A comprehensive geospatial analysis of the Pemberton Icefield's glacial retreat over the last 40 years using multi-decadal Landsat imagery.
This technical report provides a comprehensive geospatial assessment of the Pemberton Icefield’s cryospheric evolution over a 40-year longitudinal study period (1985–2025). Utilizing a multi-sensor Landsat time series, the analysis integrates binary glacial status classifications with SWIR1-NIR-Red false-color composites to quantify surface area loss and morphological change. The findings indicate a profound and accelerating contraction of the icefield, characterized by significant terminal retreat and the thinning of lateral margins. Spectral analysis confirms that the transition from perennial ice to proglacial exposure is most pronounced at lower elevations, where the high-contrast visualization of the false-color imagery highlights a systematic replacement of high-albedo glacial surfaces with bedrock and pioneer vegetation.
The spatiotemporal dynamics of this retreat are further elucidated through pixel-level transition matrices, which reveal a critical decline in Glacier-to-Glacier (G→G) stability. While the 1985 baseline depicts a largely contiguous ice mass, the 2025 terminal state shows extensive fragmentation, particularly among the peripheral glacial tongues. The transition data suggests that the rate of ice-to-non-ice conversion has not remained linear but has accelerated in the latter half of the study period. Notable spatial patterns include the isolation of formerly tributary glaciers and the expansion of nunataks, signaling a shift from a unified icefield to a collection of disconnected alpine glaciers. These results underscore the high sensitivity of the Pemberton Icefield to regional climatic forcing and provide a high-resolution dataset for future hydrological and glaciological modeling.
The Pemberton Icefield has undergone a profound cryospheric contraction over the 40-year study period, characterized by a sustained and accelerating decline in ice extent. From a 1985 baseline of 70,212 hectares, the total glacial area has diminished to 40,686 hectares by 2025. This represents a net loss of approximately 42% of the icefield’s area. While the retreat was relatively linear during the late 20th century, the data indicates a significant acceleration in mass loss following 2010; nearly half of the total observed area loss occurred in the final 15 years of the time series.
A notable statistical anomaly is observed in the year 2000, where the classified glacial area reaches a period maximum of 71,506 hectares. This value deviates from the multi-decadal downward trend and is attributed to extensive seasonal snowpack at the time of image acquisition. This spectral interference temporarily masked the underlying glacial recession, highlighting the challenges of differentiating perennial ice from late-season snow in automated classifications. Following this anomaly, the 2005 data (61,688 ha) confirms a return to the long-term recession trajectory.
Spatially, the time-series animation reveals a distinct pattern of marginal recession and increasing fragmentation. The most pronounced changes are concentrated at lower-elevation valley tongues and along the periphery of the central ice mass. Key spatial trends include:
| Non-Ice | Glacier | |
|---|---|---|
| 1985 | 71350.0 | 70212.0 |
| 1990 | 80478.0 | 61084.0 |
| 1995 | 81908.0 | 59654.0 |
| 2000 | 70056.0 | 71506.0 |
| 2005 | 79874.0 | 61688.0 |
| 2010 | 80304.0 | 61258.0 |
| 2015 | 89972.0 | 51590.0 |
| 2020 | 96223.0 | 45339.0 |
| 2025 | 100876.0 | 40686.0 |
The false-color time series, spanning from 1985 to 2023, utilizes Landsat SWIR1-NIR-Red composites to provide a high-contrast visualization of cryospheric retreat. In this spectral rendering, snow and ice are characterized by a bright cyan to brilliant white signature. This distinct appearance is driven by the high reflectance of ice in the visible and near-infrared (NIR) spectra coupled with its strong absorption in the short-wave infrared (SWIR1) band. This contrast allows for the precise differentiation of glacial boundaries from the surrounding terrain, which appears in darker, muted tones of brown and grey representing exposed lithology and dormant vegetation.
The 38-year sequence reveals a significant and progressive reduction in total glacial area across the study site. Key observations include:
Spatially, the retreat is most pronounced in the lower-elevation valley glaciers, where the transition from clean ice (cyan) to debris-covered ice (pinkish-grey) precedes total ice loss. The geographic concentration of loss is highest along the southern aspects, likely due to increased solar radiation and thermal forcing. By 2023, the spatial extent of the ice cover represents a fraction of the 1985 baseline, illustrating a profound shift in the regional hydrological and cryospheric landscape.
Analysis of the pixel-level transition matrices from 1985 to 2025 reveals a sustained and accelerating contraction of glacial extent. The Glacier-to-Glacier (G→G) stability—representing the persistent glacial core—exhibited a significant downward trend, declining from 23.15% in the 1985–1990 period to a terminal value of 13.89% by 2025. This represents a nearly 40% reduction in the stable ice core over four decades. While the core remained relatively resilient through 2010, the post-2010 epochs show a marked collapse in stability, with the G→G value dropping below the 20% threshold for the first time during the 2010–2015 interval.
The marginal dynamics are characterized by high volatility in the Glacier-to-Non-Ice (G→NI) transition, which serves as a primary indicator of ablation and terminal retreat. Two distinct periods of rapid melting are observed: * 2000–2005: A peak transition of 6.90% from Glacier to Non-Ice, the highest recorded loss in the study period. * 2010–2015: A secondary surge of 4.74%, which initiated a decade of consistent decline in total ice cover.
Conversely, the Non-Ice-to-Glacier (NI→G) transitions, representing potential glacial advance or seasonal snow persistence, remained negligible (typically <1.0%) for most of the study. A notable anomaly occurred during the 1995–2000 interval, where a 6.57% gain was recorded; however, this gain was transient, as it was immediately offset by the record losses in the subsequent five-year period.
By the 2020–2025 interval, the landscape reached a state of pronounced deglaciation. The Non-Ice-to-Non-Ice (NI→NI) stability reached a study-high of 82.51%, indicating that the vast majority of the study area is now permanently ice-free. The narrowing gap between the stable core and marginal loss suggests that the glacier is not only retreating at its edges but is also undergoing internal thinning, leading to a fragmented and increasingly vulnerable cryospheric footprint.
Values in % Area
| 1985 (rows) | 1990 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 71.81 | 0.18 |
| Glacier | 4.86 | 23.15 |
Values in % Area
| 1990 (rows) | 1995 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 75.3 | 1.37 |
| Glacier | 1.85 | 21.48 |
Values in % Area
| 1995 (rows) | 2000 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 70.58 | 6.57 |
| Glacier | 0.36 | 22.49 |
Values in % Area
| 2000 (rows) | 2005 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 70.13 | 0.81 |
| Glacier | 6.9 | 22.15 |
Values in % Area
| 2005 (rows) | 2010 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 74.74 | 2.29 |
| Glacier | 1.94 | 21.02 |
Values in % Area
| 2010 (rows) | 2015 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 76.31 | 0.38 |
| Glacier | 4.74 | 18.57 |
Values in % Area
| 2015 (rows) | 2020 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 80.52 | 0.53 |
| Glacier | 2.98 | 15.97 |
Values in % Area
| 2020 (rows) | 2025 (columns) | Non-Ice | Glacier |
|---|---|---|
| Non-Ice | 82.51 | 0.99 |
| Glacier | 2.62 | 13.89 |