ads:
SÍLVIA DE NAZARÉ MONTEIRO YANAGI
ALBEDO DE UMA FLORESTA TROPICAL AMAZÔNICA: MEDIÇÕES DE
CAMPO, SENSORIAMENTO REMOTO, MODELAGEM, E SUA INFLUÊNCIA NO
CLIMA REGIONAL.
Tese apresentada à Universidade Federal de
Viçosa, como parte das exigências do Programa
de Pós-graduação em Meteorologia Agrícola,
para obtenção do título de Doctor Scientiae.
VIÇOSA
MINAS GERAIS-BRASIL
2006
ads:
Livros Grátis
http://www.livrosgratis.com.br
Milhares de livros grátis para download.
Ficha catalográfica preparada pela Seção de Catalogação e
Classificação da Biblioteca Central da UFV
T
Yanagi, Sílvia de Nazaré Monteiro, 1973-
Y21a Albedo de uma floresta tropical Amazônica: medições de
2006 campo, sensoriamento remoto, modelagem e sua influência
no clima regional / Sílvia de Nazaré Monteiro Yanagi.
– Viçosa : UFV, 2006.
xxii, 128f. : il. ; 29cm.
Texto em inglês.
Orientador: Marcos Heil Costa.
Tese (doutorado) - Universidade Federal de Viçosa.
Referências bibliográficas: f. 116-128.
1. Climatologia agrícola. 2. Amazônia - Clima - Modelos
matemáticos. 3. Desmatamento. 4. Sensoriamento remoto.
5. Microclimatologia florestal. I. Universidade Federal de
Viçosa. II.Título.
CDD 22.ed. 630.2515
ads:
SÍLVIA DE NAZARÉ MONTEIRO YANAGI
ALBEDO DE UMA FLORESTA TROPICAL AMAZÔNICA: MEDIÇÕES DE
CAMPO, SENSORIAMENTO REMOTO, MODELAGEM, E SUA INFLUÊNCIA NO
CLIMA REGIONAL.
Tese apresentada à Universidade Federal de
Viçosa, como parte das exigências do Programa
de Pós-graduação em Meteorologia Agrícola,
para obtenção do título de Doctor Scientiae.
APPROVADA: 31 de outubro de 2006
Dr. Yosio Edemir Shimabukuro Prof. Gilberto C. Sediyama
Co-Orientador
Dra. Francisca Zenaide de Lima
Dr. Humberto Alves Barbosa
Prof. Marcos Heil Costa
(Orientador)
ii
Este trabalho é dedicado à
meus pais José e Nazaré,
minha avó Francisca “in memorian”,
meu marido Tadayuki
e minha sogra Conceição.
iii
AGRADECIMENTOS
À Universidade Federal de Viçosa (UFV), especialmente ao Departamento de
Engenharia Agrícola, pela oportunidade de realizar o curso.
À Coordenadoria de Aperfeiçoamento de Pessoal de Ensino Superior (CAPES) e ao
Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), pela concessão
de bolsa de estudo.
Aos meus pais José e Nazaré Albuquerque, pela educação nos princípios da
verdade, pelos grandes incentivos, apoio e força para vencer as dificuldades encontradas
durante a realização deste trabalho, por tudo.
Aos meus irmãos, tio e sobrinho: Patrícia, José Walter, Nivaldo e Matheus,
especialmente, a minha avó Francisca Monteiro (in memoriam) pela grande força na
continuidade da minha vida profissional.
Ao meu marido Tadayuki Yanagi Junior pelo grande apoio nas horas difíceis em
que me encontrei, pelos incentivos profissionais, pelo amor, força e amizade e, acima de
tudo pela compreensão dos dias em que a distancia física nos separou.
A minha sogra e amiga Conceição Yanagi, pelo carinho e cuidado durante as
minhas horas de estudo, e principalmente pela compreensão de minha ausência.
Ao professor Dr. Marcos Heil Costa, pela orientação, amizade e apoio profissional.
Aos meus conselheiros professores Drs. Humberto Ribeiro da Rocha e Yosio
Edemir Shimabukuro, pelas valiosas sugestões e contribuições.
iv
Aos membros do comitê de minha banca de defesa de tese Dr. Yosio Edemir
Shimabukuro, Prof. Gilberto Chohaku Sediyama, Dr. Humberto Alves Barbosa, e
Dra. Francisca Zenaide de Lima, pelas valiosas contribuições para finalização deste
trabalho.
Ao professor Dr. Antônio Carlos Lôla da Costa da Universidade Federal do Pará,
pelos conselhos e incentivos profissionais.
A todos os Professores do curso de Meteorologia Agrícola, em especial aos
professores Drs. Gilberto C. Sediyama, José Maria Nogueira da Costa, Sérgio Zolnier,
Luiz Cláudio Costa e Everardo C. Mantovani, pelos valiosos conhecimentos, atenção e
amizade.
Aos funcionários do Departamento de Engenharia Agrícola, pelo suporte, em
especial Marcos, Galinari, “Tia Maria” e as secretárias Kelly e Edna, pelo carinho e
dedicação.
Aos estudantes do Grupo de Pesquisa em Climatologia, Christiane, Varejão,
Hewlley, Santiago, Lucía, Francisca, Clever, Tomé, Edson, Márcia, Fabrício e Luciana
pelo coleguismo e pelas contribuições no desenvolvimento deste trabalho.
A todos os meus colegas da Área de Meteorologia, em especial Welliam, Raniére,
José Luiz, Danilo Filho, Vanda, Hernani, Michelly, Rochane, Rosandro, José de Paula,
Evandro, Raquel, Evaldo pela força e amizade.
Aos meus queridos amigos Marcos Paulo (“iuiu”), Kelly, Welliam e Adriane,
Christiane e Raniére, Hewlley e Luizinho, Olívio e Tânia, Sandra e André, Bergson, Leila
e Mauro, Gleidson e outros a minha eterna amizade.
Agradeço especialmente a minha amiga Christiane Leite, pelo apoio nas horas
incertas e pela sua grande ajuda durante as minhas simulações, quando estava fora da
cidade de Viçosa participando de uma Conferência Internacional.
A todos os demais professores, colegas e funcionários que, direta ou indiretamente,
participaram da realização deste trabalho, o meu sincero agradecimento.
v
BIOGRAFIA
SÍLVIA DE NAZARÉ MONTEIRO YANAGI, filha de José C. Moraes de
Albuquerque e Maria de Nazaré Monteiro de Albuquerque, nasceu em 06 de maio de 1973,
na cidade de Belém-Pará-Brazil.
Em dezembro de 1997 concluiu o curso de graduação em Meteorologia pela
Universidade Federal do Pará (UFPA).
Em julho de 1998 concluiu o curso de especialização em Sensoriamento
Remoto pela Universidade Federal do Pará (UFPA).
Em junho de 2001 concluiu o curso de pós-graduação, em nível de mestrado,
em Meteorologia Agrícola na Universidade Federal de Viçosa (UFV).
Em setembro de 2002 iniciou o curso de pós-graduação, em nível de doutorado,
em Meteorologia Agrícola na Universidade Federal de Viçosa (UFV).
vi
SUMÁRIO
LISTA DE FIGURAS............................................................................................... ix
LISTA DE QUADROS............................................................................................. xv
LISTA DE ABREVIATURAS................................................................................. xvii
RESUMO.................................................................................................................. xix
ABSTRACT.............................................................................................................. xxi
GENERAL INTRODUCTION................................................................................. 01
CHAPTER 1 – SOURCES OF VARIATION OF ALBEDO OF AMAZONIAN
VEGETATION……………………………………………………………………. 04
1.1. INTRODUCTION………………………………………………………….. 04
1.2. Sites, instrumentation and data……………………………………………... 07
1.3. Sources of variation of hourly albedo……………………………………… 11
1.4. Sources of variation of monthly albedo…………………………………….. 17
1.5 CONCLUSIONS………………………………………………………….... 21
1.6 NOMENCLATURE………………………………………………………... 23
CHAPTER 2 – MODELING RADIATIVE TRANSFER IN TROPICAL
RAINFOREST CANOPIES: SENSITIVITY OF SIMULATED ALBEDO
TO CANOPY ARCHITECTURAL AND OPTICAL PARAMETERS…………... 24
2.1. INTRODUCTION………………………………………………………...... 24
2.2. MATERIALS AND METHODS…………………………………………... 26
vii
2.2.1. IBIS description………………………………………………….. 26
2.2.2. Experimental site and data……………………………………….. 32
2.2.3. Sensitivity to canopy architectural and optical parameters………. 33
2.3. RESULTS AND DISCUSSION……………………………………………. 34
2.4. SUMMARY AND CONCLUSIONS………………………………………. 38
2.5 NOMENCLATURE…………………………………. 39
CHAPTER 3 – SIMULATIONS OF TROPICAL RAINFOREST ALBEDO: IS
CANOPY WETNESS IMPORTANT? …………………………………………… 41
3.1. INTRODUCTION………………………………………………………….. 41
3.2. MATERIALS AND METHODS…………………………………………... 43
3.2.1. Sites, instrumentation and data…………………………………... 43
3.2.2. Description of the IBIS model…………………………………… 45
3.2.3. Experiment design……………………………………………….. 46
3.3. RESULTS AND DISCUSSION……………………………………………. 49
3.4. SUMMARY AND CONCLUSIONS………………………………………. 57
3.5 NOMENCLATURE………………………………………………………... 58
CHAPTER 4 – COMPARISON OF SEASONAL AND SPATIAL
VARIATIONS OF ALBEDO ESTIMATED BY A CLIMATE MODEL AND
ALBEDO DERIVED FROM REMOTE SENSORS DATA FOR THE AMAZON
TROPICAL RAINFOREST ………………………………………………………. 59
4.1. INTRODUCTION………………………………………………………….. 59
4.2. ALBEDO DATA…………………………………………………………… 61
4.2.1. Land surface albedo simulations…………………………………. 61
4.2.2. Remote sensing albedos………………………………………….. 61
4.2.3. Field measurements albedo………………………………………. 66
4.3. RESULTS AND DISCUSSION……………………………………………. 66
4.4. SUMMARY AND CONCLUSIONS………………………………………. 73
CHAPTER 5 – RADIATIVE PROCESSES OF THE PRECIPITATION
CHANGE AFTER TROPICAL DEFORESTATION…………………………….. 75
5.1. INTRODUCTION………………………………………………………….. 75
5.2. MATERIALS AND METHODS………………………………………....... 77
5.2.1. Description of the CCM3-IBIS model…………………………… 77
5.2.2. Experiment design……………………………………………….. 78
viii
5.3 RESULTS…………………………………………………………………... 79
5.3.1. Precipitation dependence on surface radiation balance………….. 79
5.3.2. Clouds dependence on surface radiation balance and feedbacks
on the incoming solar radiance…………………………………... 84
5.3.3. Convection dependence on surface radiation balance…………… 89
5.4. DISCUSSION………………………………………………………………. 99
5.5. CONCLUSIONS…………………………………………………………… 106
5.6 NOMENCLATURE………………………………………………………... 107
CHAPTER 6 – GENERAL CONCLUSIONS…………………………………….. 108
6.1. OVERVIEW……………………………………………………………….. 108
6.2. CONCLUSIONS…………………………………………………………... 111
6.3. RECOMMENDATIONS FOR FUTURE RESEARCH…………………… 114
GENERAL REFERENCES...................................................................................... 116
ix
LISTA DE FIGURAS
Figure 1.1 – Orientation map………………………………………………………. 08
Figure 1.2 – Forest albedo as a function of solar zenith angle for four
transmissivity ranges (0-25%, 25-50%, 50-75% e 75-100%) and for
(a) dry canopy and (b) wet canopy…………………………………… 15
Figure 1.3 – Pasture albedo as a function of solar zenith angle for four
transmissivity ranges (0-25%, 25-50%, 50-75% e 75-100%) and for
(a) dry canopy and (b) wet canopy…………………………………… 16
Figure 1.4 – Average monthly variability of (a) surface albedo and (b)
atmospheric transmissivity for tropical rainforest and pasture………. 19
Figure 1.5 – (a) Average monthly albedo for the forest and pasture ecosystems as
a function of atmospheric transmissivity (τ), and regression
equations; (b) residuals between the observed albedo and the albedo
estimated by the regression equations in Figure 5a…………………...
20
Figure 1.6 – Average monthly variability of surface albedo for tropical rainforest
and pasture and its difference, considering the canopy wetness,
where α
F wet
and α
F dry
are the wet and dry forest albedos,
respectively; α
P wet
and α
P dry
are the wet and dry pasture albedos; and
α'
P-F wet
and α'
P-F dry
are the difference between surface albedos for
x
wet and dry pasture and forest canopy, respectively…………………. 21
Figure 2.1 – Schematic representation of the radiative transfer model in IBIS……. 28
Figure 2.2 –
Simulated albedo as a function of ρ
NIR,up
and χ
up
, for the Cuieiras
Biological Reserve (K34)……………………………………………..
35
Figure 2.3 – Temporal variation of the albedo observed in the Cuieiras Biological
Reserve (K34) and the albedo simulated by the IBIS model,
according to the upper canopy element orientation parameter……….. 36
Figure 2.4 – Root Mean Square Error (RMSE) between the observed and
simulated albedo as a function of the canopy optical parameters
ρ
NIR,up
and χ
up
, for the Cuieiras Biological Reserve (K34)…………...
37
Figure 3.1 – Orientation map………………………………………………………. 44
Figure 3.2 – Diurnal variation of the simulated and observed surface albedo in the
Cuieiras Reserve for selected days…………………………………… 52
Figure 3.3 – Diurnal variation of the simulated and observed surface albedo in the
Ducke Reserve for selected days……………………………………... 53
Figure 3.4 – Diurnal variation of the simulated and observed surface albedo in the
Jaru Reserve for selected days………………………………………... 54
Figure 3.5 – Monthly profile of the observed and simulated surface albedo,
monthly precipitation and frequency of rainfall events, at three
Amazon rainforest sites: (a) Cuieiras Reserve, from June 1999 to
September 2000, (b) Ducke Reserve, from January to December of
1995 and (c) Jaru Reserve, from January to December of 1993……... 56
Figure 4.1 – Spatial variability of the surface albedo simulated by the (a) CCM3
model and albedo from six remote sensing products for Amazon
basin: (b) CG99, (c) ERBE, (d) SRB/ISLSCP2, (e) UMD, (f)
MODIS Black-sky and (g) MODIS white-sky and it respective
anomalies: (h) CCM3-CG99, (i) CCM3-ERBE, (j) CCM3-
SRB/ISLSCP2, (k) CCM3-UMD, (l) CCM3-MODIS Black-sky and
(m) CCM3-MODIS white-sky ………………………………………. 68
Figure 4.2 – Seasonal variability of the surface albedo simulated by the CCM3
model and albedo from six remote sensing products for Amazon
basin, and measured at three study sites………………………………. 70
Figure 4.3 – Seasonal variability of the surface albedo simulated by the CCM3
xi
model and albedo from six remote sensing products for the pixels
comprising the three study sites, and albedo measured at the (a)
Manaus (Cuieiras and Ducke) and (b) Jaru sites……………………... 72
Figure 5.1 – Annual mean precipitation profile as a function of albedo changes for
different climatic experiments………………………………………... 76
Figure 5.2 – Annual mean precipitation anomaly as a function of albedo anomaly
(a) and net radiation anomaly (b), for different levels of pastureland
and soybean cropland expansions……………………………………. 80
Figure 5.3 – Semester mean precipitation anomaly as a function of albedo
anomaly for the dry semester (a), rainy semester (b), and both (c),
and as a function of net radiation anomaly for the dry semester (d),
rainy semester and both (f). White squares represent pastureland and
gray squares represent soybean cropland…………………………….. 82
Figure 5.4 – Trimester mean precipitation anomaly as a function of albedo
anomaly for the January to March trimester (a), April to June
trimester (b), July to September trimester (c), October to December
trimester (d) and all trimesters (e), and as a function of net radiation
anomaly for the January to March trimester (f), April to June
trimester (g), July to September trimester (h), October to December
trimester (i) and all trimesters (j). White squares represent
pastureland and gray squares represent soybean cropland…………… 83
Figure 5.5 – Annual mean total cloud anomaly as a function of albedo anomaly
(a) and net radiation anomaly (b), for different levels of pastureland
and soybean cropland expansions……………………………………. 84
Figure 5.6 – Semester mean total cloud anomaly as a function of albedo anomaly
for the dry semester (a), rainy semester (b), and both (c), and as a
function of net radiation anomaly for the dry semester (d), rainy
semester and both (f). White squares represent pastureland and gray
squares represent soybean cropland………………………………….. 86
Figure 5.7 – Trimester mean total cloud anomaly as a function of albedo anomaly
for the January to March trimester (a), April to June trimester (b),
July to September trimester (c), October to December trimester (d)
and all trimesters (e), and as a function of net radiation anomaly for
xii
the January to March trimester (f), April to June trimester (g), July to
Se89ptember trimester (h), October to December trimester (i) and all
trimesters (j). White squares represent pastureland and gray squares
represent soybean cropland…………………………………………... 87
Figure 5.8 – Annual mean incoming solar radiation anomaly as a function of total
cloud anomaly (a) for different levels of pastureland and soybean
cropland expansions. Semester mean incoming solar radiation
anomaly as a function of total cloud anomaly for the dry semester
(a), rainy semester (b), and both (c). Trimester mean incoming solar
radiation anomaly as a function of total cloud anomaly for the
January to March trimester (a), April to June trimester (b), July to
September trimester (c), October to December trimester (d) and all
trimesters (e). White squares represent pastureland and gray squares
represent soybean cropland…………………………………………... 88
Figure 5.9 – Annual mean vertical velocity anomaly as a function of albedo
anomaly (a) and net radiation anomaly (b), for different levels of
pastureland and soybean cropland expansions……………………….. 90
Figure 5.10 – Semester mean vertical velocity anomaly as a function of albedo
anomaly for the dry semester (a), rainy semester (b), and both (c),
and as a function of net radiation anomaly for the dry semester (d),
rainy semester and both (f). White squares represent pastureland and
gray squares represent soybean cropland……………………….......... 91
Figure 5.11 – Trimester mean vertical velocity anomaly as a function of albedo
anomaly for the January to March trimester (a), April to June
trimester (b), July to September trimester (c), October to December
trimester (d) and all trimesters (e), and as a function of net radiation
anomaly for the January to March trimester (f), April to June
trimester (g), July to September trimester (h), October to December
trimester (i) and all trimesters (j). White squares represent
pastureland and gray squares represent soybean cropland…………… 92
Figure 5.12 – Annual mean precipitation anomaly as a function of vertical velocity
anomaly (a) for different levels of pastureland and soybean cropland
expansions. Semester mean precipitation anomaly as a function of
xiii
vertical velocity anomaly for the dry semester (a), rainy semester (b),
and both (c). Trimester mean precipitation anomaly as a function of
vertical velocity anomaly for the January to March trimester (a),
April to June trimester (b), July to September trimester (c), October
to December trimester (d) and all trimesters (e). White squares
represent pastureland and gray squares represent soybean cropland… 93
Figure 5.13 –
Seasonal profile of vertical velocity at 500 hPa, in forest (F
ab
) for
the February to April trimester (rainy season) (a) for the June to
August trimester (dry season) (b) for the vertical velocity anomalies
in different levels of pastureland expansions: P
ab
%25
- F
ab
(c), P
ab
%50
-
F
ab
(d) and P
ab
%75
- F
ab
(e) for the February to April trimester,
respectively, and for the anomalies P
ab
%25
- F
ab
(c), P
ab
%50
-F
ab
(g) and
P
ab
%75
- F
ab
(h) for the June to August trimester, respectively. Positive
values are represented by solid line and indicate decrease in vertical
motion and negative values are represented by dashed line and
indicate increase in vertical motion…………………………………...
95
Figure 5.14 –
Seasonal profile of vertical velocity at 500 hPa, in forest (F
ab
) for
the February to April trimester (rainy season) (a) for the June to
August trimester (dry season) (b) for the vertical velocity anomalies
in different levels of soybean cropland expansions: S
25%
- F
ab
(c),
S
50%
- F
ab
(d) and S
75%
- F
ab
(e) for the February to April trimester,
respectively, and for the anomalies S
25%
- F
ab
(c), S
50%
- F
ab
(g) and
S
75%
- F
ab
(h) for the June to August trimester, respectively. Positive
values are represented by solid line and indicate decrease in vertical
motion and negative values are represented by dashed line and
indicate increase in vertical motion…………………………………...
96
Figure 5.15 –
Seasonal profile of precipitation at 500 hPa, in forest (F
ab
) for the
February to April trimester (rainy season) (a) for the June to August
trimester (dry season) (b) for the precipitation anomalies in different
levels of pastureland expansions: P
ab
%25
- F
ab
(c), P
ab
%50
-F
ab
(d) and
xiv
P
ab
%75
- F
ab
(e) for the February to April trimester, respectively, and for
the anomalies P
ab
%25
- F
ab
(c), P
ab
%50
-F
ab
(g) and P
ab
%75
- F
ab
(h) for the
June to August trimester, respectively. Negative values are
represented by dashed line and indicate decrease in pasture
precipitation and positive values are represented by solid line and
indicate increase in pasture precipitation……………………………..
97
Figure 5.16 –
Seasonal profile of precipitation at 500 hPa, in forest (F
ab
) for the
February to April trimester (rainy season) (a) for the June to August
trimester (dry season) (b) for the precipitation anomalies in different
levels of soybean cropland expansions: S
25%
- F
ab
(c), S
50%
- F
ab
(d)
and S
75%
- F
ab
(e) for the February to April trimester, respectively,
and for the anomalies S
25%
- F
ab
(c), S
50%
- F
ab
(g) and S
75%
- F
ab
(h)
for the June to August trimester, respectively. Negative values are
represented by dashed line and indicate decrease in pasture
precipitation and positive values are represented by solid line and
indicate increase in pasture precipitation……………………………..
98
Figure 5.17 – Diagram of anomalies of radiative processes………………………… 101
Figure 5.18 – Schematic representation of the precipitation changes with and
without cloud feedbacks……………………………………………... 102
xv
LISTA DE TABELAS
Table 1.1 – Main characteristics of the remote sensing albedo products …………... 08
Table 1.2 – Covariance analysis for hourly albedo data……………………………. 13
Table 1.3 – Albedo differences for forest and pasture for dry- and wet canopy
condition……………………………………………………………..… 14
Table 1.4 – Covariance analysis for monthly albedo data………………………….. 18
Table 2.1 – Parameters used by the model………………………………………….. 34
Table 3.1 – Optical parameters used by the DF99 calibration, and by the new
calibrations using the dry-canopy (DC
i
) and wet-canopy (WC
i
)
versions of the model, where i is equal to M for the Manaus-nearby
sites (Ducke and Cuieiras Reserves) or J for the Jaru Reserve. χ
leaf-up
is
the upper canopy leaf orientation, χ
leaf-lo
is the lower canopy leaf
orientation,
Leaf
loVIS −
α is the lower canopy visible leaf reflectance,
Leaf
upVIS −
α
is the upper canopy visible leaf reflectance,
Leaf
loNIR−
α is the lower canopy
NIR leaf reflectance,
Leaf
upNIR−
α
is the upper canopy NIR leaf reflectance,
f
wetmax
is maximum fraction of water cover on two-sided leaf, τ
drip
is
the decay time for intercepted liquid drip off and
ν is the ratio of the
scattering coefficients of the canopy surfaces wet by water and dry
canopy surfaces, applied individually to leaves and stems. All values
xvi
are dimensionless, except for τ
drip
, which is in seconds………………...
48
Table 3.2 – Statistics of observed and simulated surface albedo for the entire time
series and for precipitation hours at three Amazon rainforest sites
(Cuieiras, Ducke and Jaru reserves), for the simulations based on the
DF99 calibration, for the new calibration using the dry-canopy (DC)
and wet-canopy (WC) versions of the model.
X
is the average albedo,
ε is the mean relative error and RMSE is the root mean square error….
50
Table 4.1 – Main characteristics of the remote sensing albedo products…………… 62
Table 5.1 – Annual mean of the variables and of some parameters: Rn (net
radiation), Sin (downward solar radiation), Sout (upward solar
radiation), Lin-Lout (longwave balance), P (precipitation), E
(evaporation), LE (latent heat flux) and H (sensible heat flux) for the
simulations F
ab
(Forest, control run), P
ab
%25
(75% forest and 25%
pasture), P
ab
%50
(50% forest and 50% pasture), P
ab
%75
(25% forest and
75% pasture), S25% (75% forest and 25% soybean), S50% (50%
forest and 50% soybean), S75% (25% forest and 75% soybean), P
total
(extrapolating to 100% of forest replaced by pasture), S
total
(extrapolating to 100% of forest replaced by soybean), P
total
- F
ab
(
difference between pasture and forest) and S
total
- F
ab
(difference
between soybean and forest)……………………………………………
100
Table 5.2 – Coefficient of determination (R
2
) of the linear regression analysis
between Rn’ (surface net radiation),
ω’ (wind vertical velocity), C’
(total cloud) and P’ (precipitation) versus the variables in the left bar…
99
xvii
LISTA DE ABREVIATURAS
ABRACOS – Anglo Brazilian Amazonian Climate Observation Study
ADM – Angular distribution model
ANCOVA – Analysis of covariance
AVHRR – Advanced Very High Resolution Radiometer
BB – Blackbody
BRDF – Bidirectional Reflectance Distribution Function
CCM3 – NCAR Community Climate Model
CG99 – Csiszar and Gutman (1999)
COLA – Center for Ocean-Land Atmosphere Studies
ERBE – Earth Radiation Budget Experiment
GCM – General Circulation Model
GEMEX – Global Energy and Water Cycle Experiment
GOES – Geostationary Operational Environmental Satellite
IBIS – Integrated Biosphere Simulator
INPA – Instituto Nacional de Pesquisas da Amazônia
ISLSCP2 – The International Satellite Land Surface Climatology Project
LBA – Large-Scale Biosphere-Atmosphere Experiment in Amazonia
METEOSAT – Meteorological satellite
MODIS – Moderate Resolution Imaging Spectroradiometer
xviii
NASA – National Aeronautics and Space Administration
NIR – Near-infrared band
NOAA – National Oceanic and Atmospheric Administration
RMSE – Root mean square error
SD – Solar Diffuser
SDSM – Solar Diffuser Stability Monitor
SRB – Surface Radiative Budget
SRCA – Spectroradiometric Calibration Assembly
TOA – Top-of-the atmosphere
UMD – University of Maryland
VIS – Visible band
xix
RESUMO
YANAGI, Sílvia de Nazaré Monteiro, D.Sc., Universidade Federal de Viçosa, outubro de
2006. Albedo de uma floresta tropical Amazônica: medições de campo,
sensoriamento remoto, modelagem, e sua influência no clima regional. Orientador:
Marcos Heil Costa. Co-Orientadores: Humberto Ribeiro da Rocha e Yosio Edemir
Shimabukuro.
O albedo é definido como a razão entre a radiação solar refletida e a radiação solar
incidente em uma superfície. É o principal fator que afeta o balanço de radiação terrestre e
tem sido freqüentemente considerado em estudos do clima global e regional. A Amazônia
é uma das regiões do planeta onde a resposta da circulação atmosférica regional a
mudanças do albedo superficial é mais intensa. Estudos de simulação utilizando diversos
modelos mostram que a conversão da floresta tropical em pastagem causa uma redução na
precipitação local, a qual é principalmente dependente da mudança no albedo da superfície.
O presente trabalho tem como objetivos investigar as fontes de variação espacial e
temporal do albedo de uma floresta tropical amazônica, usando medições de campo,
modelagem e produtos de sensoriamento remoto; e investigar o papel da mudança do
albedo da superfície, após o desmatamento Amazônico, no clima regional. Neste trabalho
foram utilizados dados de albedo observados em quatro áreas florestais (Reservas Florestais de
xx
Caxiuanã, Cuieiras, Ducke e Jaru) e duas pastagens (Dimona e Nossa Senhora Aparecida)
pertencentes aos Projetos LBA (Experimento de Grande Escala da Biosfera-Atmosfera na
Amazônia) e ABRACOS (Estudos de Observação do Clima na Amazônia Anglo-Brasileira).
