In September 2026, University of Utah scientists using TROPOMI saw Western US forest fluorescence fade two years before bark-beetle mortality showed in aerial surveys.

A satellite instrument built to track air pollution has turned out to be a useful early-warning sensor for dying forests. In work covered on 13 September 2026 by Scienmag and published earlier that month by the University of Utah, a Utah-led team reported that solar-induced chlorophyll fluorescence measured by TROPOMI — the Tropospheric Monitoring Instrument aboard Europe’s Copernicus Sentinel-5P — faded in Western U.S. conifer stands roughly two years before USDA Forest Service aerial detection surveys logged bark-beetle mortality in the same places.

The study, led by Lewis Kunik and published in Remote Sensing of Environment (DOI 10.1016/j.rse.2026.115550), is not a forecasting product and does not claim to name the trees that will die. What it does claim is narrower and, for land managers, still consequential: at landscape scale, a physiological signal invisible to the eye and largely invisible to conventional greenness indices shifted first, and it shifted in a way that drought conditions alone did not account for.

The faint red glow that greenness indices miss

Solar-induced chlorophyll fluorescence, or SIF, is a byproduct of photosynthesis. When chlorophyll absorbs sunlight, a small fraction of that energy is re-emitted as a faint red glow — a signal far too weak to see from the ground, but detectable from orbit by spectrometers sensitive enough to resolve it against reflected sunlight. Crucially, the glow tracks how efficiently a plant is actually using the light it absorbs. When trees are stressed, that efficiency drops, and the glow dims.

That distinction matters for evergreens in particular. Pines, spruces and firs can hold onto their needles while effectively dormant, so canopy color and structure stay broadly intact even as the stand’s photosynthetic machinery slows. Indices built on greenness or canopy geometry — NDVI among them — therefore register a forest that still looks like a forest. The Utah team reports that SIF, and a derived physiological metric the paper calls SIFyield, responded earlier and more sensitively than the other canopy products tested, including land surface temperature and NDVI.

TROPOMI was chosen for coverage and sampling frequency rather than fine detail. The instrument was designed for atmospheric chemistry, not vegetation, but its wide swath and near-daily revisits produce the kind of repeat time series that a multi-year physiological trend requires. The paper’s analysis works with TROPOMI SIF gridded at 0.05° — roughly 5 km — combined with MODIS vegetation products and tree-mortality grids for the western United States covering 2018 to 2023. At that resolution, each measurement is a stand-scale average, not a tree.

A two-year lead, and a 10–20% gap against drought-matched controls

The comparison design is what keeps the drought explanation from swallowing the result. The researchers matched disturbed areas against biogeographically similar control areas that saw little wildfire or bark-beetle mortality — the University of Utah release cites controls spanning 2011 to 2023 — and used bootstrapping to test whether the differences held up. Those control forests experienced comparable drought. Their SIF still declined, as drought-stressed vegetation does. But according to the paper, the decline was 10–20% less severe than in the stands that beetles later infested.

In other words, dry conditions were part of the story everywhere, and something additional was happening in the stands that would go on to die. Across moderate-to-severe bark-beetle mortality, the paper’s abstract reports growing-season SIF dropping to 60–70% of pre-drought levels and staying there over multiple years.

Wildfire served as the validation testbed, for a practical reason: burn severity can be quantified far more cleanly than beetle activity, which spreads unevenly, over years, and is often only catalogued once crowns turn red. Against fire, SIF declines scaled with the amount of vegetation lost, falling as low as 20% of pre-fire levels in the most heavily affected areas. The paper describes wildfire effects as more predictable than those from beetle mortality — which is precisely why fire made a reasonable yardstick before the same method was pointed at insects.

The author list reflects that mix of remote sensing and forest-health expertise: Kunik, who recently completed his doctorate at the University of Utah under the joint supervision of John Lin in Atmospheric Sciences and David Bowling in the School of Biological Sciences, worked with Brett Raczka (Utah and NCAR), Jeffrey Hicke (University of Idaho), Christian Frankenberg (Caltech), Rui Cheng (Claremont McKenna College) and Michèle Slaton of the Inyo National Forest, USDA Forest Service. Funding came from NASA’s Carbon Monitoring System, an NSF Graduate Research Fellowship, Utah’s Wilkes Center for Climate Science & Policy and a USDA Forest Service agreement.

What the signal is not

It is worth being blunt about the limits, and the authors are. As described in the Utah coverage, Kunik frames the objective not as predicting which individual tree will die, but as flagging areas of concern early enough that managers can investigate on the ground, mobilize crews, allocate funding or otherwise prepare before mortality becomes widespread. That is a triage signal, not a diagnosis.

SIF also moves for many reasons. Drought, insects, canopy dieback, phenological timing, illumination geometry and understory composition can all push the number around. The team accounted for a range of these factors, but attributing a single year’s SIF change to one cause remains difficult. The paper reports a consistent pattern across this analysis; it does not establish a universal detector, and nothing here amounts to an operational system or a Forest Service deployment.

The carbon framing is the authors’ own forward look rather than a settled finding. Because SIF functions as a fingerprint of plant CO2 uptake at regional and global scales, they argue it could help track whether repeated disturbance weakens the uptake capacity of Western forests — and, over time, whether those forests tip from carbon sink toward source. Lin points to ESA’s FLEX mission as the next step, bringing higher-spatial-resolution fluorescence measurements than TROPOMI can deliver today.

For now, the claim is deliberately modest: a coarse, satellite-derived physiological signal that dimmed in Western conifer stands roughly two years before aerial surveyors recorded the dead trees — a landscape-scale flag worth investigating, not a map of which trees are next.