PowerPoint-Pr

1
How do we know that human influence is changing
(regional) climate?
Hans von Storch Institute for Coastal Research,
GKSS Research Center, Geesthachtand Meteorological
Institute, Hamburg University
2
Questions about ongoing non-natural change

• Global anthropogenic change - An argument
  for efforts to mitigate climate change by
  diminishing drivers (political asset)
• Regional anthropogenic change need to
  discriminate between global and regional
  drivers. An argument for efforts to mitigate
  regional change by diminishing regional drivers
  (political asset), or an argument to
  implement adaptive measures to deal with
  changing risks and opportunities (information
  for stakeholders)

3
Detection and attribution of ongoing change
Omstedt, 2005
4
Detection and attribution of non-natural ongoing
change

• Detection of the presence of non-natural
  signals rejection of null hypothesis that recent
  trends are drawn from the distribution of trends
  given by the historical record. Statistical
  proof.
• Different definition Detection is the process
  of demonstrating than an observed change is
  significantly different (in a statistical sense)
  than can be explained by natural internal
  variability (IPCC, TAR, 2001 see also IDAG,
  2005)
• Attribution of cause(s) Non-rejection of the
  null hypothesis that the observed change is made
  up of a sum of given signals. Plausibility
  argument.
• History
• Hasselmann, K., 1979 On the signal-to-noise
  problem in atmospheric response studies.
  Meteorology over the tropical oceans (B.D.Shaw
  ed.), pp 251-259, Royal Met. Soc., Bracknell,
  Berkshire, England. Hasselmann, K., 1993
  Optimal fingerprints for the detection of time
  dependent climate change. J. Climate 6, 1957 -
  1971 Hasselmann, K., 1998 Conventional and
  Bayesian approach to climate change detection and
  attribution. Quart. J. R. Meteor. Soc. 124
  2541-2565IDAG, 2005 Detecting and attributing
  external influences on the climate system. A
  review of recent advances. J. Climate 18,
  1291-1314

5
Global
6
Cases of Global Climate Change Detection Studies
of strong, well documented signals Examples
1) Rybski et al. (2006) 2)
Counting recent extremes of weak, not well
documented signals. Example Near-globally
distributed air temperature IDAG (2005), Hegerl
et al. (1996), Zwiers (1999) Rybski, D., A.
Bunde, S. Havlin,and H. von Storch, 2006
Long-term persistence in climate and the
detection problem. Geophys. Res. Lett. 33,
L06718, doi10.1029/2005GL025591IDAG, 2005
Detecting and attributing external influences on
the climate system. A review of recent advances.
J. Climate 18, 1291-1314 Hegerl, G.C., H. von
Storch, K. Hasselmann, B.D. Santer, U. Cubasch,
P.D. Jones, 1996 Detecting anthropogenic climate
change with an optimal fingerprint method. J.
Climate 9, 2281-2306 Zwiers, F.W., 1999 The
detection of climate change. In H. von Storch
and G. Flöser (Eds.) Anthropogenic Climate
Change. Springer Verlag, 163-209, ISBN
3-540-65033-4
7
The Rybski-et al. study
Rybski, D., A. Bunde, S. Havlin,and H. von
Storch, 2006 Long-term persistence in climate
and the detection problem. Geophys. Res. Lett.
33, L06718, doi10.1029/2005GL025591
8
Rybski, D., A. Bunde, S. Havlin,and H. von
Storch, 2006 Long-term persistence in climate
and the detection problem. Geophys. Res. Lett.
33, L06718, doi10.1029/2005GL025591
- Statistics of ?T(m,L) which is the difference
of two m-year NH temperature means, separated by
L years. - Temperature variations are modeled as
Gaussian long-memory process, fitted to the
various reconstructions.
9
Counting extremely warm years

• Among the last 16 years, 1991-2006, there were
  the 12 warmest years since 1881 (i.e., in 126
  samples) how probable is such an event if the
  time series were stationary?
• Monte-Carlo simulations taking into account
  serial correlation, either AR(1) (with lag-1
  correlation ?) or long-term memory process (with
  Hurst parameter H0.5d).
• Best guesses
• ? 0.8
• H 0.5 d ? 0.50.3 (??)

