• Barnett, T. P., , L. Dümenil, , U. Schlese, , E. Roeckner, , and M. Latif, 1989: The effect of Eurasian snow cover on regional and global climate variations. J. Atmos. Sci., 46 , 661685.

    • Search Google Scholar
    • Export Citation
  • Bougeault, P., 1985: A simple parameterization of the large-scale effects of cumulus convection. Mon. Wea. Rev., 113 , 21082121.

  • Calvet, J-C., , J. Noilhan, , J-L. Roujean, , P. Bessemoulin, , M. Cabelguenne, , A. Olioso, , and J-P. Wigneron, 1998: Towards interactive vegetation in meteorological models, including physiology and plant growth in a soil–vegetation–atmosphere transfer scheme. Agric. For. Meteor., 92 , 7395.

    • Search Google Scholar
    • Export Citation
  • Charney, J. G., 1975: Dynamics of deserts and drought in the Sahel. Quart. J. Roy. Meteor. Soc., 101 , 193202.

  • Choudhury, B. J., 1993: Reflectivities of selected land surface types at 19 and 37 GHz from SSM/I observations. Remote Sens. Environ., 46 , 117.

    • Search Google Scholar
    • Export Citation
  • Cunnington, W. M., , and P. R. Rowntree, 1986: Simulations of the Saharan atmosphere-dependence on moisture and albedo. Quart. J. Roy. Meteor. Soc., 112 , 971999.

    • Search Google Scholar
    • Export Citation
  • Deardorff, J. W., 1977: A parameterization of the ground surface moisture content for use in atmospheric prediction models. J. Appl. Meteor., 16 , 11821185.

    • Search Google Scholar
    • Export Citation
  • ——,. 1978: Efficient prediction of ground surface temperature and moisture, with inclusion of a layer of vegetation. J. Geophys. Res., 83 , 18891903.

    • Search Google Scholar
    • Export Citation
  • Delworth, T. L., , and S. Manabe, 1988: The influence of potential evaporation on the variabilities of simulated soil wetness and climate. J. Climate, 1 , 523547.

    • Search Google Scholar
    • Export Citation
  • Déqué, M., , C. Dreveton, , A. Braun, , and D. Cariolle, 1994: The ARPEGE/IFS atmosphere model, A contribution to the French community climate modelling. Climate Dyn., 10 , 249266.

    • Search Google Scholar
    • Export Citation
  • —— and Coauthors, 1999: Documentation ARPEGE-CLIMAT. Centre National de Recherches Météorologiques, Météo-France, Toulouse, France.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., 1999: Assessing GCM sensitivity to soil wetness using GSWP data. J. Meteor. Soc. Japan, 77 , 367385.

  • ——, and Shukla, J., 1993: Observational and modeling studies of the influence of soil moisture anomalies on the atmospheric circulation. Predictions of Interannual Climate Variations, J. Shukla, Ed., NATO Series I, Vol. 6, Springer-Verlag, 1–23.

    • Search Google Scholar
    • Export Citation
  • Douville, H., 1998: Validation and sensitivity of the global hydrologic budget in stand-alone simulations with the ISBA land surface scheme. Climate Dyn., 14 , 151171.

    • Search Google Scholar
    • Export Citation
  • ——, and Royer, J-F., 1996: Sensitivity of the Asian summer monsoon to an anomalous Eurasian snow cover within the Météo-France GCM. Climate Dyn., 12 , 441448.

    • Search Google Scholar
    • Export Citation
  • ——, and Chauvin, F., 2000: Relevance of soil moisture for seasonal climate predictions: A preliminary study. Climate Dyn.,16, 719–736.

    • Search Google Scholar
    • Export Citation
  • ——, Royer, J-F., , and J-F. Mahfouf, 1995a: A new snow parameterization for the Météo-France climate model. Part I: Validation in stand-alone experiments. Climate Dyn., 12 , 2135.

    • Search Google Scholar
    • Export Citation
  • ——, ——, and ——,. 1995b: A new snow parameterization for the Météo-France climate model. Part II: Validation in a 3D GCM experiment. Climate Dyn., 12 , 3752.

    • Search Google Scholar
    • Export Citation
  • ——, Bazile, E., , P. Caille, , D. Giard, , J. Noilhan, , L. Peirone, , and F. Taillefer, 1999: Global Soil Wetness Project: Forecast and assimilation experiments performed at Météo-France. J. Meteor. Soc. Japan, 77 , 305316.

    • Search Google Scholar
    • Export Citation
  • ——, Royer, J-F., , D. B. Stephenson, , S. Tyteca, , J. Polcher, , P. Cox, , N. Gedney, , and P. Valdes, 2000: Impact of CO2 doubling on the Asian summer monsoon: Robust versus model-dependent responses. J. Meteor. Soc. Japan,78, 421–439.

    • Search Google Scholar
    • Export Citation
  • Eltahir, E. A. B., , and C. Gong, 1996: Dynamics of wet and dry years in West Africa. J Climate, 9 , 10301042.

  • Fennessy, M. J., , and J. Shukla, 1999: Impact of initial soil wetness on seasonal atmospheric prediction. J. Climate, 12 , 31673180.

  • Ferranti, L., , J. M. Slingo, , T. N. Palmer, , and B. J. Hoskins, 1999: The effects of land-surface feedbacks on the monsoon circulation. Quart. J. Roy. Meteor. Soc., 125 , 15271550.

    • Search Google Scholar
    • Export Citation
  • Folland, C. K., , T. N. Palmer, , and D. E. Parker, 1986: Sahel rainfall and worldwide sea temperatures. Nature, 320 , 602607.

  • Gates, W. L., 1992: AMIP: The Atmospheric Model Intercomparison Project. Bull. Amer. Meteor. Soc., 73 , 19621970.

  • Geleyn, J-F., and Coauthors, 1995: Atmospheric parameterization schemes in Météo-France's Arpege NWP model. Proc. Seminar on Parameterization of Subgrid-Scale Physical Processes, Reading, United Kingdom, ECMWF, 385–402.

    • Search Google Scholar
    • Export Citation
  • International GEWEX Project Office (IGPO), 1998: Global Soil Wetness Project: Preliminary report on the pilot phase. IGPO Publication Series, No. 29, 48 pp.

    • Search Google Scholar
    • Export Citation
  • Jarvis, P. G., 1976: The interpretation of leaf water potential and stomatal conductance found in canopies in the field. Philos. Trans. Roy. Soc. London, 273B , 593610.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., , and M. J. Suarez, 1996: The influence of land surface moisture retention on precipitation statistics. J. Climate, 9 , 25512567.

    • Search Google Scholar
    • Export Citation
  • Legates, D. R., , and C. J. Willmott, 1990: Mean seasonal and spatial variability in gauge-corrected global precipitation. Int. J. Climatol., 10 , 111127.

    • Search Google Scholar
    • Export Citation
  • Mahfouf, J-F., , and J. Noilhan, 1996: Inclusion of gravitational drainage in a land surface scheme based on the force-restore method. J. Appl. Meteor., 35 , 987992.

    • Search Google Scholar
    • Export Citation
  • Manzi, A. O., , and S. Planton, 1994: Implementation of the ISBA parameterization scheme for land surface processes in a GCM—an annual cycle experiment. J. Hydrol., 155 , 353387.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., 1994: Influence of the land surface in the Asian summer monsoon: External conditions versus internal feedbacks. J. Climate, 7 , 10331049.

    • Search Google Scholar
    • Export Citation
  • ——,. 1997: The south Asian monsoon and the tropospheric biennial oscillation. J. Climate, 10 , 19211943.

  • Mintz, Y., , and Y. V. Serafini, 1992: A global monthly climatology of soil moisture and water balance. Climate Dyn., 8 , 1327.

  • Morcrette, J-J., 1990: Impact of changes to the radiation transfer parameterizations plus cloud optical properties in the ECMWF model. Mon. Wea. Rev., 118 , 847873.

    • Search Google Scholar
    • Export Citation
  • Noilhan, J., , and S. Planton, 1989: A simple parameterization of land surface processes for meteorological models. Mon. Wea. Rev., 117 , 536549.

    • Search Google Scholar
    • Export Citation
  • Palmer, T. N., , C. Brankovic, , P. Viterbo, , and M. Miller, 1992: Modeling interannual variations of summer monsoons. J. Climate, 5 , 399417.

    • Search Google Scholar
    • Export Citation
  • Robock, A., , A. C. Schlosser, , K. A. Vinnikov, , N. A. Speranskaya, , J. K. Entin, , and S. Liu, 1998: Evaluation of the AMIP soil moisture simulations. Global Planet. Change, 19 , 181208.

    • Search Google Scholar
    • Export Citation
  • ——, Vinnikov, K. Y., , G. Srinivasan, , J. K. Entin, , S. E. Hollinger, , N. A. Speranskaya, , S. Liu, , and A. Namkhai, 2000: The global soil moisture data bank. Bull. Amer. Meteor. Soc., 81 , 12811299.

    • Search Google Scholar
    • Export Citation
  • Rowell, D. P., , C. K. Folland, , K. Maskell, , and M. N. Ward, 1995: Variability of summer rainfall over tropical North Africa (1906–1992): Observations and modelling. Quart. J. Roy. Meteor. Soc., 121 , 669704.

    • Search Google Scholar
    • Export Citation
  • Schär, C., , D. Lüthi, , U. Beyerle, , and E. Heise, 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12 , 722736.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and and Coauthors, 1997: Modeling the exchanges of energy, water, and carbon between the continents and the atmosphere. Science, 275 , 502509.

    • Search Google Scholar
    • Export Citation
  • Serafini, Y. V., 1990: The time scale of land surface hydrology in response to initial soil moisture anomalies: a case study. Tellus, 42A , 390400.

