• Collins, W. D., and Coauthors, 2006: The Community Climate System Model, version 3 (CCSM3). J. Climate, 19, 21222143.

  • Cook, K. H., 1999: Generation of the African easterly jet and its role in determining West African precipitation. J. Climate, 12, 11651184.

    • Search Google Scholar
    • Export Citation
  • Gent, P. R., and Coauthors, 2011: The Community Climate System Model, version 4. J. Climate, 24, 49734991.

  • Grodsky, S. A., J. A. Carton, and S. Nigam, 2003: Near surface westerly wind jet in the Atlantic ITCZ. Geophys. Res. Lett., 30, 2009, doi:10.1029/2003GL017867.

    • Search Google Scholar
    • Export Citation
  • Gutzler, D. S., and Coauthors, 2009: Simulations of the North American monsoon: NAMAP2. J. Climate, 22, 67166740.

  • Hagos, S. M., and K. H. Cook, 2007: Dynamics of the West African monsoon jump. J. Climate, 20, 52645284.

  • Huffman, G. J., and Coauthors, 2007: The TRMM multi-satellite precipitation analysis: Quasi-global, multi-year, combined-sensor precipitation estimates at fine scale. J. Hydrometeor., 8, 3855.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., and J. M. Arblaster, 1998: The Asian–Australian monsoon and El Niño–Southern Oscillation in the NCAR climate system model. J. Climate, 11, 13561385.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., J. M. Arblaster, D. M. Lawrence, A. Seth, E. K. Schneider, B. P. Kirtman, and D. Min, 2006: Monsoon regimes in the CCSM3. J. Climate, 19, 24822495.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., J. M. Arblaster, J. Caron, H. Annamalai, M. Jochum, A. Chakraborty, and R. Murtugudde, 2012: Monsoon regimes and processes in CCSM4. Part I: The Asian–Australian monsoon. J. Climate, 25, 25832608.

    • Search Google Scholar
    • Export Citation
  • Pu, B., and K. H. Cook, 2010: Dynamics of the West African westerly jet. J. Climate, 23, 62636276.

  • Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Riddle, E. E., and K. H. Cook, 2008: Abrupt rainfall transitions over the Greater Horn of Africa: Observations and regional model simulations. J. Geophys. Res., 113, D15109, doi:10.1029/2007JD009202.

    • Search Google Scholar
    • Export Citation
  • Rowell, D. P., and J. R. Milford, 1993: On the generation of African squall lines. J. Climate, 6, 11811193.

  • Simmons, A., S. Uppla, D. De, and S. Kobayashi, 2006: ERA-Interim: New ECMWF reanalysis products from 1989 onwards. ECMWF Newsletter, No. 110, ECMWF, Reading, United Kingdom, 25–35.

    • Search Google Scholar
    • Export Citation
  • Sultan, B., and S. Janicot, 2003: The West African monsoon dynamics. Part II: The preonset and onset of the summer monsoon. J. Climate, 16, 34073427.

    • Search Google Scholar
    • Export Citation
  • Vera, C., and Coauthors, 2006: Toward a unified view of the American monsoon systems. J. Climate, 19, 49775000.

  • Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate, 9, 840858.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Hovmöller diagrams of daily rainfall (mm day−1) averaged from 10°W to 10°E and smoothed using a 5-day running mean from the (a) TRMM, (b) CCSM4, and (c) CAM4 climatologies.

  • View in gallery

    (a) MJ climatological precipitation (contours; mm day−1) from TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis. (b) Precipitation contours (mm day−1) and 930-hPa wind vectors from the CCSM4 climatology for MJ, and (c) precipitation contours (mm day−1) and 930-hPa wind vectors from the CAM4 climatology for MJ.

  • View in gallery

    The (a) 1998–2009 JAS climatological precipitation (contours, mm day−1) from TRMM with 900-hPa wind vectors from the ECMWF-Interim reanalysis. (b) Precipitation contours (mm day−1) and 930-hPa wind vectors from the 1986–2005 CCSM4 climatology for JAS, and (c) precipitation contours (mm day−1) and 930-hPa wind vectors from the 1986–2005 CAM4 climatology for JAS.

  • View in gallery

    Precipitation climatology for 1981–2000 (mm day−1) averaged over eight ensemble members from CCSM3 for (a) MJ and (b) JAS.

  • View in gallery

    Surface temperature (K) for MJ from (a) the ERA-Interim reanalysis and (b) CCSM4; and for JAS from (c) the ERA-Interim reanalysis and (d) CCSM4. Shading interval is every 2 K while contour interval is every 1 K.

  • View in gallery

    The 600-hPa zonal wind speed (m s−1) for JAS from the (a) ERA-Interim reanalysis, (b) CCSM4 simulation, and (c) CAM4 simulation. The 200-hPa zonal wind speed for JAS from the (d) ERA-Interim reanalysis, (e) CCSM4 simulation, and (f) CAM4 simulation. Negative values are shaded.

  • View in gallery

    As in Fig. 1, but for 30°–45°E.

  • View in gallery

    DJF climatological precipitation (contours; mm day−1) from (a) TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis, (b) CCSM4, and (c) CAM4. JAS climatological precipitation (contours; mm day−1) from (d) TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis, (e) CCSM4, and (f) CAM4.

  • View in gallery

    Intraseasonal variation of rainfall (mm day−1) averaged 5°S–5°N and 35°–45°E (black line) and 10°–20°N and 35°–45°E (gray line) in the (a) TRMM, (b) CCSM4, and (c) CAM4 climatologies.

  • View in gallery

    JJA climatological precipitation (shading; mm day−1) from (a) TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis, (b) CCSM4 with 930-hPa wind vectors, and (c) CAM4 with 930-hPa wind vectors.

