• Arkin, P. A., and B. N. Meisner, 1987: The relationship between large-scale convective rainfall and cold cloud over the Western Hemisphere during 1982–84. Mon. Wea. Rev.,115, 51–74.

  • ——, and P. E. Ardanuy, 1989: Estimating climatic-scale precipitation from space: A review. J. Climate,2, 1229–1238.

  • Augustine, J. A., and F. Caracena, 1994: Lower-tropospheric precursors to nocturnal MCC development over the central United States. Wea. Forecasting,9, 116–135.

  • Badan-Dangon, A., C. E. Dorman, M. A. Merrifield, and C. D. Winant, 1991: The lower atmosphere over the Gulf of California. J. Geophys. Res.,96 (C9), 877–896.

  • Bell, G. D., and J. E. Janowiak, 1995: Atmospheric circulation associated with the Midwest floods of 1993. Bull. Amer. Meteor. Soc.,76, 681–695.

  • Berbery, E. H., E. M. Rasmusson, and K. E. Mitchell, 1996: Studies of North American continental-scale hydrology using Eta model forecast products. J. Geophys. Res.,101 (D3), 7305–7319.

  • Bluestein, H. B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes. Vol. 2, Observations and Theory of Weather Systems, Oxford University Press, 594 pp.

  • Broccoli, A. J., and S. Manabe, 1992: The effects of orography on midlatitude Northern Hemisphere dry climates. J. Climate,5, 1181–1201.

  • Bryson, R. A., and F. K. Hare, 1974: The climates of North America. World Survey of Climatology, Vol. 2, Climates of North America, R. A. Bryson and F. K. Hare, Eds., Elsevier, 1–36.

  • Chen, T.-C., 1985: Global water vapor flux and maintenance during FGGE. Mon. Wea. Rev.,113, 1801–1819.

  • Douglas, A. V., and P. J. Englehart, 1995: An analysis of the starting date for the summer monsoon in western Mexico and southeast Arizona. Proc. 20th Annual Climate Diagnostics Workshop, Seattle, WA, Climate Prediction Center, 207–211.

  • ——, and ——, 1996: Variability of the summer monsoon in Mexico and relationships with drought in the United States. Proc. 21st Annual Climate Diagnostics Workshop, Huntsville, AL, Climate Prediction Center, 296–299.

  • Douglas, M. W., R. A. Maddox, K. W. Howard, and S. Reyes, 1993:The Mexican monsoon. J. Climate,6, 1665–1677.

  • Ebisuzaki, W., 1995: Consistency of the diabatic heating in the NMC/NCAR reanalysis. Proc. 20th Annual Climate Diagnostics Workshop, Seattle, WA, Climate Prediction Center, 232–234.

  • Fritsch, J. M., R. J. Kane, and C. R. Chelius, 1986: The contribution of mesoscale convective weather systems to the warm-season precipitation in the United States. J. Climate Appl. Meteor.,25, 1333–1345.

  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc.,106, 447–462.

  • Hahn, D. G., and S. Manabe, 1975: The role of mountains in the south Asian monsoon circulation. J. Atmos. Sci.,32, 1515–1540.

  • Heideman, K. F., and J. M. Fritsch, 1988: Forcing mechanisms and other characteristics of significant summertime precipitation. Wea. Forecasting,3, 115–130.

  • Hering, W. S., and T. R. Borden Jr., 1962: Diurnal variations in the summer wind field over the central United States. J. Atmos. Sci.,19, 81–86.

  • Higgins, R. W., Y. Yao, E. S. Yarosh, J. E. Janowiak, and K. C. Mo, 1997: Influence of the Great Plains low-level jet on summertime precipitation and moisture transport over the central United States. J. Climate,10, 481–507.

  • Holton, J. R., 1992: An Introduction to Dynamic Meteorology. Academic Press, 510 pp.

  • Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci.,38, 1179–1196.

  • ——, H. H. Hsu, I. N. James, M. Masutani, P. D. Sardeshmukh, and G. H. White, 1989: Diagnostics of the global atmospheric circulation based on ECMWF analyses 1979–1989. WMO/TD-No. 326, World Meteorological Organization, 217 pp.

  • Houze, R. A., Jr., and P. V. Hobbs, 1982: Organization and structure of precipitation cloud systems. Advances in Geophysics, Vol. 24, Academic Press, 225–315.

  • Janowiak, J. E., and P. A. Arkin, 1991: Rainfall variations in the Tropics during 1986–1989, as estimated from observations of cloud-top temperature. J. Geophys. Res.,96 (Suppl.), 3359–3373.

  • Johnson, R. H., 1984: Partitioning tropical heat and moisture budgets into cumulus and mesoscale components: Implications for cumulus parameterization. Mon. Wea. Rev.,112, 1590–1601.

  • ——, 1986: Lower-tropospheric warming and drying in tropical mesoscale convective systems: Implications for the problem of cumulus parameterization. J. Meteor. Soc. Japan,64, 721–726.

  • ——, and G. S. Young, 1983: Heat and moisture budgets of tropical mesoscale anvil clouds. J. Atmos. Sci.,40, 2138–2147.

  • Kalnay E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc.,77, 437–471.

  • Kiehl, J. T., 1992: Atmospheric general circulation modeling. Climate System Modeling, K. E. Trenberth, Ed., Cambridge University Press, 319–368.

  • Krishnamurti, T. N., 1971: Tropical east–west circulations during the northern summer. J. Atmos. Sci.,28, 1342–1347.

  • Lindzen, R. S., and S. Nigam, 1987: On the role of sea-surface temperature gradients in forcing low-level winds and convergence in the Tropics. J. Atmos. Sci.,44, 2418–2436.

  • Maddox, R. A., 1980: Mesoscale convective complexes. Bull. Amer. Meteor. Soc.,61, 1374–1387.

  • Means, L. L., 1952: On thunderstorm forecasting in the central United States. Mon. Wea. Rev.,80, 165–189.

  • Mo, K. C., J. Nogues-Paegle, and J. Paegle, 1995: Physical mechanisms of the 1993 summer floods. J. Atmos. Sci.,52, 879–895.

  • Mock, C. J., 1996: Climatic controls and spatial variations of precipitation in the western United States. J. Climate,9, 1111–1125.

  • Mosiño, P. A., and E. García, 1974: The climate of Mexico. World Survey of Climatology, Vol. 2, Climates of North America, R. A. Bryson and F. K. Hare, Eds., Elsevier, 373–404.

  • Negri, A. J., R. F. Adler, R. A. Maddox, K. W. Howard, and P. R. Keehn, 1993: A regional rainfall climatology over Mexico and the southwest United States derived from passive microwave and geosynchronous data. J. Climate,6, 2144–2161.

  • Nigam, S., 1994: On the dynamical basis for the Asian summer monsoon rainfall–El Niño relationship. J. Climate,7, 1750–1771.

  • ——, 1997: The annual warm to cold phase transition in the eastern equatorial Pacific: Diagnosis of the role of stratus cloud top cooling. J. Climate,10, 2447–2467.

  • Okabe, I. T., 1994: The North American monsoon. Ph.D. dissertation, University of British Columbia, 146 pp. [Available from Dept. of Geography, University of British Columbia, 2075 Wesbrook Place, Vancouver, BC V6T 1W5, Canada.].

  • Pan, H.-L., and W.-S. Wu, 1994: Implementing a mass flux convective parameterization package for the NMC medium-range forecast model. Preprints, 10th Conf. on Numerical Weather Prediction, Portland, OR, Amer. Meteor. Soc., 96–98.

  • Pitchford, K. L., and J. London, 1962: The low-level jet as related to nocturnal thunderstorms over the Midwest United States. J. Appl. Meteor.,1, 43–37.

  • Rasmusson, E. M., and K. C. Mo, 1996: Large-scale atmospheric moisture cycling as evaluated from global NMC analysis and forecast products. J. Climate,9, 3276–3297.

  • Reyes, S., and D. L. Cadet, 1986: Atmospheric water vapor and surface flow patterns over the tropical Americas during May–August 1979. Mon. Wea. Rev.,114, 582–593.

  • ——, and ——, 1988: The southwest branch of the North American monsoon during summer 1979. Mon. Wea. Rev.,116, 1175–1187.

  • Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analyses using optimum interpolation. J. Climate,7, 929–948.

  • Rodwell, M. J., and B. J. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc.,122, 1385–1404.

  • Rowson, D. R., and S. J. Colucci, 1992: Synoptic climatology of thermal low-pressure systems over south-western North America. Int. J. Climatol.,12, 529–545.

  • Schmitz, J. T., and S. L. Mullen, 1996: Water vapor transport associated with the summertime North American monsoon as depicted by ECMWF analyses. J. Climate,9, 1621–1634.

  • Song, J.-L., and W. M. Frank, 1983: Relationships between deep convection and large-scale processes during GATE. Mon. Wea. Rev.,111, 2145–2160.

  • Stensrud, D. J., 1996: Importance of low-level jets to climate: A review. J. Climate,9, 1698–1711.

  • ——, R. L. Gall, S. L. Mullen, and K. W. Howard, 1995: Model climatology of the Mexican monsoon. J. Climate,8, 1775–1794.

  • Tang, M., and E. R. Reiter, 1984: Plateau monsoons of the Northern Hemisphere: A comparison between North America and Tibet. Mon. Wea. Rev.,112, 617–637.

  • Tiedtke, M., 1989: A comprehensive mass flux scheme for cumulus parameterization in large-scale models. Mon. Wea. Rev.,117, 1779–1800.

  • Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev.,103, 406–419.

  • Williams, E., and N. Renno, 1993: Analysis of the conditional instability of the tropical atmosphere. Mon. Wea. Rev.,121, 21–36.

  • Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate,9, 840–858.

  • Xu, K.-M., and K. A. Emanuel, 1989: Is the tropical atmosphere conditionally unstable? Mon. Wea. Rev.,117, 1471–1479.

  • Yanai, M., S. Esbensen, and J. H. Chu, 1976: Response of deep and shallow tropical maritime cumuli to large-scale processes. J. Atmos. Sci.,33, 976–991.

  • View in gallery

    (a) Mean (1979–85) July GHCN station precipitation gridded on a 2.5° lat × 2.5° long grid. The contour interval is 1 mm day−1 and values greater than 2 mm day−1 are shaded. (b) Mean (1979–85) July residually derived ECMWF diabatic heating at 500 hPa. The contour interval and shading threshold is 0.5 K day−1, and the zero contour is omitted. (c) As in (a) but for the NCEP reanalysis model precipitation. (d) As in (a) but for the ECMWF reanalysis model precipitation.