    • Search Google Scholar
    • Export Citation
  • Shukla, J., , and Y. Mintz, 1982: Influence of land-surface evapotranspiration on the earth's climate. Science, 215 , 14981501.

  • Stephenson, D. B., , K. R. Kumar, , F. J. Doblas-Reyes, , J-F. Royer, , F. Chauvin, , and S. Pezzulli, 1999: Extreme daily rainfall events and their impact on ensemble forecasts of the Indian monsoon. Mon. Wea. Rev., 127 , 19541966.

    • Search Google Scholar
    • Export Citation
  • Sud, Y. C., , and W. E. Smith, 1985: Influence of local land-surface processes on the Indian monsoon: A numerical study. J. Climate Appl. Meteor., 24 , 10151036.

    • Search Google Scholar
    • Export Citation
  • Vinnikov, K. Y., , and I. B. Yeserkepova, 1991: Soil moisture, empirical data and model results. J. Climate, 4 , 6679.

  • Walker, J., , and P. R. Rowntree, 1977: The effect of soil moisture on circulation and rainfall in a tropical model. Quart. J. Roy. Meteor. Soc., 103 , 2946.

    • Search Google Scholar
    • Export Citation
  • Webster, P. J., 1983: Mechanisms of monsoon low-frequency variability: Surface hydrological effects. J. Atmos. Sci., 40 , 21102124.

  • ——, Palmer, T. N., , M. Yanai, , J. Shukla, , V. Magnana, , T. Yasunari, , and R. Tomas, 1998: The monsoon: Processes, predictability and prediction. J. Geophys. Res., 103 , 14 45114 510.

    • Search Google Scholar
    • Export Citation
  • Wilson, M. F., , and A. Henderson-Sellers, 1985: A global archive of land cover and soils data for use in general circulation climate models. J. Climatol., 5 , 119143.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Geographical domains in which SM has been controlled (relaxed or limited); the control is gradually released from the inner boundary to the outer boundary of the domain.

  • View in gallery

    JJAS horizontal wind (m s−1) at 850 hPa over (top) south Asia and (bottom) Africa: (left) ECMWF 15-yr reanalysis, (right) ARPEGE control experiment.

  • View in gallery

    JJAS total precipitation (mm day−1) over (top) south Asia and (bottom) Africa: (left) Legates and Willmott climatology, (right) ARPEGE control experiment.

  • View in gallery

    Histograms of daily JJAS rainfall rates (RR: mm day−1) over India (5°–25°N, 70°–95°E) and west Sudan (10°–15°N, 20°W–20°E): (shaded) histogram of simulated precipitation, (solid line) histogram of observed precipitation (4-yr dataset over India, 23-yr dataset over Sudan)

  • View in gallery

    JJAS total soil moisture (kg m−2) over (top) south Asia and (bottom) Africa: (left) GSWP climatology (1987–88), (right) ARPEGE control experiment.

  • View in gallery

    Change in JJAS total soil moisture (kg m−2) prescribed over south Asia in WI0, WI1, WI7, and WI8, respectively (dashed lines represent negative values).

  • View in gallery

    Response of the JJAS 2-m air temperature (°C) over south Asia in WI0, WI1, WI7, and WI8, respectively.

  • View in gallery

    As in Fig. 7 but for JJAS sea level pressure (PMER; hPa).

  • View in gallery

    As in Fig. 7 but for JJAS horizontal wind (m s−1) at 850 hPa.

  • View in gallery

    As in Fig. 7 but for JJAS total precipitation (mm day−1).

  • View in gallery

    Change in JJAS total soil moisture (kg m−2) prescribed over Sudan–Sahel in WA0, WA1, WA7, and WA8, respectively.

  • View in gallery

    Response of the JJAS surface air temperature (°C) over south Asia in WA0, WA1, WA7, and WA8, respectively.

  • View in gallery

    As in Fig. 12 but for JJAS horizontal wind (m s−1) at 850 hPa.

  • View in gallery

    As in Fig. 12 but for total precipitation (mm day−1)

  • View in gallery

    Histograms of daily JJAS surface evaporation (mm day−1): (shaded) control experiment (WIF), (dashed line) dry experiment (WI0 or WA0), (solid line) wet experiment (WI1 or WA1). (Note change in ordinate scales between some panels.)

  • View in gallery

    As in Fig. 15 but for daily JJAS precipitation (mm day−1)

  • View in gallery

    Coefficient of variation of daily JJAS precipitation vs mean JJAS soil moisture (kg m−2) over the same domain: (top) India, (bottom) Sudan–Sahel; for each ensemble, the result of each of the six members is shown (thin symbols) as well as the six-member average (thick symbols)

  • View in gallery

    (top) Average JJAS evaporation, (middle) moisture convergence, and (bottom) precipitation over Sudan–Sahel vs mean JJAS soil moisture; for each ensemble, the result of each of the six members is shown (thin symbols) as well as the six-member average (thick symbols)

  • View in gallery

    As in Fig. 18 but over India

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 240 240 25
PDF Downloads 90 90 17

Influence of Soil Moisture on the Asian and African Monsoons. Part I: Mean Monsoon and Daily Precipitation

View More View Less
  • 1 Météo-France, Toulouse, France
© Get Permissions
Full access

Abstract

Soil moisture responds to precipitation variability but also affects precipitation through evaporation. This two-way interaction has often been referred to as a positive feedback, since the water added to the land surface during a precipitation event leads to increased evaporation, and this in turn can lead to further rainfall. Various numerical experiments have suggested that this feedback has a major influence on tropical climate variability from the synoptic to the interannual timescale. In the present study, ensembles of seasonal simulations (March–September) have been performed in order to investigate the sensitivity of the Asian and African monsoon rainfall to regional soil moisture anomalies. After a control experiment with free-running soil moisture, other ensembles have been performed in which the soil water content is strongly constrained over a limited area, either south Asia or Sudan–Sahel. Besides idealized simulations in which soil moisture is limited by the value at the wilting point or at the field capacity, more realistic experiments are relaxed toward the Global Soil Wetness Project (GSWP) soil moisture climatology. The results show a different sensitivity of the Asian and African monsoons to the land surface hydrology. Whereas African rainfall increases with increasing soil moisture, such a clear and homogeneous response is not found over the Indian subcontinent. Precipitation does increase over northern India as a consequence of wetter surface conditions, but the increased evaporation is counterbalanced by a reduced moisture convergence when averaging the results over the whole Indian peninsula. This contrasted behavior is partly related to the more dynamical and chaotic nature of the Asian monsoon, for which moisture convergence is about 2 times that found over Sudan–Sahel so that water recycling has a weaker influence on seasonal rainfall. It is also due to a different response of the frequency distribution of daily precipitation, and particularly to an increased number of strong convective events with decreasing soil moisture over India. Part II of the study will investigate how soil moisture also affects the interannual variability of the Asian and African monsoons.

Corresponding author address: Hervé Douville, CNRM/GMGEC/UDC, Météo-France, 42 Avenue Coriolis, 31057 Toulouse Cedex, France. Email: herve.douville@meteo.fr

Abstract

Soil moisture responds to precipitation variability but also affects precipitation through evaporation. This two-way interaction has often been referred to as a positive feedback, since the water added to the land surface during a precipitation event leads to increased evaporation, and this in turn can lead to further rainfall. Various numerical experiments have suggested that this feedback has a major influence on tropical climate variability from the synoptic to the interannual timescale. In the present study, ensembles of seasonal simulations (March–September) have been performed in order to investigate the sensitivity of the Asian and African monsoon rainfall to regional soil moisture anomalies. After a control experiment with free-running soil moisture, other ensembles have been performed in which the soil water content is strongly constrained over a limited area, either south Asia or Sudan–Sahel. Besides idealized simulations in which soil moisture is limited by the value at the wilting point or at the field capacity, more realistic experiments are relaxed toward the Global Soil Wetness Project (GSWP) soil moisture climatology. The results show a different sensitivity of the Asian and African monsoons to the land surface hydrology. Whereas African rainfall increases with increasing soil moisture, such a clear and homogeneous response is not found over the Indian subcontinent. Precipitation does increase over northern India as a consequence of wetter surface conditions, but the increased evaporation is counterbalanced by a reduced moisture convergence when averaging the results over the whole Indian peninsula. This contrasted behavior is partly related to the more dynamical and chaotic nature of the Asian monsoon, for which moisture convergence is about 2 times that found over Sudan–Sahel so that water recycling has a weaker influence on seasonal rainfall. It is also due to a different response of the frequency distribution of daily precipitation, and particularly to an increased number of strong convective events with decreasing soil moisture over India. Part II of the study will investigate how soil moisture also affects the interannual variability of the Asian and African monsoons.

Corresponding author address: Hervé Douville, CNRM/GMGEC/UDC, Météo-France, 42 Avenue Coriolis, 31057 Toulouse Cedex, France. Email: herve.douville@meteo.fr

1. Introduction

a. Soil moisture and climate variability

Numerical studies based on general circulation models (GCMs) have provided more and more evidence that climate is determined by a dynamic equilibrium in which the atmosphere affects the land surfaces and is affected by them. The pioneering work of Charney (1975), suggesting that desertification could be explained by an irreversible feedback between precipitation and surface albedo, is a meaningful illustration of this interaction. Other modeling studies have focused on soil moisture (SM), showing that the land surface hydrology is a crucial component of the climate system (for a review, see Dirmeyer and Shukla 1993). Back in the early 1980s, Shukla and Mintz (1982) suggested that SM anomalies could persist long enough to modify the atmospheric circulation over seasonal to interannual timescales. Such anomalies could be sustained through an evaporation feedback mechanism, which would be particularly efficient in the interior of the continents due to the strong recycling of precipitation (Serafini 1990). According to Cunnington and Rowntree (1986), this positive feedback could be more important than the Charney's hypothesis for explaining the maintenance of deserts.