  • View in gallery

    Daily rainfall (mm day−1) averaged 103°–112°W and 20°–37°N for the TRMM, CCSM4, and CAM4 climatologies.

  • View in gallery

    Daily precipitation (mm day−1) averaged 65°–45°W and 20°S–0° plus 75°–65°W and 10°S–0°, smoothed with a 3-day running mean. The black line is for the TRMM climatology, the red line is for CCSM4, and the blue line is for CAM4.

  • View in gallery

    DJFM precipitation (mm day−1) and 900-hPa winds (m s−1) from the (a) ERA-Interim and TRMM climatologies, (b) CCSM4, and (c) CAM4.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 115 106 6
PDF Downloads 57 55 2

Monsoon Regimes and Processes in CCSM4. Part II: African and American Monsoon Systems

View More View Less
  • 1 The University of Texas at Austin, Austin, Texas
  • | 2 National Center for Atmospheric Research,* Boulder, Colorado
  • | 3 National Center for Atmospheric Research,* Boulder, Colorado, and CAWCR, Bureau of Meteorology, Melbourne, Victoria, Australia
© Get Permissions
Full access

Abstract

This is the second part of a two part series studying simulation characteristics of the Community Climate System Model, version 4 (CCSM4) for various monsoon regimes around the global tropics. Here, the West African, East African, North American, and South American monsoons are documented in CCSM4. Comparisons are made to an Atmospheric Model Intercomparison Project (AMIP) simulation of the atmospheric component in CCSM4 (CAM4), to deduce differences in the monsoon simulations run with observed SSTs and with ocean–atmosphere coupling. These simulations are also compared to a previous version of the coupled model (CCSM3) to evaluate progress. In most, but not all instances, monsoon rainfall is too heavy in the uncoupled AMIP run with the Community Atmosphere Model, version 4 (CAM4), and monsoon rainfall amounts are generally better simulated with ocean coupling in CCSM4. Some aspects of the monsoon simulations are improved in CCSM4 compared to CCSM3. Early-season rainfall in the West African monsoon is better simulated in CAM4 than in CCSM4 presumably because of the specification of SSTs in the Gulf of Guinea, but the Sahel rainfall season is captured better in CCSM4 as are the African easterly jet and the tropical easterly jet. Improvements in the simulation of the Sahel rainy season (July, August, and September) in CCSM4 compared with CCSM3 are significant, but problems remain in the simulation of the early season (May and June) in association with the misrepresentation of eastern Atlantic (Gulf of Guinea) SSTs. Precipitation distributions and the southwesterly low-level inflow in the North American monsoon are improved in CCSM4 compared to CCSM3. Both CAM4 and CCSM4 reproduce the seasonal evolution of rainfall over the South American monsoon region, but the location of maximum rainfall is misplaced to the northeast in both models.

The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Corresponding author address: Dr. Kerry H. Cook, Dept. of Geological Sciences, Jackson School of Geosciences, C1100 University Station, The University of Texas at Austin, Austin, TX 78712. E-mail: kc@jsg.utexas.edu

This article is included in the CCSM4 Special Collection.

Abstract

This is the second part of a two part series studying simulation characteristics of the Community Climate System Model, version 4 (CCSM4) for various monsoon regimes around the global tropics. Here, the West African, East African, North American, and South American monsoons are documented in CCSM4. Comparisons are made to an Atmospheric Model Intercomparison Project (AMIP) simulation of the atmospheric component in CCSM4 (CAM4), to deduce differences in the monsoon simulations run with observed SSTs and with ocean–atmosphere coupling. These simulations are also compared to a previous version of the coupled model (CCSM3) to evaluate progress. In most, but not all instances, monsoon rainfall is too heavy in the uncoupled AMIP run with the Community Atmosphere Model, version 4 (CAM4), and monsoon rainfall amounts are generally better simulated with ocean coupling in CCSM4. Some aspects of the monsoon simulations are improved in CCSM4 compared to CCSM3. Early-season rainfall in the West African monsoon is better simulated in CAM4 than in CCSM4 presumably because of the specification of SSTs in the Gulf of Guinea, but the Sahel rainfall season is captured better in CCSM4 as are the African easterly jet and the tropical easterly jet. Improvements in the simulation of the Sahel rainy season (July, August, and September) in CCSM4 compared with CCSM3 are significant, but problems remain in the simulation of the early season (May and June) in association with the misrepresentation of eastern Atlantic (Gulf of Guinea) SSTs. Precipitation distributions and the southwesterly low-level inflow in the North American monsoon are improved in CCSM4 compared to CCSM3. Both CAM4 and CCSM4 reproduce the seasonal evolution of rainfall over the South American monsoon region, but the location of maximum rainfall is misplaced to the northeast in both models.

The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Corresponding author address: Dr. Kerry H. Cook, Dept. of Geological Sciences, Jackson School of Geosciences, C1100 University Station, The University of Texas at Austin, Austin, TX 78712. E-mail: kc@jsg.utexas.edu

This article is included in the CCSM4 Special Collection.

1. Introduction

In this paper we document the West African, East African, North American, and South American monsoon regimes and associated processes for the Community Climate System Model, version 4 (CCSM4). This is the second of a two part series, with the first part (Meehl et al. 2012, hereafter Part I) studying the Asian–Australian monsoon in CCSM4. Output from the fully coupled CCSM4 simulation is compared to the atmosphere-only the Community Atmosphere Model, version 4 (CAM4) runs, which differ from the CCSM4 simulations in having observed SSTs prescribed and excluding atmosphere–ocean interactions. Additionally, comparisons to the previous generation of the coupled model (CCSM3) are used to document changes and improvements. The monsoons in CCSM3 were described by Meehl et al. (2006, hereafter M06), and can also be compared to monsoon simulations in a previous version of the model (Meehl and Arblaster 1998).

