The Fuzzy Nature of Climate Change Scenarios Maps
Carlos Gay García and Oscar Sánchez Meneses
Centro de Ciencias de la Atmósfera, Universidad Nacional Autónoma de México,
Ciudad Universitaria, Mexico. D.F., Mexico
Keywords: Climate Change Uncertainties, Fuzzy Temperature and Precipitation, Linear Emission Pathways.
Abstract: The most important uncertainties present in the global change scenarios are the climate sensibility,
represented by the wide variety of GCM´s available, and the uncertainty that comes from the different GHG
emission scenarios. Starting from a fuzzy climate model constructed with concentrations of GHG, obtained
as a result of linear emission pathways, and output temperatures obtained with a deterministic simple
climate model (MAGICC) it has been determinate the output fuzzy set of global delta T thresholds such as
1, 2, 3 and 4 °C for 2100 and a medium sensibility of 3.0 °C/W/m2. These fuzzy sets are used for assign
uncertainties to values of temperature increase and precipitation change percentage taken from a map of
regional climate change and for interpret the map in a fuzzy sense. We present some maps of temperature
increase and precipitation change percentage for Mexico.
1 INTRODUCTION
Many previous works have been published about the
topic of climate change scenarios caused by global
warming, as can be verified by reviewing the latest
three assessment reports of the Working Group I of
the Intergovernmental Panel on Climate Change
(IPCC 2001, 2007 and 2013, all of them available at
http://www.ipcc.ch/).
There have been various proposals for physical
climate change scenarios based primarily on
different ways to estimate future greenhouse gas
(GHG) and, also, a collection of general circulation
models with atmosphere and ocean coupled
(AOGCM) or updated versions of these.
The scenarios are regularly presented as maps,
grid or contours, with a certain spatial resolution,
and in different spatial domains as global, regional
or local. Different GHG emission scenarios,
AOGCM's and time horizons for climatic variables,
such as temperature and precipitation, are also
considered (Conde et al., 2011).
However, the manner to interpret the scope of
these projections is complicated by the fact that
there are different sources of uncertainty associated
with the various inputs used in the development of
climate change scenarios. It is especially difficult for
both, scientists and decision makers, to take into
account the predominantly epistemic nature of
uncertainties in climate change to design adaptation
strategies or mitigation measures, so sometimes,
statistical methodologies, that may not be the
appropriate, are used (Gay and Estrada, 2009).
In previous works (Gay, et al., 2012, 2013, Gay
and Sánchez, 2013) it has been explored the use of
fuzzy logic in the representation and interpretation
of the uncertainties of climate change scenarios,
because this formulation allows the natural
inclusion, through linguistic rules, of the different
sources of uncertainty; for fuzzy logic no concept
have precise limits (Zadeh 1965).
Using models of type FIS (Fuzzy Inference
System) has been achieved to relate, via linguistic
rules of IF-THEN form, fuzzy sets associated to
values of climate sensitivity and GHG emissions
with the values of the global temperature provided
as output variable from AOGCM´s (Gay, et al.,
2012, 2013).
It is by means of changes in temperature global,
regional or local, that other climatic variables and
the different climate subsystems show the effects of
climate change (Gay and Sánchez, 2013).
Based on the fuzzy model presented by Gay et
al., (2013) and the simple climate model contained
in Magicc/Scengen (Wigley, 2008) we show how
the global mean temperature increase is distributed
on the globe for the significant thresholds of 1, 2, 3
and 4 °C. The linear emission pathways, used to
build the fuzzy model, include all the possibilities
mentioned in successive reports of IPCC.
863
Gay García C. and Sánchez Meneses O..
The Fuzzy Nature of Climate Change Scenarios Maps.
DOI: 10.5220/0005141208630873
In Proceedings of the 4th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (MSCCEC-2014), pages
863-873
ISBN: 978-989-758-038-3
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
In this work we consider the possibility of analyzing
the impacts, at the regional level, of temperature
increase and precipitation changes from the
perspective of the year in which some temperature is
reached. Two sources of uncertainty are taken into
account, the emissions of GHG and the climate
sensitivity.
We have learned that the larger concentration
and sensitivity the sooner the successive thresholds
of temperature will be reached. If the sensitivity is 6
°C/W/m2 there is no way of staying at two degrees
unless the concentrations of CO2 had followed the -
2CO2 trajectory: negative emissions that means very
strong subtraction of CO2 from the atmosphere. We
think that it is easier to consider a degree by degree
strategy than one based on dates.
