Glaciers worldwide are retreating rapidly1,2, with losses projected to continue throughout the twenty-first century3,4 and beyond5. Observations and projections assessed by the IPCC6 have primarily focused on changes in glacier mass and area, particularly in relation to sea-level rise7 and water availability8. In contrast, the evolution of the total number of glaciers has received comparatively little attention3, partly because glacier number is a less clearly defined metric and is influenced by observational limitations9. Yet knowing where and when individual glaciers will disappear is important from touristic, cultural and spiritual perspectives.
Glaciers attract millions of visitors each year, offering opportunities for recreation, education and outreach10,11,12. Many ski resorts also depend on glaciers, meaning their disappearance could affect winter tourism economies13,14. Beyond recreation, glaciers hold deep cultural, historical and symbolic importance. In many regions, they are iconic features tied to local traditions, spiritual practices and communal identity15,16,17,18. Across cultures, glaciers have also inspired stories, rituals and legends19,20. At the same time, even small glaciers can provide essential meltwater for communities and livelihoods, adding another dimension to their societal importance.
As glaciers shrink, communities are confronted with these changes, sometimes marking their loss with symbolic rituals, such as the ‘glacier funerals’ for Okjökull glacier (Iceland, 2019)21, Pizol glacier (Switzerland, 2019) (ref. 22) and Yala glacier (Nepal, 2025) (ref. 23). These ceremonies highlight the emotional and societal dimensions of glacier loss20,23,24,25. Iceland has even established a global glacier graveyard26, while initiatives such as the Global Glacier Casualty List aim to preserve the names and histories of vanishing glaciers21.
Here we offer a distinct perspective on glacier loss by quantifying the disappearance of each of the world’s more than 200,000 glaciers included in the global Randolph Glacier Inventory version 6.0 (RGI v.6.0)27, under four policy-relevant global warming scenarios by 2100 (relative to pre-industrial levels): +1.5 °C and +2.0 °C (Paris Agreement targets), +2.7 °C (current policy pledges) and +4.0 °C (a higher-emission pathway). Using three global glacier models, we introduce the concept of peak glacier extinction—the year in which the largest number of glaciers is projected to disappear between now (2025) and the end of the century. In this study, a glacier is classified as disappeared, or extinct, when either its projected area falls below 0.01 km2 (the standard inventory threshold9) or its remaining volume declines to less than 1% of its initial value. While this number-based framing provides an intuitive and complementary perspective to traditional mass loss metrics, it must be interpreted with care. Glacier number is sensitive to the inventories, which vary in resolution, completeness and treatment of small or fragmented ice bodies.
Our results show that glacier extinction (the number of individual glaciers disappearing) will peak around mid-century, with both its timing and magnitude depending on the warming level. Under +1.5 °C, global peak glacier extinction is projected to reach ~2,000 glaciers per year around 2041 (Fig. 1). Under +4.0 °C, this peak shifts to the mid-2050s and intensifies to ~4,000 per year. This delayed peak under higher warming reflects the associated longer and stronger glacier area and volume loss.
The points show the projected number of glaciers lost each year (values on the left y axis), according to the median year-scenario across all glacier and climate model combinations (Methods). The solid lines represent 11-year running means. The dotted lines indicate the cumulative percentage of glaciers lost since 2025 (values on the right y axis), with shaded bands showing the interquartile range (25th–75th percentiles). For clarity, shaded areas are shown only for the +1.5 °C and +2.7 °C scenarios. n indicates the total number of glaciers in 2025.
