Continental Scale Carbon Synthesis

Pedro L. Silva Dias

Institute of Astronomy, Geophysics and Atmospheric Sciences

University of São Paulo

•From local measurements to basin wide:

• tower to biometric measurements to satellite based estimates

•Inverse techniques

•Can we constrain the regional C balance in the Amazon? - BARCA

•Are our models good enough for modeling the climate control on the biosphere and feedbacks?

•Fire

Some ideas for discussion this afternoon

Ecotone (forest-Cerrado transition) – seasonally flooded

Tannus (2004)

Cerrado s. s. Bananal ~ 5 m

Cerradão Bananal ~ 18 m

Wet season

Cerrado sensu strictoNegrón-Juarez (2004)Rocha et al (2002)Miranda et al (1996)Vourlitis et al (2001, 2004)

Tropical rain forestGoulden et al (2004)Rocha et al (2004)Saleska et al (2003)Araujo et al (2001)Von Randow et al (2004)

Santarém km83

Dry seasonTower 67m

CO2 flux (NEE) daily mean (kg C ha-1 day-1), over Cerrado s.s (Gleba Pé de Gigante, from October 2000 to November 2001. Source: Rocha et al. (2002).

Cerrados sensu strictu (woodland savanna) : strong seasonality

Leaves senesceny, grass dormency during the winter

Grows in the wet season

Looses C in the dry season

Tropical forest (Santarém km83, km67) Saleska et al. (2003) (Science), Goulden et al. (2004) (Ecol App)

SeasonalityGEE (photosynthesis), tree growth : increases in the wet seasonSoil Resp (Soil CO2 eflux) : increases in the wet season

Net CO2 ecosystem flux – shows the opposite: absorbs in the dry season

Amazon Basin : wetlands sum ~17% (Mellack et al. 2004)

1.C functionality in land biota

- ecotone (wetlands) in Bananal Island

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

-12-10

-8-6-4-2024

Fluxo CO2 médio diurno

(mol m-2 s-1)

Cerrado s.s .

Floresta Manaus

Floresta km 83Santarém

Ecótono Bananal

Floresta Rondonia

Dry season pattern drives the seasonal tropical rain forests productivity

Contr.: H. Rocha

EXAMPLE: Remote sensing Data (MODIS Enhanced Vegetation Index, EVI) shows that broad areas of

Amazon forest “green up” in the dry season:

Green-up region

Brown-down

region

Source: Huete, Shimabukuro et al., (in review)

Eddy flux tower site

Caxiuana

São Gabriel Cachoeira

Sinop

Scaling Carbon Fluxes with models and remote sensing

Eddy flux data from a dozen towers spans:

• 1.5 – 3.5 m annual precip

• from 0 - 21S

• from primary wet forest to Cerrado, including pasture and agricuture sitesGoals for Data

• Cross-site comparisons of ecosystem properties

• parameterize and test models

• data-grounded extrapolations with remote sensing

CO2 flux variability over tropical forests in Amazonia (t C / ha / yr)

(1) Phillips (1998); (2) Meir (1996), Trumbore (1998), Davidson (2000)(3) Trumbore (2000); (4) Grace (1995), Malhi (1998), Miller (2002)(5) Houghton (2000); (6) Richey (2002)

Above ground biomass increment ( 1 – 3 ) (1)

Net ecosystem productivity ( + 2 to – 6) (3)

Soil C increment 0,5 (isotope studies)

(4)

CO2 outgassing 1 (6)

TOC exported to ocean ( 0,05 ) (6)

However, we have a very limited sample of the carbon balance ....

Marengo, 2003

Variabilidade interanual e interdecadal das chuvas na Amazônia – Marengo, 2003

Parcel studies with tree growth monitoring

- show positive trends all across Amazonia

- is larger in western Amazonia partly to the shortest dry season + fertile soils

Malhi et al (2004) (Global change biology)

How about models?

Month

NE

E (

flux

to a

tmos

pher

e) k

gC/h

a/m

onth

-100

0-5

000

500

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Model output is mean of 4 gridpoints: -54.5 > longitude > -55.5, -2.5 > latitude > -3.5, for neutral years 1980-81,1984-85,1990, & 1993-95.Data is from Tapajos, km67 site (2.85 S, 55 W, from 10-Apr-01 to 08-May-02) & km83 site (3.05 S, 55 W, from 1-Jul-00 to 1-Jul-01).

0

20

40

60

prec

ip

(cm

per

mon

th)

TEM / neutral yrsIBIS \ (1980-95)Data (7/00-7/02)

neutral yrs, 1980-95Site (7/00-7/02)

Model Output from:

TEM (Tian et al, 1998, 2000) & IBIS (Botta &

Foley, 2002) models (8 years, colored lines)

Data from Tapajos Forest eddy flux sites (black line is moving average monthly NEE

SD, points are average weekly NEE)

Mean seasonal NEE and precipitation ( SD of interannual variation) in the Flona Tapajós, Para

Slide and analysis provided by Scott Saleska, Harvard University

What’s missing in the models? Vertical resolution of respiration-relevant moisture? Diagnosis of moisture levels of litter and CWD? Moisture dependence of surface soil respiration?

And how about the future based on interactive biosphere models? Is the Amazon Dieback hypothesis solid?