The relationship between SM and precipitation variability on daily to seasonal timescales was also emphasized by several studies. Delworth and Manabe (1988) analyzed the temporal variability of a 50-yr GCM simulation. They showed that the spectrum of SM was red but with significant spatial variations. The red nature of the spectrum was more pronounced at higher latitudes due to the lower potential evaporation. In other words, SM excesses were dissipated more slowly at high latitudes where the energy available for evaporation was small. Through the use of a second GCM integration with prescribed SM, they demonstrated that interactive SM contributed significantly to near-surface atmospheric variability (Delworth and Manabe 1988). Koster and Suarez (1996) performed series of GCM simulations in which the timescale of SM retention was carefully controlled. As this timescale increased, the variance of daily precipitation decreased and the correlation between consecutive monthly precipitation increased. However, due to the lack of multiyear measurements of daily evaporation, the authors recognized that it was unclear how their results applied to the real world, so that it is still a challenge to determine the actual timescale of SM anomalies on the global scale. Recently, more realistic numerical experiments were designed in order to demonstrate the relevance of initial SM on seasonal atmospheric predictions (Fennessy and Shukla 1999; Douville and Chauvin 2000).

While precipitation has been extensively documented over the last few decades, SM is measured routinely only in a few locations and these ground measurements are generally not representative on the regional scale (Vinnikov and Yeserkepova 1991). Despite the recent initiative of Robock et al. (2000) to create a SM data bank with observations from about 600 stations covering a large variety of climates, it is not feasible to derive a global SM climatology from in situ measurements due to the high spatial variability of both precipitation and land surface properties. Remote sensing techniques, such as microwave measurements, may be used in order to obtain a better spatial coverage, but they have strong limitations and are still unable to provide reliable observations of subsurface SM (Choudhury 1993). SM fields can be derived from GCM simulations but are unreliable due to the significant biases that are still found in GCM precipitation, radiation, and low-level parameters. Using the data of the Atmospheric Model Intercomparison Project (AMIP; Gates 1992), Robock et al. (1998) compared simulated SM with in situ measurements from the former Soviet Union and the United States. Simulated SM was found to be quite different from the observations and to be strongly model dependent, especially in the Tropics.

For these reasons, most SM climatologies have been derived indirectly, using simple water budget models and observed monthly mean precipitation and radiation (see for instance, Mintz and Serafini 1992). Such products have a low spatial and temporal resolution that is a major obstacle for designing climate sensitivity experiments to SM anomalies. In the near future, more reliable SM climatologies could be obtained in the framework of the Global Soil Wetness Project (GSWP). This international initiative is aimed at producing a high-resolution SM climatology, by driving state-of-the-art land surface schemes (LSSs) with meteorological observations and analyses [International GEWEX Project Office (IGPO) 1998, hereinafter IGPO 1998]. As a pilot study, a 2-yr climatology (January 1987–December 1988) has been already produced at a 1° by 1° horizontal resolution by various LSSs. The results obtained at the Center for Ocean–Land–Atmosphere Studies and the Centre National de Recherches Météorologiques (CNRM) of Météo-France have been used to prescribe SM in global atmospheric simulations (Dirmeyer 1999; Douville and Chauvin 2000). These have shown a significant impact of SM on the simulated precipitation.

b. Soil moisture and monsoon precipitation

In the Northern Hemisphere, south Asia and tropical North Africa are the two main areas that experience the so-called monsoon climate. The year is divided into two distinct phases, the “dry” and the “wet” seasons, due to an oscillation in the large-scale circulation associated with the seasonal reversal of the land–sea temperature contrast (Webster et al. 1998). During the wet season, warm, moist, and disturbed winds blow inland from the tropical oceans and bring most of the annual precipitation. This rainy season is short (June–September) and unreliable, with large intraseasonal and interannual variations leading to severe droughts and floods that are critical for countries still having predominantly agrarian economies.

While many numerical studies have demonstrated the sensitivity of the Asian and African monsoons to sea surface temperatures (SSTs), showing significant correlations between SST and precipitation anomalies (see, e.g., Folland et al. 1986; Palmer et al. 1992), the sensitivity to SM variations has been much less investigated. However, other land surface parameters have been shown to influence the monsoon. Several GCM studies have confirmed the observational evidence that the Asian summer monsoon might be affected by the Eurasian snow cover of the former spring season (see, e.g., Barnett et al. 1989 or the more recent study by Douville and Royer 1996). The surface albedo has also been emphasized as an important factor for simulating the monsoon. Cunnington and Rowntree (1986) showed that the use of more realistic albedos over North Africa tended to transfer rainfall from region of higher albedos (deserts) to regions of lower albedos (forests). Sud and Smith (1985) noticed that both an increase in the regional surface albedo or a reduction in surface roughness were likely to weaken the Indian monsoon.

Sud and Smith (1985) also did an experiment in which they prescribed a zero evapotranspiration over India. It was found that the rainfall was essentially unaltered because the increased moisture convergence produced by the enhanced sensible heating of the planetary boundary layer largely compensated for the reduction in evapotranspiration. This conclusion was opposite to the results of Sikka and Gadgil (1980), also based on idealized experiments (dry or wet surface conditions), which indicated that the Indian monsoon rainfall was rather sensitive to regional SM. More recently, Meehl (1994) compared the relative contributions of external conditions (involving surface albedo) and internal feedbacks (involving SM) in a number of atmospheric GCMs. He concluded that there was a positive feedback between SM and precipitation over south Asia, with increased SM providing a surface moisture source for further monsoon precipitation. Meehl (1997) also suggested that SM could not contribute to the biennial signal found in the Indian monsoon since the SM anomalies contribute to latent heat flux anomalies for only one season after the summer monsoon in his coupled ocean–atmosphere GCM.

Although probably less frequently investigated, the sensitivity of the African monsoon to land surface conditions has also been emphasized by several authors. Using a simplified zonally symmetric GCM, Walker and Rowntree (1977) showed that the rainfall simulated over an idealized West African domain are very sensitive to the initialization of SM. They also analyzed observed rainfall data for the northern Sahel. They found a significant persistence of precipitation anomalies on the interannual timescale and discussed the possible role of SM in contributing to this low-frequency variability. Using a more realistic atmospheric GCM (despite prescribed zonal mean cloud amounts), Rowell et al. (1995) analyzed the variability of summer rainfall over tropical North Africa. They distinguished three main sources of variability: SST forcing, internal atmospheric variability, and SM feedback, which they assumed to be positive. By replacing the normal interactive SM scheme with a model-derived climatology, they showed that in some years SM plays a key role in the rainfall variability, but that in general SST forcing dominates. Eltahir and Gong (1995) also emphasized the possible role of land surface humidity in controlling the interannual variability of the West African monsoon. Using a more theoretical framework, they suggested that the distribution of SM could affect the meridional gradient of boundary layer entropy, and thereby the monsoon circulation.

Other studies paid more attention to the intraseasonal variability of the monsoon. Webster (1983) suggested that moist processes at the surface of the continents could induce significant feedbacks playing an important role in the intraseasonal variability of the monsoon rainfall and in the propagation of the convective events over the tropical landmasses. This idea was recently revisited by two numerical studies about the Indian monsoon variability. Ferranti et al. (1999) conducted idealized “perpetual July” global experiments with either prescribed or interactive SM in the zonal band between 40°N and 40°S. They concluded that the intraseasonal variability of the monsoon is a robust feature that is primarily related to an internal (purely atmospheric) mode of variability rather than to a response to land surface feedbacks. However, the hydrological surface feedbacks were shown to enhance the low-frequency monsoon variability (succession of active and break regimes) and thereby to affect the mean monsoon circulation. Kumar et al. (2000), manuscript submitted to Geophys. Res. Lett.) compared various atmospheric simulations using contrasted prescribed soil depths over India. Besides a control experiment, they performed sensitivity experiments with halved or doubled soil depths. They showed that increasing the soil depth led to a stronger inertia of SM and thereby to an increase in the low-frequency (from 20 to 60 days) variability of the Indian monsoon precipitation.

Last, SM and more generally land surface processes are important not only for simulating the present-day monsoon climate, but also for predicting the evolution of the monsoon precipitation for the decades to come. The response of the Asian monsoon to increased atmospheric concentration of greenhouse gases was shown to be sensitive to the surface hydrology, which can partly explain the spread found in the predictions of different GCMs (Douville et al. 2000).

c. Objectives of the study

The purpose of the present study is to investigate the relevance of SM for simulating the Asian and African monsoons. The focus is on the daily to seasonal timescales, while the influence of SM on the interannual variability will be explored in Part II of the article. In the continuation of previous numerical works, the study is based on global atmospheric simulations in which SM is either free running (control experiment) or controlled. The atmospheric model is version 2 of the Action de Recherche Petite Echelle Grande Echelle (ARPEGE) spectral model of Météo-France used at a T31 truncation. Since the focus is on the boreal summer monsoons, the study is based on ensembles of seasonal simulations from March to September.

In past studies, the lack of a reliable global SM dataset, has been a major obstacle to design realistic sensitivity experiments. The SM was prescribed using either a poor climatology such as Mintz and Serafini (1992) or the even more questionnable climatology derived from the considered GCM. Other simulations were conducted with more extreme conditions, the soil being set to saturation or being completely dry. The present study is based on the results of the recent GSWP that has led to presumably more realistic though model-dependent SM climatologies (Douville 1998). As it will be explained later, the monthly SM product that is here considered is the climatology that would be simulated by ARPEGE if there was no bias in the atmosphere. It is therefore meaningful to use this product in order to control the SM evolution over the monsoon areas. However, the GSWP product is only a 2-yr climatology (1987–88) so that it does not allow to explore a wide range of surface conditions. For this reason, less realistic experiments have been also performed using uniformly wet or dry soils.

