GC41A-0110
Assessment of 21st Century Climate Change Projections in the Tropical Andes
Rocio Urrutia & Mathias Vuille
rurrutia@geo.umass.edu, mathias@geo.umass.edu
Climate System Research Center, Department of Geosciences,University of Massachusetts Amherst, 611 North Pleasant Street, Amherst, MA
01003-9297, United States
b)
a)
1.Introduction
a)
b)
8oN
Latitude
Longitude
o
64oW
56oW
−1600
−1200
−800
−400
400
800
1200
80oW
1600
o
8 S
72oW
64oW
56oW
o
72oW
64oW
56oW
48oW
o
o
80 W
o
72 W
o
64 W
Longitude
−250
a)
−200
−150
48 W
2
3
4
72oW
64oW
−50
0
56oW
50
100
6
7
150
200
b)
o
8 N
8oN
0o
Longitude
Latitude
o
8 S
16oS
o
24oS
o
24 S
80oW
72oW
64oW
56oW
80oW
48oW
−16
−12
−8
−4
0
72oW
64oW
56oW
48oW
Latitude
Longitude
−20
o
8 S
4
8
12
16
20
Figure 7. a) Changes in mean annual relative humidity between the A2 and control run for the 30 year period (in
percent) b) Map indicating significant differences between the two scenarios. Areas in dark red are significantly
different with 95% confidence.
a)
b)
c)
d)
Western slope
72oW
64oW
56oW
Eastern slope
Longitude
8oS
16oS
24oS
2
3
o
4
o
o
64 W
Latitude
56 W
5
o
48 W
6
5
0
−5
−10
Latitude
−15
−20
Altitude (m)
Difference in temperature
Difference in temperature
−25
b)
c)
6. Discussion and conclusion
The projected warming in South America is more pronounced along the eastern Amazon basin, relatively close to
the mouth of the Amazon river, where a temperature increase of more than 7ºC is projected in the A2 scenario.
Warming is more widespread during JJA than DJF, and the simulated warming in the Andes fluctuates between
2ºC and 7ºC, being especially high in the Cordillera Blanca and the Altiplano region. When comparing A2 and B2
projections, it is apparent that mainly the magnitude changes, with up to approximately 3ºC warmer temperature
in A2, but the spatial pattern is very similar. In the A2 scenario no grid cells with temperature below freezing
remain during DJF, indicative of the significant rise of the freezing levels. This rise has significant ramifications
for glacier mass balance as the tropical glacier equilibrium line altitude (ELA) is closely tied to the 0°C isotherm.
Predicted changes in precipitation indicate an overall decrease in northeastern South America, following a similar
pattern as the observed increase in temperature in that area. This is consistent with the observed changes in
relative humidity (RH), which also closely follow the pattern observed for temperature. Clearly changes in
relative humidity reflect a temperature increase, rather than changes in specific humidity, indicating that changes
in specific humidity are insufficient to keep relative humidity constant under strong warming scenarios.
Precipitation changes in the Andes, do not present a clear pattern, but the Monsoon System appears to be stronger
in the future with increased summer precipitation and an enhanced seasonality in the South American outer
tropics.
The PDFs indicate a stronger warming along the eastern slopes of the Andes, although temperature remains
higher on the western slopes. Changes in the standard deviation are also larger on the eastern slopes. The annual
mean projected temperature is about 1ºC higher in A2 than in B2 for both slopes. The standard deviation changes
little in the B2 scenario, while A2 shows a pronounced change with a significant increase in year-to-year
temperature variability.
Changes in free tropospheric temperature indicate an increased warming with altitude for both scenarios, but less
pronounced in the B2, consistent with results based on IPCC AR4 GCMs (Bradley et al., 2006)
Changes in zonal and vertical wind indicate a strengthening of the westerly component and a weakening of the
easterly upper-level flow, which would suggest drier conditions in the Altiplano at the end of the 21st century
(Garreaud et al., 2003).
In summary, these results reveal that predicted climate change is characterized by a general increase in surface
and free tropospheric temperature, a decrease in relative humidity and, in the case of the Altiplano, a decrease in
precipitation, all of which will adversely affect the glacier mass and energy balance of the remaining tropical
glaciers along the Andes Cordillera.
