Coastal ice domes in West Antarctica preserve snowfall records that reflect past climate conditions. Using weather reanalysis from 1979 to 2022, this study identifies which domes best capture different climate drivers affecting the region. Western sites respond mainly to hemisphere-wide wind shifts, while eastern sites reflect regional storm patterns. These results guide where future ice cores should be drilled to reconstruct past atmospheric and oceanic changes in this vulnerable region.
The West Antarctic Ice Sheet (WAIS) is a dynamic system where interactions between ice, ocean, and atmosphere drive significant ice mass loss, raising concerns of irreversible retreat and sea-level rise. Long-term observational records of variability and change along the WAIS coast are largely restricted to satellite observations, but more direct observations are needed, given this region's present and future societal impact. Coastal ice rises, grounded domes of ice embedded in or along the margins of ice shelves, preserve in their accumulated snowfall high-resolution records of past climate variability that can be recovered by ice coring. These potential ice core sites offer unique opportunities to reconstruct key drivers of regional change, including modes of atmosphere-ocean variability described by the Southern Annular Mode (SAM), the Amundsen Sea Low (ASL), and El Niño–Southern Oscillation (ENSO) – and warrant exploration via climate reanalysis to assess the current relative balance of climate controls at any potential ice core site. This study uses ERA5 and MERRA-2 reanalysis to evaluate the climate controls on interannual snowfall variability at thirteen WAIS coastal ice rises over the satellite era (1979–2022). Results highlight longitudinal differences in how interannual snowfall variability at coastal ice rises is influenced by SAM, ENSO, and Bellingshausen Sea atmospheric pressure anomalies. Snow accumulation (precipitation) as resolved in atmospheric reanalysis suggests that, as potential ice core sites, Dean Island and Guest Peninsula, located in the West Sector of the WAIS coast, are strongly influenced by broad Southern Hemisphere westerly wind anomalies suppressing local precipitation, making them ideal for isolating this mode of variability in paleoclimate reconstructions. In contrast, Farwell Island in the East Sector exhibits a positive relationship between precipitation and cyclonic activity associated with Bellingshausen Sea pressure variability, making it a key site for reconstructing the influence of synoptic-scale pressure systems on coastal accumulation in this region. These findings inform future ice core studies aimed at understanding WAIS climate dynamics, with implications for projections of Antarctic stability and global sea-level rise.
Cliff erosion in periglacial coastal environments is governed by complex interactions between thermal, hydrological, and marine processes, often resulting in non-linear and spatially heterogeneous behaviour. However, most existing studies rely on linear models that may not adequately capture threshold responses and episodic dynamics, particularly in data-scarce polar regions such as Antarctica. This study investigates long-term cliff erosion patterns in a periglacial coastal setting by applying linear and non-linear statistical modelling to a multitemporal dataset. The analysis focuses on the coastal bluffs of Port Foster, Deception Island (South Shetland Islands, Antarctica), using a unique 66 year record (1956–2022) derived from historical aerial imagery and high-resolution satellite data. Shoreline positions were extracted through photogrammetric processing and analysed using a transect-based framework, incorporating both linear and non-linear (quadratic and sigmoidal) least squares regression models with uncertainty-weighted parameters. Results reveal a transition from relatively stable conditions to accelerated retreat after ∼2000, with maximum change rates reaching up to 5 m yr−1
Understanding the controls on seasonal velocity change for tidewater glaciers may provide insight into long-term retreat and acceleration. Leveraging recent high-resolution satellite data, we examine changes in surface elevation, velocity, and terminus position for four glaciers in Central Western Greenland over 2015–2021. Our approach uses a simplified force balance focused at the terminus to model the expected response in upstream velocity caused by the observed terminus changes. We find that seasonal velocities are strongly controlled by terminus advance/retreat for two glaciers. Residuals between modeled and observed velocities reveal two distinct signals: summertime pulses coincident with peak runoff and wintertime speedup extending several kilometers inland of the terminus. We evaluate the sensitivity of terminus-driven velocity to elevation change by incorporating observed seasonally varying surface topography and applying controlled modifications to the profile, specifically uniform vertical shifts and variations in surface slope. We find surface slope changes impact velocity response to terminus changes more than spatially uniform changes in elevation. Increased surface slope amplifies velocity response to terminus changes. While simplified, our model could be applied to other glaciers to assess the importance of terminus position change as a driver of seasonal velocity.