  • View in gallery

    Mean June to July (onset phase) changes in (a) gridded GHCN precipitation, (b) GPI precipitation, (c) the NCEP reanalysis model precipitation, and (d) the ECMWF reanalysis model precipitation. Climatological periods are 1950–85 for the GHCN data, 1986–96 for the GPI data, and 1979–93 for both the ECMWF and NCEP reanalyses. The contour interval and shading threshold is 1 mm day−1, and the zero contour is omitted in all panels. The GPI estimate is not produced for latitudes north of 40°N, and the GHCN data, as station data, is not available over the oceans.

  • View in gallery

    (a) The mean (1986–93) pentad evolution of precipitation averaged over the Mexican monsoon region (outlined in Fig. 1a) is shown along with the standard deviation. (b) The mean evolution of precipitation averaged over the central United States (also outlined in Fig. 1a) is shown using filled circles, while that over the Mexican monsoon region is shown with open circles, with a +1/−1 pentad running mean applied to both.

  • View in gallery

    (a) Mean July CAPE, calculated from the ECMWF monthly mean reanalysis, 1979–93, and shown using a contour interval of 200 J kg−1, with values greater than 1000 J kg−1 shaded. (b) Mean June to July change in the ECMWF CAPE, shown using a contour interval and shading threshold of 200 J kg−1. (c) Mean July 500-hPa pressure vertical velocity (omega), calculated from ECMWF monthly means, 1979–93. The contour interval and shading threshold is 0.005 Pa s−1, note that negative values denote upward motion. (d) Mean July omega at 27°N, from the ECMWF reanalysis, 1979–93. The contour interval and shading threshold is 0.005 Pa s−1. Blank areas represent orography. The zero contour is omitted in all panels.

  • View in gallery

    (a) The mean (1979–93) vertical structure of heating along a diagonal swath (line AB in Fig. 1a) whose west side is representative of the monsoon region; middle part, of the south-central United States; and east side, of the southeastern United States and northern Florida: (a) for the residually derived heating from the ECMWF reanalysis; (b) for the residually derived heating from the NCEP reanalysis; and (c) for the NCEP model produced diabatic heating. The contour interval and shading threshold is 0.5 K day−1, and the zero contour is omitted.

  • View in gallery

    The vertical profile for mean (1979–93) July conditions at 22°N, 105°W of residually derived heating (solid line with + marks), the adiabatic cooling (−N2ω) (dotted line with o marks), the NCEP model produced heating (solid line with no marks), and omega (dotted line with + marks) for (a) ECMWF reanalysis and (b) NCEP reanalysis. (c) and (d) As in (a) and (b) but for a midlatitude location (42°N, 100°W) with horizontal temperature advection (V·T) shown instead of the adiabatic cooling term (dotted line with o marks). The x axis markings to the left of the origin are at 0.02 Pa s−1 intervals. The markings to the right of the origin are at 1.0 K day−1. (e) The July mean (1979–93) adiabatic cooling (−N2ω) at 500 hPa from the ECMWF reanalysis. (f) The July mean horizontal temperature advection (V·T) at 500 hPa from the same dataset. The contour interval and shading threshold is 0.5 K day−1, and the zero contour is omitted in these panels.

  • View in gallery

    The mean July and the June to July change in the ECMWF wind and geopotential height at (a), (b) 200 hPa, (c), (d) 700 hPa, and (e), (f) 925 hPa. The blank area in the 925-hPa panels represents the intersection of orography with that pressure level. The height interval and the vector wind scale (in m s−1) are indicated for each panel.

  • View in gallery

    The mean July and the June to July change in the vertically integrated moisture flux convergence from (a), (d) ECMWF reanalysis (1979–93) and (b), (c) NCEP reanalysis (1979–93). The vectors represent the divergent part of the vertically integrated flux. The contour interval and shading threshold is 2 mm day−1, and the zero contour is omitted in all panels. The vertical integral was computed from pressure level data with the surface pressure as the lower boundary. Note that the same vector length denotes a divergent moisture flux of 60 kg m−1 s−1 in the July fields, and only 40 kg m−1 s−1 in the difference fields.

  • View in gallery

    (a) Mean (1982–95) July SST and (b) the mean June to July change. In (a) the contour interval is 1°C, with values over 28°C shaded. In (b) the contour interval is 0.5°C, with magnitudes greater than 1.5°C shaded. The OISST dataset was used for SSTs.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 335 335 22
PDF Downloads 224 224 21

Evolution of the North American Monsoon System

View More View Less
  • 1 Cooperative Institute for Climate Studies, Department of Meteorology, University of Maryland, College Park, Maryland
© Get Permissions
Full access

Abstract

A dynamically oriented description of the North American summer monsoon system, which encompasses the Mexican monsoon and the associated large-scale circulation over the continental United States, is provided by developing an evolution climatology of the precipitation, tropospheric circulation, moisture fluxes, diabatic heating, convective environment, and the adjoining basin SSTs.

A distinguishing aspect of this study is the amount of independent data analyzed, such as the newly available European Centre for Medium-Range Weather Forecasts (ECMWF) reanalyses, the National Centers for Environmental Prediction (NCEP) reanalyses, both satellite-derived and station data–based precipitation estimates, and the heating diagnosed from both reanalyses. This also provides a preliminary evaluation and comparison of the newly available NCEP and ECMWF reanalyses at the regional level, including the model-generated precipitation and heating distributions. The principal findings are the following.

  • The accompaniment of the Mexican monsoon onset by decreased precipitation to the east is shown to be a robust climatological feature. This striking linkage is also evident in the associated tropospheric circulation and, notably, in the upper-level heating fields. The climatological phasing of the precipitation between the two areas is coherent even at the pentad timescale.
  • While the Mexican monsoon onset is closely associated with thermodynamic favorability, the linkage to the central United States, as reflected in the vertical velocity and the low-level height fields, appears to be consistent with several possible forcings: the monsoon deep heating, the elevated heating of the North American cordillera and plateau, and orographic forcing associated with the seasonal movement of the easterlies encroaching on the North American cordillera.
  • Although both reanalyses yield a tropical-type deep tropospheric heating distribution in the Mexican monsoon region and, therefore, a potentially prominent role for the monsoon in the regional circulation, the considerable differences in the diagnosed heating vertical structure, thermodynamic balance, and the overall heating magnitude between the two reanalyses, and even between the NCEP reanalysis-consistent heating and the NCEP model-produced heating, suggest potentially significant differences in the implied dynamics of the North American monsoon system.

Corresponding author address: Mathew Barlow, Department of Meteorology, University of Maryland, College Park, MD 20742-2425.

Email: barlow@atmos.umd.edu

Abstract

A dynamically oriented description of the North American summer monsoon system, which encompasses the Mexican monsoon and the associated large-scale circulation over the continental United States, is provided by developing an evolution climatology of the precipitation, tropospheric circulation, moisture fluxes, diabatic heating, convective environment, and the adjoining basin SSTs.

A distinguishing aspect of this study is the amount of independent data analyzed, such as the newly available European Centre for Medium-Range Weather Forecasts (ECMWF) reanalyses, the National Centers for Environmental Prediction (NCEP) reanalyses, both satellite-derived and station data–based precipitation estimates, and the heating diagnosed from both reanalyses. This also provides a preliminary evaluation and comparison of the newly available NCEP and ECMWF reanalyses at the regional level, including the model-generated precipitation and heating distributions. The principal findings are the following.

  • The accompaniment of the Mexican monsoon onset by decreased precipitation to the east is shown to be a robust climatological feature. This striking linkage is also evident in the associated tropospheric circulation and, notably, in the upper-level heating fields. The climatological phasing of the precipitation between the two areas is coherent even at the pentad timescale.
  • While the Mexican monsoon onset is closely associated with thermodynamic favorability, the linkage to the central United States, as reflected in the vertical velocity and the low-level height fields, appears to be consistent with several possible forcings: the monsoon deep heating, the elevated heating of the North American cordillera and plateau, and orographic forcing associated with the seasonal movement of the easterlies encroaching on the North American cordillera.
  • Although both reanalyses yield a tropical-type deep tropospheric heating distribution in the Mexican monsoon region and, therefore, a potentially prominent role for the monsoon in the regional circulation, the considerable differences in the diagnosed heating vertical structure, thermodynamic balance, and the overall heating magnitude between the two reanalyses, and even between the NCEP reanalysis-consistent heating and the NCEP model-produced heating, suggest potentially significant differences in the implied dynamics of the North American monsoon system.

Corresponding author address: Mathew Barlow, Department of Meteorology, University of Maryland, College Park, MD 20742-2425.

Email: barlow@atmos.umd.edu

1. Introduction

The summertime circulation centered on northwest Mexico has been frequently identified as monsoonal due to the following characteristic features: the precipitation is largely confined to a single season, the boreal summer (Mosiño and García 1974; Tang and Reiter 1984; Douglas et al. 1993); the highest surface temperatures occur immediately before rainfall onset (Douglas et al. 1993);the surface winds in the northern Gulf of California undergo seasonal reversal (Tang and Reiter 1984; Badan-Dangon et al. 1991); a local surface low pressure center exists during the summer season (Rowson and Colucci 1992; Okabe 1994); and reasonably strong upper-level outflow (Krishnamurti 1971) accompanies the upper-level high (e.g., Bryson and Hare 1974). The Mexican monsoon, the land–sea regime along the gulf coasts of Mexico and the United States (Tang and Reiter 1984), and the associated summertime circulation features over the continental United States are together referred to as the North American monsoon system.

The summertime precipitation over North America exhibits a considerable range of magnitudes, as evident from Fig. 1a, which displays the observed mean July precipitation for the 1979–85 period. Although primarily convective in nature even in the more northerly latitudes of the contiguous United States (Heideman and Fritsch 1988), the precipitation is associated with a diverse array of processes—for example, a monsoonal land sea regime along the U.S. gulf coast (Tang and Reiter 1984), the nocturnal formation of the Great Plains low-level jet between the Rockies and the Mississippi River (Means 1952; Pitchford and London 1962; Hering and Borden 1962; Wallace 1975; Augustine and Caracena 1994; Stensrud 1996), and frontal and mesoscale (e.g., convective feedback, dryline) processes in the cental and eastern United States (Heideman and Fritsch 1988). Mesoscale convective complexes, developing regions of stratiform rain in their mature stages (Maddox 1980), in fact account for about 50%–60% of the warm season rainfall over the central United States (Fritsch et al. 1986).

A significant area of intense precipitation is also centered on northwest Mexico (Mosiño and García 1974; Negri et al. 1993; Douglas et al. 1993), with mean values exceeding 5 mm day−1 at several stations; the precipitation is even more intense over southern Mexico where mean precipitation rates can be as large as 10 mm day−1. At the same time, a well-defined geographical minimum in precipitation is observed over the highland plateau in north-central Mexico (Mosiño and García 1974). Precipitation is also large over parts of the United States, with the largest values (∼5 mm day−1) along the Gulf Coast and the eastern seaboard; a relative maximum in summertime precipitation is also evident over the Midwest (e.g., Bryson and Hare 1974). The reasons for such a climatological distribution, particularly the interrelationships between the various elements, are not fully understood at the present time.