Também foram utilizados seis produtos de sensoriamento remoto: CG99, ERBE,
SRB/ISLSCP2, UMD e MODIS céu-claro e MODIS céu-escuro. Para a simulação do
albedo em floresta tropical amazônica foi utilizado o modelo IBIS (Integrated Biosphere
Simulator) em duas versões: pontual e acoplado ao modelo climático CCM3. Os resultados
deste trabalho mostram que: as principais fontes de variação do albedo em escala horária e
mensal são a cobertura vegetal e transmissividade atmosférica, e que o molhamento do
dossel é uma importante fonte de variação em escala horária e deve ser incluída em
modelos; que o albedo simulado mostrou uma forte sensibilidade aos parâmetros do dossel
superior de orientação da folha e de refletância no infravermelho e nenhuma sensibilidade
aos parâmetros do dossel inferior, consistente com a estrutura do dossel; que a
incorporação do molhamento do dossel nos cálculos de transferência radiativa melhora os
resultados de simulação em escala horária, diminuindo o albedo durante as horas de
precipitação, mas não em escala sazonal, excluindo o molhamento do dossel como uma
fonte principal da variabilidade sazonal do albedo em florestas tropicais; que os diferentes
produtos de sensoriamento remoto apresentaram uma variação considerável em relação aos
albedos observados em campo, incluindo grandes diferenças sazonais; que anomalias de
precipitação podem ser explicadas por uma relação linear entre os processos radiativos,
onde mudanças no albedo da superfície e na radiação líquida explicam aproximadamente
96% e 99% da variação anual da precipitação e que a substituição da floresta por soja afeta
mais o clima regional que a conversão em pastagem, provavelmente devido ao alto valor
do albedo das áreas de soja.
xxi
ABSTRACT
YANAGI, Sílvia de Nazaré Monteiro, D.Sc., Universidade Federal de Viçosa, October of
2006. Albedo of an Amazon tropical rainforest: field measurements, remote
sensing, modeling, and its influence on the regional climate. Adviser: Marcos Heil
Costa. Co-Advisers: Humberto Ribeiro da Rocha and Yosio Edemir Shimabukuro.
Albedo is defined as a ratio between reflected solar radiation and incoming solar
radiation over a surface. It is a main factor that affects the terrestrial radiation balance and
has been frequently considered in the global and regional climate studies. Amazon is one
of planet’s regions where the response of regional atmospheric circulation is more intense.
Simulation studies using several models show that a conversion of tropical rainforest in
pasturelands causes reduction in the local precipitation, which is mainly dependent of
surface albedo changes. The objective of present work is to investigate the sources of
spatial and temporal variation of an Amazon tropical rainforest surface albedo, using field
measurements, modeling and satellite products; and to investigate the role of surface
albedo changes, after Amazon tropical deforestation, in the regional climate. In this work it
was used observed data from four forests (Caxiuanã, Cuieiras, Ducke and Jaru Reserves) and
xxii
two pastures (Dimona and Nossa Senhora Aparecida) belonging to LBA (Large-Scale
Biosphere-Atmosphere Experiment in Amazonia) and ABRACOS (Anglo Brazilian
Amazonian Climate Observation Study) projects. It was also used six different satellite
systems: CG99, ERBE, SRB/ISLSCP2, UMD and MODIS white-sky and MODIS black-
sky. For the simulation of Amazon tropical rainforest surface albedo the point IBIS
(Integrated Biosphere Simulator) model and it coupled to CCM3 climate model was used.
The results of this work show that: the main sources of variation of albedo in hourly and
monthly scale are the vegetation cover and atmospheric transmissivity, and that the canopy
wetness is an important source of variation in hourly scale and should be included in the
models; that the simulated albedo show strong sensitivity to superior canopy parameters of
leaf orientation and infrared reflectance and no sensitivity was found for the lower canopy
parameters, consistent with the canopy structure; that an incorporation of canopy wetness
in the radiative transfer calculation improve the results in a hourly scale, reducing the
albedo during the hours with precipitation, but not in a seasonal scale, excluding the
canopy wetness as a main source of the seasonal variability in tropical rainforest; that the
different satellite systems present a considerable variation related to the field measured
albedos, including great seasonal differences; that the precipitation anomalies can be
explained by a linear relationship between radiative processes, where the changes in
surface albedo and net radiation explain approximately 96% and 99% of the annual
variation of precipitation and that the replacement of forest by soybean cropland affects
more the regional climate than the conversion in pasture, probably due to the high value of
soybean cropland albedo.
1
GENERAL INTRODUCTION
Surface albedo, or shortwave reflection coefficient, is defined as the ratio between
the total reflected solar radiation and the total incoming solar radiation at a surface. Land
cover albedo is an important land physical parameter and has frequently been considered in
studies of global and regional climate. It controls not only the amount of net radiation
available for heating the ground and lower atmosphere and for evaporating water (Rowe,
1991), but also affects the regional climate.
In tropical areas climatic simulations have demonstrated that the climate is
sensitive to changes in the surface albedo caused by natural and anthropogenic activities,
such as desertification and tropical deforestation (Charney et al., 1977; Dorman and
Sellers, 1989; Lean and Warrilow, 1989; Xue et al., 1990; Hahmann and Dickinson, 1997;
Costa and Foley, 2000; Wei et al., 2001; Zao et al., 2001; Berbet and Costa, 2003).
Amazonia is one of the planet regions where the response of regional atmospheric
circulation to changes in surface albedo is more intense, because the rainfall in that region
is mainly of convective origin. Simulation studies for Amazonian basin using several
models show that the conversion of tropical rainforest to pasturelands causes a reduction in
local precipitation, which is mainly dependent of changes in surface albedo. Dirmeyer and
2
Shukla (1994) showed, through simulations with the COLA (Center for Ocean-Land
Atmosphere Studies) model, that the annual mean precipitation decreases when the
increase of albedo is greater than 0.03, being linearly proportional to the increase of
albedo. Until recently, researches have only concerned in the effects of conversion of
tropical rainforest to pasture on the annual mean climate, without considering the eventual
cropland expansions and the seasonal effects that these changes may cause on the regional
climate (Nobre et al., 1991, Hahmann and Dickinson, 1997; Zeng et al., 1996). However,
Berbet and Costa (2003) studied the effect of the seasonal variation of surface albedo
associated to the substitution of tropical rain forest by pasture in the pattern of seasonal
precipitation, observing that the seasonal variations of the reflected radiation difference are
related to the fluctuations in seasonal scale of precipitation anomalies caused by
deforestation.
A more realistic representation of albedo in climate models will significantly
improve the accuracy of climate simulation and prediction, because land surface albedo is
an important source of uncertainty related to General Circulation models (GCMs).
Some authors discussed the albedo of a tropical rainforest in the context of climatic
models (Culf et al., 1995; Wright et al., 1996; Berbet and Costa, 2003), but only the latter
discussed the spatial and seasonal variation of albedo of a tropical rainforest and a pasture.
Berbet and Costa (2003) verified that 28% of variance in the change of precipitation after
tropical rainforest deforestation is explained by the variance of the reflected radiation. The
authors also observed that the surface model used (IBIS) does not correctly represent the
surface albedo during part of rainy season, although it does during the dry season. Culf et
al. (1995) suggested that the seasonal variation of a tropical rainforest albedo is mainly
correlated to the soil moisture, but it is believed that in fact they were referring to leaf
wetness, once the presence of liquid water on the leaves considerably alters its optical
properties (the liquid water presents strong absorption in the near infrared and high
reflectance in the blue and green bands) (Asner, 1998; Asner et al., 2000).
3
Validation of simulated albedo can be done in two ways, against field
measurements or against satellite products. Field data have the advantage of using high
frequency data to validate simulated albedo measurements. Surface albedo has been
measured in Amazonian ecosystems at four forests and two pastures belonging to LBA
(Large-Scale Biosphere-Atmosphere Experiment in Amazonia) and ABRACOS (Anglo
Brazilian Amazonian Climate Observation Study) projects. On the other hand, satellite data
allow evaluation the spatial and temporal variability of surface characteristics. However,
the satellite systems present several problems, such as a single measurement per day, loss
of data in cloud presence, aerosols contamination and incomplete cover of the solar
spectrum. Land surface albedo has been retrieved from several space-based satellites such
as AVHRR (Advanced Very High Resolution Radiometer), ERBE (Earth Radiation
Budget Experiment), and MODIS (Moderate Resolution Imaging Spectroradiometer).
The objective of this dissertation is to investigate the sources of spatial and
temporal variation of an Amazon tropical rainforest surface albedo, using field
measurements, modeling and satellite products; and to investigate the role of surface
albedo changes, after Amazon tropical deforestation, in the regional climate. It is organized
as follows: Chapter 1 presents the investigation of the sources of variation of surface
albedo of Amazonian vegetation. In Chapter 2, the radiative transfer in tropical rainforest
canopies is modeled; analyzing the sensitivity of simulated albedo to canopy architectural
and optical parameters. In Chapter 3, the importance of canopy on the tropical rainforest
albedo is evaluated using modeling. In chapter 4, a comparison of seasonal and spatial
variations of albedos by a climate model and six different satellite products for the Amazon
tropical rainforest is presented; and in chapter 5 the radiative processes of the precipitation
change after tropical deforestation is analyzed.
4
CHAPTER 1
SOURCES OF VARIATION OF ALBEDO OF AMAZONIAN VEGETATION
1.1. INTRODUCTION
Surface albedo is defined as the ratio of reflected to incident radiation in the total
solar in a certain wavelength interval. It is the main factor that affects the land radiation
balance and has frequently been considered in studies of global and regional climate. The
main identified sources of variation of land surface albedo are land cover, solar elevation
angle, canopy wetness, and cloud cover (Pinker et al., 1980; Bastable et al., 1993; Culf et
al., 1995).
The albedo of different tropical land covers has been studied for over 30 years. In
one of the first studies comparing forest and non-forest albedos in Nigeria, Oguntoyinbo
(1970) found an average albedo of 12% for forest and varying from 15 to 21% for non-
forested areas. Pinker et al. (1980) compared dry evergreen forest and pasture albedos in
Thailand, finding, during the winter monsoon season, an average midday albedo of 10.6%
5
for the forest and 13.4% for the clearing, whereas, during the summer monsoon the authors
found albedos of 12.0% and 14.6% for the forest and clearing, respectively. In this study,
difference between forest and clearing for winter and summer were 2.8% and 2.6%,
respectively. The first measurements in Amazonia, during ARME – Amazon Region
Micrometeorological Experiment, indicated an average albedo of 12.3±0.2% for a tropical
forest near Manaus, Brazil (Shuttleworth et al., 1984). Later, during ABRACOS – Anglo
Brazilian Amazonian Climate Observation Study, Bastable et al. (1993) verified an
average albedo of 13.1% for the same site and 16.3% for a nearby pasture, a difference of
3.2%. Synthesizing the measurements at three Amazonian forest sites and three pasture
sites, Culf et al. (1996) found average albedos of 13.4% and 18%, respectively (4.6%
difference).
Solar elevation, as well as variations in cloud cover, can also cause changes in
surface albedo. Pinker et al. (1980) studied the diurnal variation of the average albedo over
a tropical dry evergreen forest and a nearby clearing, finding a strong dependence of the
albedo on the zenith angle, although this dependence is weaker during cloudy days.
Shuttleworth et al. (1984) found a quadratic dependency of the tropical rainforest surface
albedo to the solar elevation angle. Usually, surface albedo present higher values in the
morning and afternoon, and the minimum value occur at solar noon. However, Song (1998)
verified that wind may change the leaves inclination causing asymmetric variation between
morning and afternoon albedos. Giambelluca et al. (1999) observed a strong diurnal and
annual cycle of albedo associated to changes in the solar elevation angle. At high solar
angles, sunlight can penetrate to greater depths within a forest canopy. This effect is
overall considered to increase with vegetation height, giving forests the lowest albedo and
grassland the highest of any type of vegetation. Diffuse sunlight more effectively
6
penetrates the incident radiation has a high proportion of diffuse sunlight (i.e., cloudy
skies), the solar angle effect on albedo is less significant (Nouvellon et al., 2000).
Most of results reviewed before are derived from observations collected over a
period of a few days, and explain the sources of variation of surface albedo at the hourly
time scale. At the monthly time scale, however, the sources of variation may be different.
Culf et al. (1995), for example, presented results of the seasonal variability of Amazonian
vegetation albedo that are inconsistent with the hourly-scale results. They suggest that the
seasonal variability of the monthly mean albedo of the Amazon forest is correlated with
soil moisture and is not correlated with solar elevation angle. In addition, they did not find
a clearly defined seasonal trend of albedo for the pasture sites. It is unclear, however,
whether this seasonal variability is driven by soil moisture-correlated variables, as
discussed by Berbet and Costa (2003), or whether it is driven by other sources of variation
correlated with rainfall, like canopy wetness and fraction of direct and diffuse radiation.
Moreover, Berbet and Costa (2003) also verified that a complex, state-of-the-art,
land surface model is unable to reproduce correctly the seasonal variability of surface
albedo of forests and pastures, although it reproduces well the annual mean and some
aspects of the seasonal variability. This indicates that there is still much to be learned – and
incorporated into models – about the sources of variation of albedo at both the hourly and
monthly time scale. This is also important because, in tropical areas, such as Amazon
forest, climatic simulations have demonstrated that the climate is sensitive to the surface
albedo variations due to the deforestation, desertification and anthropogenic activity. As a
small change in the surface albedo results in significant climate change, it is extremely
important that climate models represent the land surface with a very high accuracy,
especially in a seasonal basis.
7
Here, I investigate the sources of variation of surface albedo of tropical
vegetation, at both the hourly and monthly time scales, with the goal to provide the
necessary knowledge that will be, later, incorporated in land surface models. In particular,
I investigate the role of canopy wetness and cloud cover, in addition to the traditional
sources of variation (land cover and zenith angle), as well as the interactions among the
sources of variation.
This chapter is divided in three parts. The next section describes the sites,
instrumentation and data used in the analysis. Then, using statistical techniques, I analyze
and discuss the sources of variation of surface albedo of Amazonian vegetation, at both the
hourly and monthly time scales. I conclude discussing the relevance of the processes
involved to the remote sensing and modeling of surface albedo of Amazonian vegetation.
1.2. Sites, instrumentation and data
Field data used in this paper were measured in six experimental sites in the Amazon
during the ABRACOS (Anglo Brazilian Amazonian Climate Observation Study) and LBA
projects (Large-Scale Biosphere-Atmosphere Experiment in Amazonia). Four of the sites have
primary tropical rainforest as main land cover (Reserva Caxiuanã, Reserva Cuieiras (K34),
Reserva Ducke, and Reserva Jaru), whereas the other two are pasture sites (Fazenda Dimona and
Fazenda Nossa Senhora Aparecida). Table 1.1 and Figure 1.1 show other (additional)
information about the sites and their location.
8
Table 1.1. Description of the experimental sites.
Experimental site Coordinates Location
Measurement period
(dd/mm/yy)
Reserva Caxiuanã
Reserva K34
Reserva Ducke
Reserva Jaru
Fazenda Dimona
Fazenda Nossa Senhora
Aparecida
01º 42’S, 51º 31' W
02º 35’S, 60º 06' W
02º 57’S, 59º 57' W
10º 05’S, 61º 55”W
02º 19’ S, 60º 19’W
10º 45’S, 62º 22’W
Melgaço, PA
Manaus, AM
Manaus, AM
Ji-Paraná, RO
Manaus, AM
Ji-Paraná, RO
16/04/99 to 10/02/00
15/05/99 to 28/11/01
31/12/90 to 31/12/96
24/10/91 to 31/12/96
01/10/90 to 31/12/96
04/10/91 to 031/12/96
Figure 1.1. Orientation map.
9
Reserva Caxiuanã is an experimental area of primary tropical evergreen forest with
approximately 330 km
2
and is located in Pará State about 400 km west of Belém. Reservas
Ducke and Cuieiras (K34) are protected primary forest areas, located 25 km and 60 km
north of Manaus, Amazonas State, respectively. These sites are surrounded by undisturbed
forest for at least 5 km. Reserva Jaru is a forest reserve owned by the Brazilian
Environmental Protection Agency (IBAMA) and is located about 80 km north of Ji-Paraná,
in Rondônia State. Fazenda Dimona is cattle ranch with 10 km
2
of clearing, located about
100 km north of Manaus. There is little forest remaining in this area. Fazenda Nossa
Senhora Aparecida is a ranch located 50 km northeast of Ji-Paraná and situated in a strip of
cleared land about 4 km wide and several tens of kilometers long. The ranch is in the
center of an area of about 50 km in radius, which has undergone large-scale clearing.
At Reserva Caxiuanã, incident and reflected solar radiation were measured with a
net radiometer (Kipp & Zonen, Delft, Netherlands) mounted at 45.5 m on the top of a
tower. Radiation data were measured every 10 s and averaged half-hourly and stored on a
datalogger (21X, Campbell Scientific). At Reserva Cuieiras a piranometer (Kipp & Zonen
CM 21, Delft, Netherlands), connected to a datalogger (CR10, Campbell Scientific,
Shepshed, UK) was used to measure incident and reflected solar radiation each 30 seconds,
storing the averages every 30 minutes. For the remaining sites, the incident and reflected solar
radiation were measured using two solarimeters (Kipp and Zonen, Delft, the Netherlands). These
instruments are part of an automatic weather station (Didcot Instruments, Abingdon, UK)
connected to a datalogger (CR10, Campbell Scientific, Shepshed, UK) and hourly-averaged data
were recorded.
Data of incident and reflected global solar radiation used in this study are available on-
line through www.cptec.inpe.br/abracos/available.html and www.lba.cptec.inpe.be/beija-flor.
10
Data for the Reservas Jaru and Ducke, and Fazendas Dimona and Nossa Senhora Aparecida are
the same used by Culf et al. (1995, 1996). Hourly averages of surface albedos are determined
through the ratio between reflected and incident solar radiation. Hourly albedo data of six
experimental sites (Table 1.1) are used to study both the hourly and monthly sources of
variation of tropical rainforest and pasture albedo.
For the hourly-scale, I consider four possible sources of variation: type of land
cover (forest, pasture), solar zenith angle (0º to 90º), canopy wetness (dry, wet) and
atmospheric transmissivity (0 to 1), as an indicator of the partition of the incident radiation
between direct and diffuse components. To avoid the undesirable effects of aerosols,
common on this region, I considered only data collected a few days after a large storm
(rainfall greater than 10 mm), when I assumed that the atmosphere would be clean of
aerosols. With the help of hourly precipitation data, I also filtered the data to represent wet-
and dry-canopy conditions: data collected between the beginning of a rain event and up to
three hours after the end of a rain are used to represent wet-canopy conditions; data
collected at least 24 hours after the end of a rain event are used to represent dry-canopy
conditions; while data collected between three and 24 hours after a rain event are discarded
from the analysis, as I am unsure about the canopy wetness status. Hourly atmospheric
transmissivity is calculated as the ratio between incident solar radiation at the surface and
calculated incident solar radiation at an horizontal surface at the top of the atmosphere.
For the monthly time scale, I considered three possible sources of variation: type
of land cover, monthly-averaged solar zenith angle, and monthly-averaged atmospheric
transmissivity. Data were not filtered for the presence of aerosols or for the canopy
wetness conditions, except where noted.
11
Other possible sources of variation of land surface albedo are not included in this
analysis, such as variation among sites that have the same land cover, and interannual
variability due to changes in soil moisture or leaf area index.
1.3. Sources of variation of hourly albedo
An analysis of covariance (ANCOVA) evaluates and analyzes the effects of land
cover, canopy wetness, atmospheric transmissivity and zenith angle on the surface albedo
in Amazon region. I chose the analysis of covariance due to the combined effects of
qualitative (vegetation and canopy wetness) and quantitative (transmissivity and solar
zenith angle) factors on the albedo. Data were analyzed using the PROC GLM (SAS,
2001) with the model including a factorial combination of two levels of land cover, L,
(forest and pasture), two levels of canopy wetness, ω, (dry and wet) and the fixed effects of
two regression factors (atmospheric transmissivity, τ, and zenith angle, Z) on the analysis,
as follows:
ijklijijkijkijkijkijk
ijkijkijkijkijkijkijjiijkl
LLLL
LLbaLL
εωτωττωωτω
ω
τ
τ
τ
τ
ω
ω
μ
α
Z
+⋅⋅⋅+⋅Ζ⋅+⋅Ζ⋅+⋅⋅Ζ+⋅⋅+⋅Ζ
+⋅Ζ+
⋅
+
⋅
+
Ζ
⋅
+
Ζ
⋅
+
⋅+⋅
+
++=
(1)
where α
ijkl
is the estimated albedo; μ is the overall mean; L
i
is the effect of the ith level of
factor L;
ω
j
is the effect of the jth level of factor ω; a is the true common slope of the
covariate
ijk
τ
;
ijk
τ
is the overall average of the covariate τ for the observations in the study;
b is the true common slope of the covariate Z
ijk
; Z
ijk
is the overall average of the covariate
Z for the observations in the study; the products of the sources of variation represent the
effect of the interaction between them (e.g. L·ω
ij
, etc.);
ε
ijkl
is the random effect due to
sampling, normally and independently distributed with mean zero and variance σ
2
(
)
[
]
2
,0 NID
σε
ijkl
.
12
In Table 1.2, I present results of the ANCOVA based on the type I error (Cochran
and Cox, 1992). I chose the type I error because it quantifies the contribution of each factor
and its interactions for the model, adjusted to the previous factors. At the hourly-scale, land
cover (forest or pasture) explains 43.99% of the variance of the albedo. The addition of
canopy wetness (ω) to the model increases the explanation of the variance by 3.42%.
Further inclusion of the factors τ and Z, adjusted to the previous factors; improve the
variance explained by the model by 6.77% and 0.06%, respectively. Overall, the model
explains 57.08% of the variance of the albedo. The sources of variation L, ω, τ and Z, and
the interactions τ·Z and τ·ω explain 56.25% of the total variance of the model, while the
rest of the interactions explain only 0.83% (Table 1.2). Below, I discuss the effect of these
main sources of variation and interactions.
As expected, tropical rainforest albedo is significantly different from pasture
albedo, P<10
-10
(Table 1.2), with average values of 12.1% and 16.7%, respectively. These
results are in agreement with previous publications by Oguntoyinbo (1970), Shuttleworth
et al. (1984), Bastable et al. (1993) and Culf et al. (1996).
13
Table 1.2. Covariance analysis for hourly albedo data.
Source of variation
Degrees of
Freedom
Mean
Square
F Value
Probability
F*
Variance
explained
(%)
Model 15 0.2573 490.47 < 10
-10
–
Land cover (L) 1 2.9744 5670.77 < 10
-10
43.99
Canopy wetness (ω)
1 0.2313 441.07 < 10
-10
3.42
L·ω
1 0.0004 0.72 0.3962 0.01
Atmospheric
transmissivity (τ)
1 0.4577 872.59 < 10
-10
6.77
Zenith angle (Z) 1 0.0043 8.25 0.0041 0.06
τ·Z
1 0.0279 53.26 < 10
-10
0.41
τ·L
1 0.0002 0.43 0.5120 0.003
τ·ω
1 0.1084 206.66 < 10
-10
1.60
Z·L 1 0.0064 12.11 0.0005 0.09
Z·ω
1 0.0198 37.68 < 10
-10
0.29
τ·L·ω
1 0.0014 2.71 0.0998 0.02
Z·L·ω
1 0.0006 1.06 0.3033 0.01
τ·Z·L
1 0.0208 39.60 < 10
-10
0.31
τ·Z·ω
1 0.0051 9.66 0.0019 0.07
τ·Z·L·ω
1 0.0003 0.53 0.4666 0.004
Error 5533 0.0005 – – –
Total variance explained 57.08%
CV = 17.23
* Indicates significance at the specified level of probability.
Albedos determined for wet and dry canopies are significantly different (P<10
-10
)
independent of vegetation type. The average difference between albedos for dry and wet
canopy is 0.4% (Table 1.3), while there is no significant interaction (P∼0.4) between the
factors land cover and canopy wetness. Overall, reflectance of a wet canopy is smaller than
14
the reflectance of a dry canopy for both ecosystems, forest and pasture, because water has
a very high absortance, especially in the near infrared band. The same difference between
pasture albedo and forest albedo (0.046) is found when only dry canopy or only wet
canopy albedos are analyzed, which confirms the non-significant interaction between land
cover (L) and canopy wetness (ω).
Table 1.3. Albedo differences for forest and pasture for dry- and wet canopy condition.
Vegetation All Dry Wet
Difference
(Dry – Wet)
Forest
0.121 0.123 0.119 0.004
Pasture
0.167 0.169 0.165 0.004
Difference (Pasture – Forest)
0.046 0.046 0.046 -
Surface albedo is also influenced by zenith angle and by the interaction between τ
and Z as verified in previous studies (Pinker et al., 1980; Shuttleworth et al., 1984;
McCaughey, 1987; Giambelluca et al., 1999). Surface albedo only depends on zenith angle
when direct solar radiation is predominant, which explains why the effect of the zenith
angle (P = 0.0041) is less significant than the effect of the interaction between τ and Z
(P<10
-10
, Table 1.2).
Figure 1.2 and Figure 1.3 show the albedo of a tropical rainforest and pasture,
respectively, for four transmissivity ranges (0-25%, 25-50%, 50-75% e 75-100%) as a
function of solar zenith angle for dry- (Figure 1.2a and 1.3a) and wet-canopy (Figure 1.2b
and 1.3b) conditions. The 0-25% transmissivity interval characterizes overcast sky
condition with diffuse solar radiation predominance, whereas the 75-100% transmissivity
interval indicates clear sky condition with predominance of direct solar radiation. Overall,
15
forest and pasture dry albedos under clear sky tend to increase with increasing solar zenith
angle as observed by other authors (Pinker et al., 1980; McCaughey, 1987), while albedo is
independent of solar zenith angle under overcast conditions, as also observed by Eck and
Deering (1992) and Bégué et al. (1996).
Figure 1.2. Forest albedo as a function of solar zenith angle for four transmissivity ranges
(0-25%, 25-50%, 50-75% e 75-100%) for (a) dry forest canopy and (b) wet
forest canopy.
16
Figure 1.3. Pasture albedo as a function of solar zenith angle for four transmissivity ranges
(0-25%, 25-50%, 50-75% e 75-100%) for (a) dry pasture canopy and (b) wet
pasture canopy.
Figures 1.2 and 1.3 also confirm that atmospheric transmissivity affects surface
albedo independently of land cover type (P<10
-10
, Table 1.2). For example, lower values of
albedo for wet-canopy are observed for 0-25% transmissivity interval (Figure 1.2b and
1.3b). Albedo tends to be approximately constant when diffuse irradiance predominates,
because the diffuse light beams penetrate more deeply into the vegetation canopy and is
more effectively absorbed (Nouvellon et al., 2000).
17
1.4. Sources of variation of monthly albedo
I also used an analysis of covariance (ANCOVA), through PROC GLM (SAS,
2001), to analyze the monthly albedo data via a factorial design. The factors land cover
(L), atmospheric transmissivity (τ), zenith angle (Z) and the interactions τ·Z, τ·L, Z·L and
τ·Z·L are considered, but the canopy wetness is not included. The statistical model tested
by the ANCOVA is
ijklijkijkijkijkijkijkiijkl
LLLbaL
ε
τ
τ
τ
τ
μ
α
+⋅
Ζ
⋅
+
⋅
+
Ζ
⋅
+
⋅
Ζ
+
Ζ⋅+⋅++=
(2)
Analogously to the previous section, I also used the type I error to compose the
mean square for the monthly albedo analysis (Table 1.4). At the monthly-scale, land cover
(L) explains 69.74% of the variance of the albedo. Addition of the effect of the
atmospheric transmissivity (τ) to the model increases the explanation of the variance by
3.20%, while inclusion of the zenith angle (Z) adds an explanation of 0.91% of the
variance. Overall, the model explains 76.91% of the variance of the model. The sources of
variation L, τ and Z explain 73.85% of the total variance of the model, while the
interactions among them explain only 3.06%.
Similar to the hourly scale analysis, forest and pasture albedos are significantly
different (P<10
-10
, Table 1.4), with average values of 12.3% and 17.4%, respectively.
Atmospheric transmissivity and zenith angle also affect surface albedo (P<10
-10
and P =
0.00094, respectively) independently of land cover type.