In Greenland, mass wasting through frontal ablation occurs not only at the ice-ocean interface but also at the ice-lake intersection. Recent studies have found that lakes cover 10 % of the entire ice sheet margin and stress the importance of understanding frontal dynamics in lacustrine settings. However, relatively little is known about how lake-terminating glaciers compare to marine-terminating glaciers under the same climatic conditions. At a unique study site in South Greenland, a lake and a marine terminus are part of the same glacier system (Qooqqup Sermia), subject to the same regional climate forcings and fed by the same upstream ice masses. In this study, we analyse the drivers of change at both glacier fronts and compare their dynamics with a comprehensive remote sensing dataset supported by climate model output. Furthermore, during two field campaigns, we collected lake bathymetry data alongside temperature and lake level measurements. We find that despite being subject to the same regional climate forcing and fed by the same upstream ice masses, the two termini show contrasting front dynamics in the long- and short-term. We infer extremely low subaqueous melt rates in the lake as one of the main differences between the two environments, likely contributing to the observed contrast in dynamics. A massive disintegration event of more than 3 km of the lake terminus showcases the possibility of rapid mass loss at lake-terminating glaciers in Greenland. Our results stress that lake- and marine-terminating glaciers require separate estimates of frontal ablation through subaqueous melt when included in model simulations of the Greenland Ice Sheet.
The current pace of glacier retreat in the European Alps is unprecedented in the observational record and has significant implications for water resources and downstream ecosystems. Quantifying the future evolution of these systems requires physically based glacier models that are calibrated against observational data. Using the open-source Framework for assimilating Remote-sensing Observations for Surface mass balance Tuning (FROST), we infer mean Equilibrium Line Altitudes (ELAs) and other surface Mass Balance (SMB) parameters for 409 Alpine glaciers for the time period 2000–2019 using an Ensemble Kalman Filter. The method combines an elevation-dependent SMB model with ice dynamics from the Instructed Glacier Model (IGM). Validation against ELA estimates from in-situ measurements and end-of-summer snowline data shows good agreement, with Pearson correlation coefficients of r=0.74r=0.64, respectively. These results demonstrate that FROST
Understanding the mechanisms of Antarctic sea ice variability, as well as its predictability, remains a central challenge in climate modelling due to the sparseness of observations and the complex processes involved. This study assesses how incorporating sea ice thickness (SIT) observations can improve the reanalysis and prediction skills of Antarctic sea ice over a period long enough to yield robust conclusions. Two 30-year reanalyses are produced using the Norwegian Climate Prediction Model (NorCPM), with and without LEGOS SIT assimilation, and they are used to initialise year-long hindcasts from 1995–2022 beginning in January, April, July, and October. Assimilation of SIT observations improved estimates of Antarctic sea ice trends, seasonal cycle, and interannual variability – particularly in the West Antarctic and the West Pacific, where sea ice is thick and LEGOS SIT is reliable. The integrated ice edge error (IIEE) was also reduced in the reanalysis during the austral winter and spring, but a degradation was observed during the austral summer. Hindcasts revealed a long SIT memory, with October initialisation resulting in substantial sea ice extent (SIE) skill gains up to 12 months and January initialisation extending prediction skill by 2–3 months in the pan-Antarctic, with strong improvement in the Weddell Sea and the Amundsen–Bellingshausen Seas. The SIE and SIT prediction skill was also improved in the West Pacific during the austral summer and autumn, a region that previously posed a challenge for prediction skill. We show that SIT observations are important for improving Antarctic SIE predictions, especially for minimum SIE forecasts in the austral summer and at longer lead times.