In Gay et al., (2013) were presented maps for 2
GCM´s (as an example) with the necessary
concentration to reach 1, 2, 3 and 4 °C limits to
2100. The maps show the spatial distribution of the
temperature increase over the globe.
Emissions and sensitivity introduce uncertainties
in the temperature that in turn must be reflected in
the scaled temperature displayed in a map. Other
source of uncertainty considered is the GCM. The
maps constructed for different GCM´s illustrate all
possibilities for a region of the globe.
Additionaly, we present maps on a regional scale
for Mexico, corresponding to the global maps
mentioned above. It is evident that maps were
constructed from Magicc/Scengen for each GCM.
In this work we show how the GCM´s introduce
uncertainty in the estimates of precipitation change
in global and regional scales.
2 METHOD
We use the results reported by Gay and Sanchez
(2013) consisting in linear emission paths (Figure 1),
and the concentrations, forcings and temperatures
calculated with the use of the Magicc/Scengen up to
the year 2100, to discuss the timing of reaching a
warming of 1, 2, 3 and 4 degrees centigrade. To
illustrate how this can be done we observe from the
temperature profile that corresponds to the emission
path labeled 5CO2 (that is the linear profile whose
value in 2100 is five times the emissions in 1990),
when the curve crosses the 1, 2, 3 and 4 °C,
thresholds and look at the time when this happens
(see Figure 2). These dates depend on the sensitivity
used in the model and occur sooner as the sensitivity
increases.
Figure 1: Linear emission pathways as proposed by Gay et
al., (2012). As noted there, (-2) CO2 means -2 times the
emission (fossil + deforestation) of CO2 of 1990 by 2100
and so for -1, 0, 1, to 5 CO2. All the linear pathways
contain the emission of non CO2 GHG as those of the
A1FI and were inserted in MAGICC V.5.3 (Wigley,
2008). Here we include the emission scenario
corresponding to RCP 8.5 obtained from Magicc V.6.0
(Meinshausen, et al., 2011).
Figure 2: The corresponding global temperature
increments for emission pathways of Figure 1, calculated
from MAGICC V.5.3 and 6.0. Note the similarity between
curves A1FI and 5CO2, and also between curves RCP8.5
and 4CO2.
For example, one degree may happen as soon as
2021, 2 °C as soon 2039 as it is shown in Tables 1
and 2 of Gay and Sanchez (2013) reproduced here
for clarity. If the emissions followed more moderate
paths the dates of crossing the thresholds would be
delayed. The lesson from this is obvious: the lower
the emissions the later the thresholds would be
crossed and the more time we would have to adapt
to increased temperatures. Then again using the
Magicc/Scengen, maps were found for 1 to 4 °C.
These maps may serve as planning tools for
adaptation studies.
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
864
Since the temperature and concentration of the
A1FI and RCP8.5, of the IPCC are very close to our
profile corresponding to large emissions then the
timing can be applied to those profiles as well (as we
can see in Figure 2).
Table 1: Dates to achieve the 1 °C threshold following
linear emission trajectories from -2CO2 to 5CO2.
Emission
Trajectory
Sensitivity (deg C/W/m2)
1.5 3.0 6.0
-2CO2 2049
-1CO2 2057 2039
0CO2 2079 2048 2033
1CO2 2063 2042 2029
2CO2 2056 2038 2027
3CO2 2051 2035 2024
4CO2 2047 2032 2023
5CO2 2044 2030 2021
B1-IMA 2090 2043 2027
A1FI-MI 2046 2033 2024
Table 2: Dates to achieve the 2 °C threshold following
linear emission trajectories from -2CO2 to 5CO2.
Emission
Trajectory
Sensitivity (deg C/W/m2)
1.5 3.0 6.0
-2CO2
-1CO2 2073
0CO2 2100 (1.98°C) 2059
1CO2 2072 2052
2CO2 2064 2048
3CO2 2093 2058 2045
4CO2 2081 2054 2042
5CO2 2053 2051 2039
B1-IMA 2057
A1FI-MI 2076 2053 2042
In this work we emphasize the possibility of
analyzing the impacts of temperature increase from
the perspective of the year in which some
temperature is reached at the regional level. We
think that it is easier to consider a degree by degree
strategy than one based on dates. In other words
having information of the approximate dates in
which different thresholds, for example, one degree
would be crossed, would enable a policy maker to
act on those sectors or activities that would be
affected by one degree leaving for later those that
would be affected by larger temperature changes.
