ALASKA / ARCTIC COAST / 70.6385° N
Wainwright.
The changing edge, one year at a time.
BETWEEN ADJACENT REFERENCE DATES
A year of coastal change.
A small shift. A longer story.
Rates are drawn on two DNR reference display tracks, offset ±55 m for readability. These tracks are not shoreline locations. Click the map to inspect a reference point; drag to pan.
ONE FIXED REFERENCE POINT
Follow a place through time.
Where is this transect?
T0201Filled dots: end reference date. Open dots: start. WLI is blue; VL is gold. The inset locates this cut on the full coast.
Position through time · T0201
WLI VL — Smoothed model ○ Observed annual medianSigned position (m), positive toward land · gray band = selected interval
What does “position” mean?
Distance from a fixed zero to the boundary intersection, measured along this transect. A larger value means farther inland.
Along-coast distance chooses the transect.
Position measures across the coast.
Where is zero in these data?
The zero is fixed 600 m from the landward transect start. Position = 600 m − distance from that start. The same zero is used for every year. Position is a horizontal distance, not elevation or distance along the coast.
READ THE CHANGE, AND THE LIMITS
Three views.
One coastal record.
The dates matter.- 01 / Yearly interval rate
- Position change between adjacent reference dates, divided by the actual elapsed years.
- 02 / Cumulative displacement
- Total position change from the 1995 reference to each later reference, in metres. It is not divided by elapsed time.
- 03 / Compare any two years
- Total position difference between the two reference years you choose, in metres. It is not divided by the interval duration.
Data, quality, and calculation details
These are the existing spatially and temporally smoothed Wainwright experiment curves, sampled at 725 fixed transects spaced 10 m apart. The 10 m spacing is not positioning accuracy. Alongshore Gaussian smoothing is 75 m; the second temporal smoothing uses λ = 24.
Reference dates are in autumn for 1995–2024. The available model ends at mid-2025, so the last step spans 0.75 years; it is not a full calendar-year measurement. Values are never extrapolated beyond a transect’s supported model record. Missing endpoints stay missing.
Faded estimates retain source quality bits: observation gap >3 years, proximity to a supported record edge, low spatial support, or a smoothing adjustment >15 m. The 1996–1999 curves have no accepted annual observations. Some VL changes near 3.3 km have large smoothing adjustments.
The direction is positive toward land (apparent retreat / erosion) and negative toward sea (apparent advance / accretion). Water level, waves, vegetation seasonality, sensor changes, and registration can also affect positions. This experiment has no tide or sensor-offset correction and does not establish absolute erosion accuracy.
The satellite image is a fixed 2025-08-15 Sentinel-2C background in UTM 4N (EPSG:32604), not a historical scene animation. WLI and VL use different fixed color scales across all three modes. Gray is missing; lower-opacity colors show limited support.
All yearly + cumulative values · CSV ↓