• Bellenger, H., , and J. P. Duvel, 2009: An analysis of ocean diurnal warm layers over tropical oceans. J. Climate, 22 , 36293646.

  • Chen, S. S., , and R. A. Houze, 1997: Diurnal variation and life cycle of deep convective systems over the tropical Pacific warm pool. Quart. J. Roy. Meteor. Soc., 123 , 357388.

    • Search Google Scholar
    • Export Citation
  • Crook, A., 1996: Sensitivity of moist convection forced by boundary layer processes to low-level thermodynamic fields. Mon. Wea. Rev., 124 , 17671785.

    • Search Google Scholar
    • Export Citation
  • Duvel, J. P., , H. Bellenger, , P. Xavier, , and F. Doblas-Reyer, 2008: Air-sea processes in the Indian Ocean and the intraseasonal oscillation. Proc. ECMWF Workshop on Atmosphere-Ocean Interaction, Reading, United Kingdom, ECMWF, 83–92.

    • Search Google Scholar
    • Export Citation
  • Fairall, C. W., , E. F. Bradley, , J. S. Godfrey, , G. A. Wick, , J. B. Edson, , and G. S. Young, 1996: Cool-skin and warm-layer effects on sea surface temperature. J. Geophys. Res., 101 , 12951308.

    • Search Google Scholar
    • Export Citation
  • Fairall, C. W., , E. F. Bradley, , J. E. Hare, , A. A. Grachev, , and J. B. Edson, 2003: Bulk parameterization of air–sea fluxes: Updates and verification for the COARE algorithm. J. Climate, 16 , 571591.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., , and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105 , 11711188.

  • Hashiguchi, H., , S. Fukao, , Y. Moritani, , T. Wakayama, , and S. Watanabe, 2004: A lower troposphere radar: 1.3 GHz active phased-array type wind profiler with RASS. J. Meteor. Soc. Japan, 82 , 915931.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., , P. A. Arkin, , and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122 , 22962311.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., , T. M. Rickenbach, , S. A. Rutledge, , P. E. Ciesielski, , and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12 , 23972418.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., , P. E. Ciesielski, , and J. A. Cotturone, 2001: Multiscale variability of the atmospheric mixed layer over the western Pacific warm pool. J. Atmos. Sci., 58 , 27292750.

    • Search Google Scholar
    • Export Citation
  • Kawai, Y., , and A. Wada, 2007: Diurnal sea surface temperature variation and its impact on the atmosphere and ocean: A review. J. Oceanogr., 63 , 721744.

    • Search Google Scholar
    • Export Citation
  • Parsons, D. B., , K. Yoneyama, , and J-L. Redelsperger, 2000: The evolution of the tropical western Pacific atmosphere-ocean system following the arrival of a dry intrusion. Quart. J. Roy. Meteor. Soc., 126 , 517548.

    • Search Google Scholar
    • Export Citation
  • Randall, D. A., , Harshvardhan, , and D. A. Dazlich, 1991: Diurnal variability of the hydrologic cycle in a general circulation model. J. Atmos. Sci., 48 , 4062.

    • Search Google Scholar
    • Export Citation
  • Steiner, M., , R. A. Houze, , and S. E. Yuter, 1995: Climatological characterization of three-dimensional storm structure from operational radar and rain gauge data. J. Appl. Meteor., 34 , 19782007.

    • Search Google Scholar
    • Export Citation
  • Sui, C. H., , K. M. Lau, , Y. N. Takayabu, , and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54 , 639655.

    • Search Google Scholar
    • Export Citation
  • Takayabu, Y., 2002: Spectral representation of rain profiles and diurnal variations observed with TRMM PR over the equatorial area. Geophys. Res. Lett., 29 , 1584. doi:10.1029/2001GL014113.

    • Search Google Scholar
    • Export Citation
  • Yasunaga, K., , M. Fujita, , T. Ushiyama, , K. Yoneyama, , Y. N. Takayabu, , and M. Yoshizaki, 2008: Diurnal variations in precipitable water observed by shipborne GPS over the tropical Indian Ocean. SOLA, 4 , 97100.

    • Search Google Scholar
    • Export Citation
  • Yoneyama, K., , M. Fujita, , N. Sato, , M. Fujiwara, , Y. Inai, , and F. Hasebe, 2008a: Correction for radiation dry bias found in RS92 radiosonde data during the MISMO experiment. SOLA, 4 , 1316.

    • Search Google Scholar
    • Export Citation
  • Yoneyama, K., and Coauthors, 2008b: MISMO field experiment in the equatorial Indian Ocean. Bull. Amer. Meteor. Soc., 89 , 18891903.

