North America Carbon Program Synthesis Objectives

Managed by UT-Battellefor the Department of Energy

Estimating the North American Carbon Balance

Using Inter-Comparison Among Inversions, Regional Terrestrial Biogeochemistry Models, and Observational

Data

Wilfred Post, Deborah Huntzinger, Kenneth Davis, Brett Raczka, Daniel Hayes, Anna Michalak, Yaxing Wei,

Andrew Jacobson , Robert Cook, and North American Carbon Program Regional-

Interim Synthesis Participants

2 Managed by UT-Battellefor the Department of Energy Presentation_name

North America Carbon Program Synthesis Objectives Utilize new and emerging NACP research findings to

develop a picture of the current state of the North America carbon cycle

Identify continental scale carbon sources/sinks Characterize continental scale inter-annual variation Assess our capability of modeling and measuring

large scale carbon cycle dynamics


4 Managed by UT-Battellefor the Department of Energy

Data sets assembled and regridded for comparison with models MODIS monthly GPP, annual NPP Soil C from surveys Inventory based C pool and flux estimates

– Forest biomass from inventory– Crop NPP from crop yield data

Comparison with eddy flux – crossover model runs with Site Synthesis

Presentation_name


Participating Models

19 TBMs Models differ in:

– Prognostic versus diagnostic– Driver data– Vegetation and soil properties– Photosynthetic formulation– # of carbon pools, soil carbon decomposition dynamics– Processes included, etc.

25 Inverse Models– 20 models with TRANSCOM results– 8 models with post-TRANSCOM results resolved to 1x1

degree

Presentation_name


TBM Model Net Flux for North America

Across model mean net flux; 2000-2005

NEP = 0.66 PgC/yr (1.8 to -0.25 PgC/yr)

NPP = 9.2 PgC/yr (6.2 to 13.8 Pg C/yr)

GPP = 18.4 PgC/yr (9.9 to 31.7 Pg C/yr)

Rh = 8.6 PgC/yr (5.8 to 13.1 Pg C/yr)

Presentation_name


Long-term Mean Summer (June, July, August) Net Ecosystem Productivity

Presentation_name


Can we diagnose the reasons for the lack of consensus in TBM performance? Driver data Photosynthesis formulation Phenology Decomposition – N limitation Regional differences Missing important processes

Presentation_name


Crossover Models - Regional model runs are more positively biased for GPP. Site model runs closer to observations.

Raczka et al. (in preparation)


Long-Term Mean NEE – TBMs and Inversions

Presentation_name



Net NPP over U.S. Agricultural Lands:Models Compared to NASS Inventory-Based Data

Presentation_name


Net NPP Spatial Pattern Over U.S. Agricultural Lands: Models Compared to NASS Inventory-Based Data

Presentation_name


Inventory – Model Comparison

Presentation_name

Data Sources Years of Data Location Fluxes and Pools Available Source

Canada Mgd. Forest (CBM-CFS3) 2000 – 2006 Canada (n = 15) NEE, NPP, VegC, Fire

EmissionsKurz et al.Stinson et al.

Canada Croplands 2000 – 2006 (average) Canada (n = 8) NEE, NPP McConkey

U.S. Forest (FIA / CCT) 2000 – 2006 Continental U.S. plus Alaska (n = 49) NEE, VegC Smith & Heath

U.S. Croplands 2000 – 2005 Continental U.S. (n = 49) NEE, NPP West et al.

Mexico 2000 - 2005 Mexico (n = 32) NEE, VegC, Fire Emissions deJong et al. 2010


Accounting for Lateral Fluxes

Presentation_name

CO 2 CO 2

Residu

e

VEGC

SOILC

Harvest

LivestockC

CH4

HumanC

CH4

Consumption

Imports – Exports

Land Use Change

PRODC

Fire RH

NPP

Fire

NEE

Grasslands /

Settlements

CO

2 U

ptake

CO

2 R

elea

se

NEEForest / Cropland

Sector“Other” Lands Sector

Export


INVE

NTO

RY

DA

TA

TOTAL FORESTLAND CROPLAND OTHER LANDSFO

RW

AR

D

MO

DEL

SIN

VER

SE

MO

DEL

S

Mean average annual NEE (Tg C yr-1), 2000 to 2006


Summary Regional TBM comparisons indicate a lack of

consensus for NEP and component fluxes – GPP, Ra, Rh

Analyses to diagnose the causes reveal:– Model formulation plays a significant role.– Different weather driver data sets greatly impact

GPP TBM Re tends to be tightly related to GPP –

this dampens the NEE seasonal cycle and IAV

Annual NEE cycle amplitude is small for TBM compared to inversions.

Presentation_name


Conclusions Regional C modeling enterprise needs a more rigorous

approach to development and evaluation. Large disparities remain in estimates based on temporal and

spatial extrapolations of experimental and site based understanding.

Additional data, especially based on spatially extensive measurements, needs to be integrated into the modeling system approach.– Improve model algorithms/parameters– Improved diagnostic and predictive usefulness– Evaluate model skill– Develop useful benchmarks

An transition from data-poor to data-rich approach is emerging for developing multi-stream observing systems and modeling system analyses.

The NACP regional synthesis has contributed by exploring how a wider range of data sources can be used.

Presentation_name


Next Steps

RECCAP– If you have model results for 1990-2008 we would

be interested in incorporating them into the North American RECCAP chapter.

– Contact Mac Post <[email protected]>

MsTMIP– If you didn’t make it to the Side Meeting yesterday

you are still invited to participate– Contact Deborah Huntzinger

<[email protected]>– Check the NACP web site for a MsTMIP e-mail list

to be started soon.

Presentation_name


Acknowledgments

NASA Terrestrial Ecology support the Modeling and Synthesis Thematic Data Center (MAST-DC)

DOE Office of Science for workshop support NOAA support for collecting inversion model

results Generous contributions of time and

resources from any modeling teams and data providers that made these analyses possible

Presentation_name

North America Carbon Program Synthesis Objectives

Documents