Data Assimilation Strategies and Land-Surface Heterogeneity Effects in the Planetary Boundary Layer
Open Access
- Author:
- Reen, Brian P.
- Graduate Program:
- Meteorology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 28, 2007
- Committee Members:
- David R Stauffer, Committee Chair/Co-Chair
Kenneth James Davis, Committee Member
Toby Nahum Carlson, Committee Member
Christopher J Duffy, Committee Member - Keywords:
- spectra
ABL heterogeneity
obs nudging
mesoscale modeling
data assimilation
PBL heterogeneity - Abstract:
- Accurate simulation of the planetary boundary layer (PBL) is important for many meteorological applications including triggering convection, air chemistry, and transport and dispersion. Data assimilation of near-surface (~2 m above ground level) and within-PBL observations should improve simulation of the PBL, but due to difficulties including spatial and temporal variability, mass-field (temperature and water vapor mixing ratio) observations at these heights are often not utilized. <P> This thesis investigates 1D and 3D assimilation of near-surface and within-PBL mass-field observations using observation (obs) nudging and uses the best 3D forecast from these techniques to investigate the scales in the relationship between surface forcing and the PBL. These experiments use the nonhydrostatic Fifth-Generation Pennsylvania State University / National Center for Atmospheric Research Mesoscale Model MM5 for 29 May, 6 June, and 7 June 2002 during the International H2O Project over the Southern Great Plains. <P> It was found in 1D experimentation that near-surface mass-field observations should be compared to diagnostic model values at the same height to calculate the difference between the model and observations (the innovation) used in the assimilation. This innovation was best applied throughout the model-diagnosed PBL during free-convective conditions because of the well-mixed nature of the PBL; however, the innovation should otherwise only be applied at the lowest model level. Near-surface air temperature innovations were also applied to the top soil layer temperature to minimize disruption to the surface energy balance. Methods using model-diagnosed PBL height to assimilate within-PBL mass-field observations were tested but none of these methods performed consistently better than the default MM5 method of simply applying the innovation at the same height at each time. <P> The 3D forecasts created using the near-surface and within-PBL mass-field obs nudging techniques tested here were analyzed using both Fourier analysis and multiresolution decomposition to determine the correlation between surface forcing (surface buoyancy flux) and PBL height at different scales. This analysis indicated correlations less than 0.50 even at 128 km and increasing correlations at larger scales, but large day-to-day variability. The large scale PBL height structure was found to be largely dependent on surface forcing and atmospheric conditions. <P>