Glacier evolution modelling with deep learning: challenges and opportunities

2020 ◽  
Author(s):  
Jordi Bolibar ◽  
Antoine Rabatel ◽  
Isabelle Gouttevin ◽  
Clovis Galiez ◽  
Thomas Condom ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Deep Learning ◽  
Mass Balance ◽  
Open Source ◽  
Data Science ◽  
Data Availability ◽  
Alternative Methods ◽  
Surface Mass Balance ◽  
Surface Mass ◽  
Challenges And Opportunities

<div>Glacier surface mass balance (SMB) and glacier evolution modelling have traditionally been tackled with physical/empirical methods, and despite some statistical studies very few efforts have been made towards machine learning approaches. With the end of this past decade, we have witnessed an impressive increase in the available amount of data, mostly coming from remote sensing products and reanalyses, as well as an extensive list of open-source tools and libraries for data science. Here we introduce a first effort to use deep learning (i.e. a deep artificial neural network) to simulate glacier-wide surface mass balance at a regional scale, based on direct and remote sensing SMB data, climate reanalysis and multitemporal glacier inventories. Coupled with a parameterized glacier-specific ice dynamics function, this allows us to simulate the evolution of glaciers for a whole region. This has been developed as the ALpine Parameterized Glacier Model (ALPGM), an open-source Python glacier evolution model. To illustrate this data science approach, we present the results of a glacier-wide surface mass balance reconstruction of all the glaciers in the French Alps from 1967-2015. These results were analysed and compared with all the available observations in the region as well as another physical/empirical SMB reconstruction study. We observe some interesting differences between the two SMB reconstructions, which further highlight the interest of using alternative methods in glacier modelling. Due to (relatively) recent advances in data availability and open tools (e.g. Tensorflow, Keras, Pangeo) this research field is ripe for progress, with many interesting challenges and opportunities lying ahead. To conclude, some perspectives on data science glacier modelling are discussed, based on the limitations of our current approach and on upcoming tools and methods, such as convolutional and physics-informed neural networks. </div>

scholarly journals A deep learning reconstruction of mass balance series for all glaciers in the French Alps: 1967–2015

2020 ◽  
Author(s):  
Jordi Bolibar ◽  
Antoine Rabatel ◽  
Isabelle Gouttevin ◽  
Clovis Galiez
Keyword(s):  
Deep Learning ◽  
Mass Balance ◽  
Data Science ◽  
Regional Analysis ◽  
Average Error ◽  
High Mountain ◽  
Surface Mass Balance ◽  
Surface Mass ◽  
French Alps ◽  
Mass Losses

Abstract. Glacier surface mass balance (SMB) data are crucial to understand and quantify the regional effects of climate on glaciers and the high-mountain water cycle, yet observations cover only a small fraction of glaciers in the world. We present a dataset of annual glacier-wide surface mass balance of all the glaciers in the French Alps for the 1967–2015 period. This dataset has been reconstructed using deep learning (i.e. a deep artificial neural network), based on direct and remote sensing SMB observations, meteorological reanalyses and topographical data from glacier inventories. This data science reconstruction approach is embedded as a SMB component of the open-source ALpine Parameterized Glacier Model (ALPGM). An extensive cross-validation allowed to assess the method’s validity, with an estimated average error (RMSE) of 0.49 m w.e. a−1, an explained variance (r2) of 79 % and an average bias of +0.017 m w.e. a−1. We estimate an average regional area-weighted glacier-wide SMB of −0.72 ± 0.20 m w.e. a−1 for the 1967–2015 period, with moderately negative mass balances in the 1970s (−0.52 m w.e. a−1) and 1980s (−0.12 m w.e. a−1), and an increasing negative trend from the 1990s onwards, up to −1.39 m w.e. a−1 in the 2010s. Following a topographical and regional analysis, we estimate that the massifs with the highest mass losses for this period are the Chablais (−0.90 m w.e. a−1) and Ubaye and Champsaur ranges (−0.91 m w.e. a−1 both), and the ones presenting the lowest mass losses are the Mont-Blanc (−0.74 m w.e. a−1), Oisans and Haute-Tarentaise ranges (−0.78 m w.e. a−1 both). This dataset (available at: https://doi.org/10.5281/zenodo.3663630) (Bolibar et al., 2020a) – provides relevant and timely data for studies in the fields of glaciology, hydrology and ecology in the French Alps, in need of regional or glacier-specific meltwater contributions in glacierized catchments.

