight et al. also identify topics that require additional research. The first factor mentioned are social & ecological contexts of restoration strategies. Social context will guide the formulation of a more strategic approach — setting goals, addressing such questions as the public perception & value of American elm in urban & forest areas, forest manager goals for incorporation of American elm, & municipal requirements for urban trees. It is essential to determine the long-term durability of resistance. Also need to explore how best to promote spread of DED-tolerant genes given the high numbers of local, non-resistant elms across the landscape. scale strategies.
Knight et al. note that need experimental plantings in additional parts of the species’ enormous range to identify potential problems, test performance on different soil types and in different climates. Need experiments to identify interactions among elm genetics and abiotic & biotic environ variables to guide silvicultural, site preparation, and planting strategies. I have observed apparently thriving American elms along roadways in the Washington, D.C. metropolitan area.
I believe no one is protecting them. Certainly other elms in the area have died. So far I have not found people trying to find “lingering” or “survival” elms here. I seek people who want to work with these trees!

The USDA Forest Service is not the only entity engaged in breeding American elms. The University of Minnesota is supporting an American elm breeding program through its Minnesota Invasive Terrestrial Plants and Pest Center (MITPPC) (see Bernardt citation at end of blog). Scientists are identifying DED-resistant elms in the wild, cloning and testing them, and replanting the strongest candidates across urban and natural landscapes. Their goal is to reintroduce the more resilient clones across Minnesota’s urban and natural landscapes, restoring lost canopy and biodiversity while preparing forests for a future stressed by climate change.
Bernardt describes the usual four essential steps: identification of trees that appear to be resistant; propagation of clones from those trees; growing sufficient numbers of these; and testing them for resistance. The Minnesota program – like many similar programs for breeding the many tree species being killed by non-native pests – ask the public to help in searching for “survivor” trees—American elms that appear to be withstanding Dutch elm disease even as others around them succumb.
The article summarizes the next steps and challenges. It notes, for example, that using clones rather than seedlings is essential because resistance is not reliably passed on during sexual reproduction. (However, Cornelia Pinchot Wilson has told me that colleagues should soon publish articles demonstrating that resistance is heritable.) Furthermore, the clones must be grown for several years—often five or more—until the trees are large enough to be tested for resistance. The article does not indicate whether the earlier step of propagating elm clones is challenging or easy. It has been difficult for other species, e.g., chestnuts, whitebark pine, ash, and koa.
To confirm whether a specific tree is resistant, the team typically tests each tree twice, since environmental factors like location and weather can influence outcomes. Trees that pass these tests move on to the next stage: reintroduction plantings in natural areas and parks. These field-growing trees serve two important roles. First, they contribute to restoring elm populations in natural and urban landscapes. Second, these trees can be observed over the long term to confirm whether they exhibit persistent resistance and are adapted to local environmental conditions.
The project’s success, the researchers say, hinges on collaboration. State agencies, local governments, and community members all play critical roles. Among those helping have been the Minnesota Department of Natural Resources, the Izaak Walton League, and Three Rivers Park District.
The article reminds us that resistance breeding is a long-term process. As noted above, clones must be grown for years before they can be tested.
Also, resistant trees aren’t immune to the pathogen. Instead, they survive despite the disease in sufficient numbers to restore the species to some of its former range and ecological role.
Finally, the trees must also survive the every-day challenges of life as a tree: storms, animal feeding, and other pests and diseases – native and non-native. The article mentions elm yellows disease but not elm zig-zag sawfly which has been moving West (it was detected in Ohio in 2023 and Wisconsin in 2024). Nor does it mention the fungus Plenodomus tracheiphilus, which is killing American elms in Alberta. Breeding program staff can help – for example, the Minnesota program now uses larger protective tubes to better shield the young trees from wildlife.
The Minnesota program plans to establish seed orchards. They hope that by planting trees confirmed to be resistant near to each other, they will cross-pollinate and produce seeds that are more likely to carry resistance, possibly even combining different resistance genes. Trees in these orchards would capture a broad range of resistance traits, helping future generations of elms stand strong against Dutch elm disease.
Program leaders Ryan Murphy and Ben Held also hope new technologies for studying genes will enable discovery of the genetic basis for resistance to DED. Identifying resistance genes or markers would make producing resistant trees a lot easier. It would also enable breeders to build up genetic diversity more deliberately.
SOURCES
Bernhardt, C. 2025. “Reviving a Giant” July 2025. University of Minnesota. Minnesota Invasive Terrestrial Plants and Pest Center. Website?
Knight, K.S., L.M. Haugen, C.C. Pinchot, P.G. Schaberg, & J.M. Slavicek. Undated. American elm (Ulmus americana) in restoration plantings: a review.
Wei, X.; Zhang, J.; Conrad, A.O.; Flower, C.E.; Pinchot, C.C.; Hayes-Plazolles, N.; Chen, Z.; Song, Z.; Fei, S.; Jin, J. Machine Learning-based Spectral and Spatial Analysis of Hyper-and multi-spectral Leaf Images for Dutch Elm Disease Detection and Resistance Screening. Artif. Intell. Agric. 2023, 10, 26–34. https://doi.org/10.1016/j.aiia.2023.09.003.
Posted by Faith Campbell
We welcome comments that supplement or correct factual information, suggest new approaches, or promote thoughtful consideration. We post comments that disagree with us — but not those we judge to be not civil or inflammatory.
For a detailed discussion of the policies and practices that have allowed these pests to enter and spread – and that do not promote effective restoration strategies – review the Fading Forests report at http://treeimprovement.utk.edu/FadingForests.htm
Or
































