{"id":576,"date":"2026-10-05T12:47:03","date_gmt":"2026-10-05T12:47:03","guid":{"rendered":"https:\/\/meratch.com\/blog\/?p=576"},"modified":"2026-10-05T12:53:42","modified_gmt":"2026-10-05T12:53:42","slug":"a-catastrophic-glacier-collapse-in-nepal-shows-us-where-the-limits-of-our-adaptation-to-climate-change-lie","status":"publish","type":"post","link":"https:\/\/meratch.com\/blog\/2026\/10\/05\/a-catastrophic-glacier-collapse-in-nepal-shows-us-where-the-limits-of-our-adaptation-to-climate-change-lie\/","title":{"rendered":"A catastrophic glacier collapse in Nepal shows us where the limits of our adaptation to climate change lie"},"content":{"rendered":"\n

What 65 years of data say about the link between such events and climate change<\/em><\/p>\n\n\n\n

This article presents an analysis of long-term trends in temperature, precipitation, freezing-level height, snow cover, glaciers and glacial lakes across the Hindu Kush Himalaya and, separately, Nepal for 1961\u20132025, complemented by grid-cell trend maps and by an overlap analysis in glacierised terrain. The strongest and most robust signal is the shift from snow to rain at high elevations and the associated increase in days when heavy precipitation falls while the freezing level sits above the glaciers. By contrast, a trend in precipitation extremes themselves cannot be demonstrated from the available data. The results are placed in the context of the rock and ice collapse above the Lende Khola in Rasuwa District on 26 August 2026.<\/em><\/p>\n\n\n\n\n\n\n\n

The minute that changed a valley<\/strong><\/h2>\n\n\n\n

On 26 August 2026 a mass of rock and ice detached above the Lende Khola valley in Nepal’s Rasuwa District. According to the World Weather Attribution reconstruction it fell from about 5,150 m to the valley floor at roughly 3,750 m \u2014 a drop of some 1,400 m \u2014 and involved approximately two square kilometres of rock wall and glacier ice. The resulting debris flood reached Rasuwagadhi, 22 km downstream, in about seven minutes, an average speed of roughly 190 kilometres per hour.<\/p>\n\n\n\n

In the first days after the disaster, precise figures for the detachment area circulated, along with claims of two successive collapses. It was August, the height of the monsoon, and only twenty-three percent of the valley floor was cloud-free in both the pre- and post-event scenes, so the collapse volume cannot be quantified from open imagery. We found no support in the seismic record for the alleged second large collapse twenty-two minutes before the main event; the USGS, however, recorded a further signal with energy equivalent to a magnitude 4.2 earthquake about three hours after the main event, which itself was equivalent to magnitude 5.2. The trigger was not a cloudburst: precipitation in those days was weak. The USGS also notes that it remains unclear whether the initial failure was a landslide that entrained part of a glacier or a collapse of the glacier itself, so the mechanism cannot be treated as settled.<\/p>\n\n\n\n

Something else is more telling. The freezing level \u2014 the height at which temperature crosses zero \u2014 stood at about 5,900 metres in the days before the collapse, roughly 700 metres above the surrounding glaciers. It was raining on the ice in the source zone, if only lightly. It is precisely this combination, rather than torrential rain, that has changed most markedly at Himalayan altitudes in recent decades.<\/p>\n\n\n\n

How much energy the falling mass carried<\/strong><\/h3>\n\n\n\n

This is a scenario calculation, not a measurement. We cannot determine the collapse volume from our own data; for an area of about two square kilometres and a thickness of fifty to one hundred metres it amounts to roughly one hundred to two hundred million cubic metres. With a rock\u2013ice mixture density of 1,800\u20132,400 kg\/m\u00b3 and a fall of 1,400 m (5,150 to 3,750 m after WWA), this corresponds to a mass of 180\u2013480 million tonnes and a released potential energy of about 2.5\u20136.6 petajoules, or 0.6\u20131.6 megatonnes of TNT. Most of it is consumed by friction and fragmentation. The observed quantity is the average front speed of roughly 190 kilometres per hour over the 22 km reach.<\/p>\n\n\n\n

