PI: Nicolas H. Younan Surya S. Durbha, Fengxiang Han, Roger L. King, Jian Chen, Zhiling Long
Post on 19-Mar-2016
45 Views
Preview:
DESCRIPTION
Transcript
Using Simulated OCO Measurements for Assessing
Terrestrial Carbon Poolsin the Southern United States
PI: Nicolas H. YounanSurya S. Durbha, Fengxiang Han, Roger L. King,
Jian Chen, Zhiling LongGeoResources Institute (GRI)
Institute for Clean Energy TechnologyMississippi State University
NASA RPC- 10 Jul 2007
Orbiting Carbon Observatory (OCO)
First global, space-based measurements of atmospheric carbon dioxide (CO2) with the precision, resolution, and coverage needed to characterize CO2 sources and sinks on regional scales.
Uncertainties in the atmospheric CO2 balance could be reduced substantially if data from the existing ground based CO2 network were augmented by spatially resolved, global, measurements of the column integrated dry air mole fraction (X CO2 ) with precisions of ~1 ppm (0.3% of 370 ppm (Crisp et al 2004)
Source:http://oco.jpl.nasa.gov/images/ground_track-br.jpg
NASA RPC- 10 Jul 2007
Scope of the Research This research is focused on the assessment of This research is focused on the assessment of
terrestrial carbon pools in the southeast and south terrestrial carbon pools in the southeast and south central United States. central United States.
In particular, this investigation intends to leverage In particular, this investigation intends to leverage upon:upon: Multiple NASA sensors Multiple NASA sensors The terrestrial ecosystem model (CASA) and The terrestrial ecosystem model (CASA) and Transport model GISS: GCM Model ETransport model GISS: GCM Model E
Undertake a Rapid Prototyping (RPC) experiment to Undertake a Rapid Prototyping (RPC) experiment to address the need to quantify the carbon exchange address the need to quantify the carbon exchange over different ecosystems.over different ecosystems.
Test how well data from Test how well data from OCO observations and CO2 measurement networks constrain CO2 fluxes at at model-grid resolution.model-grid resolution.
NASA RPC- 10 Jul 2007
Science Questions Proposed RPC experiment seeks to address the Proposed RPC experiment seeks to address the
following questions:following questions: What information about carbon exchange can be obtained What information about carbon exchange can be obtained
from OCO high-precision column measurements of from OCO high-precision column measurements of CO2?? How can we integrate top-down OCO measurements with How can we integrate top-down OCO measurements with
ground based measurements, atmospheric and terrestrial ground based measurements, atmospheric and terrestrial ecosystem models to quantify carbon exchange over ecosystem models to quantify carbon exchange over different ecosystems?different ecosystems?
What are the current annual rates of terrestrial carbon What are the current annual rates of terrestrial carbon sequestration in each state of the Southeast and South-sequestration in each state of the Southeast and South-central U.S.?central U.S.?
What is the current baseline in the region for possible carbon What is the current baseline in the region for possible carbon trading?trading?
What is the potential for enhancing terrestrial carbon What is the potential for enhancing terrestrial carbon sequestration?sequestration?
NASA RPC- 10 Jul 2007
What are the current annual rates of terrestrial carbon sequestration in each state of the region?
What's the overall contribution of terrestrial carbon sequestration in each state of the region to mitigating its total greenhouse gas emission?
What's the current baseline for possible carbon trading in the region?
What's the potential of further enhancing terrestrial carbon sequestration in the region? And
What are the overall economic impacts of current and potential terrestrial carbon sequestration on the region?
Currently funded DOE project for leverage
NASA RPC- 10 Jul 2007
Total terrestrial carbon storage and pools in the region
Soil Organic C:16535 Tg C, 76%
Forest C:4454 TgC, 20.5%
Housing/Furniture C:661 Tg C, 3.0%Crop C: 85 Tg C,
0.4%Pasture C:
27.8 Tg C, 0.13%
Total Terrestrial C Storage: 21762 Tg C
NASA RPC- 10 Jul 2007
Current annual terrestrial carbon sink in the region
Soil OM, 3.8 Tg C/yr,
1.8%
Crop C: 85 TgC/yr, 41.4%
Forest C:76 Tg C/yr,
36.9%
Housing/Furniture C:13.2 Tg C/yr, 6.4%
Pasture C:27.8 Tg/yr,
13.5%
Total Annual Terrestrial C Sink: 206 Tg C/yr
NASA RPC- 10 Jul 2007
The potential terrestrial carbon sequestration in the region
Forestland29.4 Tg C/yr
54.6%
Cropland15.1 Tg C/yr
28%
Grassland9.4Tg C/yr
17.5%
Potential Annual C Sink: 53.9 Tg C/yr
NASA RPC- 10 Jul 2007
Findings
Current annual terrestrial carbon sequestration (soil, forest, crop, pasture and house/furniture) in the region can offset 40% of the total annual greenhouse gas emission.
