MULTIPLE BIOME-CLIMATE EQUILIBRIA IN AMAZONIA AND PERSPECTIVES FOR THE FUTURE OF THE RAINFOREST. Carlos A. Nobre

Transcrição

1 MULTIPLE BIOME-CLIMATE EQUILIBRIA IN AMAZONIA AND PERSPECTIVES FOR THE FUTURE OF THE RAINFOREST Carlos A. Nobre CPTEC/INPE, Cachoeira Paulista, SP- Brazil

2 Environmental Changes in Amazonia and the Hypothesis of Savannization Carlos A Nobre, Marcos Oyama, Manoel Cardoso, Gilvan Sampaio, Luis Salazar, David Lapola, and Marcos Costa Slide courtesy: IPAM

3 Focus on Amazonia

4

5 Vegetation-Climate Interactions Climate Vegetation Bidirectional on various times scales

6 Does vegetation matter for the Earth System? Is vegetation distribution a fingerprint of climate? or Does vegetation distribution influence and participate to climate (state and changes)?

7 Does vegetation matter for the Earth System? The impacts of human activity on the Amazon rainforest could result in the collapse of large portions of the rainforest and significant loss of biodiversity within 30 to 50 years. A comparison is made with similar events in the Saharan ecosystem, which was once a region of richer vegetation, before its abrupt collapse about 6000 years ago

8 Remove vegetation from the continents large changes will happen in the water cycle Kleidon et al. (2000)

9 Remove vegetation from the continents large changes will happen in the water cycle Land=desert Land=forest Kleidon et al. (2000)

10 Remove vegetation from the continents large changes will happen in the water cycle Land=desert Land=forest ATMOSPHERE OCEANS CONTINENTS Kleidon et al. (2000)

11 Remove vegetation from the continents large changes will happen in the water cycle Land=desert Land=forest ATMOSPHERE km km km km 3 OCEANS km 3 CONTINENTS Kleidon et al. (2000)

12 Remove vegetation from the continents large changes will happen in the water cycle Land=desert Land=forest ATMOSPHERE km km km km km km km km 3 OCEANS km km 3 CONTINENTS Kleidon et al. (2000)

13 Vegetation partitions net radiation into more latent and less sensible heat (figure taken from Kabat et al.: Vegetation, Water, Humans, and the Climate, IGBP BAHC)

14 The ecosystems of Amazonia are subjected to a suite of environmental drivers of change LUCC Fire Climate Change Climate Extremes

15 Balanço Hídrico P = E + C P = Precipitação E = Evapotranspiração C = Convergência de Umidade E P C

16 Modeling Deforestation and Biogeography in Amazonia Current Biomes Post-deforestation 1 Tropical Forest 6 Savanna Nobre et al. 1991, J. Climate

17 Biome-Climate Bi-Stability for the Sahel The second equilibriun state depend mostly on vegetation (albedo) feedback and secondarily on ocean feedbacks SCHEFFER EL AL., NATURE VOL OCTOBER 2001 Current State Second State

18 Figure 6 Over the past 9,000 years, average Northern Hemisphere summer insolation (upper panel) has varied gradually owing to subtle variation in the Earth's orbit. About 5,000 years before present (yr BP), this change in solar radiation triggered an abrupt shift in climate and vegetation cover over the Sahara, as reffected in the contribution of terrigenous (land-eroded) dust to oceanic sediment at a sample site near the African coast (lower panel). Modified from ref. 61. SCHEFFER EL AL., NATURE VOL OCTOBER 2001

19

20 Figure 2 Two ways to shift between alternative stable states. a, If the system is on the upper branch, but close to the bifurcation point F2, a slight incremental change in conditions may bring it beyond the bifurcation and induce a catastrophic shift to the lower alternative stable state (`forward shift'). If one tries to restore the state on the upper branch by means of reversing the conditions, the system shows hysteresis. A backward shift occurs only if conditions are reversed far enough to reach the other bifurcation point, F1. b, A perturbation (arrow) may also induce a shift to the alternative stable state, if it is suf ciently large to bring the system over the border of the attraction basin (see also Fig. 3).

21 Figure 3 External conditions affect the resilience of multi-stable ecosystems to perturbation. The bottom plane shows the equilibrium curve as in Fig. 2. The stability landscapes depict the equilibria and their basins of attraction at five different conditions. Stable equilibria correspond to valleys; the unstable middle section of the folded equilibrium curve corresponds to a hill. If the size of the attraction basin is small, resilience is small and even a moderate perturbation may bring the system into the alternative basin of attraction.