Early 21st-century glacier surface mass balance across High Mountain Asia derived from remote sensing

2020 ◽  
Author(s):  
Evan Miles ◽  
Michael McCarthy ◽  
Amaury Dehecq ◽  
Marin Kneib ◽  
Stefan Fugger ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Mass Balance ◽  
Surface Velocity ◽  
High Mountain ◽  
Surface Mass Balance ◽  
Equilibrium Line Altitude ◽  
Glacier Surface ◽  
Surface Mass ◽  
Early 21St Century ◽  
Debris Cover

<p>Glaciers in High Mountain Asia have experienced intense scientific scrutiny in the past decade due to their hydrological and societal importance. The explosion of freely-available satellite observations has greatly advanced our understanding of their thinning, motion, and overall mass losses, and it has become clear that they exhibit both local and regional variations due to debris cover, surging and climatic regime. However, our understanding of glacier accumulation and ablation rates is limited to a few individual sites, and altitudinal surface mass balance is essentially unknown across the vast region.</p><p>Here we combine recent assessments of ice thickness and surface velocity to correct observed glacier thinning rates for mass redistribution in a flowband framework to derive the first estimates of altitudinal glacier surface mass balance across the region. We first evaluate our results at the glacier scale with all available glaciological field measurements (27 glaciers), then analyze 4665 glaciers (we exclude surging and other anomalous glaciers) comprising 43% of area and 36% of mass for glaciers larger than 2 km<sup>2</sup> in the region. The surface mass balance results allow us to determine the equilibrium line altitude for each glacier for the period 2000-2016.  We then aggregate our altitudinal and hypsometric surface mass balance results to produce idealised profiles for distinct subregions, enabling us to consider the subregional heterogeneity of mass balance and the importance of debris-covered ice for the region’s overall ablation.</p><p>We find clear patterns of ELA variability across the region.  9% of glaciers accumulate mass over less than 10% of their area on average for the study period. These doomed  glaciers are concentrated in Nyainqentanglha, which also has the most negative mass balance of the subregions, whereas accumulation area ratios of 0.7-0.9 are common for glaciers in the neutral-balance Karakoram and Kunlun Shan. We find that surface debris extent is negatively correlated with ELA, explaining up to 1000 m of variability across the region and reflecting the importance of avalanching as a mass input for debris-covered glaciers at lower elevations. However, in contrast with studies of thinning rates alone, we find a clear melt reduction for low-elevation debris-covered glacier areas, consistent across regions, largely resolving the ‘debris cover anomaly’.  </p><p>Our results provide a comprehensive baseline for the health of the High Asian ice reservoirs in the early 21<sup>st</sup> Century. The estimates of altitudinal surface mass balance and ELAs will additionally enable novel strategies for the calibration of glacier and hydrological models. Finally, our results emphasize the potential of combined remote-sensing observations to understand the environmental factors and physical processes responsible for High Asia’s heterogeneous patterns of recent glacier evolution.</p>

scholarly journals COSIPY v1.3 – an open-source coupled snowpack and ice surface energy and mass balance model

2020 ◽  
Vol 13 (11) ◽  
pp. 5645-5662
Author(s):  
Tobias Sauter ◽  
Anselm Arndt ◽  
Christoph Schneider
Keyword(s):  
Energy Balance ◽  
Surface Energy ◽  
Mass Balance ◽  
Open Source ◽  
Surface Energy Balance ◽  
Balance Model ◽  
Surface Mass Balance ◽  
Mass Balance Model ◽  
Surface Mass ◽  
Ice Surface