\"\"
Figure 1 |<\/strong> Nepal and its hazards. Top, the cryosphere: glaciers, glacial lakes coloured by area change between 2016\u201318 and 2023\u201325, recorded glacial lake outburst floods and rock\u2013ice avalanches. Bottom, landslides and floods reported in the BIPAD system, 2011\u20132025; circles mark events with fatalities. Shaded relief: GMTED2010.<\/figcaption><\/figure>\n\n\n\n

Data and methods<\/strong><\/h2>\n\n\n\n

Domains, periods and grids<\/strong><\/h3>\n\n\n\n

The domain is the Hindu Kush Himalaya as delineated by ICIMOD, the Himalaya after the polygon of Liu et al. (2022), and Nepal by its national boundary. The base period is 1961\u20132025 with 1961\u20131990 as the reference climatology. Trends are also computed for 1991\u20132025, which is more reliable in reanalysis because satellite observations enter the assimilation after 1979. Spatial analyses use a 0.25\u00b0 grid for reanalysis, 0.05\u00b0 for CHIRPS, 0.1\u00b0 for IMERG and 0.025\u00b0 for MODIS snow cover.<\/p>\n\n\n\n

Data sources<\/strong><\/h3>\n\n\n\n

Temperature and freezing-level height come from the ERA5 and ERA5-Land reanalyses, with Berkeley Earth and CRU TS 4.09 used as independent checks. ERA5-Land has a native resolution of about nine kilometres (0.1\u00b0); all analyses were carried out on a common 0.25\u00b0 grid onto which it was resampled, so the resolution quoted here is that of the analysis, not of the source. Precipitation is assessed from four independent sources simultaneously, because products diverge strongly in mountainous terrain. Glacier mass balance is taken from the WGMS database and glacial lakes from the HMA_GLI inventory, which we extended to 2025 through our own Landsat 8 tracking. Recorded hazards come from HMAGLOFDB, from the rock\u2013ice avalanche inventory of Zhong et al. (2024) and, for Nepal, from the BIPAD system.<\/p>\n\n\n\n

Statistical treatment<\/strong><\/h3>\n\n\n\n

Trends are estimated with the robust Theil\u2013Sen slope and tested with the non-parametric Mann\u2013Kendall test. Confidence intervals are 95 %; for area shares they are obtained by a one-degree block bootstrap that accounts for spatial correlation. For maps that test thousands of cells at once we add field significance following the Benjamini\u2013Hochberg procedure at a level of 0.10, as recommended by Wilks (2016). Cells with p < 0.05 are hatched, and area shares are weighted by the cosine of latitude.<\/p>\n\n\n\n

Indicators developed for this work<\/strong><\/h3>\n\n\n\n

Beyond standard ETCCDI indices we constructed three indicators closer to the mechanism by which hazards form. The first is the rain share of precipitation, computed as total precipitation minus snowfall in ERA5-Land. The second is the number of days on which precipitation exceeds the 95th percentile of wet days in the reference period while the daily freezing level sits above the area-weighted median glacier elevation of the cell \u2014 in other words, days when it rains on ice. The third is an overlap score: the number of six adverse indicators met simultaneously in a glacierised cell.<\/p>\n\n\n\n

Daily freezing-level heights were obtained from the ARCO-ERA5 archive on Google Cloud at 09 UTC, which corresponds to early afternoon in Nepal. Heights above ground were converted to elevations above sea level by adding the model orography. As a check, the trend of this daily series agrees with the independently processed monthly freezing-level series.<\/p>\n\n\n\n

Validation against observations<\/strong><\/h3>\n\n\n\n

Reanalysis is a model, not a measurement. A comparison with the high-altitude Pyramid station below Everest (5,035 m) shows that ERA5-Land is about five degrees Celsius too cold at that elevation and underestimates the number of days above freezing by roughly ninety-six per year. Its night-time trend, however, is almost exact: +0.29 \u00b0C per decade at the station against +0.31 \u00b0C in the model. Daily maxima diverge. We therefore refrain from interpreting absolute values in high mountains and work with trends.<\/p>\n\n\n\n