22 Externally driven equilibrium change

23 Amazonian Vegetation: Multiple Equilibria, Persistence & Climate A After Wang & Eltahir 2000 B A complication: How does the system get to one or the other? C Another complication Climate change shifts equilibria Vegetation, like climate, can have more than one state that is persistent and resilient, in analogy with movement of a ball on a landscape. Small disturbances lead to adjustments and return to the initial state. Large disturbances may cause the system to change to a new stable state, possibly to revert at a later time (cf. C. Nobre). A shift in climate, due to natural or anthropogenic causes, can change the landscape, as well as the frequency and magnitude of disturbance. The change in relative system stability might make a vegetation change irreversible (e.g. Cox et al, 2001), but it might take a disturbance for the shift to occur. Leads to the concept of instability.

24 Atlantic rainforest

25 Savanna? The Holdridge Life-Zone Classification System (Holdridge, 1947; 1964)

26 Holdridge Life Zones and potential vegetation: the way most models deal with climatic effects on vegetation cover. Data courtesy of D. Skole drying Holdridge life zones (Holdridge 1967)

27 Mean annual precipitation in mm Tmean > 24 C 13 C < Tcoldest month < 18 C P (3 driest months) < 50 mm P (6 wettest months) > 600 mm 1000 mm < Pannual < 1500 mm Annual precipitation Mean climatic equator Climatic Conditions for Savannas Growing season Growing season length in months Arid Savanna Rainforest Savanna Arid South Equator North Latitude A scheme of the relationship between mean annual precipitation and growing season length in tropical climates (from Newman, 1977)

28 Fig. 1 Regions of the Amazon basin that can potentially be converted to savanna after some deforestation. Black regions represent regions in the Amazon basin with tropical forest and having d.s. precipitation > 100 mm. Dark grey regions represent regions having tropical forest with d.s. precipitation 100 mm. This region could potentially be converted to savanna, given enough deforestation. Light grey regions represent other types of vegetation but mainly savannas having precipitation during the dry season 100 mm. The dry season precipitation isoline was derived from Nix (1983). Sternberg, 2001, Global Ecology & Biogeography, 10,

29 Vegetation Types in Brazilian an Amazonia Radam -70,0-70,5-65,5-60,5-55,5-50,5 Area of Study -45,0-45,-10 Vegetation A, Aa, Ab, As C, Cs D, Da, Db, Dm, Ds F, Fa, Fs LO, La, Ld, Lg ON P, Pa, Pf S, SM, SN, SO, ST, Sa, Sd, Sg, Sp Td, Tp WATER rm M km Scale 1: Projection longitude of central meridian ,-15 W N S E

30 Map. no. 1 Annual rainfall (mm) Sombroek 2001, Ambio

31 DEZ-FEV MAR-MAI JUN-AGO Precipitação (mm) > <300 SET-NOV Nilo and Nobre, 1991, Climanálise

32 Amazon River Discharge ( m 3 /s) station: Óbidos (01 S, 55 W) month Year Large interannual variability in the hydrological cycle

33 Sombroek 2001, Ambio

34 Sombroek 2001, Ambio

35

36 The Hypothesis of Savannization Nobre et al. (1991) proposed that a post-deforestation climate in Southern Amazonia would be warmer, drier and with longer dry season, typical of the climate envelope of the tropical savanna (Cerrado) domain of Central South America. Savannization in this context is a statement on regional climate change and not intended to describe complex ecological processes of vegetation replacement.

37 Biomes of tropical south America and precipitation seasonality Biomes of Brazil Tropical Forest-Savanna Boundary Number of consecutive months with less than 50 mm rainfall Tropical Forest Shrubland Savanna Annual Rainfall The importance of rainfall seasonality (short dry season) for maintaining tropical forests all over Amazonia Sombroek 2001, Ambio

38 Evapotranspiration seasonality in the Amazon tropical forest and savanna Source: Rocha (2004) Latent Heat flux (W m -2 ) Forest Savanna mm day -1 Cerrado s.s. SP Floresta trop RO Floresta trop Manaus Floresta trop Santarém Net Radiation (W m -2 ) Forest Savanna

39 Map of dry season length (DSL) (data after Sombroek, 2001), expressed as the number of months with <100 mm of rain. Steege et al., Biodiversity and Conservation 12 (in press), 2003 Kluwer Academic Publishers

40 TROPICAL FOREST COVER BIOME DISTRIBUTION RESPONDS TO CLIMATE! CLIMATE STATE (ANNUAL PRECIPITATION AND LENGTH OF DRY SEASON)

41 TROPICAL FOREST COVER BIOME DISTRIBUTION RESPONDS TO CLIMATE! CLIMATE STATE (ANNUAL PRECIPITATION AND LENGTH OF DRY SEASON)

