Interactive framework for global dialogue, connecting science

1
Java Climate Model jcm.chooseclimate.org
Interactive framework for global dialogue,
connecting science policy, emissions
impacts, UNFCCC articles 2 and 6.
Steering our ship, to stabilise concentrations
of greenhouse gases... to avoid dangerous
anthropogenic interference in the climate system
(the aim of UNFCCC, Article 2) requires balancing
many risk and value judgements. Computer models
help by providing a quantitative framework to
resolve complex interactions, but remain a
mysterious black box to all but a few experts.
Yet the ultimate integrated assessment model is
the global network of human heads, and effective
implementation of any global agreement relies on
active support of citizens around the world. So
to democratise climate science and policy we
need to improve the accessibility of models, as
well as outreach and feedback with the public
(the task of UNFCCC Article 6).
The author, responding to changing weather to
steer to a safe landing?
Dr Ben Matthews ben_at_chooseclimate.org
(mobile)32 472 987028 (ring me to arrange a
demonstration) Currently based at Institut
d'Astronomie et de Géophysique, Université
Catholique de Louvain, Belgium (working with Prof
Jean-Pascal vanYpersele) JCM was also developed
with Klima UmweltPhysik, University
Bern DEA-CCAT/Energimiljoradet, Copenhagen
UNEP-GRID Arendal, Norway
The Java Climate Model (JCM) was created for this
process. It enables anybody to explore both
climate policy options and scientific
uncertainties, using the same models as IPCC-TAR,
simply by adjusting parameter controls with a
mouse in a web browser. The instant mechanical
response shown on linked graphics demonstrates
cause and effect, and the sensitivity to choices,
assumptions, and risk/value judgements.
Everything is interconnected... People want to
know how local emissions which they can control,
may influence the vast global natural and human
systems, and so change local climate impacts
which affect them directly. In the example below,
the main cause-effect chain goes anti-clockwise
from the top, but total emissions are calculated
by an inverse iteration method aiming to
stabilise temperature at 2C above preindustrial
(as below).
...from regional emissions... (upper-right plots,
curves for 12 regions, colours as map, from
1900-2100) JCM currently includes socio-economic
data for 12 regions, for each SRES scenario
(adapted from IMAGE). This may be combined with
various distribution formulae to calculate
regional CO2 emissions. In the example, CO2
emissions follow Kyoto targets to 2012, then
converge to equal per-capita levels by 2050,
using the population fromSRES A1B.
...via core global science... (lower-left plots
, from 1750 to 2300). JCM includes an efficient
java implementation of the simple models and
formulae as used for IPCC-TAR projections.
Atmospheric chemistry and radiative forcing
includes all greenhouse gases, plus ozone and
aerosol precursors (total 35 gases). The carbon
cycle comes from the Bern model. There is a
feedback from climate to carbon, coupled within
one time-loop. The heat flux calculations are
based on the Wigley-Raper UDEB model with
parameters fitted to seven GCMs as in TAR. The
slow penetration of carbon and heat into the
oceans is calculated by upwelling-diffusion
models solved by an eigenvector method. This
includes the temperature-upwelling feedback
affecting thermal expansion, to which ice-melt is
added to calculate sea-level rise.
Correspondence with IPCC predictions may be
demonstrated by superimposed data.
The attribution plot (upper right) illustrates an
integrated application of JCM, for a recent
UNFCCC model intercomparison exercise regarding
the attribution of responsibility for climate
change due to different sources of emissions (the
Brazilian proposal). As well as being a
scientific challenge, attribution of effects from
processes which combine non-linearly raises some
interesting intergenerational and inter-regional
equity issues. (note, the attribution
calculations are not yet adapted for
stabilisation scenarios, so this plot is not
consistent with others in this example)
Multipurpose tool JCM has an flexible structure
of interacting modules, efficiently recalculating
only what is both needed for plots and changed by
controls. The interactions may be shown by a
self-adjusting flowchart, clicking on which also
summons relevant web documentation and java
source code. The code is entirely open source.