Abstract. Glacier changes are a vivid example of how environmental systems react to a changing climate. Distributed surface mass balance models, which translate the meteorological conditions on glaciers into local melting rates, help to attribute and detect glacier mass and volume responses to changes in the climate drivers. A well-calibrated model is a suitable test bed for sensitivity, detection, and attribution analyses for many scientific applications and often serves as a tool for quantifying the inherent uncertainties. Here, we present the open-source COupled Snowpack and Ice surface energy and mass balance model in PYthon (COSIPY), which provides a flexible and user-friendly framework for modeling distributed snow and glacier mass changes. The model has a modular structure so that the exchange of routines or parameterizations of physical processes is possible with little effort for the user. The framework consists of a computational kernel, which forms the runtime environment and takes care of the initialization, the input–output routines, and the parallelization, as well as the grid and data structures. This structure offers maximum flexibility without having to worry about the internal numerical flow. The adaptive subsurface scheme allows an efficient and fast calculation of the otherwise computationally demanding fundamental equations. The surface energy balance scheme uses established standard parameterizations for radiation as well as for the energy exchange between atmosphere and surface. The schemes are coupled by solving both surface energy balance and subsurface fluxes iteratively such that consistent surface skin temperature is returned at the interface. COSIPY uses a one-dimensional approach limited to the vertical fluxes of energy and matter but neglects any lateral processes. Accordingly, the model can be easily set up in parallel computational environments for calculating both energy balance and climatic surface mass balance of glacier surfaces based on flexible horizontal grids and with varying temporal resolution. The model is made available on a freely accessible site and can be used for non-profit purposes. Scientists are encouraged to actively participate in the extension and improvement of the model code.

scholarly journals Reversed Surface-Mass-Balance Gradients on Himalayan Debris-Covered Glaciers Inferred from Remote Sensing

2020 ◽  
Vol 12 (10) ◽  
pp. 1563 ◽  
Author(s):  
Rosie R. Bisset ◽  
Amaury Dehecq ◽  
Daniel N. Goldberg ◽  
Matthias Huss ◽  
Robert G. Bingham ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Mass Balance ◽  
Critical Role ◽  
Regional Scale ◽  
High Mountain ◽  
Surface Mass Balance ◽  
Equilibrium Line Altitude ◽  
Ice Dynamics ◽  
Surface Mass ◽  
Debris Coverage

Meltwater from the glaciers in High Mountain Asia plays a critical role in water availability and food security in central and southern Asia. However, observations of glacier ablation and accumulation rates are limited in spatial and temporal scale due to the challenges that are associated with fieldwork at the remote, high-altitude settings of these glaciers. Here, using a remote-sensing-based mass-continuity approach, we compute regional-scale surface mass balance of glaciers in five key regions across High Mountain Asia. After accounting for the role of ice flow, we find distinctively different altitudinal surface-mass-balance gradients between heavily debris-covered and relatively debris-free areas. In the region surrounding Mount Everest, where debris coverage is the most extensive, our results show a reversed mean surface-mass-balance gradient of −0.21 ± 0.18 m w.e. a−1 (100 m)−1 on the low-elevation portions of glaciers, switching to a positive mean gradient of 1.21 ± 0.41 m w.e. a−1 (100 m)−1 above an average elevation of 5520 ± 50 m. Meanwhile, in West Nepal, where the debris coverage is minimal, we find a continuously positive mean gradient of 1.18 ± 0.40 m w.e. a−1 (100 m)−1. Equilibrium line altitude estimates, which are derived from our surface-mass-balance gradients, display a strong regional gradient, increasing from northwest (4490 ± 140 m) to southeast (5690 ± 130 m). Overall, our findings emphasise the importance of separating signals of surface mass balance and ice dynamics, in order to constrain better their contribution towards the ice thinning that is being observed across High Mountain Asia.

scholarly journals 25 years (1981–2005) of equilibrium-line altitude and mass-balance reconstruction on Glacier Blanc, French Alps, using remote-sensing methods and meteorological data