The timing of peak glacier extinction varies markedly across regions, reflecting differences in glacier size distributions (Fig. 2), inventories and response times (Extended Data Fig. 1). In regions dominated by small and rapidly responding glaciers, such as the Caucasus, the Subtropical Andes (Low Latitudes), North Asia and the European Alps (Central Europe), over 50% of the glaciers are projected to disappear within the next two decades (Figs. 1 and 2). As a result, peak glacier extinction in these regions occurs early, typically before or around 2040, and is largely insensitive to the warming level. The glaciers lost during this period are predominantly small and contribute minimally to the region’s total ice volume and area. In contrast, regions with a higher proportion of larger glaciers, including the ice sheet peripheries (Greenland and Antarctic/Sub-Antarctic), Svalbard and the Russian Arctic, exhibit a delayed peak in glacier extinction that occurs later in the twenty-first century or potentially even beyond in some regions such as Arctic Canada North. This latter region combines a large median glacier size with long response timescales (Extended Data Fig. 1). In these regions, the timing of peak extinction is also more sensitive to the warming level (Fig. 1), consistent with their slower dynamic responses and dominance of larger glaciers (Fig. 2). In some cases, it may also reflect inventory limitations, as small glaciers that are abundant in regions such as the Alps may be underrepresented in other regions9, delaying the apparent timing of peak extinction. High-mountain Asia hosts more than one third of all glaciers around the globe (~90,000 glaciers out of ~210,000 (ref. 27)), making it a key contributor to the global glacier distribution (Fig. 2). Due to the predominance of glaciers with intermediate sizes (Fig. 2), this region exhibits a distinct mid-century peak in glacier extinction, which is strongly reflected in the global pattern.
The pie charts show the number of glaciers that are expected to disappear in each global region. The colour gradient indicates when this loss is projected to occur: red hues denote earlier loss, and yellow hues represent later loss. The dashed lines mark the years 2050 and 2075. For clarity, the colours are grouped into five-year intervals. In the central world map, glacier locations at the inventory date are shown in blue. The histograms around the world map compare the number and size of glaciers in 2025 (white bars) with projections for the year 2100 under the same four global warming levels (coloured bars). The height of each bar indicates the total number of glaciers within the corresponding size interval. A solid black line highlights the height of the white bars, representing the present-day distribution. The numbers of glaciers in 2025 (black font) and 2100 (Shared Socio-economic Pathway (SSP) colours) under different warming levels are listed to the top right of each histogram. Basemap from Natural Earth (https://www.naturalearthdata.com) with glacier outlines from RGI v.6.0 (ref. 27).
At peak glacier extinction, between ~2,000 individual glaciers per year under +1.5 °C warming and ~4,000 per year under +4.0 °C are projected to disappear globally (Fig. 1). This rate is equivalent to losing the entire glacier population of the European Alps in just one year and represents a rate three to five times higher than the present-day modelled global loss of 750–800 glaciers annually (Fig. 1). Following this peak, the annual loss rate gradually declines, reaching 700–1,200 glaciers per year by the end of the century. However, this does not mark the end of glacier disappearance: substantial glacier mass loss is expected to continue beyond 21005,28, suggesting that many additional glaciers will disappear in the twenty-second century.
Regionally, maximum glacier extinction rates are influenced by both the present-day glacier number, particularly the abundance of small glaciers, and the warming level. In regions with relatively few glaciers, such as Iceland, peak annual loss remains limited to 5–10 glaciers. In contrast, Central Asia, which hosts the largest glacier population, currently loses 200–300 glaciers per year. This rate is projected to peak at ~500 per year under +1.5 °C warming, increasing to ~1,100 per year under +4.0 °C (Fig. 1). Expressed in relative terms, the maximum glacier extinction rates correspond to ~1.5% ± 0.7% of present-day glaciers per year in most regions, with lower values in high-latitude regions (for example, Antarctic and Sub-Antarctic, ~0.5%) and higher values in regions with many small glaciers (for example, Central Europe, ~3.3%).
While glaciers are expected to shrink significantly this century, many could still survive, especially if global warming is limited (Fig. 2). Under a +1.5 °C scenario, the projected number of glaciers lost over the next 20–30 years is roughly half of that expected under a +4.0 °C trajectory. By 2100, nearly 50% of today’s glaciers may remain (Figs. 1 and 2). In contrast, under current climate pledges (+2.7 °C of warming by 2100 (ref. 29)), annual glacier losses remain for a long time about 1,000 glaciers per year higher than for +1.5 °C warming, but lower than under the +4.0 °C pathway until around 2070. Beyond that point, glacier loss rates under +2.7 °C converge with those under +4.0 °C, ultimately leaving only ~20% of the initial glacier count by 2100 (Fig. 1). Under +4.0 °C, extinction rates remain high for decades, with 3,000 to 4,000 glaciers disappearing annually between 2035 and 2065. Fewer than 10% of present-day glaciers are projected to survive by 2100, even though larger glaciers may split up into several smaller ones while retreating (Fig. 1). These differences highlight the importance of mitigation: limiting warming to +1.5 °C could more than double the number of glaciers surviving by 2100 compared with +2.7 °C and prevent the near-complete loss expected under +4.0 °C warming, where fewer than 20,000 glaciers are projected to persist (Fig. 1 and Extended Data Table 1).
