The low-level circulation of the North American Monsoon as revealedby QuikSCATSimona Bordoni,1Paul E. Ciesielski,2Richard H. Johnson,2Brian D. McNoldy,2and Bjorn Stevens1Received 16 March 2004; revised 15 April 2004; accepted 22 April 2004; published 29 May 2004.[1] Five years (1999–2003) of near-surface QuikSCATocean winds over the Gulf of California and northeastPacific Ocean are used to characterize the changes in thelow-level circulation associated with the North AmericanMonsoon. Our analysis shows that the onset of the summerseason is accompanied by a seasonal reversal of the flowalong the Gulf of California, with the establishment of atime-mean southerly wind throughout the gulf. Thisreversal, not evident in the global reanalysis products,occurs in late sprin g and precedes the onset of themonsoonal rains. In the core of the monsoon, the time-mean flow is found to be modulated by transient events,namely gulf surges, detected in the near-surface wind fieldas periods of enhanced southerly flow which typicallyoriginate at the southern end of the gulf and propagatenorthward. The histogram of the summertime along-shorewinds identifies these surges as a distinct population ofevents, readily distinguishable from the backgroundflow.INDEX TERMS: 3322 Meteorology and AtmosphericDynamics: Land/atmosphere interactions; 3329 Meteorology andAtmospheric Dynamics: Mesoscale meteorology; 3360Meteorology and Atmospheric Dynamics: R emote sensing.Citation: Bordoni, S., P. E. Ciesielski, R. H. Johnson, B. D.McNoldy, and B. Stevens (2004), The low-level circulation of theNorth American Monsoon as revealed by QuikSCAT, Geophys.Res. Lett., 31, L10109, doi:10.1029/2004GL020009.1. Background[2] The North American Monsoon (NAM) is a regionalscale circulation that develops during the summer monthsover large areas of northwestern Mexico and the southwest-ern United States and is associa ted with a pronouncedincrease in rainfall over an otherwise very arid region.Although less impressive than its Asian counterpart, such acirculation is regarded as a monsoon circulation, based onseasonal reversal of pressure and wind patterns, energy andmass transfers, and typical regimes of rainfall and temper-ature [Adams and Comrie, 1997]. From an historical per-spective, attention originally focused on the observed abruptincrease in rainfall between June and July over Arizona. As aresult, the NAM literature has a strong geographical biastowards the southwestern United States (SW USA) and onlyrecently has the increased summertime precipitation over theSW USA been recognized as the northernmost and mostvariable extremity of a much broader phenomenon, centeredover northwestern Mexico [Douglas et al., 1993]. Becauseof the lack of data with sufficient horizontal resolution toadequately sample the atmospheric conditions over themonsoon area and because of the original focus on marginalregions rather than on the core of the monsoon, there are stillgaps in our basic understanding of the NAM, let alone itspotential relation to other features in the regional circulation.[3] Observations mainly collected during field studies aspart of t he SouthWest Area Monsoon Project (SWAMP)suggest the existence of a persistent time-mean southerlywind over the Gulf of California (GoC), believed to be theprimary mode of moisture transport to the western flanks ofthe Sierra Madre Occidental and to the arid southwest deserts[Douglas, 1995]. Superimposed on this mean moisture flux,transien t events, namely gulf surges, are observed. Gulfsurges along the GoC are northward surges of relatively cool,moist maritime air from the Tropical Pacific into the south-western desert region, which occur every summer during theNAM season. They were initially described in the early 1970s[Ha les, 1972] and their anomalous northward moisturetransport from the GoC is believed to support periods ofmonsoon bursts over the SW USA. Because of the sparsity oflower tropospheric observations over the GoC, a completedescription and understanding of the surge phenomenon arestill lacking. Most recent works have involved case studies,both from observational and modeling approaches, whereasstudies aimed at determining their structure and frequencyhave been relatively few [Douglas and Leal, 2003].[4] Modern scatt erometry (QuikSCAT) provides nearlydaily maps of the near-surface winds over the oceans at25-km spatial resolution and makes it possible for the firsttime to characterize with a high degree of accuracy thedynamical state of the lower troposphere over otherwisepoorly sampled regions [Liu, 2002]. In this study, we use fivesummers of QuikSCAT wind data (1999 –2003) to quantifythe changes in the low-level circulation associated with themonsoon, with particular focus on the GoC region, nowrecognized as the dominant low-level moisture source for themonsoon convection [Schmitz and Mullen, 1996; Berbery,2001]. Since the width of the GoC ranges from 150 to300 km, the spatial r esolution of QuikSCAT winds providesa considerable number of data points within the span of theGulf, thus allowing unprecedented spatial and temporalresolution of the summertime low-level circulation.[5] This study begins by constructing mean climatologiesfrom QuikSCAT of the low-level winds over the GoC, toexplore the time-mean structure of the flow associated withthe onset of the summertime precipitation and its relation-ship to the larger scale circulation over the northeast PacificGEOPHYSICAL RESEARCH LETTERS, VOL. 31, L10109, doi:10.1029/2004GL020009, 20041Department of Atmospheric Sciences, University of California at LosAngeles, Los Angeles, California, USA.2Department of Atmospheric Science, Colorado State University, FortCollins, Colorado, USA.Copyright 2004 by the American Geophysical Union.0094-8276/04/2004GL020009$05.00L10109 1of4in which it is embedded. The seasonal march of themonsoonal circulation is then examined by developing timeseries of the winds for the available years. Finally, histo-grams of the low-level winds are used to identify gulfsurges. Hovmo¨ller diagrams have been constructed tocharacterize their spatial and temporal structure and todetect their propagation speed up the GoC.2. Analysis[6] The QuikSCAT winds used in this study are obtainedfrom the NASA/JPL Level 3 daily, gridded wind vectorswith a horizontal grid spacing of 0.25°. In-situ studies[Ebuchi et al., 2002; Pickett et al., 2003; Bourassa et al.,2003] have shown that QuikSCAT wind retrievals areaccurate