42 TROPICAL FOREST COVER BIOME DISTRIBUTION RESPONDS TO CLIMATE! SAVANNA CLIMATE STATE (ANNUAL PRECIPITATION AND LENGTH OF DRY SEASON)

43 FOREST TROPICAL FOREST COVER BIOME DISTRIBUTION RESPONDS TO CLIMATE! SAVANNA CLIMATE STATE (ANNUAL PRECIPITATION AND LENGTH OF DRY SEASON)

44 FOREST TROPICAL FOREST COVER BIOME DISTRIBUTION RESPONDS TO CLIMATE! SAVANNA CLIMATE STATE (ANNUAL PRECIPITATION AND LENGTH OF DRY SEASON)

45 FOREST TROPICAL FOREST COVER BIOME DISTRIBUTION RESPONDS TO CLIMATE! SAVANNA CLIMATE STATE (ANNUAL PRECIPITATION AND LENGTH OF DRY SEASON)

46 PRECIPITATION TROPICAL FOREST COVER

47 PRECIPITATION TROPICAL FOREST COVER

48 PRECIPITATION CLIMATE RESPONDS TO VEGETTION! TROPICAL FOREST COVER

49 Focus of LBA Field Research PRECIPITATION TROPICAL FOREST COVER

50 Focus of LBA Field Research PRECIPITATION CLIMATE RESPONDS TO VEGETTION! TROPICAL FOREST COVER

51 Focus of LBA Field Research PRECIPITATION CLIMATE RESPONDS TO VEGETTION! TROPICAL FOREST COVER

52 PRECIPITATION CLIMATE RESPONDS TO VEGETTION! Deforestation Simulations TROPICAL FOREST COVER

53 PRECIPITATION CLIMATE RESPONDS TO VEGETTION! Deforestation Simulations TROPICAL FOREST COVER

54 PRECIPITATION CLIMATE RESPONDS TO VEGETTION! TROPICAL FOREST COVER

55 P present PRECIPITATION CLIMATE RESPONDS TO VEGETTION! 0.8 P present TROPICAL FOREST COVER P present 2 to 2.5m

56 Simple Climate - Biome Interaction Model CLIMATE STATE TROPICAL FOREST COVER

57 Simple Climate - Biome Interaction Model CLIMATE STATE TROPICAL FOREST COVER

58 Simple Climate - Biome Interaction Model FOREST COVER = f (CLIMATE) CLIMATE STATE TROPICAL FOREST COVER

59 Simple Climate - Biome Interaction Model CLIMATE STATE FOREST COVER = f (CLIMATE) TROPICAL FOREST COVER

60 Simple Climate - Biome Interaction Model CLIMATE STATE CLIMATE= = f (FOREST COVER) FOREST COVER = f (CLIMATE) TROPICAL FOREST COVER

61 Simple Climate - Biome Interaction Model CLIMATE STATE CLIMATE= = f (FOREST COVER) FOREST COVER = f (CLIMATE) TROPICAL FOREST COVER

62 Simple Climate - Biome Interaction Model STABLE CLIMATE STATE CLIMATE= = f (FOREST COVER) FOREST COVER = f (CLIMATE) TROPICAL FOREST COVER

63 Simple Climate - Biome Interaction Model STABLE CLIMATE STATE CLIMATE= = f (FOREST COVER) FOREST COVER = f (CLIMATE) TROPICAL FOREST COVER

64 Simple Climate - Biome Interaction Model STABLE CLIMATE STATE CLIMATE= = f (FOREST COVER) FOREST COVER = f (CLIMATE) UNSTABLE TROPICAL FOREST COVER

65 Fig. 3 Establishment of relative forest area in a savanna region as a function of precipitation. Sternberg, 2001, Global Ecology & Biogeography, 10,

66 Forest Vegetation = f (climate) Climate = f (vegetation) Savanna Sternberg, 2001, Global Ecology & Biogeography, 10,

67 1.1 Mudanças globais: aspectos climato-ecológicos Como os biomas da América do Sul seriam afetados? R: (i) Experimentos in loco, e.g. FACE (caros e difíceis de controlar) (ii) Modelos de Vegetação Potencial (MVPot; captam interações bioma-clima) BIOME3 (Haxeltine & Prentice 1996) BIOME (Prentice et al. 1992) TRIFFID (Cox 2001) Simple TRIFFID (Huntingford et al. 2000) MAPSS (Neilson 1995) CPTEC PVM (Oyama & Nobre 2004)

68 2. CPTEC PVM (Oyama & Nobre 2004) entradas: Temperatura, Precipitação saída: um bioma (através de 5 variáveis ambientais) 1. G 0 ( C dia mês -1 ) : growing degree-days (em T basal 0 C) 2. G 5 ( C dia mês -1 ) : growing degree-days (em T basal 5 C) 3. T c ( C): temperatura do mês mais frio 4. H (adimensional): índice hídrico (mod. balanço hídrico) 5. D (adimensional): índice de sazonalidade (mod. balanço hídrico) algoritmo identifica bioma em equilíbrio climático MVPot com melhor desempenho na Am. Sul!