  • Zeng, X., , and A. Beljaars, 2005: A prognostic scheme of sea surface skin temperature for modelling and data assimilation. Geophys. Res. Lett., 32 , L14605. doi:10.1029/2005GL023030.

    • Search Google Scholar
    • Export Citation
  • Zhang, G. J., 2002: Convective quasi-equilibrium in midlatitude continental environment and its effect on convective parameterization. J. Geophys. Res., 107 , 4220. doi:10.1029/2001JD001005.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    SST evolution from R/V Mirai, the two m-TRITON buoys, and the corresponding average (black solid line). For m-TRITON, SSTs are computed using the Fairall et al. (2003) algorithm. Days with strong DWL corresponding to average DSA of more than 0.5 K are spotted with thick black lines.

  • View in gallery

    Total cumulative hourly distributions (6 radar time steps) of convective features at 3 km from the R/V Mirai Doppler radar for (a) (left axis) the whole period (thick line) and (right axis) for the 19 days with strong DWL corresponding to DSA of more than 0.5 K (thin solid line) and the 11 days without strong DWL (thin dashed line); and (b) convective features with an echo-top (left axis) higher than 9 km (thick lines) and (right axis) lower than 9 km (thin lines) for days with (solid) and without DWL (dashed).

  • View in gallery

    Average diurnal evolution for days with (solid) and without (dashed) strong DWL of (a) the mixed-layer height (m, see text for details) from R/V Mirai soundings, (b) the 300–800-m height average vertical wind (m s−1) from SB-LTR, (c) R/V Mirai precipitation rate (mm h−1), and the average from R/V Mirai and the two m-TRITON buoys of (d) DSA (K) due to DWL formation, (e) horizontal wind (m s−1), (f) surface latent, and (g) sensible heat fluxes (W m−2). For surface fluxes thick lines represent heat fluxes computed with DWL-induced SST variations and thin lines represent heat fluxes computed with measured bulk SST from which the DWL effect has been removed; during the day SST is simply linearly interpolated.

  • View in gallery

    Average diurnal evolution of (a) CAPE (solid) and CIN (dashed) for days with (crosses) and without strong DWL; (b) as in (a), but for CAPE and CIN generation rate from changes in boundary layer parcel (averaged between 100 and 400 m) characteristics and (c) from changes in environment characteristics. The time series are smoothed over 3 consecutive time steps (one sounding every 3 h).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 1 1 1
PDF Downloads 2 2 2

Role of Diurnal Warm Layers in the Diurnal Cycle of Convection over the Tropical Indian Ocean during MISMO

View More View Less
  • 1 Center for Climate System Research, The University of Tokyo, Kashiwa, Japan
  • 2 Center for Climate System Research, The University of Tokyo, Kashiwa, and Research Institute for Global Change, Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan
  • 3 National Institute for Agro-Environmental Sciences, Tsukuba, Japan
  • 4 Research Institute for Global Change, Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan
© Get Permissions
Full access

Abstract

The role of air–sea interaction in the diurnal variations of convective activity during the suppressed and developing stages of an intraseasonal convective event is analyzed using in situ observations from the Mirai Indian Ocean cruise for the Study of the Madden–Julian oscillation (MJO)-convection Onset (MISMO) experiment. For the whole period, convection shows a clear average diurnal cycle with a primary maximum in the early morning and a secondary one in the afternoon. Episodes of large diurnal sea surface temperature (SST) variations are observed because of diurnal warm layer (DWL) formation. When no DWL is observed, convection exhibits a diurnal cycle characterized by a maximum in the early morning, whereas when DWL forms, convection increases around noon and peaks in the afternoon. Boundary layer processes are found to control the diurnal evolution of convection. In particular, when DWL forms, the change in surface heat fluxes can explain the decrease of convective inhibition and the intensification of the convection during the early afternoon.

* Current affiliation: Laboratoire de Météorologie Dynamique, Institut Pierre-Simon Laplace, Ecole Normale Supérieure, Paris, France

Corresponding author address: Hugo Bellenger, Laboratoire de Météorologie Dynamique, Ecole Normale Supérieure, 24 rue Lhomond, 75005 Paris, France. Email: bellenger@lmd.ens.fr

Abstract

The role of air–sea interaction in the diurnal variations of convective activity during the suppressed and developing stages of an intraseasonal convective event is analyzed using in situ observations from the Mirai Indian Ocean cruise for the Study of the Madden–Julian oscillation (MJO)-convection Onset (MISMO) experiment. For the whole period, convection shows a clear average diurnal cycle with a primary maximum in the early morning and a secondary one in the afternoon. Episodes of large diurnal sea surface temperature (SST) variations are observed because of diurnal warm layer (DWL) formation. When no DWL is observed, convection exhibits a diurnal cycle characterized by a maximum in the early morning, whereas when DWL forms, convection increases around noon and peaks in the afternoon. Boundary layer processes are found to control the diurnal evolution of convection. In particular, when DWL forms, the change in surface heat fluxes can explain the decrease of convective inhibition and the intensification of the convection during the early afternoon.

