| Simulation of Atmospheric CO2 Concentration with SPEEDY |
| IMAGE & MSRI Summer Graduate Workshop | |
| on Data Assimilation for the Carbon Cycle Boulder, CO 8-13 July 2007 |
|
| Ji-Sun Kang | |
| University of Maryland |
| Goal |
| Simulate transport of atmospheric CO2 in a simple but realistic model | |
| Add coupling with VEGAS | |
| Perform LETKF (EnKF) testing methodologies using a nature run with VEGAS and a forecast model without it |
| WhatÕs SPEEDY? |
| SPEEDY (F. Molteni, 2003) | ||
| (Simplified Parametrizations, primitivE-Equation DYnamics) | ||
| General Circulation Model of the Atmosphere | ||
| Prognostic variables: U, V, T, q, Ps | ||
| Spectral model with T30 resolution (96X48) | ||
| Seven layers in the vertical | ||
| Six-hour forecast (modified by Takemasa Miyoshi) | ||
| To Predict CO2 Concentration |
| Add one additional prognostic variable, CO2 | ||
| Only advection and diffusion in the model | ||
| Source/Sink from the surface | ||
| No feedback between the integrated CO2 and radiative properties in SPEEDY | ||
| No chemical processes | ||
| Problem of Mass Conservation |
| Imperfect conservation of mass (spectral dynamics) | |||
| Empirical correction | |||
| Sum up CO2 forcing from the whole surface | |||
| A=·(CO2 forcing) for every timestep (40 min) | |||
| Accumulate total mass of CO2 in the simulation: B | |||
| B= · (CO2 concentration) over all grids for every timestep | |||
| Multiplying the ratio of A/B to every grid | |||
| A>B (loss of mass in SPEEDY) | |||
| Simulation |
| Three-year run from Jan. 1st in 1982 | |||
| Initial condition | |||
| Atmospheric fields (u, v, T, q, Ps) from SPEEDY | |||
| Surface CO2 flux (kg/m2/sec) | |||
| Fossil fuel emission: Andres et al. (1995) | |||
| Constant forcing | |||
| 5.0 PgC/yr | |||
| Conversion of Units |
| From the surface flux (kg/m2/sec) to the concentration in the atmosphere (ppmv) | |||
| Δc=C_sflx(kg/m2/sec)*(g/ΔP)*29/12 | |||
| where ΔP= Psfc*Δσ, (Psfc=1000 hPa, Δσ=0.2) | |||
| CO2 Transport in the Atmosphere |
| The lowest layer of the model | ||
| From the very first step of simulation to the end of the first year | ||
| Only source from fossil fuel emission | ||
| No sink | ||
| CO2 Transport in the Atmosphere |
| Vertical Cross section of Zonal Mean | ||
| Annual Mean of CO2 Distribution |
| Comparison |
| Current Work |
| Coupling VEGAS to SPEEDY (done) | |||
| VEGAS (VEgetation Global Atmosphere and Soil) | |||
| Zeng, 2003 (University of Maryland) | |||
| Vegetation dynamics and terrestrial carbon cycle | |||
| Coupled to SLand | |||
| SLand (Simple-Land) | |||
| Land surface energy and water budget, flux exchange with lower atmosphere | |||
| Further Plan |
| Local Ensemble Transform Kalman Filter (LETKF) | ||
| Couple model with LETKF | ||
| Use model coupled with VEGAS (nature run) to create simulated observations of surface fluxes and concentrations, as well as regular atmospheric observations | ||
| Perform observing system simulation experiments (OSSEs) to the forecast model of CO2 , SPEEDY, using simulated observations from the coupled model | ||
| Thank you |
| Structure of SPEEDY |
| References |
| Andres, R. J., et al. 1996: A 1¡ x 1¡ distribution of carbon dioxide emissions from fossil fuel consumption and cement manufacture, 1950-1990, Global Biogeochem. Cycles, 10, 419-429. | |
| http://earthobservatory.nasa.gov/Library/CarbonCycle/carbon_cycle.html | |
| http://www.cmdl.noaa.gov/gallery2/d/26551-4/ccggmap.png | |
| Lecture Notes from the workshop of ÔData Assimilation for the Carbon CycleÕ at Berkeley, July 17-28, 2006 | |
| Molteni, F., 2003: Atmospheric simulations using a GCM with simplified physical parameterizations. I: model climatology and variability in multi-decadal experiments. Clim. Dyn., 20, 175-191. | |
| NASA-CASA Project, http://geo.arc.nasa.gov/sge/casa/ | |
| Takahashi, T., et al. 2002: Global sea-air CO2 flux based on climatological surface ocean PCO2, and seasonal biological and temperature effects, Deep-Sea Research, II, 49, 1601-1622 |
| Slide 16 |