This study quantifies uncertainties of the observed Arctic and Antarctic Sea-Ice Area (SIA). Uncertainties in SIA estimates are derived from a single product, using a refined method to propagate local sea-ice concentration (SIC) uncertainties to hemispheric SIA estimates. The method accounts for spatial and temporal error correlations. The SIA uncertainty time-series based on the EUMETSAT Ocean and Sea Ice Satellite Application Facility (OSI SAF) SIC record is relatively stable over time, even though SIA itself shows notable seasonal and long-term variability. The seasonal cycle of the uncertainty is instead linked largely to the distribution of the ice. In the growing season, the SIC fields are more compact with a shorter sea-ice edge separating high and low sea ice concentrations. In the melting season the sea-ice edge is in comparison more diffuse. This seasonal evolution of the sea-ice edge leads to a relatively large SIA uncertainty in the melting season and a smaller uncertainty in the growing season.The new single-product time series is compared with the spread across several SIA satellite products. The spread in the latter is characterized by seasonally varying biases. After accounting for these biases, the remaining differences are consistent with our new single-product SIA uncertainty. The two approaches are complementary: The inter-product approach provides insights into the influence of the product development while the new single-product SIA uncertainty allows for dynamic daily and monthly estimates which do not rely on the selection and availability of other products. It represents the non-systematic (bias-free) component of the uncertainties, which, among other things, determines the significance of new SIA extremes.The single-product, non-systematic uncertainties in SIA trends from 1979 to 2025 are estimated to be 11⋅103 km214⋅103 km2
Daily, basin-scale snow depth maps are needed for forecasting and operations, yet airborne lidar typically provides only episodic snapshots. We present a portable relative-depth machine-learning framework that converts a small number of lidar acquisitions plus a single daily driver time series (in-situ station or ERA5-Land) into temporally coherent, per-pixel daily snow depth maps. A random forest model is trained on lidar–driver differences where lidar supplies the spatial pattern of departures and the driver supplies temporal evolution; learning is constrained to observed conditions using a valid-pixel mask and synthetic zero-depth maps at season start and end. We evaluate the approach in two contrasting regimes – Mores Creek, Idaho, and Hubbard Brook, New Hampshire – using multi-year lidar records. Across both basins, performance is fit for purpose (R2
The future behavior of the Antarctic Ice Sheet is considered to be one of the largest uncertainties in global climate projections, with its stability fundamentally governed by grounding-zone processes, bed geometry, and sensitivity to oceanic forcing. However, observational records only reflect a short moment when considering the length of a full cycle of ice sheet expansion and retreat. Therefore, paleo-data present a valuable extension to the observational period. East Antarctica's deglaciation history remains largely understudied compared to the West Antarctic margin. This emphasizes the urgent need for reliable long-term spatiotemporal data on ice sheet change, particularly for sectors that play key roles in supplying the world's oceans with dense bottom water. In this study, we performed a multi-proxy analysis on geophysical and geological data recovered from two prominent glacial cross-shelf troughs on the Mac. Robertson continental shelf. We classified submarine glacial landforms on the continental shelf along both troughs from combined multibeam swath bathymetry and sub-bottom profiler data to infer past grounding line extent and the pattern of subsequent grounding line retreat. Additionally, combined sedimentological, sediment-physical, and geochemical analyses, including foraminifer radiocarbon dating, reveal the style and timing of this retreat across the shelf. Our study concludes that grounded ice reached the Mac. Robertson continental shelf break until just before ∼12.7 cal. ka BP, hence preventing the formation of Dense Shelf Water (DSW) in its current form. We therefore infer a different formation mechanism for DSW as an important precursor of Antarctic Bottom Water under such full glacial conditions before continued grounding line retreat exposed the middle continental shelf by ∼10.8 cal. ka BP and set the stage for more modern-like conditions.
