| 10 Feb 2026
Glacier thinning causes warmer and drier regional climate at the Jostedalsbreen ice cap in western Norway
Abstract. Glacier recession gives rise to changes in land surface type and topography that are poorly represented in atmospheric models but may have important local impacts on climate. Implementing these changes in the Weather Research and Forecasting (WRF) model for the Jostedalsbreen ice cap in western Norway results in warmer and drier regional climate with less snow that can amplify glacier recession through a positive feedback effect. Most of the climatic response to glacier recession is related to the surface lowering associated with ice melt, resulting in reduced orographic lifting of moist air masses and higher surface pressure. The climatic response to glacier recession is largest where the ice melts but is also evident in adjacent valleys several kilometers away from the ice cap. While the warming by glacier recession amplifies effects of global warming, reduced precipitation counteracts the projected regional increase in precipitation. These findings should be included in estimates of glacier mass balance and have implications for agriculture, hydropower, tourism, and biodiversity around glacierised landscapes.
General comments
The manuscript presents a series of sensitivity experiments conducted with a kilometre-scale numerical weather prediction model with the aim of disentangling the effects of glacier loss on local climate. Changes in surface cover type and elevation associated with complete glacier loss are investigated as are combined effects for a more realistic glacier loss scenario. The effect of glacier lakes under a complete glacier loss scenario are also presented. The results indicate that the topographic changes make the larger difference to local and regional climate than surface type changes. This is through direct adiabatic effects of surface lowering as well as indirect effects on regional moisture flux and orographic processes. The introduction and study objectives are well set out. The model and simulations are adequately described. Appropriate model evaluation is made, though additional figures and tables could be included in supplementary material to aid closer interrogation of the results and increase the robustness of the study. The length and spatial resolution of the simulations provide a robust basis for interpretation and attempts are made to further understand how the effects vary with wind regime. The figures are of a good quality and quantity. The conclusions are supported by the results, though the discussion of the implications for future glacier climate interactions is, necessarily, qualitative.
There are some aspects of the methodology that could have implications for the results shown that need attention -namely whether the topography was additionally smoothed in some simulations and whether the NOAH land surface model adequately represents the energy and mass balances of glacier ice and snow surfaces.
A major limitation of the study is that is does not consider the effect of changing snow cover during winter, spring and summer that would accompany glacier retreat (only the extent of exposed glacier ice surfaces). Because of this, full effect of changing surface type on the local and regional climate has not been assessed. This should be discussed further in Section 4.
Overall, the study provides some useful results and with revisions to the text and additional details should make a good contribution to the literature surrounding glacier-climate interactions. There are more opportunities for analyses with the datasets, but perhaps these could come in a later manuscript that could also include explicit pseudo-global warming simulations to quantitatively compare the competing effects of changing glacier ice cover, topography and regional climate.
Line comments
13 – there are earlier studies looking at the effect of including glaciers in RCM simulations which is essentially the opposite comparison to what is done here. E.g.
https://link.springer.com/article/10.1007/s00382-009-0685-6
58 “is therefore representative for many other glacierised areas of the world.” This is perhaps overstated as the overarching climate (maritime, continental, polar, temperate, tropical) will have a large bearing on what aspects of glacier-topography interactions are important. This is mentioned at the end of the conclusions, but this statement should be tempered here and discussed further section 4.
67: “Higher model resolution will likely also result in limited additional scientific insight” agreed but this doesn't negate the fact that the finest scales aren't well represented and that simulations that include these scales may yield different results from the interactions that occur. Please revise.
85: “mild and wet winters and cool summers” some regionally average values here would be insightful to those not as familiar with the region (e.g., seasonal temperature range, total annual precipitation.
101: It is worth noting here that all the simulations are performed with the same boundary conditions i.e., there is no global warming induced regional changes in temperature or precipitation.
113: Please clarify if small glaciers appear in the default runs as well or just the Future glacier outlines run?
119: Did you consider running a 2100-future with lakes? Or will the lakes only form with full glacier retreat?
123: Some more detail on the land surface model is appropriate here (e.g. what is the albedo of glacier ice in the NOAH scheme? How is snow handled?
123: The Noah (and Noah-MP) model has known issues related to cold biases and poor representation of glaciers, snowpack and frozen ground. how might this impact the results of the study? Some further discussion is warranted – some papers of use may be:
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2010JD015139
https://www.sciencedirect.com/science/article/pii/S0165232X24000302
https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/2019JD030823
https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2025JD044230
https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2025JD044191
134: A table with station name, location, elevation, surface type, variables measured, measurement dates, % data gaps would be very useful to the reader.
154: It would be useful to indicate which valley was studied previously and how it relates to the AWS locations used here.
161: unclear and long sentence. Is the prevailing wind from the South? and how does this relate to the highest elevation and the SM and SB stations? Do you mean “in the lee of the prevailing southerly winds (as shown at the models highest elevation…)”?
166: “Modelled snow over the entire ice cap (not shown)” any reason for not showing this – it would be a useful addition to the paper especially given the focus on surface cover
170: I don't think you can say that temperature is well represented as biases are significant - 'adequately' perhaps given the terrain.
171: Figures A1 and A3 need to be in the main body of the manuscript and the mean values from Figure A2 in a table.
176: The best performance at the glacier station seems to be in winter with summer not performing as well. please revise
Figure 2: the extent of the ice surface is confusing between this figure and the next two (which use dots for ice loss between simulations). Consider using a different symbol in this figure (e.g., an x or cross) or a different symbol in Figures 3 and 4
183: “wind speed is too high” A table or figure showing histograms of wind speed for each site would be very useful for model evaluation. The wind roses are too small to compare directly. Overestimation of wind speed by NWP is well known, but it is useful to quantify, especially given the paper concerns the interaction of wind systems with surface type and topography.