Results<\/h2>\n\n\n\n

Warming is fastest at night and at altitude<\/strong><\/h3>\n\n\n\n

Annual temperature across the Hindu Kush Himalaya has risen by 0.22\u20130.32 \u00b0C per decade since 1961 and by 0.29\u20130.42 \u00b0C since 1991, depending on the source. Global land warmed by 0.27 and 0.33 \u00b0C over the same periods. As a whole, then, the region is not warming exceptionally fast; what matters is how the change is distributed in space and time.<\/p>\n\n\n\n

Warming peaks between three and five kilometres and is carried mainly by night-time minima. Extremes change faster than the mean: the coldest night of the year warms three times faster than the warmest day (0.36 against 0.12 \u00b0C per decade), and warm nights have increased by seventeen per decade since 1991, significantly so across almost the entire region. Days above freezing increase by six to seven per decade between three and five kilometres, and the melt season above 4,000 metres lengthens by almost eight days per decade. Ice is therefore losing the nights during which it used to refreeze.<\/p>\n\n\n\n

The foothills and the Terai are an exception: the warmest day of the year is declining there, by 0.3 \u00b0C per decade below one thousand metres in Nepal. A similar phenomenon in northern India is attributed to aerosols, irrigation and cloudiness; confirming it would require station data from Nepal’s hydrometeorological service.<\/p>\n\n\n\n

\"\"
Figure 2 |<\/strong> Grid-cell trends in temperature extremes, 1961\u20132025, from ERA5-Land: melt days, frost days, warm days, the warmest day of the year, the coldest night of the year and melt-season length above 4,000 m. Hatching marks cells with p < 0.05.<\/figcaption><\/figure>\n\n\n\n

The freezing level is rising ever faster<\/strong><\/h3>\n\n\n\n

During the monsoon the freezing level over the Himalaya has risen by twenty-four metres per decade since 1961. Over shorter, more recent periods the rate increases: forty-nine metres since 1991, eighty-one since 2001 and one hundred and thirty-three since 2011. Comparing nested periods that share an end year is not in itself a test of acceleration, so we added two formal tests to the 1950\u20132025 series. The quadratic term in the regression is positive and highly significant (p < 0.001), and a piecewise linear model with a break in 1990 yields a slope change of +50 metres per decade for the Himalaya and +53 for Nepal (p < 0.001). The difference between Theil\u2013Sen slopes before and after 1990 is +52 and +63 metres per decade respectively, with 95 % confidence intervals of +26 to +77 and +34 to +86. The rise is therefore genuinely accelerating. The figure of about one hundred metres per decade quoted by WWA corresponds to the past fifteen to twenty years, not to the period since the mid-twentieth century.<\/p>\n\n\n\n

\"\"
Figure 3 |<\/strong> Monsoon freezing level (June\u2013September). Left, the anomaly relative to 1961\u20131990 for the Himalaya and Nepal; right, the trend as a function of start year with 95 % confidence intervals. All periods end in 2025.<\/figcaption><\/figure>\n\n\n\n

At altitude, it rains instead of snowing<\/strong><\/h3>\n\n\n\n

Above five thousand metres, slightly less than forty-eight percent of precipitation fell as rain in 1961\u20131990; in 2011\u20132025 the figure is fifty-five percent. The rain share is increasing across ninety-four percent of the area above three kilometres, significantly so on seventy percent. In glacierised cells it has risen by 2.6 percentage points per decade since 1991.<\/p>\n\n\n\n

Closer still to the mechanism is the count of days on which heavy precipitation coincides with a freezing level above the glaciers. Across the region these have increased from 3.3 to 4.1 per year and in Nepal from 4.5 to 7.1; since 1991 they rise in eighty-nine percent of Nepal’s glacierised cells. In the Karakoram the trend is close to zero. An independently constructed indicator that uses the modelled liquid fraction instead of the freezing level yields an almost identical series, which strengthens the result.<\/p>\n\n\n\n

Physically, rain on snow and ice acts in three ways: it delivers sensible heat, percolates into fractures and raises pore-water pressure, reducing slope stability. Laboratory experiments show that warming ice-filled rock joints towards the melting point sharply reduces their shear strength, and field measurements document water pressure within fractured permafrost. Individual disasters such as Kedarnath in 2013 or Melamchi in 2021 were complex cascades with several contributing factors, so we cite them not as proof of a single mechanism but as cases in which this factor was present. A check in our data illustrates it: on 16 June 2013 the Kedarnath cell received 129 mm of precipitation, ninety percent of it as rain, with the freezing level at about 5,360 m. At Chamoli in February 2021, by contrast, there was no precipitation at all and the freezing level lay below the glaciers, consistent with a mechanical trigger without a meteorological cause.<\/p>\n\n\n\n

\"\"
Figure 4 |<\/strong> Days with heavy precipitation while the freezing level sits above the glaciers, in glacierised cells. The maps show trends for 1961\u20132025 and 1991\u20132025 (hatching p < 0.05); the graphs show area means for the Hindu Kush Himalaya and for Nepal. Stars: Kedarnath, Chamoli, Melamchi and Rasuwa.<\/figcaption><\/figure>\n\n\n\n

Precipitation extremes: three products, three answers<\/strong><\/h3>\n\n\n\n

Monsoon totals show no change over 1961\u20132016 and, in Nepal, a slight decline. For the annual maximum daily total, three independent sources give three different answers: CHIRPS a decrease of 7.2 percent per decade for Nepal, the ERA5-Land reanalysis a slight increase, and satellite-based IMERG an increase of 9.4 percent per decade since 2001.<\/p>\n\n\n\n

Both strong signals most likely have a technical origin. CHIRPS steps down abruptly in Nepal around 1991, consistent with a change in input gauges, and IMERG jumps at the TRMM-to-GPM transition in 2014, when maxima rose by ten to twelve percent while annual totals rose only two. What is robust instead is the spatial pattern: the heaviest daily totals fall in the foothills below two kilometres and decrease with altitude, while at high elevations a single hour accounts for up to a third of the daily maximum. The dispute between products could be settled by the gauge-based APHRODITE dataset, which is available only on individual request.<\/p>\n\n\n\n

\"\"
Figure 5 |<\/strong> Trend in the annual maximum daily precipitation according to CHIRPS, ERA5-Land and IMERG, with a map of product agreement. The spread between products exceeds the signal itself.<\/figcaption><\/figure>\n\n\n\n

Glaciers, lakes and snow cover<\/strong><\/h3>\n\n\n\n

Nepal’s glaciers have lost 21.2 metres of water equivalent since 1976, roughly twenty-three metres of ice. The pace has accelerated in the past decade: the regional mean balance is \u22120.51 metres per year, equivalent to a loss of thirty-one gigatonnes annually against thirteen in 2010\u20132019. The Karakoram remains close to balance, yet its losses are deepening too.<\/p>\n\n\n\n

Between the 1990s and 2015\u201318 the number of glacial lakes grew from 1,223 to 1,778 and their area by fifty-four percent. Our own tracking shows the growth continuing: +3.2 percent across the region and +4.3 percent in Nepal between 2016\u201318 and 2023\u201325. This is a lower bound, because only lakes that already existed at the start of the period are tracked.<\/p>\n\n\n\n

MODIS snow observations for 2001\u20132025 show a shortening snow season in every region, but with twenty-five years and large interannual variability none of the regional trends is statistically significant. The clearest decline is above five thousand metres, with a median of minus nine days per decade in Nepal and a significant decrease over a quarter of that area.<\/p>\n\n\n\n

\"\"
Figure 6 |<\/strong> Mean annual mass balance of individual glaciers, 2016\u20132025, from the WGMS database. Symbol size scales with glacier area. Both the Karakoram anomaly and the heavy losses in the eastern Himalaya and south-eastern Tibet are visible.<\/figcaption><\/figure>\n\n\n\n

Hazard records are not measurements<\/strong><\/h3>\n\n\n\n

Recorded glacial lake outburst floods rose from eighteen to thirty per decade before 2000 to forty-five to forty-seven afterwards. A large share of the increase, however, consists of supraglacial lakes \u2014 the type that became reliably detectable only in the satellite era \u2014 while moraine-dammed outbursts, usually the most destructive, show no increase per decade. In Nepal, reported landslides have grown tenfold since 2011 while fatalities show no trend, which points to a change in reporting after the BIPAD system was launched.<\/p>\n\n\n\n

This asymmetry deserves to be taken seriously in any interpretation of inventories. The statement that ‘hazards are increasing’ is, for recorded events, partly a statement about the observing system rather than about nature.<\/p>\n\n\n\n

