Effect of Land Surface Processes on Meteorological Simulations

1
Institute for Multidimensional Air Quality Studies
Effect of Land Surface Processes on
Meteorological Simulations
Fang-Yi, Bonnie, Cheng, Dr. Daewon Byun and Dr.
Sharon Zhong Institute for Multi-dimensional Air
Quality Studies (IMAQS), University of Houston
Introduction
PBL Height Distribution
Configuration of Domain
CAMS Measurement Sites

• Currently MM5 uses a 1-km USGS dominant land use
  and land cover (LULC) database. HGA is
  represented as a totally impervious surface which
  causes unrealistic meteorological simulations.
• Meteorological simulations for the HGA provided
  by Texas AM (Dr. John Nielsen-Gammon) accurately
  predict maximum temperatures and reproduce the
  large-scale temperature patterns using a simple
  5-layer slab soil model.
• To better understand Houstons high ozone
  problem, and to understand the impact of urban
  vegetation, we use NOAH LSM in order to provide
  better meteorological simulations.
• In this study, a different meteorological
  simulation was conducted based on the usage of
  land surface modeling which provides lower
  boundary conditions.

UTC 2200, Aug. 25, 2000
D4
UTC 2200, Aug. 30, 2000
D3
D2
a. Analysis nudging for d1,d2,d3 b. observation
nudging of wind in d4 c. d1, d2 2-way nesting
d. d3,d4 continuous one-way nesting e. MRF PBL
Parameterization f. Dudhia explicit moisture
scheme g. RRTM radiation scheme

• 10 urban sites
• 18 rural sites
• 6 coast sites

Objective

• TAMU simulation the 2-m temperature is
  under-predicted from Aug. 29 to Aug. 31 causing
  the lower PBL height.
• UH simulation the added canopy water into the
  urban area successfully decreases the PBL height
  in the urban area. Also the added emissivity
  value successfully resolves the minimum
  temperature problem.

D4
Study the effects of land surface processes (land
use data, land surface related parameters) on the
meteorological conditions and on Houstons high
ozone concentration problem.
Design of Meteorological Simulation
Comparison of 2-m Temperature
Scattered Diagram of 2-m Temperature
Conclusion

• First (Base case) MM5 simulation coupled with
  the slab soil model (case S1). Special
  characteristics GOES satellite data assimilation
  technique was used in the simulation. The
  datasets were provided by TAMU.

• With USGS LULC, the HGA urban area was treated as
  a totally impervious surface. Therefore, we have
  large diurnal variations in temperature and very
  low latent/sensible heat flux ratio in the urban
  areas in the MM5 simulation.
• UH simulation a bias in daytime temperature is
  mostly fixed by adding canopy water in the urban
  area a lower bias in minimum temperature was
  fixed by increasing the emissivity value.
• GOES data assimilation uses the clear-air
  infrared skin temperature to successfully
  reproduce the earths surface characteristics.
  But under-prediction of temperature happens from
  Aug. 29 Aug. 31.

• Second Use the recently developed NOAH Land
  Surface Model (NOAH LSM) (EK, 2001) with
  identical inputs and model configurations as in
  the S1 case except using different land-surface
  parameterizations (S2).
• (Fei Chen modified land surface characteristics
  in urban areas heat capacity, thermal
  conductivity and shading factor)

Urban CAMS site average
Rural CAMS site average
Time sires of 2-m temperature averaged all over
CAMS sites
Urban Sites

• Third Based on the S2 simulation but modified
  land surface characteristics to compensate the
  simulation problem in the S2 simulation. This
  simulation was S3.
• added canopy water into the urban area 3x10-6
  from Aug. 22Aug. 28 2x10-6 Aug. 29 Sep 02
  (meter of available water per second)
• b. added emissivity in the simulation domain and

Future Work
Rural Sites

• In the future, we plan to replace the current
  1-km USGS 25 category databases with a new land
  use database. (inside the 8 HGA counties)
  (Steven, 2003).
• The land use related parameters (roughness
  length, albedo and shading factor, etc.) will be
  reallocated based on newer land use database.

• Urban sites Both simulations show good
  comparison with CAMS measurement data. For the UH
  simulation, the added emissivity improves the
  minimum temperature the added canopy water
  improves the maximum temperature. But on Aug. 31,
  maximum temperature is 2 degree higher than CAMS
  data in the UH simulation. The TAMU simulation
  has lower bias in min. and max. temperature from
  Aug. 29 to Aug. 31.
• Rural sites Both simulations have values close
  to the CAMS measurement sites. Again a similar
  problem in the TAMU simulation is that the
  maximum temperature is under-predicted from Aug.
  29 to Aug. 31.
• Coast sites TAMUs simulation has a good
  prediction in most of the coast sites. But the UH
  simulation gives a higher diurnal temperature
  variation.

Coastal Sites
Stephen Stetsen, GEM Pete Smith, Texas Forest
Service
Time sires of 2-m temperature for individual
sites