CBP's Powerpoint template for scientific posters

1
Impacts of Land-Cover Change on the
Hydrometeorology of the Amazon Renato Ramos da
Silva and Roni Avissar Department of Civil
Environmental Engineering, Duke University,
Durham, NC 27708
44.14
Introduction Hydroclimate change resulting from
the replacement of natural forest by degraded
vegetation in the Amazon has yet to be fully
understood and quantified. Previous studies using
coarse grid global circulation models (GCMs) had
shown that continental deforestation reduces
rainfall and can affect the climate of other
regions, such as North America (Nobre et al.,
1991 Werth and Avissar, 2002). In this study the
Regional Atmospheric Modeling System (RAMS) is
used to understand and quantify the impacts of
deforestation on the hydrometeorology of the
Amazon basin during the wet season based on
realistic scenarios of deforestation. Three
scenarios of land covers are considered (Fig. 1)
(1) current vegetation (2) degraded land cover
as a result of deforestation expansion, as
estimated by Soares Filho et al. (2004), for
2030 and (3) same as (2) but for 2050.

• Results
• RAMS simulated relatively well the rainfall
  accumulation for January and February 1999. In
  general results from the model as compared with
  observations show that
• The model simulates well the accumulation of
  rainfall in Rondonia as compared with
  observations from raingauges (Fig. 3).
• The simulated spatial accumulation agrees
  relatively well with TRMM satellite derived data,
  but the model overestimates rainfall near the
  Andes and underestimates it near the coast of the
  state of Para (Fig. 4).
• Empirical Orthogonal Function analysis shows
  that the model is able to capture the principal
  modes of rainfall variability (Fig. 5).

Fig. 5. Empirical Orthogonal Function Analysis
applied to daily rainfall accumulation for
January and February 1999, as derived from the
TRMM satellite (left), and simulated with RAMS
(center), and their respective principal
components (PCs) (right). The percentage of
explained variance is provided for each EOF, and
numbers in parenthesis are correlation
coefficient between model and observation PCs.
Observations
Fig. 1 Land cover in the Amazon Basin (a)
current vegetation, (b) 2030, (c) 2050. Dark
green represents forest, yellow is pasture and
magenta is mixed woodland (cerrado). Scenarios
are from Soares Filho et al. (2004).

• Wind strengthens as a result of the replacement
  of forest by pasture in the state of Para and
  Mato Grosso, while it weakens in Tocantins and
  north of Amazonas, emphasizing complex
  land-atmosphere interactions that remained to be
  explained
• These results emphasize that the region west of
  Para has a high climatic susceptibility to
  land-cover change. In this study, deforested
  areas were assumed to be pasture. A more
  realistic scenario needs to recognize other
  non-natural land cover such as agricultural
  fields, artificial lakes and urban areas. To
  account for these landscape heterogeneities
  higher resolution simulations need to be
  performed and evaluated.

(a)
RAMS
(b)
Fig. 3. Simulated rainfall accumulation in
Rondonia compared to observations from a
raingauge network consisting of 40 stations. The
bars represent the standard deviation in the
observations.
(c)
Methodology The Regional Atmospheric Modeling
System (RAMS) is used to simulate the
hydrometeorology of the region for January and
February 1999. Results are compared with
observations from the LBA TRMM/Wet AMC (Avissar
et al., 2002). The two-month numerical
integration was conducted using a 20-km grid size
covering the domain shown in Figs. 1-2. The NCEP
Reanalysis was used as atmospheric boundary
conditions (Kalnay et al. 1996). First, a control
simulation using current land cover (Fig. 1a) was
performed. Then, two other simulations were
produced using the same boundary conditions as in
the control case, but assuming projected land
covers for 2030 and 2050 (Figs. 1b-c).
Rainfall near the Andes is overestimated (Fig.
4). This is caused by the extreme gradient of
topography, which is particularly challenging to
simulate with current state-of-the-art numerical
models.
Literature cited Avissar R., P. L. Silva Dias,
M. A. Silva Dias, and C. Nobre, 2002 The
large-Scale Biosphere-Atmosphere Experiment in
Amazonia (LBA) Insights and future research
needs. J. Geophys. Res., 107, (doi
10.129/2002JD002704). Huffman, G.J., R.F. Adler,
B. Rudolph, U. Schneider, and P. Keehn, 1995
Global Precipitation Estimates Based on a
Technique for Combining Satellite-Based
Estimates, Rain Gauge Analysis, and NWP Model
Precipitation Information, J. Clim., 8,
1284-1295. Kalnay, E., et al., 1996 The
NCEP/NCAR 40-year reanalysis project, Bull. Amer.
Meteor. Soc., 77 (3), 437-471. Nobre, C. A., P.
J. Sellers, and J. Shukla, 1991 Amazonian
deforestation and regional climate change, J.
Clim., 4(10), 957-987. Pielke, R. A., et al.,
1992 A Comprehensive Meteorological Modeling
System - Rams. Meteorology and Atmospheric
Physics, 49, 69-91. Soares Filho, B. S., H.
Rodrigues, D. Nepstad, G. Cerqueira, E. Voll , A.
Alencar, and P. Lefebvre, 2004 A spatially
explicit simulation model of deforestation for
the Amazon Basin. III LBA Scientific Conference
(Brasilia, DF). Werth, D., and R., Avissar, 2002
The local and global effects of Amazon
deforestation. J. Geophys. Res., 107, (doi
10.129/2001JD000717).
Fig. 6. Rainfall anomaly (mm) relative to the
control simulation for 2030 (top) and 2050
(bottom) and percentage change for some of the
sub-regions
Rainfall anomalies shown in Fig. 6 are associated
with changes in wind speed and temperature (Fig.
7). Wind speed stronghtens in deforested areas
such as the state of Para and Mato Grosso, but
decreases in Tocantins and north of the Amazonas
state. Temperatures increase in the deforested
areas.
Fig. 4. Accumulated rainfall for January and
February 1999 simulated with RAMS (top) and
derived from the TRMM satellite (bottom) (Huffman
et al., 1995). Units are mm.
Fig. 2. Topography of the simulated domain. The
terrain changes abruptly from near sea level in
the Amazon Basin to heights on the order of 6000
meters in the Andes, making this region
particularly challenging to simulate with
numerical models.
Fig. 7. Wind speed anomalies (shaded, left), for
2030 (top) and 2050 (down). Vectors represent the
mean wind. Surface temperature anomalies are
shown in the right panel.
The analysis of the first three principal
components shows good correlation between model
results and observations (Fig. 5).

• Conclusions
• These preliminary simulations presented here
  indicate that
• Rainfall gradually decrease as deforestation
  increase in the heavily deforested regions of the
  Amazon Basin
• Rainfall accumulation seem to significantly
  increase near the heavily deforested regions,
  e.g., in Bolivia and the state of Tocantins

• Comparing the simulations for 2030 and 2050 with
  the control simulations, we find that
• rainfall in the Amazon decreases by about 6 and
  10 for 2030 and 2050, respectively (Fig. 6)
• some regions such as Bolivia and Tocantins had
  higher accumulations of rainfall.

For further information Please contact
renato_at_duke.edu.