The Foundation-Patuxent-Academy system (FPAS) is a major Antarctic ice stream system, draining both East and West Antarctica, with a global sea level potential of ∼3 m. We provide a holistic catchment-scale overview of the FPAS reviewing its glaciological and hydrological systems, its glacial history, and its modelled response to past and future climate change. FPAS may be vulnerable to future change because of: (i) a deep (∼2.4 km below sea level) low-gradient retrograde bed that encourages grounding-zone retreat; (ii) a low-gradient ice surface and high tidal range, which are likely to promote flotation of grounded ice and seawater intrusion; (iii) an active and dynamic subglacial hydrological system; (iv) complex ice-meltwater-ocean interactions at the grounding zone; (v) potential for substantive expansion of the across-flow length – and cross sectional area – of the grounding zone; and (vi) susceptibility to ice flow-switching and water piracy (e.g. via the adjacent Support Force Glacier). Despite such potential vulnerabilities, existing numerical model simulations of FPAS grounding-zone retreat produce a wide and divergent range of past and future scenarios. Uncertainties in the future response of the FPAS to a warming climate result from poor constraints on its topography and hydrology, processes of ice-ocean interaction, interlinkages with the surrounding ice sheet and ice shelf, and a shortage of FPAS-specific modelling experiments. This review outlines and evaluates these critical gaps in our knowledge of the FPAS and develops a strategy to address them. This strategy would provide: (i) the first robust and comprehensive evaluation of the FPAS's vulnerability to current and near-future climate
Landfast ice in Alaska is experiencing rapid changes in extent and duration, impacting the safety and utility of the ice for Arctic coastal communities. Current datasets of landfast ice only distinguish landfast ice from mobile pack ice, omitting crucial information regarding the relative safety of the ice. InSAR (Interferometric Synthetic Aperture Radar) holds promise for identification of landfast ice and measurement of centimeter-scale deformation from a spaceborne sensor. We use two InSAR-derived properties: coherence to identify areas of landfast ice, and the interferometric phase gradient to approximate a new metric called apparent strain (ϵa), which acts as a proxy for estimating the relative stability. Apparent strain is defined as the horizontal gradient of interferometric phase in the line-of-sight displacement. We built on a previous study Dammann et al.2019), by assigning quantitative apparent strain values to identify three distinct stability classifications of landfast ice: Bottomfast (
At least since we began measuring in detail, the West Antarctic Ice Sheet has lost a lot of ice, but we don't know precisely how important climate change is in this. Here, we put a number on the role of climate change in retreat of a glacier in this ice sheet, for the first time. We show that climate change made the shrinking of this glacier much worse. Our work also suggests that what happened on very long timescales (the last 10,000 years) might also matter for retreat of the ice sheets today.
The West Antarctic Ice Sheet (WAIS) has undergone rapid change over the satellite era, characterized by significant thinning, grounding-line retreat, and mass loss. More than a third of the ice loss from this region is from Pine Island Glacier (PIG). However, robust causal links between anthropogenic climate change and PIG ice loss have yet to be established. Here we attempt to quantify the role of anthropogenic climate change in observed retreat of PIG over the 20th century and how this may evolve up to 2200. To do so, we use an ensemble Kalman inversion data assimilation method embedded into an uncertainty quantification framework, called calibrate-emulate-sample (CES). This procedure, which assimilates observations of grounding-line retreat and ice volume, yields observationally constrained probability distributions of both model and climate forcing parameters. Our analysis suggests that it is unlikely that the extent of 20th century PIG retreat would have taken place without anthropogenically driven trends in ice-sheet forcing and that anthropogenic forcing exacerbated the extent of PIG retreat over the 20th century, by approximately 18 %. These results are, importantly, conditional on our choice of initial state. For our chosen initial state, we find that the parameter combinations compatible with these observational constraints require PIG to lose mass (but not experience grounding-line retreat) over the entire simulated period since 1750, not just after the 1940s when grounding-line retreat was initiated. This preconditioned ice mass loss introduces significant uncertainty into our quantification of 20th century forcing contributions. In simulations with no anthropogenic trend in forcing, we still observe significant retreat; this may result either from a larger-than-actual initial state, or may suggest that the earlier ice state preconditioned the industrial era retreat, possibly implicating longer term changes to WAIS in the present retreat.