Through proper policies and best management, about 10.1% of the total greenhouse gas in the region can be further offset by terrestrial sequestration.
Terrestrial carbon sequestration proves to be the most cost-effective option for sequestering carbon in the region.
Han, F.X., J. Lindner, and C. Wang. 2007. Making carbon sequestration a paying proposition. Naturwissenschaften 94: 170-182.
DOI 10.1007/s00114-006-0170-6.
Han, F.X., M. J. Plodinec, Y. Su, D.L. Monts, and Z. Li. 2007. Terrestrial carbon pools in southeast and south-central United States.
Climatic Change. DOI 10.1007/s10584-007-9244-5.
Han F.X., Z.P. Li, J. Lindner, Y. Su, D. L. Monts, R. King, B. Xing, and J.M. Plodinec. 2007. Role of soils and soil management for
mitigating greenhouse effect. In B. Xing, F. Wu (eds) Natural Organic Matter and Its Significance in the Environment. The Science
Press, Beijing and Brill Academic Publisher, Leiden, Boston and Tokyo.
NASA RPC- 10 Jul 2007
Rapid Prototyping Using Simulated Data Sets RPC using:RPC using:
Simulated Orbiting Carbon Observatory (OCO) Simulated Orbiting Carbon Observatory (OCO) obtained through the Observing System obtained through the Observing System Simulation Experiment (OSSE)Simulation Experiment (OSSE)
Perform various sensitivity studies and Perform various sensitivity studies and understand their suitability.understand their suitability.
NASA Carbon Query and Estimation Tool NASA Carbon Query and Estimation Tool (CQUEST) is the target DSS.(CQUEST) is the target DSS.
NASA RPC- 10 Jul 2007
Rapid Prototyping Concept (RPC)
NASA RPC- 10 Jul 2007
RPC Experimental Design
• Assimilation of aircraft measurements, satellite data (precipitable water, surface winds)
• Vegetation Indices• Biome type• Soil properties• Weather Reanalysis
MeteorologyMeteorology(e.g. GOES (e.g. GOES
data data analysis)analysis)
• 1 year spinup• Monthly
• Terrestrial CO2 surface flux
• Winds, cloud mass fluxes, model Parameters
• Forward Transport Model
• Fossil Fuels
1 year spinup (2002)Land Land
Surface Surface Model Model (CASA)(CASA)
Transport Transport Model (GISS Model (GISS GCM Model GCM Model E)E)
[CO[CO22] OBS] OBS
• OCO, Networks
InversionInversion
NASA RPC- 10 Jul 2007
RPC Experimental Design
• Vegetation Indices• Biome type• Soil properties• Weather Reanalysis
• 1 year spinup• Monthly
• Terrestrial CO2 surface flux
Land Land Surface Surface Model Model (CASA)(CASA)
• Assimilation of aircraft measurements, satellite data (precipitable water, surface winds)
MeteorologyMeteorology(e.g. GOES (e.g. GOES
data data analysis)analysis)
• Winds, cloud mass fluxes, model Parameters
• Forward Transport Model
• Fossil Fuels
1 year spinup (2002)
Transport Transport Model (GISS Model (GISS GCM Model GCM Model E)E)
[CO[CO22] OBS] OBS
• OCO, Networks
InversionInversion
NASA RPC- 10 Jul 2007
Evaluating the Suitability of Vegetation Parameters
Evaluate the usefulness of multi-angle Evaluate the usefulness of multi-angle measurements from MISR data sets to measurements from MISR data sets to assess the model predictions in the event of assess the model predictions in the event of using NDVI and LAI observation from multi-using NDVI and LAI observation from multi-angle data sets. angle data sets.