The ultimate objective of the study is to assess the importance of a possible SM feedback onto the monsoon precipitation. This question is still a matter of debate since some studies conclude about a significant positive feedback through evaporation (Meehl 1994) while other numerical experiments suggest that the monsoon precipitation is not sensitive to SM (Sud and Smith 1985). Note that this soil–precipitation feedback has also been described as a relevant process in the interior of the mid and high-latitude continents. However, the understanding of this feedback has changed dramatically over the past decades. The early simplified view in which precipitation was primarily due to local evaporation was challenged in the 1960s, when it was realized that the average residence time of water molecules in the atmosphere was about one week. The role played by the water vapor transport was then emphasized but the classic water recycling scenario is still widely recognized as a major process, although the physical mechanisms that underly this process have not been thoroughly investigated until recently (Schär et al. 1999).

In the following section of the present study, the ARPEGE climate model and the Interactions between Soil Biosphere and Atmosphere (ISBA) land surface scheme will be briefly described, as well as the design of the seasonal climate simulations. Section 3 will discuss the control monsoon climate. Section 4 will analyze the sensitivity of the mean monsoon circulation and rainfall to SM, while section 5 will focus on daily precipitation. Section 6 will summarize the results and give the conclusions.

2. Methodology

a. The ARPEGE climate model

The ARPEGE climate model originates from the ARPEGE/Integrated Forecasting System (IFS) numerical weather prediction model developed jointly by Météo-France and the European Centre for Medium-Range Weather Forecasts (ECMWF). It is a spectral atmospheric model using a triangular truncation and a hybrid vertical coordinate (Déqué et al. 1994). In the present study, version 2 of the model is used with a T31 truncation (horizontal resolution of about 3.8°) and 19 vertical levels. Deep convection is parameterized according to the mass flux scheme of Bougeault (1985). Shallow convection is parameterized as part of the stability-dependent computation of the turbulent exchange coefficients by a modification of the Richardson number using the vertical gradient of specific humidity (Geleyn et al. 1995). The main novelties in comparison with version 1 are the implementation of an improved radiation scheme (Morcrette 1990) and of a convective gravity wave drag, as well as the introduction of a new spectral orography (Déqué et al. 1999).

At the earth's surface, the ISBA scheme is used to provide a boundary condition to temperature and moisture (Noilhan and Planton 1989; Manzi and Planton 1994; Mahfouf and Noilhan 1996). In version 1, ISBA had only two temperature levels in the soil. The deep soil temperature was restored toward climatological values in order to avoid spurious drift at high latitudes during the polar night. This method was rather restrictive and the deep soil temperature relaxation has been replaced by a four-layer heat diffusion scheme in version 2. Last, the treatment of snow-covered surfaces has also been improved by adding snow density and snow albedo as prognostic variables, thereby allowing the description of snow-aging processes and a more accurate calculation of the snow thermal and radiative properties (Douville et al. 1995a,b).

Despite containing the basic physics of land surface processes, ISBA needs only a few parameters depending on the types of soil and vegetation. The vegetation types were derived by Manzi and Planton (1994) from the global classification of Wilson and Henderson-Sellers (1985). Heat transfer in the ground is based on the force–restore method (Deardorff 1978). The treatment of the canopy has been simplified to avoid the numerical resolution of a specific foliage temperature. A single surface temperature is computed, which is representative of the whole soil–snow–canopy system. The scheme has three prognostic variables for liquid water: the reservoir of rain intercepted by the canopy, the surface volumetric water content, and the total volumetric water content. The soil hydrology assumes the same type of force–restore method for water as for heat conduction (Deardorff 1977), namely, a diffusive exchange between the surface and the deep soil layer:
i1520-0442-14-11-2381-e1
where Pg is the flux of liquid water reaching the soil, Eg is the evaporation at the soil surface, Etr is the transpiration rate, ρw is the density of liquid water, and τ is the duration of 1 day. Note that ws is the surface volumetric water content (arbitrary normalization depth d1), while wp is the total volumetric water content (the total soil depth d2 is prescribed according to the soil and vegetation types).

Two SM thresholds play an important role in the computation of the surface evapotranspiration, which are estimated as a function of the soil textural properties: the wilting point wwilt and the field capacity wfc. Below the wilting point, the SM stress becomes so high that the plants are unable to extract water from the ground through transpiration. Beyond the field capacity, the SM stress disappears so that the bare ground evaporation and the plants' transpiration reach the so-called potential rate. In Eq. (1), wseq represents the surface moisture when gravity balances the capillarity forces. Like the coefficients C1 and C2, it is a function of soil texture and soil moisture. The deep drainage coefficient C3 characterizes the velocity at which the water profile is restored to the field capacity wfc and also takes account of the soil texture (Mahfouf and Noilhan 1996).

Note finally that in the present version of ISBA, transpiration is controlled by a Jarvis-type stomatal resistance (Jarvis 1976; Noilhan and Planton 1989). Over recent years, a new generation of LSSs has appeared, which computes the stomatal conductance as a function of the assimilation of carbon dioxide (Sellers et al. 1997). Such a scheme has been also developed at Météo-France (Calvet et al. 1998), but is not used in the present study because further calibrations are needed in order to validate it on a wide range of climate and vegetation types. The authors believe that the conclusions of the study would not be greatly modified by the additional feedbacks (the scheme also simulates the biomass evolution through an allocation of the assimilated carbon) induced by a more interactive vegetation. These feedbacks are likely to be significant on the seasonal timescale, but are assumed to be weak in comparison with the direct effect of the SM anomalies that will be prescribed in the ARPEGE model.

b. Relaxation toward the GSWP climatology

The GSWP was aimed at studying the feasibility of producing a high-resolution soil wetness climatology, using meteorological observations and analyses to drive state-of-the-art LSSs (IGPO 1998). Global hydrometeorology and radiation datasets were provided on a 6-hourly basis in a common 1° × 1° format for the period 1987–88. ISBA, like each LSS participating in GSWP, was driven by these data (offline simulations) in order to produce a 2-yr SM climatology (Douville 1998; Douville et al. 1999). Such a climatology is obviously LSS dependent and is not a “universal” SM product. However, the 2-yr climatology produced by ISBA is fully consistent with the ARPEGE climate model: it has been obtained with the LSS and the soil and vegetation boundary conditions used in the GCM. It is the climatology that would be simulated by ARPEGE if its radiative and precipitation fluxes were close to the observations.

It is therefore particularly interesting to use the GSWP SM product in order to correct the SM evolution simulated in the ARPEGE model, which is affected by the systematic biases of the atmospheric part of the model (on the other hand, the GSWP climatology is still affected by the possible deficiencies and uncertainties in the ISBA scheme). Like for SSTs, prescribing SM at the lower boundary of the atmospheric model is a simple but questionable method. The simulated atmosphere is not necessarily in equilibrium with the prescribed SM so that the water and energy fluxes between the land surface and the low atmosphere can be unrealistic.

In order to mitigate this problem, the total soil water content of the 3D model is not really prescribed but only relaxed toward GSWP, while keeping the other water reservoirs of the ISBA scheme (the surface soil layer and the canopy) free to interact with the atmosphere. As a consequence, the land surface is still affected by, and can feed back onto, the atmosphere. For instance, the leaves and the ground surface are still affected by rain, and the subsequent increase in surface evaporation has a cooling impact on the low atmosphere, like in a fully interactive experiment. However, the memory of the rainy event is much shorter since the impact of precipitation on the total soil water content (and therefore on transpiration) is strongly limited because of the relaxation toward GSWP. This relaxation is introduced as a nudging in Eq. (2):
i1520-0442-14-11-2381-e3
where wclim is the daily total SM obtained by linear interpolation from the monthly GSWP climatology and τclim is the timescale of the relaxation, which has been fixed at 6 h so that the monthly average of wp remains close to the GSWP climatology.

c. Control and perturbed experiments

Since the purpose of the study is to analyze the daily to seasonal timescales rather than the interannual variability, all atmospheric simulations have been forced with prescribed climatological monthly mean SSTs. The SST dataset is the AMIP I climatology (Gates 1992). Two sets of parallel experiments are conducted to investigate separately the influence of SM on the Asian and African monsoons. In each set, the SM conditions are controlled over a limited domain: south Asia (5°–25°N, 60°–120°E) or Sudan–Sahel (10°–20°N, 20°W–40°E). In order to ensure a smooth transition between the area with controlled SM and the other continents with free-running SM, a 5° intermediate zone is defined around the controlled domain in which the SM constraint gradually disappears (Fig. 1). The two sets of experiments share the same control run, WIF, with globally interactive SM.

Since the focus is on the summer monsoons, it is useless to run full annual cycles. All the simulations start at the end of February thru mid-March and are carried on up to the end of September. In order to detect the SM signal versus the internal atmospheric variability, each experiment is actually an ensemble of six 7-month integrations using the same boundary conditions but different initial conditions (at 5-day intervals). The initial conditions are derived from a former ARPEGE experiment using climatological SSTs and interactive SM. Averaging the six integrations of each ensemble will allow us to assess the statistical robustness of our results. In the following, only the significant impacts of SM will be discussed. The ensembles will be compared over the June–July–August–September (JJAS) period, which can be roughly considered as the monsoon season in the Northern Hemisphere. Since the model is initialized between 23 February and 20 March, it runs at least two and a half months before the onset of the monsoon, which is hopefully enough to avoid any spurious behavior related to the spinup of the atmosphere.