The model-produced precipitation fields from the National Centers for Environmental Prediction (NCEP) and European Centre for Medium-Range Weather Forecasts (ECMWF) reanalyses during 1979–85 (Figs. 1c and 1d, respectively) have a similar large-scale pattern, but notable differences in both the magnitude of the various features and the small-scale structure are evident throughout the region. The ECMWF reanalysis model precipitation modestly underpredicts the precipitation in northwest Mexico and overpredicts the precipitation in the southeastern United States, whereas the NCEP reanalysis model precipitation overpredicts the precipitation in both areas, by as much as a factor of 2.

The distribution of the residually diagnosed diabatic heating from the ECMWF reanalysis at the level at which deep convective heating is expected to be a maximum in the Tropics–subtropics (500 hPa)1 is shown in Fig. 1b.2 Over Mexico, and particularly along its western coast, the significant values of deep heating attest to the vigor of deep convection in generating local precipitation. Similarly, the prominent 500-hPa heating maximum over the north-central United States can be attributed to the well-known vigor of summertime convection in that area. The 500-hPa heating features over the western United States, on the other hand, are not necessarily indicative of the deep convective nature of heating (precipitation), in view of the significant elevations in this region.

The source of moisture for the summertime precipitation has been difficult to determine, particularly over the southwestern regions where the strong diurnal cycle and rapid variations of orography severely limit the accuracy of the moisture transport calculations, which already suffer from the poor vertical sampling of the lower troposphere (e.g., Berbery et al. 1996). As yet, there is little agreement on the source of moisture for the Mexican monsoon, with conclusions ranging from the eastern Pacific as the primary source, with significant but not dominant contribution from the Gulf of Mexico [Reyes and Cadet (1988), from analysis of FGGE3 observations; Schmitz and Mullen (1996), from diagnosis using ECMWF analysis] to the Gulf of California as the primary source, with essentially no contribution from the eastern Pacific and little contribution from the Gulf of Mexico [Stensrud et al. (1995), from mesoscale modeling].

The transition into the summertime regime is rather dramatic over Mexico and the southern United States. The Mexican monsoon has a rapid onset that is accompanied by broad belts of decreased precipitation to the north and east (Tang and Reiter 1984; Mock 1996). This out-of-phase relationship is also present in the accompanying 200-hPa divergence field, and with opposite polarity (particularly, over Mexico and the south-central United States) during the Mexican monsoon decay phase (Okabe 1994). A qualitatively similar out-of-phase relationship between the Mexican monsoon and central United States precipitation is apparently evident even in the interannual variability linear correlation statistics (Douglas and Englehart 1996) and, interestingly, also in the recent analysis of contrasting drought and flood conditions in the U.S. midwest [see Fig. 12 in Mo et al. (1995) or Fig. 16 in Bell and Janowiak (1995)].

The correspondence between the precipitation patterns, boundary layer structure, and orography of the North American Plateau and the Tibetan Plateau led Tang and Reiter (1984) to characterize northwest Mexico, the western United States, and portions of the central United States as under the influence of a plateau monsoon regime. The Tibetan Plateau has been shown to be critical for the abrupt nature of the Asian summer monsoon onset (Hahn and Manabe 1975), which, given Tang and Reiter’s (1984) assertion regarding similarities between the Tibetan Plateau monsoon and the western North American circulation, raises the question of the North American Plateau’s role in the Mexican monsoon onset. Indeed, the mountain/no-mountain general circulation model experiments of Broccoli and Manabe (1992) indicate that orography is essential for a vigorous Mexican monsoon circulation (cf. their Fig. 10), but it is difficult to determine from these experiments if it is the mechanical or the thermodynamic impact of orography (through elevated sensible heating) that is critical for the occurrence of a vigorous Mexican monsoon.

The objective of this dynamically oriented study of the evolution of the North American monsoon system is to (a) determine the robust features of the circulation and precipitation climatology over the entire North American monsoon region, (b) develop stable climatologies of the derived quantities such as moisture fluxes and the 3D diabatic heating using both the ECMWF and NCEP reanalysis data, (c) ascertain the dominant thermodynamic balances operating in various sectors of the monsoon system, and (d) advance the dynamical understanding of the out-of-phase relationship between interannual fluctuations of the Mexican monsoon and the U.S. great plains (and south-central United States) precipitation by analyzing the circulation change and its potential forcing during the climatological onset and decay phases of the Mexican monsoon.

The datasets used in this study are described in section 2, which also contains a brief description of the method used in diagnosing diabatic heating from the reanalyses. The seasonal evolution of precipitation over the entire North American monsoon region is presented in section 3 from four different estimates; this section is followed by an examination of the changes in the convective environment during the Mexican monsoon onset period in section 4. The heating vertical structure of both the diagnosed heating from the ECMWF and the NCEP reanalyses and the NCEP reanalysis model-generated heating is described in section 5. The regional extent in which the thermodynamic balance is of predominantly tropical nature (i.e., vertical motions balancing diabatic heating) is ascertained in section 5b for both heating fields. The evolution of winds and geopotential height in the planetary boundary layer (925 hPa), and in the lower and upper troposphere (700 hPa and 200 hPa, respectively) are displayed in section 6. The seasonal evolution of the moisture fluxes and SST are described in sections 7 and 8, respectively. Discussion and conclusions follow in section 9.

2. Datasets

a. Precipitation estimates

The monthly precipitation station data used in this study are extracted from the Global Historical Climatology Network (GHCN) dataset, available from the National Climatic Data Center. To maximize both the number of stations and the time period for Mexican stations while retaining station consistency, all stations that began reporting on or before 1950, and ended reporting on or after 1985, were chosen. While this choice excludes several years of data from the original GHCN records, the resulting 274 stations for Mexico and 1473 stations for the United States have consistent data coverage. These point measures of precipitation were gridded onto a 2.5° × 2.5° spatial grid using Cressman analysis, with radii of influence empirically determined to minimize the influence of adjacent stations over areas of no data, so that there is almost no smoothing included in the process. The climatological monthly mean values were computed from the gridded values, and these differed little from the gridded values of the climatological monthly mean station data. Due to Mexico’s rapidly varying orography, the relatively small number of stations, and the clustering of the precipitation recording stations, and because of the difficulties in relating point precipitation measurements to areal averages (Arkin and Ardanuy 1989), the gridded station precipitation data should be considered to be only an approximate representation of the actual precipitation field.

The GOES Precipitation Index (GPI) rainfall estimates (Arkin and Meisner 1987; Janowiak and Arkin 1991) are based on geostationary satellite infrared coverage, and are available on a 2.5° × 2.5° spatial grid in the 40°S–40°N lat band, both as pentad (5 days) values (1986–93) and as monthly values (1986–96). Other datasets, such as the merged monthly precipitation data (Xie and Arkin 1996) were also examined, but the GPI estimates were used in this study as they compared favorably with the other datasets, and had the additional advantage of depicting variability at the pentad timescale.

b. ECMWF reanalyses

The 6-h ECMWF’s surface and upper-air reanalyses archived at the National Center for Atmospheric Research on a 2.5° × 2.5° global grid and at 17 pressure levels during 1979–93 were used to compute the monthly values of geopotential height, wind, temperature, specific humidity, and model-produced precipitation. Since observations are very sparse over Mexico, the low-level wind analysis, in particular, reflects the model’s bias, for example, off the western coast of Mexico, there are significant differences between the ECMWF reanalysis and Douglas et al.’s (1993) analysis based on a denser network of rawinsonde observations. The ECMWF reanalysis uses intermittent statistical (optimum interpolation) analysis with 6-h cycling, 1D variational physical retrieval of TIROS Operational Vertical Sounder cloud cleared radiances, and diabatic, nonlinear normal mode initialization, in conjunction with a T106 resolution (∼120 km) spectral model, with 31 vertical hybrid levels. Other potentially pertinent features of the model are a prognostic scheme for clouds, and a mass flux convection scheme (Tiedtke 1989) (see the ECMWF September 1996 newsletter for a full description of the ECMWF reanalysis).

In the absence of an ECMWF counterpart to the NCEP model-produced heating, the diabatic heating that is consistent with the analyzed large-scale circulation was residually diagnosed (at 2.5° lat × 5° long resolution) in order to compare the heating fields in the NCEP and ECMWF reanalyses.

c. NCEP reanalyses

NCEP 6-h reanalysis fields for the 1979–93 period, available on a 2.5° × 2.5° global grid and at 17 pressure levels (Kalnay et al. 1996), were used to provide another estimate of winds, geopotential height, moisture, (model produced) precipitation, and diabatic heating. As with the ECMWF reanalysis, the low-level winds in the Gulf of California are not in close agreement with observations. The NCEP reanalysis uses a spectral statistical interpolation scheme in conjunction with a T62 resolution (∼210 km) spectral model with 28 vertical sigma levels (Kalnay et al. 1996). Other potentially pertinent features of the model are a diagnostic scheme for clouds, and a simplified Arakawa–Schubert cumulus convection scheme (Pan and Wu 1994) (see Kalnay et al. 1996 for a detailed description of the NCEP reanalysis).

The NCEP reanalysis additionally contains the 3D diabatic heating field that is generated during a 6-h model forecast starting from each time step’s analysis, and is available partitioned into six components: large-scale condensation, deep convective, shallow convective, vertical diffusion, and shortwave and longwave heating rates. However, in view of potential differences between the model-produced and the analysis-consistent heating (e.g., Ebisuzaki 1995), and, in order to directly compare with the ECMWF analysis-consistent heating, the diabatic heating that is consistent with the NCEP analyzed large-scale circulation was also residually diagnosed.

d. Residual diagnosis of diabatic heating

The 3D diabatic heating is residually diagnosed from the thermodynamic equation (e.g., Hoskins et al. 1989;Nigam 1994) using the analyzed vertical velocity (ω):
i1520-0442-11-9-2238-e1
This diagnosis, however, does not provide any information about the constituent sensible, latent, and radiative heating components whose knowledge should be helpful in understanding the evolution, and particularly, the abrupt onset feature of the Mexican monsoon.

In the absence of an observational counterpart, the credibility of a heating estimate can only be convincingly evaluated by the extent of its dynamical consistency with the observed/analyzed large-scale circulation—that is, only through modeling, and such a modeling evaluation4 shows the diagnosed 3D heating to be a rather accurate estimate (Nigam 1994, 1997).

e. Sea surface temperature

The optimum interpolation sea surface temperature (OISST, Reynolds and Smith 1994) available from NCEP contains weekly SST values on a 1° lat × 1° long grid for the 1982 onward period. Monthly means were formed from this weekly data for the 1982–95 period.

