Support the Arctic Sea Ice Forum and Blog

Author Topic: Assessing uncertainties in Sea Ice Extent products  (Read 189 times)

Niall Dollard

  • Grease ice
  • Posts: 711
    • View Profile
  • Liked: 288
  • Likes Given: 60
Assessing uncertainties in Sea Ice Extent products
« on: July 03, 2020, 12:59:30 AM »
A good question raised by Paul in the melting thread on the differences between Jaxa and NSIDC this summer.

A good few of us have noted/mentioned this before. This summer it is quite noticeable.

I have found a good technical article which strives to assess differences between these products. Here:

https://iopscience.iop.org/article/10.1088/1748-9326/aaf52c

Some snippets :

There are clear differences in the estimates from the products and these differences vary seasonally (figure 2). These relate to two primary aspects of the products. First is the sensitivity of the algorithms to emission by different ice conditions, most notably thin ice and surface melt. Surface melt generally reduces concentration (the algorithm interprets the liquid water as open water) and thin ice is underestimated because of emission from the water beneath the ice (Steffen et al 1992). Second, as noted above, the spatial resolution is a key factor in the estimate of extent. Higher resolution products, such as those from AMSR2 will more precisely detect the edge, other factors being equal.

The low JAXA values are likely due to the higher spatial resolution, resulting in less 'smearing' of the ice edge; however, this doesn't necessarily mean that JAXA is more accurate—it may miss thin or melting ice and underestimate the ice edge location.

Attached Image shows the seasonal changes and that Jaxa's are lowest.

oren

  • Moderator
  • First-year ice
  • Posts: 5503
    • View Profile
  • Liked: 1684
  • Likes Given: 1590
Re: Assessing uncertainties in Sea Ice Extent products
« Reply #1 on: July 03, 2020, 04:13:22 AM »
The weird order of things is:
NSIDC extent > AMSR2 extent
AMSR2 area > NSIDC area

I browsed through the data files provided by Wipneus for NSIDC and for UH AMSR2 (not the same as JAXA, but hopefully a good proxy for finding differences from NSIDC). On July 1st this year NSIDC overestimated extent by 725k, while last year on July 1st the overestimate was 665k. The difference seems small but with some years bunched together this may explain the difference in the rankings. There are positive and negative differences in several regions, but it seems the biggest difference is in the Greenland Sea, where last year the overestimate was 37k and this year it's 101k.
Looking at the NSIDC and JAXA extent maps the visual difference does not jump out at me. Of course it does not help that the orientation is upside down. But in any case, I think this is all just noise. The rankings magnify small differences which will evolve through the season in any case.

ajouis

  • New ice
  • Posts: 63
    • View Profile
  • Liked: 24
  • Likes Given: 23
Re: Assessing uncertainties in Sea Ice Extent products
« Reply #2 on: July 03, 2020, 05:42:53 AM »
If the Greenland sea is where the biggest divergence happened, maybe it is linked to terminal glacier ice and/or lower resolution ability with higher artefacts due to the large body of land ice nearby?

interstitial

  • Frazil ice
  • Posts: 361
    • View Profile
  • Liked: 113
  • Likes Given: 67
Re: Assessing uncertainties in Sea Ice Extent products
« Reply #3 on: July 03, 2020, 07:32:09 AM »
Recently there was a large area in greenland sea that showed ice were non existed. I think it was jaxa but I am not sure. I don't think it is there anymore.

Paul

  • Frazil ice
  • Posts: 134
    • View Profile
  • Liked: 25
  • Likes Given: 2
Re: Assessing uncertainties in Sea Ice Extent products
« Reply #4 on: July 03, 2020, 12:28:40 PM »
The weird order of things is:
NSIDC extent > AMSR2 extent
AMSR2 area > NSIDC area

I browsed through the data files provided by Wipneus for NSIDC and for UH AMSR2 (not the same as JAXA, but hopefully a good proxy for finding differences from NSIDC). On July 1st this year NSIDC overestimated extent by 725k, while last year on July 1st the overestimate was 665k. The difference seems small but with some years bunched together this may explain the difference in the rankings. There are positive and negative differences in several regions, but it seems the biggest difference is in the Greenland Sea, where last year the overestimate was 37k and this year it's 101k.
Looking at the NSIDC and JAXA extent maps the visual difference does not jump out at me. Of course it does not help that the orientation is upside down. But in any case, I think this is all just noise. The rankings magnify small differences which will evolve through the season in any case.

That's the thing though, it has the same problem with most if not all years of recording ice where there is none so its not like this year is the odd one out.

As it happens, I trust the JAXA data more but usually they are both more or less in agreement on what the current extent is.