Where the changes coincide<\/strong><\/h3>\n\n\n\n

For every glacierised cell we counted how many of six adverse indicators are met simultaneously: a significant increase in melt days, in the rain share of precipitation and in rain-on-glacier days; a glacier mass balance below minus half a metre of water equivalent per year; a growing glacial lake; and a recorded outburst flood or rock\u2013ice avalanche within seventy-five kilometres.<\/p>\n\n\n\n

Nepal stands well above the regional average: fifty-one percent of its glacierised cells meet at least four indicators, against twelve percent across the Hindu Kush Himalaya as a whole. The difference is driven mainly by rain on glaciers. The darkest bands lie along the main Himalayan arc from western Nepal to Bhutan and across south-eastern Tibet, while the Karakoram stays light. Rasuwa lies within the dark band. The score is descriptive rather than a risk index: the indicators are not independent, hazard records are biased, and exposure of people and infrastructure is not included.<\/p>\n\n\n\n

\"\"
Figure 7 |<\/strong> Overlap of changes in glacierised cells, 1991\u20132025. The map gives the number of indicators met simultaneously; the panels show the share of cells for each indicator and the distribution of scores for the Hindu Kush Himalaya and Nepal.<\/figcaption><\/figure>\n\n\n\n

Discussion<\/strong><\/h2>\n\n\n\n

What can be said about attribution<\/strong><\/h3>\n\n\n\n

For an individual event, the question of whether climate change caused a particular slope collapse has an honest answer only in conditional form. Our data cannot establish whether the mass above the Lende Khola would have detached without warming. Nor does the World Weather Attribution study claim so. A single avalanche depends on geological structure, the fracture network, the ice within it and often on a local trigger that data at tens of kilometres resolution cannot capture.<\/p>\n\n\n\n

What can be said is that the environment in which such cascades form is changing in one direction, and that this change is unambiguous in the data. The freezing level is rising and its rise is accelerating; melt days above four thousand metres are increasing; the melt season is lengthening; rain is replacing snow at altitude; and glaciers are retreating at record rates. These are exactly the conditions that degrade the permafrost binding steep rock walls and that increase the volume of water available to debris flows.<\/p>\n\n\n\n

The missing link in the chain is direct measurement and modelling of permafrost temperature in rock walls. Without it, predicting which wall will fail next lies beyond this analysis. That is the natural next step and at the same time the main uncertainty to keep in mind when reading these results.<\/p>\n\n\n\n

The limits of adaptation<\/strong><\/h3>\n\n\n\n

Nepal operates a functioning flood early warning system built on rainfall and water-level stations, with lead times measured in hours. This collapse was simply too fast, and too unusual in its genesis, for such a system to detect in time: settlements twenty kilometres downstream had five to seven minutes. WWA reaches the same conclusion, stating that no existing early warning system could have provided sufficient lead time. Optical satellites are nearly blind during the monsoon and radar imagery arrives hours to days apart; the only source that captured the event in real time was seismic.<\/p>\n\n\n\n

Part of the risk therefore falls into the band of residual risk that existing measures do not cover. Design floods derived from historical rainfall records do not contain cascades of rock, ice and sediment, as Chamoli in 2021 and the Teesta-III dam in 2023 showed. What genuinely works is mostly slow and spatial: lowering hazardous lakes, as at Tsho Rolpa and Imja Tsho, keeping permanent structures off debris fans and out of narrow valley reaches, exchanging data across borders and preparing the population. The limits of adaptation here are not abstract: they are the physical limit of lead time and the design limits of particular structures. They should be named openly rather than hidden behind an engineered measure that creates a false sense of safety.<\/p>\n\n\n\n

Limitations<\/strong><\/h2>\n\n\n\n

ERA5-Land downscales temperature from the coarser ERA5 grid using a lapse rate, so it remains systematically biased in highly dissected terrain, and the partitioning of precipitation into rain and snow is model-based. Daily freezing-level heights come from a single time of day. Satellite records of precipitation and snow are short and sensitive to changes in the satellite constellation. Hazard inventories are not homogeneous in time. Finally, trends computed from 1961 are less certain in reanalysis than those from 1991, because the earlier period rests on a considerably sparser observing network.<\/p>\n\n\n\n