Under high warming, many regions are expected to become nearly glacier-free by 2100 (Fig. 2). For example, under +4.0 °C, only ~20 glaciers are projected to remain in Central Europe, a 99% reduction compared with the present day (Fig. 1 and Extended Data Table 1). Similar losses are expected in Western Canada and USA and in the Low Latitudes. In contrast, regions such as the Antarctic and Sub-Antarctic are projected to retain 47–64% of their glaciers under +4.0 °C and +2.7 °C warming, and up to 84% under +1.5 °C. Other regions, including Central Asia and Alaska, show a steep increase in glacier disappearance with higher warming, losing two to three times more glaciers at the peak extinction year under +4 °C than under +1.5 °C. These contrasts highlight how ambitious climate policy can make a substantial difference in preserving glaciers.
Although many glaciers are projected to persist through the end of the century, their number, size and distribution will vary substantially by region, initial glacier number and warming level (Fig. 2). Our results point to a systematic shift in the glacier size distribution, with an increasing dominance of very small glaciers and a decline in larger ones (Fig. 2). This trend primarily reflects the continued shrinkage of today’s larger glaciers, which, despite substantial mass loss, are still projected to survive by 2100 in diminished form. This may, in part, also reflect limitations in current inventories, which often struggle to detect and delineate small or debris-covered ice bodies and apply varying size thresholds9. Small glaciers are well captured in some regions but underrepresented in others, introducing inconsistencies in counts, and influencing both extinction timing and totals. The omission of glacierets and other very small ice bodies results in wide-ranging estimates of global glacier numbers, with a central estimate of ~435,000 (ref. 9). While this uncertainty influences our results, it predominantly concerns very small, minimally dynamic glaciers, which are expected to disappear rapidly. With our approach, we are confident in having a very comprehensive image of the number of glaciers that currently have a decent size (exceeding a threshold of 0.01 km2). Additionally, current glacier models do not simulate fragmentation. Each glacier evolves as a single unit, and the formation of smaller glaciers through splitting is not accounted for (Extended Data Fig. 2). Our analysis therefore tracks the number of disappearing glaciers as inventoried around the year 2000. Given that glacier number is shaped by observational limitations and subjective classification choices, it should be interpreted with greater caution than glacier area or mass, which are more directly and objectively measurable.
The introduction of peak glacier extinction reframes the narrative of glacier change, offering a metric that complements traditional measures of glacier mass and area loss. This peak loss of individual glaciers is more than a numerical milestone: it marks a turning point with profound implications for ecosystems, water resources and cultural heritage. As mourning rituals, memorials and glacier graveyards emerge, glacier loss turns from a scientific concern into a human story of vanishing landscapes, fading traditions and disrupted daily routines. This transition also highlights the urgent need for adaptation, particularly in regions dependent on meltwater from small glaciers, which are often the first to disappear.
In the context of the UN’s International Year of Glacier Preservation 2025 and the UN Decade of Action for Cryospheric Sciences 2025–2034, our results underscore the urgency of ambitious climate policy. An earlier peak in glacier extinction, as seen under low-warming scenarios, does not mean that losses are harmless but that fewer glaciers vanish than in a later and larger peak under high-warming scenarios. An earlier peak thus reflects a more optimistic trajectory. The timeline of peak glacier extinction is not yet decided. The difference between losing 2,000 and 4,000 glaciers per year by the middle of the century is determined by near-term policies and societal decisions taken today.