Satellite records from 1999 to 2022 show the upper edge of continuous alpine plant cover rising in all six Himalayan regions studied. The movement is real and measurable, but the reasons behind it are subtler than melting snow opening new ground.
Above the last trees of the Himalaya the ground does not turn bare all at once. Dwarf juniper and rhododendron thin into grasses and cushion plants, then into lichen on scree, then into permanent snow and ice. Ecologists call the upper limit of continuous plant cover the vegetation line. A study published in Ecography in April 2026 has tracked its movement from Ladakh to Bhutan over 24 years, at a spatial resolution the authors describe as a first for the range.
Ruolin Leng and colleagues at the University of Exeter, working with the Swiss Federal Institute for Forest, Snow and Landscape Research and Tribhuvan University in Kathmandu, found that the line rose in every region between 1999 and 2022. The pace varied more than fivefold: 6.95 metres a year in Manthang, in the central Himalaya, 6.02 in Reckong, 4.60 in Ladakh, 2.40 in Ngari, 1.45 in Bhutan and 1.42 in Khumbu, the region around Mount Everest. Averaged across the six, the line climbed about 38 metres a decade, within the range reported for plant life on mountains elsewhere.
What is being measured
A vegetation line is defined by pixels, not by plants. Satellites cannot identify species; they record how strongly a patch of ground reflects red and near-infrared light, which is combined into the Normalised Difference Vegetation Index, or NDVI. The Exeter team calibrated an NDVI threshold of 0.22 against a two-metre land-cover map of the Khumbu region, at which roughly a tenth of a 30-metre Landsat pixel is covered by plants. The line is the elevation above which pixels fall below that threshold.
It therefore marks where plant cover is dense enough to be seen from orbit, and it sits below the highest elevations at which individual plants survive. Field surveys in Ladakh have recorded vascular plants at 6,150 metres; the satellite-derived median for the region is 5,625 metres.
The processing is demanding. From May to October the monsoon fills the high Himalaya with cloud, so the researchers took each year’s maximum NDVI from Landsat 5, 7 and 8 and corrected it with a phenology model that estimates the seasonal peak when clear images are missing. They masked cloud, snow and terrain shadow, restricted the analysis to ground above 4,000 metres, and located the line each year from the elevation histogram of vegetated pixels. A straight-line fit through the annual values gives each region’s rate of shift.
What upward movement means on the ground
Upward migration is easily misread as plants marching uphill. What happens is slower and less orderly. Seedlings establish where it was previously too cold, too dry or too snowbound for them, existing shrubs thicken, and warmth-loving species can spread among the cold-adapted plants that already occupy the alpine belt. The authors note the risk that such encroachment poses to native alpine species. Because a mountain narrows towards its summit and gives way to rock and ice, room for expansion is finite, and species already at the edge have nowhere higher to go.
What the climate data show
It is tempting to read the result as plants following warming uphill as snow retreats. The paper’s own numbers do not support so simple a story. Using ERA5 reanalysis data at roughly 30-kilometre resolution, the authors found no statistically significant trend in skin temperature in any region between 1999 and 2022. The only significant climate trend was a rise in annual precipitation in Manthang, about six millimetres a year. Trends in winter snow depth were small and not statistically significant anywhere.
Snow depth nevertheless features in the conclusions, but not as melt. Manthang, where the line rose fastest, showed the largest increase in snow depth and in precipitation. Khumbu and Bhutan, where it rose slowest, had weaker precipitation gains and more browning, meaning declining greenness, in 3.6 to 4.3 per cent of pixels against 0.5 to 2.1 per cent elsewhere. The authors suggest that increased snow accumulation may promote greening. Yet Reckong, the second-fastest region, saw snow depth decline slightly, so the pattern across six regions is suggestive rather than clean.
What the study can and cannot show
The study establishes, with a method tested against independent data, that satellite-detectable plant cover has advanced upslope across the range over a generation. The fit is stronger in some places than others: the trend explains two-thirds of the year-to-year variation in Manthang but only a fifth in Khumbu. The detection method is also less reliable in Bhutan, where the line it identifies falls below the upper limit of vegetated pixels.
The authors say they did not set out to pinpoint the drivers of greening and browning. The design cannot separate climate from grazing, land use or nutrient deposition, and the 30-kilometre climate grid may miss the microclimates that decide what grows on an individual slope. Method matters as well. An earlier study using a lower threshold and two snapshots, in 2000 and 2014, estimated shifts of 7 to 28 metres a year, far above these rates.
A signal, not a hazard
A rising vegetation line is not itself a danger. Its value is diagnostic. Plant cover changes how much water evaporates, how much snow is retained and how warm the soil becomes, so a persistent shift can alter the alpine water cycle that feeds downstream rivers, although this study did not measure those effects. It is also one of the few continuous records available at altitude, where weather stations are sparse and long-term ground measurements scarce, as the authors acknowledge.
A satellite indicator that behaves differently in Khumbu and Manthang is a reminder that the Himalaya is not changing as a single block. That regional variation, rather than any one average, is what later assessments of snow, glaciers and water supply will need to explain.
Amuluru G Manoj