This study attempts to determine the robust features of the climatological evolution of the North American monsoon system through intercomparison of datasets, for which the datasets should ideally be of the same duration. For the data-sparse regions considered in this study, such a restriction would be severe as it would preclude the use of some datasets in developing a convincing analysis. As such, we have undertaken comparisons of climatologies despite the absence of complete overlap between the concerned datasets.

3. Seasonal evolution of precipitation

The Mexican monsoon rainfall is heaviest during July and August. The seasonal transition into and out of this heavy rainfall period is accompanied by striking large-scale precipitation and circulation variations over the North American monsoon region. Figure 2 displays the precipitation change during the Mexican monsoon onset, as captured by the June to July change, manifest in four different estimates of precipitation: the GHCN gridded station data, the GPI IR-derived data, and the model precipitation from the two reanalyses. As the Mexican monsoon onset occurs in late June (Douglas and Englehart 1995), the June to July monthly change accurately captures the monsoon onset.

a. Onset phase

As the GPI and GHCN data periods do not overlap, and neither has more than an 8-yr overlap with the reanalyses, the June to July change is shown in Fig. 2 for the entire time period of each dataset (1950–85 for GHCN, 1986–96 for GPI, and 1979–93 for the two reanalyses). Comparisons with a fifth dataset that spans part of the time periods of all the data [the gridded 1969–93 hourly precipitation data (HPD), available from the Climate Prediction Center, not shown] suggest that, although there is interdecadal variablity in the June to July change pattern, the differences between time periods are significantly smaller than the differences between the datasets. The noncontemporaneous nature of the comparison in Fig. 2, thus, does not critically affect a qualitative comparison of the different estimates.

Given the uncertainties in the observed estimates of precipitation and the differing time periods, there is heartening agreement on the primary features of precipitation change over the United States and Mexico: a dramatic increase over northwest Mexico, scattered5 decreases to the north and east, and modest increases over the northeastern coast of the Gulf of Mexico. Even the dipole structure over Central America is stable throughout the different periods and captured in both models. The magnitude and specific location of the maximum change does vary considerably from dataset to dataset, but, given the considerable uncertainties in even the observed estimates, it is not possible to quantitatively evaluate the accuracy of the various estimates. Qualitatively, the ECMWF reanalysis seems to have an overvigorous evolution of the eastern Pacific ITCZ, while the NCEP reanalysis seems to have an overly weak evolution in the same area. The NCEP reanalysis slightly overpredicts the precipitation change in the southeast United States and, in fact, greatly exaggerates the total precipitation in that area (see Fig. 1c). Higgins et al. (1997) have linked the overprediction of the southeastern United States precipitation in the NCEP reanalysis (by a factor of ∼2) to the model’s excessively large diurnal cycle, which seems reasonable, as northwest Mexico, another area with a significant diurnal cycle, also appears to have excessive NCEP model precipitation, at least in the northern extent (cf. Fig. 1c).

The decay phase of the monsoon, not shown for reasons of brevity, qualitatively resembles the reverse of the onset, with decreases in precipitation over northwest Mexico accompanied by increases to the east. Again, the different estimates disagree on the exact placement and magnitude of the changes but agree on the qualitative pattern.

b. Pentad evolution

Given the rapid onset of the Mexican monsoon, it is of interest to examine the climatological evolution of monsoon precipitation at timescales shorter than the monthly timescale. Figure 3a shows the area-averaged monsoon precipitation evolution in the climatological (1986–93) GPI pentad data, along with the precipitation standard deviation in each pentad. The monsoon region is defined to consist of all grid boxes in western Mexico that experienced a June to July increase in GPI precipitation of at least 2 mm day6, and is marked in Fig. 1a. Although areal averaging facilitates description, it can potentially blur the rapid onset feature as the averaging region extends over 10 degrees of latitude.

Figure 3a shows the Mexican monsoon onset to occur primarily in June—in agreement with the findings of Douglas and Englehart (1995), despite the possible blurring of the onset—followed by relatively constant areally averaged precipitation through July and August. The monsoon decay is comparatively more gradual, beginning in early September and continuing until mid-October. The standard deviation remains large throughout the monsoon’s life cycle, particularly during the transition phases, but is seldom larger than half of the peak mean value (∼10 mm day−1). Examination of the precipitation evolution during each of the eight summers (1986–93) revealed that, except for 1986, none of the years experienced a “monsoon break” during July and August, where a break is defined here as at least one pentad during which the average over the whole monsoon area is less than 25% of the peak mean value (∼2.5 mm day−1). In fact, this remains true even if the break condition is weakened to 50% of the peak mean value, demonstrating the striking consistency of the monsoon precipitation when considered in aggregate.

The climatological (1986–93) evolution of the out-of-phase relationship between the Mexican monsoon and the south-central United States (28°–40°N, 102°–90°W, and marked in Fig. 1a) precipitation is examined by overlaying the two areally averaged pentad evolutions in Fig. 3b. Though not as marked as the precipitation increase over Mexico, the south-central U.S. precipitation weakens with the beginning of the Mexican monsoon onset in early June, and the out-of-phase evolution is evident until August. This coherent phasing of the central U.S. precipitation decrease with the monsoon increase at the pentad timescale was verified with the longer-term (1974–96) outgoing longwave radiation pentad data. The evolution of this linkage during the monsoon decay phase is not well represented in this figure as the eastern center-of-action of the linkage shifts southward of the chosen averaging region.

4. The convective environment

To investigate the observed distribution and evolution of precipitation, both the thermodynamic potential of the environment for convection and the large-scale dynamic forcing of the environment, which may impede or enhance the actualization of that potential, are examined in this section. Convective available potential energy (CAPE), the work done by the buoyancy force on a parcel once it has been lifted to the level where it becomes positively buoyant, is calculated as an estimate of the thermodynamic potential, a necessary but not sufficient condition for deep convection [e.g., the analysis of Song and Frank (1983), which showed that CAPE values of around 970 J kg−1 were necessary for convection in the eastern Atlantic GATE area]. The large-scale vertical velocity, an important factor in the occurrence of precipitation, particularly in the Tropics (e.g., Yanai et al. 1976), is examined to provide an estimate of the large-scale dynamic control.

As noted in the introduction, a variety of smaller-scale, rapidly evolving systems, such as the Great Plains low-level jet, are known to be important in determining the convective activity throughout the study area. Since the resolution in both time and space of the reanalyses and of their underlying observations is unlikely to resolve more than the overall character of these smaller-scale features, CAPE and large-scale vertical velocity are examined here as estimates of the more general, larger-scale mechanisms that modulate deep convection in these regions.

a. Convective available potential energy

The mean July CAPE, calculated from ECMWF data, is shown in Fig. 4a (cf. the precipitation in Fig. 1); the computational details are given in the appendix. There are large values over the tropical ocean and coastal areas, where the temperature and the moisture content near the surface is large,7 and a local maximum in the central United States, where there is a moderate amount of moisture and relatively high surface temperatures. The computed values are consistent with the 1000–3000 J kg−1 range typical of moderate to strong convection in the midlatitudes (Bluestein 1993). As tropical ocean convection has been shown to be strongly modulated by both the diurnal cycle and larger-scale forcing (Song and Frank 1983), the preference for coastal precipitation despite the large values of CAPE over open ocean regions suggests a prominent role for land–sea circulations in actualization of the thermodynamic potential in these regions. It is noteworthy that, despite problems with the representation of the planetary boundary layer (PBL) structure in both reanalyses, particularly over the Gulf of California, the computed CAPE appears reasonable. The NCEP CAPE, not shown, is broadly similar but with considerably larger maximum values (by a factor of 2 over northwest Mexico and the southeastern United States) and a well-defined local maximum over the southeastern United States.

The June to July change in CAPE (Fig. 4b) shows that the monsoon onset is well captured, with the large values of July CAPE in northwest Mexico resulting from an increase from near-zero values in June. The change in CAPE in the monsoon area is essentially controlled by the large increase in low-level moisture, as the surface temperatures are already high in June. Elsewhere, the only large deviations from the June CAPE occur over the north-central region of the United States and off the east coast. The out-of-phase relationship between NW Mexico and the region to the east is reproduced, but, while the monsoon increase is well captured, the small values in the region of negative change to the east, which results in July CAPE values still above 2000 J kg−1 along the coastal regions of Texas and NE Mexico, suggest that CAPE, alone, is not sufficient to diagnose the out-of-phase linkage in the June to July precipitation change field.

When examining CAPE, it is also necessary to evaluate the work required to raise the parcel to the level where it becomes positively buoyant, the Convective Inhibition Energy (CINE) (e.g., Bluestein 1993), which must be overcome before the CAPE can be realized. The June to July change in CINE (not shown) over northwest Mexico reflects a northwestward shift of an area of considerable inhibition from over NW Mexico in June to over Baja California and the northern Gulf of California in July. In both June and July, CINE has values greater than 20 J kg−1 for much of the central United States, and large values (>40 J kg−1) in a narrow region along the eastern boundary of the Rockies and in Texas, partially intersecting the maxima of CAPE in those areas. CINE, then, reflects a considerable barrier to the utilization of the large values of CAPE in the south-central United States, in good agreement with the relative minimum in July precipitation in that area. The small June to July change in CINE, however, shows that it is not an important factor in the onset, except possibly on the submonthly timescale for NW Mexico, if large values of CAPE develop before the CINE decreases.

The correspondence between CAPE and precipitation, then, is in accordance with CAPE acting as a necessary but not sufficient condition throughout the region, except for the monsoon development, which is strictly represented, with negligible values of CAPE in June and large values in July.

b. Large-scale vertical velocity

The July large-scale 500-hPa vertical velocity field (Fig. 4c) shows reasonably strong values of ascent over the regions of monsoonal and coastal precipitation and over the North American Plateau, and considerable values of descent off the west coast, over the central and eastern areas of the northern United States, and, notably, over the central United States and Texas. The June to July change (not shown) closely mirrors the precipitation changes (see Fig. 2). A correspondence between precipitation and vertical velocity is expected in areas of tropical convection, where precipitation is closely related to larger-scale ascent. On the other hand, the atmosphere can respond quite strongly to deep heating, with several areas of descending motion in adjoining regions (e.g., Rodwell and Hoskins 1996).

To investigate the relationship between the ascending (and, hence, either heated or orographically forced) and descending areas at the southern maximum of the descent over Texas, the vertical structure of the vertical motion field is shown across the domain at 27°N in Fig. 4d. The descent is seen to be concentrated at low levels to the east and west of the ascent over the cordillera and, at upper levels, to the east and west of the ascent associated with the monsoon heating (cf. Fig. 1b). This departure from the relatively homogenous descent that might be expected from radiative cooling, as well as the proximity to regions of strong heating and significant orography, is highly suggestive of a dynamic influence on vertical velocity. Given Rodwell and Hoskins’ (1996) investigation of the linear response to deep heating at 25°N, which showed forced descent to the northwest of the heating and little impact to the east, due to the small projection of the off-equatorial heating onto the first Kelvin mode, it is perhaps surprising that there is an area of significant elevated descent immediately to the east of the upper-level heating, as well.