The retreat of Greenlandic glaciers through calving has major implications for the ice sheet's mass balance and future sea-level rise contributions. Despite its importance, the implementation of calving in ice sheet models remains contested, with several calving laws suggested to parametrise this process. While the performance of some of these calving laws has been tested for Antarctic ice shelves and Greenland's grounded outlet glaciers, it is unclear which calving law would best capture the observed behaviour of Greenland's ice shelves. Petermann, Ryder, and Nioghalvfjerdsbræ (79N) glaciers terminate as Greenland's three largest ice shelves, accounting for 90 % of the remaining floating ice and buttressing ∼ 15 % of the ice sheet's mass. Here we build on other systematic calving studies by comparing five calving laws at Greenland's three largest ice shelves using the Ice-sheet and Sea-level System Model (ISSM). We begin by constraining the performance of each law against observed terminus fluctuations between 2008 and 2024, and continue with projections to 2300 under various climate forcings. When evaluated against observed terminus changes, we recommend the use of a von Mises or Crevasse Depth calving law owing to their consistent performance and similar tuning parameters across the three ice shelves. However, in our extended projection runs, we find that calving parametrisations have little influence on grounding line discharge rates, which are instead driven by the choice of climate forcings. Large ice shelf calving or collapse events are scarce, and only in these rare cases do we find any pronounced increase in ice discharge. Our results indicate either continued buttressing potential from Greenland's ice shelves into the coming centuries or fundamental flaws in the current set of calving laws when calibrated to contemporary ice-shelf behaviour.
Winter Arctic polynyas, openings in the pack ice, play a crucial role for the climate from sea ice production to cloud formation and are hotspots for the ecosystem and human activity. Their area has significantly increased since satellite records began. Yet their representation has yet to be evaluated in any generation of global climate models, most likely because their automatic multi-model retrieval is challenging. We here use a newly-developed machine-learning based method and evaluate polynya activity against the satellite-derived one over 1979–2024 in the 18 CMIP6 models with daily sea ice concentration available. We find that models overestimate winter Arctic polynya area but underestimate its frequency, and limit their opening to the seasonally ice-covered regions. Polynya area is increasing in most models, but the bias of these trends are inconsistent. Although the model with the highest resolution has both the highest areas and frequencies, the sea ice model component is a more robust predictor of polynya activity, with most activity in the models whose thermodynamics scheme enhances ice growth/melt. Accordingly, we found more polynyas in the models with a larger seasonal cycle, in particular those with a warmer autumn that would delay ice growth or melt early. Finally, we confirm preliminary findings that polynya activity does not seem to impact the representation of the water column; if anything, we find less-dense water at the bottom of the continental shelf following larger polynya activity. Overall, our results suggest that in the Arctic, CMIP6 models unrealistically open only sensible-heat polynyas.