Two sets of perturbed experiments (WI0, WI1, WI7, WI8 for the Asian monsoon; WA0, WA1, WA7, WA8 for the African monsoon) have been achieved. Since GSWP is only a 2-yr dataset, it was decided to conduct experiments in which SM is relaxed either toward GSWP 1987 or GSWP 1988 but also to perform more extreme ensembles in which the soil is either dry or wet. All experiments are summarized in Table 1. The WI7/WA7 and WI8/WA8 are relaxed toward GSWP 1987 and 1988, respectively. Remember that the relaxation is activated only over a limited domain (Fig. 1). In a previous study, Douville and Chauvin (2000) have performed similar ensembles, but with a global relaxation (W87 and W88). Here, it has been decided to control the SM evolution only over the focused domain in order to analyze the regional SM feedback and to avoid the complexity due to possible teleconnections.

In WI0/WA0, SM is not allowed to exceed the wilting point but is still freely evolving below wwilt. The vegetation is unable to extract water from the soil. Although weak, the bare ground evaporation can remain significant and somewhat variable. In WI1/WA1, SM is not allowed to be less than the field capacity, so that the surface evaporation is always potential (no soil water stress). Beyond wfc SM is still free running but is strongly relaxed toward wfc because of the deep drainage term in Eq. (2). Like for previous experiments, the dry or wet conditions are imposed only over the domains shown in Fig. 1, in order to avoid the unrealistic climate that would be simulated with entirely dry or wet continents.

3. Control climate

The sensitivity of a complex system is not simply related to, but is likely to depend on, its basic state. Before analyzing the climate response to changes in SM, it is therefore important to validate the control climate. Remember that the model is used with a T31 truncation so that we do not expect to capture the details of the observed climate. The purpose is rather to verify that the ARPEGE climate model is able to depict the main features of the Asian and African monsoons in a reasonable way. For the sake of simplicity, most of the validation is based on JJAS seasonal means.

Figure 2 compares the wind simulated at 850 hPa with the climatology derived from the 15-yr ECMWF reanalysis. The large-scale features of both the Asian and African summer monsoon circulation are captured by the model. However, the Asian low-level westerly jet is too strong and too zonal so that it continues too far eastward. Over North Africa, the intertropical convergence zone (ITCZ) is well located around 10°N, but the eastward flow is slightly overestimated on both sides of this area.

Figure 3 validates the distribution of the JJAS precipitation against the Legates and Willmott (1990) climatology. Over south Asia, the model simulates the strong precipitation observed over southeast Asia, but underestimates the Indian monsoon rainfall as well as the equatorial precipitation belt over the Indian Ocean. On the other hand, the control experiment shows overly strong rainfall over the southern slopes of the Tibetan Plateau compared to the Legates and Willmott climatology. The African monsoon is fairly well represented. The main deficiencies are the exaggerated northward extent of the equatorial precipitation, the overestimated rainfall over East Africa, and the underestimated rainfall over West Africa.

Thanks to the rain gauge data collected by the Indian meteorological department over India, and by the French Institut de Recherche pour le Développement, it is also possible to conduct a partial validation of the daily precipitation. Figure 4 displays the histograms of daily rainfall averaged over two continental areas where observations are available: the Indian peninsula (5°–25°N, 70°–95°E) and the west of Sudan (10°–15°N, 20°W–20°E). Over both regions, the distribution of the simulated precipitation rates shows significant biases. The all-India rainfall rates are underestimated with frequent intensities around 5 mm day−1 while the observed data suggests a broader spectrum with maximum frequencies between 4 and 10 mm day−1. On the other hand, the number of both weak and strong precipitation events seems to be reasonable. This is not true over West Africa where the model tends to overestimate the number of both dry and heavy rainfall (more than 10 mm day−1) days, while the frequency of medium precipitation rates (between 2 and 8 mm day−1) is underestimated.

Such a validation of GCM precipitation on a daily basis is somewhat unusual and must be considered with caution because of the possible limitations in the rain gauge data (problems for measuring weak and strong precipitation, spatial coverage of the rain gauge network). Nevertheless, it should be done more systematically when it is possible since it provides a more detailed picture of the model's behavior than the monthly mean values. In the present study, the simulated distribution of the daily rainfall rates shows significant deficiencies and is probably biased enough to affect the precipitation response to changes in SM. For this reason, we do not claim to deliver a quantitative estimate of the sensitivity of the monsoon rainfall, but only a qualitative assessment of the soil–precipitation feedback.

Figure 5 compares the total SM simulated in WIF with the GSWP climatology which has been used to relaxe SM in some of the perturbed experiments. Despite the previously discussed deficiencies in the simulated rainfall, the main features of the GSWP SM distribution are captured both over Asia and North Africa. However, some of the precipitation errors are reflected in the SM simulation, with, for example, overestimated SM over Sahel and East Africa, or underestimated SM over western India. These significant discrepancies between the ARPEGE results and the GSWP climatology will be strongly reduced in the experiments where the simulated soil water content will be relaxed toward GSWP.

4. Impact of SM on the mean monsoon

a. Asian monsoon

Figure 6 shows the JJAS SM anomalies that are prescribed over south Asia in the perturbed experiments WI0, WI1, WI7, and WI8, in comparison with the control experiment, WIF, with free-running SM. Because of the control SM distribution, and especially the dry bias over India, the dry and wet experiments, WI0 and WI1, are not symmetric. WI0 shows a negative SM anomaly over southeast Asia, while WI1 is wetter than WIF over India. The relaxed experiments, WI7 and WI8, have similar SM anomaly patterns, which means that the SM biases of the control simulation are stronger than the differences between 1987 and 1988 in the GSWP climatology. These patterns are a combination of the anomalies prescribed in the dry and wet experiments, but with smaller magnitude.

Figure 7 shows the SM impact on the simulated surface air temperatures. As expected, positive (negative) SM anomalies are associated with negative (positive) temperature anomalies due to the effect of evapotranspiration on the surface energy budget. India is more sensitive than southeast Asia, as proved by the larger impact found in WI1 than in WI0, and by the absence of temperature anomalies over southeast Asia in WI7 and WI8. This is mainly due to the vegetation distribution. While the northwest of the Indian subcontinent is relatively arid, southeast Asia is partly covered with rain forests that intercept and reevaporate a significant amount of the precipitation. Since only the dry part of the canopy contributes to the plants' transpiration, the interception loss limits the impact of SM on the surface energy budget.

The monsoon being the atmospheric response to the reversal of the land–sea temperature contrast between the Asian continent and the Indian Ocean, the previously discussed temperature anomalies are likely to have significant impacts on the large-scale circulation. This is confirmed by Figs. 8 and 9 illustrating the changes in the sea level pressure and the wind at 850 hPa. Over India, a dry (wet) SM anomaly is associated with a strengthening (weakening) of the monsoon trough and of the southwest monsoon flow. In keeping with previous studies such as Meehl (1994), the Indian thermal low is very sensitive to changes in the surface energy budget. Over southeast Asia, the atmospheric response is less clear due to the limited changes in the 2-m temperature. As expected, the sea level pressure decreases in WI0, but this is also somewhat true in WI1. A negative sea level pressure anomaly is centered over the Pacific Ocean east of south China, which is probably the result of an indirect effect whereby a weakening of the Indian monsoon leads to a zonal shift in the Walker circulation. In keeping with the size of the SM anomalies, the atmospheric response is weaker in WI7 and WI8, but is consistent with the results found in WI1 over northern India. Compared to WI7, WI8 indicates a stronger monsoon flow over southern India, which is consistent with the observed difference between 1987 and 1988 (respectively an El Niño and a La Niña year), suggesting that SM contributes to the interannual variability of the monsoon. This question will be further investigated in Part II of the present study.

Figure 10 shows the response of the monsoon precipitation to the prescribed SM anomalies. Remember that in the control experiment, WIF, the Asian summer monsoon rainfall is underestimated over northern India, but overestimated over Tibet and southeast Asia. The changes in precipitation found in experiments WI7 and WI8 indicate a partial but significant correction of these deficiences. Over India, a distinction must be done between the northern and southern parts of the peninsula, since stronger precipitation in the north is often associated with lower precipitation in the south (except in WI8 and it will be discussed later). Note that positive SM anomalies over northern India (in WI1, WI7, and WI8) lead to a local enhancement of precipitation despite the weakening of the monsoon flow. Two processes are competing in this area: 1) increased SM means more evaporation and therefore more precipitation, 2) more evaporation also leads to surface cooling which decreases the meridional temperature gradient and thereby the monsoon circulation and precipitation. In the present study, the first mechanism seems to be dominant over northern India, but it will be shown later that the competition between both processes is more obvious when averaging the results over the whole Indian peninsula. Over southeast Asia, both dry and wet experiments produce a negative precipitation anomaly, suggesting that it is not only the local SM and surface temperature anomalies that are important, but also their gradient between India and southeast Asia, which is similar in WI0 and WI1.

b. African monsoon

The SM anomalies prescribed over Sudan–Sahel in experiments WA0, WA1, WA7, and WA8 are shown in Fig. 11. Note that WA0 and WA1 are not symmetric due to the SM gradient between equatorial and subtropical North Africa. Experiments WA7 and WA8 show very similar patterns, thereby confirming that the model biases often exceed the interannual variability estimated from the GSWP climatology. Except in the extreme western and eastern parts of the perturbed domain, the SM anomalies are negative, but their magnitude is weaker than in WA0 over West Africa.

