How glaciers gain and lose mass—the foundation of glacier dynamics
2026-02-27
Athabasca Glacier, the most visited glacier in North America, has retreated more than 1.5 km since measurements began in 1870 and lost roughly half its volume. Signs at the roadside parking lot mark where the glacier’s terminus stood in 1890, 1910, 1940, 1970, 1992. Each sign is a little further from the ice. The Columbia Icefield that feeds it holds enough water that if it melted completely, global sea levels would rise by a measurable fraction of a millimetre. Whether it continues to retreat, and at what rate, is determined by a single annual number: the specific mass balance — the difference between snow gained and ice lost over the year.
Glacier mass balance is the accountant’s view of a glacier: income (accumulation of snow and ice from snowfall, avalanches, and refreezing) versus expenditure (ablation through surface melt, sublimation, calving into water). When income exceeds expenditure the glacier advances; when expenditure exceeds income it retreats. The balance varies with elevation — the upper reaches of a glacier accumulate snow while the lower tongue melts vigorously — and the altitude at which the two exactly cancel is the equilibrium line altitude (ELA). As climate warms, the ELA rises, shrinking the accumulation zone. This model quantifies that relationship and connects it to the ice dynamics that determine how a glacier responds over decades to a step change in climate.
Will this glacier exist in 50 years?
Glacier mass balance determines fate:
\frac{dM}{dt} = \dot{b}
Where: - M = glacier mass (kg) - \dot{b} = mass balance rate (kg/m²/year or m w.e./year) - w.e. = water equivalent
Components:
\dot{b} = \dot{c} - \dot{a}
Where: - \dot{c} = accumulation rate (snowfall, avalanches, refreezing) - \dot{a} = ablation rate (melt, sublimation, calving)
Three scenarios: - \dot{b} > 0: Glacier growing (positive balance) - \dot{b} = 0: Steady state (equilibrium) - \dot{b} < 0: Glacier shrinking (negative balance)
Current reality: Most glaciers worldwide have \dot{b} < 0 (retreating)
Accumulation zone (upper glacier): - Snowfall > melt - Net gain each year - Ice flows downward
Ablation zone (lower glacier): - Melt > snowfall - Net loss each year - Ice flows in from above
Equilibrium Line Altitude (ELA): - Where accumulation = ablation - Divides two zones - Typically 3000-4000m in mid-latitudes
Typical profile:
With elevation: - High elevation: Large positive balance (+2 m w.e./year) - ELA: Zero balance (0 m w.e./year) - Low elevation: Large negative balance (-4 m w.e./year)
Gradient:
\frac{d\dot{b}}{dz} \approx 0.005-0.010 \text{ year}^{-1}
(5-10 mm w.e. per meter elevation)
Winter (Oct-May): - Accumulation dominates - Snow accumulates - Little/no melt (cold)
Summer (Jun-Sep): - Ablation dominates - Snow and ice melt - Maximum mass loss
Net annual balance:
\dot{b}_{\text{annual}} = \dot{b}_{\text{winter}} + \dot{b}_{\text{summer}}
Typically: \dot{b}_{\text{winter}} > 0, \dot{b}_{\text{summer}} < 0
At elevation z, time t:
\dot{b}(z,t) = P_{\text{snow}}(z,t) - M(z,t) - S(z,t) - C(z,t)
Where: - P_{\text{snow}} = solid precipitation - M = surface melt - S = sublimation - C = calving (if at terminus)
Melt from energy balance:
M = \frac{Q_{\text{net}}}{L_f \rho_w}
Where Q_{\text{net}} from Model 42 (shortwave, longwave, sensible, latent).
