This document discusses global patterns of moraine-dammed glacial lake outburst floods (GLOFs) and the challenges of attributing them to climate change. It notes that while temperatures are rising in mountain regions worldwide, the recorded number of GLOFs has fallen with no clear trend in magnitude. Attributing GLOFs to anthropogenic global warming is difficult due to issues modeling small-scale mountain processes and a lack of complete global data on GLOF events over time. The development of improved models at appropriate scales and a more robust global GLOF inventory are needed to better understand the relationship between climate and these extreme geomorphic events.

1. Global pattern of moraine-dammed
Gacial Lake Outburst Floods
Stephan Harrison
John Reynolds (Aberystwyth)
Christian Hugel (Zurich)
Jeff Kargel (Arizona)
Andy Wiltshire (UKMO)
Richard Betts (UKMO)
Liam Reinhardt (Exeter)
hulme Trust Helix

2. Rationale and structure
• AR5 stress on adaptation to climate change
• New focus on Detection and Attribution as important for
extreme events (includes litigation as well as development)
• Important debates on Equilibrium Climate Sensitivity,
Transient Climate Response and policy implications
• Geomorphological sensitivity and D&A for GLOFs
• Stress global scale approach needed

3. Climate Change: Detection
Bottom: estimate of uncertainty
for one dataset (black). Anomalies
are relative to the mean of
1961-1990.
Energy accumulation (1021 J on
Earth from 1971-2010. Upper
ocean is 700m. Estimates below
2000m started in 1992.

4. The Radiative Forcing of
WMGHG in AR5 is 2.83 ±
0.28 Wm–2
Milankovitch cycles
around 1.5 Wm–2
The range of
climate sensitivities in
climate models used for
IPCC AR4 is a warming of
0.6–1.2K per Wm–2 of
forcing.
Attribution: Global average radiative forcing
estimates and ranges

5. We can show a response to this:
• Glacier recession
• Paraglaciation
• GLOFs
– Important as responsible for thousands of deaths per
year
– Huge infrastructure costs
– Implications for economic and social development in
remote mountain regions

6. Global glacier recession
From 169 glacier length records
(Oerlemans 2006)

7. West Rongbuk Glacier in
1921 and 2008.
(RGS & David Breashears)
Imja Glacier, central
Himalaya in 1955 and
2009
(Fritz Müller and Alton Byers).

8. Does paraglaciation become the
dominant mechanism of
sediment movement in
glaciated mountain systems?
Implications for changes in the
magnitude/frequency of natural hazards?

9. Paraglacial....
landsystems and
processes driven by
deglaciation
Knight and Harrison,
2013,2014

10. GLOFs seen as paraglacial
phenomena and as a non-linear
response to warming and glacier
recession
No global assessment
Major problem in parameterising
RCMs……so development of predictive
tools difficult
ICIMOD

11. 11
Nare Glacier GLOF, Khumbu Himal, Nepal ASTER image, Bhutan & Tibet
Questions: Can we detect a climate signal in GLOF
development? Will Climate change affect the
magnitude/frequency of GLOFs?

12. Three issues have to be addressed
• Climate sensitivity
• Geomorphological sensitivity
• Attribution and glacial systems

13. How much will the temperature warm? i.e.
what is climate sensitivity?
AR5 definitions:
Equilibrium climate sensitivity (ECS): “defined as the equilibrium
change in global surface temperature following a doubling of the
atmospheric equivalent CO2 concentration”*
Transient climate response (TCR): “the change in global surface
temperature in a global coupled climate model in a 1%yr-1 CO2
increase experiment at the time of CO2 doubling”*
*Note: Units are in degrees Celsius (°C).

14. PDFs, distribution and ranges
for ECS (IPCC AR5)
• no central figure given
• lower limit may be reduced
(problem of using recent data sets to
constrain this?)
• long tail…..although figures greater
than 6°C unlikely (and palaeo record
doesn’t suggest this)
• TCR lower than this….but may be
more relevant for adaptation policy

15. What is
‘geomorphological
sensitivity’?
GLOFs become valuable
metric with which to test
‘geomorphological
sensitivity’ and to discuss
uncertainties
1. Is there a warming
trend in mountain
regions?
2. Can we detect trends
in magnitude and
frequency of GLOFs?ICIMOD

16. However, Attribution problem....How can we
Attribute Extreme Events to Human Forcing
• Difficult problem to solve because:
– Relates to a set of small scale events
– These may be stochastic and/or display chaos
– Understanding them depends on assessing nature of
initial conditions as well as forcing (i.e. the precise
conditions of the system before the event began as
well as whatever might be acting on the system)
– To model events we need probabilistic framework...lots
of model runs (ensembles)….and with a suitable
model!

17. Event based attribution
Approach:
Find out how many times
a specific event of interest
occurs in a model with
predindustrial GHG forcing
(‘natural world’).
Find out how many times
a specific event of interest
occurs in a model with
current levels of GHG
forcing (‘actual world’).
Assess the difference
(Fraction of Attributable
Risk).
Requires:
a model capable of simulating
the event(s) of interest
many runs of the model
FAR = 1 – P0 / P1

18. Problem of running meaningful model
ensembles which can help develop attribution
for GLOFs
Very difficult in mountains:
• Large changes in albedo, water vapour fluxes, ice temperature,
supraglacial debris fluxes, landcover etc over small spatial scales
• Topography difficult to model and model resolution still much too
coarse
• How to model paraglacial triggers?
• How to model initial conditions?

19. Tasman Glacier, New
Zealand
Charles Warren
Attribution and debris
covered glaciers
Debris obscures the
glacier-climate
signal...and difficult to
model and parameterise
Baltoro Glacier

20. The GLOF record

21. • We need a Global record of GLOFs and a more complete data
set than hitherto.
In Zailiskiy Alatau GLOFs
concentrated in 1960s-
90s…few in recent years
Passmore, Harrison et al
2008. AAAR

22. Developing a record for D&A was the
motivation for the GLOF inventory
• 567 glacier hazard events had to be cross-checked
• Problem is that many events named after the
glacial lake in some datasets; in others the same
event named after the glacier or mountain region
• Eventually these whittled down to 191 GLOFs
associated with moraine dam failure…best candidates
for Atrribution

24. We can show that
world’s mountains are
warming….but data
resolution is 5°
Black crosses are
glaciers from WGI-XF
Harrison et al. in
submission

25. • For Attribution….we need integration between
glacier and climate models (RCMs) at scales
where debris supply, lake initiation, trigger
mechanisms can be modelled.
• And we need an uncensored data base of GLOF
magnitude and frequency in the past.
• Pattern so far shows rising temperatures in the
world’s mountains but the number of GLOFs has
fallen and there is no trend in magnitude.

26. Conclusions
• Attributing GLOFs to AGW too early because attribution
studies have not been done (and are problematic).
• Modelling at scale required and in mountain regions also
difficult. Parameterisation of crucial processes needed.
• No global GLOF inventory and data sets are incomplete
(so Detection difficult)
• However, an important task….for adaptation policy,
litigation issues (Climate Justice Programme) and
assessing impact of climate on extreme geomorphological
events.