A description of the CCSM4 and the experiments analyzed in this paper are described briefly in section 2. Sections 36 include documentation of the West African, East African, North American, and South American monsoon systems, respectively. Conclusions follow in section 7.

2. Model and observed data descriptions

The standard CCSM3 (e.g., Collins et al. 2006) is compared to the new CCSM4 (Gent et al. 2011). The CCSM3 had a T85 atmospheric model with 26 levels in the vertical and was coupled to land and sea ice components as well as a nominal 1°-resolution ocean model going down to about ¼° in the equatorial tropics. As noted above, characteristics of the worldwide monsoon simulations in CCSM3 were described by M06.

CCSM4 includes a finite-volume 1° version of the atmospheric model CAM4, with improved components of ocean, land, and sea ice compared to CCSM3 (Gent et al. 2011). Grid points in the atmosphere are spaced roughly every 1° latitude and longitude, and there are 26 levels in the vertical. The ocean is a version of the Parallel Ocean Program (POP) with a nominal latitude–longitude resolution of 1° (down to ¼° in the equatorial tropics) and 60 levels in the vertical. No flux adjustments are used in either CCSM3 or CCSM4. Experiments analyzed include twentieth-century simulations with a combination of anthropogenic and natural forcings and a multicentury preindustrial control run (Gent et al. 2011). Atmospheric Model Intercomparison Project (AMIP) simulations with CAM4 were run with observed monthly mean SSTs from 1979 to 2005 and the same anthropogenic and natural radiative forcings as CCSM4 as well as the same resolution. The observed SST data are from the Hadley Centre Sea Ice and SST dataset (HadISST; Rayner et al. 2003). The model climatologies for CAM4 and CCSM4 are formed by averaging output from 1990 to 2005.

The Tropical Rainfall Measuring Mission (TRMM) precipitation data (3B42) are used for validating model output. The climatology is calculated from daily values at 1°-resolution for 1998–2010 (Huffman et al. 2007). While the TRMM data are well suited for tropical applications at higher resolution, the current 12-yr record may be somewhat short for establishing a climatology. For this reason the model and TRMM climatologies are also compared with the Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP) dataset monthly mean values on a 2.5° grid for 1979–2008 (Xie and Arkin 1996). These comparisons are not shown in the figures, but they indicate that the TRMM data are in general agreement with the CMAP data on that product’s coarser resolution, and that the TRMM data provide additional detail needed for validating the 1° CCSM4 output.

The primary standard of comparison used here for circulation, geopotential height, and temperature fields is the European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA)-Interim reanalyses (Simmons et al. 2006), calculated from daily values for 1990–2005 at 1.5° resolution. It is a different time period than that used for forming the TRMM climatology, which is necessary because TRMM does not extend back to 1990.

3. The West African monsoon

Figure 1a displays the time series of daily precipitation over West Africa, averaged from 10°E to 10°W and smoothed using a 5-day running mean, from the TRMM climatology (1998–2009). In the Northern Hemisphere winter months (November– February), the rainfall maximum can be characterized as a marine ITCZ, located close to the equator over the Gulf of Guinea. It begins to strengthen and move northward to the Guinean coast (located at about 4°N) in March. From mid-May until mid-July, maximum rainfall rates of 10 mm day−1 or more remain in place over the Guinean coast. In mid-to-late July the rainfall maximum shifts to the north abruptly, an event known as the West African monsoon jump that marks monsoon onset across the Sahel (Sultan and Janicot 2003; Hagos and Cook 2007). It remains centered near 12°N, with rainfall rates generally sustained above 6 mm day−1 from 8° to 15°N, until early September. After that time, maximum rainfall rates gradually decrease and the rainfall maximum moves equatorward smoothly.

Fig. 1.
Fig. 1.

Hovmöller diagrams of daily rainfall (mm day−1) averaged from 10°W to 10°E and smoothed using a 5-day running mean from the (a) TRMM, (b) CCSM4, and (c) CAM4 climatologies.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

The time series of precipitation over West Africa from the CCSM4 climatology is displayed in Fig. 1b. During the boreal winter months, a marine ITCZ forms with realistic rainfall rates, but located a few degrees south of its observed latitude (Fig. 1a). Rather than moving north and becoming stationary over the Guinean coast region in March through June, the CCSM4 precipitation maximum moves north continuously in May and June and does not cross into the Northern Hemisphere until late June. Beginning in April a secondary precipitation maximum, which is not observed, forms inland near 8°N. In late June, 2–3 weeks earlier than in the observations, the CCSM4 rainfall executes a jump to the north but, again, the precipitation maximum is about 5° south of its observed location. Similar to the observed rainfall, the southward retreat of the rainfall maximum is smooth and maximum rainfall rates decrease in the fall.

In the CAM4 climatology (Fig. 1c), the time series of precipitation is more realistic than that simulated by CCSM4 in February through June, with the latitude of the precipitation maximum accurately captured, presumably due to the specification of observed SSTs. However, CAM4 does not produce a discernible monsoon jump, and Sahel rainfall rates in boreal summer are low. In addition, the CAM4 simulation produces an unobserved precipitation enhancement in the fall and winter (October–January).

From the observations (Fig. 1a) it is clear that there are two distinct warm-season precipitation regimes over West Africa, namely, the May–June (MJ) period of strong Guinean coast rainfall and the July–September (JAS) period with rainfall in the Sahel, following the monsoon jump. Both model simulations also represent two distinct rainfall regimes approximately in MJ and JAS, so these averaging periods are used in the following figures of rainfall distributions and the regional circulation.