Figure 1.4a shows the seasonal variability of albedo for forest and pasture
ecosystems, showing also that the smaller difference of albedos occur between September
and November, as also observed by Culf et al. (1995, 1996) and Wright et al. (1996). It is
fundamental to understand what causes this variability. In particular, I discuss two
possibilities raised in the introduction: seasonal changes in atmospheric transmissivity and
seasonal changes in canopy wetness.
18
Table 1.4. Covariance analysis for monthly albedo data.
Source of variation
Degrees of
Freedom
Mean
Square
F Value
Probability
F*
Variance
explained
(%)
Model 7 0.0291 135.65 < 10
-10
–
Land cover (L) 1 0.1846 860.90 < 10
-10
69.74
Atmospheric
transmissivity (τ)
1 0.0085 39.57 < 10
-10
3.20
Zenith angle (Z) 1 0.0024 11.18 0.00094 0.91
Z·L 1 0.0017 7.99 0.00504 0.65
τ·Z
1 0.0019 8.79 0.00328 0.71
τ·L
1 0.0004 2.02 0.15633 0.16
τ·Z·L
1 0.0041 19.07 0.00002 1.54
Error 285 0.0002 – – –
Total variance explained 76.91%
CV = 10.26
* Indicates significance at the specified level of probability.
The role of atmospheric transmissivity is demonstrated by Figures 1.4b and 1.5a.
Figure 1.4b shows the variation of atmospheric transmissivity above each land cover type
throughout the year. Atmospheric transmissivity over the forest sites is slightly greater (i.e.
less cloudiness) than over the pasture sites during the dry season (June to November) and
smaller during the wet season (December to May). The lower atmospheric transmissivity
over the pasture sites during the dry season is probably related to the presence of aerosols
(dust, soot, etc.) due to the constant fires this time of the year. During the dry season in
Marabá County, Pará State, Fisch et al. (1994) found mean surface albedo of 0.19 and 0.08
during the days without and with burning occurrence, respectively. In addition, Cutrim et
al. (1995) and Negri et al. (2004) verified an increase in shallow cumulus over deforested
areas during the dry season in the Amazon. Figure 1.5a shows that α'
(α' = α
P
– α
F
,
19
difference between the pasture and forest albedos) decreases with the increase of τ. So, dry
season (τ>0.45) albedos tend to be closer than rainy season (τ<0.45) albedos. The residuals
between the observed albedo and the albedo estimated by the regression equations in
Figure 1.5a are shown in Figure 1.5b. The seasonal variability of the residuals for the
pasture is probably related to the seasonal variability of pasture leaf area index, as
discussed by Wright et al. (1996). So far, I do not have an explanation for the seasonal
variability of the residuals for the forest.
Figure 1.4. Average monthly variability of (a) surface albedo and (b) atmospheric
transmissivity for tropical rainforest and pasture.
20
Figure 1.5. (a) Average monthly albedo for the forest and pasture ecosystems as a function
of atmospheric transmissivity (τ), and regression equations; (b) residuals
between the observed albedo and the albedo estimated by the regression
equations in Figure 1.5a.
Using the dataset of the previous section, I also analyzed the effect of canopy
wetness on the monthly variability of the surface albedo (Figure 1.6). At the monthly time
scale, although the canopy wetness causes a small reduction in the albedo of both forest
and pasture, its effect on the difference of albedo (α') is independent of the canopy wetness
status.
21
Figure 1.6. Average monthly variability of surface albedo for tropical rainforest and
pasture and its difference, considering the canopy wetness, where α
F wet
and
α
F dry
are the wet and dry forest albedos, respectively; α
P wet
and α
P dry
are the
wet and dry pasture albedos; and α'
P-F wet
and α'
P-F dry
are the difference
between surface albedos for wet and dry pasture and forest canopy,
respectively.
1.5. CONCLUSIONS
Data from six experimental sites were used to analyze the sources of variation of
hourly and monthly tropical vegetation albedo. At the hourly-scale data, land cover,
atmospheric transmissivity and canopy wetness are the most important sources of
variation; the effect of canopy wetness on the albedo is independent of land cover and
reduces the albedo in 0.004 for both the forest and pasture ecosystems. The interactions τ·Z
and τ·Z·L are more important as drivers of variations on the albedo than the zenith angle by
itself.
At the monthly time scale, land cover and atmospheric transmissivity are the major
sources of variation. The results presented here let me conclude that, although the
22
difference in surface albedo (α') is partially dependent on τ (a consequence of the
interaction L·τ), it is independent of the canopy wetness.
The classical paper of Sellers (1985), which provided the theoretical basis for most
of the land surface models that simulate land surface albedo, assumes that the canopy
reflectance is dependent mainly on land cover parameters, zenith angle, and fraction of
direct and diffuse radiation. Such models usually incorporate only the reflectance effects of
canopy snow, neglecting canopy wetness. The results of this study indicate that the canopy
wetness is the third most important source of variation of hourly surface albedo, and can be
included in such models. In addition, although the effect of direct and diffuse radiation is
represented separately in such models, the partition between them from the incident solar
radiation is usually poorly represented, and may be one of the main reasons why state-of-
the-art land surface models are still unable to reproduce correctly most of existing
variability in the data (Berbet and Costa, 2003).
It has been demonstrated that even a 0.03 change in tropical land surface albedo is
sufficient to drive important large-scale changes in the regional atmospheric circulation
(Dirmeyer and Shukla, 1994) and, hence, provoke climate change. This very small signal
introduces a significant problem to our ability to correctly simulate the climate change
after a tropical deforestation, because a hypothetical uncertainty of 0.01 in the simulated
albedo would represent 1/3 of the climate change signal. In this context, considerable
efforts are needed to improve the parameterizations for the tropical rainforest albedo and
the land surface radiation balance, and reduce the uncertainty in the forcing of climate
models. A rigorous incorporation of additional sources of variation in land surface models
would improve the accuracy of land surface albedo simulations and would allow for more
consistent experiments to evaluate the climate change after a tropical deforestation.
23
1.6. NOMENCLATURE
L Land cover (pasture and forest)
Ζ
solar zenith angle (0º to 90º)
ω
canopy wetness (dry, wet)
τ
atmospheric transmissivity (0 to 1)
α
ijkl
estimated albedo
μ
overall mean
L
i
effect of the ith level of factor L
ω
j
effect of the jth level of factor ω
a
true common slope of the covariate
ijk
τ
ijk
τ
overall average of the covariate τ
b
true common slope of the covariate Z
ijk
Z
ijk
overall average of the covariate Z
ε
ijkl
random effect due to sampling, normally and independently distributed with
mean zero and variance σ
2
σ
2
Variance
α’
difference between the pasture and forest albedos
α
p
pasture albedos
α
f
forest albedos
α
F wet
wet forest albedos
α
F dry
dry forest albedos
α
P wet
wet pasture albedos
α
P dry
dry pasture albedos
α'
P-F wet
difference between surface albedos for wet pasture and forest canopy
α'
P-F dry
difference between surface albedos for dry pasture and forest canopy
24
CHAPTER 2
MODELING RADIATIVE TRANSFER IN TROPICAL RAINFOREST
CANOPIES: SENSITIVITY OF SIMULATED ALBEDO
TO CANOPY ARCHITECTURAL AND OPTICAL PARAMETERS
2.1. INTRODUCTION
Global climate models (GCMs) simulate the evolution of climate based on
physical principles as well as initial and boundary conditions. To do so, these models must
represent the exchanges of radiation, heat, momentum and mass between the atmosphere
and the underlying surface, in particular terrestrial environments. Practical considerations,
however, require that small scale processes be parameterized in terms of larger scale
variables.
The parameterization of terrestrial surface processes in climate models is
confined to modules implementing vertical exchange models. The radiation component of
these modules relies on solutions derived from two-stream approaches (Dickinson, 1983;
25
Sellers, 1985), which follow developments made in the field of atmospheric physics
(Coakley and Chylek, 1975; Meador and Weaver, 1980). Compared to the atmosphere
analogue, the radiation transfer in plant canopies is rendered complex because the
elementary scatterers – leaves and stems – are large compared to the typical wavelength of
solar radiation, they can be oriented and clumped and they exhibit complex variable optical
properties. The two-stream formulations thus have to be adapted to represent, at least in
simplified forms, the effects of these complexities. Solutions have been developed for
multiple possibilities of leaves and stems orientations, from strictly vertical (eroctophile) to
strictly horizontal (planophile), including spherical and any other orientation in between,
allowing a more realistic representation of vegetation canopies (Pinty et al., 2006).
Despite the several solutions available, model exercises usually do not take full
advantage of the theoretical developments in canopy radiative transfer. For example, leaf
orientation, an important architectural parameter, has usually been assumed to have a
spherical orientation for broadleaf trees, and upright orientation for grasses, but the actual
average orientation or its probability density function are generally not determined.
GCM-grade radiation transfer schemes are constrained by several limitations:
they must be computer efficient and numerically stable, must use measurable or retrievable
variables or parameters, must provide sufficiently accurate estimations of the radiant
fluxes; and must respect energy conservation principles, which require that reflected,
transmitted and absorbed fluxes sum up to the incident radiation, independent of the
assumed canopy structure inside the domain. Radiation schemes should therefore be able
to simulate accurately both the flux reflected from the top of the canopy, that is its albedo,
and the flux transmitted to the ground underneath the vegetation layer. In this modeling
context, the albedo is a prime candidate for validation exercises.
26
Here in this study I use the above canopy albedo as the main indicator of model
performance, and evaluate the sensitivity of the simulated albedo of a tropical rainforest to
a set of canopy architectural and optical parameters, with the goals of (a) understanding the
response of the model to the several canopy optical parameters, and (b) obtaining the best
set of parameters to be used in climate models, in particular leaf and stem orientation. I
modeled the radiative transfer using the radiative transfer module of the Integrated
Biosphere Simulator (IBIS), and I validate the simulations against albedo measurements
taken at the tropical rainforest site at the Cuieiras Biological Reserve (K34).
2.2. MATERIALS AND METHODS
2.2.1. IBIS description
In this study a 0-D version of the Integrated Biosphere Simulator–IBIS (Foley et
al., 1996) is used to model the radiative transfer in a tropical rainforest canopy. Although
the model includes representations of several land surface processes (energy, water and
momentum exchange between the soil-vegetation-atmosphere system, canopy physiology,
vegetation phenology, vegetation dynamics, and terrestrial carbon balance), this study
concerns only the solar radiation balance.
The albedo of a vegetation canopy layer (
α) is defined as the ratio of the upwelling
to the downwelling solar radiant fluxes at the top of the canopy, both depending on the
location of the source, i.e., the cosine of the Sun zenith angle (
μ) and type of illumination
(normally both direct and diffuse). Representation of the albedo of such system requires
the adoption of assumptions or simplifications. IBIS assumes that the surface albedo is
approximated by a simple weighting of two distinct surface albedo types, each associated
with an incident irradiance field: the directional hemispherical reflectance (α
d
), associated
with an incident irradiance field that is purely collimated (
d
in
I ) and the indirect
27
hemispherical reflectance (α
i
), associated with an incident irradiance field that is purely
isotropic (
i
in
I
). Both albedo types can be combined to approximate the surface albedo as
follows (Kondratyev, 1972):
i
in
d
in
ii
in
dd
in
II
II
+
+
=
αα
α
(1)
In IBIS the exchange of solar radiation between the soil-vegetation-atmosphere
system is calculated following the standard two-stream approximation, with separate
calculations for direct and diffuse radiation in both visible and near-infrared bands. It
solves the canonical radiative transfer problem of two-stream vegetation layer plus
underlying surface of known albedo. Starting from the known soil albedo (α
g
), the method
is to first calculate the albedo of the combined lower canopy-ground system (α
g-lo
), then
the albedo of the combined upper canopy-lower canopy-soil system (α
lo-up
).
As shown in Figure 2.1, effects of clumping and partial vegetation cover are also
treated. When fluxes are passed between the upper and lower story, or between the lower
story and the ground, the two-stream fluxes passing through the canopy are merged with
the unmodified fluxes passing through the gaps (weighting with respect to the appropriate
fractional cover f
up
or f
lo
). Assuming that there is no snow in the canopy, the albedo of each
spectral band Λ is given by:
(
)
lo
d
lolo
d
g
d
log
ff
ΛΛΛ−
+−⋅=
ααα
1
(2)
(
)
lo
i
lolo
i
g
i
log
ff
ΛΛΛ−
+−⋅=
ααα
1
(3)
(
)
up
d
upup
d
log
d
uplo
ff
ΛΛ−Λ−
+−⋅=
ααα
1
(4)
(
)
up
i
upup
i
log
i
uplo
ff
ΛΛ−Λ−
+−⋅=
ααα
1
(5)
28
L
o
w
e
r
C
a
n
o
p
y
U
p
p
e
r
C
a
n
o
p
y
I
in
(
VIS,NIR
)
χ
Llo
χ
Lup
ρ
VIS,up
ρ
NIR,up
τ
VIS,up
τ
NIR,up
ground
ρ
VIS,lo
ρ
NIR,lo
τ
VIS,lo
τ
NIR,lo
d
I
in
I
I
I
I
I
out
I
I
I
f
lo
f
up
ground-lo interface
lo-up interface
Figure 2.1. Schematic representation of the radiative transfer model in IBIS.
The two-stream algorithm uses several canopy architectural and optical parameters.
The canopy architectural parameters include the upper and lower canopy element
orientation (χ
up
and χ
lo
); fraction of ground area covered by lower and upper canopy (f
lo
and f
up
), leaf area index (L) and stem area index (S). The canopy optical parameters include
the upper and lower canopy leaf reflectance by visible (VIS) and near-infrared (NIR)
spectral band Λ (ρ
Λ
: ρ
VIS,up
,ρ
VIS,lo
, ρ
NIR,up
and ρ
NIR,lo
), and the upper and lower canopy leaf
transmittance by visible (VIS) and near-infrared (NIR) spectral band Λ (τ
Λ
: τ
VIS,up
,τ
VIS,lo
,
τ
NIR,up
and τ
NIR,lo
). The procedure below describes the major theoretical aspects of the two-
stream approximation, and must be repeated for each vegetation layer.
29
Within each vegetation layer, the upward and downward diffuse fluxes obey
()
()
[]
()
SLKdii
eKII
SLd
dI
+−
↓
↑
↑
=−−−+
+
−
βμωωβωβμ
11
(6)
()
()
[]
()
()
SLKdii
eKII
SLd
dI
+−↑
↓
⋅−=−−−+
+
−
↓
βμωωβωβμ
111
(7)
where I
↑
and I
↓
are the upward and downward diffuse radiative fluxes per unit incident
flux, K= G(µ)/µ is the optical depth of direct beam per unit leaf and stem area, G(µ) is the
relative projected area of leaf and stem elements in the direction cos
-1
µ,
μ
is the average
inverse diffuse optical depth per unit leaf and stem area, ω is a scattering coefficient, β
i
and
β
d
are upscatter parameters for indirect (diffuse) and direct beam radiation, respectively, L
is the single-sided leaf area index, and S is the single-sided stem area index. Given the
direct beam ground albedo
d
g Λ
,
α
and diffuse ground albedo
i
g Λ
,
α
for each spectral band Λ,
these equations are solved to calculate the fluxes, per unit incident flux, absorbed by the
vegetation, reflected by the vegetation, and transmitted through the vegetation for direct
and diffuse radiation, for visible (Λ < 0.7 µm) and near-infrared (Λ ≥ 0.7 µm) spectral
bands.
Equations (6) and (7) are then solved analytically for a two-stream layer underlain
by a non-specular surface of known albedo (i.e., the ground), with prescribed incoming
downward direct and diffuse fluxes. This canonical solution (Equations 8 and 9) is applied
first to the lower story underlain by soil to obtain the effective albedo of that system, then
to the upper story underlain by the lower story and soil system, yielding the overall canopy
albedo. The analytical solutions for the single scattering albedo of direct (
d
Λ
α
) and indirect
(
i
Λ
α
) fluxes are
30
()
()
()
()
()
⎪
⎭
⎪
⎬
⎫
⎪
⎩
⎪
⎨
⎧
+
−
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⋅
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−+⋅
⎥
⎦
⎤
⎢
⎣
⎡
+−−
⎪
⎭
⎪
⎬
⎫
⎪
⎩
⎪
⎨
⎧
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
+−
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−−⋅
⎥
⎦
⎤
⎢
⎣
⎡
+−+=
+−
+−
Λ
ΛΛ
ΛΛ
Λ
ΛΛ
ΛΛ
+−
Λ
SLh
SLh
d
g
i
d
d
g
i
d
SLh
d
e
hb
ehbKb
h
K
d
hbhbKb
h
K
ed
h
μ
α
βω
μμ
σ
βμω
μ
α
βω
μμ
σ
βμω
σ
α
1
1
1
1
1
1
1
(8)
()()
()
[]
()
()
[]
1
2
2
−
Λ
ΛΛ
Λ
ΛΛ
+−
+−
Λ
ΛΛ
Λ
ΛΛ
ΛΛΛ
⎪
⎪
⎭
⎪
⎪
⎬
⎫
⎥
⎥
⎥
⎥
⎥
⎦
⎤
⎢
⎢
⎢
⎢
⎢
⎣
⎡
+−
⎟
⎟
⎟
⎟
⎟
⎠
⎞
⎜
⎜
⎜
⎜
⎜
⎝
⎛
−+
−−
⋅
+
−
⎪
⎪
⎩
⎪
⎪
⎨
⎧
⎥
⎥
⎥
⎥
⎥
⎦
⎤
⎢
⎢
⎢
⎢
⎢
⎣
⎡
⋅
⎟
⎟
⎟
⎟
⎟
⎠
⎞
⎜
⎜
⎜
⎜
⎜
⎝
⎛
−−
−+
⋅−−+=
hb
hb
hb
e
hb
e
hb
hb
hbhb
i
g
i
i
g
i
SLh
SLh
i
g
i
i
g
i
ii
μ
α
βω
μ
α
βω
μ
μ
α
βω
μ
α
βω
μ
μμβωα
(9)
where:
(
)
i
b
ΛΛ
−−=
βω
11,
()
(
)
[
]
()
()
[]
2
2
2
1
1
i
did
bK
KbKh
ΛΛ
ΛΛΛΛΛ
−−
−+−−
=
βωμ
ββωμβμω
σ
,
()
μ
βω
2
2 i
b
h
ΛΛ
−
=
and
()
()
()
()
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⋅
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−+⋅−−
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−−⋅
+
=
+−
Λ
ΛΛ
Λ
ΛΛ
+−
SLh
d
g
i
d
g
i
SLh
ehbhbhb
e
hb
d
α
βω
μμ
α
βω
μ
μ
1
The remaining issue to be described concerns the relevant expressions for G(µ), ω,
β
i
, β
d
, and
μ
. In the specific case of structurally homogeneous vegetation canopy layers,
the elementary scatterers are modeled as oriented plates of finite small size. Depending on
the vegetation type and environmental conditions, the orientation probability of the
normals to these plates may follow various distributions including planophile, erectophile,
or even heliotropic. Once the leaf angle probability distribution function is given, it
31
becomes feasible to express the extinction coefficient of any elementary volume and thus
the total extinction of the vertically homogeneous vegetation layer (e.g., Ross, 1981;
Dickinson, 1983; Verstraete, 1987). This extinction coefficient, traditionally expressed
with Ross (1981) G function, modulates the optical thickness of the homogeneous
vegetation layer. The generic expression (Equation 10) for leaves with an orientation
parameter χ is
()
hv
G
χ
μ
χ
π
πμ
μ
⎟
⎠
⎞
⎜
⎝
⎛
−
+
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−−
+=
2
12
2
14
2
1
2
(10)
where the leaf orientation χ, also defined as the departure of leaf angles from a random
distribution, equals +1 for horizontal leaves, 0 for random leaves, and –1 for vertical
leaves. χ
h
and χ
v
are the positive and negative parts of χ, respectively.
The ω β
i
and ω β
d
parameters should thus be expressed via the G function and the
leaf orientation probability distribution. The ω β parameter is the integral over the
appropriate leaf orientation probability distribution, performed between 0 and π/2, of the
scatter parameter of an individual scattering element, leaf or stem (Norman and Jarvis,
1975). Solving the integral, the upscatter parameter for diffuse radiation is (Equation 11)
⎟
⎠
⎞
⎜
⎝
⎛
+
+
⎟
⎠
⎞
⎜
⎝
⎛
−
+
⎟
⎠
⎞
⎜
⎝
⎛
+
=
ΛΛΛΛΛΛ
ΛΛ
3
2
36
τρτρ
χ
τρ
χβω
hvi
(11)
and the upscatter parameter for direct beam radiation is (Equation 12)
()
()()() ()
[]
()
μτμρτρμχτρ
μ
χβω
ccc
c
hvd
ΛΛΛΛΛΛΛΛ
+−+−⋅+−
−
= 1
2
12
(12)
where
()
μ
c is a transmittance coefficient that varies between 0.5 for
μ
= 0 and 0.1667 for
μ
= 1.
By definition, the scattering coefficient is the sum of the reflectance and
transmittance of the scattering element (Equation 13):
32
ΛΛΛ
+=
τρω
(13)
where
Λ
ρ
is a weighted combination of the leaf and stem reflectances
(
)
stemleaf
Λ
Λ
ρρ
,
(Equation 14):
stem
stem
leaf
leaf
ww
ΛΛΛ
+=
ρρρ
(14)
where
()
SLLw
leaf
+
= / and
(
)
SLSw
stem
+
=
/;
Λ
τ
is a weighted combination of the leaf
and stem transmittances
(
)
stemleaf
Λ
Λ
ττ
, (Equation 15)
stem
stem
leaf
leaf
ww
ΛΛΛ
+=
τττ
(15)
Finally, the average inverse diffuse optical depth per unit leaf and stem area (
μ
) is
given by Equation 16
∫
=
1
0
'
)'(
'
μ
μ
μ
μ
d
G
(16)
after Dickinson (1983).
μ
varies between 0.90 and 1.04 for the planophile and erectophile
cases, and 1.00 for the spherical cases.
2.2.2. Experimental site and data
Field data used in this study were measured at K34 site in the Cuieiras Biological
Reserve (2º 35’S, 60º 07’W, 90 m above sea level), during the LBA project (Large-Scale
Biosphere-Atmosphere Experiment in Amazonia). The Cuieiras Biological Reserve is an
INPA (Instituto Nacional de Pesquisas da Amazônia) protected forest reserve, about 60 km
north of Manaus, embedded in a vast area of pristine rainforest. The 50 m tall K34 tower,
erected in 1999, is located on a medium sized plateau (2°36’32. 67’’S, 60°12’33. 48’’W,
130 m) (Araújo et al., 2002). The natural vegetation and topography of this site are
representative of much of Central Amazonia.
33
I used incident and reflected solar radiation data, collected from June 1999 to
September 2000 by a piranometer (Kipp & Zonen CM 21, Delft, Netherlands), installed at
a height of 44.6 m above the forest ground surface, connected to a datalogger (CR10,
Campbell Scientific, Shepshed, UK). The data were measured every 30 seconds, and the
averages stored every 30 minutes.
2.2.3. Model sensitivity to canopy architectural and optical parameters
I carried out a sensitivity analysis of the albedo simulated by the model to six
canopy architectural and optical parameters: the upper canopy leaf and stem orientation
(χ
up
), the lower canopy leaf orientation (χ
lo
), upper and lower canopy leaf reflectance on
visible (VIS) and near-infrared (NIR) spectral band Λ (ρ
Λ
: ρ
VIS,up
,ρ
VIS,lo
, ρ
NIR,up
and ρ
NIR,lo
,
respectively). Other canopy architectural and optical parameters are set to the values in
Table 2.1. The model was then run several times with different combinations of the
parameters above, to determine in detail the sensitivity of the model to these parameters.
34
Table 2.1. Parameters used by the model.
Architectural parameters
Value
single-sided leaf area index (L) 6.175
single-sided stem area index (S) 0.025
fraction of overall area covered by lower canopy (
f
lo
) 0.500
fraction of overall area covered by upper canopy (
f
up
) 0.975
Optical parameters
direct and diffuse beam ground albedo on visible (VIS) spectral band (a
g,VIS
) 0.10
direct and diffuse beam ground albedo on near-infrared (NIR) spectral
band(
a
g,NIR
)
0.40
lower canopy leaf transmittance on visible (VIS) spectral band (
ρ
VIS,lo
)
0.07
upper canopy leaf transmittance on visible (VIS) spectral band (
ρ
VIS,up
)
0.05
lower canopy leaf transmittance on near-infrared (NIR) spectral band (
ρ
NIR,lo
)
0.25
upper canopy leaf transmittance on near-infrared (NIR) spectral band (
ρ
NIR,up
)
0.20
2.3. RESULTS AND DISCUSSION
Initial tests (not shown) demonstrated that the above canopy albedo is not sensitive
to the lower canopy parameters. This is because of the high leaf and stem area indices of
the upper canopy. I use ρ
VIS,lo
= 0.062, ρ
NIR,lo
= 0.60 and χ
lo
= 0.10 and concentrate the
discussion on the sensitivity of simulated albedo to the model upper canopy architectural
and optical parameters χ
up
and,ρ
VIS,up
and ρ
NIR,up
, respectively.
Several combinations of χ
up
, ρ
VIS,up
,ρ
NIR,up
were tested looking for the best mean
simulated value of surface albedo for the study period. A strong sensitivity of simulated
albedo to these parameters was observed, resulting in optimal values of simulated albedo
for ρ
VIS
between 0.05 and 0.07, for ρ
NIR,up
between 0.27 and 0.28, and for χ
up
between 0.6
and 0.9 (Figure 2.2). This range of χ
up
characterizes a predominance of upper canopy
elements (leaves and stems) with low inclination angle with respect to the horizontal.
35
0.095
0.105
0.115
0.125
0.135
0.145
0.155
0.26 0.28 0.30 0.35
Albedo
obs
0.095
0.105
0.115
0.125
0.135
0.145
0.155
0.26 0.28 0.30 0.35
Albedo
obs
χ
up
= 0.86
ρ
NIR,up
ρ
NIR,up
χ
up
= -1
χ
up
= -0.6
χ
up
= -0.2
χ
up
= 0.2
χ
up
= 0.6
χ
up
= 1
ρ
VIS,up
= 0.052
ρ
VIS,up
= 0.062
ρ
VIS,up
= 0.072
ρ
VIS,up
= 0.082
(a)
(b)
-20%
-10%
0%
10%
20%
30%
ρ
NIR,up
= 0.28
χ
up
χ
up
ρ
VIS,up
= 0.052
ρ
VIS,up
= 0.062
ρ
VIS,up
= 0.072
ρ
VIS,up
= 0.082
ρ
NIR,up
= 0.26
ρ
NIR,up
= 0.27
ρ
NIR,up
= 0.28
ρ
NIR,up
= 0.29
ρ
NIR,up
= 0.30
ρ
NIR,up
= 0.35
0.095
0.105
0.115
0.125
0.135
0.145
0.155
-1 -0.6 -0.2 0.2 0.6 1
Albedo
0.095
0.105
0.115
0.125
0.135
0.145
0.155
-1 -0.6 -0.2 0.2 0.6 1
Albedo
obs
obs
(d)(c)
-20%
-10%
0%
10%
20%
30%
-20%
-10%
0%
10%
20%
30%
-20%
-10%
0%
10%
20%
30%
Figure 2.2. Simulated albedo as a function of
ρ
NIR,up
and χ
up
, for the Cuieiras Biological
Reserve (K34).
Figure 2.3 shows the temporal variation of the observed and simulated albedos
considering different values of the optical parameter χ
up
for five selected days. In these
simulations, ρ
NIR,up
= 0.275 and ρ
VIS,up
= 0.062. The model best represents the diurnal
cycle with χ
up
= 0.86.
36
χ
up
= 0.86(Best fit)
Observed
χ
up
= -1 χ
up
= -0.6
χ
up
= -0.2
χ
up
= 0.2
χ
up
= 0.6
χ
up
= 1
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
27/8/1999 29/8/1999 15/9/1999 19/9/1999 21/9/1999
Date
Albedo
Figure 2.3. Temporal variation of the albedo observed in the Cuieiras Biological Reserve
(K34) and the albedo simulated by the IBIS model, according to the upper
canopy element orientation parameter.
