GLIMS Update: Chamonix Workshop Summary
Jeffrey S Kargel
jkargel at usgs.gov
Mon Sep 16 18:26:13 MDT 2002
PENULTIMATE DRAFT. To go out at end of day if possible. Rick, can you go
through a paragraph that pertains to image aquisition statistics (just
after first dashed line). And you were wondering whether to include any
info on wide pointing, since you did not actually discuss that. Since
Bruce and I both forgot a summary of Andy's talk, which was one of the
highlights for me, I wonder what other crucial stuff we're missing.
Dear GLIMS community:
These are notes compiled by Bruce Raup, Jeff Kargel, and Andy Kaeaeb on
the GLIMS workshop held in Chamonix, 2002-09-02/03 following a successful
IGS Symposium the week before, also in Chamonix. Apologies are extended
in advance for any omissions or inaccuracies. There are four issues
which require a GLIMS-wide evaluation and commentary feedback from each
of you:
(1) Whether to push for a journal special issue on remote sensing of
changing glaciers and field validation of remote sensing determinations
of changes.
(2) Future GLIMS meeting opportunities, including theme-based meetings
and training classes in use of algorithms for glacier mapping and change
determination.
(3) Each RC is asked to develop web pages for inclusion on the GLIMS
website or hot linkable to and from the GLIMS site.
(4) VERY IMPORTANT-- the NSIDC database is ready for ingest of GLIMS
data; we are mandating a near-term ingest from each RC-- data for a
few as a single glacier. These matters are detailed below. And there
are additional issues which must soon be tackled by GLIMS subsets
comprising two GLIMS working groups.
Attendees: Brian Anderson, Hernan De Angelis, Jorge Arigony, Etienne
Berthier, Jaume Calvet, Luke Copland, Jean-Pierre Dedieu, David Garcia,
Arthur Greene, Jan Ove Hagen, Ricardo Jana, Andreas Kaeaeb, Jeff Kargel,
Mikael Luthje, Fabian Mauz, Frank Paul, Frank Rau, Bruce Raup, Niels
Reeh, Paul Sirota, Leigh Stearns, Rick Wessels
Representatives of sixteen institutions in ten nations were present.
Seven Regional Centers, plus some of their stewards, and the NSIDC
Database Center and the GLIMS Coordination Center were represented.
During the meeting the GLIMS European Alps RC adopted a new France
steward (Jean-Pierre Dedieu), in addition to a new Spanish Pyrenees
steward adopted just before the workshop (Javier Chueca).
---------------------------------------------------------------------------
A number of image classification algorithms were presented, all of which
were promising for clean ice. Frank Paul presented an algorithm which
uses band ratioing, then expands this area by selecting areas adjacent to
"glacier" (as determined by the band ratios) that have slopes less than
23 degrees.
Rick Wessels presented "glimsview", a tool for analyzing ASTER imagery and
producing GLIMS data. The first major release of glimsview is set for Oct.
31.
Rick also summarized the status of ASTER acquisitions over glaciers.
There are currently about 7300 L1A and 2300 L1B ASTER scenes with the
correct gain settings and potentially low cloud cover over our GLIMS
regions. Of these, about 6800 L1A and 2100 L1B cover the arctic and
antarctic areas. Over 920 "wide-pointing" L1A images have been
acquired. (The set of GLIMS data acquisition requests (DARs) includes
regions that require the wide-pointing capability of the VNIR system, in
order to image more pole-ward than otherwise possible.) Of these, only
three have been processed to L1B and most show scene cloud-cover
estimates of > 90%.
Bruce Raup presented the status of the NSIDC glacier database, and the
data transfer format. There was significant discussion of additional
database segment types and definitions to be considered by the GLIMS
Database/Parameters Working Group.
Luke Copland presented the Canadian glacier database and interface (at
http://icedata.eas.ualberta.ca). One interesting feature of their
interface is that the user can upload their own data and display on top of
their glacier maps.