69 Modelagem de Distribuição Geográfica de Espécies Geografia Ecologia Algoritmo Pontos de Ocorrência Temperatura Modelo do Nicho Ecológico Precipitação Previsão da Distribuição

70 Modelagem de Distribuição Geográfica de Biomas Geografia Ecologia Algoritmo Área de Ocorrência Variável Ambiental A Variável ambientaql B Modelo de Biomas Previsão da Distribuição

71 A Potential Biome Model that uses 5 climate parameters to represent the (SiB) biome classification was developed (CPTEC-PBM). CPTEC-PBM is able to represent quite well the world s biome distribution. A dynamical vegetation model was constructed by coupling CPTEC-PBM to the CPTEC Atmospheric GCM (CPTEC-DBM).

72 Simple Land Surface Model Pr: rain Ps: snow T: sfc air temperature Ts: soil temperature S: soil water storage N: overland snow storage E: evapotranspiration R: runoff M: snowmelt Oyama and Nobre, 2002

73 T i, i+1 Ts i, i+1 (soil freezing) S i S i+1 Pr i,, i+1 R i, i+1 E i, i+1 P i, i+1 Ps i,, i+1 M i, i+1 N i N i+1

74 Five climate parameters drive the potential vegetation model Monthly values of precipitation and temperature Water Balance Model Potential Vegetation Model SSiB Biomes Oyama and Nobre, 2002

75 growing degree-days on 0 o C base growing degree-days on 5 o C base Figure 6. Environmental variables used in CPTEC PVM: growing degree-days on 0 o C base (a), growing degree-days on 5 o C base (b), mean temperature of the coldest month (c), wetness index (d), seasonality index (e). Growing degree-days in oc day month -1, and temperature in o C. Oyama and Nobre, 2002

76 mean temperature of the coldest month Wetness index Oyama and Nobre, 2002

77 Oyama and Nobre, 2002 seasonality index

78 Tropical Forest The potential vegetation model algorithm Oyama and Nobre, 2002

79 Visual Comparison of CPTEC-PBM PBM versus Natural Vegetation Map CPTEC-PBM SiB Biome Classification 62% agreement on a global 2 deg x 2 deg grid Oyama and Nobre, 2002

80 Visual Comparison of CPTEC-PBM versus Natural Vegetation Map NATURAL VEGETATION POTENTIAL VEGETATION SiB Biome Classification Oyama and Nobre, 2002

81 Statistic κ (Monserud e Leemans 1992) agrement perfect excel. v. good good agreement good regular poor agreement poor v.poor none Oyama and Nobre, 2002

82 Objective verification of CPTEC-PBM bioma nome p 0 (%) κ concordância 1 floresta tropical 71 0,73 muito boa 2 floresta temperada 52 0,49 regula 3 floresta mista 26 0,26 pouca 4 floresta de coníferas 55 0,56 boa 5 lariços 70 0,65 boa 6 savana 56 0,60 boa 7 campos extratropicais 76 0,50 regular 8 caatinga 50 0,40 regular 9 semi-deserto 57 0,55 boa 10 tundra 62 0,67 boa 11 deserto 70 0,74 muito boa Global Mean Literature Tropical Forest Temperate Forest Mixed Forest Boreal Forest Larch Savannas Grasslands Dry shrubland Semi-desert Desert média global 62 0,58 boa literatura ~ 40 0,40-0,50 regular agreement Very Good Regular Poor Good Good Good Regular Regular Good Good Very Good Good Oyama and Nobre, 2002

83 Searching for Multiple Biome-Climate Equilibria

84 Climate Equilibrium States Vegetation = f (climate) Climate = f (vegetation) Oyama, 2002

85 Vegetation = f 1 (climate variables) = f 1 (g 0, g 5, Tc, h, s) g 0 = degree-days above 0 C g 5 = degree-days above 5 C Tc = mean temperature of the coldest month h = aridity index s = sesonality index f 1 is a highly nonlinear function Climate = f 2 (vegetation) = f 2 (AGCM coupled to vegetated land surface scheme) f 2 is also a nonlinear function