The sensitivity of the root mean square error (RMSE) between simulated and
observed albedos as a function of the variation of the canopy optical parameters is shown
in Figure 2.4. This analysis indicates that the combination of the canopy optical parameters
that minimize the RMSE are χ
up
= 0.86, ρ
VIS,up
= 0.062, and ρ
NIR,up
= 0.275, confirming the
results obtained in Figures 2.2 and 2.3. Again, these parameters were obtained assuming
that ρ
VIS,lo
= 0.062, ρ
NIR,lo
= 0.60 and χ
lo
= 0.10, L = 6.175, S = 0.025, f
up
= 0.975 and f
lo
=
0.5.
37
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.26 0.28 0.30 0.35
RMSE
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.26 0.28 0.30 0.35
RMSE
ρ
NIR,up
ρ
NIR,up
ρ
VIS,up
= 0.052
ρ
VIS,up
= 0.062
ρ
VIS,up
= 0.072
ρ
VIS,up
= 0.082
χ
up
= -1
χ
up
= -0.6
χ
up
= -0.2
χ
up
= 0.2
χ
up
= 0.6
χ
up
= 1
(a)
(b) χ
up
= 0.86
ρ
NIR,up
= 0.28
0.010
0.015
0.020
0.025
0.030
0.035
0.040
-1 -0.6 -0.2 0.2 0.6 1
χ
up
RMSE
ρ
VIS,up
= 0.052
ρ
VIS,up
= 0.062
ρ
VIS,up
= 0.072
ρ
VIS,up
= 0.082
(d)
0.010
0.015
0.020
0.025
0.030
0.035
0.040
-1 -0.6 -0.2 0.2 0.6 1
χ
up
RMSE
ρ
NIR,up
= 0.26
ρ
NIR,up
= 0.27
ρ
NIR,up
= 0.28
ρ
NIR,up
= 0.29
ρ
NIR,up
= 0.30
ρ
NIR,up
= 0.35
(c)
Figure 2.4. Root Mean Square Error (RMSE) between the observed and simulated albedo
as a function of the canopy optical parameters
ρ
NIR,up
and
χ
up
, for the Cuieiras
Biological Reserve (K34).
The results indicate that the albedo simulated by IBIS is mostly sensitive to the
parameters χ
up
and ρ
NIR,up
. Although Bonan (1996) and Oleson et al. (2004) reported that
the good fit is obtained for χ
up
in the range -0.4 to 0.6, in this study, the minimization of
the mean relative error and the RMSE is achieved for χ
up
=
0.86. This parameter reduces
significantly the effect of the zenith angle on the albedo during sunrise and sunset (Figure
2.3). Lower daily amplitudes of surface albedo are expected when the canopy architectural
parameter, χ
up
, describes the upper canopy elements as mainly horizontally distributed, as
38
shown in Figure 2.3. For χ
up
= 0.6, the daily amplitude of the surface albedo is
overestimated, while for χ
up
= 1.0 it has a flat daily profile.
2.4. SUMMARY AND CONCLUSIONS
This study evaluated the sensitivity of the surface albedo simulated by IBIS to a set
of tropical rainforest canopy optical parameters. The results are evaluated against albedo
measurements taken above the K34 site at the Cuieiras Biological Reserve. Sensitivity
analysis indicates a strong response to the parameters χ
up
and
upNIR,
ρ
, a smaller sensitivity
to ρ
VIS,up
and no sensitivity at all to the lower canopy parameters χ
lo
,
loVIS,
ρ
and
loNIR,
ρ
,
which is consistent with the canopy structure. The combination of parameters that
minimize the RMSE and mean relative error RMSE are χ
up
= 0.86, ρ
VIS,up
= 0.062, and
ρ
NIR,up
= 0.275. From the analysis of Figures 2.2, 2.3 and 2.4, however, it seems reasonable
to conclude that values of χ
up
in the range of 0.8 to 0.9, of ρ
VIS,up
in the range of 0.05 to
0.07 and ρ
NIR,up
in the range of 0.26 to 0.28 all yield results that provide the better values of
RMSE, with little variation among them.
The successful simulations of tropical rainforest albedo by IBIS indicate its
potential to simulate the canopy radiative transfer for narrow spectral bands, for close
comparison with remote sensing products. Additional parameterizations of the canopy
architectural and optical parameters according to the plant species, soil types, plant
phenology, leaf water content, and soil surface wetness may improve considerably the
scope of such modeling exercises, building a solid basis for stronger interactions among
field observations, climate models and remote sensing products.
39
2.5. NOMENCLATURE
α
albedo of a vegetation canopy layer
μ
cosine of the Sun zenith angle
α
d
directional hemispherical reflectance
d
in
I
incident irradiance field that is purely collimated
α
i
indirect hemispherical reflectance
i
in
I
incident irradiance field that is purely isotropic
α
g
soil albedo
α
g-lo
albedo of the combined lower canopy-ground system
α
lo-up
albedo of the combined upper canopy-lower canopy-soil system
χ
up
upper canopy element orientation
χ
lo
lower canopy element orientation
f
lo
fraction of ground area covered by lower canopy
f
up
fraction of ground area covered by upper canopy
ρ
Λ
upper and lower canopy leaf reflectance by visible (VIS) and near-infrared
(NIR) spectral band Λ
ρ
VIS,up
upper canopy leaf reflectance by visible (VIS) spectral band Λ
ρ
VIS,lo
lower canopy leaf reflectance by visible (VIS) spectral band Λ
ρ
NIR,up
upper canopy leaf reflectance by near-infrared (NIR) spectral band Λ
ρ
NIR,lo
lower canopy leaf reflectance by near-infrared (NIR) spectral band Λ
leaf
Λ
ρ
leaf reflectances
stem
Λ
ρ
stem reflectances
τ
Λ
upper and lower canopy leaf transmittance by visible (VIS) and near-infrared
(NIR) spectral band Λ
τ
VIS,up
upper canopy leaf transmittance by visible (VIS) spectral band Λ
τ
VIS,lo
lower canopy leaf transmittance by visible (VIS) spectral band Λ
τ
NIR,up
upper canopy leaf transmittance by near-infrared (NIR) spectral band Λ
τ
NIR,lo
lower canopy leaf transmittance by near-infrared (NIR) spectral band Λ
leaf
Λ
τ
leaf transmittances
stem
Λ
τ
stem transmittances
I
↑
upward diffuse radiative fluxes per unit incident flux
I
↓
downward diffuse radiative fluxes per unit incident flux
K
optical depth of direct beam per unit leaf and stem area
G(µ) relative projected area of leaf and stem elements in the direction cos
-1
µ
μ
average inverse diffuse optical depth per unit leaf and stem area
ω
Λ
scattering coefficient
β
i
upscatter parameters for indirect (diffuse) beam radiation
β
d
upscatter parameters for direct beam radiation
L
single-sided leaf area index
S single-sided stem area index
d
g
Λ,
α
direct beam ground albedo for each spectral band Λ
i
g
Λ,
α
diffuse ground albedo for each spectral band Λ
d
Λ
α
single scattering albedo of direct fluxes
i
Λ
α
single scattering albedo of indirect fluxes
40
χ
h
leaf orientation for horizontal leaves
χ
v
leaf orientation for vertical leaves
b work variable
h1 work variable
h work variable
d1 work variable
41
CHAPTER 3
SIMULATIONS OF TROPICAL RAINFOREST ALBEDO: IS CANOPY
WETNESS IMPORTANT?
3.1. INTRODUCTION
Surface albedo is the main factor that affects the land radiation balance, not only
controlling the amount of solar energy available for heating the ground and lower
atmosphere and for evaporating water (Rowe, 1991), but also affecting the atmospheric
circulation and climate. Particularly in the tropics, where the solar radiation balance is
stronger, changes in surface albedo have been found to influence the regional climate, as
tropical deforestation studies demonstrate (Nobre et al., 1991; Dirmeyer and Shukla, 1994;
Costa and Foley, 2000). An accurate simulation of the solar radiation balance is also
important. For example, Berbet and Costa (2003) demonstrated that an uncertainty of 10 W
⋅ m
-2
in the seasonal solar radiation balance translates into an uncertainty of 30 mm/month
in the simulated rainfall, a very significant uncertainty, especially in the dry season in that
42
region. In this context, a good representation of albedo by climate models is essential to
correctly address the tropical deforestation climate change problem.
Berbet and Costa (2003), however, also verified that even a complex, state-of-the-
art; land surface scheme coupled to a climate model is unable to correctly reproduce details
of seasonal variability of albedo of tropical rainforests, although it reproduces well the
annual mean and some aspects of the seasonal variability. This indicates that there is still
much to be learned – and incorporated into models – about the sources of variation of
albedo at both hourly and monthly time scales.
At the monthly time scale, Culf et al. (1995) analyzed the surface albedo at forest
sites in Amazonia. The authors reported that albedo seasonality at these sites is not related
to the variation of solar elevation angle or to cloudiness, but suggested a relationship to soil
moisture. Although soil moisture affects ground albedo, most likely the changes in forest
albedo are related to soil moisture-correlated variables: smaller soil exposure, darker leaves
(associated with the leaf water potential) and higher canopy wetness (Berbet and Costa,
2003). Canopy wetness, in particular, is a strong candidate for changing canopy albedo,
because reflectance of liquid water varies, depending on the wavelength, between 0 and
5%, much lower than the 12-13% usually measured above tropical rainforest canopies.
Hence the presence of liquid water on the canopy increases the absorption of solar
radiation, reducing the canopy overall albedo.
Here, I investigated the role of canopy wetness in the simulated albedo of a
tropical rainforest. In this study, I run simulations using three versions of the land
surface/ecosystem model IBIS: the standard version using the original calibration used by
Delire and Foley (1999), the same version recalibrated to fit the tropical rainforests albedo
data, and a modified version that incorporates the effects of canopy wetness on calculated
surface albedo. The next section describes the sites and instrumentation that measured the
43
albedo data used here to validate the model simulations. Sections 3 and 4 describe the IBIS
model, the modifications to incorporate canopy wetness and the experiment design, while
Section 5 presents and discusses the simulation results.
3.2. MATERIALS AND METHODS
3.2.1. Sites, instrumentation and data
Field data used in this study were measured at three sites in the Amazon during the
ABRACOS (Anglo Brazilian Amazonian Climate Observation Study) and LBA (Large-
Scale Biosphere-Atmosphere Experiment in Amazonia) projects. Ducke Reserve (02º 57’S,
59º 57' W, 80 m above sea level) is an area of protected primary forest with mean height
forest canopy of 35 m, with some trees reaching up 40 m. The forest in Reserva Ducke is
composed by a large variety of tree species. The tallest species in the area around the tower
are
Piptadenia suaveolens Miq., Licania micrantha Miq., Bocoa viridiflora (Ducke) Cowan
and so on, with more details in Shuttleworth et al. (1984) and in Roberts et al. (1990).
Cuieiras Biological Reserve (2º 35’S, 60º 07’W, 90 m above sea level) (K34) is an INPA
(Instituto Nacional de Pesquisas da Amazônia) protected forest reserve, about 60 km north
of Manaus, embedded in a vast area of pristine rainforest. The natural vegetation and
topography of this site are representative of much of central Amazonia. These sites are
surrounded by undisturbed forest for at least 5 km. Jaru Reserve (10º 05'S, 61º 55'W, at
120 m above sea level) is an IBAMA (Instituto Brasileiro do Meio Ambiente e dos
Recursos Naturais Renováveis) forest reserve and is located about 80 km north of Ji-
Paraná. The mean height of the forest canopy is 33 m. The tallest tree species in the area
immediately surrounding the tower are
Cedrella odorata, Dioclea cf bicolor Bth.,
Strychnos amazonicus Krukoff (McWilliam et al., 1996; Roberts et al., 1996) (Figure 3.1).
44
Figure 3.1. Orientation map.
At the Cuieiras Reserve, a piranometer (Kipp & Zonen, Delft, Netherlands),
connected on a datalogger model (21X, Campbell Scientific) measured incident and
reflected solar radiation each minute, storing the averages every 20 minutes. For the
remaining sites, the incident and reflected solar radiation were measured using two
solarimeters (Kipp and Zonen, Delft, the Netherlands). These instruments are part of an
automatic weather station (Didcot Instruments, Abingdon, UK) connected to a datalogger
(CR10, Campbell Scientific, Shepshed, UK) and hourly-averaged data were recorded.
I used incident and reflected solar radiation and precipitation data, collected from
June 1999 to September 2000 at Cuieiras Reserve, from January to December of 1995 at
Ducke Reserve and from January to December of 1993 at Jaru Reserve. Data used in this
study are available on-line through www.cptec.inpe.br/abracos/available.html and
http://lba.cptec.inpe.br/beija-flor.
45
3.2.2. Description of the IBIS model
To simulate the diurnal albedo of an Amazon tropical forest, I used the Integrated
Biosphere Simulator–IBIS (Foley et al., 1996). It includes representations of land surface
processes, like energy, water and momentum exchange between the soil-vegetation-
atmosphere system, canopy physiology, vegetation phenology, vegetation dynamics, and
terrestrial carbon balance. Originally, IBIS was globally calibrated (more details in Delire
and Foley, 1999; Kucharik et al., 2000), and since then, it has been used in several studies
of the biosphere-atmosphere interactions in Amazonia (Costa and Foley, 2000; Botta and
Foley, 2002; Foley et al., 2002; Berbet and Costa, 2003).
One of the processes simulated by IBIS, of main interest here, is the exchange of
solar radiation between the soil-vegetation-atmosphere system. Solar radiation transfer is
calculated following the two-stream approximation, with separated calculations for direct
and diffuse radiation in both visible and near-infrared bands. The canopy radiative transfer
code of IBIS is standard in the literature (Norman and Jarvis, 1975; Sellers, 1985; Sellers
et al., 1986; Pollard and Thompson, 1995 – Appendix A; Bonan, 1996; Oleson et al.,
2004). A detailed description of the radiation transfer and albedo algorithms and model
sensitivity to several parameters is presented by Yanagi and Costa (2006). In this study, I
modified the IBIS canopy radiation transfer code to incorporate the effects of canopy
wetness on the vegetation reflectance. Although IBIS already calculates the canopy
wetness, in this work the parameters ω (scattering coefficient, eq. 3 in Sellers, 1985; eq. 11
in Sellers et al., 1986; eq. 3.5 in Oleson et al., 2004), β (upscatter parameter for diffuse
radiation, eq. 3 in Sellers, 1985; eq. 11 in Sellers et al., 1986; eq. 3.6 in Oleson et al., 2004)
and β
o
(upscatter parameter for direct radiation, eq. 4 in Sellers, 1985; eq. 12 in Sellers et
al., 1986; eq. 3.7 in Oleson et al., 2004) were modified to include the radiative effects of
canopy wetness, according to Equations 1 to 4:
46
ω = ω
dry
·(1 – f
wet
) +
ω
water
·f
water
+
ω
snow
·f
snow
(1)
β = ω
dry
·
β
dry
(1 – f
wet
) +
ω
water
·
β
water
·f
water
+
ω
snow
·
β
snow
·f
snow
/
ω
(2)
β
ο
= ω
dry
·
β
ο
dry
(1 – f
wet
) +
ω
water
·
β
ο
water
·f
water
+
ω
snow
·
β
ο
snow
·f
snow
/
ω
(3)
ω
water
=
ν
·
ω
dry
(4)
where f
wet
is the total wet (water and snow) fraction of the canopy (f
wet
= f
water
+f
snow
), f
water
is
the fraction of the canopy wet by liquid water and
f
snow
is the fraction of the canopy wet by
snow. The superscripts
dry, water and snow denote dry, wet by water and wet by snow
canopies, respectively.
ν
is the ratio of the scattering coefficients of the canopy surfaces
wet by water and dry canopy surfaces, applied individually to leaves and stems.
3.2.3. Experiment design
In this study, I conducted three simulations of the diurnal surface albedo for each
of the three Amazon forest sites using the off-line IBIS version, as follows:
1)
DF99: for reference to previous studies, this simulation uses the set of optical
parameters used by Delire and Foley (1999);
2)
DC
i
, the dry-canopy (control) simulations: use the original code, without modifications
to incorporate the effects of wetness on canopy reflectance. It uses a set of leaf optical
parameters chosen to better fit the simulated results to the experimental data of the sites
studied, minimizing the RMSE between the observed and simulated albedos. The
subscript
i may be equal to M for the Manaus-nearby sites (Ducke and Cuieiras
Reserves), or
J for Jaru Reserve.
3)
WC
i
, the wet-canopy simulations: similar to DC
i
, but including the modifications
described in Equations (1) through (4).
The terms dry-canopy (DC) and wet-canopy (WC), when referring to versions of the
IBIS code, denote only the status of the canopy during the radiative transfer calculations. It
47
should be noted that both versions simulate the interception of water by the canopy, and
the evaporation and dripping of the canopy-stored water.
To calibrate the DC simulations, initially I do a sensitivity analysis of the simulated
albedo to several canopy optical parameters. The most sensitive parameters are the upper
and lower canopy leaf orientation (
χ
leaf-up
and
χ
leaf-lo
, respectively), and upper and lower
canopy visible and near-infrared (NIR) leaf reflectance (
Leaf
upVIS
−
α
,
Leaf
loVIS −
α
,
Leaf
upNIR
−
α
and
Leaf
loNIR
−
α , respectively). The model is run several times with different combinations of the
parameters above. Table 3.1 shows, for the DC simulations, the parameter combinations
that provide the best adjustment of the diurnal albedo without considering the effect of
canopy wetness. Next, this set of parameters is added to the WC version of the code, and
the parameters
f
wetmax
(maximum fraction of water cover on two-sided leaf),
τ
drip
(decay
time for intercepted liquid drip off) and ν (ratio of the scattering coefficients of the canopy
surfaces wet by water and dry canopy surfaces, applied individually to leaves and stems)
are calibrated.
48
Table 3.1. Optical parameters used by the DF99 calibration, and by the new calibrations using the dry-canopy (DC
i
) and wet-canopy (WC
i
)
versions of the model, where
i is equal to M for the Manaus-nearby sites (Ducke and Cuieiras Reserves) or J for the Jaru Reserve.
χ
leaf-up
is the
upper canopy leaf orientation,
χ
leaf-lo
is the lower canopy leaf orientation,
Leaf
loVIS −
α is the lower canopy visible leaf reflectance,
Leaf
upVIS
−
α
is the upper
canopy visible leaf reflectance,
Leaf
loNIR−
α
is the lower canopy NIR leaf reflectance,
Leaf
upNIR
−
α
is the upper canopy NIR leaf reflectance, f
wetmax
is
maximum fraction of water cover on two-sided leaf,
τ
drip
is the decay time for intercepted liquid drip off and
ν
is the ratio of the scattering
coefficients of the canopy surfaces wet by water and dry canopy surfaces, applied individually to leaves and stems. All values are dimensionless,
except for
τ
drip
, which is in seconds.
upleaf −
χ
loleaf −
χ
Leaf
loVIS
−
α
Leaf
upVIS
−
α
Leaf
loNIR
−
α
Leaf
upNIR
−
α
f
wetmax
τ
drip
ν
All Reserves DF99 0.00 -0.50 0.062 0.062 0.60 0.40 0.25 7200 s –
DC
M
0.86 0.10 0.062 0.062 0.60 0.28 0.25 7200 s –
Manaus
(Ducke,
Cuieiras)
WC
M
0.86 0.10 0.062 0.062 0.60 0.28 0.80 3600 s 0.10
DC
J
0.86 0.10 0.082 0.082 0.60 0.30 0.25 7200 s –
Jaru
WC
J
0.86 0.10 0.082 0.082 0.60 0.30 0.80 3600 s 0.40
49
3.3. RESULTS AND DISCUSSION
The three IBIS versions are run for each of the three study sites. Figures from 3.2
to 3.4 show the diurnal profile of the surface albedo for Cuieiras, Ducke and Jaru Reserves,
respectively, for selected days of the year. To facilitate the interpretation, the days selected
represent either no-rain days – charts on the left side (a, b, c) of the figures – or days with a
single daytime rainfall event – right side (d, e, f) of the figures. Figures from 3.2 and 3.3
show results for the Manaus (
M) set of parameters, while Figure 3.4 shows results for the
Jaru (
J) set of parameters, according to Table 3.1.
The DF99 simulations overestimate the surface albedo, particularly when zenith
angle is high (Figures 3.2 to 3.4). The albedo simulated by the calibrated model (DC
M
and
DC
J
) fit better to the observed data in days without precipitation occurrence (left side of
Figures 3.2 to 3.4), but the DC version of the model does not represent the observed albedo
drop during precipitation events (right side of Figures 3.2 to 3.4).
WC simulations, however, show that the modified version of IBIS is able to
reproduce the considerable decrease in surface albedo during precipitation hours (right side
of Figures 3.2 to 3.4). This reduction in the surface albedo is consistent with an increase of
the absorption of the incident solar radiation by the liquid water deposited on the leaves,
increasing solar radiation absortance and reducing solar radiation reflectance.
Table 3.2 shows the statistics of observed and simulated surface albedo for the
precipitation hours and for the entire time series at three Amazon rainforest sites (Cuieiras,
Ducke and Jaru reserves), for the simulations based on the DF99 calibration, for the new
calibration using the dry-canopy (DC) and wet-canopy (WC) versions of the model.
50
Table 3.2. Statistics of observed and simulated surface albedo for the entire time series and
for precipitation hours at three Amazon rainforest sites (Cuieiras, Ducke and Jaru
reserves), for the simulations based on the DF99 calibration, for the new calibration using
the dry-canopy (DC) and wet-canopy (WC) versions of the model.
X
is the average
albedo, ε is the mean relative error and RMSE is the root mean square error.
Entire time series Precipitation hours
X
ε (%)
RMSE
X
ε (%)
RMSE
DF99 0.146 23.44 0.0398 0.139 41.58 0.0550
DC
M
0.121 2.14 0.0219 0.120 22.13 0.0365
WC
M
0.118 -0.03 0.0210 0.105 6.11 0.0302
Ducke
Reserve
Observed 0.118 – – 0.099 – –
DF99 0.146 24.11 0.0384 0.142 40.79 0.0521
DC
M
0.121 2.58 0.0169 0.120 19.16 0.0320
WC
M
0.117 -0.73 0.0142 0.104 2.91 0.0234
Cuieiras
Reserve
Observed 0.118 – – 0.101 – –
DF99 0.146 12.26 0.0332 0.145 19.57 0.0387
DC
M
0.121 -7.15 0.0232 0.121 -0.35 0.0222
WC
M
0.117 -9.68 0.0243 0.108 -10.93 0.0247
Using
set of optical
parameters
calibrated for
the sites near
to Manaus
Jaru
Observed 0.130 – – 0.121 – –
DF99 0.146 12.26 0.0332 0.145 19.57 0.0387
DC
J
0.133 1.98 0.0215 0.132 9.43 0.0250
WC
J
0.129 -0.50 0.0209 0.121 0. 40 0.0207
Using
set of optical
parameters
calibrated for
Jaru site
Jaru
Observed 0.130 – – 0.121 – –
Mean relative error (ε) and root mean square error (RMSE) are defined according to
Equations 5 and 6:
(
)
∑
=
−
=
n
i
o
i
o
i
s
i
X
XX
n
1
1
ε
(5)
51
()
n
XX
RMSE
n
i
o
i
s
i
∑
=
−
=
1
2
(6)
where X
s
and X
o
are simulated and observed albedo and n is the number of data points.
Given my interest here in the effects of canopy wetness on the simulated rainforest albedo,
I analyzed the “precipitation hours” data separately from the entire data (Table 3.2). For
the entire time series, for the two sites near Manaus, the new calibration (DC
M
) represents
a considerable improvement over the DF99 calibration, lowering the mean relative error by
an order of magnitude, and the RMSE by nearly half. The inclusion of the effects of
canopy wetness on the radiative transfer code further reduces ε and RMSE, consistent with
the results shown in Figures 3.2 to 3.4. Looking for a single parameterization for
Amazonia, I also tested the Manaus parameters at the Jaru site. However, it turns out that
the average observed albedo of the Jaru site is much higher than at the Manaus sites, and
the parameters used for Manaus are not suitable to Jaru. I then selected a new set of
parameters for the Jaru site (DC
J
). Results for Jaru site, using a site-specific calibration,
show an improvement similar to the one obtained at the Manaus sites.
Although the WC simulation shows an improvement in the simulation statistics for
the entire time series, the true effect of it is best seen in the analysis of data during the
precipitation hours. Initially, I can verify that the DF99 and DC simulation results during
precipitation hours are much worse than the equivalent result for the entire time series,
which by itself indicates that a precipitation-related process is missing in the model. In
addition, the WC simulations for the precipitation hours improve considerably the results
when compared to the control (DC), with ε dropping by an order of magnitude and RMSE
decreasing by about one-fifth.
52
8 13 18
0.00
0.05
0.10
0.15
0.20
0.25
Albedo
0
5
10
Precipitation (mm/h)
(a) 7/7/1999 (d) 7/14/1999
0.00
0.05
0.10
0.15
0.20
0.25
Albedo
0
5
10
Precipitation (mm/h)
(b) 9/15/1999 (e) 8/6/1999
0.00
0.05
0.10
0.15
0.20
0.25
Local time
Albedo
0
5
10
Precipitation (mm/h)
Observed
DF99 DC
M
WC
M
Precipitation
(c) 3/16/2000 (f) 5/22/2000
8 13 18
8 13 18 8 13 18
8 13 18 8 13 18
Figure 3.2. Diurnal variation of the simulated and observed surface albedo in the Cuieiras
Reserve for selected days.
53
8 13 18
0.00
0.05
0.10
0.15
0.20
0.25
Albedo
0
5
10
Precipitation (mm/h)
(a) 1/20/1995 (d) 5/7/1995
0.00
0.05
0.10
0.15
0.20
0.25
Albedo
0
5
10
P
r
ecipitation (mm/h)
(b) 1/30/1995 (e) 5/11/1995
0.00
0.05
0.10
0.15
0.20
0.25
Local time
Albedo
0
5
10
Precipitation (mm/h)
Observed
DF99 DC
M
WC
M
Precipitation
(c) 2/11/1995 (f) 12/28/1995
8 13 18
8 13 18 8 13 18
8 13 18 8 13 18
Figure 3.3. Diurnal variation of the simulated and observed surface albedo in the Ducke
Reserve for selected days.
54
8 13 18
0.00
0.05
0.10
0.15
0.20
0.25
Albedo
0
5
10
Precipitation (mm/h)
(a) 1/3/1993 (d) 1/22/1993
0.00
0.05
0.10
0.15
0.20
0.25
Albedo
0
5
10
Precipitation (mm/h)
(b) 7/17/1993
(e) 1/31/1993
0.00
0.05
0.10
0.15
0.20
0.25
Local time
Albedo
0
5
10
Precipitation(mm/h)
Observed
DF99 DC
J
WC
J
Precipitation
(c) 12/12/1993 (f) 12/29/1993
8 13 18
8 13 18 8 13 18
8 13 18 8 13 18
Figure 3.4. Diurnal variation of the simulated and observed surface albedo in the Jaru
Reserve for selected days.
55
These results demonstrate that, at the hourly time scale, the incorporation of canopy
wetness on the radiative transfer calculations improves substantially the simulation results,
in particular for the times when the canopy is wet, but also brings an improvement to the
simulation of the entire period.
I also compare the model results at the monthly time scale for the three sites studied
(Figure 3.5). In all cases, the WC albedo is smaller than the DC albedo, but the changes
introduced do not substantially modify the simulated albedo. As seen in Figures 3.2 to 3.4,
because of evaporation and dripping, the effect of canopy wetness on the surface albedo is
restricted to the duration of a rainfall event plus one or two hours. Although Amazonia is
one of the rainiest climates on the Earth, the frequency of rainfall events (7.1% at Ducke,
14.2% at Cuieiras, and 17.0% at Jaru) is relatively low for a more significant effect of
canopy wetness on albedo seasonality.
An exception is observed at the Ducke Reserve (Figure 3.5b), where a more
pronounced drop in the WC-simulated albedo is observed in April. Even in this extreme
case, the simulated decrease in monthly albedo accounts for less than half of the observed
change, which let me conclude that canopy wetness alone is not sufficient to represent
correctly the seasonal variability of the albedo for a tropical rainforest.