Joint unpublished work by Zorita, Stocker and von
Storch, 2007
10
How do we determine the control climate?
In general, the data base for the
control/undisturbed climate is not good

• With the help of the limited empirical evidence
  from instrumental observations, possibly after
  suitable extraction of the suspected
  non-natural signal.
• By projection of the signal on a proxy data
  space, and by determining the stats of the latter
  from geoscience indirect evidence (e.g., tree
  rings).
• By accessing long control runs done with
  quasi-realistic climate models

11
Signal or noise?
Trend in air temperature 1965-1994 1916-1945
Hegerl, G.C., H. von Storch, K. Hasselmann, B.D.
Santer, U. Cubasch, P.D. Jones, 1996 Detecting
anthropogenic climate change with an optimal
fingerprint method. J. Climate 9, 2281-2306
12
Reducing the degrees of freededom
Specific problem in climate applications usually
very many (gt103) degrees of freedom, but the
signal of change resides in a few of these
degrees of freedom. Example Signal (2, 0, 0,
...0) with all components independent. Power of
detecting the signal, depends on degrees of
freedom.
Thus, the dimension of the problem must be
reduced before doing anything further. Usually,
only very few components are selected, such as 1
or 2.
13
Guess patterns
The reduction of degrees of freedom is done by
projecting the full signal S one or a few several
guess patterns Gk, which are assumed to
describe the effect of a driver. S ?k ?k Gk
n with n undescribed part. When Gk orthonormal
then ?k ST?Gk.
Example guess pattern supposedly representative
of increased CO2 levels
14
Hegerl, G.C., H. von Storch, K. Hasselmann, B.D.
Santer, U. Cubasch, P.D. Jones, 1996 Detecting
anthropogenic climate change with an optimal
fingerprint method. J. Climate 9, 2281-2306
15
Optimizing s/n ratio
Optimization of the expected signal to noise
ratio with the inverse covariance matrix of
the internal climate variability.
Hegerl, G.C., H. von Storch, K. Hasselmann, B.D.
Santer, U. Cubasch, P.D. Jones, 1996 Detecting
anthropogenic climate change with an optimal
fingerprint method. J. Climate 9, 2281-2306
16
The attribution problem

• Attribution is considered to be obtained, when
• the suspected link between forcing and response
  is theoretically established, and
• the data do not contradict that ?k1 in the
  assumed representation S ?k ?k Gk n.
• A contradiction prevails if the null hypothesis
  ?k1 is rejected.
• Thus, a non-contradiction is a plausibility-argume
  nt. It may be due to a too small data base.

17
Attribution

• 2-patterns problem (Hegerl et al. 1997)
• guess patterns for climate change mechanisms
  taken as first EOFs of a climate change
  simulation on that mechanism.
• only CO2 increase
• increase of CO2 and industrial aerosols as well.
• orthogonalisation of the two patterns
• estimation of natural variability through GCM
  control simulations done at MPI in Hamburg, GFDL
  in Princeton and HC in Bracknell.

18
Example Attribution
Attribution diagram for observed 50-year trends
in JJA mean temperature.
Zwiers, F.W., 1999 The detection of climate
change. In H. von Storch and G. Flöser (Eds.)
Anthropogenic Climate Change. Springer Verlag,
163-209, ISBN 3-540-65033-4
The ellipsoids enclose non-rejection regions for
testing the null hypothesis that the
2-dimensional vector of signal amplitudes
estimated from observations has the same
distribution as the corresponding signal
amplitudes estimated from the simulated 1946-95
trends in the greenhouse gas, greenhouse gas plus
aerosol and solar forcing experiments.
19
Attribution - plausibility
From Hadley Center, IPCC TAR, 2001
20
Regional the Baltic Sea catchment
21
The Baltic Sea Catchment Assessment BACC
An effort to establish which knowledge about
anthropogenic climate change is available for the
Baltic Sea catchment. Working group BACC of GEWEX
program BALTEX. Approximately 80 scientist from
10 countries have documented and assessed the
published knowledge. Assessment has been accepted
by intergovernmental HELCOM commission as a basis
for its future deliberations.
22
The Baltic Sea Catchment Assessment BACC

• Summary of BACC ResultsBaltic Area Climate
  Change Assessment
• Presently a warming is going on in the Baltic
  Sea region.
• No formal detection and attribution studies
  available.
• BACC considers it plausible that this warming is
  at least partly related to anthropogenic factors.
• So far, and in the next few decades, the signal
  is limited to temperature and directly related
  variables, such as ice conditions.
• Later, changes in the water cycle are expected
  to become obvious.
• This regional warming will have a variety of
  effects on terrestrial and marine ecosystems
  some predictable such as the changes in the
  phenology others so far hardly predictable.