Figure 12 depicts the response of the surface air temperature. Not surprisingly, WA0, WA7, and WA8 show a significant warming over Sudan–Sahel, which is proportional to the prescribed SM anomaly. Conversely, a cooling is found over Sudan–Sahel in the wet experiment WA1. Note that this cooling is weak in the center of the domain, due to the very low vegetation cover and the limited effect of the deep SM on the bare ground evaporation (through the exchange of water between the deep and the surface soil layers). The sea level pressure anomalies (not shown) are clearly correlated to the temperature changes, with higher temperatures resulting in lower pressures and vice versa.

The SM impact on the 850-hPa circulation is illustrated in Fig. 13. A significant intensification of the ITCZ appears over Sudan in the dry experiment WA0, as indicated by the stronger latitudinal shear of the zonal wind. This intensification is also found in WA7 but is less clear in WA8. In the wet Sahel experiment, the eastward monsoon flow is reduced over Soudan, but the convergence is stronger over West Africa. In keeping with this result, WA1 shows increased rainfall in this area as well as over East Africa (Fig. 14). On the other hand, the SM increase is not sufficient to induce precipitation over northern Sahel. The other experiments indicate a decrease in the monsoon rainfall over West Africa, particularly clear in WA0. The patterns of the precipitation response is more complex over East Africa with a succession of positive and negative anomalies.

5. Impact of SM on the daily precipitation

a. Daily evaporation and precipitation rates

Figure 15 shows how the daily surface evaporation rates simulated by ARPEGE are affected, regionally, by the prescribed SM conditions from June to September. Only the extreme experiments (either dry or wet) are compared with the control experiment, WIF. The histograms are constructed by dividing the distribution in intervals of 0.25 mm day−1 and estimating the frequency of each interval from the six time series of 122 days (JJAS) available for each ensemble. Four areas are distinguished: the overall south Asian (5°–25°N, 60°–120°E) and North African (10°–20°N, 20°W–40°E) domains, as well as the two subdomains where daily observations of precipitation have been available (Fig. 4), namely, India (5°–25°N, 70°–95°E) and west Sudan (10°–15°N, 20°W–20°E).

Not surprisingly, SM has a strong effect on the frequency distribution of the daily evaporation rates. This effect is regionally dependent and is not only a change in the mean evaporation. The reader is reminded that the total soil water content is not the only difference between WIF and the other experiments that use prescribed SM instead of free-running SM. In these experiments, there is still a significant variability in the daily evaporation. This is due to the variability of the atmospheric demand and to the fact that the surface SM and the reservoir of rain intercepted by the canopy are still evolving freely, so that the bare ground evaporation and the interception loss are much less controlled than the plants' transpiration. Over south Asia, the average free-running SM is between the wilting point and the field capacity, so that the dry and wet experiments show symmetric histograms compared to the control ensemble. Over India, the control climate has a dry bias and the average free-running SM is relatively dry. As a result, the evaporation rates are more affected in WI1 than in WI0.

Over Sudan–Sahel, SM has a strong impact on the shape of the histogram. Decreasing SM leads to reduced mean evaporation but also to increased evaporation variability. This impact also appears over India, but is more obvious over Sudan–Sahel. It is probably due to the larger extent of bare soil (nonvegetated surfaces) in North Africa. Surface evaporation is made up of three components in the Tropics: the bare soil evaporation, the interception loss (evaporation from the wet part of the canopy), and the transpiration (dry part of the canopy). While transpiration depends on the deep SM, which is controlled in the dry and wet experiments, the other components are not directly influenced by the deep SM and are therefore more variable. Both SM and vegetation control the variability of surface evaporation, and a dry and bare soil is a stronger source of high-frequency variability for the atmosphere than a wet and vegetated surface.

Figure 16 shows histograms of daily precipitation with intervals of 2 mm day−1. Over India, the mean precipitation rate is not very sensitive to SM, and this despite the strong sensitivity of the mean evaporation. On the other hand, SM has a significant impact on the precipitation variability. The WI1 shows an increased number of days with medium rainfall (between 4 and 8 mm day−1) in comparison with WIF, while, on the contrary, WI0 shows an increased number of days with low or heavy precipitation. Wet surface conditions tend to guarantee a minimum precipitation rate but also to limit the frequency of strong convective events in the ARPEGE GCM. The results are less clear when averaged over the whole south Asian domain, but there is still a flattening of the histogram as a response to decreased SM. Over west Sudan and Sudan–Sahel, SM affects both the mean and the variability of daily precipitation. As over India, low rainfall rates (below 4 mm day−1) are more frequent over dry than over wet soils, but dry soils do not favor the occurrence of precipitation rates exceeding 8 mm day−1.

b. Coefficient of variation

Figure 17 summarizes how SM affects the distribution of the daily precipitation. For each integration of each ensemble, it shows the scatterplot of the estimated coefficient of variation (CV: standard deviation normalized by mean daily precipitation) against the spatial (within the considered domain) and temporal (JJAS) mean SM. All experiments are considered: interactive SM, dry and wet cases, relaxation toward GSWP 1987 and 1988. As far as the Asian monsoon is considered, it was decided to focus on the Indian domain where SM was more contrasted between each experiment than over southeast Asia. The coefficient of variation of daily rainfall is an interesting parameter to analyze, since it allows one to detect a change in standard deviation that is not simply related to a variation in mean precipitation. Besides the values obtained for each individual integration of a given ensemble, the average coefficient of the ensemble is also plotted.

Note that the range of SM that is explored (between the wilting point and the field capacity) over both the Asian and African domains is much wider than the range that is shown by the GSWP climatology between 1987 (WI7 and WA7, respectively) and 1988 (WI8 and WA8, respectively). Note also that all members of the ensemble with interactive SM (WIF) lie outside the 1987–88 interval, and this despite the fact that 1987 and 1988 represent a contrasted sample of observed monsoon precipitation (especially over India) due to the opposite SST anomalies in the equatorial Pacific. These remarks raise a number of issues that are common to many GCM studies. First, SM results from a balance between various contributions (precipitation, evaporation, runoff) that have similar magnitude and are not perfectly simulated, so that it is extremely difficult to achieve a realistic simulation of this quantity, even in a state-of-the-art GCM. Second, most numerical sensitivity studies, in which SM is prescribed at extreme values (either very low or very high) or from a GCM climatology are far beyond the range of the observed SM variability.

Figure 17 indicates that the daily monsoon rainfall produced by the ARPEGE model is quite sensitive to the simulated surface hydrology. The coefficient of variation decreases with increasing SM. This trend is more pronounced over India than over Africa, which is related to the higher sensitivity of the mean precipitation over Sudan–Sahel. Over both domains, the experiments relaxed toward GSWP (especially WI8 and WA8 relaxed toward year 1988 of GSWP) do not fit very well the monotonic trend indicated by the extreme (dry and wet) ensembles. This result raises some doubts about the conclusions that can be drawn from idealized experiments. In other words, the change in daily precipitation found between WI0 and WI1 (or WA0 and WA1) is not necessarily a good estimate of the precipitation sensitivity to SM, not only in the real world, but also in the GCM when a more realistic range of SM is explored.

Note, however, that the decrease of the coefficient of variation between WI7 and WI8 (WA7 and WA8) is in qualitative agreement with the decrease found between WI0 and WI1 (WA0 and WA1). Note also that the specific behavior of WI8 and WA8 could be simply a “sampling artifact” since the average coefficient of variation can be sensitive to some outliers in six-member ensembles. The only robust qualitative result is that the variability of the daily monsoon rainfall decreases with increasing SM. Over India, this is mainly due to the flattening of the frequency distribution of the precipitation rates with decreasing SM. Over Sudan–Sahel, this is rather the consequence of increased precipitation with increasing SM.

6. Summary and conclusions

The soil-precipitation two-way interaction is commonly referred to as a positive feedback, since the water added to the land surface during a precipitation event leads to increased evaporation, and this in turn can lead to further rainfall. Yet, the relevance and magnitude of this potential feedback is poorly known, and this limited knowledge is mainly based on idealized GCM sensitivity studies with unrealistic SM conditions.

While precipitation has been extensively documented over the last few decades, SM is measured routinely only in a few locations and cannot be retrieved from current satellite measurements. For this reason, SM climatologies still have to be derived indirectly, using water budget models. Such a work has been recently achieved at CNRM in the framework of the GSWP. The ISBA land surface scheme used in the ARPEGE GCM has been driven by 6-hourly meteorological analyses in order to produce a 2-yr (1987–88) global SM climatology at a 1° by 1° horizontal resolution. This dataset is not a universal climatology but is fully consistent with ARPEGE. It is the climatology that the GCM would produce if its atmospheric biases were close to zero.

In the present study, this climatology has been used to prescribe SM over a limited domain (either south Asia or Sudan–Sahel) in ensembles of global seasonal atmospheric simulations (March–September). The main objective is to investigate the sensitivity of the Asian and African monsoon rainfall to regional SM anomalies. Besides a control experiment with free-running SM, idealized simulations in which SM is limited by the value at the wilting point or at the field capacity have been performed, as well as presumably more realistic experiments in which SM is relaxed toward the GSWP climatology. The sensitivity of the monsoon precipitation has been analyzed both on the daily and seasonal timescales.

On the seasonal timescale, the Indian and African monsoons do not respond in a similar way. Over Sudan–Sahel (Fig. 18), the monsoon rainfall increases with increasing SM, which is due to an increase in surface evaporation without any significant change in the atmospheric moisture convergence. Such a response is not found over the Indian subcontinent (Fig. 19). Although positive SM anomalies lead to a local enhancement of monsoon precipitation over northern India, the rainfall sensitivity to SM is weak on the continental scale, since the increased evaporation is counterbalanced by a decreased moisture convergence. As compared with Sudan–Sahel, India is embedded in a more intense monsoon flow. The atmospheric moisture convergence is stronger and is also more sensitive to the land surface conditions. SM affects the Indian monsoon precipitation through two competing processes: 1) a recycling effect whereby larger evaporation leads to larger rainfall and 2) a dynamical effect whereby surface evaporation cools the land surface, hampers the deepening of the north Indian thermal low, and weakens the monsoon flow.

Note, however, that experiment WI8 has a specific behavior and shows stronger moisture convergence than WI7 despite wetter surface conditions. This result will be further investigated in Part II of the present study and will be related to regional circulation patterns, which do not appear if the relaxation toward GSWP is implemented not only over south Asia but over all the continents, as in experiments W87 and W88 (see Table 1) discussed by Douville and Chauvin (2000). These regional patterns also illustrate the fact that the precipitation anomalies are less spatially homogeneous over south Asia than over Africa. Despite the use of ensemble simulations, it might reflect the dynamical and chaotic nature of the Asian monsoon, where a single storm can bring extreme amounts of water over a few days and have a sizable impact on the estimated seasonal rainfall (Stephenson et al. 1999). Note also the apparent contradiction between the absence of change in precipitation from WI0 to WI1 and the link between SM and precipitation when considering the various members of WIF. In WIF, the soil water content is not controlled so that the soil–precipitation feedback is masked by the dominant impact of precipitation on SM. In the other experiments, SM is not driven by precipitation so that the feedback is highlighted.

On the daily timescale, the sensitivity of the Indian and African precipitation is also somewhat different. Both domains show an increase in the coefficient of variation (standard deviation normalized by the mean precipitation rate) with decreasing SM, but this response is not related to the same modification of the daily rainfall histograms. Over India, the increasing variability is due to the flattening of the frequency distribution, while it is mainly related to a decrease in the mean precipitation rate over Sudan–Sahel. In other words, dry soil tend to promote extreme daily rainfall rates over India, where the atmospheric moisture convergence can be strong enough to sustain intense convective events despite the dry surface conditions. This change in the tail of the daily rainfall distribution is not found over Sudan–Sahel, where moisture convergence is lower and precipitation depends more on surface evaporation. Note that the African response could have been different if the Guinean coast had been considered instead of Sudan–Sahel.

In summary, the soil–precipitation feedback is regionally dependent. In keeping with the results of Sud and Smith (1985), the present study suggests that the feedback is weak over India because of the competitive impacts of SM on surface evaporation and moisture convergence. On the other hand, the feedback is clearly positive over Sudan–Sahel where the moisture convergence is weaker and less sensitive to the surface hydrology, so that the recycling of precipitation through local evaporation is a dominant mechanism. The study also shows that this contrasted behavior of the seasonal monsoon precipitation is partly related to a different response of the daily precipitation rates. While both domains show an increased frequency of low precipitation rates with decreasing SM, India also shows an increased frequency of strong convective events so that the mean precipitation rate is not modified much.

It is difficult to assess how model dependent these numerical results are and how they apply to the real world. As illustrated in Fig. 4, the ARPEGE climate model shows some deficiencies in simulating the histogram of daily rainfall over India and west Sudan, which raises some doubts about the sensitivity of this distribution to the prescribed SM. In the future, more attention should be paid to the daily and synoptic timescales in GCM simulations. Many GCM studies focus on the monthly to annual timescales, but these scales are partly controlled by higher-frequency processes. It is therefore necessary to validate the climate models on a wide range of scales since their climatology can be realistic despite being the result of an unrealistic simulation of the synoptic timescale.

In a symmetric way, it is also crucial to validate the atmospheric models on the interannual timescale, in order to verify if their sensitivity to an observed perturbation in the SST forcing or in the land surface boundary conditions, is realistic or not. This issue will be the focus of Part II of the present study. Although tropical SSTs represent a major source of interannual variability, it will be shown that SM also contributes to the variability of the Asian and African monsoons and that the soil–precipitation feedback does not necessarily reinforce the atmospheric response to the prescribed SST anomalies.

Acknowledgments

The authors are indebted to Michel Déqué and Pascal Marquet who are in charge of the ARPEGE climate model. They wish to thank the anonymous reviewers for their useful comments. Thanks are also due to Rupa Kumar and David Stephenson, as well as to Henri Laurent and Serge Janicot for providing the daily precipitation data.

REFERENCES

  • Barnett, T. P., , L. Dümenil, , U. Schlese, , E. Roeckner, , and M. Latif, 1989: The effect of Eurasian snow cover on regional and global climate variations. J. Atmos. Sci., 46 , 661685.

    • Search Google Scholar
    • Export Citation
  • Bougeault, P., 1985: A simple parameterization of the large-scale effects of cumulus convection. Mon. Wea. Rev., 113 , 21082121.

  • Calvet, J-C., , J. Noilhan, , J-L. Roujean, , P. Bessemoulin, , M. Cabelguenne, , A. Olioso, , and J-P. Wigneron, 1998: Towards interactive vegetation in meteorological models, including physiology and plant growth in a soil–vegetation–atmosphere transfer scheme. Agric. For. Meteor., 92 , 7395.

    • Search Google Scholar
    • Export Citation
  • Charney, J. G., 1975: Dynamics of deserts and drought in the Sahel. Quart. J. Roy. Meteor. Soc., 101 , 193202.

  • Choudhury, B. J., 1993: Reflectivities of selected land surface types at 19 and 37 GHz from SSM/I observations. Remote Sens. Environ., 46 , 117.

    • Search Google Scholar
    • Export Citation
  • Cunnington, W. M., , and P. R. Rowntree, 1986: Simulations of the Saharan atmosphere-dependence on moisture and albedo. Quart. J. Roy. Meteor. Soc., 112 , 971999.

    • Search Google Scholar
    • Export Citation
  • Deardorff, J. W., 1977: A parameterization of the ground surface moisture content for use in atmospheric prediction models. J. Appl. Meteor., 16 , 11821185.

    • Search Google Scholar
    • Export Citation
  • ——,. 1978: Efficient prediction of ground surface temperature and moisture, with inclusion of a layer of vegetation. J. Geophys. Res., 83 , 18891903.

    • Search Google Scholar
    • Export Citation
  • Delworth, T. L., , and S. Manabe, 1988: The influence of potential evaporation on the variabilities of simulated soil wetness and climate. J. Climate, 1 , 523547.

    • Search Google Scholar
    • Export Citation
  • Déqué, M., , C. Dreveton, , A. Braun, , and D. Cariolle, 1994: The ARPEGE/IFS atmosphere model, A contribution to the French community climate modelling. Climate Dyn., 10 , 249266.

    • Search Google Scholar
    • Export Citation
  • —— and Coauthors, 1999: Documentation ARPEGE-CLIMAT. Centre National de Recherches Météorologiques, Météo-France, Toulouse, France.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., 1999: Assessing GCM sensitivity to soil wetness using GSWP data. J. Meteor. Soc. Japan, 77 , 367385.

  • ——, and Shukla, J., 1993: Observational and modeling studies of the influence of soil moisture anomalies on the atmospheric circulation. Predictions of Interannual Climate Variations, J. Shukla, Ed., NATO Series I, Vol. 6, Springer-Verlag, 1–23.

    • Search Google Scholar
    • Export Citation
  • Douville, H., 1998: Validation and sensitivity of the global hydrologic budget in stand-alone simulations with the ISBA land surface scheme. Climate Dyn., 14 , 151171.

    • Search Google Scholar
    • Export Citation
  • ——, and Royer, J-F., 1996: Sensitivity of the Asian summer monsoon to an anomalous Eurasian snow cover within the Météo-France GCM. Climate Dyn., 12 , 441448.

    • Search Google Scholar
    • Export Citation
  • ——, and Chauvin, F., 2000: Relevance of soil moisture for seasonal climate predictions: A preliminary study. Climate Dyn.,16, 719–736.

    • Search Google Scholar
    • Export Citation
  • ——, Royer, J-F., , and J-F. Mahfouf, 1995a: A new snow parameterization for the Météo-France climate model. Part I: Validation in stand-alone experiments. Climate Dyn., 12 , 2135.

    • Search Google Scholar
    • Export Citation
  • ——, ——, and ——,. 1995b: A new snow parameterization for the Météo-France climate model. Part II: Validation in a 3D GCM experiment. Climate Dyn., 12 , 3752.

    • Search Google Scholar
    • Export Citation
  • ——, Bazile, E., , P. Caille, , D. Giard, , J. Noilhan, , L. Peirone, , and F. Taillefer, 1999: Global Soil Wetness Project: Forecast and assimilation experiments performed at Météo-France. J. Meteor. Soc. Japan, 77 , 305316.

    • Search Google Scholar
    • Export Citation
  • ——, Royer, J-F., , D. B. Stephenson, , S. Tyteca, , J. Polcher, , P. Cox, , N. Gedney, , and P. Valdes, 2000: Impact of CO2 doubling on the Asian summer monsoon: Robust versus model-dependent responses. J. Meteor. Soc. Japan,78, 421–439.

    • Search Google Scholar
    • Export Citation
  • Eltahir, E. A. B., , and C. Gong, 1996: Dynamics of wet and dry years in West Africa. J Climate, 9 , 10301042.

  • Fennessy, M. J., , and J. Shukla, 1999: Impact of initial soil wetness on seasonal atmospheric prediction. J. Climate, 12 , 31673180.

  • Ferranti, L., , J. M. Slingo, , T. N. Palmer, , and B. J. Hoskins, 1999: The effects of land-surface feedbacks on the monsoon circulation. Quart. J. Roy. Meteor. Soc., 125 , 15271550.