56
Precipitation
(mm/month)
Frequency (%)
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0.15
0.16
JFMAMJJASOND
Albedo
0
100
200
300
400
Precipitation
Observed
DF99
DC
J
WC
J
(c)
0
10
20
30
40
Frequency
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0.15
0.16
JFMAMJJASOND
Albedo
0
100
200
300
400
(a)
0
10
20
30
40
Precipitation
(mm/month)
Frequency (%)
Precipitation
Observed
DF99
DC
M
WC
M
Frequency
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0.15
0.16
JFMAMJJASOND
Albedo
0
100
200
300
400
(b)
Precipitation
(mm/month)
0
10
20
30
40
Frequency (%)
Precipitation
Observed
DF99
DC
M
WC
M
Frequency
Figure 3.5. Monthly profile of the observed and simulated surface albedo, monthly
precipitation and frequency of rainfall events, at three Amazon rainforest
sites: (a) Cuieiras Reserve, from June 1999 to September 2000, (b) Ducke
Reserve, from January to December of 1995 and (c) Jaru Reserve, from
January to December of 1993.
57
3.4. SUMMARY AND CONCLUSIONS
The goal of this work was to study the effect of canopy wetness on the simulated
albedo of a tropical rainforest. Simulations were run using three versions of the IBIS
model: the standard version, the same version recalibrated to fit the tropical rainforests
albedo data, and a modified version that incorporates the effects of canopy wetness on
surface albedo, for three Amazon forest study sites in both hourly and monthly time scales.
The incorporation of canopy wetness on the radiative transfer calculations improves
the simulation results at the hourly time scale, reproducing the observed decrease in
surface albedo during precipitation hours, when the canopy is wet. Although the canopy
wetness has an important effect, this effect is restricted to the times when the canopy is
actually wet, a relatively short period of time at the monthly or longer time scales.
Therefore, the changes introduced are not sufficient to substantially improve the
representation of albedo seasonality.
While these results exclude the role of canopy wetness as a main source of
seasonal variability of tropical rainforests albedo, this study narrows the choice of sources
of albedo seasonal variation. Following the discussion of Culf et al. (1995) and Berbet and
Costa (2003) on the subject, I recommend that future studies investigate the role of
photoinhibition and leaf water potential on the seasonality of the tropical rainforest albedo.
The clear definition of these roles, and their incorporation into climate models, will
eventually allow us to do much more detailed studies of the climatic effects of tropical
deforestation.
58
3.5. NOMENCLATURE
ω
scattering coefficient
β
upscatter parameter for diffuse radiation
β
o
upscatter parameter for direct radiation
f
wet
total wet (water and snow) fraction of the canopy
f
water
fraction of the canopy wet by liquid water
f
snow
fraction of the canopy wet by snow
ν
ratio of the scattering coefficients of the canopy surfaces wet by water and dry
the canopy surfaces
χ
leaf-up
upper canopy leaf orientation
χ
leaf-lo
lower canopy leaf orientation
Leaf
upVIS
−
α
upper canopy visible leaf reflectance
Leaf
loVIS
−
α
lower canopy visible leaf reflectance
Leaf
upNIR
−
α
upper canopy near-infrared (NIR) leaf reflectance
Leaf
loNIR
−
α
lower canopy near-infrared (NIR) leaf reflectance
f
wetmax
maximum fraction of water cover on two-sided leaf
τ
drip
decay time for intercepted liquid drip off
X
s
simulated albedo
X
o
observed albedo
ε
mean relative error
59
CHAPTER 4
COMPARISON OF SEASONAL AND SPATIAL VARIATIONS OF ALBEDO
ESTIMATED BY A CLIMATE MODEL AND ALBEDO DERIVED FROM
REMOTE SENSING DATA FOR THE AMAZON TROPICAL RAINFOREST
4.1. INTRODUCTION
Albedo is the ratio of upwelling radiant energy relative to the downwelling
irradiance incident upon a surface. Land surface albedo is a key land physical parameter in
characterizing the surface energy budget and has frequently been considered in studies of
global and regional climate. General Circulation Model (GCM) studies have demonstrated
that the climate presents a strong sensitivity to surface albedo changes, originated by
natural action and anthropogenic land use changes, such as desertification (Charney et al.,
1977; Xue et al., 1990) and tropical deforestation (Dickinson and Henderson-Sellers, 1988;
Hahmann and Dickinson, 1997; Berbet and Costa, 2003). Thus, the better representation of
60
surface albedo in climate models will enhance the accuracy of climate simulation and
prediction.
Albedo simulated by climate models has been evaluated with field observations and
satellite measurements. Field data are fundamental to validate model results, mainly
because they represent more realistically a specific site studied. However, they present
operational difficulties, are expensive to obtain, and only provide sparse and eventual
measurements.
On the other hand, remote sensing is an important tool to obtain coherent temporal
and spatial variations of surface characteristics such as albedo (Lewis et al., 1999), thus
aiding the validation and improvement of model parameterization. A number of land
surface albedo products have been produced based on several remote sensing devices,
including the polar orbiting NOAA (National Oceanic and Atmospheric Administration)
with AVHRR (Advanced Very High Resolution Radiometer) sensor, ERBE (Earth
Radiation Budget Experiment), MODIS (Moderate Resolution Imaging
Spectroradiometer), and the geostationary satellites GOES and METEOSAT. In many
cases, data from different remote sensors were combined to generate a single product.
This chapter presents a comparison of simulated albedo computed by the NCAR
Community Climate Model (CCM3) coupled to the Integrated Biosphere Simulator (IBIS)
with derived albedo measurements from six remote sensing products. A comparison against
field data collected at three study sites is also performed. Section 2 describes the albedo
datasets analyzed. Section 3 presents and discusses the results and conclusions are
summarized in Section 4.
61
4.2. ALBEDO DATA
4.2.1. Land surface albedo simulations
The land surface albedo simulations were produced by the National Center for
Atmospheric Research (NCAR) Community Climate Model version 3 (CCM3, Kiehl et al.,
1998) coupled with an updated version of the Integrated Biosphere Simulator (IBIS) of
Foley et al. (1996) and Kucharik et al. (2000). I refer to this coupled model as CCM3–IBIS
(Delire et al., 2002). CCM3 is a spectral atmosphere model with spatial resolution of T42
(the spectral representation of the horizontal fields is truncated at the 42
nd
wavenumber
using a triangular truncation; horizontal fields are converted to a 2.81° x 2.81° grid) and
has 18 vertical levels and operates with a 20-min time step (Kiehl et al., 1998). IBIS is an
integrated model of land surface processes, vegetation dynamics and carbon cycle. IBIS
has two vegetation layers (i.e., trees and short vegetation) and a variable number of soil
layers. In IBIS, the exchange of solar radiation between the soil-vegetation-atmosphere
system is calculated following the standard two-stream approximation, with separate
calculations for direct and diffuse radiation in both visible (VIS) and near-infrared (NIR)
bands. A detailed description of the albedo algorithm and model sensitivity to several
parameters is presented by Yanagi and Costa (2006). In these simulations, IBIS operates
on the same T42 spatial grid as the CCM3 atmospheric model.
4.2.2. Remote sensing albedo
In this work, I used land surface albedo from six remote sensing products: Csiszar
and Gutman (1999) – CG99, ERBE, SRB-ISLSCP2, UMD (GOES/METEOSAT/AVHRR),
MODIS white-sky, and MODIS black-sky (Table 4.1).
62
Table 4.1. Main characteristics of the remote sensing albedo products.
Remote sensing product
and sensors involved
Spectral resolution
Spatial
resolution
Temporal
resolution
Used
period
Sky
condition
Average
albedo
Source
CG99 (AVHRR – Advanced
Very High-Resolution
Radiometer)
Channel 1
(0.58-0.68 μm, VIS)
and Channel 2
(0.72-1.1μm, NIR)
0.5ºx0.5º
1.1 km
monthly 1985-1991 Clear-sky
11.0 %
Csiszar and
Gutman,1999
ERBE (Earth Radiation
Budget Experiment)
Shortwave channel 1
(0.2-5.0 μm)
1ºx1º
35-50km
monthly 1986-1990 Clear-sky
15.5 %
Barkstrom, 1984
SRB-ISLSCP2
Shortwave channel
(0.3-5.0 μm)
1ºx1º monthly 1986-1995 All-sky 15.0 %
Stackhouse et al., 2000;
Stackhouse et al., 2003
UMD (GOES-METEOSAT-
AVHRR)
Shortwave channel
(0.3-5.0 μm)
0.5ºx0.5º
30 km
monthly 1990-1992 All-sky 15.0 % Pinker, 2002
White-sky 15.3 %
MODIS (Moderate
Resolution Imaging
Spectroradiometer). Albedo
product (MOD43B)
bb3 brodband
(0.3-5.0 μm)
0.5ºx0.5º
(1 km)
16-day 2000-2001
Black-sky 13.5 %
Lucht et al.,2000a;
Schaaf et al., 2002
63
AVHRR visible (0.58-0.68 μm, channel 1) and near-infrared (0.72-1.1 μm, channel 2)
reflectances at 0.5ºx0.5º of spatial resolution were used to calculate the Amazon surface
albedo. AVHRR is flown onboard the polar orbiter “NOAA” operational environmental
satellites series. AVHRR has no on-board calibration capability for the VIS and near-IR
channels used for generating the albedo product. Pre-launch calibration coefficients should
not be used because of the degradation of the instrument sensitivity. The channel 1 and 2
digital counts were converted to reflectance values using post-launch calibration coefficients,
using the formulae derived by Rao and Chen (1995, 1996). AVHRR uses top-of-the
atmosphere (TOA) narrow-to-broadband conversion equation developed by Hucek and
Jacobowitz (1995), and ERBE (Suttles et al., 1988) model to account for the bi-directional
effects. The atmospheric correction was done using the model by Li and Garand (1994) and
the normalization to overhead Sun illumination angle was done by the formula proposed by
Briegleb et al. (1986).
I uses the ERBE shortwave (0.2-5.0 μm) channel reflectance at 1ºx1º spatial
resolution obtained from re-gridded of original 2.5º x 2.5º resolution. ERBE sensor was
launched on the ERBS (Earth Radiation Budget Satellite). Ground calibration sources consist
of a reference black-body and an integrating sphere in a vacuum chamber. In flight, an
internal black-body, evacuated tungsten lamps, and observation of the Sun are used to check
the stability and precision of instruments. Hourly clear-sky radiative fluxes in a grid box are
calculated as the averaged fluxes of the clear pixels, which are then converted from radiance
through the ERBE angular-directional models (Suttles et al., 1988). Cloud-free scenes are
identified through an algorithm that adopts a maximum likelihood estimation technique
(Wielicki and Green, 1989).
Surface Radiative Budget (SRB) data for all-sky condition albedo, based on
ISLSCP2 (The International Satellite Land Surface Climatology Project) observations, were
64
calculated from shortwave (0.3-5.0 μm) reflectances at 1ºx1º spatial resolution. No
instruments were directly used in the generation of this NASA/GEWEX SRB dataset. All
inputs used were higher level products provided by other institutions, as example, all satellite
derived radiance and cloud data are from the ISCCP DX data set. Calibration information for
ISCCP is available in Brest et al. (1997). Conversion of both narrowband visible cloud and
clear radiances into broadband radiances is computed by the Pinker/Laszlo algorithm (Pinker
and Laszlo, 1992) and then an angular distribution model (ADM) is applied to convert this
radiance to a top-of-atmosphere (TOA) albedo. The TOA albedos for the all-sky conditions
are then matched using a radiative transfer model with the meteorological inputs. The
advantage of this technique is that the resulting surface fluxes are consistent with the TOA
albedos.
Albedos derived from the product GOES/METEOSAT/AVHRR (UMD) and
ancillary data, as available from the Global Energy and Water Cycle Experiment (GEWEX)
ISCCP DX, were calculated through the shortwave reflectance (0.3-5.0 μm) at 0.5ºx0.5º
spatial resolution. These data are also available from the NASA/GEWEX/ISCCP DX
observations, as produced in the Department of Meteorology, University of Maryland (UMD)
(Pinker, 2002) as part of the LBA (Large-Scale Biosphere-Atmosphere Experiment in
Amazonia) Project. Since the ISCCP cloud analysis uses only radiances from the spectral
channels common to all radiometers, namely ‘‘visible’’ (wavelength approximately 0.6 μm,
called VIS) and ‘‘window’’ infrared (wavelength approximately 1.1 μm, called IR), only
these radiances have been calibrated and normalized by ISCCP (Brest and Rossow 1992;
Rossow et al. 1992; Desormeaux et al. 1993; Rossow et al. 1996). Data are corrected for
Rayleigh scattering and daily variations of solar irradiance and ozone absorption as a function
of illumination and viewing geometry (Brest et al., 1997). Indeed, the ISCCP calibration
65
procedure differs from most in that instead of using one small selected site (e.g., a desert
target), it is used a wide variety of targets and the entire globe itself as a target.
MODIS MOD43B albedo product, white-sky (diffuse) and black-sky (direct at solar
noon), were determined from shortwave reflectance (0.3-5.0 μm, bb3 broadband) at 0.5ºx0.5º
spatial resolution. The MODIS instrument was launched on board NASA’s Terra platform.
MODIS have four main On-Board Calibrators (Blackbody - BB, Solar Diffuser - SD, Solar
Diffuser Stability Monitor - SDSM, and the Spectroradiometric Calibration Assembly - SRCA)
that generate various stimuli to provide radiometric, spectral and spatial calibration of the
MODIS instrument. The BB is the prime calibration source for the thermal bands located from
3.5 µm to 14.4 µm, while the SD provides a diffuse, solar-illuminated calibration source or the
visible, near infrared, and shortwave infrared bands (0.4 µm ≤ λ < 2.2 µm). The SDSM tracks
changes in the reflectance of the SD via reference to the Sun so that potential instrument
changes are not incorrectly attributed to changes in this calibration source. The SRCA is a
multifunction calibration instrument that provides in-flight spectral, radiometric and spatial
calibration. Two additional calibration techniques that MODIS uses are views of the moon and
deep space. The MODIS albedo algorithm adopts a semi-empirical, kernel-driven linear
Bidirectional Reflectance Distribution Function (BRDF) to characterize the anisotropy of the
global surface (Lucht et al., 2000a; Schaaf et al., 2002). The BRDF model relies on the
weighted sum of three parameters that are retrieved from the multi-spectral, multi-date, cloud-
free atmospherically-connected surface reflectances associated.
To plot the anomaly maps, I aggregated the half-degree and one-degree products to a
2.8º x 2.8º grid, compatible with the CCM3 output. When necessary, a linear interpolation
was used to fill the missing values for the ERBE, MODIS white-sky and MODIS black-sky
products.
66
4.2.3. Field albedo measurements
Field data used in this study were measured at three sites in the Amazon region during
the ABRACOS (Anglo Brazilian Amazonian Climate Observation Study) and LBA (Large-
Scale Biosphere-Atmosphere Experiment in Amazonia) projects. Ducke and Cuieiras (K34)
Reserves (02º 57’S, 59º 57' W and 02º 35’S, 60º 06' W) are INPA (Instituto Nacional de
Pesquisas da Amazônia) protected forest reserves, about 25 km north-east and 60 km north of
Manaus, respectively. Jaru Reserve (10º 05’S, 61º 55”W) is an IBAMA (Instituto Brasileiro
do Meio Ambiente e dos Recursos Naturais Renováveis) Forest Reserve and is located about
80 km north of Ji-Paraná in Rondônia State. At the Cuieiras Reserve a piranometer (Kipp &
Zonen, Delft, Netherlands), connected to a datalogger model (CR10, Campbell Scientific)
measured incident and reflected solar radiation each 30 minutes. For the remaining sites, the
incident and reflected solar radiation were measured using two solarimeters (Kipp and Zonen,
Delft, the Netherlands). These instruments are part of an automatic weather station (Didcot
Instruments, Abingdon, UK) connected to a datalogger (CR10, Campbell Scientific,
Shepshed, UK) and hourly-averaged data were recorded. I used incident and reflected solar
radiation data, collected from June 1999 to September 2000 at Cuieiras Reserve, from January
to December of 1995 at Ducke Reserve and from January to December of 1993 at Jaru
Reserve. Data used in this study are available on-line through
www.cptec.inpe.br/abracos/available.html and http://lba.cptec.inpe.br/beija-flor.
4.3. RESULTS AND DISCUSSION
Initially, IBIS was calibrated against data measured at Jaru Biological Reserve, with an
absolute error of - 0.002 and a root mean square error of 0.007.
Figure 4.1 shows, for the Amazon region, the spatial variability of the average annual
albedo (a) simulated by IBIS, and derived remotely sensed products: (b) CG99, (c) ERBE, (d)
67
SRB/ISLSCP2, (e) UMD, (f) MODIS Black-sky and (g) MODIS white-sky and the respective
differences: (h) IBIS–CG99, (i) IBIS–ERBE, (j) IBIS–SRB/ISLSCP2, (k) IBIS–UMD, (l)
IBIS–MODIS Black-sky and (m) IBIS–MODIS white-sky.
It can be seen through the anomaly maps (Figures 4.1h to 4.1n) that the remote sensing
albedo products that present best agreements related to CCM3 were MODIS-bsa (difference -
1) and the CG99 (difference +1.5), as showed in the Figures 1b and 1e, respectively. The other
products SRB/ISLSCP2, UMD, MODIS-wsa and ERBE present average anomalies of -2.5, -
2.5, -2.8 and -3.0, respectively. The smallest differences between CCM3 model and MODIS-
bsa that calculate wholly direct radiation probably occur because CCM3 may be giving higher
weighting to direct radiation rather than to diffuse radiation in the algorithm used to calculate
the albedo. In addition, the worst simulation of MODIS white- compared to black-sky albedos
may be explained by the overestimation the increase of albedo with solar zenith angle (Oleson
et al., 2003). The authors found 1% to 5% to VIS and 1% to 15% in the NIR, while in this
work MODIS white-sky albedo overestimated the black-sky albedo in 13% approximately.
Differences among the remotely sensed albedos from several products studied may be
attributed by the different spectral distribution of incoming radiation and the capacity of the
algorithms of each product to filter the confounding effects such as atmospheric scattering and
absorption anisotropy, inadequate spatial sampling and narrowband to broadband conversions
(Oleson et al., 2003). In addition, although all of these instruments have been carefully
calibrated, some differences among them are expected (Smith et al., 2006), including the
continuous degradation while they are in orbit (eg. degradation of spectral response of each
channel).
68
CCM3
60ºW80ºW 70ºW
50ºW
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
CG99
ERBE
ISLSCP2
UMD
MODIS-BSA
MODIS-WSA
-6 -3
03
6
difference
9
11 13 15 17
60ºW80ºW 70ºW
50ºW
9
11 13 15 17
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
60ºW80ºW 70ºW
50ºW
9
11 13 15 17
60ºW80ºW 70ºW
50ºW
9
11 13 15 17
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
60ºW80ºW 70ºW
50ºW
9
11 13 15 17
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
60ºW80ºW 70ºW
50ºW
9
11 13 15 17
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
60ºW80ºW 70ºW
50ºW
9
11 13 15 17
difference
difference
difference
difference
difference
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
5ºN
0ºN
5ºS
10ºS
15ºS
20ºS
60ºW80ºW 70ºW
50ºW
-6 -3
03
6
60ºW80ºW 70ºW
50ºW
-6 -3
03
6
60ºW80ºW 70ºW
50ºW
-6 -3
03
6
60ºW80ºW 70ºW
50ºW
-6 -3
03
6
60ºW80ºW 70ºW
50ºW
-6 -3
03
6
60ºW80ºW 70ºW
50ºW
Figure 4.1. Spatial variability of the surface albedo simulated by the (a) CCM3 model and
albedo products derived from six remote sensing for Amazon basin: (b) CG99, (c)
ERBE, (d) SRB/ISLSCP2, (e) UMD, (f) MODIS Black-sky and (g) MODIS white-
sky and it respective anomalies: (h) CCM3-CG99, (i) CCM3-ERBE, (j) CCM3-
SRB/ISLSCP2, (k) CCM3-UMD, (l) CCM3-MODIS Black-sky and (m) CCM3-
MODIS white-sky.
69
Figure 4.2 shows the seasonal profile of the CCM3 simulated albedo and albedo
products derived from six remote sensing (CG99, ERBE, SRB-ISLSCP2, UMD, MODIS
White-sky and black-sky) for the Amazon basin and also representation of field albedo data at
the Biological Reserves of Cuieiras (K34), Ducke and Jaru. Overall, an overestimation of the
albedo derived from the satellites products in relation to the CCM3 model was observed,
except for CG99 product that underestimated it. Brest et al. (1997) examining a complete 8-
year record of cloud and surface properties obtained with the first calibration reveals a
systematic “step-down” in monthly global mean VIS reflectance at each transition of polar
orbit from NOAA-7 to NOAA-9 between January and February 1985, and from NOAA-9 to
NOAA-11 between October and November 1988 (Klein and Hartman, 1993; Rossow and
Cairns, 1995). On the other hand, Csiszar and Gutman (1999) affirm that heavy aerosol
loading introduces errors for extremely low or high albedos. It can be observed during the end
of the dry-season, because the biomass burning aerosols. In addition, the initial monthly
product showed obvious aerosol signal in tropical evergreen regions; hence monthly values
were replaced by yearly minima in those areas. The amplitudes among the simulated albedo
(CCM3) and those generated by the satellite products as well those measured punctually at the
field by meteorological tower were smaller under the less rainy period (from June to August),
except for ERBE product. The better agreement for this period is probably due to best chance
for obtaining cloud- and smoke-free background values.
70
MODIS-wsa
MODIS-bsa
CCM3-IBIS
UMD (GOES-METEOSAT-AVHRR)
SRB-ISLSCP2
ERBE
CG99 (AVHRR)
Months
Observed-Ducke
Observed-Jaru
Observed-Cuieiras (K34)
K
D
J
W
B
A
E
I
U
C
%
Figure 4.2. Seasonal variability of the surface albedo simulated by the CCM3 model and
albedo products derived from six remote sensing for Amazon basin, and measured
at three study sites.
71
Figure 4.3 shows the seasonal profile of the albedo simulated by the CCM3 model and
the albedo derived by the six remote sensing products for the representative pixels which
include the micrometeorological towers of the Biological Reserves of (a) Cuieiras (K34) and
Ducke and (b) Jaru. Observing the isolated pixel, a good adjustment of the field measured data
at the three sites: Cuieiras and Ducke, located at Amazonas State (Figura 4.4a); and at Jaru
(Figura 4.4b), located at Rondônia State. Overall, MODIS-bsa presented the smallest
amplitude when compared to field measurements and CCM3 simulation data, again. The
smallest albedo differences among the satellite products and simulated by the CCM3 model
were observed from June to August (less rainy period), probably because there are less
cloudiness as confirmed by amount of pixels contaminated by cloud (Durieux et al., 2003),
reducing the number of pixels contaminated by clouds, except for ERBE product (Figure 4.3b).
Highest albedo differences were found for the Biological Reserves of Cuieiras and Ducke,
probably due to both are located nearby urban-areas, causing higher vegetation cover
variability (Figure 4.3a).
As can be seen in Figures 4.2 and 4.3, the greatest differences between the albedos
computed by CCM3 and ERBE, from June to August, probably are due to aerosol effects not
included in the calculation and some due to the fact that ERBE may designate scenes as clear
when they have a small amount of sub-pixel cloudiness in their view. For the other months,
higher amplitude values may be attributed by the cloud contamination of the scenes (especially
during the rainy season, from December to May), the instruments errors, the sampling errors,
and the uncertainties in models used in the data processing. Indeed, the increase of biomass
burning aerosols in the atmosphere, with peak in September, contributes to the increase of the
amplitude among the satellite albedo products and CCM3 albedo model, especially for the
pixel comprising the Jaru Reserve (Ji-Paraná, RO), which is located in the Amazon arc of
72
(a)
(b)
MODIS-wsa
MODIS-bsa
CCM3-IBIS
UMD (GOES-METEOSAT-AVHRR)
SRB-ISLSCP2
ERBE
CG99 (AVHRR)
Observed-Ducke
Observed-Jaru
Observed-Cuieiras (K34)
K
D
J
W
B
A
E
I
U
C
Months
Figure 4.3. Seasonal variability of the surface albedo simulated by the CCM3 model and
albedo products derived from six remote sensing for the pixels comprising the
three study sites, and albedo measured at the (a) Manaus (Cuieiras and Ducke)
and (b) Jaru sites.
73
deforestation. In this specific site, UMD and SRB/ISLSCP2 products presented greatest
variation in relation to the CCM3 model and the field-observed albedos.
In spite of field measurement albedo data is adequate to characterize a specific site;
some differences are foreseen when compared to remotely sensed data, because some spatial
variability is expected in the vicinity of any tower even in relatively homogenous locations
(Lucht et al., 2000b).
Although validation of a satellite derived albedo is almost impossible, once aircraft
measurements cover large areas, ground truth albedo can be used as a procedure to assess the
uncertainty of the estimated product (Lucht et al., 2000b; Justice et al., 2000; Jin et al. 2003).
In spite of some improvements of the remote sensing albedo products have been
observed during the last years, it is important to point out that incorporation of more accurate
models and atmospheric information at different steps of the retrieval procedure may improve
the quality of albedo products.
4.4. SUMMARY AND CONCLUSIONS
In this work, simulated albedo is computed by CCM3 coupled with IBIS and compared
against six different albedo products derived from remote sensing: CG99, ERBE, SRB-
ISLSCP2, UMD, MODIS white-sky and MODIS black-sky), including also representation of
field albedo data at three study sites (Cuieiras, Ducke, and Jaru Reserves). A considerable
variation among the albedo estimated by the different remote sensing products was observed,
including great seasonal differences. MODIS black-sky albedo product presents the smallest
amplitude in relation to the CCM3 simulated and field observed albedos. Biomass burning
aerosols released to atmosphere, especially during the end of dry season, introduces increased errors
to the surface albedo retrieved from UMD, SRB-ISLSCP2 and ERBE. These results suggest that
some caution be exercised when using satellite albedo products over the Amazon basin and
74
more accurate sensors and algorithms need to be developed for reducing the uncertainties
involved in the data processing.
75
CHAPTER 5
RADIATIVE PROCESSES OF PRECIPITATION CHANGE AFTER TROPICAL
DEFORESTATION
5.1. INTRODUCTION
Amazon is one of the regions where the regional climate response to the underneath
vegetation is most intense. According to several studies, that studied slightly different
scenarios using different climate models, a large-scale conversion of tropical rainforest to
pastureland leads to a decrease in local precipitation, which is linearly related to surface
albedo changes (Figure 5.1).
Despite this general agreement, there are still many unanswered questions
regarding the effects of tropical deforestation on the regional climate. For example, it has
not been explained why some experiments, like Nobre et al. (1991), show a consistent
decrease in precipitation over all the deforested area, while others (eg. Lean and Rowntree,
1997; Costa and Foley, 2000) show different spatial patterns of decrease/increase of
76
precipitation, with a predominance of decrease. In addition, there is a contradiction
between large-scale numerical deforestation studies, which predict a decrease in cloudiness
and precipitation, and the observations of the current climate above deforested areas, which
show increased cloudiness (Cutrim et al., 1995; Negri et al., 2004). The answer to these
questions may lie on the competing effects that non-radiative and radiative processes have
on the determination of precipitation change after a tropical deforestation, as well as on the
linear response of precipitation to surface albedo (Dirmeyer and Shukla, 1994; Berbet and
Costa, 2003, among others).
Figure 5.1. Annual mean precipitation profile as a function of albedo changes for different
climatic experiments.
In this context, I run eleven simulations using the NCAR Community Climate
Model (CCM3) coupled to the Integrated Biosphere Simulator (IBIS) to carefully
investigate these processes. This chapter concentrates on the radiative processes of the
precipitation change after tropical deforestation, while a companion paper will investigate
the effects of non-radiative processes on the Amazon precipitation changes. The numerical
77
experiment is designed to cover a wide range of changes in land surface albedo and in
surface net radiation, so that reliable relationships between precipitation change and
radiative budget could be obtained.
This chapter is organized in 8 sections: Sections 2 and 3 describe the models used
and the experiment design, Sections 4 to 6 present the results of the numerical climate
experiment and Sections 7 and 8 present a discussion and conclusions.