5. Diabatic heating

A defining characteristic of vigorous monsoons is the evolution of the associated diabatic heating, which is dominated, in the mature phase, by the strong latent heating of the middle–upper troposphere resulting from the deep convective nature of monsoon precipitation. In the Mexican monsoon case, the deep heating of the troposphere during the monsoon’s mature phase was noted earlier in Fig. 1b, which showed the 500-hPa heating over northwest Mexico to be as large as 1 K day−1 in the ECMWF reanalysis derived heating. The dominant thermodynamic balance in the Tropics is between diabatic heating and adiabatic cooling (−N2ωQ/cp) (e.g., Holton 1992) so that, by virtue of the continuity equation, the horizontal divergence is intimately related to the vertical derivative of heating h·V = ∂/∂p(Q/cpN2). Thus, by impacting tropical divergence, the heating vertical structure can profoundly influence the vorticity dynamics, and the accompanying rotational circulation. Understanding the relationship between the vigorous convection of the North American monsoon system and the large-scale circulation, then, requires characterization of both the vertical structure of the associated heating and, since the domain encompasses both tropical and extratropical latitudes, the nature of the thermodynamic balance of the heating.

a. Heating vertical structure

The three different estimates of July heating, along a diagonal swath from northwest Mexico through the south-central United States and into the southeastern United States (the line AB shown in Fig. 1) are shown in Fig. 5: residually derived ECMWF heating in Fig. 5a, residually derived NCEP heating in Fig. 5b, and NCEP model-produced heating in Fig. 5c. All three fields exhibit similar large-scale characteristics: deep heating over the monsoon region, with one maximum around 400 hPa and another one around 150 hPa; considerable surface heating on the eastern flank of the Cordillera; modest values of cooling at low levels over the south-central United States; and deep heating over the southeastern United States. Although all the estimates yield significant values of deep heating over the monsoon region (1.5–2.5 K day−1 at 500 hPa), there are considerable differences between the estimates over large parts of the domain, even between the heating consistent with the NCEP reanalysis and the 6-h NCEP model forecast from that reanalysis. The primary differences between the ECMWF estimate and the NCEP estimates are (a) the larger NCEP values, particularly over the southeastern United States (consistent with the overprediction of the NCEP model precipitation in that region), and (b) the relatively stronger heating above 350 hPa in the NCEP estimates.

Focusing on the monsoon region, the vertical profile of heating in northwest Mexico (at 22°N, 105°W) from 850 hPa8 to 100 hPa is shown in Fig. 6a for the ECMWF residually derived heating (solid line with +), and in Fig. 6b for the NCEP residually derived heating (solid line with +) and the NCEP model produced heating (thicker solid line, no marks). As previously noted, the ECMWF heating shows the expected midlevel maximum at 500 hPa, but also has an upper-level maximum at 150 hPa for which a satisfactory explanation has yet to be advanced. The NCEP-derived heating has a much stronger upper-level maximum, at 5 K day−1 at 150 hPa, and a poorly defined midtropospheric maximum. The NCEP model heating is similar to the NCEP residually derived heating, but with a lower upper-level maximum, at 200 hPa, and somewhat better definition of the midlevel maximum. These profile differences result in heating differences between the model and the residually derived estimate of as much as 1 K day−1 at 700 hPa and 200 hPa [in agreement with Ebisuzaki’s (1995) finding of significant differences between the NCEP model heating and the residually derived heating for January 1993].

Examination of the partitioning of the NCEP model heating into radiative and convective components, shows that both the 400-hPa maximum and the 200-hPa maximum are attributed to the “deep convective heating” component. The NCEP reanalysis uses a model with a simplified Arakawa–Schubert cumulus convection scheme, while the ECMWF reanalysis uses a model with a mass flux convection scheme. The tendency of mass flux schemes to produce lower maxima of heating (Kiehl 1992) may partially explain the lower levels of maximum heating in the ECMWF reanalysis. It is not clear why the NCEP model produces such large values of heating at 200 hPa and, unfortunately, the ECMWF model heating is unavailable for comparison. Significant values of heating above 350 hPa have yet to be noted in high-resolution, limited-area budget studies of tropical heating. The height of maximum diabatic heating in the Tropics, as derived from numerous small-scale budget studies (summarized in, e.g., Johnson 1984), ranges from 500 hPa to 350 hPa, with only the east Atlantic GATE area showing a significantly different level of maximum heating, at 600 hPa. Although difficult to understand, the double-maximum vertical structure has been noted before in a large-scale diagnosis of tropical heating (Nigam 1994).

The heating vertical structure over the central United States (at 100°W and 42°N)—a U.S. great plains region with a local July precipitation maximum and an upper-level heating maximum in the ECMWF reanalysis (see Fig. 1)—is shown in Fig. 6c for the ECMWF residually derived estimate and in Fig. 6d for the NCEP estimates. Both residually derived estimates show considerable heating between 550 hPa and 200 hPa underlaid by considerable values of cooling between 850 hPa and 550 hPa.9 The NCEP model profile is similar, but with the heating occurring at higher levels (maximizing at 250 hPa), and a change from the lower-level cooling to heating at 850 hPa. Examination of the NCEP model’s deep convective and radiative contributions reveals that, at least in the model, this profile is the result of deep convective heating, but of smaller magnitude than the monsoon heating, so that the convective heating is overwhelmed by radiative cooling at lower levels.

More interesting, however, is the close correspondence between the residually derived heating profiles and the profile diagnosed in stratiform regions of tropical mesoscale convective cloud systems (e.g., Johnson and Young 1983). The deep stratiform clouds (nimbostratus) are associated with the mesoscale convective systems common to both the Tropics and summertime midlatitudes (Houze and Hobbs 1982), and produce substantial amounts of precipitation. Given the significance of mesoscale convective complexes in producing the warm-season precipitation in the central United States (Fritsch et al. 1986), the diagnosed profiles suggest that the stratiform precipitation and its associated heating—due to condensation and freezing in the upper troposphere and evaporation and melting in the lower troposphere (Johnson and Young 1983)—may be the dominant physical process contributing to heating in the northerly latitudes.

b. Thermodynamic balance

The thermodynamic balance operative in the two residually derived heatings is examined by displaying both the vertical velocities (ω, so negative values represent upward motion) and the adiabatic cooling (−N2ω) in the same panels as the vertical profiles of heating (Figs. 6a–d). The thermodynamic balance of the heating in the monsoon region is “tropical” in both estimates in the sense that there is close correspondence between the adiabatic cooling and the diabatic heating. Above 400 hPa, however, the variations in heating are balanced by variations in static stability rather than vertical velocity, as seen from a comparison of the omega and the adiabatic cooling profiles; the upper-level heating maximum, in particular, is balanced in both cases by strong adiabatic cooling resulting from increased static stability at the tropical tropopause levels.

Deep heating in the midlatitudes, on the other hand, is expected to be balanced by horizontal advection of temperature and accompanied by large-scale descent (Hoskins and Karoly 1981) and, indeed, in the north-central United States the heating is largely balanced, away from the surface, by horizontal temperature advection in both reanalysis (see Figs. 6c and 6d, the dotted line with ∘’s represents horizontal temperature advection). Both reanalyses also show descending motions in this region, but the descent is more deeply extended in the NCEP reanalysis.

The thermodynamic balance over the rest of the domain has also been investigated, and both the adiabatic cooling (Fig. 6e) and the horizontal advection of temperature (Fig. 6f) are shown at 500 hPa. South of 35°N, the adiabatic cooling largely balances the diabatic heating, whereas over the northern half of the United States, the horizontal advection term becomes dominant (cf. with the total diabatic heating field in Fig. 1b). It is interesting to note that the thermodynamic balance in the northern latitudes is achieved with the horizontal temperature advection offsetting not only diabatic heating but also the warming generated by the dynamically induced subsidence.

6. Seasonal evolution of tropospheric circulation

a. Climatological July circulation

The North American monsoon circulation during July, and the June to July circulation change associated with Mexican monsoon onset are examined at the upper- and lower-tropospheric levels to understand the dynamical impact of the dramatic precipitation increase over Mexico. As the large-scale flow, dominated by the rotational component, is well represented in most global analyses, only the ECMWF’s rendition is discussed here. The climatological (1979–93) ECMWF wind and geopotential height fields10 for July, and their June to July change at 200, 700, and 925 hPa are displayed in Fig. 7; 925 hPa is chosen as the lowest level as the Great Plains low-level jet is, perhaps, best captured at this ECMWF’s analysis output level (e.g., Berbery et al. 1996). Although the 925-hPa level is within the boundary layer and the 700-hPa level intersects the boundary layer over regions of western Mexico during the diurnal cycle, examination of the monthly mean values for each analysis time separately (0000, 0600, 1200, 1800 UTC) indicates that the diurnal cycle is not strongly represented in these fields and the discussion as given is relevant to all analysis times.

The July upper-tropospheric circulation (Fig. 7a) is dominated by the well-known North American anticyclone (∼40 m high) centered over northwest Mexico and a ridge extending northward over the central United States, both consistent with the northward retreat of westerlies over continental longitudes. The anticyclone is centered at about the same latitude (∼30°N) as the Tibetan anticyclone associated with the Asian summer monsoon, but is only about one-fourth as strong. The 200-hPa July circulation also contains a weak trough over the Caribbean.

The 700- and 925-hPa July circulations (Figs. 7c and 7e) show the considerable influence of the two subtropical anticyclones11 (located over adjoining oceans) on the flow over the North American monsoon region. The western edge of the Atlantic anticyclone (the Bermuda high) profoundly impacts the circulation (and climate) over Mexico and the central and eastern United States by bestowing a southerly component to the local flow, whereas the shallower Pacific anticyclone primarily influences the circulation over the western United States. The 925-hPa flow in particular contains a vigorous Great Plains jet, which is responsible for a substantial part of the northward moisture transport from the Gulf of Mexico into the central United States. The low-level flow over the Gulf of California, on the other hand, is weak and few wind analyses correctly represent even the flow direction over this and the adjoining Mexican monsoon region (Douglas et al. 1993; Stensrud et al. 1995; Schmitz and Mullen 1996)—for example, the ECMWF surface winds (not shown) do not agree well with Douglas et al.’s (1993) analysis in this area. In view of such uncertainties, a definitive characterization of the near-surface winds in this region must await the availability of higher density observations.

b. Onset phase circulation

The June to July change at 200 hPa (Fig. 7b) broadly reflects the northward retreat of westerlies over the continental longitudes (noted earlier), and their weakening over the eastern North Pacific. The geopotential height change consists of a ridge12 over northwest Mexico and the southwestern United States (also noted before), and a trough over the south-central United States, which was not evident in the July map (Fig. 7a). While the 200-hPa ridge over northwest Mexico can be ascribed, in part, to the vorticity forcing from the divergent outflow resulting from the considerable deep convective monsoon heating (as discussed in section 5), the processes leading to trough development over the south-central United States remain to be identified and understood.