86 Modelo Atmosférico Global para Previsão de Tempo: Código computacional (centenas de milhares de linhas de código) que representa aproximações numéricas de equações matemáticas, equações estas representativas das Leis Físicas que regem os movimentos da atmosfera e as interações com a superfície; o cálculo é feito para até 10 dias de previsão. 38 níveis verticais 1º lat 100 km 1º long 100 km Número de elementos: Número de elementos: x x x x 28= 28= 2,24 2,24 milhões milhões E-W N-S Vertical E-W N-S Vertical Calcula-se para cada um destes volumes: Calcula-se para cada um destes volumes: Temperatura, Temperatura, umidade, umidade, direção direção e e velocidade velocidade do do vento, vento, altura altura geopotencial. geopotencial. Domínio Geográfico Interações entre camadas Interações laterais Interações com a superfície

87 Modelo Atmosférico Regional para Previsão de Tempo: Semelhante ao Modelo Global, porém para um domínio geográfico limitado; o cálculo é feito para 3 dias de previsão. Número Número de de elementos: elementos: x x x x 46= 46= 5 milhões milhões E-W N-S Vertical E-W N-S Vertical 46 níveis verticais 1º lat 20 km 1º long 20 km Calcula-se Calcula-se para para cada cada um um destes destes volumes: volumes: Temperatura, Temperatura, umidade, umidade, direção direção e e velocidade velocidade do do vento, vento, altura altura geopotencial. geopotencial. Domínio Geográfico Interações entre camadas Interações laterais Interações com a superfície globo

88 Modelo Acoplado Atmosfera-Oceano Global para Previsão Climática: Código computacional (centenas de milhares de linhas de código) que representa aproximações numéricas de equações matemáticas, equações estas representativas das Leis Físicas que regem os movimentos da atmosfera, dos oceanos e as interações entre estes dois fluídos e entre a superfície dos continentes e a atmosfera; o cálculo é feito para um período de poucos meses a anos. Atmosfera 28 níveis verticais 1,8 lat 200 km 1,8 long 200 km Número de elementos: Número de elementos: 200 x 100 x 28= 0,56 milhões 200 x 100 x 28= 0,56 milhões E-W N-S Vertical E-W N-S Vertical Calcula-se para cada um dos volumes Calcula-se para cada um dos volumes atmosféricos: atmosféricos: Temperatura, Temperatura, umidade, umidade, direção direção e e velocidade velocidade do do vento, vento, altura altura geopotencial. geopotencial. Calcula-se Calcula-se para para cada cada um um dos dos volumes volumes atmosféricos: atmosféricos: Temperatura, Temperatura, salinidade, salinidade, direção direção e e velocidade velocidade da da corrente, corrente, pressão. pressão. El Niño 125 km 20 níveis verticais Oceano -5km de profundidade

89 How to find numerically Multiple Vegetation-Climate Equilibrium States? Oyama, 2002

90

91

92

93

94

95

96

97 Results of CPTEC-DBM for two different Initial Conditons: all land areas covered by desert (a) and forest (b) a c Initial Conditions b Biome-climate equilibrium solution with IC as forest (a) is similar to current natural vegetation (c); when the IC is desert (b), the final equilibrium solution is different for Tropical South America Oyama, 2002

98 Is the current Climate-Biome equilibrium in Amazonia the only possible one?

99 Two Biome-Climate Equilibrium States found for South America! -- current state (a) -- second state (b) Soil Moisture Rainfall anomalies Oyama and Nobre, 2003

100 Precipitation mechanism in the Amazon Unconditional probability of a wet day. The daily data spans 1979 to mm < P < 5 mm 03 UTC 18 UTC Annual Precipitation 5 mm < P < 25 mm Instability lines SACZ Sea Breezes Source: Obregon, 2001

101 Ecological adaptation I: Deep rooting Fraction of water ex tracted by roots Wet season Depth (m) 68% 84% Source: Bruno et al., 2005 Tropical forest data in Santarem km83

102 Ecological adaptation I: Deep rooting Fraction of water ex tracted by roots Wet season Depth (m) 68% 84% Source: Bruno et al., 2005 Tropical forest data in Santarem km83

103 Ecological adaptation I: Deep rooting Fraction of water ex Dry season tracted by roots Wet season Depth (m) 68% 84% Source: Bruno et al., 2005 Tropical forest data in Santarem km83

104 Ecological adaptation I: Deep rooting Fraction of water ex Wet season Depth (m) Dry season tracted by roots Source: Bruno et al., 2005 Tropical forest data in Santarem km83