Durbha, S.S., R.L. King, and N.H. Younan, Support vector machines regression for retrieval of leaf area Durbha, S.S., R.L. King, and N.H. Younan, Support vector machines regression for retrieval of leaf area
index from multiangle imaging spectroradiometer, index from multiangle imaging spectroradiometer, Remote Sensing of Environment Special Issue: Multi-Remote Sensing of Environment Special Issue: Multi-
angle Imaging SpectroRadiomenter (MISR)angle Imaging SpectroRadiomenter (MISR), Volume 107, Issues 1-2, pp. 348-361, March 2007. , Volume 107, Issues 1-2, pp. 348-361, March 2007.
NASA RPC- 10 Jul 2007
CASA Model Tasks Sensitivity analysis of how much NPP increase is required Sensitivity analysis of how much NPP increase is required
to sustain the regional terrestrial carbon sink of the study to sustain the regional terrestrial carbon sink of the study area.area.
Net Ecosystem ProductivityNet Ecosystem Productivity (NEP) defined as Net Primary (NEP) defined as Net Primary Production (NPP) minus the heterotrophic soil respiration Production (NPP) minus the heterotrophic soil respiration predictions would be used to infer variability in regional predictions would be used to infer variability in regional scale carbon fluxes and to better understand patterns scale carbon fluxes and to better understand patterns over terrestrial carbon sinksover terrestrial carbon sinks..
The CASA model estimates of carbon products would be The CASA model estimates of carbon products would be calibrated with field-based measurements ofcalibrated with field-based measurements of
Crop production, Crop production, Forest ecosystem fluxes, andForest ecosystem fluxes, and Inventory estimates of carbon pool sizes at multiple locations in south Inventory estimates of carbon pool sizes at multiple locations in south
eastern and south central United States.eastern and south central United States.
NASA RPC- 10 Jul 2007
RPC Experimental Design
• Terrestrial CO2 surface flux
• Winds, cloud mass fluxes, model Parameters
• Forward Transport Model
• Fossil Fuels
1 year spinup (2002)
• Assimilation of aircraft measurements, satellite data (precipitable water, surface winds)
• Vegetation Indices• Biome type• Soil properties• Weather Reanalysis
MeteorologyMeteorology(e.g. GOES (e.g. GOES
data data analysis)analysis)
• 1 year spinup• Monthly
Land Land Surface Surface Model Model (CASA)(CASA)
Transport Transport Model (GISS Model (GISS GCM Model GCM Model E)E)
[CO[CO22] OBS] OBS
• OCO, Networks
InversionInversion
NASA RPC- 10 Jul 2007
OCO Data Assimilation: Problem Formulation
Compare predictions from atmospheric transport model (e.g. GISS Model E) and measurements of atmospheric carbon abundances from OCO and at observation sites distributed over the regions of interest.
Spatial pattern of the observed and predicted differences can be used to infer the spatial distribution of sources and sinks of carbon dioxide by seeking a distribution of fluxes that in a least squares sense minimizes the difference between the model predictions and observation, as well as any prior information used to constrain the problem.
NASA RPC- 10 Jul 2007
OCO Data Assimilation: Techniques and Strategies Commonly employed
technique to estimate carbon fluxes is Bayesian synthesis inversion.
A cost function is formulated that has two terms:
One involving the observations and one involving a prior estimate of the fluxes.
Resulting flux estimates are constrained both by observations and prior estimates.
(Baker et al., 2006)Cost function to minimizeCost function to minimize
NASA RPC- 10 Jul 2007
OCO Data Assimilation: Techniques and Strategies Improved Kalman Smoother for atmospheric inversion.
Produces estimates of fluxes at a particular time using observations from that time step as well as observations from subsequent times.
Normal Kalman filter would use only past observations to estimate fluxes at a particular time step
Ensemble Kalman filters allows for application on large problem.
Adjoint-based descent methods for variational data assimilation
We are exploring the possibility of developing a Support Vector Regression-based technique for this purpose
NASA RPC- 10 Jul 2007
Original Timeline Official start date Official start date March 12, 2007March 12, 2007
NASA RPC- 10 Jul 2007
Significant Issues
OCO simulated data acquisition problems Any leads would be highly appreciated!
NASA RPC- 10 Jul 2007
Summary: OCO Data Assimilation for Assessing Terrestrial Carbon Pools in the Southern US High-density data (e.g. OCO) should allow us to
address a variety of science and policy questions that have remained previously unanswered.
Resolving surface fluxes to the regional biome level will help to quantify the relative importance of the key driving processes.
Resolving them to regional levels helps in the carbon management and verification of carbon credits, compliance, etc.
NASA RPC- 10 Jul 2007
Questions?Questions?