BACC Group Assessment of climate change for the
Baltic Sea basin, Springer-Verlag, in press
23
Significant trends

• Often,an anthropogenic influence is assumed to be
  found when trends are found to be significant.
• In many cases, the tests for assessing the
  significance of a trend are false as they fail to
  take into account serial correlation.
• If the null-hypothesis is correctly rejected,
  then the conclusion to be drawn is if the data
  collection exercise would be repeated, then we
  may expect to see again a similar trend.
• Example N European warming trend April July
  as part of the seasonal cycle.
• It does not imply that the trend will continue
  into the future (beyond the time scale of serial
  correlation).
• Example. Usually September is cooler than July.

24
Significant trends
Establishing the statistical significance of a
trend is a necessary condition for claiming that
the trend would represent evidence of
anthropogenic influence. Claims of a continuing
trend require that the dynamical cause for the
present trend is identified, and that the driver
causing the trend itself is continuing to
change. Thus, claims for extension of present
trends into the future require- empirical
evidence for ongoing trend, and- theoretical
reasoning for driver-response dynamics, and-
forecasts of future driver behavior.
25
Consistency analysisattribution without detection
The check of consistency of recent and ongoing
trends with predictions from dynamical (or other)
models represents a kind of attribution without
detection. This is in particular useful, when
time series of insufficient length are available
or the signal-to-noise level is too low. The idea
is to estimate the driver-related change E from a
(series of) model scenarios (or predictions), and
to compare this expected change E with the
recent trend R. If R ? E, then we may conclude
that the recent change is not due to the changing
driver, at least not completely.
26
Measures of similarity
27
Consistency analysis Seasonal precip in the
Baltic Sea catchment
Example Changing DJF-mean precipitation in the
Baltic Sea catchment Trend of precip according
to different data sources.
28
Consistency analysis

• Expected signals E
• six simulations with regional coupled
  atmosphere-Baltic Sea model RCAO (Rossby-Center,
  Sweden)
• three simulations run with HadCM3 global
  scenarios, three with ECHAM4 global scenarios
  2071-2100
• two simulation exposed to A2 emission scenario,
  two simulations exposed to B2 scenario 2071-2100
• two simulations with present day GHG-levels
  1961-90
• Climate change in the four scenarios relatively
  similar.

29
Consistency analysis
30
Consistency analysis
Patterns correlations between observed (CRU)
trends in DJF seasonal precipitation in the
Baltic Sea catchment and expected signals
derived from scaled RCM changes.
Global model scenario Pattern correlations Pattern correlations without NAO
HadAM A2 0.83 0.75
HadAM B2 0.75 0.64
ECHAM A2 0.85 0.75
ECHAM B2 0.84 0.74
The pattern correlations are all significantly
larger than pattern correlations between random
combinations of trends.
31
Consistency analysis
Ratio of intensities between observed (CRU)
trends in DJF seasonal precipitation in the
Baltic Sea catchment and expected signals
derived from scaled RCM changes. All model
predictionsresult in too largetrends for the
past years.When taking out theNAO the
situationslightly improves.
Global model scenario Intensity-ratio ??R??/??E?? Intensity-ratio without NAO
HadAM A2 2.96 2.53
HadAM B2 4.50 3.98
ECHAM A2 1.94 1.57
ECHAM B2 2.50 2.07
32
Consistency analysis Baltic Sea catchment

1. In DJF patterns of trends in observed CRU data
   consistent with patterns of trends predicted by
   RCAO-scenario.
2. Intensity of trends in 1976-2005 considerably
   larger than predicted by RCAO scenarios.
3. When talking out the NAO, the dominant pattern of
   natural variability, the over-prediction of the
   intensities is somewhat smaller.
4. Possible causes- scenarios inappropriate
   (false)- drivers other than CO2 at work
   (industrial aerosols?)- natural variability much
   larger than signal (signal-to-noise ratio ?
   0.2-0.5).
5. Similar situation in spring.
6. No consistency in summer and fall.

33
Overall summary

• How do we know that human influence is changing
  (regional) climate?
• Statistical analysis of ongoing change with
  distribution of naturally occurring changes
  detection, statistical proof.- ok für global and
  continental scale temp.
• Consistency of continental temp change with
  change in regions such as Baltic Sea catchment
  (temp and related variables see Jonas
  presentation)

34
Overall summary

• How do we know that human influence is changing
  (regional) climate?
• Attribution (of causal drivers) is a
  plausibility argument determine consistency of
  ongoing change with expected changes.
• Done for global and continental scale temp (and
  related) variables (see IDAG).
• First efforts on regional scales (see JonasÄ
  presentation).

35
(No Transcript)