5.2. MATERIALS AND METHODS
5.2.1. Description of the CCM3-IBIS model
The NCAR Community Climate Model – CCM3 (Kiehl et al., 1998) is a spectral
atmosphere model of the horizontal fields. The model operates at a resolution of T42, at
about 2.81° x 2.81° and has 18 vertical levels and operates with a 20-min step. The
diagnosis of cloud fraction represents a generalization of the scheme introduced by Slingo
(1987) and depends on relative humidity, vertical pressure velocity, ω; atmospheric
stability; and the convective mass flux associated with parameterized moist convection.
Three types of cloud are diagnosed by the scheme: convective cloud, layered cloud, and
low-level marine stratus (Kiehl et al., 1998). Precipitation occurs in two forms: stable and
convective. When the column is stable to moist convection, only stable precipitation occurs
whenever saturated conditions exist. The phase of precipitation for surface accumulation
depends on low-level temperature. Convective precipitation occurs when the column is
unstable to moist convection and this precipitation process is monthly independent of the
cloud prediction scheme and of the radiative properties of clouds (Kiehl et al., 1998).
The land surface radiation budget is computed by the Integrated Biosphere
Simulator – IBIS (Foley et al., 1996). IBIS has two vegetation layers (i.e., trees and short
vegetation), a variable number of soil layers, and includes representations of land surface
78
processes, like energy, water and momentum exchange between the soil-vegetation-
atmosphere system, and also includes vegetation dynamics component that, in this study is
disabled (use only static vegetation). In IBIS, the exchange of solar radiation between the
soil-vegetation-atmosphere system is calculated following the standard two-stream
approximation, with separate calculations for direct and diffuse radiation in both visible
(VIS) and near-infrared (NIR) bands. In these simulations, IBIS operated on the same T42
spatial resolution as the CCM3 atmospheric model. I also modified the IBIS model to
introduce a new land cover type, a soybean cropland. This new land cover type is based on
the physiology of a C3 grass, but has specific phenology parameterizations that emulate a
soybean crop that grow in Amazonia.
5.2.2. Experiment design
A climate experiment is designed to elucidate the radiative effects of precipitation
change after tropical deforestation on regional climate in Amazon. All of the simulations
are run for 12 years, using the same initial conditions, but only the last 10 years are
averaged to analyze the results. The first two years of simulation are discarded, and are
used to let the model reaches an equilibrium state, specifically with respect to soil
moisture. In all simulations, sea surface temperature and atmospheric composition are set
to the average of 1990 decade.
In this study, I conducted eleven simulations, described below:
(a) Control runs (Rainforest land cover): We run two repetitions (F
a
, F
b
), where the only
difference between them is the albedo of the rainforest, which is set to 0.125 (F
a
) and
0.129 (F
b
), respectively (Culf et al., 1995). LAI and biomass of the rainforest are set
to 5.94 m
2
m
-2
and 10.36 kg C m
-2
(Roberts et al., 1996; Medina and Cuevas, 1996);
79
(b) Pasture land expansion. It is divided in three different simulations, where pastureland
partially replaces the original rainforest in each Amazonia grid cell, increasing
gradually from 0% (control run) to 25%, 50% and 75%. For each level of
deforestation, I run two repetitions (P
a
25%
, P
b
25%
, P
a
50%
, P
b
50%
, P
a
75%
, P
b
75%
) where the
unique difference between them is the albedo of the pasture, which is set to 0.177 and
0.182, respectively (Culf et al., 1995). Average LAI of the pastureland is set to 3.18
m
2
m
-2
(Roberts et al., 1996; Medina and Cuevas, 1996);
(c)
Soybean land expansion. It is divided in three simulations, where each Amazonia grid
cell partially replaces the original rainforest by pastureland, increasing gradually from
0% (control run) to 25% (S
25%
), 50% (S
50%)
and 75% (S
75%
).
In this climate experiment, I assumed a planting date of January 5
th
as most
representative for the region, representing an increasing of albedo and LAI during harvest
with maximum peak of 0.26 and 5.98 m
2
m
-2
, respectively, in March 19
th
, decreasing after
the harvest in May when the leaves fall and dry, completing, this way, the seasonal cycle
of soybean crop. In our model representation of soybean albedo, I chose to represent the
values of albedo according to the literature that suggest a peak of 0.26 (Blad and Baker,
1972; André and Viswanadhan, 1983; Fontana et al., 1991).
5.3. RESULTS
5.3.1. Precipitation dependence on surface radiation balance
In this study, I evaluated the radiative processes of precipitation change after
tropical deforestation through the analysis of the anomaly fields of precipitation (P’),
surface reflected radiation (Sr’), surface net radiation (Rn’), total cloud cover (C’) and
vertical wind velocity (ω’) between deforested (pasture and soybean) and forested
conditions, considering the deforestation levels of 25%, 50% and 75% for both pasture and
80
soybean cropland. The use of pasture and soybean land covers for the deforested cases, as
well as the use of different deforestation levels, allows the study of a wider range of
surface radiation anomalies, helping to establish the relationships between the variables.
Figure 5.2 shows the linear relationship between the precipitation anomalies, P’
(difference between P
ab
%25
, P
ab
%50
, P
ab
%75
and F
ab
) against the annual mean albedo anomalies,
α’ (Figure 5.2a) and against the annual mean net radiation, Rn’ (Figure 5.2b) for the
(a) Annual mean
y = -17.205*x + 0.47
R
2
= 0.96
-2.0
-1.0
0.0
1.0
2.0
0.00 0.02 0.04 0.06 0.08 0.10
α
'
P' (mm day
-1
)
Pasture
25%
50%
75%
25%
50%
75%
Soybean
(b) Annual mean
y = 0.052*x - 0.14
R
2
= 0.99
-2.0
-1.0
0.0
1.0
2.0
-20 -15 -10 -5 0 5 10 15 20
Rn' (W m
-2
)
P' (mm day
-1
)
Pasture
Soybean
25%
50%
75%
25%
75%
50%
Figure 5.2. Annual mean precipitation anomaly as a function of albedo anomaly (a) and
net radiation anomaly (b), for different levels of pastureland and soybean
cropland expansions.
81
different types of land cover (pasture and soybeans) and different levels of deforestation
(25%, 50% and 75%). Changes in land surface albedo and in surface net radiation explain
about 96% (P<10
-15
) and 99% (P<10
-21
) of the precipitation anomaly variance. These
results are similar to those presented in Figure 5.1, although they have less dispersion, due
to the use of a single climate model.
To represent the seasonal variability in both incoming radiation (S
in
) and surface
albedo, I studied the changes in the precipitation after deforestation against changes in the
reflected radiation after deforestation (Sr’,
Sr = S
in
α
.) and surface net radiation (Rn’) at
the semester and trimester time scales. In these analyses, each point represents an average
of each band of latitude, totaling eight bands throughout the studied area.
In a semester basis (Figure 5.3), it is observed a linear decrease of precipitation
anomaly with the increase of Sr’ variance (R
2
varies from 42 to 47%) and with the
decrease of Rn’ (R
2
varies from 56 to 73%). The same profile is observed at a trimester
basis (Figure 5.4). Overall, 26% (P<10
-13
) of variance in P’ is explained by the variance in
Sr’ when all trimesters are analyzed (Figure 5.4e), agreeing with the 28% found by Berbet
and Costa (2003). For each 10 W m
-2
increase in Sr’, the precipitation is reduced by
approximately 0.63 mm day
-1
, while a reduction of 10 W·m
-2
in Rn’ induces a decrease in
the range of 0.31- 0.38 mm day
-1
.
82
(a) Dry semester (May-Oct)
y = -0.0566*x + 0.48
R
2
= 0.42
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(d) Dry semester (May-Oct)
y = 0.0299*x - 0.21
R
2
= 0.73
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
(b) Rainy semester (Nov-Apr)
y = -0.0709*x + 0.54
R
2
= 0.47
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(e) Rainy Semester (Nov-Apr)
y = 0.0801*x - 0.02
R
2
= 0.66
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
(c) Dry and rainy semesters
y = -0.0629*x + 0.50
R
2
= 0.43
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(f) Dry and rainy semesters
y = 0.0378*x - 0.20
R
2
= 0.56
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
Figure 5.3. Semester mean precipitation anomaly as a function of albedo anomaly for the
dry semester (a), rainy semester (b), and both (c), and as a function of net
radiation anomaly for the dry semester (d), rainy semester and both (f). White
squares represent pastureland and gray squares represent soybean cropland.
83
(a) Trimester (Jan -Mar)
y = -0.0527*x + 0.20
R
2
= 0.30
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(c) Trimester (Jul-Sep)
y = -0.0444*x + 0.03
R
2
= 0.37
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(f) Trimester (Jan-Mar)
y = 0.0557*x - 0.22
R
2
= 0.28
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
(h) Trimester (Jul-Sep)
y = 0.0031
NS
x - 0.65
R
2
= 0.01
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
(e) All trimester
y = -0.0632*x + 0.50
R
2
= 0.26
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(j) All trimester
y = 0.0308*x - 0.24
R
2
= 0.24
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
(b) Trimester (Apr-Jun)
y = -0.0451*x + 0.62
R
2
= 0.42
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(d) Trimester (Oct-Dec)
y = -0.1266*x + 1.29
R
2
= 0.29
-4.0
-2.0
0.0
2.0
4.0
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
P' (mm day
-1
)
(g) Trimester (Apr-Jun)
y = 0.0277*x + 0.02
R
2
= 0.56
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
(i) Trimester (Oct-Dec)
y = 0.1065*x + 0.31
R
2
= 0.75
-4.0
-2.0
0.0
2.0
4.0
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
P' (mm day
-1
)
Figure 5.4. Trimester mean precipitation anomaly as a function of albedo anomaly for the
January to March trimester (a), April to June trimester (b), July to September
trimester (c), October to December trimester (d) and all trimesters (e), and as a
function of net radiation anomaly for the January to March trimester (f), April
to June trimester (g), July to September trimester (h), October to December
trimester (i) and all trimesters (j). White squares represent pastureland and gray
squares represent soybean cropland.
84
5.3.2. Clouds dependence on surface radiation balance and feedbacks on the incoming
solar radiance
Figure 5.5 shows the annual mean cloudiness anomalies (C’) as a function of
albedo and net radiation anomalies. Changes in land surface albedo and in annual mean net
radiation explain about 97% (P<10
-16
) and 88% (P<10
-10
) of the cloudiness variance,
respectively, when different types of land cover and levels of deforestation were
considered. An increase in α’ or a decrease in Rn’ induces a decrease in C’.
(a) Annual mean
y = -0.7188*x + 0.01
R
2
= 0.97
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
0.00 0.02 0.04 0.06 0.08 0.10
α
'
C'
Pasture
Soybean
25%
50%
75%
25%
50%
75%
(b) Annual mean
y = 0.002*x - 0.016
R
2
= 0.88
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
-20-15-10-5 0 5 101520
Rn'
(W m
-2
)
C'
Pasture
Soybean
25%
50%
75%
25%
50%
75%
Figure 5.5. Annual mean total cloud anomaly as a function of albedo anomaly (a) and net
radiation anomaly (b), for different levels of pastureland and soybean cropland
expansions.
85
Similar results are found at the semester (Figure 5.6) and trimester (Figure 5.7)
basis. Variations in C’ are better explained by Sr’ and Rn’ during the rainy season than
during the dry season (Figures 5.6b, 5.6c, 5.7a, 5.7f and 5.7i). Note that positive values of
C’ are common for low levels of pastureland expansion (low Sr’). This is consistent with
the results found by several authors. Initially, Chu et al. (1994) analyzed outgoing
longwave radiation (OLR) from 1974 to 1990 and monthly rainfall totals at Belém and
Manaus. Both analyses showed increasing trends in cloudiness and precipitation (related to
an increase in convection) associated with deforestation over almost the entire Amazon
basin. The most significant increase in convection was found in the western equatorial part
of Amazonia, along the eastern slope of the Andes, where rainfall is most abundant. In
addition, Cutrim et al. (1995) using the GOES (Geostationary Operational Environmental
Satellite) visible imagery, demonstrated that an enhanced frequency of shallow cumulus
cloudiness is formed primarily over deforested regions and over topographic elevations
during the dry season in the Amazon.
Furthermore, Negri et al. (2004) observed that in the dry season, when the effects of
the surface are not overwhelmed by large-scale weather disturbances, cumulus cloudiness,
deep convective cloudiness, and rainfall occurrence are larger over the deforested and
savanna regions than over areas of dense forest. This can be in response to a local
circulation initiated by the differential heating of the region’s varying forestation. Finally,
Durieux et al. (2003) verified during the wet season that convective cloudiness is stronger
over deforested areas, while a significant decrease in convective cloudiness is seen during
the dry season, together with an increase in low-level clouds, mainly made up of shallow
cumulus clouds, that with subsequent expansion of deforestation these clouds tend to
stabilize.
86
(b) Rainy semester (Nov-Apr)
y = -0.0023*x + 0.02
R
2
= 0.68
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(a) Dry semester (May-Oct)
y = -0.0029*x + 0.01
R
2
= 0.32
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(c) Dry and rainy semesters
y = -0.0028*x + 0.02
R
2
= 0.37
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(e) Rainy semester (Nov-Apr)
y = 0.0024*x + 0.001
R
2
= 0.77
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
(d) Dry semester (May-Oct)
y = 0.0002
NS
x - 0.04
R
2
= 0.01
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
(f) Dry and rainy semesters
y = 0.0005*x - 0.02
R
2
= 0.05
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
Figure 5.6. Semester mean total cloud anomaly as a function of albedo anomaly for the
dry semester (a), rainy semester (b), and both (c), and as a function of net
radiation anomaly for the dry semester (d), rainy semester and both (f). White
squares represent pastureland and gray squares represent soybean cropland.
87
(c) Trimester (Jul-Sep)
y = -0.0031*x - 0.005
R
2
= 0.18
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(a) Trimester (Jan-Mar)
y = -0.0018*x + 0.01
R
2
= 0.73
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(e) All trimester
y = -0.0028*x + 0.02
R
2
= 0.23
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(d) Trimester (Oct-Dec)
y = -0.0059*x + 0.06
R
2
= 0.50
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(b) Trimester (Apr-Jun)
y = -0.0007
NS
x- 0.001
R
2
= 0.05
-0.20
-0.10
0.00
0.10
0.20
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
C'
(h) Trimester (Jul-Sep)
y = -0.0006
NS
x - 0.06
R
2
= 0.05
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
(f) Trimester (Jan-Mar)
y = 0.0019*x - 0.002
R
2
= 0.65
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
(j) All trimester
y = 0.0005*x - 0.02
R
2
= 0.03
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
(i) Trimester (Oct-Dec)
y = 0.0041*x + 0.007
R
2
= 0.89
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
(g) Trimester (Apr-Jun)
y = -0.0004
NS
x - 0.01
R
2
= 0.05
-0.20
-0.10
0.00
0.10
0.20
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
C'
Figure 5.7. Trimester mean total cloud anomaly as a function of albedo anomaly for the
January to March trimester (a), April to June trimester (b), July to September
trimester (c), October to December trimester (d) and all trimesters (e), and as a
function of net radiation anomaly for the January to March trimester (f), April
to June trimester (g), July to September trimester (h), October to December
trimester (i) and all trimesters (j). White squares represent pastureland and gray
squares represent soybean cropland.
88
(g) Trimester (Jul-Sep)
y = -238.78*x + 12.32
R
2
= 0.56
-10
10
30
50
70
90
-0.20 -0.10 0.00 0.10 0.20
C'
Sin'(W m
-2
)
(e) Trimester (Jan -Mar)
y = -212.83*x + 9.15
R
2
= 0.40
-10
10
30
50
70
90
-0.20 -0.10 0.00 0.10 0.20
C'
Sin'(W m
-2
)
(i) All trimester
y = -169.25*x + 14.83
R
2
= 0.35
-10
10
30
50
70
90
-0.20 -0.10 0.00 0.10 0.20
C'
Sin' (W m
-2
)
(h) Trimester (Oct-Dec)
y = -18.674
NS
x + 17.69
R
2
= 0.02
-10
10
30
50
70
90
-0.20 -0.10 0.00 0.10 0.20
C'
Sin'(W m
-2
)
(f) Trimester (Apr-Jun)
y = -300.28*x + 17.13
R
2
= 0.40
-10
10
30
50
70
90
-0.20 -0.10 0.00 0.10 0.20
C'
Sin'(W m
-2
)
(c) Rainy semester (Nov-Apr)
y = -123.94*x + 11.98
R
2
= 0.23
0
10
20
30
40
50
60
70
-0.20 -0.10 0.00 0.10 0.20
C'
Sin'(W m
-2
)
(b) Dry semester (May-Oct)
y = -221.48*x + 16.28
R
2
= 0.38
0
10
20
30
40
50
60
70
-0.20 -0.10 0.00 0.10 0.20
C'
Sin'(W m
-2
)
(d) Dry and rainy semesters
y = -242.59*x + 13.07
R
2
= 0.46
0
10
20
30
40
50
60
70
-0.20 -0.10 0.00 0.10 0.20
C'
Sin' (W m
-2
)
(a) Annual mean
y = -234.25*x + 13.32
R
2
= 0.80
5
10
15
20
25
30
-0.06 -0.04 -0.02 0 0.02 0.04 0.06
C'
Sin' (W m
-2
)
Pasture
Soybean
25%
50%
75%
25%
50%
75%
Figure 5.8. Annual mean incoming solar radiation anomaly as a function of total cloud
anomaly (a) for different levels of pastureland and soybean cropland
expansions. Semester mean incoming solar radiation anomaly as a function of
total cloud anomaly for the dry semester (a), rainy semester (b), and both (c).
Trimester mean incoming solar radiation anomaly as a function of total cloud
anomaly for the January to March trimester (a), April to June trimester (b), July
to September trimester (c), October to December trimester (d) and all
trimesters (e). White squares represent pastureland and gray squares represent
soybean cropland.
89
The incoming radiation cloud feedbacks are shown in Figure 5.8. Changes in
annual mean cloudiness explain about 80% (P<10
-7
) of the annual mean incoming solar
radiation variance. A linear relationship is observed for all studied cases, presenting a
positive anomaly that decreases with the increase of cloudiness anomaly. The higher
dispersion of the data, compared to the previous analyses, is probably due to the several
types of clouds and optical properties present in the region
5.3.3. Convection dependence on surface radiation balance
Figure 5.9 shows the relationship between vertical wind velocity anomaly at 500
hPa (ω’) and albedo and net radiation anomalies for different types of land cover and
different levels of deforestation. A positive anomaly indicates a stronger subsidence over
the studied area. Changes in land surface albedo and in net radiation explain about 89%
(P<10
-11
) and 97% (P<10
-17
) of the vertical wind variance, respectively.
Similar results are found at the semester (Figure 5.10) and trimester basis (Figure
5.11). Following the results for the cloudiness, ω’ is more dependent on Sr’ and Rn’ during
the rainy season than during the dry season (Figure 5.10), in particular during Oct-Dec
(Figure 5.11d and 5.11i).
Figure 5.12 shows the linear relationship of the precipitation anomalies as a
function of ω’ in an annual, semester and trimester basis. At the annual basis, 98% (P<10
-
17
) of variation of the precipitation anomaly is explained by the changes in ω’. The
precipitation anomaly reduces with the increase of the ω’ anomaly. Changes in Rn’ and α’,
due to the replacement of forest in different types of land cover (pasture and soybean crop),
cause changes on precipitation and vertical wind velocity. Similar profile is verified at the
semester and trimester analysis (Figures 5.12b to 5.12i).
90
(a) Annual mean
y = 0.1441*x - 0.01
R
2
= 0.89
-0.02
-0.01
0.00
0.01
0.02
0.00 0.02 0.04 0.06 0.08 0.10
α
'
ω
' (Pa s
-1
)
Pasture
Soybean
25%
50% 75%
25%
50%
75%
(b) Annual mean
y = -0.0004*x - 0.001
R
2
= 0.97
-0.02
-0.01
0.00
0.01
0.02
-20 -15 -10 -5 0 5 10 15 20
Rn' (W m
-2
)
ω
' (Pa s
-1
)
Pasture
Soybean
25%
50%
75%
25%
50%
75%
Figure 5.9. Annual mean vertical velocity anomaly as a function of albedo anomaly (a)
and net radiation anomaly (b), for different levels of pastureland and soybean
cropland expansions.
91
(b) Rainy semester (Nov-Apr)
y = 0.0007*x - 0.01
R
2
= 0.33
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(a) Dry semester (May-Oct)
y = 0.0005*x - 0.01
R
2
= 0.17
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(c) Dry and rainy semesters
y = 0.0006*x - 0.01
R
2
= 0.21
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(e) Rainy semester (Nov-Apr)
y = -0.0008*x - 0.001
R
2
= 0.51
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
(d) Dry semester (May-Oct)
y = -0.0004*x - 0.002
R
2
= 0.75
-0.04
-0.02
0.00
0.02
0.04
-30-20-100102030
Rn' (W m
-2
)
ω
' (Pa s
-1
)
(f) Dry and rainy semesters
y = -0.0005*x - 0.001
R
2
= 0.56
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
Figure 5.10. Semester mean vertical velocity anomaly as a function of albedo anomaly for
the dry semester (a), rainy semester (b), and both (c), and as a function of net
radiation anomaly for the dry semester (d), rainy semester and both (f). White
squares represent pastureland and gray squares represent soybean cropland.
92
(c) Trimester (Jul-Sep)
y = 0.0003*x - 0.002
R
2
= 0.22
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(a) Trimester (Jan-Mar)
y = 0.0005*x - 0.003
R
2
= 0.26
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(e) All trimester
y = 0.0006*x - 0.006
R
2
= 0.16
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(d) Trimester (Oct-Dec)
y = 0.0013*x - 0.01
R
2
= 0.21
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(b) Trimester (Apr-Jun)
y = 0.0004*x - 0.01
R
2
= 0.15
-0.04
-0.02
0.00
0.02
0.04
0 5 10 15 20 25 30 35 40
Sr' (W m
-2
)
ω
' (Pa s
-1
)
(h) Trimester (Jul-Sep)
y = -0.0001*x + 0.002
R
2
= 0.26
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
(f) Trimester (Jan-Mar)
y = -0.0005*x + 0.002
R
2
= 0.23
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
(j) All trimester
y = -0.0004*x - 0.0002
R
2
= 0.31
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
(i) Trimester (Oct-Dec)
y = -0.0012*x - 0.006
R
2
= 0.70
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
(g) Trimester (Apr-Jun)
y = -0.0004*x - 0.003
R
2
= 0.57
-0.04
-0.02
0.00
0.02
0.04
-30 -20 -10 0 10 20 30
Rn' (W m
-2
)
ω
' (Pa s
-1
)
Figure 5.11. Trimester mean vertical velocity anomaly as a function of albedo anomaly for
the January to March trimester (a), April to June trimester (b), July to
September trimester (c), October to December trimester (d) and all trimesters
(e), and as a function of net radiation anomaly for the January to March
trimester (f), April to June trimester (g), July to September trimester (h),
October to December trimester (i) and all trimesters (j). White squares
represent pastureland and gray squares represent soybean cropland.
93
(e) Trimester (Jan -Mar)
y = -92.002*x - 0.11
R
2
= 0.93
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(g) Trimester (Jul-Sep)
y = -84.183*x - 0.43
R
2
= 0.54
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(i) All trimester
y = -83.521*x - 0.25
R
2
= 0.89
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
(Pa s
-1
)
P' (mm day
-1
)
(f) Trimester (Apr-Jun)
y = -62.59*x - 0.15
R
2
= 0.81
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(h) Trimester (Oct-Dec)
y = -84.532*x - 0.20
R
2
= 0.97
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(b) Dry semester (May-Oct)
y = -67.525*x - 0.35
R
2
= 0.85
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(c) Rainy semester (Nov-Apr)
y = -86.968*x - 0.14
R
2
= 0.93
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(d) Dry and rainy semesters
y = -74.816*x - 0.26
R
2
= 0.87
-4.0
-2.0
0.0
2.0
4.0
-0.04 -0.02 0 0.02 0.04
ω
' (Pa s
-1
)
P' (mm day
-1
)
(a) Annual mean
y = -113.86*x - 0.22
R
2
= 0.98
-2.0
-1.0
0.0
1.0
2.0
-0.02 -0.01 0.00 0.01 0.02
ω
'(Pa s
-1
)
P' (mm day
-1
)
Pasture
25%
50%
75%
25%50%
75%
Soybean
Figure 5.12. Annual mean precipitation anomaly as a function of vertical velocity anomaly
(a) for different levels of pastureland and soybean cropland expansions.
Semester mean precipitation anomaly as a function of vertical velocity
anomaly for the dry semester (a), rainy semester (b), and both (c). Trimester
mean precipitation anomaly as a function of vertical velocity anomaly for the
January to March trimester (a), April to June trimester (b), July to September
trimester (c), October to December trimester (d) and all trimesters (e). White
squares represent pastureland and gray squares represent soybean cropland.
94
Figures 5.13 and 5.14 show the seasonal profile of vertical wind velocity at 500
hPa, ω, for forest (F
ab
) (Figures 5.13a, 5.13b, 5.14a and 5.14b) and for their anomalies
(ω’) over different types of land cover and different levels of deforestation (P
ab
25%
- F
ab
:
Figures 5.13c and 5.13d, P
ab
%50
- F
ab
: Figures 5.13e and 5.13f, P
ab
75%
- F
ab
: Figures 5.13g and
5.13h, S
25%
- F
ab
: Figures 5.14c and 5.14d, S
50%
- F
ab
: Figures 5.14e and 5.14f, S
75%
- F
ab
:
Figures 5.14g and 5.14h). The left column represents the rainy season (February to April),
and for the right column represents the dry season (July to August). A strong convection is
observed in the tropical forest throughout of the rainy season, decreasing in dry season. On
the other hand, Amazon land surface deforestation result in a reduction in convection and
precipitation, generating stronger subsidence motions over the region (ω’>0), once the
albedo values increase and cause reduction in Rn’ as a response to the reduction of land
cover, as verified for the soybean cropland analyzed in this study (Figures 5.13c to 5.13h
and 5.14c to 5.14h).
Seasonal precipitation profile is also observed in the Forest (F
ab
) and in the
anomalies for different types of land cover (P
ab
25%
P
ab
50%
P
ab
75%
, S
25%
, S
50%
, S
75%
) and in
different levels of deforestation at 500 hPa, considering the three rainiest months (February
to April) and driest months (June to August) (Figures 5.15 and 5.16). In the Amazon region
the rainfall is mainly of convective origin. In the forest the evapotranspiration is
continuous throughout the year, presenting deep roots and also more available radiation,
associated mainly to the smaller value of albedo. This profile can be evidenced in the rainy
season (Figures 5.15a and 5.16a) and a little less in the dry season, where a subsidence
motions begins to be evidenced. That motion expands with the increase of deforestation
from 25% to 75%, modifying the rainfall pattern. The combined effects of changes in α’,
ω’ and C’ results changes in the rainfall. Figures 5.15c to 5.15h and 5.16c to 5.16h show a
95
F
ab
F
ab
P
25%
- F
ab
ab
P
25%
- F
ab
ab
P
50%
- F
ab
ab
P
50%
- F
ab
ab
P
75%
- F
ab
ab
P
75%
- F
ab
ab
Figure 5.13. Seasonal profile of vertical velocity at 500 hPa, in forest (F
ab
) for the
February to April trimester (rainy season) (a) for the June to August trimester
(dry season) (b) for the vertical velocity anomalies in different levels of
pastureland expansions: P
ab
%25
- F
ab
(c), P
ab
%50
-F
ab
(d) and P
ab
%75
- F
ab
(e) for the
February to April trimester, respectively, and for the anomalies P
ab
%25
- F
ab
(c),
P
ab
%50
-F
ab
(g) and P
ab
%75
- F
ab
(h) for the June to August trimester, respectively.
Positive values are represented by solid line and indicate decrease in vertical
motion and negative values are represented by dashed line and indicate
increase in vertical motion.