* Current affiliation: Laboratoire de Météorologie Dynamique, Institut Pierre-Simon Laplace, Ecole Normale Supérieure, Paris, France

Corresponding author address: Hugo Bellenger, Laboratoire de Météorologie Dynamique, Ecole Normale Supérieure, 24 rue Lhomond, 75005 Paris, France. Email: bellenger@lmd.ens.fr

1. Introduction

The diurnal cycle of convection has strong radiative effects and may impact climate variability at longer time scales. It is thus of particular importance to understand the processes that drive the diurnal cycle of convection in order to represent it correctly in the models. Over tropical continents, convection usually peaks in the early evening after being triggered during the afternoon by the enhancement of surface buoyancy flux. On the contrary, convection usually peaks in the early morning over tropical open-oceanic regions (e.g., Takayabu 2002). Many studies focus on this characteristic of the diurnal cycle, proposing different processes to explain it. Gray and Jacobson (1977) propose that it could be due primarily to the day and night variation of the clear-sky tropospheric radiative cooling surrounding organized convection. The clear-sky region radiative cooling is indeed greater during the night when there is no balance from shortwave radiation absorption. This nocturnal increase in radiative cooling results in an increase in the subsidence in the cloud-free region. Finally, the increased subsidence induces an increase in low-level mass and moisture convergence into disturbed regions, which leads to stronger convection. Using numerical models, Randall et al. (1991) show that the stabilization due to the absorption of shortwave radiation by high clouds during the day could explain the convective minimum during the afternoon. Sui et al. (1997) suggest that this diurnal variation of convection could be due to diurnal change in available precipitable water associated to the diurnal variations of air temperature.

However, convection over oceans can also show a secondary maximum in the late afternoon that has received less attention (Janowiak et al. 1994). Using observations from the Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE), it has been suggested that this afternoon maximum would occur during relatively calm periods when convective activity is reduced (e.g., Chen and Houze 1997). Parsons et al. (2000) suggested that the absorption of shortwave radiation by boundary layer moist air could destabilize the atmosphere and explain the increase of convection in the afternoon for days following a dry intrusion. However, air–sea interaction may also act in the same manner and could also explain this convective peak. Periods of reduced convective activity are characterized by relatively calm near-surface conditions. The solar radiation then induces a stable thermal stratification in the upper ocean that is not destroyed by the wind-induced vertical mixing and a diurnal warm layer (DWL) forms. DWLs can correspond to a sea surface temperature (SST) increase of up to several degrees during daylight (see Kawai and Wada 2007 for a review). This daily SST anomaly (DSA) induces a rise in surface heat fluxes of around 50 W m−2 for a DWL of 2 K (Fairall et al. 1996). Using TOGA COARE observations, previous studies suggested that these SST variations might play an important role in triggering convection during the afternoon (Chen and Houze 1997; Sui et al. 1997; Johnson et al. 2001).

The Mirai Indian Ocean cruise for the Study of the Madden–Julian oscillation (MJO)-convection Onset (MISMO) field experiment (Yoneyama et al. 2008b) took place during the developing stage of an intraseasonal convective event in the Indian Ocean. Using these observations, Yasunaga et al. (2008) showed that the surface latent heat flux increase associated with DWL could account for an important part of the diurnal precipitable water diurnal variation. Here, we emphasize the role of DWLs in the diurnal evolution of static stability and in triggering the convection in the afternoon during the developing stage of an intraseasonal oscillation (ISO). In the following sections, we present MISMO data and then depict the diurnal variations of SST and convective activity observed during MISMO. The impact of DWLs on surface fluxes and local stability is then analyzed in order to test the hypothesis that DWLs can account for the observed increase of convection during the afternoon.

