Marine atmospheric boundary layer height over the Eastern Pacific: Data analysis and model evaluation

Xubin Zeng, Michael A. Brunke, Mingyu Zhou, Chris Fairall, Nicholas A. Bond, Donald H. Lenschow

Research output: Contribution to journalArticlepeer-review

68 Scopus citations


The atmospheric boundary layer (ABL) height (h) is a crucial parameter for the treatment of the ABL in weather and climate models. About 1000 soundings from 11 cruises between 1995 and 2001 over the eastern Pacific have been analyzed to document the large meridional, zonal, seasonal, and interannual variations of h. In particular, its latitudinal distribution in August has three minima: Near the equator, in the intertropical convergence zone (ITCZ), and over the subtropical stratus/stratocumulus region near the west coast of California and Mexico. The seasonal peak of h in the ITCZ zone (between 5.6° and 11.2°N) occurs in the spring (February or April), while it occurs in August between the equator and 5.6°N. Comparison of these data with the 10-yr monthly output of the Community Climate System Model (CCSM2) reveals that overall the model underestimates h, particularly north of 20°N in August and September. Directly applying the radiosonde data to the CCSM2 formulation for computing h shows that, at the original vertical resolution (with the lowest five layers below 2.1 km), the CCSM2 formulation would significantly underestimate h. In particular, the correlation coefficient between the computed and observed h values is only 0.06 for cloudy cases. If the model resolution were doubled below 2.1 km, however, the performance of the model formulation would be significantly improved with a correlation coefficient of 0.78 for cloudy cases.

Original languageEnglish (US)
Pages (from-to)4159-4170
Number of pages12
JournalJournal of Climate
Issue number21
StatePublished - Nov 1 2004

ASJC Scopus subject areas

  • Atmospheric Science


Dive into the research topics of 'Marine atmospheric boundary layer height over the Eastern Pacific: Data analysis and model evaluation'. Together they form a unique fingerprint.

Cite this