96
S
25%
-
F
ab
F
ab
F
ab
S
25%
-
F
ab
S
50%
-
F
ab
S
50%
-
F
ab
S
75%
-
F
ab
S
75%
-
F
ab
Figure 5.14. Seasonal profile of vertical velocity at 500 hPa, in forest (F
ab
) for the
February to April trimester (rainy season) (a) for the June to August trimester
(dry season) (b) for the vertical velocity anomalies in different levels of
soybean cropland expansions: S
25%
- F
ab
(c), S
50%
- F
ab
(d) and S
75%
- F
ab
(e)
for the February to April trimester, respectively, and for the anomalies S
25%
-
F
ab
(c), S
50%
- F
ab
(g) and S
75%
- F
ab
(h) for the June to August trimester,
respectively. Positive values are represented by solid line and indicate
decrease in vertical motion and negative values are represented by dashed line
and indicate increase in vertical motion.
97
F
ab
F
ab
P
25%
-
F
ab
ab
P
25%
-
F
ab
ab
P
50%
-
F
ab
ab
P
50%
-
F
ab
ab
P
75%
-
F
ab
ab
P
75%
-
F
ab
ab
Figure 5.15. Seasonal profile of precipitation at 500 hPa, in forest (F
ab
) for the February
to April trimester (rainy season) (a) for the June to August trimester (dry
season) (b) for the precipitation anomalies in different levels of pastureland
expansions: P
ab
%25
- F
ab
(c), P
ab
%50
-F
ab
(d) and P
ab
%75
- F
ab
(e) for the February to
April trimester, respectively, and for the anomalies P
ab
%25
- F
ab
(c), P
ab
%50
-F
ab
(g) and P
ab
%75
- F
ab
(h) for the June to August trimester, respectively. Negative
values are represented by dashed line and indicate decrease in pasture
precipitation and positive values are represented by solid line and indicate
increase in pasture precipitation.
98
F
ab
S
75%
-
F
ab
F
ab
S
75%
-
F
ab
S
50%
-
F
ab
S
50%
-
F
ab
S
25%
-
F
ab
S
25%
-
F
ab
Figure 5.16. Seasonal profile of precipitation at 500 hPa, in forest (F
ab
) for the February
to April trimester (rainy season) (a) for the June to August trimester (dry
season) (b) for the precipitation anomalies in different levels of soybean
cropland expansions: S
25%
- F
ab
(c), S
50%
- F
ab
(d) and S
75%
- F
ab
(e) for the
February to April trimester, respectively, and for the anomalies S
25%
- F
ab
(c),
S
50%
- F
ab
(g) and S
75%
- F
ab
(h) for the June to August trimester,
respectively. Negative values are represented by dashed line and indicate
decrease in pasture precipitation and positive values are represented by solid
line and indicate increase in pasture precipitation.
99
fairly significant decrease in rainfall over the deforested region, implicating in reduced
ascent at 500 hPa. This decrease is statistically significant at 95% level of confidence
(Table 5.2).
Table 5.2.
Coefficient of determination (R
2
) of the linear regression analysis between Rn’
(surface net radiation), ω’ (wind vertical velocity), C’ (total cloud) and P’ (precipitation)
versus the variables in the left bar.
Rn’
ω’
C’ P’
Sr’
0.93 0.85 0.98 0.93
Rn’
– 0.97 0.88 0.97
ω
’
–
– 0.82 0.98
C’ –
– – 0.91
5.4. DISCUSSION
This climate experiment was designed to cover a wide range of variation in land
surface albedo and in the surface net radiation, by replacing a low-albedo rainforest by two
land cover types, a mid-albedo pastureland and a high-albedo soybean cropland,
considering different levels of deforestation (25%, 50% and 75%). The choice of low
deforestation levels also allows a closer comparison to the (current) observed variations.
A summary of all experiments is presented in Table 5.1. Simulated mean values of
precipitation and evaporation for tropical rainforest are 5.44 and 3.85 mm day
-1
. Several
authors, reviewed by Costa and Foley (1998), found mean precipitation values from 5.19 to
6.36 mm day
-1
, with an average of 5.84 mm day
-1
. Annual mean evapotranspiration from
LBA towers values range from 3.30 to 3.68 mm day
-1
(Biajoli, 2006). Deforestation had a
100
Table 5.1. Annual mean of the variables and of some parameters: Rn (net radiation), Sin (downward solar radiation), Sout (upward solar
radiation), Lin-Lout (longwave balance), P (precipitation), E (evaporation), LE (latent heat flux) and H (sensible heat flux) for the simulations
F
ab
(Forest, control run), P
ab
%25
(75% forest and 25% pasture), P
ab
%50
(50% forest and 50% pasture), P
ab
%75
(25% forest and 75% pasture), S25% (75%
forest and 25% soybean), S50% (50% forest and 50% soybean), S75% (25% forest and 75% soybean), P
total
(extrapolating to 100% of forest
replaced by pasture), S
total
(extrapolating to 100% of forest replaced by soybean), P
total
- F
ab
( difference between pasture and forest) and S
total
-
F
ab
(difference between soybean and forest).
Experiment
F
ab
P
ab
%25
P
ab
%50
P
ab
%75
S
25%
S
50%
S
75%
P
total
S
total
P
total
- F
ab
S
total
- F
ab
Rn (W m
-2
) 139.0 143.0 141.0 139.2 133.3 127.0 122.0 137.5 116.0 -1.4 -23.0
LE (W m
-2
) 112.8 107.7 101.7 95.7 100.8 91.6 82.8 89.7 73.8 -23.1 -39.0
H (W m
-2
) 27.9 36.3 40.2 44.6 33.7 36.4 39.7 48.7 42.7 -20.8 -14.8
Sin (W m
-2
) 216.2 229.2 233.4 237.8 231.9 236.8 242.0 242.0 247.1 +25.8 +30.9
Sout (W m
-2
) 27.5 34.0 37.2 40.2 40.2 46.3 52.1 43.3 58.1 +15.8 +30.6
Lin-Lout (W m
-2
) 49.8 52.7 55.6 57.1 58.4 63.4 68.2 59.3 73.1 +9.5 +23.5
P (mm day
-1
) 5.44 5.52 5.38 5.28 5.01 4.66 4.42 5.16 4.13 -0.28 -1.31
E (mm day
-1
) 3.85 3.67 3.48 3.28 3.43 3.13 2.83 3.09 2.53 -0.76 -1.31
α
0.127 0.148 0.159 0.169 0.174 0.196 0.216 0.179 0.237 +0.052 +0.110
C 0.729 0.725 0.716 0.707 0.710 0.691 0.676 0.698 0.658 -0.032 -0.071
101
substantial effect on the climate of the tropical forest region examined as a reduction in
precipitation of -0.3 mm day
-1
and in evaporation of -0.8 mm day
-1
.
The radiative mechanism for precipitation reduction after tropical deforestation
goes as follows: The increased albedo reduces the surface net radiation. In a region where
convection is strong, this weakens the convective activity, reducing cloudiness and
precipitation. The cloud reduction, however, increases the incoming radiation, introducing
a negative feedback on the process. The final precipitation decrease is smaller in the
presence of the cloud feedback than it would be in their absence (Figures 5.17 and 5.18).
Sin’
Rn’
ω’
C’
Radiative Processes
Cloud
Feedbacks
Sin’ = 13.32 - 234.25C’
R
2
=0.80, P<10
-7
ω’ = -0.001 - 0.0004Rn’
R
2
=0.97, P<10
-17
C’ = -0.02 - 4.3364ω’
R
2
=0.82, P<10
-8
α’
Sr’
P’
P’ = 0.20 - 22.976C’
R
2
=0.91, P<10
-11
Rn’ = 12.82 - 1.2234Sr’
R
2
=0.93, P<10
-12
Sin’
Rn’
ω’
C’
Radiative Processes
Cloud
Feedbacks
Sin’ = 13.32 - 234.25C’
R
2
=0.80, P<10
-7
ω’ = -0.001 - 0.0004Rn’
R
2
=0.97, P<10
-17
C’ = -0.02 - 4.3364ω’
R
2
=0.82, P<10
-8
α’
Sr’
P’
P’ = 0.20 - 22.976C’
R
2
=0.91, P<10
-11
Rn’ = 12.82 - 1.2234Sr’
R
2
=0.93, P<10
-12
Figure 5.17. Diagram of anomalies of radiative processes.
102
P’
α’
without cloud feedbacks
with cloud feedbacks
negative feedbacks
P’
α’
without cloud feedbacks
with cloud feedbacks
negative feedbacks
Figure 5.18. Schematic representation of the precipitation changes with and without cloud
feedbacks.
The results of this study not only agree with previous observational, modeling and
theoretical studies on the subject, but also help explain the dichotomy between the
decreasing precipitation results from full large-scale deforestation climate experiments, and
the observed increase in cloudiness and precipitation over mesoscale deforested regions.
The next paragraphs discuss each of these points.
Results of incoming solar radiation and surface net radiation from this experiment
diverge of the results observed from the ABRACOS experiment, which was especially
designed to contrast the microclimatology of forested and deforested conditions. In this
study incoming solar radiation assumed the values of 216.2 and 242.0 W m
-2
for forest and
pasture,
respectively and, surface net radiation assumed the values of 139.0 and 137.5 W
m
-2
(Table 5.1). On the other hand, Culf et al. (1996) analyzing results from ABRACOS
experiment found 193.63 and 192.59 W m
-2
for incoming solar radiation and 131.25 and
116.67 W m
-2
for surface net radiation for forest and pasture in the Jí-Paraná region,
respectively; while Nobre et al. (1996) encountered 228.01 and 218.75 W m
-2
for incoming
solar radiation and, 138.89 and 122.69 W m
-2
for surface net radiation, respectively. These
103
divergences probably occur due to the scale which the observations were made. Incoming
solar radiation and surface net radiation increase over deforested area in the model in study
because the cloudy feedbacks. On the other hand, contiguous area of pasture in the
ABRACOS experiment probably was very little so that these feedbacks occurred.
The results also help explain several inconsistencies among previous published
modeling results. I will separate the discussion in two: climate sensitivity to increased
albedo, and the role of albedo in determining spatial patterns of increase/decrease in
precipitation.
Most Amazon deforestation climate experiments show an average decrease in
precipitation over the deforested area. Some experiments, however, like Nobre et al.
(1991), show a consistent decrease in precipitation over all the deforested area, while
others (eg. Lean and Rowntree, 1997; Costa and Foley, 2000; Voldoire and Royer, 2004)
show different spatial patterns of decrease/increase of precipitation, but with a
predominance of decrease. This different behavior is easily explained by Figures 5.3 and
5.4. In these figures, the mid-albedo pastureland is represented by gray boxes, while the
high-albedo soybean cropland is represented by the white boxes. Each box represents the
average state for a group of deforested Amazon grid cells over a band of latitude. From
these figures, it is clear that, the higher the increase in Sr (α), the lower the likelihood of a
group of grid cells to show a positive anomaly of precipitation. Thus, the low values of
land surface albedo anomaly in Costa and Foley (2000) – 0.035 – and in Lean and
Rowntree (1997) – 0.05 – are consistent with a spatial patterns of mixed decrease and
increase of precipitation, while the high land surface albedo anomaly in Nobre et al. (1991)
– 0.09 – is consistent with a regular pattern of precipitation decrease. Thus, the
discrepancies in the regional precipitation patterns in the literature are mainly radiatively
controlled, and are, in fact, a consequence of the land surface albedo choice of the modeler.
104
Our results are also consistent with the theoretical studies published. The theoretical
background about the precipitation changes after a tropical deforestation can be traced to
three main articles. Charney (1975) hypothesized that the subsidence over the Sahara
desert, at that time explained only by the descendent branch of the Hadley cell, might be
strengthened by its high albedo. Charney theory considered a dry atmosphere, and could
not be applied to humid tropical regions like the Amazon rainforest. Later, Eltahir (1996)
proposed a similar theory that considered a moist atmosphere. According to the author, the
increase of surface albedo after deforestation cause a reduction in the net surface radiation,
which cools the upper atmosphere over the deforested area, inducing a direct thermal
circulation causing a descendent motion over the deforested region, which results in
subsidence. Zeng and Neelin (1999) completed Eltahir’s theory, using linear assumptions
to describe the radiative processes, and the energy, water balance and cloud-radiative
feedback mechanisms in the atmospheric column. His final formulation proposes a linear
relationship between P’ and (S
in
·α)’, or Sr’, similar to the statistically significant regression
equations in Figures 5.3 and 5.4.
The anomaly in precipitation is explained by radiative and non-radiative processes.
The precipitation change after a tropical deforestation may be explained by the simplified
equation 1
P’ = a
NR
+ b
R
Rn’ (1)
where P’ is the anomaly in precipitation, Rn’ is the increase in surface net radiation, a
NR
is
a linear coefficient, related to non-radiative processes, and
b
R
is an angular coefficient,
related to radiative processes (Figure 5.1). From our analysis, usually the non-radiative
term (a
NR
) is small (< 0.3 mm.day
-1
), while the radiative term dominates (b
R
·Rn’ may be as
high as -2 mm·day
-1
). To facilitate comparisons among the experiments, I may choose to
write P’ as a function of the land surface albedo (α’) using the simplified equation 2
105
P’ = a
NSR
+ b
SR
α’
(2)
where a
NSR
is a linear coefficient related to non-solar radiation processes (includes non-
radiative processes and long wave radiation feedbacks), and b
SR
is an angular coefficient
related to the solar radiation processes only, which may also be defined as the Amazon
climate sensitivity to increased land surface albedo. From Figure 5.2, the precipitation
sensitivity from this experiment is -0.17 mm.day
-1
per 0.01 increase in land surface albedo.
This is consistent with the sensitivity calculated using the results found by Dirmeyer and
Shukla (1994) – about -0.19 mm.day
-1
per 0.01 increase in land surface albedo (using the
COLA/SSiB GCM), and with the slope of Figure 5.1 regression line (-0.16 mm.day
-1
per
0.01 increase in land surface albedo). Note that Equation (2) may only be used at the
annual mean scale. The linear coefficient a
NSR
may also be estimated using the results
above. From my analysis, a
NSR
~ 0.47 mm day
-1
; from Dirmeyer and Shukla (1994) data, I
calculate that it is 0.27 mm day
-1
.
Considering only the non-radiative effects of deforestation (equivalent to assume
Rn’=0), the decrease in roughness length, leaf area index, and rooting depth
,
cause a
decrease in the latent heat flux, and an increase in surface temperature, sensible heat flux,
atmospheric instability, cloudiness, and precipitation. On the other hand, the radiative
effects, caused primarily by increases in surface albedo, but also by long-wave and cloud-
radiative feedbacks, cause a decrease in both surface heat fluxes (due to reduced radiation
absorbed by the surface), resulting in a loss of radiative energy from the Amazon land
surface, cooling off the atmospheric column over the deforestation which induces a
thermally driven circulation that results in subsidence, with subsequent reduction in
convection, cloudiness, and precipitation.
This competitive effect explains why, although most Amazon deforestation climate
experiments show an average decrease in precipitation over the deforested area, it is
106
possible to find experiments that actually show an average increase. Depending on the
parameterizations used, precipitation may increase or decrease. For the coefficients
obtained not only by this study, but also by Dirmeyer and Shukla (1994), an increase in
land surface albedo of about 0.03 is necessary in order to offset the non-radiative processes
that tend to increase precipitation. For most of the cases in the literature, consistent with a
full deforestation, the choices of parameters were so that it resulted in a decrease in
average precipitation. In partial deforestation scenarios, the average increase in land
surface albedo over a large grid cell is small, leading to a very small precipitation change.
5.5. CONCLUSIONS
This study evaluated the radiative processes of the precipitation change after
tropical deforestation and its effects on regional climate in Amazon, considering different
levels of deforestation (25%, 50% and 75%), and different types of land cover
replacements for the original rainforest land cover. The results showed that the changes in
precipitation are linearly related to the anomaly in net surface radiation, and the change in
precipitation due to radiative mechanisms may be an order of magnitude higher than those
due to non-radiative mechanisms. The basic mechanism proposed for precipitation
reduction after tropical deforestation starts with the decrease in surface net radiation due to
the increased albedo after deforestation. In a region where convection is strong, this
weakens the convective activity, reducing cloudiness and precipitation. The cloud
reduction, however, increases the incoming radiation, introducing a negative feedback on
the process. The final precipitation decrease is smaller in the presence of the cloud
feedback than it would be in their absence. Competitive non-radiative mechanisms tend to
increase precipitation. The non-radiative processes of the precipitation change after
tropical deforestation will be analyzed in a companion paper.
107
This study explains, for the first time, why it is consistent to observe increased
cloudiness and precipitation under partial deforestation scenarios, and to still believe that
large-scale deforestation might lead to a decrease in precipitation. Furthermore, if the
large-scale deforestation is related to the expansion of croplands, the precipitation change
may be much higher than if caused by the expansion of pasturelands. This contrast will
also be explored by a further study.
5.6. NOMENCLATURE
F
a
control run (Rainforest land cover)-1
st
repetition
F
b
control run (Rainforest land cover)-2
ns
repetition
P
a
25%
rainforest partially replaces pastureland (level of deforestation 25%) -1
st
repetition
P
b
25%
rainforest partially replaces pastureland (level of deforestation 25%)-2
nd
repetition
P
a
50%
rainforest partially replaces pastureland (level of deforestation 50%)-1
st
repetition
P
b
50%
rainforest partially replaces pastureland (level of deforestation 50%)-2
nd
repetition
P
a
75%
rainforest partially replaces pastureland (level of deforestation 75%)-1
st
repetition
P
b
75%
rainforest partially replaces pastureland (level of deforestation 75%)-2
nd
repetition
P’ anomaly fields of precipitation between deforested (pasture and soybean) and
forested conditions
S
25%
rainforest partially replaces soybean (level of deforestation 25%)
S
50%
rainforest partially replaces soybean (level of deforestation 50%)
S
75%
rainforest partially replaces soybean (level of deforestation 75%)
Sr’ anomaly surface reflected radiation between deforested (pasture and soybean) and
forested conditions
Rn’ anomaly surface net radiation between deforested (pasture and soybean) and
forested conditions
C’ anomaly total cloud cover between deforested (pasture and soybean) and forested
conditions
ω’
anomaly vertical wind velocity between deforested (pasture and soybean) and
forested conditions
S
in
’ anomaly incoming radiation between deforested (pasture and soybean) and
forested conditions
S
in
incoming radiation
a
NR
linear coefficient, related to non-radiative processes
b
R
an angular coefficient, related to radiative processes
a
NSR
linear coefficient related to non-solar radiation processes (includes non-radiative
processes and long wave radiation feedbacks)
b
SR
an angular coefficient related to the solar radiation processes only
108
CHAPTER 6
GENERAL CONCLUSIONS
6.1. OVERVIEW
It is known that a more realistic representation of the land surface is important to
improve the climate simulated by General Circulation Models (GCMs). In particular, the
surface albedo is an important source of uncertainties related to the surface radiative
budgets. In the tropics, such as in Amazon tropical rainforest, where the solar radiation
balance is stronger, changes in surface albedo caused by deforestation have been
demonstrated to influence the regional climate (Nobre et al., 1991; Dirmeyer and Shukla,
1994; Hahmann and Dickinson, 1997; Costa and Foley, 2000; Berbet and Costa, 2003).
Hence, a more realistic representation of albedo in GCMs will significantly improve the
accuracy of climate simulation and prediction.
In this Dissertation, It is used measured field data, state-of-art numerical models
and remote sensing products to investigate the sources of spatial and temporal variation of
109
surface albedo of an Amazon tropical rainforest and the role of changes in surface albedo,
after deforestation, on the regional climate. To study this challenging subject, this
dissertation was divided in five chapters, where the conclusion of each chapter is
summarized below.
In chapter 1, it was studied the main sources of variation of surface albedo of
Amazonian tropical vegetation, at both the hourly and monthly time scales. In addition to
the traditional sources of variation (land cover and zenith angle) it was examined the role
of canopy wetness and atmospheric transmissivity on the surface albedo and the respective
interactions. Field data used in this analysis were collected at six micrometeorological sites
(four forests and two pasturelands) during ABRACOS (Anglo Brazilian Amazonian
Climate Observation Study) and LBA projects (Large-Scale Biosphere-Atmosphere
Experiment in Amazonia). At the hourly-scale, land cover variation (L) (forest, pasture),
atmospheric transmissivity (τ) and canopy wetness (ω) are the most important sources of
variation, while at monthly scale, only land cover and atmospheric transmissivity are
important. Reduction of 0.004 in the surface albedo due to canopy wetness is observed for
both ecosystems, forest and pastureland. The results presented here let me to conclude
that, although the difference in surface albedo (α') is partially dependent on τ (a
consequence of the interaction L·τ), it is independent of the canopy wetness. It is also
noted that, when simulating the land surface albedo, neither the role of atmospheric
transmissivity nor the role of canopy wetness is well represented in land surface models
today.
In chapter 2, it was evaluated the sensitivity of the surface albedo simulated by the
Integrated Biosphere Simulator (IBIS) to a set of tropical rainforest canopy architectural
and optical parameters. The parameters tested in this study are the orientation and
reflectance of the leaves of upper and lower canopies in the visible (VIS) and near-infrared
110
(NIR) spectral bands. The results were evaluated against albedo measurements taken above
the K34 site at the INPA (Instituto Nacional de Pesquisas da Amazonia) Cuieiras
Biological Reserve. The sensitivity analysis indicates a strong response to the upper
canopy leaves orientation (χ
up
) and to the reflectivity in the near-infrared spectral band
(ρ
NIR,up
), a smaller sensitivity to the reflectivity in the visible spectral band (ρ
VIS,up
) and no
sensitivity at all to the lower canopy parameters, which is consistent with the canopy
structure. The combination of parameters that minimized the RMSE and mean relative
error are χ
up
= 0.86, ρ
VIS,up
= 0.062, and ρ
NIR,up
= 0.275. These parameterizations allow the
improvement of the accuracy of the land surface albedo simulations obtained by IBIS
model, indicating its potential to simulate the canopy radiative transfer for narrow spectral
bands, allowing in the future, a close comparison with remote sensing products.
In chapter 3, the role of canopy wetness on the simulated albedo of Amazon
tropical rainforest was investigated, once it is the third most important source of variation
at hourly scale. Simulations were run using three versions (0-D) of the land
surface/ecosystem IBIS model. The results demonstrated that, while the incorporation of
canopy wetness on the radiative transfer calculation improves the simulated surface
albedos at hourly time scale, no substantial improvement was verified at monthly or longer
time scales, because radiative effect of the canopy wetness is restricted to very short times,
when the canopy is actually wet. However, these results exclude the role canopy wetness
as a main source of seasonal variability of tropical rainforest albedo
In chapter 4, six different products of land surface albedo derived from different
remote sensors are compared with the albedo simulated by the Community Climate Model
(CCM3) coupled to the Integrated Biosphere Simulator (IBIS). Field measurements
collected at three sites in Amazon are also included as reference data. The results presented
in this chapter let me to conclude that the albedos estimated from different remote sensing
111
systems data vary considerably in the range between 0.10 and 0.20, and there are
substantial differences in seasonality. Field measurements and modeled albedo vary
seasonally from 0.11 to 0.14, and only the black-sky product albedo from MODIS matched
to the field-observed albedo seasonality.
In chapter 5, eleven simulations using the CCM3-IBIS were run to investigate the
radiative processes of the precipitation change after deforestation and its effects on the
regional climate in Amazon, considering different levels of deforestation (25%, 50% and
75%), and different types of land cover replacements (pasture and soybean) for the original
rainforest land cover. Results show that the changes in precipitation are linearly related to
the anomaly in net surface radiation, and the change in precipitation due to radiative
mechanisms may be an order of magnitude higher than those due to non-radiative
mechanisms. Radiative processes dominate because precipitation in the region is largely
convective. A reduction in net radiation weakens the convective activity, reducing
cloudiness and precipitation. While cloud-radiative negative feedbacks inhibit a larger
climate change* (precipitation decrease), competitive non-radiative mechanisms tend to
increase precipitation.
6.2. CONCLUSIONS
An overview on the results and individual conclusions of each chapter lead me to
argue that the regional climate of Amazon is strongly controlled by the radiative processes
changes in surface albedo and net radiation – caused by changes in land cover due to the
replacement of forest in pasturelands and croplands – are fundamental, explaining over
90% of the variance in the precipitation change.
* The term climate change refers only to the precipitation change.
112
The degree of importance of each source of variation of surface albedo found in
Chapter 1 and the parameterization of the model obtained from the sensitivity analyses in
Chapter 2 are important to define priority improvements in the representation of albedo in
GCMs, once the state-of-the-art land surface models are still unable to reproduce
realistically the seasonal variability of surface albedo of forested and deforested areas
(Bebert and Costa, 2003). The results found in this dissertation, show that land cover and
atmospheric transmissivity are the two most important sources of surface albedo variance
and are in agreement with the significantly differences between surface albedos of forest
and pasture found by several authors (Oguntoyinbo, 1970; Shuttleworth et al., 1984;
Bastable et al., 1983 and Culf et al., 1996) and with the influence of zenith angle and
interaction between atmospheric transmissivity and zenith angle on surface albedo found
by Pinker et al. (1980), Shuttleworth et al. (1984), McCaughey (1987) and Giambelluca et
al. (1999).
A third source of variation of surface albedo found in this dissertation is the canopy
wetness. Although Culf et al. (1995) suggested that surface albedo seasonality is related to
soil moisture, Berbet and Costa (2003) suggested that most likely the changes in forest
albedo are related to soil moisture-correlated variables: smaller soil exposure, darker leaves
(associated with the leaf water potential) and higher canopy wetness. Simulations using the
IBIS model with changes in the radiative transfer code to incorporate the radiative effects
of the canopy wetness show that canopy wetness is not evidenced as a possible source of
variation of surface albedo at seasonal-scale time, excluding the role canopy wetness as a
main source of seasonal variability of tropical rainforest albedo as evidenced in the
Chapter 3. However, these modifications in the IBIS code reproduced well the significant
decrease in surface albedo during the precipitation hours, confirming the results found in
Chapter 1.
113
Land surface parameterizations have been poorly represented because of limited
observations (Zhou et al., 2003), and the only practical and economical approach to large-
scale characterization of land surface parameters is remote sensing (Hu et al., 2000).
Although the unquestionable importance of remote sensing as key tool to study the global
environment and its evolution throughout the last decades, in Chapter 4 of this dissertation
it is found that satellite products for the Amazon tropical rainforests present severe
limitations, due to the significant difference among the six different remote sensing
systems studied. For the Amazon region, the albedos estimated vary considerable in the
range between 0.10 and 0.20, and there are substantial differences in seasonality.
Numerous studies with GCMs suggest that tropical deforestation can result in
regional-scale climate change, increasing particularly reducing precipitation. Simulation
results from the CCM3 GCM (Chapter 5) show that an increase in the surface albedo
(caused by the replacement of forest by pastureland and soybean cropland) reduces energy
balance, convection and cloud formation generating an anomalous subsidence motion
resulting in the reduction of rainfall over the Amazon. These results agree with most of
investigations about the effect of deforestation over Amazon basin (Nobre et al., 1991;
Eltahir, 1996; Werth and Avissar, 2002; Hoffmann et al., 2003). The cloud reduction,
however, increases the incoming radiation, introducing a negative feedback on the
radiative process. The final precipitation decrease is smaller in the presence of the cloud
feedback than it would be in their absence.
114
6.3. RECOMMENDATIONS FOR FUTURE RESEARCH
The results presented in this dissertation elucidated some points related to the
sources of spatial and temporal variation of an Amazon tropical rainforest surface albedo,
and to the role of surface albedo changes, after Amazon tropical deforestation, in the
regional climate. It, however, provoked several new lines of research:
•
Following the discussion in Chapter 1 and by Culf et al. (1995) and by Berbet and Costa
(2003), I recommend that future work should investigate the role of senescence, leaf age
and leaf water potential on the seasonality of the tropical rainforest albedo. It is
expected with the inclusion of these factors and other additional sources of variation in
the land surface models, further improvements in accuracy providing more detailed
studies of the climatic effects of tropical deforestation. A specific field experiment setup
is needed, though.
•
Future additional parameterizations of the canopy architectural (e.g., single-sided leaf
and stem area indexes, fraction of overall area covered by lower and upper canopies)
and optical (e.g., direct and diffuse beams ground albedo on visible and near-infrared
spectral bands, lower and upper canopies leaf transmittances on visible and near-
infrared spectral bands) parameters according to the plant species, soil types, plant
phenology, leaf water content, and soil surface wetness may improve considerably the
scope of such modeling exercises, building a solid basis for stronger interactions
between field observations, climate models and remote sensing products.
