Sea ice age distribution for the Arctic for week 22 of the years 1995, 2005, and 2015. Numbers 1, 2, 3, etc. denote first-year ice, i.e. sea ice of the current freezing season, second-year ice (has survived one summer melt season), third-year ice (has survived two summer melt seasons), etc.
This data set is only available for a restricted user group, please contact us if you want to access these data.
RESTRICTED only accessable in CEN/MPI net or via CliSAP login What does that mean?
Data access via file system: /data/icdc/ice_and_snow/arctic_sea_ice_age
Due to visualisation reason, seasonal sea ice, i.e. the sea ice of the current winter, is assigned an age of 1 year. The reading of the sea-ice age is as follows: first-year sea ice (from the current winter, i.e. is in its first year, therefore = 1), second-year sea ice (has survived on melting season, is in its second year, therefore = 2), third-year sea ice (has survived two melting seasons), and so on.
Sea ice concentration and sea ice drift based on satellite data are used to investigate sea ice drift trajectories. First, the sea ice extent is calculated for every week. Only grid cells with at least 15% sea ice concentration contribute to the extent. Each of these grid cells is treated as a Lagrange particle which moves according to the mean weekly sea ice drift (derived from the NSIDC ice drift data set). The position of that cell is tracked for every week and the number of weeks the particle remains inside the area given by the sea ice extent, which is calculated anew for every week, is summed up which gives the age. If the particle is located outside this area after its translation due to the sea ice drift then it is assumed that the sea ice in that grid cell has melted. Every year in September, between week 37 and 38, upon commence of freezing conditions the age of all still existing grid cells is rounded up to the successive full year, e.g. a grid cell of age 45 weeks with > 15% first-year ice (rest is open water) becomes 2nd-year ice (images at top: "1" --> "2") while a grid cell of the age 2 years and 50 weeks i.e. 3rd-year ice of > 15% (rest is open water) becomes 4th-year ice (in images at top: "3" --> "4").
Co-existence of ice of different age in one grid cells results in the survive of the older one in order to take into account that younger and hence thinner sea ice deforms and melts more easily. Also here a threshold value of 15% is used: A grid cell with 15% ice 6 years old and 85% first-year ice is assigned 6-year old ice.
See the references for more details about the method and possible applications.
This is the NSIDC Arctic Sea Ice Age data set version 4.1 (see: http://nsidc.org/data/nsidc-0611), downloaded May 7, 2019 (until 2018-12-31).
Last data set update at ICDC: August 6, 2020.
|Sea ice age||years (0 ... 16)||0 = open water|
|1 = first-year ice|
|2 = second-year ice|
|16 = 16-year old ice|
|20 = land|
|21 = lakes / too close to coast|
Period and temporal resolution:
- 1984-01-01 until 2019-12-31
Coverage and spatial resolution:
- Northern hemisphere
- Spatial resolution: 12.5 km x 12.5 km, EASE-Projection (http://nsidc.org/data/ease/) but a different grid (see Na12500-CF.gpd)
- Geographic longitude: 0°E to 360°E
- Geographic latitude: 48.4°N to 90°N
- Dimension: 721 columns x 721 rows
- Altitude: 0.0 m
The data set does not contain uncertainty estimates. The approach uses weekly values of sea ice concentration and sea ice drift; random uncertainties in these daily products are therefore reduced.
Largest uncertainties can be expected during melt conditions and also in the marginal ice zone when and where both sea ice concentration and drift have the largest uncertainties.
It cannot be excluded that the ice age is - in general - positively biased because of the preferential assignment of the age of the oldest class to the entire grid cell - for concentrations of that oldest class down to 15% at the end of summer.
It cannot be excluded that potential inconsistencies in the sea ice drift data set used have an impact on the sea ice age derived. In this context we note, that the sea ice drift data set is based on a heterogeneous mixture of different data sources.
We note that for the first years of the data set the typical distribution of sea ice age classes is possibly not yet fully developed due to the methodology used. It lacks data of sea ice older than 7 years.
For further details and discussion of uncertainties and error sources we refer to the references.
National Snow and Ice Data Centre (NSIDC)
email: nsidc (at) nsidc.org
NASA Goddard Space Flight Center (GSFC)
email: walt.meier (at) nasa.gov
CCAR, University of Colorado, UCB 449, Boulder, CO
email: scottincolorado (at) gmail.com
CCAR, University of Colorado, UCB 431, Boulder, CO
email: mark.tschudi (at) colorado.edu
ICDC / CEN / University of Hamburg
email: stefan.kern (at) uni-hamburg.de
- Fowler, C., et al., Satellite-derived evolution of Arctic sea ice age: October 1978 to March 2003. Geosci. Rem. Sens. Lett., 1(2), 71-74, 2004, doi: 10.1109/LGRS.2004.824741
- Rigor, I.G., and J.M. Wallace, Variations in the age of Arctic sea-ice and summer sea-ice extent. Geophys. Res. Lett., 31(L09401), 2004, doi: 10.1029/2004GL019492
- Maslanik, J., et al., Distribution and trends in Arctic sea ice age through spring 2011. Geophys. Res. Lett., 38(L13502), 2011, doi: 10.1029/2011GL047735
- Tschudi, M. A., et al., Tracking the movement and changing surface characteristics of Arctic sea ice. J. Selected Topics in Earth Obs. Rem. Sens., 3(4), 2010, doi: 10.1109/JSTARS.2010.2048305
- Korosov, A. A., et al., A new tracking algorithm for sea ice age distribution estimation. The Cryosphere, 12, 2073-2085, 2018, doi: 10.5194/tc-12-2073-2018
Tschudi, M., W. N. Meier, J. S. Stewart, C. Fowler, and J. Maslanik. 2019 to present. EASE-Grid Sea Ice Age, Version 4.1, [indicate subset used]. Boulder, Colorado USA: NASA National Snow and Ice Data Center Distributed Active Archive Center. https://nsidc.org/data/nsidcc-0611. https://doi.org/10.5067/UTAV7490FEPB [last access date: July 30, 2020].