•The first two GCM climate projections to include an interactive carbon cycle showed a positive climate-carbon cycle feedback mostly due to the negative impacts of climate change on land carbon storage (Cox et al. 2000; Friedlingstein et al. 2001);

Interactive CO2 and dynamic vegetation 2090s - 1990s Haddley Center

•However, the magnitude of the feedback varied markedly between the models (Friedlingstein et al. 2003).

•Other recent coupled simulations: Thompson et al. 2004, Matthews et al. 2004, Zeng et al. 2004, Brovkin et al. 2004), as part of the Coupled Climate-Carbon Cycle Model Intercomparison Project (C4MIP).

•Dufresne et al. (2002) presented results with a fully coupled version of the IPSL where by 2100 the atmospheric CO2 would be 18% higher due to inclusion of carbon cycle feedbacks than without. This positive feedback from the carbon cycle is about 1/3 that reported in Cox et al. Preliminary fully coupled simulations with the Earth Simulator in Japan yield results similar to those of Cox et al. (Kawamiya, pers. comm.).

Carbon-Climate Futures

Carbon Flux: Ocean to Air

-10-8-6-4-202468

10

1850 1900 1950 2000 2050 2100

Pg

C/y

r

Cox et al (2000)

Friedlingstein et al (2001)

Carbon Flux: Land to Air

-10-8-6-4-202468

10

1850 1900 1950 2000 2050 2100

Pg

C/y

r

Atmospheric CO2

200300400500600700800900

1000

1850 1900 1950 2000 2050 2100

pp

m

Global Mean Temperature

13

14

15

16

17

18

19

20

1850 1900 1950 2000 2050 2100

Ce

lsiu

s

Coupled climate—vegetation models project dramatically different futures (CO2, vegetation, T) using different ecosystem models.

~ 2º Kin 2100 T=5T=3

Coupled simulations of climate and the carbon cycle

Slide: A.S. Denning

CO2 => 700 => 500

Summary •How well do coupled climate—vegetation models simulate the response of forest CO2 fluxes and carbon sequestration to warming? Most models seem to over-estimate the sensitivity of ecosystem C storage to T: they use T responses from short-term studies (“zero-order” or “first-order” models), conflate seasonal changes in ecosystem structure with a response to T, and neglect short term/long-term decoupling (e.g. soil warming)

• What are the strongest determinants of vegetation transitions (forest—non-forest) ? [current content of forests (as wood): 350 ppm CO2 equivalent] Precipitation and T together, in concert with sources of ignition, appear to dominate in the tropics, at least under current climate.

• What do coupled climate—vegetation models tell us about covariance among climate change parameters and drivers?Basic model climate sensitivity is critical non-linearity of Carbon – Climate coupling. Warmer—wetter may be likely at mid– and high—latitudes and warmer—dryer in the tropics, implying less positive vegetation—climate feedback and damage to tropical ecosystems.

Contr: Steve Wofsy

Conclusions ● Seasonal water stress is a major determinant of seasonal

carbon exchange in tropical forests in the Amazon region

● Overestimation of drought stress may lead to severe climate simulation errors in coupled models

● Correct coupled climate simulation requires resolution of vertical gradients of soil moisture, and probably also of soil respiration

Contr.: Scott Denning

Another problem in the models: “SAVANIZATION”

SavannaRain forest

Atmopshericdrying

Lessprecipitation

Hig

hse

nsib

lehe

atf

lux

Atmosphericmoistening

Moreprecipitation

High

latenth

eatflux

Contribution: Jon Lloyd

Modeling needs

● Better understanding of determinants between savanna and forest globally

● Better representation of savanna diversity

● (In one often used global data base there are 16 types of temperate/boreal forest and one type of savanna!)

Jon Lloyd contribution

CONCLUSIONS

•In order to reduce the significant uncertainties in climate-carbon cycle projections it is critically important that carbon cycle models are more completely constrained by observational data.

•In order to do this the models need to be more complete so that they are more obviously comparable to the real world. In the case of the land carbon cycle current limitations include the lack of consistent modelling of the effects of land-use change and regrowth (Sitch et al., 2005), and the neglect of carbon-nitrogen interactions.

•C fluxes in tropical amazonian rain forests depend strongly on -the dry season pattern and soil type

•Some places loose and some absorb C on the 1 km2 scale

•Net primary productivity shows positive trend in parcel studies

•Flooded areas are single biomes in the Amazon, where the flood pulse control the emission and sink. They respond for a substantial fraction of the Amazon basin

•Understanding the C cycle mechanisms is a remarkable contribution to the vegetation-climate couple models (a nice tool to assess environmental sustainability over the next decades)

•The C in Amazonia can substantially strengthen the greenhouse effect, given the current IPCC scenarios and deforestation.

Need for BARCA!!! (answer some questions but...others....)

Strategy:

• Lagrangian regional experiments: diurnal airborne measurements of CO2, CO, and H2O within and above the Planetary Boundary Layer (PBL).

• Eulerian experiments: vertical profiles at different times of the day over selected locations

• Large-scale surveys: sampling of large-scale CO2 distribution along the synoptic flow pattern, combined with knowledge of diurnal variations from the Eulerian experiments, gives Basin-scale flux constraints

•to promote simultaneous observation activities in and around the Amazon Basin as a basis for the budgets of heat, water, carbon.•conduct the regular surface measurements in towers and operational networks with high quality during a full year starting in August 2004

Activities: