The Diurnal Cycle of Convection in NAME: Observations Versus Reanalysis

Previous studies of the North American Monsoon (NAM) have been

hampered by the lack of observations, particularly over Mexico.

During the summer of 2004, the North American Monsoon Experiment

(NAME) established an enhanced observational network over Mexico

and the southwestern United States aimed at determining the sources

and limits of predictability of warm season precipitation associated

with the NAM. Nested within the larger NAME souding network,

an enhanced budget array with 4-6 daily sonde launches was

deployed over the southern Gulf of California extending eastward

into northwestern Mexico to study the diurnal cycle over the

core of the monsoon region. These sonde data, along with profiler and

pibal observations, were objectively analyzed at 4-h, 25 hPa, 1 degree

resolution over the core of the monsoon region. In addition,

observations from 157 surface sites in this region along with twice-daily

QuikSCAT oceanic winds were quality controlled and processed into a gridded

data set covering this domain at 1-h, 0.25 degree resolution.

Using these data sets, the mean, temporal variability, and diurnal

characteristics of the monsoon flow are documented with detail not

previously possible.

Independent of model data over land, these objectively analyzed

products are compared to similar analyses from a special North American

Regional Reanlysis (NARR) which was produced for the same period. This

comparison reveals the NARR's tendency to initiate the daily convective

cycle too early in the day despite the strong constraints placed on the

NARR by assimilating observed precipitation. General circulation models

over this region exhibit a similar behavior reaching their convective

maxima 2-3 hours earlier than observed. Using primarily objectively

analyzed surface data and observed surface fluxes, we explore

possible reasons for this erroreous behavior in the reanalyses and

models in this monsoon region.