Figure 2a shows shaded contours of TRMM precipitation for MJ with vectors of the 900-hPa wind from the ERA-Interim reanalysis. The precipitation maximum stretches just south of Africa’s Guinean coast, supported by cross-equatorial southerly flow. There are maxima near the Cameroon highlands (near 8°E and 5°N) and just off the west coast near 10°W and 5°N. The 1 mm day−1 precipitation contour lies near 13°N. The West African westerly jet (Grodsky et al. 2003; Pu and Cook 2010) has not yet formed on the west coast at 10°N.

Fig. 2.
Fig. 2.

(a) MJ climatological precipitation (contours; mm day−1) from TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis. (b) Precipitation contours (mm day−1) and 930-hPa wind vectors from the CCSM4 climatology for MJ, and (c) precipitation contours (mm day−1) and 930-hPa wind vectors from the CAM4 climatology for MJ.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

In the MJ average, the CCSM4 simulation fails to bring the rainfall maximum to the Guinean coast (Fig. 2b). Maximum rainfall rates reside just south of the equator and a secondary band of enhanced precipitation forms inland near 8°N. The 1 mm day−1 contour over the Sahel is near 16°N, a few degrees north of its observed position. The model produces precipitation maxima over the Cameroon highlands, but the west coast maximum is barely represented. The southwesterly direction of the monsoon flow on the Guinean coast is exaggerated, to too strong, compared to the reanalysis (Figs. 2a), and the flow over the Sahara west of the Greenwich meridian is northeasterly instead of northerly. The West African westerly jet is captured, but it seems to be fully formed and it is quite wide in latitude.

The CAM4 simulation, with prescribed SSTs, produces a much more accurate simulation over the ocean in MJ than CCSM4. As seen in Fig. 2c, the MJ precipitation maximum has realistic magnitudes. It is well positioned along the Guinean coast, and the two maxima are realistic. Over land, the precipitation distribution is similar to that of the CCSM4 simulation. The low-level flow crossing the Guinean coast is stronger than observed, and the West African westerly jet has formed too early as in CCSM4.

The observed and modeled precipitation and low-level flow for JAS are displayed in Fig. 3. In the TRMM observations (Fig. 3a), the zonally oriented band of precipitation between 10°W and 5°E has moved north and is centered at about 10°N in the southern Sahel, while the central Guinean coast experiences a summer reduction in rainfall. The orographic rainfall maximum over the Cameroon highlands remains in place throughout the warm season and there is a maximum on the west coast near 10°N that is collocated with the West African westerly jet. The 1 mm day−1 precipitation contour is located at the northern edge of the Sahel, at about 18°N on average. Over the Sahara, the prevailing winds are northerly (from the Mediterranean region) east of 10°E, easterly from 5°W to 10°E, and northerly along the coast.

Fig. 3.
Fig. 3.

The (a) 1998–2009 JAS climatological precipitation (contours, mm day−1) from TRMM with 900-hPa wind vectors from the ECMWF-Interim reanalysis. (b) Precipitation contours (mm day−1) and 930-hPa wind vectors from the 1986–2005 CCSM4 climatology for JAS, and (c) precipitation contours (mm day−1) and 930-hPa wind vectors from the 1986–2005 CAM4 climatology for JAS.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

The CCSM4 and CAM4 simulations (Figs. 3b,c, respectively) are somewhat similar to each other during JAS; however, overall, CCSM4 produces a more accurate representation than CAM4 of the observed precipitation climatology. Both models move the zonal-mean precipitation north into the Sahel, but CCSM4 maintains a more zonally uniform rainfall band across the Sahel, similar to the observations. Also, CAM4 does not maintain the observed precipitation maximum over the Cameroon highlands into the summer months as in the observations and, to a lesser extent, CCSM4. The west coast maximum is stronger than observed in both simulations.

To evaluate possible improvements in CCSM4 compared with CCSM3, Figs. 4a and 4b display rainfall distributions for MJ and JAS from an 8-member ensemble of CCSM3 simulations. (The figures in M06 use different months for averaging and so are not directly comparable here.) CCSM3 has the same tendency as CCSM4 to place the precipitation maximum over the equatorial Gulf of Guinea in MJ instead of bringing it to the coast (~5°N), but in CCSM4 the erroneous southward extension of the rainband near 10°E is not reproduced (cf. Figs. 4a and 2b). In CCSM3, the 1 mm day−1 precipitation contour is located near 20°N over West Africa, but in CCSM4 it is located near 16°N, which is more accurate. CCSM4 also captures the observed orographic precipitation maximum over the Cameroon highlands in MJ while CCSM3 does not.

Fig. 4.
Fig. 4.

Precipitation climatology for 1981–2000 (mm day−1) averaged over eight ensemble members from CCSM3 for (a) MJ and (b) JAS.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

In JAS, improvements in CCSM4 (Fig. 3b) relative to CCSM3 (Fig. 4b) are striking. CCSM4 simulates a rainband stretching across the continent centered near 12°N, similar to the observations (Fig. 4a), but CCSM3 does not capture the zonal uniformity of rainfall and strong precipitation remains over the Gulf of Guinea.