191: Presumably ‘surface air temperature’ is meant here (as opposed to surface skin temperature that is also commonly called surface temperature). Please add ‘air’
Figure 3/4: The changes in temperature (and in some areas precipitation) outside of the glacier domains in the 2100-volume and no-ice-volume simulations indicate very similar patterns. Are these related to elevation changes in the model? e.g., were these surfaces smoothed further than the no-ice-surface and control simulations? If so, what are the implications for the results and other comparison? If not, what could be the mechanism for such a result?
Figure 3/4: The blue frame is exceptionally hard to see in the figures. It also makes the squares appear purple and thus not aligned with the colorbar scale in Figure 3. Consider zooming this panel in and/or using a similar symbol to the other panels to denote surface type change.
Figure 3/4: A distance scale on the figure would be useful when distances are referred to in the text
197: “likely mainly related to the albedo feedback” do you mean surface heating of bare surfaces not covered in snow? Please expand a little here.
212: “more moisture available for precipitation further inland” it would strengthen your argument if you could show this by comparing integrated moisture flux metrics for each simulation?
214: “The local pattern of changes in precipitation on the western side is similar to the changes in temperature…” as per first comment for figure 3/4 - are these related to changes in smoothed topography between the runs?
224: consider adding "TB' to figure 3a and 4a so the reader does not have to refer to figure 1.
228: it is worth highlighting here that the simulations only deal with the topographic/surface cover changes and you haven't done any psuedo-global warming simulations with different boundary conditions to assess the combined effect with regional climate changes.
230: “same emission scenario” do you mean the same emission scenario as the projected glacier outlines? please clarify.
234: “feedback effects” where are these discussed? Please expand
236: “that impacts by changes in land use are more than one order of magnitude smaller than those related to changes in elevation”. Agreed but this only takes into account changes in glacier ice and not the seasonal snowpack changes that lead to the loss in glacier ice. Changes to snowpack will have a large impact on winter and spring land surface-atmosphere feedbacks. This may also extend to summer in the higher elevation areas that currently generate a deep snowpack. Please revise.
239: “inclusion of new lake surfaces can still have a larger impact than ice surface removal through their impact on moisture fluxes”. This interpretation would come better after the results are shown. You also later discuss how the change in both temperature and precipitation is limited so this sentence seems a bit contradictory. Please revise.
254: “coupling would also strengthen the robustness of the findings”. Agreed this is one of the major limitations of the study that needs more discussion. Perhaps posit some hypotheses as to mechanisms that may change the direction or magnitude of the interactions for future studies to assess.
269-271: Confusing - expressing the changes as a fraction of the mean precipitation in given wind directions could help here. Or stating the mean values in the text, as they are hard to interpret from the figure.
270-271: are the mean changes for all wind directions statistically significant given the large scatter and small mean values of precip?
279: these results are for the no-ice-volume experiment, so are they relevant to a smaller ice cap? Please revise
281: “other sensitivity experiments” results for these need to be shown - perhaps in appendix or supporting material.
291: this sentence is a bit confusing - you say RCMS indicate precipitation increase in the previous sentence, but that RCM show negative trends in future snowfall here? Presumably this is due to temperature increase. If so, please note this is due to temperature or revise otherwise.
295: “Improvements in climate projections due to inclusion of glacier recession” will the glacier surface height changes be resolved in the RCM? Please discuss.
305: Would be good to discuss Salerno et al. (2023) here too.
Figure A1: needs station labels or a legend for easy reference
Figure A1: Side by side plots for the 6 AWS locations would be more helpful for model validation. Or histograms of wind speed.
Figure A2: what timescale is the MAE calculated over – from the figure it looks like it could be MAE of monthly means but perhaps it is MAE of daily or hourly means?
Figure A2: Figures A1-A3 should be in the main body of the manuscript. That being said, figure A2 hints at clear seasonal pattern that is better shown in Figure A3, so Figure A2 could be excluded and table with the bias and MAE at monthly and hourly timescales included.
Figure A3: is this mean bias error? or mean absolute error?
Editorial comments
52: “models are a potent”
88: here and throughout - please avoid using parentheses to denote alternative statements. write out in full i.e. "estimate an 2.8 °C increase in mean annual temperature and 10 % increase in precipitation when comparing 1991-2020 to 2071-2100 under a high emission scenario (SSP3-7.0) (Dyrrdal et al., 2025)”" or “estimate increases in mean annual temperature and precipitation of 2.8 °C and 10 %, respectively, when comparing 1991-2020 to 2071-2100 under a high emission scenario (SSP3-7.0) (Dyrrdal et al., 2025)”
101: as per comment for line 88
163: “Oldedalen is also associated”
Figure 2 caption: as per comment for line 88: “MG and JD are from 2021 and 2016-2019, respectively.”
179: sentence is too long - please split
Figure 3 caption: as per comment for line 88.
195: refer to the run name “When the ice surface is removed, but the elevation is unchanged (no-ice-surface run), changes in…”
200: refer to the run name “…corresponding surface-lowering (no-ice-volume run), changes…)
206: confusing sentence - do you mean “Along with the overall warming in the no-ice-volume run, the annual snow-to-rain ratio decreases over the ice cap, so that the reduction in snowfall is more than twice as large as the reduction in rainfall over the ice cap" ?
225: “at these locations” which other locations are used to construct the mean values detailed?
225: as per comment line 88 – “the change in temperature and precipitation is -1K and +9%, respectively. However, the signal....”
259: “…from between the southeast…”
265: as per line 88 comments: "-1.5 and -2.6 mm on days with southerlies and westerlies, respectively
268: as per line 88 comment
286: use ‘temperature’ in place of ‘air’ as per line 191