48oW
8
0o
72 W
25
Figure 10. a) Pressure-longitude cross section of zonal and vertical winds at 16ºS for DJF in the control scenario. b)
as in a) but for the A2 scenario. c).as in a) but for difference between the A2 and control scenarios.
8oS
8oN
o
20
250
b)
16 S
Figure 2.a) Difference in mean annual temperature for 30 years between A2 and control scenario. b) As in a), but for
austral summer temperature (DJF) c) As in a), but for austral winter temperature (JJA).
80 W
15
48oW
Longitude
5
10
Latitude
Longitude
−100
Figure 6. a) Time-latitude diagram of the seasonal cycle of precipitation (in mm/day) averaged over longitude 50º-70º
W for the control scenario b) As in a), but for the A2 scenario.
8oN
80oW
o
56 W
Longitude
1
80oW
a)
2000
24oS
o
24 S
80oW
48oW
16oS
o
16 S
24 S
a)
0o
Latitude
Latitude
8oS
16oS
5
5.5 Predicted changes in zonal and vertical winds for summer in a longitudinal transect at 16º S
o
8 S
5.3 Probability Density Functions (PDF) for temperature along the western and eastern slope of the Andes
5.1 Predicted changes in surface temperature
o
0
Figure 5. a) Absolute difference (in mm) in mean annual precipitation for 30 years between A2 and control scenario, b)
Map indicating significant changes between the two scenarios. Areas in dark red are significantly different with 95%
confidence c) As in a), but for relative difference (in percentage).
5. Results
0
0
5
Figure 9. a) Difference in annual mean free tropospheric temperature between the A2 and control scenario. b) as in a)
but for difference between the B2 and control scenario.
o
8 N
Latitude
0
3
2000
4
24oS
48oW
Longitude
−2000
3.5
4000
o
16 S
24 S
72oW
4
6000
4.5
2000
o
16oS
80oW
4.5
8000
0
8oS
o
5.5
6000
5
0o
0o
8oS
24 S
Our analyses concentrate on variables known to affect glacier energy and mass balance. Analyses on surface
variables focused on determining the differences between the A2 and control run in terms of temperature,
precipitation and relative humidity. Dynamic analyses of monsoon behavior were investigated by means of timelatitude diagrams (Hovermuller diagrams). Probability Density Functions of mean annual temperature were
established for all scenarios along the eastern and western slopes of the Andes to assess changes in the width and
amplitude of the distribution. Latitudinal transects of free tropospheric temperature changes were developed for
both A2 and B2 scenarios. Of special relevance for climate and precipitation in the Andes is the uppertropospheric zonal flow aloft (Garreaud et al., 2003). This study therefore also includes an analysis of changes in
zonal wind and vertical velocity along a latitudinal cross-section from the Pacific to the Amazon basin across the
central Andes.
0o
10000
6
4000
c)
8oN
12000
5
B2
10,000
5,000
82,500
5.2 Predicted changes in surface precipitation and relative humidity
o
c)
σ (º)
0.43
0.44
6.5
Altitude (m)
A2
5,000
0
55,000
0
o
8 N
μ (º)
291.4
288.6
7
10000
0
a)
Latitude
Control
95,000
22,500
285,000
Annual
DJF
JJA
4. Analyses
b)
B2
σ (º)
0.73
0.72
b)
12000
8000
Figure 1. Study area
8oN
a)
Figure 4. a) Difference in mean annual temperature between the A2 and control scenario. b) As in a), but for B2
scenario.
Table 1. Area in km2 with temperature below 0ºC
16oS
3.2 Climate model
In this study the Hadley Centre Regional Climate Modeling System, called PRECIS, was used to assess changes in
climate between today and the end of the 21st century. The model was run in 0.44º lat. x 0.44º lon. (~50 x 50 km)
resolution. Three simulations are analyzed here. The first simulation (RCM 61-90) involves a thirty year (19611990) period that serves as a control run and provides the base-line for comparison with the projected greenhouse
gas emission scenarios. The second simulation (RCM SRES-A2) is based on a medium-high emission and high
population-growth scenario run for the 2071-2100 period. In the third simulation (RCM SRES- B2), the projected
greenhouse scenario corresponds to a lower emission and population growth, and it is also a run for the 2071-2100
period.
a)
A2
μ (º)
292.4
289.7
5.4 Predicted changes in free tropospheric temperature
3. Methods
3.1 Study area
The study area corresponds to tropical South America between approximately
13º N-30º S and 89º W-41º W (Figure 1). This corresponds to the complete
RCM domain, but after removing the grid cells where adjustments
between the GCM and the regional model take place, the actual study area
is reduced to approximately 10º N-27º S and 86º W-44º W. The analysis
has been mainly focused on the Andes region to detect and assess climatic
changes that could potentially have an impact on tropical glaciers in that area.
