NACP

1
Modeling ecosystem-level carbon fluxes using an
advanced land surface model (FAPIS) coupled to
the Regional Atmospheric Modeling System (RAMS).
Jesse N. Miller1, Lianhong Gu1, and W. Mac
Post1 1 Environmental Sciences Division, Oak
Ridge National Laboratory (ORNL), Oak Ridge,
Tennessee 37831-6407
NACP
Sample Results
Experiment Setup
Introduction
Spatial Plots
MOFlux Tower
RAMS-FAPIS, is a newly developed
ecosystem/atmospheric transport model. It
consists of the Regional Atmospheric Modeling
System (RAMS), a numerical prediction model that
is used to simulate atmospheric circulations at
the mesoscale, and the Fluxes and Pools
Integrated Simulator (FAPIS), which is a
fully-developed terrestrial ecosystem land
surface model. In the RAMS-FAPIS tool the
original land surface model of RAMS, LEAF2, is
replaced by FAPIS.
RAMS-FAPIS is run over Missouri with nested grids
centered over the MOFLUX tower location. It is
forced with meteorological data from the North
American Regional Reanalysis (NARR).

• Vegetation Forest density 583 trees ha?1.
  Canopy height 17 to 20 m. LAI 4.2 m2 m?2.
  Oak-hickory forest.
• Soil types Weller silt loam, rocky, thin soil
  covering.
• Climate Warm, humid and continental, MAT
  12.8oC, MAP 940 mm, severe droughts common in
  late summer.
• Sensors eddy covariance system at 31 m, two
  profile sampling systems CO2/water vapor
  (12-level) and air temperature/relative humidity
  (8-level), soil measurements.

Figure 4. Average vertical velocity from west to
east across MO bisecting MOFlux tower.
Figure 5. CO2 concentration from west to east
across MO bisecting MOFlux tower.
Figure 3. CO2 concentration at 22 m calculated as
a RAMS scalar.
Objectives

• To provide a modeling tool that can be used to
  diagnose and predict short-term
  spatially-explicit carbon sources and sinks.
• To provide a bottom-up approach for estimating
  terrestrial CO2 fluxes using atmospheric
  concentration (tall towers, Orbiting Carbon
  Observatory) and land surface flux data (FLUXNET)
  using data assimilation methods that could be
  comparable to results from top-down approaches.

These plots are good for showing RAMS-derived
variables, including newly-added scalars (CO2),
along with meteorological drivers over time and
space. Can watch synoptic systems and identify
sources/sinks for CO2.

• Baskett Research and Education Area, University
  Of Missouri
• 25km from Columbia, Missouri

Surface Analysis
Single-point Plots
Figure 2. Surface analysis at start of model run.
These types of plots are good for showing
within-canopy profiles of CO2 concentration,
temperature, humidity, winds, NEE, and GPP as
well as biological and soil layer variables.
Figure 1. MOFlux and other AmeriFlux locations
within MCI region.
Models
RAMS

• Numerical prediction model with a flexible
  configuration (such as two-way interactive
  grids!) good for mesoscale atmospheric
  simulations. Terrain-following.
• Full set of non-hydrostatic, compressible
  equations for atmospheric dynamics and
  thermodynamics, plus conservation equations for
  scalars such as water vapor, liquid and ice
  hydrometeor mixing ratios.
• Many parameterization options.
• Original form, with LEAF2 land surface model,
  does not calculate ecosystem exchange variables
  such as NEE.

RAMS-FAPIS Basics

• 6 hourly gridded pressure-level forcing data is
  prepared by RAMS software from NARR data. Input
  variables, updated every time step, are provided
  by RAMS to canopy top level of FAPIS. FAPIS
  calculates ecosystem variables for each grid cell
  and returns via bottom layer of RAMS. Surface
  layer fluxes (momentum, heat, moisture) are
  calculated from these using the Louis (1979)
  scheme in original RAMS surface model. Fluxes
  are fed to RAMS every time step.

Figure 6. Canopy air space CO2 concentration, as
calculated in FAPIS, for 2 days at 2 levels.
Future Work

• Compare RAMS-FAPIS output with MOFlux single
  point observations.
• Compare RAMS-FAPIS output with NARR fields.
• Conduct footprint analysis and quantify sources
  of observed scalars such as CO2 or isoprenes in
  conjunction with aircraft measurements.
• Conduct sensitivity tests with changing land
  surface characteristics, etc.
• Run over Amazon region to analyze influence of
  biomass heat storage.

Grid Parameters
Model Configuration
FAPIS

• Bulk microphysics
• Kuo-type convective parameterization
• Harrington radiation scheme (shortwave and
  longwave)
• Mellor/Yamada turbulence closure (prognostic TKE,
  Smagorinsky horizontal diffusion)
• CO2 added as scalar to RAMS. Initialized at 380
  ppm.

• 2 grids, both with 30 vertical cells expanding by
  15 from a 45 m thick surface layer to 1200 m
  near 15 km top.
• First Grid 60 x 60 horizontal cells with 10 km
  x 10 km resolution. 60 second time step.
• Second Grid 20 x 20 horizontal cells with 2 km
  x 2 km resolution. 30 second time step.

• Components of net ecosystem exchange (NEE) and
  especially gross primary productivity, is
  computed in a way that is consistent with water
  and energy fluxes.
• Couples processes of radiative transfer (both
  direct and diffuse), photosynthesis,
  transpiration, leaf energy balance, soil energy
  balance and turbulence transfer. Uses Buckley
  stomatal conductance model instead of Ball.
• Structured canopy, multiple layers, exchange
  between canopy and the atmosphere computed
  directly, taking into account within canopy
  gradients of CO2 concentration, temperature,
  humidity and winds.
• 4 pool soil organic matter dynamics.
• Can represent biomass heat storage.

Acknowledgements

• The author would like to thank ORAU/ORISE as well
  as Mac Post and Lianhong Gu for their mentorship,
  Jeff Nichols for computer help, and the rest of
  the research group (Dan Ricciuto, Bai Yang, Tony
  King, etc.) for lots of input and patience.

ORNL is managed by the University of
Tennessee-Battelle LLC under contract
DE-AC05-00OR22725 with the U.S. Department of
Energy