The superiority of the CAM4 precipitation simulation compared with the CCSM4 simulation in MJ suggests that at least some of the inaccuracy in the coupled-model simulation may be associated with either the misrepresentation of Atlantic SSTs or the processes of air–sea coupling. Figures 5a and 5b display surface temperature for the MJ climatologies from the Interim reanalysis and the CCSM4 simulation. Note that the SSTs are observations included in the reanalysis output, but land surface temperatures are reanalysis products. In the reanalysis (and in the prescribed-SST CAM4 simulation), SSTs up to 302 K lie along the Guinean coast (~5°N) with cooler temperatures to the south associated with the formation of the Atlantic cold tongue. The rainfall maximum (Fig. 2a) is located near the same latitude as the SST maximum. In the CCSM4 simulation for MJ (Fig. 5b), maximum SSTs of 302 K stretch along the equator and there is no sign of the developing Atlantic cold tongue. The strongest precipitation rates (Fig. 2b) are located just south of the SST maximum in the Gulf of Guinea. This comparison suggests that the incorrect simulation of the precipitation distribution in the CCSM4 in MJ is associated with the incorrect simulation of SSTs in the Gulf of Guinea. Causality cannot be distinguished, however, since the low-level flow and SSTs are coupled. For example, note that the low-level flow in the reanalysis (Fig. 2a) is oriented rather precisely along the west coast of Africa in the Southern Hemisphere tropics, suggesting coastal upwelling of cool water. In CCSM4 (Fig. 2b), the flow near the coast is not as favorable for upwelling.

Fig. 5.
Fig. 5.

Surface temperature (K) for MJ from (a) the ERA-Interim reanalysis and (b) CCSM4; and for JAS from (c) the ERA-Interim reanalysis and (d) CCSM4. Shading interval is every 2 K while contour interval is every 1 K.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

Inaccuracies in the simulation of Atlantic SSTs in CCSM4 persist into the summer months, when the Atlantic cold tongue is well developed according to the reanalysis (Fig. 5c). The CCSM4 simulation (Fig. 5d) captures some cooling along the equator between 30° and 5°W, suggesting the presence of equatorial upwelling, but cool SSTs do not develop along the coast. Despite these inaccuracies in the SST simulation, CCSM4 produces an improved simulation of JAS rainfall compared with MJ. During these months, West African rainfall distributions are more closely coupled with continental conditions than in MJ, and CCSM4 produces a reasonable distribution of surface temperatures over West Africa (Fig. 5d).

During the monsoon months in the Sahel (JAS), the African easterly jet plays an important role in regulating rainfall amounts since it carries moisture away from the continent below the level of condensation (Cook 1999), and it may contribute to an unstable environment in which squall lines can develop and deliver intense rainfall events (Rowell and Milford 1993). Figure 6a displays the climatological zonal wind speed at 600 hPa from the ERA-Interim reanalysis. The African easterly jet extends across most of the continent near 10°N with higher wind speeds, exceeding 12 m s−1, over West Africa. The jet is produced in the CCSM4 simulation (Fig. 6b), although it does not extend inland from the west coast as strongly as in the reanalysis and it is centered about 4° of latitude too far north. In the CAM4 simulation (Fig. 6c), the African easterly jet is confined unrealistically to the western half of the continent and the region of easterly flow extends too far north.

Fig. 6.
Fig. 6.

The 600-hPa zonal wind speed (m s−1) for JAS from the (a) ERA-Interim reanalysis, (b) CCSM4 simulation, and (c) CAM4 simulation. The 200-hPa zonal wind speed for JAS from the (d) ERA-Interim reanalysis, (e) CCSM4 simulation, and (f) CAM4 simulation. Negative values are shaded.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

Upper-level flow over West Africa in summer is dominated by the tropical easterly jet. In the reanalysis (Fig. 6d), this jet exceeds speeds of 5 m s−1 and is centered near 5°N over West Africa. North of about 12°N, the upper-level flow is westerly. In CCSM4, the magnitude of the tropical easterly jet is 2–3 times the observed speed, the easterlies extend too far north, and the jet core is at 10°N instead of 5°N as observed. These flaws are even more pronounced in the CAM4 simulation (Fig. 6f).

4. The East African monsoon

Figure 7a shows the time series of daily precipitation over East Africa, averaged from 30° to 45°E and smoothed using a 5-day running mean, from the TRMM climatology. During boreal winter (December–February), the rainfall maximum exceeds 6 mm day−1 and is centered between 10° and 15°S. In March the precipitation maximum moves northward a little, and in mid-April it jumps close to the equator and weakens. A second northward jump occurs in June to about 10°N. As was the case over West Africa, the southward retreat of the rainfall is smooth in the boreal fall. Riddle and Cook (2008) document these observed discontinuities in the rainfall over East Africa in detail.

Fig. 7.
Fig. 7.

As in Fig. 1, but for 30°–45°E.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

The CCSM4 simulation reproduces some, but not all, of these complicated precipitation features (Fig. 7b). During the boreal winter months, rainfall rates are reasonable and located south of the equator, and a gradual northward movement of the maximum is reproduced in early spring. But the first jump, from 10°S to the equator in April, is not simulated. Instead, a second precipitation maximum forms between 5° and 10°N, and this second maximum is similar to that observed except it has formed too early in the year. The boreal fall retreat of the rainfall maximum is smooth, as is observed, but rainfall rates are exaggerated.

CAM4 (Fig. 7c) captures two jumps in the position of the precipitation maximum in boreal spring and summer, but the positions of the precipitation maxima are too far north. Also, the magnitudes of the precipitation rates are double or even triple the observed rates in summer. Overall, the CCSM4 simulation is more realistic than the CAM4 simulation over East Africa, although neither is perfect.

As shown in Fig. 7, rainfall over East Africa in boreal winter is greatest in the Southern Hemisphere. Figure 8a shows the December–February (DJF) precipitation climatology over East Africa from TRMM with 900-hPa wind vectors from the ERA-Interim reanalysis. On the continent, rainfall is greatest over northern Zambia and Mozambique, with stronger precipitation over the mountainous island of Madagascar to the east, and the strongest low-level flow is from the northeast. The CCSM4 climatology reproduces this distribution of rainfall quite well (Fig. 8b), but with rates that are higher than those in the TRMM climatology. The region’s northeasterly flow is also captured but, again, with higher velocity than in the reanalysis. CAM4, shown in Fig. 8c, has more realistic rainfall rates over the Indian Ocean, reflecting the fact that the calculated SSTs in this region in CCSM4 are 1–2 K warmer than observed (not shown).

Fig. 8.
Fig. 8.

DJF climatological precipitation (contours; mm day−1) from (a) TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis, (b) CCSM4, and (c) CAM4. JAS climatological precipitation (contours; mm day−1) from (d) TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis, (e) CCSM4, and (f) CAM4.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

During boreal summer (JAS), the greatest observed rainfall rates occur over northern Ethiopia (Fig. 8d). The Somali jet is in place, with wind speeds of 10–15 m s−1 forming the eastern portion of the anticyclonic flow about the Indian Ocean subtropical high centered near 65°E and 3°N (not shown). CCSM4 places the subtropical high at a similar location and captures the precipitation maximum over the Ethiopian highlands. However, the model produces an unobserved precipitation maximum north of Lake Victoria and the low-level wind speed is, again, stronger than in the reanalysis. CAM4 (Fig. 8c) produces a precipitation distribution and low-level wind field over the region that is much more similar to that in CCSM4 than to the observations. CAM4 also has a greatly exaggerated precipitation maximum over the southwestern Saudi Arabian peninsula in association with low-level flow that impinges too strongly on the region’s topography and the more general pattern of rainfall over the western Indian Ocean that extends too far west (see also Part I).

Of particular importance over East Africa is the timing of the rainy seasons since these dictate agricultural practice. Over Kenya, for example, there are two rainy seasons—the “long rains” of spring and the “short rains” of fall. These two periods are clearly seen in the time series of the TRMM precipitation climatology (black line), averaged 5°S–5°N and 35°–45°E to capture the rainfall maxima (see Fig. 8), in Fig. 9a. Farther north, for example, over Ethiopia (10°–20°N and 35°–45°E), there is a single summer rainfall season. While this structure in the rainfall time series is often attributed to having a “double pass” of the ITCZ over Kenya and a “single pass” over Ethiopia, an examination of the observed rainfall distributions in Figs. 8a and 8d makes it clear that the rainfall is not zonally uniform and so cannot be characterized by the same dynamics as a marine ITCZ. It is more correct to evaluate the response of the monsoon system to the distribution of solar heating.

Fig. 9.
Fig. 9.

Intraseasonal variation of rainfall (mm day−1) averaged 5°S–5°N and 35°–45°E (black line) and 10°–20°N and 35°–45°E (gray line) in the (a) TRMM, (b) CCSM4, and (c) CAM4 climatologies.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

Precipitation from CCSM4 averaged over the same two regions is shown in Fig. 9b. There are two precipitation maxima simulated over the southern region as in the observations, but the timing is inaccurate and the spring long rains are weaker than the fall short rains and the dry summer period is not as distinct. Over the northern averaging region the model simulates a peak in rainfall during the summer, but there are also strong rainfall rates in the spring which are not observed. CAM4 precipitation (Fig. 9c) is similar to the CCSM4 precipitation over these regions. The high rainfall rates simulated in the northern averaging region are associated with the anomalous high rainfall centered over the southwestern Arabian Peninsula in this model.

5. The North American monsoon

Previous simulations of the North American monsoon with global as well as regional models (e.g., Gutzler et al. 2009) showed some comparable systematic errors that were documented in CAM3 and CCSM3 (M06). These included less than observed summer rainfall over New Mexico and Arizona and no evidence of southwesterly low-level flow over that region. CAM4 and CCSM4 in Fig. 10 show notable improvement in both of those aspects of the North American monsoon simulation. There is also a better representation of the orographic precipitation over western Mexico, along with low-level convergence from the westerly and southwesterly winds from the Pacific, and the southeasterly winds from the Gulf of Mexico as observed.

Fig. 10.
Fig. 10.

JJA climatological precipitation (shading; mm day−1) from (a) TRMM with 900-hPa wind vectors (m s−1) from the ERA-Interim reanalysis, (b) CCSM4 with 930-hPa wind vectors, and (c) CAM4 with 930-hPa wind vectors.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

The annual cycle of precipitation over the North American monsoon region (103°–112°W, 20°–37°N; following Vera et al. 2006) is shown in Fig. 11 for the TRMM observations, CAM4, CCSM4, and CCSM3. All three model simulations overestimate the North American monsoon rainfall amounts throughout the year. The overestimate is similar in CCSM3 and CCSM4, but smaller in CAM4. While CCSM4 especially overestimates the rainfall in boreal fall and winter, this bias is reduced compared to CCSM3.

Fig. 11.
Fig. 11.

Daily rainfall (mm day−1) averaged 103°–112°W and 20°–37°N for the TRMM, CCSM4, and CAM4 climatologies.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

A distinct seasonal peak in boreal summer is observed, with rapid development in early June and decay over September and October. All three models capture the seasonality well, but they tend to begin the development one month early. The decline of the monsoon rains in the fall in the coupled models is not as distinct as observed because of the misrepresentation of winter rainfall.