Knowing the timing can help the planning process.
Postponing the action complicates matters
because the uncertainties become larger. For
example the projections of temperatures and
precipitation in 2100 depend on two sources of
uncertainty, the emissions path of GHG and the
climate sensitivity. For example, in 2100 the
uncertainty associated to a one degree of global
increase extends to more than two degrees, then for
a 1 °C global increase, maps for one and two degrees
are to be considered. For 4 °C and sensitivity 3,
uncertainty can expand the temperature range to
6.41 °C.
Emissions and sensitivity introduce uncertainties
in the temperature that in turn must be reflected in
the scaled temperature and precipitation displayed in
a map. Other source of uncertainty considered is the
GCM itself. What we mean here is that the map for a
one degree increase given for the GFDL 2.0
(Geophysical Fluid Dynamics Laboratory Coupled
Model, version 2.0) is going to be slightly (or
seriously when talking of precipitation) different that
the one coming from the HADGEM1 (Hadley
Centre Global Environmental Model version 1). The
choice of AOGCM´s has been somewhat arbitrary.
3 REGIONAL VIEW OF GLOBAL
T OVER TEMPERATURE
AND PRECIPITATION IN
MEXICO
Maps on a regional scale for Mexico, corresponding
to the global maps mentioned above are constructed
by interpolation methods for temperature and
precipitation; we show some of them (Figures 3 to
10). Arguments mentioned above for global maps
apply as well to these.
The size of the grid in the data obtained from
Magicc/Scengen, is relatively big (2.5° x 2.5°) and,
for purposes of regionalization, we reduced it to 0.5’
x 0.5’ (about 10 Km x 10 Km) applying the method
of splines as presented in Conde et al., (2011).
The maps of regional climate change scenarios
for Mexico, over temperature and precipitation, are
presented according to each T threshold.
TheFuzzyNatureofClimateChangeScenariosMaps
865
Figure 3: Regional scenario for temperature at T
global
=
1.01 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
In this work we show how the GCM´s introduce
uncertainty in the estimates of precipitation change
in global and regional scales. This is simply due to
the fact that modeling strategies and
parameterizations differ from model to model. For
example Table 3 and 4 illustrate for 3 points in the
map the differences in the values of the temperature
and precipitation (for the same global temperatures)
at the same geographical position produced by
different models.
Figure 4: Regional scenario for temperature at T
global
=
2.02 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
866
Figure 5: Regional scenario for temperature at T
global
=
3.0 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
Figure 6: Regional scenario for temperature at T
global
=
4.02 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
TheFuzzyNatureofClimateChangeScenariosMaps
867
Figure 7: Regional scenario for precipitation at T
global
=
1.01 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
Figure 8: Regional scenario for precipitation at T
global
=
2.02 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
868
Figure 9: Regional scenario for precipitation at T
global
=
3.0 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
Figure 10: Regional scenario for precipitation at T
global
=
4.02 °C threshold, according to GFDL 2.0 (upper panel)
and HADGEM1 (lower panel) for 5CO2 emission
trajectory. Maps were obtained using Magicc/Scengen V.
5.3 data and MATLAB script.
TheFuzzyNatureofClimateChangeScenariosMaps
869
Table 3: Uncertainty produced by using model GFDL 2.0,
projected to 2100 with emission pathway of 5CO2.
Temperature and precipitation increments calculated for 3
points located NW (-108.75, 31.25), SW (-101.25, 21.25)
and Central (-88.75, 21.25) Mexico. Data obtained from
Magicc/Scengen V.5.3.
Delta T threshold (°C)
1 2 3 4
Temperature (°C)
0.90 1.92 2.80 3.75
0.92 1.87 2.80 3.77
0.61 1.32 2.12 2.87
Precipitation (%)
19.96 38.54 54.8 72.96
1.33 5.8 11.88 16.71
9.69 18.25 22.31 28.97
Table 4: Uncertainty produced by using model
HADGEM1, projected to 2100 with emission pathway of
5CO2, Temperature and precipitation increments
calculated for 3 points located NW (-108.75, 31.25), SW
(-101.25, 21.25) and Central (-88.75, 21.25) Mexico. Data
obtained from Magicc/Scengen V.5.3.