The 98-curve plots (left) were generated by JCM
in only a few minutes, using internet explorer,
and a simple text scripting code. As well as for
batch calculations in problem-solving frameworks,
this multi-purpose code may also be used to
create automatic demonstrations, interacting with
web documentation to illustrate specific points.
Such demonstrations or snapshots may be sent
between users across the web, allowing some
remote controlwhich may help with teaching or
global dialogue. The code uses the same
object-label system as the pop-up information
about controls, curves, and units, which is now
available in ten (human!) languages. So JCM may
be used both to investigate new questions (as
left), and to explain existing knowledge in a
novel interactive way. It has already been used
for teaching students in several countries, as
well as at policy meetings. Maintaining a balance
between scientific robustness and flexibility on
the one hand, and simplicity and interactivity on
the other, is not easy. However it is essential
that we make such an effort, to ensure that
serious climate models become more transparent
and accessible. JCM works on most web browsers
and is very compact, loading in seconds. You can
also download packages, including documentation
and source code, to use offline. The
interactive model illustrates itself and the
climate system much better than a static poster.
So please experiment for yourself!
...to regional impacts The regional climate map
(centre) shows recent HadCM3 GCM data, scaled to
JCM global average, and added to the baseline
climatology. The colors mix maximum-temperature
(red), minimum temperature (blue), and
preciptitation (green). Averages are calculated
for each country or subregion. Everything
recalculates instantly, as you adjust any
parameter. Many GCM datasets are available from
IPCC-DDC, whose visualiser is now being updated
using the same java code.
Stabilisation Scenarios Interpretation of Article
2 (above) includes many factors nonlinear
thresholds in the global climate response,
regional impacts on ecological, agricultural and
socioeconomic systems, and the achievability of
emissions mitigation pathways. The global debate
to balance these factors requires insight from
complex models, however it is likely to focus on
a few simple indicators. So JCM includes options
to stabilise atmospheric CO2, radiative forcing
(all gases), or global temperature, at any level
and year. For example, European Union
policymakers proposed that warming should not
exceed 2C (above preindustrial), and therefore,
that CO2 should be stabilised at less than
550ppm. Using average TAR parameters, and
mitigating 21 gases, JCM shows that to be
scientifically credible this target must be
interpreted as 550ppm CO2 equivalent, including
forcing from all gases (with CO2 alone about
450-500ppm). Considering also convergence between
regions, this implies reducing EU emissions by
about 75 by 2050. Large uncertainties affect
such inverse calculations. The figures (right)
show 98 CO2 emissions pathways leading
tostabilisation of temperature at 2C, combining
the seven GCMs (parameterised as in the TAR),
with the six SRES scenarios (plus one using 2000
levels), with an without mitigation of other
gases. The wide range poses a challenge for
policymakers! However, selecting a temperature
rather than an equivalent concentration target
shifts the burden of managing uncertainty away
from the receivers of climate impacts
(particularly more vulnerable poor countries),
towards the controllers of emissions (mostly from
rich countries). When designing an efficient
iteration algorithm for inverse calculations, the
correction-feedback process is more important
than the initial guess. This also applies to the
global iteration between scientists, policymakers
and citizens, essential for interpreting Article
2. So we should not fear making bold guesses, but
need to design better feedback in the global
dialogue.
The redder set of 49 curves have no mitigation of
non-CO2 gases, whilst the other 49 assume that
emissions of each gas (including aerosol and
ozone precursors) are reduced by an equal
proportion, compared to the SRES baseline in each
year.
CO2 Emissions (fossil landuse) GtC/yr
Global Average Temperature
The same stes of curves as above. The greener
curves correspond to cooler GCMs. The baseline is
preindustrial, hence there is already divergence
at 2000. For a few scenarios (hottest model
GFDL large other gas emissions in SRESA2) the
iteration failed to find a pathway below 2C