Control
μ (º)
σ (º)
288.5
0.41
285.5
0.33
Western slope
Eastern slope
2. Objectives
The main objective of this study is to provide an assessment of predicted changes in climate in tropical South
America and especially in the Andes region, by comparing output from a regional climate model (PRECIS) from
the end of the 21st century (2071-2100) with a control run from the end of the 20th century (1961-1990).
This simulation provides a baseline for estimating future climate change impacts in a medium-high (A2) emission
scenario. In addition, mainly for temperature change assessments, a medium-low (B2) scenario is also considered
for comparison.
Table 2. Average temperature (µ) and standard deviation (σ ) in each PDF for each scenario and slope of the Andes
Latitude
The tropical Andes are one of the regions where recent climate change has been most evident (Vuille et al.,
submitted). This is consistent with the notion that high-elevation tropical mountains, extending to the midtroposphere, will be more affected by warming (Bradley et al., 2006). One of the main impacts of this warming in
the Andes is the retreat of glaciers, a process that may affect the availability of water for human consumption,
irrigation, mining and power generation (Vergara et al., 2007).
Climate models are an important tool to determine future changes and potential impacts of climate change. Since
general circulation models (GCM) make projections at a relatively coarse resolution, they do not adequately
capture the regional-scale changes, and they are not very well suited in regions with complex topography or
diverse land cover (Jones et al., 2004; Solman et al., 2007). Regional climate models (RCM) on the other hand are
comprehensive physical models that frequently incorporate atmosphere and land surface components, and the
representation of important processes within the climate system (Jones et al., 2004). These models have the ability
of solving meso-scale features that are not evident in global models, and are especially useful in places with
complex topography, such as the Andes.
7
Figure 3.Difference in mean annual temperature between the A2 and control scenario for the Andes region.
Figure 8. a) PDF of mean annual temperature for the control (blue line) and A2 (red line) scenario, considering grid
cells >100 m along the western slope of the Andes. b) As in a) but for control and B2 scenario. c) As in a) but
considering only grid cells >400 m, and for the eastern slope of the Andes. d) as in c) but for control and B2 scenario.
In all cases black bars represent the individual years for each distribution.
7. References
Bradley, R.S., Vuille, M., Diaz, H.F., & Vergara, W. 2006. Threats to water supplies in the Tropical Andes. Science 312, 5781:
1755-1756.
Garreaud, R., Vuille, M., & Clement, A. 2003. The climate of the Altiplano: Observed current conditions and mechanisms of
past changes. Palaeogeography, Palaeoclimatology, Palaeoecology 194: 5-22.
Jones, R.G., M. Noguer, D. Hassell, D. Hudson, S. Wilson, G. Jenkins & J. Mitchell. 2004. Generating high resolution climate
change scenarios using PRECIS. Met Office Hadley Centre, Exeter, UK. 40 pp. April 2004.
Solman, S., Nunez, M., & Cabre, M. Regional climate change experiments over southern South America I: present climate.
Climate dynamics doi 10.1007/s00382-007-0304-3.
Vergara, W, Deeb, A., Valencia, A., Bradley, R., Francou, B., Zarzar, A., Grünwaldt, A. & Haeussling, S. 2007. Economic
impacts of rapid glacier retreat in the Andes. Eos, Transactions, American Geophysical Union 88: 25, 19 June 2007.
Vuille, M., B. Francou, P. Wagnon, I. Juen, G. Kaser, B.G. Mark & R.S. Bradley, Climate change and tropical Andean glaciers
– Past, present and future. Earth Science. Reviews., (submitted).
8. Acknowledgements
This study was funded by NSF (EAR-0519415 awarded to MV) and by a Fulbright Fellowship to RU.