    • Search Google Scholar
    • Export Citation
  • Folland, C. K., , T. N. Palmer, , and D. E. Parker, 1986: Sahel rainfall and worldwide sea temperatures. Nature, 320 , 602607.

  • Gates, W. L., 1992: AMIP: The Atmospheric Model Intercomparison Project. Bull. Amer. Meteor. Soc., 73 , 19621970.

  • Geleyn, J-F., and Coauthors, 1995: Atmospheric parameterization schemes in Météo-France's Arpege NWP model. Proc. Seminar on Parameterization of Subgrid-Scale Physical Processes, Reading, United Kingdom, ECMWF, 385–402.

    • Search Google Scholar
    • Export Citation
  • International GEWEX Project Office (IGPO), 1998: Global Soil Wetness Project: Preliminary report on the pilot phase. IGPO Publication Series, No. 29, 48 pp.

    • Search Google Scholar
    • Export Citation
  • Jarvis, P. G., 1976: The interpretation of leaf water potential and stomatal conductance found in canopies in the field. Philos. Trans. Roy. Soc. London, 273B , 593610.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., , and M. J. Suarez, 1996: The influence of land surface moisture retention on precipitation statistics. J. Climate, 9 , 25512567.

    • Search Google Scholar
    • Export Citation
  • Legates, D. R., , and C. J. Willmott, 1990: Mean seasonal and spatial variability in gauge-corrected global precipitation. Int. J. Climatol., 10 , 111127.

    • Search Google Scholar
    • Export Citation
  • Mahfouf, J-F., , and J. Noilhan, 1996: Inclusion of gravitational drainage in a land surface scheme based on the force-restore method. J. Appl. Meteor., 35 , 987992.

    • Search Google Scholar
    • Export Citation
  • Manzi, A. O., , and S. Planton, 1994: Implementation of the ISBA parameterization scheme for land surface processes in a GCM—an annual cycle experiment. J. Hydrol., 155 , 353387.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., 1994: Influence of the land surface in the Asian summer monsoon: External conditions versus internal feedbacks. J. Climate, 7 , 10331049.

    • Search Google Scholar
    • Export Citation
  • ——,. 1997: The south Asian monsoon and the tropospheric biennial oscillation. J. Climate, 10 , 19211943.

  • Mintz, Y., , and Y. V. Serafini, 1992: A global monthly climatology of soil moisture and water balance. Climate Dyn., 8 , 1327.

  • Morcrette, J-J., 1990: Impact of changes to the radiation transfer parameterizations plus cloud optical properties in the ECMWF model. Mon. Wea. Rev., 118 , 847873.

    • Search Google Scholar
    • Export Citation
  • Noilhan, J., , and S. Planton, 1989: A simple parameterization of land surface processes for meteorological models. Mon. Wea. Rev., 117 , 536549.

    • Search Google Scholar
    • Export Citation
  • Palmer, T. N., , C. Brankovic, , P. Viterbo, , and M. Miller, 1992: Modeling interannual variations of summer monsoons. J. Climate, 5 , 399417.

    • Search Google Scholar
    • Export Citation
  • Robock, A., , A. C. Schlosser, , K. A. Vinnikov, , N. A. Speranskaya, , J. K. Entin, , and S. Liu, 1998: Evaluation of the AMIP soil moisture simulations. Global Planet. Change, 19 , 181208.

    • Search Google Scholar
    • Export Citation
  • ——, Vinnikov, K. Y., , G. Srinivasan, , J. K. Entin, , S. E. Hollinger, , N. A. Speranskaya, , S. Liu, , and A. Namkhai, 2000: The global soil moisture data bank. Bull. Amer. Meteor. Soc., 81 , 12811299.

    • Search Google Scholar
    • Export Citation
  • Rowell, D. P., , C. K. Folland, , K. Maskell, , and M. N. Ward, 1995: Variability of summer rainfall over tropical North Africa (1906–1992): Observations and modelling. Quart. J. Roy. Meteor. Soc., 121 , 669704.

    • Search Google Scholar
    • Export Citation
  • Schär, C., , D. Lüthi, , U. Beyerle, , and E. Heise, 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12 , 722736.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and and Coauthors, 1997: Modeling the exchanges of energy, water, and carbon between the continents and the atmosphere. Science, 275 , 502509.

    • Search Google Scholar
    • Export Citation
  • Serafini, Y. V., 1990: The time scale of land surface hydrology in response to initial soil moisture anomalies: a case study. Tellus, 42A , 390400.

    • Search Google Scholar
    • Export Citation
  • Shukla, J., , and Y. Mintz, 1982: Influence of land-surface evapotranspiration on the earth's climate. Science, 215 , 14981501.

  • Stephenson, D. B., , K. R. Kumar, , F. J. Doblas-Reyes, , J-F. Royer, , F. Chauvin, , and S. Pezzulli, 1999: Extreme daily rainfall events and their impact on ensemble forecasts of the Indian monsoon. Mon. Wea. Rev., 127 , 19541966.

    • Search Google Scholar
    • Export Citation
  • Sud, Y. C., , and W. E. Smith, 1985: Influence of local land-surface processes on the Indian monsoon: A numerical study. J. Climate Appl. Meteor., 24 , 10151036.

    • Search Google Scholar
    • Export Citation
  • Vinnikov, K. Y., , and I. B. Yeserkepova, 1991: Soil moisture, empirical data and model results. J. Climate, 4 , 6679.

  • Walker, J., , and P. R. Rowntree, 1977: The effect of soil moisture on circulation and rainfall in a tropical model. Quart. J. Roy. Meteor. Soc., 103 , 2946.

    • Search Google Scholar
    • Export Citation
  • Webster, P. J., 1983: Mechanisms of monsoon low-frequency variability: Surface hydrological effects. J. Atmos. Sci., 40 , 21102124.

  • ——, Palmer, T. N., , M. Yanai, , J. Shukla, , V. Magnana, , T. Yasunari, , and R. Tomas, 1998: The monsoon: Processes, predictability and prediction. J. Geophys. Res., 103 , 14 45114 510.

    • Search Google Scholar
    • Export Citation
  • Wilson, M. F., , and A. Henderson-Sellers, 1985: A global archive of land cover and soils data for use in general circulation climate models. J. Climatol., 5 , 119143.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Geographical domains in which SM has been controlled (relaxed or limited); the control is gradually released from the inner boundary to the outer boundary of the domain.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 2.
Fig. 2.

JJAS horizontal wind (m s−1) at 850 hPa over (top) south Asia and (bottom) Africa: (left) ECMWF 15-yr reanalysis, (right) ARPEGE control experiment.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 3.
Fig. 3.

JJAS total precipitation (mm day−1) over (top) south Asia and (bottom) Africa: (left) Legates and Willmott climatology, (right) ARPEGE control experiment.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 4.
Fig. 4.

Histograms of daily JJAS rainfall rates (RR: mm day−1) over India (5°–25°N, 70°–95°E) and west Sudan (10°–15°N, 20°W–20°E): (shaded) histogram of simulated precipitation, (solid line) histogram of observed precipitation (4-yr dataset over India, 23-yr dataset over Sudan)

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 5.
Fig. 5.

JJAS total soil moisture (kg m−2) over (top) south Asia and (bottom) Africa: (left) GSWP climatology (1987–88), (right) ARPEGE control experiment.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 6.
Fig. 6.

Change in JJAS total soil moisture (kg m−2) prescribed over south Asia in WI0, WI1, WI7, and WI8, respectively (dashed lines represent negative values).

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 7.
Fig. 7.

Response of the JJAS 2-m air temperature (°C) over south Asia in WI0, WI1, WI7, and WI8, respectively.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 8.
Fig. 8.

As in Fig. 7 but for JJAS sea level pressure (PMER; hPa).

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 9.
Fig. 9.

As in Fig. 7 but for JJAS horizontal wind (m s−1) at 850 hPa.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 10.
Fig. 10.

As in Fig. 7 but for JJAS total precipitation (mm day−1).

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 11.
Fig. 11.

Change in JJAS total soil moisture (kg m−2) prescribed over Sudan–Sahel in WA0, WA1, WA7, and WA8, respectively.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 12.
Fig. 12.

Response of the JJAS surface air temperature (°C) over south Asia in WA0, WA1, WA7, and WA8, respectively.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 13.
Fig. 13.

As in Fig. 12 but for JJAS horizontal wind (m s−1) at 850 hPa.

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 14.
Fig. 14.

As in Fig. 12 but for total precipitation (mm day−1)

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 15.
Fig. 15.

Histograms of daily JJAS surface evaporation (mm day−1): (shaded) control experiment (WIF), (dashed line) dry experiment (WI0 or WA0), (solid line) wet experiment (WI1 or WA1). (Note change in ordinate scales between some panels.)

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 16.
Fig. 16.

As in Fig. 15 but for daily JJAS precipitation (mm day−1)

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 17.
Fig. 17.

Coefficient of variation of daily JJAS precipitation vs mean JJAS soil moisture (kg m−2) over the same domain: (top) India, (bottom) Sudan–Sahel; for each ensemble, the result of each of the six members is shown (thin symbols) as well as the six-member average (thick symbols)

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Fig. 18.
Fig. 18.

(top) Average JJAS evaporation, (middle) moisture convergence, and (bottom) precipitation over Sudan–Sahel vs mean JJAS soil moisture; for each ensemble, the result of each of the six members is shown (thin symbols) as well as the six-member average (thick symbols)

Citation: Journal of Climate 14, 11; 10.1175/1520-0442(2001)014<2381:IOSMOT>2.0.CO;2

Table 1.

Summary of the experiments with freely evolving, relaxed, or limited SM. All experiments are ensembles of 7-month integrations

Table 1.
Save