Delta T threshold (°C)
1 2 3 4
Temperature (°C)
1.54 3.12 4.45 5.95
1.16 2.33 3.44 4.61
0.68 1.45 2.29 3.1
Precipitation (%)
4.23 9.17 13.99 18.86
-6.76 -9.29 -9.09 -11.09
9.61 18.1 22.11 28.7
For example for the Northwest point (-108.75,
31.25) the local temperature depends on the global
one and is not necessarily the same: If the global T is
1°C the local is 0.9 for the GFDL model and is
different for the other model, 1.54 for the
HADGEM. The contrasts for the precipitation are
very large.
4 UNCERTAINTY OF T
GLOBAL
PROJECTED OVER THE MAPS
As mentioned before, (Gay and Sanchez, 2013)
considered two sources of uncertainty contributing
to the temperatures in 2100, the first coming from
the emissions: large emissions mean large
temperature changes, and the second due to our
imprecise knowledge of the climate sensitivity of the
models. The first uncertainty is for the politicians to
resolve because emissions depend on policy and if
this is oriented towards lowering them, the
temperatures could be kept within certain limits
determined in part by the uncertainty in the
sensitivity of the models. Therefore the second
source is for the scientists who need to narrow the
interval of climate sensitivity which is still too large
as it is shown in our discussion.
Once we have the global mean temperatures and
an idea of the associated uncertainty due to different
emission paths and sensitivities, using the same idea
for scaling employed in the Magicc/Scengen system
(Wigley, 2008), we convert this information to a two
dimensional maps of temperatures and
precipitations.
If we denote the uncertainty by a then we
propose the following equations:
T
new
= T
grid
/T
map
x T
magicc
(1)
P
new
= P
g
ri
d
/T
ma
p
x T
ma
g
icc
(2)
where T
magicc
, is the temperature produced by the
fuzzy model of Gay and Sanchez (2013) which is in
fact a fuzzy number and consequently T
new
and
P
new
also are. T
grid
/T
map
(or P
grid
/T
map
) represent the
normalized pattern of change for T (or P), i.e., the
change of the variable per each degree centigrade of
global warming.
The fuzzy model mentioned above, consisting of
18 fuzzy rules (Gay et al. 2013), is run to obtain
global temperatures increases in year 2100 and their
corresponding uncertainty intervals. This
information is then used to produce two-dimensional
maps depicting physically consistent geographical
distributions of temperatures which in turn are
consistent with global temperatures obtained from
our fuzzy model.
In the fuzzy model we use the value of the
sensitivity is fixed at the best estimate of 3 °C/W/m2
and varying the concentration we try to get 1, 2, 3
and 4 degrees centigrade. The temperature is a
function of the concentration, the T7, T8, … , T12
shown in the Figures 11 to 14 are the output
temperature increase fuzzy sets whose
characteristics were calculated via MAGICC data
(Gay et al., 2013). In this way we obtain the
following fuzzy values for the global temperatures:
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
870
Figure 11: Fuzzy output temperature for 1 °C, its
membership value is 0.646. The membership of 2 and 3 °C
is marginal. Data obtained from the 18 rules FIS (Gay and
Sanchez, 2013).
Figure 12: Fuzzy output temperature for 2 °C, its
membership value is 0.596. The figure also shows the
membership values for 1, 3 and 4 °C. Data obtained from
the 18 rules FIS (Gay and Sanchez, 2013).
Figure 13: Fuzzy output temperature for 3 °C, its
membership value is 0.845. The figure also shows the
membership values for 2, 4 and 5 °C. Data obtained from
the 18 rules FIS (Gay and Sanchez, 2013).
Figure 14: Fuzzy output temperature for 4 °C, its
membership value is 0.679. The figure also shows the
membership values for 2, 3, 5 and 6 °C. Data obtained
from the 18 rules FIS (Gay and Sanchez, 2013).
For an increase of one degree the concentration
of CO2 required is 220 ppmv and the uncertainty
interval is from 0.08 to 2.17 degrees, based on the
fuzzy sets characteristics reproduced here as a graph
(see Figure 11) consequently for a 1 °C global
increase, maps for one and two degrees (see Figures
3 to 10) are to be considered with their respective
membership value (1)=0.64 and (2)=0.094.
If T is 2 degrees the interval is from 0.08 to
3.27 °C; for 3 and 4 degrees the uncertainty intervals
are from 1.07 to 5.02 °C and from 1.82 to 6.41 °C
respectively (Figure 12). Therefore for a 3 °C global
increase the uncertainty extends to 5 °C so, maps
corresponding to 3, 4 and 5 degrees should be
considered.
Now we can interpret the maps obtained in the
previous section (Figures 3 to 10) in a different way.