As Arctic amplification accelerates glacier retreat, new dynamic landscapes are emerging at the interface of terrestrial and marine systems. This study identifies and analyses a distinct coastal landform: the Paraglacial Moraine Lagoon (PML). Formed by coastal barriers composed of terminal or lateral moraines deposited during the Little Ice Age, PMLs represent a critical yet understudied component of the glacier–climate change feedback system. Using a multi-decadal record (1936–2024) comprising aerial photography, satellite imagery, and the Digital Shoreline Analysis System (DSAS), we quantified the evolution of fourteen PML systems across the Svalbard Archipelago. Our results show that PMLs now occupy over 56 % of Svalbard's total lagoon area (ca. 83 km2), nearly triple the area they occupied in the 1930s. We identify two divergent evolutionary trajectories: (1) an erosional – fragmenting pathway (e.g., Tjuvfjordlaguna), where marine forcing leads to barrier narrowing and inlet expansion, and (2) a stabilizing – isolating pathway (e.g., Femtelaguna), where land-terminating glaciers drive rapid terrestrial sediment infilling and barrier progradation. We argue that PMLs function as essential “paraglacial sinks” trapping glaciogenic sediments and organic matter, thereby creating sheltered biodiversity hubs in otherwise harsh coastal environments. As transient features, the formation and eventual destruction of PMLs serve as a high-resolution proxy for the rapid paraglacial adjustment of polar coastlines.
Water stable isotopes (δ18OδD) from ice cores are widely used to reconstruct past temperature variations through their well-established relationship with local air temperature, commonly referred to as “isotopic paleothermometer”. However, depositional and post-depositional effects lead to large uncertainties in the use this proxy in Antarctica. The magnitude of these uncertainties strongly depends on site location, with larger impacts in low-accumulation regions of East Antarctic Plateau. Depositional effects include origin of moisture, which exhibits asymmetries shaped by the continent's geographical and topographical features, as well as precipitation intermittency, which introduces aliasing in the archived signal. Post-depositional processes, such as sublimation and firn-atmosphere exchange, further alter the isotopic composition of snow before its transformation into ice, thereby modifying the correlation between δ18Oδ18O-temperature relationship between different basins, with statistically significant correlations (p < 0.05) when the analysis is extended to the Antarctic dataset. This agreement further suggests the model's ability to predict the temporal slope required for calibrating isotopic ice-core records used for temperature reconstructions, even in regions influenced by multiple moisture sources. Temporal slopes based on monthly precipitation values range from 0.4 ‰ °C−1‰ °C−1δ18O
To improve the accuracy and timeliness of glacier surface-velocity retrieval in complex mountain terrain, we develop a high-spatial-resolution fusion method combining Landsat, Sentinel-1/2, and UAV (Unmanned Aerial Vehicle) data, and produce monthly velocity products for the Kangri Karpo region for 2015–2024. Compared with existing large-area public datasets, the products offer markedly higher spatial resolution and better detection of small mountain glaciers; relative to single-sensor inputs prior to fusion, the valid-pixel ratio increases by ∼50 %, the average number of valid months per pixel over the decade rises by ∼50, and spatial smoothness improves – demonstrating the method's suitability for rugged terrain. Spatially, velocities follow the canonical “fast center, slow margins” pattern, with multi-year maxima >700 m yr−1<100 m yr−1. Attribute analysis indicates significant correlations between velocity and area, slope, and aspect: larger glaciers flow faster overall; within individual glaciers, velocity responds more strongly to slope; and, with similar area and slope, south-facing glaciers are slightly faster than north-facing ones. Temporally, the intra-annual series shows clear seasonality, with peaks at the beginning and end of the melt season and sustained high speeds throughout. At the interannual scale, most pixelwise decadal trends lie within −36.5+36.5 m yr−1∼38.3 % of glaciers accelerate significantly, 25.5 % decelerate significantly, and 36.2 % show no significant trend (p≥0.05). By aspect, significant acceleration is concentrated on south- and west-facing glaciers, whereas significant deceleration occurs mainly on east- and north-facing glaciers. Month-resolved trends indicate acceleration primarily in April–May (∼54.7–73.0 m yr−1