Integrate over glacier area:
B = \int_A \dot{b}(x,y) \, dA
Where: - B = total mass balance (m³ w.e./year) - A = glacier area
Specific balance (per unit area):
\bar{b} = \frac{B}{A}
Units: m w.e./year
Typical values: - Healthy glacier (equilibrium): \bar{b} \approx 0 m/year - Retreating glacier: \bar{b} = -0.5 to -2 m/year - Advancing glacier: \bar{b} = +0.5 to +1 m/year (rare today)
ELA where \dot{b}(z_{\text{ELA}}) = 0
Linear mass balance model:
\dot{b}(z) = \beta(z - z_{\text{ELA}})
Where \beta = mass balance gradient (year⁻¹)
Example: \beta = 0.008 year⁻¹, ELA = 3200m
At 3500m: \dot{b} = 0.008(3500 - 3200) = +2.4 m/year
At 2900m: \dot{b} = 0.008(2900 - 3200) = -2.4 m/year
\text{AAR} = \frac{A_{\text{accumulation}}}{A_{\text{total}}}
Equilibrium glacier: AAR ≈ 0.65 (65% in accumulation zone)
Retreating glacier: AAR < 0.65
Advancing glacier: AAR > 0.65
Why 65%? Accumulation zone has gentler slopes (more area per elevation), ablation zone steeper.
Problem: Calculate glacier-wide mass balance.
Glacier profile:
| Elevation (m) | Area (km²) | \dot{b} (m/year) |
|---|---|---|
| 3600-3800 | 2.0 | +1.5 |
| 3400-3600 | 3.0 | +1.0 |
| 3200-3400 | 4.0 | +0.5 |
| 3000-3200 | 3.5 | -0.5 |
| 2800-3000 | 2.5 | -1.5 |
| 2600-2800 | 1.0 | -2.5 |
Total area: 16.0 km²
Find: Glacier-wide mass balance, ELA, AAR
Step 1: Mass balance by band
Band 1: B_1 = 2.0 \times 1.5 = +3.0 km² · m/year
Band 2: B_2 = 3.0 \times 1.0 = +3.0
Band 3: B_3 = 4.0 \times 0.5 = +2.0
Band 4: B_4 = 3.5 \times (-0.5) = -1.75
Band 5: B_5 = 2.5 \times (-1.5) = -3.75
Band 6: B_6 = 1.0 \times (-2.5) = -2.5
Step 2: Total balance
B_{\text{total}} = 3.0 + 3.0 + 2.0 - 1.75 - 3.75 - 2.5 = 0.0 \text{ km}^3 \text{ w.e./year}
Specific balance:
\bar{b} = \frac{0.0}{16.0} = 0.0 \text{ m/year}
This glacier is in equilibrium!
Step 3: Find ELA
ELA between bands 3 and 4 (where balance crosses zero).
Linear interpolation: - Band 3 top (3400m): +0.5 m/year - Band 4 bottom (3200m): -0.5 m/year
\text{ELA} = 3200 + 200 \times \frac{0.5}{0.5 + 0.5} = 3200 + 100 = 3300 \text{ m}
Step 4: Calculate AAR
Accumulation zone (above 3300m): Bands 1, 2, 3 = 2.0 + 3.0 + 4.0 = 9.0 km²
\text{AAR} = \frac{9.0}{16.0} = 0.56 = 56\%
Note: AAR = 56% < 65% suggests this glacier would typically be retreating, BUT net balance is zero. This could indicate: - Recent advance (geometry catching up) - Steep ablation zone (high mass turnover) - Measurement uncertainty
Typical healthy glacier: AAR ≈ 0.65, \bar{b} = 0
Below is an interactive glacier mass balance simulator.
<label>
ELA shift (m):
<input type="range" id="ela-shift" min="-200" max="200" step="20" value="0">
<span id="ela-val">0</span> m
</label>
<label>
Mass balance gradient:
<input type="range" id="mb-gradient" min="0.004" max="0.012" step="0.001" value="0.008">
<span id="gradient-val">0.008</span> year⁻¹
</label>
<label>
Terminus elevation (m):
<input type="range" id="terminus-elev" min="2400" max="3000" step="100" value="2600">
<span id="terminus-val">2600</span>
</label>
<div class="glacier-info">
<p><strong>Glacier-wide balance:</strong> <span id="glacier-balance">--</span> m/year</p>
<p><strong>ELA:</strong> <span id="ela-result">--</span> m</p>
<p><strong>AAR:</strong> <span id="aar-result">--</span></p>
<p><strong>Status:</strong> <span id="glacier-status">--</span></p>
</div><canvas id="glacier-canvas" width="700" height="400" style="border: 1px solid #ddd;"></canvas>Try this: - ELA shift +100m: Warmer climate, glacier retreats (red bars dominate) - ELA shift -100m: Cooler climate, glacier advances (blue bars dominate) - Higher gradient: Steeper mass balance profile (more sensitive) - Lower terminus: Longer glacier, more ablation zone - Blue bars: Accumulation (snow gain) - Red bars: Ablation (ice loss) - Orange line: Equilibrium Line Altitude - Gray dashed: Glacier area distribution - Watch balance go from positive → equilibrium → negative!