Andreas Kaeaeb presented a quantitative assessment of ASTER DEMs and
glacier
flow displacement mapping. ASTER DEMs can be derived with an accuracy of
approx. 20m for suitable terrain, including glacier surfaces, but errors
of hundreds of meters occur for steep/high peaks. One problem is that steep
terrain can cause a point seen in nadir to be hidden in the aft looking
image; the data do not exist for parallax determination, and so any
interpolations across the null regions must involve errors. Selection of
immobile control points will thus have to be done very carefully when
rendering digital terrain models into absolute digital elevation models or
when assessing glacier displacements. Glacier mapping from ASTER has
similar accuracy as from ETM+. Glacier flow fields can be derived from
repeated ASTER imagery with approx. 10 m accuracy. These and other
results are summarized at www.geo.unizh.ch/~kaeaeb/glims/index.html. The
flow displacement measurements for one glacier and a rock glacier were
used in an IGS poster by Kargel, Kaeaeb, Wessels, and Kieffer to extract
some useful rheological information for these icy flows.
Kaeaeb also presented a pre-dinner presentation of the development of a
supraglacial lake on Bellvedere Glacier, Italian Alps. Their group has
won a contract to study this particular lake and its potential hazards.
Rick Wessels asked how many people had simple glacier outline maps for
any or all the glaciers in their regions. Only Andy Kaeaeb and Luke
Copland raised their hands.
The need for thorough guidelines (or "cookbooks") was discussed.
Cookbook topics include:
Algorithms
Glacier Parameters (snowlines, front types, etc.)
Assigning GLIMS IDs to glaciers
Putting data together into the data submission format
Luke Copland has an 8-page document detailing the steps in their
processing. Fabian Mauz has a great start at a pictorial dictionary of
glacier parameters.
The GLIMS working groups were formerly "algorithms", "database", and
"parameters". It was recognized that much of the work of "parameters"
had already been subsumed into the "database" working group. We decided
to formalize this, so that now there are only two working groups:
Algorithms (chaired by Andreas Kaeaeb), and Database/Parameters (chaired
by Bruce Raup).
Updates from Regional Centers and other presentations
Andreas Kaeaeb (Swiss steward representing the European Alps RC)
Wide ranging progress-- though little of it ASTER-based for lack of good
cloud-free coverage-- was reported on a country-by-country basis. Among
other progress, the Swiss glacier inventory-- mainly from Landsat-- is
almost complete via a digital Hydrological Atlas.
Leigh Stearns (student of Gordon Hamilton, University of Maine; Ross Ice
shelf sector of Antarctica):
GLIMS RC activity is funded by NASA (Gordon Hamilton, PI). Activities
include feature tracking in ice streams E, F, and Byrd Glacier; ice shelf
monitoring (tracking ice wall position). Data used include DISP, MSS,
ASTER, SPOT, Landsat 7, collectively presenting a 40-year history
of ice wall position of ice shelves. Stearns will likely be starting
the ice shelf studies for her PhD dissertation. Also presented plans
for a newly funded Svalbard stewardship under the Norway islands RC.
Niels Reeh (Technical University of Denmark; Greenland):
Anker Weidick's SW Greenland glacier inventory contains around 5000
glaciers. This inventory was meticulously produced over years by hand.
Niels is investigating whether it exists in digital form.
Frank Rau (University of Freiburg; Antarctic Peninsula):
Presented some results from James Ross and Vega Islands (42.4 km2 loss of
ice area 1988 to 2001). He listed the limitations the work under as RC
for Antarctic Peninsula: image data availability; multitude of glacier
types; large glacierized area, with many ice divides; difficulty in
determining boundaries; poor geodetic database (GCPs, etc.); clouds;
polar night. Of 2000 ASTER images, 90% are cloudy and 15% of clear-sky
scenes have the wrong gain settings. Even so, ASTER coverage is fairly
good, and Landsat coverage is complete going back to mid 1980's.
Pointed out weakness in WGMS-like glacier classification system, which,
for example, makes no distinction between dry and wet calving. Glacier
changes have been detected using ASTER and archive data. 1-2 km of
retreat from 1988-2001 for most of Ross Island area, with net change in
area of -42.4 km2. Glaciers are stable in NW Antarctic Peninsula, and
retreat is observed on the NE and SW coasts of the peninsula. RC
activities are funded by Deutsche-Forschegsgemeinschaft and Volkeswagen.
Jaume Calvet (University of Barcelona):
Showed results from Livingston Island (approx. 60 W, 62.5 S), where they
have both spaceborne data and ground-based photogrammetric data. Results
showed significant glacier thinning and 4.3% retreat in area since 1956.