6. The South American monsoon

Figure 12 displays daily rainfall from the TRMM, CCSM4, and CAM4 climatologies, averaged over the South American monsoon region (covering the entire basin from 65° to 45°W and the equator to 15°S, plus 75° to 65°W and the equator to 10°S) and smoothed using a 3-day running mean. Both model simulations capture the seasonal cycle of rainfall in the Amazon basin quite well. The most notable difference between the models and the TRMM observations is that rainfall during the monsoon season [December–March (DJFM)] is more variable on synoptic time scales in the observations than in the models. Averaged over both the monsoon and dry seasons [June–September (JJAS)], the models produce slightly less rainfall than the observations in this averaging region.

Fig. 12.
Fig. 12.

Daily precipitation (mm day−1) averaged 65°–45°W and 20°S–0° plus 75°–65°W and 10°S–0°, smoothed with a 3-day running mean. The black line is for the TRMM climatology, the red line is for CCSM4, and the blue line is for CAM4.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

Precipitation from the TRMM climatology during the South American monsoon season (DJFM) is plotted in Fig. 13a along with 900-hPa wind vectors from the ERA-Interim reanalysis. The Amazon basin precipitation maximum stretches across the continent from about 2°S and 70°W to the coast near 25°S and 45°W. The South Atlantic convergence zone (SACZ) extends diagonally off the continent over the South Atlantic Ocean, and there is a coastal maximum on the Atlantic coast near the mouth of the Amazon River. Note that the monsoon rainfall is clearly distinguished from the Atlantic marine ITCZ, which is located between the equator and 5°N. The strong zonal precipitation gradient across northeastern South America (the Nordeste region) is also an important feature of the South America precipitation climatology, and the TRMM observations resolve orographic precipitation along the eastern slopes of the Andes. A secondary precipitation maximum is located in the subtropics (near 30°S and 58°W).

Fig. 13.
Fig. 13.

DJFM precipitation (mm day−1) and 900-hPa winds (m s−1) from the (a) ERA-Interim and TRMM climatologies, (b) CCSM4, and (c) CAM4.

Citation: Journal of Climate 25, 8; 10.1175/JCLI-D-11-00185.1

As suggested by the low-level flow in the reanalyses (Fig. 13a), the primary moisture source for the South American monsoon rainfall is the tropical Atlantic Ocean. The onshore, northeasterly flow turns southward to flow along the eastern foothills of the Andes. The ERA-Interim reanalysis places strong northerly flow across the continental interior, with a turn toward the west south of about 25°S. The SACZ is associated with the northerly winds on the western side of the South Atlantic subtropical high.

The excellent agreement between the models and observations averaged over the Amazon basin (Fig. 12) indicates that the seasonality of the South American monsoon system is captured properly in the models, but it is somewhat misleading in representing the structure of the simulated precipitation fields. As seen in Figs. 13b and 13c, both the CCSM4 and CAM4 simulations maintain a South American precipitation maxima during the DJFM monsoon season supported by onshore flow from the equatorial Atlantic; however, maximum rainfall rates in both model simulations are somewhat high, and the maximum is misplaced to the northeast. The models also miss the precipitation maximum on the northeast coast. The modeled flow from the Amazon basin maintains its southward direction until about 25°S as in the reanalysis, but does not turn as sharply to the west so the band of easterly flow seen in the observations near 30°S is less pronounced. The models associate orographic rainfall with the Andes Mountains, but rates are too high and the region of high rainfall is too large. The marine Atlantic ITCZ is well simulated in CAM4, but the CCSM4 simulation shows a double-ITCZ structure. The SACZ is captured well in both models, as is the pronounced zonal precipitation gradient across the Nordeste region.

Simulated rainfall distributions in CCSM4 have changed compared to CCSM3, with some improvements. In particular the tendency for excessive rainfall located too far south compared to observations in CCSM3 (M06, Fig. 10) is corrected in CCSM4 (Fig. 13). In both CCSM4 and CAM4, maximum precipitation rates in the Amazon basin reach 15 mm day−1 while in CCSM3 and CAM3 they were just over 10 mm day−1. The location of the intense rainfall is not accurate, and CCSM4 and CAM4 produce a large precipitation maximum centered near 50°W and 6°S with high rainfall stretching to the northwest instead of the observed double maximum over the Amazon (65°W and 8°S) and along the northeast coast. As was the case for CCSM3 and CAM3, the current generation of models produces high rainfall rates (>12 mm day−1) in the Andes, while the TRMM data suggest that such high rates are confined to the eastern foothills.

7. Conclusions

The simulation characteristics of the West African, East African, North American, and South American monsoon systems are documented for CCSM4. Some comparisons are made to an AMIP simulation of the atmospheric component in CCSM4, the CAM4, to suggest changes to the monsoon simulations with and without accurate simulations of SSTs and/or ocean–atmosphere coupling, and to CCSM3, an earlier version of the coupled GCM. Many aspects of the monsoon simulations are improved in CCSM4 compared to CCSM3.

Over West Africa in May and June, before the onset of the monsoon rains in the Sahel, CCSM4 places maximum precipitation south of the equator instead of along the Guinean coast (~4°N) and forms a double-rainfall structure with a secondary maximum near 8°N. CCSM4 does, however, capture the sudden movement of the precipitation maximum into the Sahel at the monsoon onset, and maintains realistic rainfall rates through July, August, and September, unlike both CAM4 and CCSM3. The simulation of the African easterly jet and the tropical easterly jet are also improved in CCSM4 compared with CCSM3 and CAM4.

CCSM4 produces an improved simulation of East African rainfall compared with CAM4. While both models produce reasonable distributions of rainfall across the region, but with rates higher than observed, intraseasonal variations in rainfall are more realistic in CCSM4. In addition, CCSM4 avoids the generation of excessive rainfall over the southeastern Arabian Peninsula that occurs in the CAM4 simulation. CCSM4 produces two distinct rainy seasons over Kenya as observed (the long rains and short rains in boreal spring and fall, respectively), but the long rains in the model are weaker than the short rains and the summer season is too wet.