An examination of the circulation change in the lower troposphere (Figs. 7d and 7f) is quite illuminative. The 700-hPa geopotential change contains rather interesting deviations from zonal symmetry in the northward advance of the easterlies, in particular, a trough over northwest Mexico and the southwestern United States and a ridge over eastern Mexico and also the south-central United States (also apparent in the 925-hPa circulation in Fig. 7f). While these features are very consistent with the June to July precipitation change (e.g., Fig. 2a), the reasons for the existence of the stationary wave itself are not clear, although there are interesting possibilities:one, which we intend to pursue in a subsequent modeling study, is to determine if this wave is in fact orographically forced by the encroachment of the seasonal easterlies onto the North American cordillera; the same orographic forcing could potentially lead to trough development at 200 hPa over the south-central United States (Fig. 7b)—a conjecture that we also propose to verify through modeling.

The June to July change at 925 hPa (Fig. 7f) reflects other interesting developments as well, such as that of the eastern Pacific ITCZ, the Bermuda high, and strengthening of the features associated with the Great Plains jet in the lee of the North American Plateau. Also, in accord with the Bermuda high development, a channel of easterly flow develops along its southern flank, extending from the Atlantic and curving south around the cordillera into northwest Mexico where it is joined by the ITCZ-related flow.

7. Seasonal evolution of moisture fluxes

The July mean (1979–93) vertically integrated moisture flux convergence and the divergent part of the vector flux is presented in Fig. 8a for the NCEP reanalysis and in Fig. 8b for the ECMWF reanalysis. Only the divergent component of the moisture flux is shown, neglecting the large rotational component, which does not contribute to changes in moisture, in order to focus on the transport of moisture directly associated with the moisture flux convergence and divergence. Despite the known problems in resolving the small-scale structure of the PBL in some areas and the significant orography, the moisture flux convergences are quite plausible, with significant convergence in the ITCZ, Central American, and northwest Mexico regions in both reanlayses. The very large values of divergence along the eastern flank of the cordillera present in both the ECMWF and the NCEP operational analyses and attributed to errors resulting from the difficulties in resolving the local orography (Rasmusson and Mo 1996; Schmitz and Mullen 1996) are greatly reduced in the reanalyses. Comparison with the July precipitation fields (Figs. 1c and 1d) shows that over some areas, notably northwest Mexico, the vertically integrated moisture flux convergence is even quantitatively consistent with the precipitation.

The June to July change in vertically integrated moisture flux convergence (NCEP in Fig. 8c, ECMWF in Fig. 8d) shows an even closer agreement with precipitation (Figs. 2c and 2d), at least in the ECMWF reanalysis, capturing all the major precipitation changes, some even quantitatively. The NCEP reanalysis shows smaller, less coherent changes, possibly as a result of a greater role for evaporation in the NCEP reanalysis precipitation. Some caution must be used when interpreting the June to July change in transport, as the evolution of more than one feature may be overlaid, as in the Pacific around 10°–20°N, where the local weakening of the subtropical anticyclone and the northward movement of the ITCZ are both occurring.

Perhaps the most striking feature of the divergent moisture flux fields (Figs. 8a and 8b) is the large-scale moisture transport into northwest Mexico and Central America from both the east and the west, unbroken by the significant orography of the Cordillera. The westward flux over the mountains of Mexico for the summer season is implicitly suggested in analyses of FGGE data (Chen 1985) and the NCEP operational data (Rasmusson and Mo 1996) and has been explicitly noted in the ECMWF operational analysis (Schmitz and Mullen 1996). The westward divergent flux is clearly evident in both the ECMWF and the NCEP reanalysis. The magnitude of the change in the divergent flux over the Cordillera in both reanalyses (Figs. 8c and 8d) reveals that this westward transport across the mountains begins with the monsoon onset period, and is a considerable change from the weak eastward transport of the premonsoon period. Also interesting is the association of the areas of moisture flux divergence in the central United States, Texas, and northeastern Mexico with transport toward northwest Mexico and the eastern Pacific ITCZ in the onset change field. In the southern extent of these divergence areas, the onset changes result in actual transport toward northwest Mexico in July.

Although broadly in agreement, the two reanlayses differ in the small-scale structure of the divergent moisture flux and its onset change: while over half of the July convergence over northwest Mexico is a result of the June to July onset in the ECMWF reanalysis, the large convergence over northwest Mexico in the NCEP reanalysis is also present in June, and the primary change during the onset in the NCEP data is confined over the Gulf of California. The large-scale source of moisture for NW Mexico in July (again, excepting local evaporation, which may be quite large over the Gulf of California) also varies between the two reanalyses, with the NCEP reanalysis favoring southerly transport into the monsoon area (Fig. 8a), in agreement with the analysis of Reyes and Cadet (1986, 1988), while the ECMWF reanalysis favors transport from the east (Fig. 8b). Examination of just the low-level (surface to 850 hPa) moisture flux (not shown) indicates westward moisture flux across Central America, northwestward along the southwestern coast of Mexico, and into the monsoon region in both reanalyses during July, in agreement with Schmitz and Mullen’s (1996) finding based on the ECMWF operational moisture flux analysis for July and August. Additionally, analysis of the June to July change reveals that this pathway exists in June up to about 18°N, with the final extension into the monsoon region occurring during the change from June to July.

8. Seasonal evolution of SST

In view of the fundamental importance of differential sensible heating (e.g., due to land–sea thermal contrasts) in setting up sea level pressure gradients, especially during the premonsoon-onset phase, we examine the seasonal evolution of SSTs in the basins adjoining Mexico. The SST gradients themselves can force low-level winds, particularly over the eastern tropical oceans, which are capped by trade inversions (Lindzen and Nigam 1987), but the suitability of this mechanism in the basins adjoining Mexico remains to be examined.

The climatological July SSTs (Fig. 9a) show the warmest water (>29°C) to be present along the western coast of Mexico and in the Gulf of California, and in the Gulf of Mexico. Comparison of the July SSTs with the corresponding 925-hPa height and wind fields (Fig. 7e) in the tropical–subtropical (θ ⩽30°) North Pacific indicates that the Lindzen–Nigam mechanism may in fact be operative, as the height contours approximately follow the isotherms there, with the 925-hPa heights (and, therefore, surface pressure) increasing toward the colder isotherms. The striking difference from the tropical situation, of course, is that the subtropical winds are nearly geostrophic, blowing along the isobars/isotherms, rather than across them as in the deep Tropics.

The June to July change in SSTs (Fig. 9b) shows the dramatic warming of the Gulf of California (up to 3°C) and the adjoining Pacific Ocean. As before, the SST change is compared to the corresponding 925-hPa height and wind changes (Fig. 7f). For the difference fields, however, the Lindzen–Nigam mechanism is not helpful in understanding the subtropical (15° ⩽ θ ⩽ 30°) circulation, perhaps because the underlying SST change is largest along the Mexican coast (Fig. 9b), where the alongshore wind component determines SSTs through Ekman upwelling dynamics, rather than the SSTs determining the low-level winds through the well-mixed atmospheric boundary layer dynamics (as in the Lindzen–Nigam mechanism).

9. Discussion and concluding remarks

The climatological evolution of the North American monsoon system has been documented from a dynamically oriented survey of several independent datasets, yielding, through their agreement, an atlas of potentially important features and, through their disagreement, an estimate of uncertainty.

The Mexican monsoon precipitation has been shown to be associated with large values of deep heating and a tropical-type thermodynamic balance between diabatic heating and adiabatic cooling. The structure of adiabatic cooling, however, is controlled by variations in the static stability in the upper-half of the troposphere, rather than by variations of vertical velocity, suggesting a more complicated relationship than just a simple proportionality between diabatic heating and rising motion. The noted out-of-phase climatological relationship between the Mexican monsoon precipitation and the surrounding central United States precipitation during the June to July transition has been shown to be robust through several different estimates of precipitation and through several time periods and, additionally, on the pentad timescale, as well. This linkage in precipitation has been shown to be associated with significant changes in the large-scale tropospheric circulation.

Changes in the thermodynamic environment, in terms of the development of an area of high CAPE, seem to capture the strength of the precipitation increase over northwest Mexico during monsoon onset quite well, but the decrease in precipitation over the central United States is not associated with any significant reduction in the local CAPE. Interestingly, the CAPE distribution is also not very helpful in understanding why precipitation favors the western coast of Mexico over the eastern coast during the monsoon season. The large values of CAPE over coastal Mexico in July are associated with both high surface temperature and high values of specific humidity, but as surface temperature is already high in June, the June to July evolution of low-level moisture likely controls the changes in CAPE in this region.

Given the correspondence with precipitation, the mutual agreement between the reanalyses, and the reduction of error in moisture transport calculations with reanalyses data, the fidelity of the reanalyses’ divergent moisture transports is sufficient to conclude that there is a considerable linkage between the various regions during the Mexican monsoon onset, despite the intervening orography. The June to July change in moisture transports associates the moisture flux divergence areas of the central United States, Texas, and northeastern Mexico with transports toward the precipitation onset regions of northwest Mexico and the eastern Pacific ITCZ. However, the need for better (higher density) observational data is highlighted by the fact that both reanalyses show considerable differences in the large-scale moisture transport into northwest Mexico. Additionally, these monthly mean results provide motivation for studies on much shorter timescales, in order to investigate the daily transports that make up the intriguing monthly mean pattern, particularly, to understand the linkage between the Mexican monsoon and the central United States precipitation.

Examination of the vertical velocity field suggests that the monsoon onset may play a role in reducing precipitation to the east by forcing descent over the adjacent areas. The dynamic inhibition of convection over the south-central United States, however, appears to be strongly affected by the elevated heating of the North American cordillera in Mexico and the North American Plateau. Additionally, the June to July change in 700-hPa geopotential height and winds suggests that the seasonal impingement of the easterlies upon the cordillera may generate an orographically forced wave that may significantly impact the large-scale circulation and precipitation fields.

In the course of this observational analysis, the newly available NCEP and ECMWF reanalysis have been intercompared over the North American monsoon system area. While the rotational component of the flows is, not surprisingly, in good agreement, there is significant disagreement between the divergent flows of the two reanalyses, particularly when assessed by the vertical profile of diagnosed heating. Furthermore, there is disagreement between the heating diagnosed from the NCEP reanalysis and the available NCEP model output heating. The tropical-type deep tropospheric heating associated with the monsoon in all the estimates suggests a prominent role for the monsoon in the regional circulation, but the considerable differences in the vertical profiles suggest that the implied dynamics13 of the North American monsoon system may be different in the ECMWF and NCEP reanalysis, and in the NCEP model.