A chronicle of Himalayan cascades<\/strong><\/h2>\n\n\n\n
Date<\/strong><\/td>Place<\/strong><\/td>Event<\/strong><\/td><\/tr>
4 Aug 1985<\/td>Dig Tsho, Khumbu (Nepal)<\/td>A moraine-dammed lake burst, destroying a hydropower plant under construction and bridges downstream.<\/td><\/tr>
5 May 2012<\/td>Seti River, Annapurna (Nepal)<\/td>A rock\u2013ice avalanche turned into a debris flood; 72 dead and missing.<\/td><\/tr>


16\u201317 Jun 2013

<\/td>		Kedarnath, Uttarakhand (India)<\/td>Extreme monsoon rain and the outburst of Chorabari lake; thousands died.<\/td><\/tr>
25 Apr 2015<\/td>Langtang (Nepal)<\/td>The Gorkha earthquake released an ice and rock avalanche that wiped out a village; about 350 died.<\/td><\/tr>

5 Jul 2016<\/td>		Bhote Koshi (Tibet and Nepal)<\/td>The outburst of Gongbatongshacuo lake damaged the Araniko highway and a hydropower plant.<\/td><\/tr>
7 Feb 2021<\/td>Chamoli, Uttarakhand (India)<\/td>A rock\u2013ice avalanche destroyed two hydropower plants; more than 200 died.<\/td><\/tr>
15 Jun 2021<\/td>Melamchi (Nepal)<\/td>A debris flood damaged the main water supply conduit for Kathmandu.<\/td><\/tr>
4 Oct 2023<\/td>South Lhonak, Sikkim (India)<\/td>A glacial lake outburst destroyed the Teesta-III dam; dozens died.<\/td><\/tr>
16 Aug 2024<\/td>Thame, Khumbu (Nepal)<\/td>A lake outburst flooded part of the village; no fatalities.<\/td><\/tr>
26 Aug 2026<\/td>Rasuwa (Nepal)<\/td>Rock and ice collapse above the Lende Khola and a debris flood in the border valley.<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Conclusions<\/strong><\/h2>\n\n\n\n

Sixty-five years of data give a coherent picture. The Himalaya is not warming more dramatically than the rest of the world’s land, but it is warming where it matters most for slope stability: at night and between three and five kilometres. The freezing level is rising ever faster, rain is replacing snow at high elevations, and with it come more days when heavy precipitation falls directly onto ice. Glaciers are losing mass at record rates and glacial lakes keep growing.<\/p>\n\n\n\n

At the same time, not everything claimed about Himalayan disasters is supported by the data. Monsoon precipitation is not increasing and a trend in downpours cannot be demonstrated. The rising count of recorded events partly reflects better observation. Drawing this line between robust and unproven matters in a field where lives and the credibility of science are at stake. The robust signal is strong enough on its own to justify investment in glacial lake monitoring, early warning in exposed valleys, and permafrost temperature measurements in high rock walls.<\/p>\n\n\n\n

References<\/strong><\/h2>\n\n\n\n

Burrows, K., et al. (2023). Retrieval of monsoon landslide timings with Sentinel-1 reveals the effects of earthquakes and extreme rainfall. Geophysical Research Letters, 50, e2023GL104720. https:\/\/doi.org\/10.1029\/2023GL104720<\/p>\n\n\n\n

Dunham, A. M., et al. (2022). Topographic control on ground motions and landslides from the 2015 Gorkha earthquake. Geophysical Research Letters, 49, e2022GL098582. https:\/\/doi.org\/10.1029\/2022GL098582<\/p>\n\n\n\n

Dussaillant, I., et al. (2025). Annual mass change of the world’s glaciers from 1976 to 2024 (WGMS Annual Mass Change Estimates). Earth System Science Data, 17, 1977\u20132012. https:\/\/doi.org\/10.5194\/essd-17-1977-2025<\/p>\n\n\n\n

Funk, C., et al. (2015). The climate hazards infrared precipitation with stations \u2014 a new environmental record for monitoring extremes. Scientific Data, 2, 150066. https:\/\/doi.org\/10.1038\/sdata.2015.66<\/p>\n\n\n\n