•
More robust algorithms to retrieve the surface albedos from remote sensing products
should be developed and tested to reduce the uncertainties related to the effects of
atmospheric scattering and absorption, anisotropy, inadequate temporal, spatial and
spectral sampling, and narrowband to broadband conversion.
115
•
Further simulations involving more realistic climate change should be performed
considering realistic scenarios of agriculture expansion in Amazonia.
•
It is also important to investigate the non-radiative climate processes due to
deforestation. It is also important to built new datasets for soybean cropland
parameterizations, which nowadays have been the crop with more expansion in Amazon
region.
116
GENERAL REFERENCES
André, R.G.B. and Y. Viswanadhan, 1983: Radiation balance of soybeans grown in
Brazil.
Agricultural Meteorology, 30, 150-173.
Araújo, A. C., A. D. Nobre, B. Kruijt, J. A. Elbers, R. Dallarosa, P. Stefani, C. von
Randow, A. O. Manzi, A. D. Culf, J. H. C. Gash, R. Valentini, P. Kabat, 2002:
Comparative measurements of carbon dioxide fluxes from two nearby towers in a
central Amazonian rainforest: The Manaus LBA site,
J. Geophys. Res., 107: D8090,
doi:10.1029/2001JD000676.
Asner, G. P., 1998: Biophysical and biogeochemical sources of reflectance in
canopy reflectance.
Remote sens. Environ., 64, 234-253.
Asner, G. P., C. A. Wessman, C. A. Bateson, and J. L. Privette, 2000: Impact of tissue,
canopy, and landscape factors on the hyperspectral reflectance variability of arid
ecosystems.
Remote sens. Environ., 74, 69–84.
Barkstrom, B. R., 1984: The Earth Radiation Budget Experiment (ERBE),
Bull. Amer.
Meteorol. Soc.
, 65, 1170-1185.
Bastable, H.G., W. J. Shuttleworth, R. L. G. Dallarosa, G. Fisher, and C. A. Nobre, 1993:
Observations of climate, albedo and surface radiation over cleared and undisturbed
Amazonia forest.
Int. J. Climatol., 13, 783-796.
Bégué, A., J. L. Roujean, N. P. Hanan, S. D. Prince, M. Thawley, A. Huete, D. Tanré,
1996: Shortwave radiation budget of Sahelian vegetation: 1. Techniques of
measurement and results during HAPEX – Sahel.
Agr. Forest Meteorol., 79, 79-96.
Berbet, M. L. C. and M. H. Costa, 2003: Climate change after tropical deforestation:
seasonal variability of surface albedo and its effects on precipitation change,
J. Climate,
16, 2099-2104.
117
Biajoli, M. C., 2006: Relatório final de atividades referentes à Bolsa FUNCATE, Categoria
Formação em C&T – Tipo II.
UFV, p.12.
Blad, B. L. and D. G. Baker, 1972: Reflected radiation from a soybean crop.
Agron. J., 64,
277-280.
Bonan, G. B., 1996:
A Land surface model (LSM version 1.0) for ecological, hydrological,
and atmospheric studies: technical description and user’s guide
. NCAR Technical note
TN-417+STR, 150p.
Botta, A. and J. A. Foley, 2002: Effect of climatic variability and disturbances on the
Amazonian terrestrial ecosystems dynamics.
Global Biogeochem. Cy., 16 (4), doi:
10.1029/2000GB001338.
Brest, C. L. and W. B. Rossow, 1992: Radiometric calibration and monitoring of NOAA
AVHRR data for ISCCP.
Int. J. Remote Sens., 13, 235–273.
Brest, C., W. B. Rossow, and M. D. Roiter, 1997: Update of Radiance Calibrations for
ISCCP,
J. Atmos. Ocean. Tech., 14, 1091-1109.
Briegleb, B. P., P Minnis, V. Ramanathan, and E. Harrison, 1986: Comparison of regional
clear-sky albedos inferred from satellite observations and model computations.
J.
Climate Appl. Meteor
., 25, 214-226.
Charney, J. G., 1975: Dynamics of deserts and drought in the Sahel.
Q. J. Roy. Meteor.
Soc.
, 101, 193-202.
Charney, J. G., W. J. Quirk, S. H. Chow, and J. Kornifield, 1977: A comparative study of
the effects of albedo change on drought in semi-arid regions
, J. Atmos. Sci., 34, 1366-
1385.
Chu, P. S., Z. R. Yu, and S. Hastenrath, 1994: Detecting climate change concurrent with
deforestation in the Amazon Basin: Which way has it gone?
Bull. Amer. Meteor. Soc.,
75, 579-582.
118
Coakley, J. A. and P. Chylek, 1975: The two-stream approximation in radiative transfer:
Including the angle of incident radiation.
J. Atmos. Sci., 32: 249-261.
Cochran, W. G. and G. M. Cox, 1992:
Experimental designs. John Wiley and Sons, 2
st
ed.,
611p.
Costa, M. H. and J. A. Foley, 1998: A comparison of precipitation datasets for the Amazon
basin.
Geophys. Res. Lett., 25, 155–158.
Costa, M. H. and J. A. Foley, 2000: Combined effects of deforestation and doubled
atmospheric CO2 concentrations on the climate of Amazonia.
J. Climate, 13, 18–34.
Csiszar, I. and G. Gutman, 1999: Mapping global land surface albedo from
NOAA/AVHRR.
J. Geophys. Res., 104, 6215-6228.
Culf, A. D., G. Fisch, and M. G. Hodnett, 1995: The albedo of Amazonian forest and ranch
land.
J. Climate, 8, 1544-1554.
Culf, A. D., J. L. Esteves, A. O.Marques Filho, H. R. Rocha, 1996: Radiation, temperature
and humidity over forest and pasture in Amazonia. In: Gash, J. H. C., Nobre, C. A.,
Roberts, J. M., Victoria, R. L.
Amazonian deforestation and climate. New York: John
Wiley & Sons, 175-191.
Cutrim, E., D. W. Martin, and R. Rabin, 1995: Enhancement of cumulus clouds over
deforested lands in Amazonia.
Bull. Amer.Meteor. Soc., 76, 1801-1805.
Delire, C. and J. A Foley, 1999: Evaluating the performance of a land surface/ ecosystem
model with biophysical measurements from contrasting environments.
J. Geophys.
Res.,
104, 16895-16909.
Delire, C., J. A. Foley, S. Thompson, 2002: Evaluating the carbon cycle of a coupled
atmosphere-biosphere model.
Global Biogeochemical Cycles, 17, article n° 1012.
119
Desormeaux, Y., W. B. Rossow, C. L. Brest, and G. G. Campbell, 1993: Normalization
and calibration of geostationary satellite radiances for ISCCP.
J. Atmos. Ocean. Tech.,
10, 304–325.
Dickinson, R. E. and A. Henderson-Sellers, 1988: Modelling tropical deforestation: A
study of GCM land-surface parametrizations.
Q. J. Roy. Meteor. Soc., 144, 439-462.
Dickinson, R. E. and P. Kennedy, 1992: Impacts on regional climate of Amazon
deforestation.
Geophys. Res. Lett., 19, 1947–1950.
Dickinson, R. E., 1983: Land surface processes and climate-surface albedos and energy
balance.
Adv. Geophys., 25:305-353.
Dirmeyer, P. A. and J. Shukla, 1994: Albedo as a modulator of climate response to tropical
deforestation.
J. Geophys. Res., 99, 20863-20877.
Dorman, J. L. and P. J. Sellers, 1989: A global climatology of albedo, roughness
length and stomatal resistance for atmospheric general circulation models as
represented by the Simple Biosphere Model (SiB).
Journal of Applied Meteorology, 28,
833-855.
Durieux, L., L. A. T. Machado, and H. Laurent, 2003: The impact of deforestation on
cloud cover over the Amazon arc of deforestation.
Remote Sens. Environ., 86, 132-140.
Eck, T. F. and D. W. Deering, 1992: Canopy albedo and transmittance in a spruce-hemlock
forest in mid-September.
Agr. For. Meteorol., .59, 237-248.
Eltahir, E. A. B., 1996: Role of vegetation in sustaining large-scale atmospheric
circulations in the tropics.
J. Geophys. Res., 101, 4255-4268.
Fisch, G., I. R. Wright, and H. G. Bastable, 1994: Albedo of tropical grass: A case study of
pre- and post-burning.
Int. J. Climatol., 14, 102–107.
120
Foley, J. A., A. Botta, M. T. Coe, and M. H. Costa, 2002: El Niño-Southern Oscillation
and the climate, ecosystems and rivers of Amazonia.
Global Biogeochem. Cy, 16,
1132. doi: 10.1029/2002GB001872.
Foley, J. A., I. C. Prentice, N. Ramankutty, S. Levis, D. Pollard, S. Sitch, and A. Haxeltine,
1996: An integrated biosphere model of land surface processes, terrestrial carbon
balance and vegetation dynamics.
Global Biogeochem. Cy., 10, 603-628.
Fontana, D. C., M. A. Berlato, and H. Bergamaschi, 1991: Soybean radiation balance in
subtropical region,
Pesquisa Agropecuaria Brasileira, 26, 411-418 (in portuguese with
abstract in English).
Giambelluca, T.W., J. Fox, S. Yarnasarn, P. Onibutr, and M. A. Nullet, 1999: Dry-season
radiation balance of land covers replacing forest in northern Thailand.
Agr. Forest.
Meteorol.
, 95, 53-65.
Hahmann, A. N. and R. E. Dickinson, 1997: RCCM2/BATS Model over Tropical South
America: Applications to Tropical Deforestation,
J. Climate, 10, 1944-1964.
Henderson-Sellers, A., R. E. Dickinson, T. B. Durbidge, P. J. Kennedy, K. McGuffie, and
A. J. Pitman, 1993: Tropical deforestation: Modeling local to regional scale climate
change.
J. Geophys. Res., 98, 7289–7315.
Hoffmann, W. A., W. Schroeder, and R. Jackson, 2003: Regional feedbacks among fire,
climate, and tropical deforestation.
J. Geophys. Res., 108, 4721,
doi:10.1029/2003JD003494.
Hu, B., W. Lucht, A. Strahler, C. Schaaf, and M. Smith, 2000: Surface albedos and angle-
corrected NDVI from AVHRR observations of South America,
Remote Sens. Environ.,
71,119-132.
Hucek, R. and H. Jacobowitz, 1995: Impact of scene dependence on AVHRR albedo
models.
J. Atmos. Ocean. Tech. 12, 697-711.
121
Jin, Y., C. B. Schaaf, C. E. Woodcock, F. Gao, X. Li, A. H. Strahler, W. Lucht, and S.
Liang, 2003: Consistency of Moderate-Resolution Imaging Spectroradiometer surface
bidirectional reflectance distribution function and albedo retrievals, 2, Validation,
J.
Geophys. Res.
, 108(D5), 4159, doi:10.1029/2002JD002804.
Justice, C. O., R. E. Wolfe, N. El Saleous, J. Descloitres, E. Vermote, D. Roy, J. Owens,
and E. Masuoka, 2000: The availability and status of MODIS land products. The Earth
Observer, 12 (6), 10– 18.
Kiehl, J. T., J. J. Hack, G. B. Bonan, B. A. Boville, D. L. Williamson, and P. J. Rasch,
1998: The National Center for Atmospheric Research Community Climate Model:
CCM3,
J. Climate, 11, 1131-1149.
Klein, S. A. and D. L. Hartmann, 1993: Spurious trend in the ISCCP C2 dataset.
Geophys.
Res. Lett.,
20, 455–458.
Kondratyev, K. Y., 1972:
Radiation processes in the atmosphere, No. 309. Geneva, WMO.
Kucharik, C. J., J. A. Foley, C. Delire, V. A. Fisher, M. T. Coe, S. T. Gower, J. D. Lenters,
C. Young-Molling, J. M. Norman, and N. Ramankutty, 2000: Testing the performance
of a dynamic global ecosystem model: Water balance, carbon balance end vegetation
structure.
Global Biogeochem. Cy., 14, 795-825.
Lean, J. and D. A. Warrilow, 1989: Simulation of the regional climatic impact of Amazon
deforestation.
Nature, 342, 411-413.
Lean, J. and P. R. Rowntree, 1997: Understanding the sensitivity of a GCM simulation of
Amazonia deforestation to the specification of vegetation and soil characteristics.
J.
Climate
, 10, 1216-1235.
Lean, J., Rowntree, P. R., 1993: A GCM simulation of the impact of Amazonian
deforestation on climate using an improved canopy representation.
Q. J. Roy. Meteor.
Soc.,
119, 509–530.
122
Lewis, P., M. I. Disney, M. J. Barnsley, and J. -P. Muller, 1999: Deriving albedo for
HAPEX-Sahel from ASAS data using kernel-driven BRDF models,
Hydrol. Earth Syst.
Sc.
, 3, 1-13.
Li, Z. and L. Garand, 1994: Estimation of surface albedo from space: A parametrization for
global application.
J. Geophys. Res., 99, 8335-8350.
Lucht, W., A. H. Hyman, A. H. Strahler, M. J. Barnsley, P. Hobson, and J.-P. Muller,
2000a: A comparison of satellite-derived spectral albedos to ground-based broadband
albedo measurements modeled to satellite spatial scale for a semi-desert landscape,
Remote Sens. Environ., 74, 85-9.
Lucht, W., C. B. Schaaf, and A. H. Strahler, 2000b: An algorithm for the retrieval of
albedo from space using semiempirical BRDF models.
IEEE T. Geosci. Remote Sens.,
38,977– 38,998.
Manzi, O., and S. Planton, 1996: Calibration of a GCM using ABRACOS and ARME data
and simulation of Amazonian deforestation. In: Gash, J.H.C., Nobre, C.A., Roberts,
J.M., Victoria, R.L. (Eds.),
Amazonian Deforestation and Climate, 1st ed. John Wiley
and Sons, Chichester, 505–530.
McCaughey, J. H., 1987: The albedo of a mature mixed forest and a clear-cut site at
Petawawa, Ontario.
Agr. Forest. Meteorol., 40, 251-263.
McWilliam, A-L.C., O. M. R. Cabral, B. M. Gomes, J. L. Esteves, and J. M. Roberts,
1996: Forest and pasture leaf-gas exchange in south-west Amazonia. In: Gash, J.H.C.,
Nobre, C.A., Roberts, J.M., Victoria, R.L. (Eds.), Amazonian Deforestation and
Climate, 1st ed. John Wiley and Sons, Chichester, 265-286.
Meador, W. E. and W. R. Weaver, 1980: Two-stream approximations to radiative transfer
in planetary atmospheres: a unified description of existing methods and a new
improvement.
J. Atmos. Sci., 37: 630-643.
123
Medina, E. and E. Cuevas, 1996: Biomass production in nutrient limited rainforest:
implications for responses to global change. In: Gash, J.H.C., Nobre, C.A., Roberts,
J.M., Victoria, R.L. (Eds.),
Amazonian Deforestation and Climate, 1st ed. John Wiley
and Sons, Chichester, pp. 221–239.
Negri, A. J., R. F. Adler, L. Xu, and J. Surratt, 2004: The Impact of Amazonian
Deforestation on Dry Season Rainfall.
J. Climate, 17, 1306-1319.
Nobre, C. A., G. Fisch, H. R. Rocha, R. F. Lyra, E. P. Rocha, A. C. L. Costa, and V. N.
Ubarana, 1996: Observations of the atmospheric boundary layer in Rondônia, In: Gash,
J.H.C., Nobre, C.A., Roberts, J.M., Victoria, R.L. (Eds.),
Amazonian Deforestation and
Climate
, 1st ed. John Wiley and Sons, Chichester, pp. 413-424.
Nobre, C., P. Sellers, and J. Shukla, 1991: Amazonian deforestation and regional climate
change,
J. Climate, 4, 957-988.
Norman, J. M. and P.G. Jarvis, 1975: Photosynthesis in Sitka Spruce. V. Radiation
penetration theory and a test case.
J. Appl. Ecol., 12:839-878.
Nouvellon, Y., A. Bégué, M. S. Moran, D. L. Seen, S. Rambal, D. Luquet, G. Chehbouni,
and Y. Inoue, 2000: PAR extinction in shortgrass ecosystems: effects of clumping, sky
conditions and soil albedo.
Agr. Forest Meteorol., 105, 21-41.
Oguntoyinbo, J. L., 1970: Reflection coefficient of natural crops and urban surfaces in
Nigeria.
Q. J. Roy. Meteor. Soc., 96, 430-441
Oleson, K.W., G. B. Bonan, C. Schaaf, F. Gao, Y. Jin, and A. Strahler, 2003: Assessment
of global climate model land surface using MODIS data,
Geophys. Res. Lett., 30,
doi:10.1029/2002GL016749.
Oleson, K.W.; Y. Dai, G. B. Bonan, M. Bosilovich, R. Dickinson, P. Dirmeyer, F.
Hoffman, P. Houser, S. Levis, G-Y. Niu, P. Thornton, M. Vertenstein, Z.-L. Yang, and
124
X. Zeng, 2004: T
echnical description of the Community Land Model (CLM). NCAR
Technical note TN-461+STR, 174p.
Pinker, R. T. and I. Laszlo, 1992: Modeling surface solar irradiance for satellite
applications on a global scale,
J. Appl. Meteor., 31, 194-211.
Pinker, R. T., O. E. Thompson, and T. F. Eck, 1980: The albedo of a tropical evergreen
forest.
Q. J. Roy. Meteor. Soc., 106, 551-558.
Pinker, R., 2002: Satellite Estimates of Radiative Fluxes in the Western Hemisphere.
College Park, Maryland: The Global Land Cover Facility.
Pinty, B., T., R. E. Lavergne, J.-L Dickinson, N. Widlowski, Gobron, and M. M.
Verstraete, 2006: Simplifying the interaction of land surfaces with radiation for relating
remote sensing products to climate models,
J. Geophys. Res., 111, D02116,
doi:10.1029/2005JD005952.
Pollard, D. and S. Thompson, 1995: The effect of doubling stomatal resistance in a global
climate model.
Global Planat. Change, 10: 129-161.
Rao, C.R.N. and J. Chen, 1995: Inter-satellite calibration linkages for the visible and near-
infrared channels of the Advanced Very High Resolution Radiometer on the NOAA-7, -
9 and –11 spacecraft.
Int. J. Remote Sens., 16, 1931-1942.
Rao, C. R. N. and J. Chen, 1996: Post-launch calibration of the visible and near-infrared
channels of the Advanced Very High Resolution Radiometer on the NOAA-14
spacecraft.
Int. J. Remote Sens., 17, 2743-2747.
Roberts, J. M., O. M. R. Cabral, J. P. Costa, A-L.C. McWilliam, and T. D. A. Sá, 1996: An
overview of the leaf area index and physiological measurements during ABRACOS. In:
Gash, J.H.C., Nobre, C.A., Roberts, J.M., Victoria, R.L. (Eds.),
Amazonian
Deforestation and Climate
, 1st ed. John Wiley and Sons, Chichester, UK. pp 287-306.
125
Roberts, J. M., O. M. R.Cabral, and L. F. Aguiar, 1990: Stomatal and boundary layer
conductances measured in a terra firme rain forest.
J. Appl. Ecol. 27, 336-353.
Ross, J. 1981:
The Radiation Regime and Architecture of Plant Stands, 391 pp., Springer,
New York.
Rossow, W. B. and B. Cairns, 1995: Monitoring changes of clouds.
Climatic Change, 31,
175–217.
Rossow, W. B., A. Walker, D. Beuschel, and M. Roiter, 1996: International Satellite Cloud
Climatology Project (ISCCP) documentation of new cloud datasets. World Climate
Research Programme (ICSU and WMO), WMO/TD-737 Geneva, Switzerland, 115 pp.
[Available from ISSCP Global Processing Center, NASA GSFC Institute for Space
Studies, 2880 Broadway, New York, NY 10025.]
Rossow, W. B., Y. Desormeaux, C. L. Brest, and A. Walker, 1992: International Satellite
Cloud Climatology Project (ISCCP) radiance calibration report. World Climate
Research Programme (ICSU and WMO), WMO/TD-520 Geneva, Switzerland, 104 pp.
[Available from ISSCP Global Processing Center, NASA GSFC Institute for Space
Studies, 2880 Broadway, New York, NY 10025.].
Rowe, C. M., 1991: Modeling land-surface albedos from vegetation canopy architecture.
Phys. Geogr., 12(2), 93– 114.
SAS, 2001: User’s guide (Release 8.2) Statistical Analysis System Institute, Inc., Cary,
NC.
Schaaf, C. B., F. Gao, A. H. Strahler, W. Lucht, X. Li, T. Tsang, N. C. Strugnell, X.
Zhang, Y. Jin, J. P. Muller, P. Lewis, M. Barnsley, P. Hobson, M. Disney, G. Roberts,
M. Dunderdale, C. Doll, R. P. D’entremont, B. Hu, S. Liang, J. L. Privette, and D. Roy,
2002: First operational BRDF, albedo nadir reflectance products from MODIS.
Remote
Sens. Environ.
, 83, 135–148.
126
Sellers, P. J., 1985: Canopy reflectance, photosynthesis and transpiration.
Int. J. Remote
Sens.,
6:1335-1372.
Sellers, P. J., Y. Mintz, Y. Sud, and A. Dalcher, 1986: A simple biosphere model (SiB) for
use within general circulation models.
J. Atm. Sci. 43, 505–531.
Shuttleworth, W.J., J. H. C. Gash, C. R. Lloyd, J. M. Roberts, A. O. Marques, G. Fisch, P.
Silva, M. N. G. Ribeiro, L. C. B. Molion, L. D. A. Sá, C. A. Nobre, O. M. R. Cabral, S.
R. Patel, and J. C. Moraes, 1984: Observations of radiation exchange above and below
Amazonian forest.
Q. J. Roy. Meteor. Soc., 110, 1163-1169.
Slingo, J. M., 1987: The development and verification of a cloud prediction scheme for the
ECMWF model,
Q. J. R. Meteorol. Soc., 113, 899–927.
Smith, G. L., Z. P. Szewczyk, D. A. Rutan, and B. L. Robert, 2006: Comparison of
measurements from satellite radiation budget instruments
J. Geophys. Res., 111,
D04101, doi:10.1029/2005JD006307.
Song, J., 1998: Diurnal asymmetry in surface albedo.
Agr. Forest. Meteorol., 92, 181-189.
Stackhouse, P. W., S. J. Cox, S. K. Grupta, J. C. Mikovitz, and M. Chiacchio, 2003: The
WCRP/GEWEX Surface Radiation Budget Data set: A 1 degree resolution, 12 year flux
climatology.
Stackhouse, P. W., S. K. Grupta, S. J. Cox, M. Chiacchio, and J. C. Mikovitz, 2000. The
WCRP/GEWEX Surface Radiation Budget Project release 2: An assessment of surface
fluxes a 1 degree resolution. In IRS 2000:
Current problems in atmospheric radiation,
W. L. Smith and Y. M. Timofeyev, Eds.,
International Radiation Symposium, St.
Petersburg, Russia, July 24-29.
Sud, Y. C., G. K. Walker, J.-H. Kim, G. E. Liston, P. J. Sellers, and W. K.-M. Lau, 1996:
Biogeophysical consequences of a tropical deforestation scenario: A GCM simulation
study.
J. Climate, 9, 3225–3247.
127
Suttles, J. T., R. N. Green, P. Minnis, G. L. Smith, W. F. Staylor, B. A. Wielicki, I. J.
Walker, D. F. Young, V. R. Taylor, and L. L. Stowe, 1988: Angular
radiation models
for Earth atmosphere system.
Volume I – Shortwave radiation. NASA Reference
Publication 1184. 147pp.
Verstraete, M. M., 1987: Radiation transfer in plant canopies: Transmission of direct solar
radiation and the role of leaf orientation.
J. Geophys. Res., 92:10985-10995.
Voldoire, A. and J. F. Royer, 2004: Tropical deforestation and climate variability.
Clim.
Dyn.
, 22, 857-874.
Wei, X., A. N. Hahmann, R. E. Dickinson, Z. L. Yang, X. B. Zeng, K. J. Schaudt, C. B.
Schaaf, and N. Strugnell, 2001: Comparison of albedo computed by land surface
models and evaluation against remotely sensed data.
J. Geophys. Res., 106, 20687-
20702.
Werth, D. and R. Avissar, 2002: The local and global effects of Amazon deforestation,
J.
Geophys. Res
., 107, doi:10.1029/2001JD000717.
Wielicki, B. A. and R. N. Green, 1989: Cloud identification for ERBE radiative flux
retrieval
. J. Appl. Meteorol., 28, 1133-1146.
Wright, I. R., C. A. Nobre, J. Tomasella, H. R. Rocha, J. m. Roberts, E. Vertamatti, A. D.
Culf, R. C. S. Avalá, and M. G. Hodnett, 1996: Towards a GCM surface
parameterization for Amazonia. In: Gash, J.H.C., Nobre, C.A., Roberts, J.M., Victoria,
R.L. (Eds.),
Amazonian Deforestation and Climate, 1st ed. John Wiley and Sons,
Chichester, 473-504.
Xue, K., N. Liou, and A. Kasahara, 1990: Investigation of the biogeophysical feedback on
the African climate using a two-dimensional model.
J. Climate, 3, 337–352.
128
Yanagi, S. N. M. and M. H. Costa, 2006. Modeling radiative transfer in tropical rainforest
canopies: Sensitivity of simulated albedo to canopy Architectural and Optical
Parameters. Submitted to
Acta Amazonica.
Zao, M., A. J. Pitman, and T. N. Chase, 2001: The impact of land cover change on the
atmospheric circulation.
Clim. Dynam., 17, 467-477.
Zeng, N. and J. D. Neelin, 1999: A land-atmosphere interaction theory for the tropical
deforestation problem.
J. Climate, 12, 857-872.
Zeng, N., R. E. Dickinson, and X. Zeng, 1996: Climatic impact of Amazon deforestation—
A mechanistic model study.
J. Climate, 9, 859–883.
Zhang, H., K. McGuffie, and Henderson-Sellers, A., 1996: Impacts of tropical
deforestation, part II, The role of large-scale dynamics,
J. Climate, 9, 2498– 2521.
Zhou, L., R. E. Dickinson, Y. Tian, X. Zeng, Y. Dai, Z. –L. Yang, C. B. Schaaf, F. Gao,;
Y. Jin, A. Strahler, R. B. Myneni, H. Yu, W. Wu, and M. Shaikh, 2003: Comparison of
seasonal and spatial variations of albedos from Moderate Resolution Imaging
Spectroradiometer (MODIS) and Common Land Model.
J. Geophys. Res., 108, doi:
10.1029/2002JD003326
Livros Grátis
( http://www.livrosgratis.com.br )
Milhares de Livros para Download:
Baixar livros de Administração
Baixar livros de Agronomia
Baixar livros de Arquitetura
Baixar livros de Artes
Baixar livros de Astronomia
Baixar livros de Biologia Geral
Baixar livros de Ciência da Computação
Baixar livros de Ciência da Informação
Baixar livros de Ciência Política
Baixar livros de Ciências da Saúde
Baixar livros de Comunicação
Baixar livros do Conselho Nacional de Educação - CNE
Baixar livros de Defesa civil
Baixar livros de Direito
Baixar livros de Direitos humanos
Baixar livros de Economia
Baixar livros de Economia Doméstica
Baixar livros de Educação
Baixar livros de Educação - Trânsito
Baixar livros de Educação Física
Baixar livros de Engenharia Aeroespacial
Baixar livros de Farmácia
Baixar livros de Filosofia
Baixar livros de Física
Baixar livros de Geociências
Baixar livros de Geografia
Baixar livros de História
Baixar livros de Línguas
Baixar livros de Literatura
Baixar livros de Literatura de Cordel
Baixar livros de Literatura Infantil
Baixar livros de Matemática
Baixar livros de Medicina
Baixar livros de Medicina Veterinária
Baixar livros de Meio Ambiente
Baixar livros de Meteorologia
Baixar Monografias e TCC
Baixar livros Multidisciplinar
Baixar livros de Música
Baixar livros de Psicologia
Baixar livros de Química
Baixar livros de Saúde Coletiva
Baixar livros de Serviço Social
Baixar livros de Sociologia
Baixar livros de Teologia
Baixar livros de Trabalho
Baixar livros de Turismo