While a credible observational analysis of the evolution of moisture flux into the Mexican Monsoon area during the onset period must await higher quality data, a linear diagnostic modeling study is planned to investigate the dynamical issues raised in this study: the impact of the monsoon diabatic heating on the flow throughout the region, both at low and midlevels; the impact of the elevated heating of the North American cordillera and plateau on the regional flow; the impact of the interaction of the cordillera and the easterlies during their northward migration; and the impact of the premonsoon heating in the ITCZ and Central America on the preconditioning of the monsoon environment.

Acknowledgments

We would like to thank Eugene Rasmusson for motivating us to study the North American monsoon system, and for his continued encouragement and support. The dissertation of Ian Okabe provided a very useful starting point for this research. We would like to thank Dr. Maddox and one anonymous reviewer for comments that considerably improved the paper. We would also like to thank Wayne Higgins of CPC for providing the gridded HPD data, Bob Hart for providing the algorithmic basis for the CAPE computations, and Ilana Stern of UCAR for aiding in the ECMWF and NCEP data retrieval. This work is part of the doctoral research being conducted by Mathew Barlow.

This research effort was supported by the Cooperative Institute of Climate Studies at the University of Maryland at College Park. Additional support was provided by NA76GP0479, and by NOAA Grant NA46GP0194 and NSF Grant ATM931-6278 to S. Nigam, and NOAA Grant NA57WCO340 to Ernesto Berbery.

REFERENCES

  • Arkin, P. A., and B. N. Meisner, 1987: The relationship between large-scale convective rainfall and cold cloud over the Western Hemisphere during 1982–84. Mon. Wea. Rev.,115, 51–74.

  • ——, and P. E. Ardanuy, 1989: Estimating climatic-scale precipitation from space: A review. J. Climate,2, 1229–1238.

  • Augustine, J. A., and F. Caracena, 1994: Lower-tropospheric precursors to nocturnal MCC development over the central United States. Wea. Forecasting,9, 116–135.

  • Badan-Dangon, A., C. E. Dorman, M. A. Merrifield, and C. D. Winant, 1991: The lower atmosphere over the Gulf of California. J. Geophys. Res.,96 (C9), 877–896.

  • Bell, G. D., and J. E. Janowiak, 1995: Atmospheric circulation associated with the Midwest floods of 1993. Bull. Amer. Meteor. Soc.,76, 681–695.

  • Berbery, E. H., E. M. Rasmusson, and K. E. Mitchell, 1996: Studies of North American continental-scale hydrology using Eta model forecast products. J. Geophys. Res.,101 (D3), 7305–7319.

  • Bluestein, H. B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes. Vol. 2, Observations and Theory of Weather Systems, Oxford University Press, 594 pp.

  • Broccoli, A. J., and S. Manabe, 1992: The effects of orography on midlatitude Northern Hemisphere dry climates. J. Climate,5, 1181–1201.

  • Bryson, R. A., and F. K. Hare, 1974: The climates of North America. World Survey of Climatology, Vol. 2, Climates of North America, R. A. Bryson and F. K. Hare, Eds., Elsevier, 1–36.

  • Chen, T.-C., 1985: Global water vapor flux and maintenance during FGGE. Mon. Wea. Rev.,113, 1801–1819.

  • Douglas, A. V., and P. J. Englehart, 1995: An analysis of the starting date for the summer monsoon in western Mexico and southeast Arizona. Proc. 20th Annual Climate Diagnostics Workshop, Seattle, WA, Climate Prediction Center, 207–211.

  • ——, and ——, 1996: Variability of the summer monsoon in Mexico and relationships with drought in the United States. Proc. 21st Annual Climate Diagnostics Workshop, Huntsville, AL, Climate Prediction Center, 296–299.

  • Douglas, M. W., R. A. Maddox, K. W. Howard, and S. Reyes, 1993:The Mexican monsoon. J. Climate,6, 1665–1677.

  • Ebisuzaki, W., 1995: Consistency of the diabatic heating in the NMC/NCAR reanalysis. Proc. 20th Annual Climate Diagnostics Workshop, Seattle, WA, Climate Prediction Center, 232–234.

  • Fritsch, J. M., R. J. Kane, and C. R. Chelius, 1986: The contribution of mesoscale convective weather systems to the warm-season precipitation in the United States. J. Climate Appl. Meteor.,25, 1333–1345.

  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc.,106, 447–462.

  • Hahn, D. G., and S. Manabe, 1975: The role of mountains in the south Asian monsoon circulation. J. Atmos. Sci.,32, 1515–1540.

  • Heideman, K. F., and J. M. Fritsch, 1988: Forcing mechanisms and other characteristics of significant summertime precipitation. Wea. Forecasting,3, 115–130.

  • Hering, W. S., and T. R. Borden Jr., 1962: Diurnal variations in the summer wind field over the central United States. J. Atmos. Sci.,19, 81–86.

  • Higgins, R. W., Y. Yao, E. S. Yarosh, J. E. Janowiak, and K. C. Mo, 1997: Influence of the Great Plains low-level jet on summertime precipitation and moisture transport over the central United States. J. Climate,10, 481–507.

  • Holton, J. R., 1992: An Introduction to Dynamic Meteorology. Academic Press, 510 pp.

  • Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci.,38, 1179–1196.

  • ——, H. H. Hsu, I. N. James, M. Masutani, P. D. Sardeshmukh, and G. H. White, 1989: Diagnostics of the global atmospheric circulation based on ECMWF analyses 1979–1989. WMO/TD-No. 326, World Meteorological Organization, 217 pp.

  • Houze, R. A., Jr., and P. V. Hobbs, 1982: Organization and structure of precipitation cloud systems. Advances in Geophysics, Vol. 24, Academic Press, 225–315.

  • Janowiak, J. E., and P. A. Arkin, 1991: Rainfall variations in the Tropics during 1986–1989, as estimated from observations of cloud-top temperature. J. Geophys. Res.,96 (Suppl.), 3359–3373.

  • Johnson, R. H., 1984: Partitioning tropical heat and moisture budgets into cumulus and mesoscale components: Implications for cumulus parameterization. Mon. Wea. Rev.,112, 1590–1601.

  • ——, 1986: Lower-tropospheric warming and drying in tropical mesoscale convective systems: Implications for the problem of cumulus parameterization. J. Meteor. Soc. Japan,64, 721–726.

  • ——, and G. S. Young, 1983: Heat and moisture budgets of tropical mesoscale anvil clouds. J. Atmos. Sci.,40, 2138–2147.

  • Kalnay E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc.,77, 437–471.

  • Kiehl, J. T., 1992: Atmospheric general circulation modeling. Climate System Modeling, K. E. Trenberth, Ed., Cambridge University Press, 319–368.

  • Krishnamurti, T. N., 1971: Tropical east–west circulations during the northern summer. J. Atmos. Sci.,28, 1342–1347.

  • Lindzen, R. S., and S. Nigam, 1987: On the role of sea-surface temperature gradients in forcing low-level winds and convergence in the Tropics. J. Atmos. Sci.,44, 2418–2436.

  • Maddox, R. A., 1980: Mesoscale convective complexes. Bull. Amer. Meteor. Soc.,61, 1374–1387.

  • Means, L. L., 1952: On thunderstorm forecasting in the central United States. Mon. Wea. Rev.,80, 165–189.

  • Mo, K. C., J. Nogues-Paegle, and J. Paegle, 1995: Physical mechanisms of the 1993 summer floods. J. Atmos. Sci.,52, 879–895.

  • Mock, C. J., 1996: Climatic controls and spatial variations of precipitation in the western United States. J. Climate,9, 1111–1125.

  • Mosiño, P. A., and E. García, 1974: The climate of Mexico. World Survey of Climatology, Vol. 2, Climates of North America, R. A. Bryson and F. K. Hare, Eds., Elsevier, 373–404.

  • Negri, A. J., R. F. Adler, R. A. Maddox, K. W. Howard, and P. R. Keehn, 1993: A regional rainfall climatology over Mexico and the southwest United States derived from passive microwave and geosynchronous data. J. Climate,6, 2144–2161.

  • Nigam, S., 1994: On the dynamical basis for the Asian summer monsoon rainfall–El Niño relationship. J. Climate,7, 1750–1771.

  • ——, 1997: The annual warm to cold phase transition in the eastern equatorial Pacific: Diagnosis of the role of stratus cloud top cooling. J. Climate,10, 2447–2467.

  • Okabe, I. T., 1994: The North American monsoon. Ph.D. dissertation, University of British Columbia, 146 pp. [Available from Dept. of Geography, University of British Columbia, 2075 Wesbrook Place, Vancouver, BC V6T 1W5, Canada.].

  • Pan, H.-L., and W.-S. Wu, 1994: Implementing a mass flux convective parameterization package for the NMC medium-range forecast model. Preprints, 10th Conf. on Numerical Weather Prediction, Portland, OR, Amer. Meteor. Soc., 96–98.

  • Pitchford, K. L., and J. London, 1962: The low-level jet as related to nocturnal thunderstorms over the Midwest United States. J. Appl. Meteor.,1, 43–37.

  • Rasmusson, E. M., and K. C. Mo, 1996: Large-scale atmospheric moisture cycling as evaluated from global NMC analysis and forecast products. J. Climate,9, 3276–3297.

  • Reyes, S., and D. L. Cadet, 1986: Atmospheric water vapor and surface flow patterns over the tropical Americas during May–August 1979. Mon. Wea. Rev.,114, 582–593.

  • ——, and ——, 1988: The southwest branch of the North American monsoon during summer 1979. Mon. Wea. Rev.,116, 1175–1187.

  • Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analyses using optimum interpolation. J. Climate,7, 929–948.

  • Rodwell, M. J., and B. J. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc.,122, 1385–1404.

  • Rowson, D. R., and S. J. Colucci, 1992: Synoptic climatology of thermal low-pressure systems over south-western North America. Int. J. Climatol.,12, 529–545.

  • Schmitz, J. T., and S. L. Mullen, 1996: Water vapor transport associated with the summertime North American monsoon as depicted by ECMWF analyses. J. Climate,9, 1621–1634.

  • Song, J.-L., and W. M. Frank, 1983: Relationships between deep convection and large-scale processes during GATE. Mon. Wea. Rev.,111, 2145–2160.

  • Stensrud, D. J., 1996: Importance of low-level jets to climate: A review. J. Climate,9, 1698–1711.

  • ——, R. L. Gall, S. L. Mullen, and K. W. Howard, 1995: Model climatology of the Mexican monsoon. J. Climate,8, 1775–1794.