South Shetland Islands still do not have a single cloud-free ASTER image,
thus prompting Rick Wessels to make an attempt to bump up imaging
priority
for this and other notoriously cloudy areas, such as the European Alps.
Frank Paul (University of Zurich, S. European Alps):
Reported on the Swiss Glacier Inventory. The mass loss rate for 1985 to
1998 was 14 %/decade, which was 7 times higher than the period 1850-1973
(which saw 2 %/decade loss). Area lost since 1973: 20 %.
Paul Sirota (University of Otago, NZ):
New student of Blair Fitzharris, Sirota reported on the New Zealand
glacier inventory, which contains approximately 3000 glaciers. Much work
is based on aerial photography.
Jeff Kargel presented still-developing plans for establishment of HIGH ICE
(Himalayan Institutes of Glaciology: Hydrology, Ice, Climate, and
Environment)
in a broad collaboration, including GLIMS regional centers covering the
High Asia region. HIGH ICE is a "daughter of GLIMS" with an applications
focus. The first priority activity is in Pakistan and India, where
HIGH ICE is intended as a "science for peace" initiative.
Mikael Luthje
Works with Niels Reeh, Presented a poster on Greenland drainage networks.
Rick Wessels
Presented several posters with spectacular ASTER images of Ellesmere
Island
and other regions.
---------------------------------------------------------------------------
Discussion:
Jeff Kargel led a general discussion, where topics included: data
ingest; new web site in the works at USGS/Flagstaff; possibility of a
special GLIMS journal issue; field campaigns; future meetings.
1. FUTURE MEETINGS: General GLIMS project input is desired, including the
likelihood that you would either (a) be there anyway, and (b) attend if
a GLIMS meeting also was scheduled. Possible future
conferences on which to piggy-back another GLIMS workshop may include:
(1) Symposium on the Mass Balnce of Andean Glaciers, March 12-13, 2003,
Valdivia, Chile. A GLIMS workshop potentially could be piggybacked,
including 1 day of GLIMS RC reports and other GLIMS business, and 1 day
of algorithm-use (training) classes. See http://www.cecs.cl/ for
information on the ICSI-IAHS-sponsored Andean glacier meeting.
(2) Joint EGS/EUG/AGU Spring meeting, Nice, France, April 7-11, 2003.
This meeting will include several glacier/cryosphere sessions, including
one on changing glaciers. See http://www.copernicus.org/egsagueug/ for
more information. The plan would be to hold in conjunction with this
meeting a 1-day GLIMS workshop, potentially with and additional 1 day
of training classes for use of algorithms.
(3) IUGG, Sapporo, Japan, June 30-July 11, 2003.See
http://www.jamstec.go.jp/jamstec-e/iugg/ There will be several cryosphere
sessions, including remote sensing of the cryosphere, cryosphere-climate
relations, and sea level in relation to polar ice sheet stability. A 1-day
GLIMS workshop plus 1 day of algorithm-use training classes may be
considered.
(4) Arctic conference, to be held in Norway, late August 2004.
Possibilities exist for field trips to Svalbard as well as to fjord
glaciers on the mainland. A suggestion, developed further after the
workshop, is to piggyback a major GLIMS theme-based meeting. In
addition to basic GLIMS business and RC reports (1 day), there would be
algorithm training classes (1 day), and a 1-day public forum on the theme
of "Anomalous Stable and Advancing Glaciers: What do they Mean?"
(5) Antarctic sea ice and glaciers, to be held in Hobart (Tasmania) in
2005.
Realistically, we would seek to hold only 1 of 2 or the possible 2003
meetings, plus a meeting in each of 2004 and 2004. Besides casting your
votes, do you know of other meetings that should be considered?
2. SPECIAL JOURNAL ISSUE.
This was discussed, but there was not overwhelming interest-- about 6
people
indicated that they would submit a paper to the Journal of Glaciology if
the
opportunity was available. If there is significant additional interest
outside of the workshop group, we can consider this route. So as not to
hold up publication, we would go ahead with a publish decision after the
first 8 papers had been accepted, and those that linger a month beyond that
would have to be published separately. We will certainly revisit the issue
of a special journal issue or book in future years, though certainly GLIMS
is not yet nearly ready for a book. Your interest (or preference
NOT to have a special issue yet) is solicited.