Hall, D. K., & Riggs, G. A. (2021). MODIS\/Terra and MODIS\/Aqua Snow Cover Daily L3 Global 500 m Grid, Version 61. NASA NSIDC DAAC, Boulder.<\/p>\n\n\n\n

Harris, I., et al. (2020). Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Scientific Data, 7, 109. https:\/\/doi.org\/10.1038\/s41597-020-0453-3<\/p>\n\n\n\n

Hersbach, H., et al. (2020). The ERA5 global reanalysis. Quarterly Journal of the Royal Meteorological Society, 146, 1999\u20132049. https:\/\/doi.org\/10.1002\/qj.3803<\/p>\n\n\n\n

Huffman, G. J., et al. (2023). Integrated Multi-satellitE Retrievals for GPM (IMERG), version 07. NASA Goddard Space Flight Center, Greenbelt.<\/p>\n\n\n\n

Hugonnet, R., et al. (2021). Accelerated global glacier mass loss in the early twenty-first century. Nature, 592, 726\u2013731. https:\/\/doi.org\/10.1038\/s41586-021-03436-9<\/p>\n\n\n\n

ICIMOD (2020). Hindu Kush Himalaya region boundary. Regional Database System, ICIMOD, Kathmandu.<\/p>\n\n\n\n

Immerzeel, W. W., et al. (2020). Importance and vulnerability of the world’s water towers. Nature, 577, 364\u2013369. https:\/\/doi.org\/10.1038\/s41586-019-1822-y<\/p>\n\n\n\n

Kraaijenbrink, P. D. A., et al. (2017). Impact of a global temperature rise of 1.5 degrees Celsius on Asia’s glaciers. Nature, 549, 257\u2013260. https:\/\/doi.org\/10.1038\/nature23878<\/p>\n\n\n\n

Krajci, P., Holko, L., Perdigao, R. A. P., & Parajka, J. (2014). Estimation of regional snowline elevation (RSLE) from MODIS images for seasonally snow covered mountain basins. Journal of Hydrology, 519, 1769\u20131778. https:\/\/doi.org\/10.1016\/j.jhydrol.2014.08.064<\/p>\n\n\n\n

Mamot, P., et al. (2018). A temperature- and stress-controlled failure criterion for ice-filled permafrost rock joints. The Cryosphere, 12, 3333\u20133353. https:\/\/doi.org\/10.5194\/tc-12-3333-2018<\/p>\n\n\n\n

Munoz-Sabater, J., et al. (2021). ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. Earth System Science Data, 13, 4349\u20134383. https:\/\/doi.org\/10.5194\/essd-13-4349-2021<\/p>\n\n\n\n

Offer, M., et al. (2025). Pressurised water flow in fractured permafrost rocks revealed by borehole temperature, electrical resistivity tomography and piezometric pressure. The Cryosphere, 19, 485\u2013506. https:\/\/doi.org\/10.5194\/tc-19-485-2025<\/p>\n\n\n\n

Pepin, N. C., et al. (2022). Climate changes and their elevational patterns in the mountains of the world. Reviews of Geophysics, 60, e2020RG000730.<\/p>\n\n\n\n

Pepin, N., et al. (2015). Elevation-dependent warming in mountain regions of the world. Nature Climate Change, 5, 424\u2013430. https:\/\/doi.org\/10.1038\/nclimate2563<\/p>\n\n\n\n

Rohde, R. A., & Hausfather, Z. (2020). The Berkeley Earth Land\/Ocean Temperature Record. Earth System Science Data, 12, 3469\u20133479. https:\/\/doi.org\/10.5194\/essd-12-3469-2020<\/p>\n\n\n\n

Schneider, U., et al. (2022). GPCC Full Data Monthly Product Version 2022. Global Precipitation Climatology Centre, Deutscher Wetterdienst, Offenbach.<\/p>\n\n\n\n

Shugar, D. H., et al. (2020). Rapid worldwide growth of glacial lakes since 1990. Nature Climate Change, 10, 939\u2013945. https:\/\/doi.org\/10.1038\/s41558-020-0855-4<\/p>\n\n\n\n