NASA RPC- 10 Jul 2007
OCO Simulated Data Evaluation
In situ observations from global surface sites In situ observations from global surface sites (NOAA/CMDL, AMERIFLUX) would be used to calculate (NOAA/CMDL, AMERIFLUX) would be used to calculate the trends in the seasonal cyclethe trends in the seasonal cycle
Test whether the true flux distribution can be retrieved Test whether the true flux distribution can be retrieved using the OCO observations and independent using the OCO observations and independent CO2 flux flux distribution.distribution.
Combine the CASA model with a global transport model Combine the CASA model with a global transport model (GISS) to identify and relate the amplitude/seasonal (GISS) to identify and relate the amplitude/seasonal cycle of biospheric cycle of biospheric CO2 from OCO observations from OCO observations Helps to understand and develop methods to reduce uncertainty in Helps to understand and develop methods to reduce uncertainty in
regional regional CO2 flux estimates. flux estimates.
NASA RPC- 10 Jul 2007
Specific Tasks The input drivers to the NASA-CASA model consists of parameters The input drivers to the NASA-CASA model consists of parameters
derived from climatic, site, vegetation, soils and resolution (e.g. derived from climatic, site, vegetation, soils and resolution (e.g. daily, monthly). The following parameters are required for model daily, monthly). The following parameters are required for model initialization.initialization. Monthly data of temperature, precipitation, PAR, NDVI Soil Monthly data of temperature, precipitation, PAR, NDVI Soil
type/soil water capacity, Vegetation type, type/soil water capacity, Vegetation type, CO2 The various parameters from different sources would be studied for The various parameters from different sources would be studied for
their suitability. their suitability. in situin situ based measurements would be assessed for their inclusion based measurements would be assessed for their inclusion
into the model input.into the model input. Combine the CASA model with a global transport model (GISS Combine the CASA model with a global transport model (GISS
Model E) to identify the changes in the terrestrial biosphere that are Model E) to identify the changes in the terrestrial biosphere that are consistent with the observed increases in the amplitude of the consistent with the observed increases in the amplitude of the seasonal cycle of atmospheric seasonal cycle of atmospheric CO2. .
NASA RPC- 10 Jul 2007
OCO Data Assimilation: Techniques and Strategies No single model or set of observations can
quantify the dynamics of terrestrial Carbon exchanges, and describe the governing processes.
Recent attempts are to develop coherent methods treating both data and models as sources of information.
Core problem is “How to combine and weight the various information sources?”
NASA RPC- 10 Jul 2007
Models and DataLand biosphere model (NASA-
CASA )Transport Model (GISS GCM model E) Surface sites (in situ networks) and Simulated
observations (OCO)
General:Parameters. table: initial parameter values for CASA variables and constants.COORDINATES: files latitude of latitudes and longitude for longitudes for nongridded data all in degreesIn situ: soil.table: soil classes (soil type or composition fractions or only silt, sand, clay etc without soil.table (1 km resolution)CO2.table: CO2 ppm data AIRT: temperature (celsius)PPT: precipitation (mm/month)SOLRAD: solar radiation (w/m2 averaged over each month) Optional:LANDMASK ( not required in our case)ANN_AGRI (annual agriculture)DEFOREST (deforestation)Remote Sensing (Monthly/Bi-monthly) Landcover table: vegetation classesNDVI (MISR, VIIRS)
General:Model configuration:http://www.giss.nasa.gov/tools/modelE/HOWTO.html#part0_2Intend to run on multiple processors using HPC cluster.Resolution Selection (4o×5o or 2 × 2.5o etc.)Boundary and initial conditions and any run-specific parameters (such as the time step, length of run etc.) needs to be determinedNeed to create ‘Deck’ Files as inputs to the model. The Deck files contain the necessary input information such as Since our interest is in monthly average fluxes using monthly average observations and responses, we intend to use coarse resolution meteorological data as input (e.g. wind, surface pressure, temperature)
Global Surface SitesNOAA/CMDL, AMERIFLUXData from GLOBALVIEW-CO2 has been extensively used in other recent studies. These are CO2 measurements from the NOAA /CMDL cooperative air sampling network and have been successfully applied to other trace gas measurement records.We can use the data from some of these stations(http://islscp2.sesda.com/ISLSCP2_1/html_pages/ groups/carbon/globalview_CO2_point.html )Consider using data from about 30-50 stations for this RPC experiment
Observing System Simulation Experiment (OSSE )- OCO data Year: 2002Region: Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Texas, and Virginia.
top related