  • Tang, M., and E. R. Reiter, 1984: Plateau monsoons of the Northern Hemisphere: A comparison between North America and Tibet. Mon. Wea. Rev.,112, 617–637.

  • Tiedtke, M., 1989: A comprehensive mass flux scheme for cumulus parameterization in large-scale models. Mon. Wea. Rev.,117, 1779–1800.

  • Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev.,103, 406–419.

  • Williams, E., and N. Renno, 1993: Analysis of the conditional instability of the tropical atmosphere. Mon. Wea. Rev.,121, 21–36.

  • Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate,9, 840–858.

  • Xu, K.-M., and K. A. Emanuel, 1989: Is the tropical atmosphere conditionally unstable? Mon. Wea. Rev.,117, 1471–1479.

  • Yanai, M., S. Esbensen, and J. H. Chu, 1976: Response of deep and shallow tropical maritime cumuli to large-scale processes. J. Atmos. Sci.,33, 976–991.

APPENDIX

Computation of CAPE

Even ignoring issues concerning the accuracy of parcel theory (Williams and Renno 1993), two primary assumptions must be made in the calculation of CAPE:an assumption of the level of origin of the parcel, and an assumption regarding the process the parcel undergoes during ascent.

A standard assumption regarding the ascent of the parcel (e.g., Bluestein 1993) is that the parcel undergoes pseudoadiabatic ascent once the parcel is saturated. While it has been argued that condensate loading must be accounted for (e.g., Xu and Emanuel 1989), Williams and Renno’s (1993) work suggests that the more accurate assumption, which includes both condensate loading and the latent heat of fusion in the ice phase, is closer to the original pseudoadiabatic assumption. Given the uncertainties (e.g., choice of the freezing level) associated with the more complicated assumptions and the large cancellation between their effects, the pseudoadiabatic assumption is used here for simplicity.

The level of origin for the parcel is assumed here to be the surface (used in diverse studies; e.g., Song and Frank 1983; Williams and Renno 1993), a choice that provides a consistent assumption throughout the domain without having to rely on the small-scale structure of the boundary layer, which is unreliably resolved in the numerous regions of significant topography. Varying the level of origin within the near-surface layer did not qualitatively change the results.

CAPE is defined herein as (e.g., Williams and Renno 1993)
i1520-0442-11-9-2238-e2
where LNB denotes the level of neutral buoyancy; LFC denotes the level of free convection, the level where the parcel first achieves positive buoyancy; PCL denotes the parcel variable; ENV denotes the environmental variable; and Tv denotes virtual temperature. CAPE may be interpreted as the work done by the buoyancy force acting on the parcel between the level of free convection and the level of neutral buoyancy. CINE is defined as the same integral, only evaluated from the surface to the level of free convection, and may be interpreted as the work necessary to raise the parcel to the level of free convection. In all calculations, the variables are first interpolated onto 10-hPa increments before computation of the integral.

CAPE, as analyzed in this study, is calculated from monthly mean climatological data. Calculations from 6-h data for a test month show that, in the given domain, CAPE computed from a monthly mean is quite similar to the monthly mean of CAPE calculated every 6 h. Additionally, 6-h CAPE was partitioned into two means, based on whether precipitation was occuring in each grid box or not, at each time. The rain and the no-rain means were similar, suggesting both that the monthly mean CAPE is, indeed, representative of general conditions, and that the CAPE is acting, as expected, as a necessary but not sufficient condition.

Fig. 1.
Fig. 1.

(a) Mean (1979–85) July GHCN station precipitation gridded on a 2.5° lat × 2.5° long grid. The contour interval is 1 mm day−1 and values greater than 2 mm day−1 are shaded. (b) Mean (1979–85) July residually derived ECMWF diabatic heating at 500 hPa. The contour interval and shading threshold is 0.5 K day−1, and the zero contour is omitted. (c) As in (a) but for the NCEP reanalysis model precipitation. (d) As in (a) but for the ECMWF reanalysis model precipitation.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 2.
Fig. 2.

Mean June to July (onset phase) changes in (a) gridded GHCN precipitation, (b) GPI precipitation, (c) the NCEP reanalysis model precipitation, and (d) the ECMWF reanalysis model precipitation. Climatological periods are 1950–85 for the GHCN data, 1986–96 for the GPI data, and 1979–93 for both the ECMWF and NCEP reanalyses. The contour interval and shading threshold is 1 mm day−1, and the zero contour is omitted in all panels. The GPI estimate is not produced for latitudes north of 40°N, and the GHCN data, as station data, is not available over the oceans.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 3.
Fig. 3.

(a) The mean (1986–93) pentad evolution of precipitation averaged over the Mexican monsoon region (outlined in Fig. 1a) is shown along with the standard deviation. (b) The mean evolution of precipitation averaged over the central United States (also outlined in Fig. 1a) is shown using filled circles, while that over the Mexican monsoon region is shown with open circles, with a +1/−1 pentad running mean applied to both.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 4.
Fig. 4.

(a) Mean July CAPE, calculated from the ECMWF monthly mean reanalysis, 1979–93, and shown using a contour interval of 200 J kg−1, with values greater than 1000 J kg−1 shaded. (b) Mean June to July change in the ECMWF CAPE, shown using a contour interval and shading threshold of 200 J kg−1. (c) Mean July 500-hPa pressure vertical velocity (omega), calculated from ECMWF monthly means, 1979–93. The contour interval and shading threshold is 0.005 Pa s−1, note that negative values denote upward motion. (d) Mean July omega at 27°N, from the ECMWF reanalysis, 1979–93. The contour interval and shading threshold is 0.005 Pa s−1. Blank areas represent orography. The zero contour is omitted in all panels.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 5.
Fig. 5.

(a) The mean (1979–93) vertical structure of heating along a diagonal swath (line AB in Fig. 1a) whose west side is representative of the monsoon region; middle part, of the south-central United States; and east side, of the southeastern United States and northern Florida: (a) for the residually derived heating from the ECMWF reanalysis; (b) for the residually derived heating from the NCEP reanalysis; and (c) for the NCEP model produced diabatic heating. The contour interval and shading threshold is 0.5 K day−1, and the zero contour is omitted.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 6.
Fig. 6.

The vertical profile for mean (1979–93) July conditions at 22°N, 105°W of residually derived heating (solid line with + marks), the adiabatic cooling (−N2ω) (dotted line with o marks), the NCEP model produced heating (solid line with no marks), and omega (dotted line with + marks) for (a) ECMWF reanalysis and (b) NCEP reanalysis. (c) and (d) As in (a) and (b) but for a midlatitude location (42°N, 100°W) with horizontal temperature advection (V·T) shown instead of the adiabatic cooling term (dotted line with o marks). The x axis markings to the left of the origin are at 0.02 Pa s−1 intervals. The markings to the right of the origin are at 1.0 K day−1. (e) The July mean (1979–93) adiabatic cooling (−N2ω) at 500 hPa from the ECMWF reanalysis. (f) The July mean horizontal temperature advection (V·T) at 500 hPa from the same dataset. The contour interval and shading threshold is 0.5 K day−1, and the zero contour is omitted in these panels.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 7.
Fig. 7.

The mean July and the June to July change in the ECMWF wind and geopotential height at (a), (b) 200 hPa, (c), (d) 700 hPa, and (e), (f) 925 hPa. The blank area in the 925-hPa panels represents the intersection of orography with that pressure level. The height interval and the vector wind scale (in m s−1) are indicated for each panel.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 8.
Fig. 8.

The mean July and the June to July change in the vertically integrated moisture flux convergence from (a), (d) ECMWF reanalysis (1979–93) and (b), (c) NCEP reanalysis (1979–93). The vectors represent the divergent part of the vertically integrated flux. The contour interval and shading threshold is 2 mm day−1, and the zero contour is omitted in all panels. The vertical integral was computed from pressure level data with the surface pressure as the lower boundary. Note that the same vector length denotes a divergent moisture flux of 60 kg m−1 s−1 in the July fields, and only 40 kg m−1 s−1 in the difference fields.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

Fig. 9.
Fig. 9.

(a) Mean (1982–95) July SST and (b) the mean June to July change. In (a) the contour interval is 1°C, with values over 28°C shaded. In (b) the contour interval is 0.5°C, with magnitudes greater than 1.5°C shaded. The OISST dataset was used for SSTs.

Citation: Journal of Climate 11, 9; 10.1175/1520-0442(1998)011<2238:EOTNAM>2.0.CO;2

1

The expected profile of deep convective heating is discussed in section 5.

2

The heating diagnosis method is discussed in section 2d, and the two reanalyses’ heating fields are compared in section 5.

3

First GARP (Global Atmospheric Research Program) Global Experiment, conducted in 1979.

4

This evaluation was performed using a steady global primitive equation model having high horizontal and vertical resolution: Δθ = 2.5°, zonal Fourier truncation at wavenumber 15 or 30, and 18 vertical sigma levels.

5

If the contour interval were decreased to 0.5 mm day−1, not shown because of the large changes in other regions, the decreases would appear as a continuous belt to the north and east of the monsoon region in both the GHCN and the ECMWF model precipitation, although for other time periods there is a modest break between the decreases to the north and the decreases to the east at around 40°N.

6

The 2 mm day−1 threshold corresponds to Douglas et al.’s (1993) monsoon index values of 60 or greater, so that at least 60% of the annual precipitation was from the July–September contributions.

7

Although variations in the temperature and moisture fields aloft can considerably impact CAPE, the greater magnitude of variations in the near-surface layer results in CAPE being frequently controlled by the near-surface environment [cf. Williams and Renno’s (1993) success in diagnosing CAPE throughout the Tropics with surface θw].

8

The surface is actually below 850 hPa; however, to avoid ambiguity due to the interpolation of data from the native sigma surfaces to the standard pressure levels and due to the differences in surface pressures of the various estimates, all estimates are shown starting at 850 hPa.

9

The low-level heating underneath the lower-tropospheric cooling region has been attributed to unsaturated descent in the mesoscale downdraft in the stratiform region (Johnson 1986), but the large-scale vertical velocity, representing the sum of convective and stratiform contributions in Figs. 6c and 6d, indicates ascent in that region.

10

The geopotential height and wind field displays are chosen as a compromise between the traditional height description in midlatitudes and the streamfunction analysis in tropical regions.

11

Although the subtropical high pressure belt is more continuous during northern winter due to higher surface pressure over land as well, the oceanic anticyclones are strongest during the northern summer for reasons discussed in Hoskins (1996).

12

As the June to July change consists of higher heights to the north—that is, opposite of the variations in climatology (e.g., Fig. 7a), troughs and ridges in Figs. 7b, 7d, and 7f are associated with inverted height structures.

13

Even a moderate vertical shift in the heating and, thus, associated winds could result in a very large difference in, for example, the character of moisture transport throughout the region, a critical factor in the evolution and maintenance of convective precipitation.

Save