For 2100 the simple fuzzy model of Gay and
Sanchez (2013) based on emission trajectories that
span from -2CO2 to 3 times the emissions in 1990
(1CO2) (that means that the paths corresponding to
4 and 5 times the 1990 emissions have been left out)
and the uncertainty of the climate sensitivity of the
models produce temperatures that span from 1 to
more than 5 degrees centigrade but with different
membership values. This information is then used to
assign a membership value to a whole map. This is
done in the following way.
Let us assume that the temperature in 2100 is 2
°C but this is a fuzzy number with a membership
value =0.51 but the fuzzy number also contains
1, 3 and 4 degrees with membership values of 0.51,
0.48, and 0.09 respectively. Therefore a warming of
two degrees would mean that we have to take into
consideration not only the map corresponding to 2
°C but also the corresponding to 1, 3, and 4 degrees
except that with different weights provided by the
TheFuzzyNatureofClimateChangeScenariosMaps
871
membership values, as it is shown in Table 5. The
same assignment should be done to the precipitation
maps. We have produced maps for 1, 2, 3 and 4 °C,
that have been associated with different dates with
the purpose of helping in the planning process. The
same maps can be used to produce fuzzy results for
2100 and somehow demonstrate that the longer we
wait the fuzzier the future becomes.
Table 5: Membership values for the delta T thresholds
projected to 2100, corresponding to fuzzy output sets
obtained from 18 rules fuzzy model. The CO2
concentrations listed are the values needed to achieve each
threshold.
Delta T thresholds (°C)
CO2
Concentration
(ppmv)
1 2 3 4
220 0.646 0.094 0.080 0
349 0.515 0.515 0.485 0.090
526 0 0.596 0.845 0.427
806 0 0.096 0.628 0.679
5 CONCLUSIONS
We have shown that, for decision-making purposes,
it is easier and more convenient to consider a
strategy based on degree by degree than one based
on dates. The different climate subsystems are
impacted by the global temperature increase, no
matter what date it occurs. The only determinant
parameter of the magnitude of impact is its climate
sensibility. Having information of the approximate
dates in which different thresholds, for example, one
degree would be crossed, would enable a policy
maker to act on those sectors or activities that would
be affected by one degree leaving for later those that
would be affected by larger temperature changes.
It has been depicted that linear emission
pathways, proposed in early works, include all the
possibilities mentioned in successive reports of
IPCC, including the RCP ones.
We have considered the uncertainty of GHG
concentrations and the uncertainty of climate system
sensibility as the more important. We know now that
the larger concentration and sensitivity the sooner
the successive thresholds of temperature will be
reached and the wider uncertainty intervals, too.
The uncertainty related to the process of
selecting AOGCM´s has its origin in the uncertainty
of climate sensibility.
The uncertainty generated from the AOGCM´s
results, available in the literature, has been estimated
from the construction of ensembles of model
projections for different dates, the range of
uncertainty in these projections is statistically
assigned by means of simple standard deviation
(Wehner, 1998) or, with a more complex procedure,
using criteria such as the performance of the actual
climatic conditions and convergence of projections,
for each AOGCM over a determined geographical
region (Giorgi and Mearns, 2002). The meaning of
averaging the results of different AOGCM´s, each
one with a different physical vision of the climate, is
not clear.
Several maps, representing climate change
scenarios, have been presented and a discussion
about them has been done. For that purpose we
select a couple of AOGCM´s and an emission
pathway of 5CO2, the more “pessimist”, but it
covered till 4 degrees of temperature increase by the
2100.
Starting from a fuzzy model of type FIS and a
simple climate model (MAGICC) we have obtained
a method for interpret the uncertainties involved in
the construction of global change scenarios of
temperature and precipitation. The key has been to
consider the uncertainty interval determinate by the
fuzzy model outputs for each one of the global
temperature thresholds, and extend it to the other
variable.
As a result of the above, it is possible see the
fuzzy perspective of the scenarios described in the
maps. It can be stated that the climate scenario for a
given global T map contains, in some degree, the
maps corresponding to other (adjacent) thresholds.
As an example, we found that the map for a T of 2
°C contains the map for 1, 3 until 4 °C, but the maps
for 1, 2 and 3 °C are almost indistinguishable from
each other.
ACKNOWLEDGEMENTS
This work was supported by the Programa de
Investigación en Cambio Climático (PINCC,
www.pincc.unam.mx) of the Universidad Nacional
Autónoma de México.
REFERENCES
Conde, C., F. Estrada, B. Martínez, O. Sánchez, C. Gay.