Shugar, D. H., et al. (2021). A massive rock and ice avalanche caused the 2021 disaster at Chamoli, Indian Himalaya. Science, 373, 300\u2013306. https:\/\/doi.org\/10.1126\/science.abh4455<\/p>\n\n\n\n

USGS (2026). 2026 Nepal debris avalanche and flash flood. Landslide Hazards Program, U.S. Geological Survey.<\/p>\n\n\n\n

Wester, P., Mishra, A., Mukherji, A., & Shrestha, A. B. (eds.) (2019). The Hindu Kush Himalaya Assessment. Springer, Cham.<\/p>\n\n\n\n

Wilks, D. S. (2016). ‘The stippling shows statistically significant grid points’: how research results are routinely overstated and overinterpreted, and what to do about it. Bulletin of the American Meteorological Society, 97, 2263\u20132273. https:\/\/doi.org\/10.1175\/BAMS-D-15-00267.1<\/p>\n\n\n\n

World Weather Attribution (2026). Rapid warming in the Himalaya exacerbates geohazard cascades beyond adaptation limits. worldweatherattribution.org<\/p>\n\n\n\n

Zhong, Y., et al. (2024). Rock\u2013ice avalanches in High Mountain Asia: an inventory and its analysis. Geomorphology, 449, 109048. https:\/\/doi.org\/10.1016\/j.geomorph.2023.109048<\/p>\n\n\n\n

Appendix: datasets used<\/strong><\/h2>\n\n\n\n
Source<\/strong><\/td>Resolution<\/strong><\/td>Period<\/strong><\/td>Use<\/strong><\/td><\/tr>
ERA5<\/td>0.25\u00b0, monthly<\/td>1950\u20132025<\/td>temperature, freezing level, precipitation, snowfall<\/td><\/tr>
ERA5-Land (GEE)<\/td>0.25\u00b0, daily<\/td>1961\u20132025<\/td>temperature and precipitation extremes, rain share<\/td><\/tr>
ARCO-ERA5<\/td>0.25\u00b0, daily<\/td>1961\u20132025<\/td>daily freezing-level height (09 UTC)<\/td><\/tr>
Berkeley Earth, CRU TS 4.09<\/td>1\u00b0, monthly<\/td>1950\u20132025<\/td>cross-check of temperature trends<\/td><\/tr>
GPCC v2022, CHIRPS v2.0, IMERG V07<\/td>0.05\u20130.25\u00b0<\/td>1951\u20132025<\/td>precipitation and its extremes<\/td><\/tr>
MODIS MOD10A1, MYD10A<\/td>500 m, daily<\/td>2001\u20132025<\/td>snow cover and snowline<\/td><\/tr>
WGMS AMCE 2026<\/td>individual glaciers<\/td>1976\u20132025<\/td>glacier mass balance<\/td><\/tr>
HMA_GLI + Landsat 8<\/td>30 m<\/td>1990\u20132025<\/td>glacial lakes and their change<\/td><\/tr>
HMAGLOFDB v1.3.0, Zhong et al.<\/td>point records<\/td>1833\u20132025<\/td>outburst floods and rock\u2013ice avalanches<\/td><\/tr>
BIPAD<\/td>point records<\/td>2011\u20132025<\/td>landslides and floods in Nepal<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Scripts, index definitions, regional masks and result tables are kept in the project repository and are available on request together with dataset versions, so that the computations can be reproduced independently.<\/p>\n","protected":false},"excerpt":{"rendered":"

What 65 years of data say about the link between such events and climate change This article presents an analysis of long-term trends in temperature, precipitation, freezing-level height, snow cover, glaciers and glacial lakes across the Hindu Kush Himalaya and, separately, Nepal for 1961\u20132025, complemented by grid-cell trend maps and by an overlap analysis in … Continued<\/a><\/p>\n","protected":false},"author":5,"featured_media":590,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[17],"tags":[],"class_list":["post-576","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-flooding"],"_links":{"self":[{"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/posts\/576","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/comments?post=576"}],"version-history":[{"count":2,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/posts\/576\/revisions"}],"predecessor-version":[{"id":588,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/posts\/576\/revisions\/588"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/media\/590"}],"wp:attachment":[{"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/media?parent=576"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/categories?post=576"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/meratch.com\/blog\/wp-json\/wp\/v2\/tags?post=576"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}