Key insight: Small ELA shifts (+100m = ~0.7°C warming) dramatically affect glacier mass balance!
Global average: ELA rising ~150m since 1980
Causes: - Temperature increase (+1°C globally) - Changes in precipitation patterns - Earlier melt onset - Longer melt season
Glacier response: - 83% of glaciers retreating - Accelerating mass loss: -280 Gt/year (2000-2019) - Contributes ~0.7 mm/year to sea level rise
Maritime glaciers (Alaska, Patagonia, Iceland): - High accumulation, high ablation - Large mass turnover - Fast response to climate (~10-20 years)
Continental glaciers (central Asia): - Low accumulation, low ablation - Slow mass turnover - Slower response (~50-100 years)
Example - Jakobshavn Glacier (Greenland): - 1985: \bar{b} \approx 0 m/year - 2000: \bar{b} = -5 m/year - 2015: \bar{b} = -12 m/year - Dramatic acceleration!
Glaciers = frozen reservoirs:
During warming: - Increased melt → more summer flow (initially) - “Peak water” occurs when glacier contribution maximal - After peak → declining flow as glaciers shrink - Eventually → minimal glacial contribution
Central Asia: - Peak water ~2020-2030 - Major impacts on irrigation, hydropower
Rock debris on glacier surface:
Thin debris (<5 cm): - Darkens surface, lowers albedo - Increases melt (more absorbed solar)
Thick debris (>10 cm): - Insulates ice - Decreases melt (less energy reaches ice)
Result: Standard energy balance overestimates melt under thick debris.
Solution: Debris thickness model, adjust conductivity.
Tidewater glaciers:
Large mass loss from icebergs breaking off.
Calving flux:
C = u \times H \times W
Where: - u = ice velocity at terminus (m/year) - H = ice thickness (m) - W = terminus width (m)
Can exceed surface melt by 2-5×!
Solution: Include calving in ablation term.
Meltwater refreezes within snow/firn:
Not recorded as ablation (water stays in glacier).
Can be 10-30% of surface melt in cold glaciers.
Solution: Model refreezing with cold content calculation.
Point measurements (stakes) 2× per year:
Miss short-term events: - Rain-on-snow - Mid-winter melt - Dust deposition (albedo change)
Solution: Automated weather stations, remote sensing, modelling.
Measure elevation change with repeat surveys:
\Delta M = \Delta V \times \rho_{\text{ice}}
Where: - \Delta V = volume change (from DEM differencing) - \rho_{\text{ice}} \approx 900 kg/m³
Methods: - Photogrammetry (aerial photos) - LiDAR (airborne laser) - InSAR (satellite radar) - ICESat (satellite laser altimetry)
Advantages: - Glacier-wide, not point - Independent validation of glaciological method - Detects systematic biases
Disadvantages: - Requires firn density assumption - Lower temporal resolution - Expensive
Validation: Geodetic vs. glaciological should agree within ±20%.
For any control volume:
\frac{\partial M}{\partial t} = \dot{M}_{\text{in}} - \dot{M}_{\text{out}}
For glacier:
\frac{\partial M}{\partial t} = \dot{b} \times A
Integrated over time:
M(t) = M(t_0) + \int_{t_0}^{t} \dot{b}(t') \, dt'
Discrete (annual):
M_n = M_0 + \sum_{i=1}^{n} \dot{b}_i
Cumulative balance tracks total mass change since reference year.
Ice thickness change:
\frac{\partial H}{\partial t} = \dot{b} - \nabla \cdot \vec{q}
Where: - H = ice thickness - \vec{q} = ice flux vector (depth-integrated velocity)
Steady state: \partial H/\partial t = 0 requires flux divergence balances mass balance:
\nabla \cdot \vec{q} = \dot{b}
Accumulation zone: Ice flows away (\nabla \cdot \vec{q} > 0)
Ablation zone: Ice flows in (\nabla \cdot \vec{q} < 0)