Subglacial topography beneath the Greenland Ice Sheet is a fundamental control on its dynamics and response to changes in the climate system. Yet, it remains challenging to measure directly, and existing representations of the subglacial topography rely on a limited number of observations. Although the use of mass conservation and the development of BedMachine Greenland substantially improved the representation of the bed topography, this approach is limited to fast-flowing sectors and is less effective in regions with complex, alpine topography. As an alternative to traditional numerical methods, recent work has explored using Physics Informed Neural Networks (PINNs), constrained by only one physical law, to solve forward and inverse problems in ice sheet modeling. Building on this work, we assess three PINN frameworks constrained by distinct conservation laws, showing that PINNs informed with a single conservation law are not sufficient for regions with sparse measurements and complex topographies. To that end, we introduce a novel approach that involves coupling two
Information about rock glacier deformation with depth is crucial for understanding the kinematic processes responsible for variations in rock glacier velocity. The majority of studies on rock glacier kinematics have been limited to surface measurements. Understanding which processes lead to the observed displacement signal on the rock glacier surface is important because it is a parameter of the essential climate variable (ECV) “permafrost”. Here, we present a unique record of the nearly eight-year-long borehole deformation data at high temporal resolution from the Murtèl rock glacier in the Swiss Alps. The extracted velocity time series with depth shows that seasonal variations are only observed in the active layer (AL), while in the ice-rich core and the shear zone the velocity remains relatively stable. Interannual variability in deformation is driven by both the AL and the ice-rich core components, while the shear zone shows minimal change over the years. The AL, ice-rich core and shear zone components of deformation make up 20 %, 24 % and 56 % of surface displacement respectively. Compared to previous borehole inclinometer data, we find an unusually high fraction of surface displacement happening in the AL at Murtèl rock glacier. There are multiple rock glacier studies that report a fast response in surface velocity to water input from snowmelt or rainfall. In contrast, at Murtèl rock glacier we find that surface acceleration begins multiple weeks after the snowmelt period ends. The largest peaks in surface velocity coincide with the years with the highest ground surface temperature rather than the years with the most snowmelt. We postulate that the deep ice-rich and cold permafrost core of Murtèl rock glacier prevents the pore water pressure in the shear zone to change at a seasonal timescale. The borehole deformation and temperature data suggest that the seasonal cycle in AL deformation is strongly related to thermal processes and the melting of refrozen seasonal ice. The comparison of three independent approaches for measuring surface displacement shows that the borehole inclinometer and geodetic survey measurements agree well over multi-annual timescales. The continuous GNSS surface observations slightly overestimate the seasonal acceleration, but match the long-term background displacement well. Our borehole deformation data provide novel information on how representative surface velocities are for rock glacier deformation at depth and on various timescales.
Wildfires are increasingly burning higher in elevations well into the seasonal snow zone, altering snow accumulation and melt dynamics. However, limited spatially distributed observations throughout the full snow season have constrained our understanding of how these changes vary across space and time. Here, we assess the impacts of fire on snow depth across nine basins in California's Sierra Nevada using a machine learning (ML) algorithm, Extreme Gradient Boosting (XGBoost), trained on 114 airborne lidar snow depth acquisitions with 50 m resolution in partially burned basins. We develop and apply an explainable ML framework by fitting an ML algorithm on snow depth for each flight as a function of spatial attributes including burn status. We then predict snow depth for counterfactual burned and unburned conditions for each flight to assess the ML-derived impact of fire on snow depth. Fire impact on snow depth evolved throughout the season, with slightly deeper predicted snow in burned forests in the accumulation season (56 % of acquisitions), and shallower predicted snow in burned forests in the ablation season (83 % of acquisitions) compared to unburned forests. Post-fire snow depth differences were smaller in the accumulation than in the ablation season. Lower elevations (< 2000 m) consistently exhibited smaller, near-zero changes in post-fire snow depth compared to higher elevations (> 2000 m). South- facing slopes experienced the largest negative post-fire snow depth changes. These results illustrate a new approach to assessing fire impacts on snow using lidar-derived snow depth and provide insights into snowpack dynamics in burned forests that are novel in their spatial extent and resolution, as well as their ability to discern fire impacts throughout the snow season.