2008 ◽  
Vol 54 (185) ◽  
pp. 307-314 ◽  
Author(s):  
Antoine Rabatel ◽  
Jean-Pierre Dedieu ◽  
Emmanuel Thibert ◽  
Anne Letréguilly ◽  
Christian Vincent
Keyword(s):  
Remote Sensing ◽  
Mass Balance ◽  
Satellite Images ◽  
Meteorological Data ◽  
Equilibrium Line ◽  
Surface Mass Balance ◽  
Mass Balances ◽  
Equilibrium Line Altitude ◽  
Surface Mass ◽  
French Alps

AbstractAnnual equilibrium-line altitude (ELA) and surface mass balance of Glacier Blanc, Ecrins region, French Alps, were reconstructed from a 25 year time series of satellite images (1981–2005). The remote-sensing method used was based on identification of the snowline, which is easy to discern on optical satellite images taken at the end of the ablation season. In addition, surface mass balances at the ELA were reconstructed for the same period using meteorological data from three nearby weather stations. A comparison of the two types of series reveals a correlation of r > 0.67 at the 0.01 level of significance. Furthermore, the surface mass balances obtained from remote-sensing data are consistent with those obtained from field measurements on five other French glaciers (r = 0.76, p < 0.01). Also consistent for Glacier Blanc is the total mass loss (10.8 m w.e.) over the studied period. However, the surface mass balances obtained with the remote-sensing method show lower interannual variability. Given that the remote-sensing method is based on changes in the ELA, this difference probably results from the lower sensitivity of the surface mass balance to climate parameters at the ELA.

THE STUDIES OF SURFACE MASS BALANCE AND ICE FLOW ON GLACIERS AUSTRELOVéNBREEN AND PEDERSENBREEN, SVALBARD, ARCTIC

2010 ◽  
Vol 22 (1) ◽  
pp. 10-22 ◽  
Author(s):  
Mingxing Xu ◽  
Ming Yan ◽  
Jiawen Ren ◽  
Songtao Ai ◽  
Jiancheng Kang ◽  
...  
Keyword(s):  
Mass Balance ◽  
Surface Mass Balance ◽  
Ice Flow ◽  
Surface Mass

scholarly journals Reconstruction of historical surface mass balance, 1984–2017 from GreenTrACS multi-offset ground-penetrating radar

2020 ◽  
pp. 1-10
Author(s):  
Tate G. Meehan ◽  
H. P. Marshall ◽  
John H. Bradford ◽  
Robert L. Hawley ◽  
Thomas B. Overly ◽  
...  
Keyword(s):  
Mass Balance ◽  
Ground Penetrating Radar ◽  
Greenland Ice Sheet ◽  
Snow Accumulation ◽  
Surface Mass Balance ◽  
Snow Density ◽  
Surface Mass ◽  
Age Model ◽  
Ground Penetrating ◽  
Good Agreement

Abstract We present continuous estimates of snow and firn density, layer depth and accumulation from a multi-channel, multi-offset, ground-penetrating radar traverse. Our method uses the electromagnetic velocity, estimated from waveform travel-times measured at common-midpoints between sources and receivers. Previously, common-midpoint radar experiments on ice sheets have been limited to point observations. We completed radar velocity analysis in the upper ~2 m to estimate the surface and average snow density of the Greenland Ice Sheet. We parameterized the Herron and Langway (1980) firn density and age model using the radar-derived snow density, radar-derived surface mass balance (2015–2017) and reanalysis-derived temperature data. We applied structure-oriented filtering to the radar image along constant age horizons and increased the depth at which horizons could be reliably interpreted. We reconstructed the historical instantaneous surface mass balance, which we averaged into annual and multidecadal products along a 78 km traverse for the period 1984–2017. We found good agreement between our physically constrained parameterization and a firn core collected from the dry snow accumulation zone, and gained insights into the spatial correlation of surface snow density.

scholarly journals Glacier Surface Mass Balance in the Suntar-Khayata Mountains, Northeastern Siberia

