Measuring Fluxes From Satellite Observations
The accurate inference of anthropogenic and natural source and sink fluxes from satellite observations of carbon dioxide is a key challenge in carbon cycle science with potential implications for global carbon monitoring, reporting, and verification. At the Jet Propulsion Laboratory, current methods are based on combining carbon dioxide observations with the GEOS-Chem transport model in a 4D-variational assimilation framework.
While this approach permits fine spatial scale resolution of carbon fluxes, it assumes assimilated fields accurately and reliably transport concentrations from surface emissions to altitude levels measured by satellite instruments. However, transport accuracy has been identified as a major source of uncertainty in “top-down” flux estimates.
Data Assimilation Framework
To address this challenge, the Center has proposed an innovative data assimilation framework that will implement an Ensemble Kalman Filter (EnKF) algorithm on the NASA GEOS-5 general circulation model in conjunction with the 4D-var GEOS-Chem CO2 inverse modeling algorithm.
A key advantage of the EnKF algorithm is a direct estimate of the transport uncertainty as defined by the ensemble spread. CO2 carbon fluxes will be estimated using GEOS-Chem driven by each ensemble member of transport. The resulting spread in carbon fluxes as a function of ensemble member will provide a direct calculation of the impact of transport error on flux estimates.
The Center will initially use TES and AIRS CO and CO2 observations and then incorporate MOPITT CO and GOSAT CO2 observations, time permitting.
Cloud and Related Water Cycle Feedbacks
The Center is focusing on the analysis of climate change simulations being performed under the auspices of the World Climate Research Programme CFMIP5/IPCC with imbedded satellite sensor simulators including CloudSat, CALIPSO and others. In this context, the Center will:
Use global data sets on clouds, precipitation and water vapor to evaluate key aspects of the atmospheric branch of the water cycle
Diagnose where major discrepancies occur and what processes are responsible for these discrepancies
Develop entirely new diagnostic approaches and satellite simulators to study and evaluate cloud feedbacks in climate in support of the CFMIP5 activity.
Climate Modeling Process Team
In addition, the Center is coordinating a Climate modeling Process Team (CPT) project which involves the University of California Los Angeles, the National Center for Atmospheric Research, the University of Washington, the Lawrence Livermore National Laboratory, and the National Centers for Environmental Prediction. This CPT is dedicated to improving the observational characterization and the representation in climate models of boundary layer clouds, namely the transition from stratocumulus to cumulus clouds.
Key deliverables of this CPT are:
More realistic characterization using observations and high-resolution models of the stratocumulus to cumulus transition
Implementation and testing of Eddy-Diffusivity/Mass-Flux (EDMF) mixing parameterization in the NCEP and NCAR global models
Implementation and testing of new cloud parameterization in the NCEP and NCAR models.
Land Processes and Feedbacks
Evapo-transpiration (ET) is a fundamental process in the climate system, influencing both water and carbon cycles. ET acts to couple these cycles, causing feedbacks between them. ET is one of the largest and most uncertain contributors in carbon and water cycles over land, and monitoring ET globally has become a key goal of NASA’s Earth Science Program and the US Climate Change Science Program.
There are currently five major methods with which to derive ET globally:
Statistical models relating ET to land and/or meteorological characteristics such as vegetation greenness or temperature
Spatial interpolation and remote sensing upscaling from eddy flux measurements
As a product of one of the components in a land surface model
Dedicated ET algorithms driven by a combination of remote sensing and/or ground observations.
The Center for Climate Sciences at JPL supports the running of five remote sensing-driven dedicated ET models and three land surface models using identical driving data and model protocols.
Land surface models are:
Joint UK Land Environment Simulator (JULES) as used in the current UKMO climate model
Lund-Potsdam-Jena (LPJ) model
Simplified Simple Biosphere model 3 (SSiB-3).
The Center will also promote the implementation of a new approach for calculating all-sky net radiation from MODIS data products (Ryu et al. 2008), to compare these fluxes to flux data currently being derived independently from A-Train data. This will allow products to be developed at 1 km spatial resolution, which in turn will allow direct comparison to eddy flux measurements.
The Center will validate the products using the new La Thuille FLUXNET database, which contains 253 sites globally where ET is directly measured in situ at ~1 km scale. In this way, satellite observations and process models are brought together to evaluate land schemes used in current climate models.