Precipitation distributions and the southwesterly low-level inflow of the North American monsoon are improved in CCSM4 compared to CCSM3, with an improvement to the simulation of orographic precipitation over western Mexico. All three models place the rainfall maximum in the summer as observed but produce too much precipitation throughout the year. Unlike CAM4, the coupled models fail to produce the winter dry season, but the overproduction of winter rain is less in CCSM4 than in CCSM3.

Both CAM4 and CCSM4 reproduce the seasonal evolution of rainfall over the South American monsoon region. The two models have very similar representations of the monsoon (DJFM) rainfall distribution, in which the region of maximum precipitation is displaced to the northeast, the rainfall maximum on the northeast coast is missed, and excessive rainfall amounts are placed over the Andes despite the increased resolution of the region’s steep topography. CAM4 does a better job in simulating the Atlantic ITCZ off the northeast coast of South America since CCSM4 produces a double-ITCZ structure. Both models are able to simulate the South Atlantic convergence zone reasonably well.

In general, there is an improvement in going from CCSM3 to CCSM4. Some of these improvements are related to the higher resolution and consequent improved representation of regional topography in CCSM4, and some are associated with improvements in the simulated SSTs in CCSM4. CAM4 provides insights into the role of ocean–atmosphere coupling and the correct simulation of SSTs. In some monsoon regions, CAM4 produces excessive monsoon precipitation that is brought down in closer agreement to observations in the CCSM4. The CAM4 simulation is better than CCSM4 for the aspects of the West African monsoon in which atmosphere–ocean coupling dominates because there are still systematic errors in the seasonal evolution of tropical SSTs (e.g., the tropical eastern Atlantic). But CCSM4 shows a marked improvement over CCSM3 when atmosphere–land surface interactions dominate (JAS).

Acknowledgments

Thanks to Dr. E. K. Vizy for technical support. We also thank two anonymous reviewers and Dr. B. Liebmann for their thoughtful comments. Portions of this study were supported by the Office of Science (BER), U.S. Department of Energy, Cooperative Agreement DE-FC02-97ER62402, and the National Science Foundation. The National Center for Atmospheric Research is sponsored by the National Science Foundation.

REFERENCES

  • Collins, W. D., and Coauthors, 2006: The Community Climate System Model, version 3 (CCSM3). J. Climate, 19, 21222143.

  • Cook, K. H., 1999: Generation of the African easterly jet and its role in determining West African precipitation. J. Climate, 12, 11651184.

    • Search Google Scholar
    • Export Citation
  • Gent, P. R., and Coauthors, 2011: The Community Climate System Model, version 4. J. Climate, 24, 49734991.

  • Grodsky, S. A., J. A. Carton, and S. Nigam, 2003: Near surface westerly wind jet in the Atlantic ITCZ. Geophys. Res. Lett., 30, 2009, doi:10.1029/2003GL017867.

    • Search Google Scholar
    • Export Citation
  • Gutzler, D. S., and Coauthors, 2009: Simulations of the North American monsoon: NAMAP2. J. Climate, 22, 67166740.

  • Hagos, S. M., and K. H. Cook, 2007: Dynamics of the West African monsoon jump. J. Climate, 20, 52645284.

  • Huffman, G. J., and Coauthors, 2007: The TRMM multi-satellite precipitation analysis: Quasi-global, multi-year, combined-sensor precipitation estimates at fine scale. J. Hydrometeor., 8, 3855.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., and J. M. Arblaster, 1998: The Asian–Australian monsoon and El Niño–Southern Oscillation in the NCAR climate system model. J. Climate, 11, 13561385.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., J. M. Arblaster, D. M. Lawrence, A. Seth, E. K. Schneider, B. P. Kirtman, and D. Min, 2006: Monsoon regimes in the CCSM3. J. Climate, 19, 24822495.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., J. M. Arblaster, J. Caron, H. Annamalai, M. Jochum, A. Chakraborty, and R. Murtugudde, 2012: Monsoon regimes and processes in CCSM4. Part I: The Asian–Australian monsoon. J. Climate, 25, 25832608.

    • Search Google Scholar
    • Export Citation
  • Pu, B., and K. H. Cook, 2010: Dynamics of the West African westerly jet. J. Climate, 23, 62636276.

  • Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Riddle, E. E., and K. H. Cook, 2008: Abrupt rainfall transitions over the Greater Horn of Africa: Observations and regional model simulations. J. Geophys. Res., 113, D15109, doi:10.1029/2007JD009202.

    • Search Google Scholar
    • Export Citation
  • Rowell, D. P., and J. R. Milford, 1993: On the generation of African squall lines. J. Climate, 6, 11811193.

  • Simmons, A., S. Uppla, D. De, and S. Kobayashi, 2006: ERA-Interim: New ECMWF reanalysis products from 1989 onwards. ECMWF Newsletter, No. 110, ECMWF, Reading, United Kingdom, 25–35.

    • Search Google Scholar
    • Export Citation
  • Sultan, B., and S. Janicot, 2003: The West African monsoon dynamics. Part II: The preonset and onset of the summer monsoon. J. Climate, 16, 34073427.

    • Search Google Scholar
    • Export Citation
  • Vera, C., and Coauthors, 2006: Toward a unified view of the American monsoon systems. J. Climate, 19, 49775000.

  • Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate, 9, 840858.

    • Search Google Scholar
    • Export Citation
Save