Water ◽  
2019 ◽  
Vol 11 (9) ◽  
pp. 1949 ◽  
Author(s):  
Yong Zhang ◽  
Xin Wang ◽  
Zongli Jiang ◽  
Junfeng Wei ◽  
Hiroyuki Enomoto ◽  
...  
Keyword(s):  
Mass Loss ◽  
Mass Balance ◽  
Surface Mass Balance ◽  
Negative Mass ◽  
Glacier Surface ◽  
Surface Mass ◽  
Northeastern Siberia ◽  
Response To Temperature ◽  
Rapid Mass

Arctic glaciers comprise a small fraction of the world’s land ice area, but their ongoing mass loss currently represents a large cryospheric contribution to the sea level rise. In the Suntar-Khayata Mountains (SKMs) of northeastern Siberia, in situ measurements of glacier surface mass balance (SMB) are relatively sparse, limiting our understanding of the spatiotemporal patterns of regional mass loss. Here, we present SMB time series for all glaciers in the SKMs, estimated through a glacier SMB model. Our results yielded an average SMB of −0.22 m water equivalents (w.e.) year−1 for the whole region during 1951–2011. We found that 77.4% of these glaciers had a negative mass balance and detected slightly negative mass balance prior to 1991 and significantly rapid mass loss since 1991. The analysis suggests that the rapidly accelerating mass loss was dominated by increased surface melting, while the importance of refreezing in the SMB progressively decreased over time. Projections under two future climate scenarios confirmed the sustained rapid shrinkage of these glaciers. In response to temperature rise, the total present glacier area is likely to decrease by around 50% during the period 2071–2100 under representative concentration pathway 8.5 (RCP8.5).

scholarly journals Long-range terrestrial laser scanning measurements of annual and intra-annual mass balances for Urumqi Glacier No. 1, eastern Tien Shan, China

2019 ◽  
Vol 13 (9) ◽  
pp. 2361-2383 ◽  
Author(s):  
Chunhai Xu ◽  
Zhongqin Li ◽  
Huilin Li ◽  
Feiteng Wang ◽  
Ping Zhou
Keyword(s):  
Mass Balance ◽  
Long Range ◽  
Tien Shan ◽  
Glacier Mass Balance ◽  
Surface Mass Balance ◽  
Mass Balances ◽  
Surface Mass ◽  
Mass Conversion ◽  
Geodetic Mass Balance

Abstract. The direct glaciological method provides in situ observations of annual or seasonal surface mass balance, but can only be implemented through a succession of intensive in situ measurements of field networks of stakes and snow pits. This has contributed to glacier surface mass-balance measurements being sparse and often discontinuous in the Tien Shan. Nevertheless, long-term glacier mass-balance measurements are the basis for understanding climate–glacier interactions and projecting future water availability for glacierized catchments in the Tien Shan. Riegl VZ®-6000 long-range terrestrial laser scanner (TLS), typically using class 3B laser beams, is exceptionally well suited for repeated glacier mapping, and thus determination of annual and seasonal geodetic mass balance. This paper introduces the applied TLS for monitoring summer and annual surface elevation and geodetic mass changes of Urumqi Glacier No. 1 as well as delineating accurate glacier boundaries for 2 consecutive mass-balance years (2015–2017), and discusses the potential of such technology in glaciological applications. Three-dimensional changes of ice and firn–snow bodies and the corresponding densities were considered for the volume-to-mass conversion. The glacier showed pronounced thinning and mass loss for the four investigated periods; glacier-wide geodetic mass balance in the mass-balance year 2015–2016 was slightly more negative than in 2016–2017. Statistical comparison shows that agreement between the glaciological and geodetic mass balances can be considered satisfactory, indicating that the TLS system yields accurate results and has the potential to monitor remote and inaccessible glacier areas where no glaciological measurements are available as the vertical velocity component of the glacier is negligible. For wide applications of the TLS in glaciology, we should use stable scan positions and in-situ-measured densities of snow–firn to establish volume-to-mass conversion.

Sign in / Sign up

Export Citation Format

Share Document