3. REGIONAL CENTER WEB PAGES.
Most RCs and stewards already have web pages. In this case, for minimal
new work, all we ask is that you provide us with the URL, and institute
a hotlink to the GLIMS website; we in turn will provide a hotlink to your
site. Ideally, the design would be such that the transition from one
site to another is fairly invisible, except that it should be clearly
indicated on your home page that this is your site, not the USGS site,
and it should be easy to return to the USGS GLIMS site without searching
hard for a return link. Alternatively, you may provide us directly with
your web pages, and we can incorporate them directly into our site; again,
proper credit to yourselves is vital.
4. DATABASE SUBMISSION NEEDED FROM EACH RC
There is a pressing need for near-term ingests of data--minimally for
one glacier per regional center-- into the GLIMS database. For this
trial ingest, the glacier analysis results can be from any method
whatsoever-- remote sensing or new field work or digitized from an old
topographic map. The important thing is that each RC should now become
familiar with the database. We have set a deadline date of October 5
for first contact with NSIDC-- either with a data submission, or with
an inquiry to Bruce Raup attempting to resolve any problem that may
exist. BRUCE: ADD CONTACT, DATABASE INFO
---------------------------------------------------------------------------
Working group breakout summaries:
Database (compoiled by Raup, Database/Parameters WG leader):
The following items were identified as high-priority issues for the
database/parameters working group:
- Identify sensible set of glacier characteristics -- improved over WGMS
and existing GLIMS. Should be able to combine different descriptors
for one glacier: both piedmont and calving. Fabian Mauz's pictorial
dictionary of glacier types, forms, is a good prototype "cookbook" to
describe glacier parameters.
- We need "cookbook" guidelines!
- Definitions of various parameters need to be made more clear: glacier
width, length, area, speed (where?), snowline. In all these cases, we
need
segments which show the future user where the quantity was measured.
Check WGMS definitions in each case.
- (mean) aspect and slope, in accumulation and ablation areas, should be
added to the database. Possibly allow different definitions of mean
slope: mean of cells in raster DEM; overall slope calculated as
rise/run.
- Defining glacier boundaries, especially where rock outcrops intersect
basin boundaries.
- Add category "basinbnd", if it's not there already.
- Category: "non-glacier-outline" (rocks, water) vs "glacier-outline"
(debris). At top level, could have 3 categories: basin boundary, ice
boundary, and non-ice boundary. This is so we can easily distinguish
between non-glacier segments (rocks) and supraglacial segments
(centerlines, lakes, debris, etc.).
--Category: Other profile data types required, such as surface profile
(segment),
basement profile, field point measurements
- When filling in data, the operator/database should make a distinction
between "unclassified" and "no data".
Algorithms (compiled by Andy Kaeaeb, Algorithms WG leader):
The algorithm working group will follow two tasks:
(1) glacier mapping techniques
(2) higher level products such as DEMs, ice velocity etc.
Task (1) has clearly the higher priority, task (2), presently, involves
mainly a few specialists who work together bilaterally.
Task (1) includes all kinds of boundaries between ice, snow, rock,
debris, ice- and snow facies etc. (Detailed list will soon be on the
algorithm WG web site).
Next steps for task (1)
- An algorithm document will be drafted by A. Kaeaeb and F. Paul, and
announced via the algorithm mailing list within the next weeks.
- Feedback and discussion by algorithm WG
- Selection of some "most promissing" algorithms and "algorithm battle"
- Detailed examples and results for the selected algorithms on the
web; discussion
- Detailed "cookbooks" for the algorithms selected suitable for GLIMS
To be included in the algorithm mailing list please contact Bruce Raup.
If you are interested in task (2) topics such as ASTER DEMs, glacier flow
fields etc. please contact directly A. Kaeaeb.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Jeffrey S. Kargel
U.S. Geological Survey
2255 N. Gemini Dr.
Flagstaff, AZ 86001
Telephone (+1) (928) 556-7034
Fax (+1) (928) 556-7014
Email: jkargel at usgs.gov
Home email: jkargel at flaglink.com
Home phone: (+1)(928) 527-4196
More information about the GLIMS
mailing list