2011. Regional Climate Change Scenarios for México.
Atmosfera, Vol. 24 (1), 125-140. ISSN: 0187-6236.
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
872
Gay, C., Sánchez, O., Martínez-López, B., Nébot, Á.,
Estrada, F. 2012. Simple Fuzzy Logic Models to
Estimate the Global Temperature Change due to GHG
Emissions. 2nd International Conference on
Simulation and Modeling Methodologies,
Technologies and Applications (SIMULTECH).
Special Session on Applications of Modeling and
Simulation to Climatic Change and Environmental
Sciences - MSCCEC 2012. July 28-31. Rome, Italy.
Thomson Reuters Conference Proceedings Citation
Index (ISI), INSPEC, DBLP and EI (Elsevier Index)
http://www.informatik.uni-
trier.de/~ley/db/conf/simultech/simultech2012.html.
Gay García, C., O. Sánchez Meneses. 2013. Natural
Handling of Uncertainties in Fuzzy Climate Models.
3rd International Conference on Simulation and
Modeling Methodologies, Technologies and
Applications (SIMULTECH). Special Session on
Applications of Modeling and Simulation to Climatic
Change and Environmental Sciences - MSCCEC
2013. July 29-31. Reykjavík, Iceland. Thomson
Reuters Conference Proceedings Citation Index (ISI),
INSPEC, DBLP and EI (Elsevier Index) Abstracts in:
http://www.simultech.org/Abstracts/2013/MSCCEC_2
013_Abstracts.htm.
Gay, C., Sánchez, O., Martínez-López, B., Nébot, Á.,
Estrada, F. 2013. Fuzzy Models: Easier to Understand
and an Easier Way to Handle Uncertainties in Climate
Change Research. In: Simulation and Modeling
Methodologies, Technologies and Applications.
Volume Editor(s): Pina, N., Kacprzyk, J. and Filipe, J.
In the series "Advances in Intelligent and Soft
Computing". Springer- Verlag GmbH Berlin
Heidelberg.
Giorgi, F., Mearns, L.O. 2002. Calculation of Average,
Uncertainty Range, and Reliability of Regional
Climate Changes from AOGCM Simulations via the
‘‘Reliability Ensemble Averaging’’ (REA) Method.
Journal of Climate, Vol. 15; 1141-1158.
IPCC-WGI, 2013: Climate Change 2013: The Physical
Science Basis. Contribution of Working Group I to the
Fifth Assessment Report of the Intergovernmental
Panel on Climate Change [Stocker, T. F., D. Qin, G.-
K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A.
Nauels, Y. Xia, V. Bex and P. M. Midgley (eds.)].
Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, 1535 pp.
IPCC-WGI, 2007: Climate Change 2007: The Physical
Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental
Panel on Climate Change [Solomon, S., D. Qin, M.
Manning, Z. Chen, M. Marquis, K. B. Averyt, M.
Tignor and H. L. Miller (eds.)] Cambridge University
Press, Cambridge, United Kingdom and New York,
NY, USA, 996 pp.
IPCC-WGI, 2001: Climate Change 2001: The Scientific
Basis, Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on
Climate Change, [Houghton, J. T.; Ding, Y.; Griggs,
D. J.; Noguer, M.; van der Linden, P. J.; Dai, X.;
Maskell, K.; and Johnson, C. A., ed.] Cambridge
University Press, ISBN 0-521-80767-0, (pb: 0-521-
01495-6).
Meinshausen, M., S. C. B. Raper and T. M. L. Wigley
(2011). "Emulating coupled atmosphere-ocean and
carbon cycle models with a simpler model,
MAGICC6: Part I – Model Description and
Calibration." Atmospheric Chemistry and Physics 11:
1417-1456. doi:10.5194/acp-11-1417-2011.
Wehner, M. F. 1998. Determination of the Sampling Size
of AGCM Ensemble Simulations. Program for
Climate Model Diagnosis and Intercomparison
(PCMDI) Report No. 47. 24 pp. (On line: http://www-
pcmdi.llnl.gov/publications/pdf/47.pdf).
Wigley T. M. L. 2008. MAGICC/SCENGEN V. 5.3: User
Manual (version 2). NCAR, Boulder, CO. 80 pp. (On
line: http://www.cgd.ucar.edu/cas/wigley/magicc/).
Zadeh, L. A. 1965. Fuzzy Sets: Information and Control.
Vol. 8(3) p. 338-353.
TheFuzzyNatureofClimateChangeScenariosMaps
873