2. Data

This study is based on measurements obtained during 30 days (27 October–25 November 2006) of observations from the MISMO campaign (Yoneyama et al. 2008b), which took place in the equatorial Indian Ocean (around 0°, 80°E). We use 10-min mean values of surface meteorology measurements from the research vessel R/V Mirai (0°, 80°E): bulk SST, air temperature, relative humidity, precipitation, horizontal wind, and incoming shortwave and longwave radiation. Hourly mean values of the same parameters (but from incoming longwave radiation) are available from two mini-Triangle Transocean buoy Network (m-TRITON) buoys (0°, 79°–82°E). Bulk SST is measured at 4.5 and 1.5 m under the surface. The corresponding skin SST is then computed using Fairall et al.’s (2003) algorithm. For R/V Mirai, these results are validated using direct skin temperature measurement by an infrared radiometer. R/V Mirai was equipped with the Ship Borne Lower Troposphere Radar (SB-LTR), an L-band (1.3 GHz) wind profiler (Hashiguchi et al. 2004) that provides 10-min mean vertical wind observations. The SB-LTR is very sensitive to raindrops, so rainy echoes have to be discarded. As we are only interested in upper boundary layer vertical winds, we discarded all rainy time steps using R/V Mirai surface precipitation measurements. A C-band (5.3 GHz) Doppler radar was also on board during the MISMO campaign. The Doppler radar retrieved spatiotemporal distribution of precipitation every 10 min within 160 km from the ship with a horizontal spatial resolution of 1 km and a vertical resolution of 500 m. Cloudy areas are detected given a minimum reflectivity of 15 dBZ. Convective features are then captured following Steiner et al. (1995) by detecting either reflectivity higher than ~40 dBZ or any reflectivity local maxima. This analysis is done at 3-km height in order to capture shallow convection as well as deep convective features. We classify the convective features according to their echo-top height into shallow (cumulus and congestus with echo-top height <9 km) and deep (echo-top height >9 km) convective features. Following Johnson et al. (1999), we take this height threshold to define deep convection. Atmospheric soundings were performed 8 times a day from R/V Mirai (every 3 h from 0000 UTC). MISMO soundings were corrected for a daytime dry bias following Yoneyama et al. (2008a). For the sake of clarity of the figures, standard deviations are not reported, though the stated differences between average diurnal maxima and minima are statistically significant at 98% or more when no additional precision is given.

3. Results

a. Diurnal cycle of convection

The first leg of the MISMO campaign took place between two intraseasonal convective events. The beginning of the leg corresponds to the end of the first event around 30 October. Observations were then mainly taken during the suppressed and developing stage of the second event that triggered around 16 November (Yoneyama et al. 2008b). During this period, large DWL-associated DSAs up to 2 K were observed from R/V Mirai and by two m-TRITON buoys (Fig. 1). DWLs can strongly vary on scales of a few kilometers; however, Bellenger and Duvel (2009) show that strong DWLs are often organized on a large scale (a region of more than 1000 km of extension). The buoy network has a typical extent of the grid of a global circulation model (GCM). To consider only the DWLs organized on this spatial scale and to link DWL characteristics to the radar observed convective activity, the average SST computed from R/V Mirai and the two m-TRITON buoys are taken into account to select the days with and without DWL. It is clear that the strongest DWLs observed from R/V Mirai are in fact organized on the scale of the buoy network (Fig. 1). Days are sorted as days with (19 days with an average SST increase of 1 K during the day, see Fig. 3d) and without (11 days) strong DWL according to the average DSA from R/V Mirai and m-TRITON buoys given by an ad hoc threshold of 0.5 K (the following results are not very sensitive to this threshold).

The total distribution of convective features clearly shows two maxima: one in the early morning and a secondary one in the middle of the afternoon (Fig. 2a). This diurnal variation of convection has already been observed: Sui et al. (1997) show for instance that during the convectively disturbed periods observed during TOGA COARE, there is a primary peak of convective rain around 0300–0500 local time (LT) and a secondary one around 1600–1800 LT. This is consistent with what is observed here during the development of an intraseasonal convective event. Here, these two maxima clearly correspond to two distinct cases: days with and days without DWL formation. During relatively calm weather, when a DWL forms, convection intensifies from 1000 to 1500–1700 LT. Then the number of convective features remains relatively high during the night. The convective maximum at 1500 LT corresponds in fact to a maximum in the number of shallow convective features (Fig. 2b). However, from 1000 to 1500 LT both numbers of deep and shallow convective features increase on days with DWL. After 1500–1700 LT the number of shallow convective features decreases. On the other hand, the number of deep convective features continues to increase (but more slowly) between 1400 and 2300 LT. When no DWL forms, convective activity intensifies during the night and peaks in the early morning (Fig. 2a). Shallow convection peaks 2 h before deep convection reaches its maximum around 0500 LT (Fig. 2b). Here, periods with and without DWLs contribute equally to the total hourly distribution of convective features. However, there are roughly twice as many days with DWL (i.e., 19) as without DWL (i.e., 11). The latter indeed correspond to days with a stronger convective activity and are characterized by a relatively higher number of deep convective features (Fig. 2b; see also Yoneyama et al. 2008b, their Fig. 7). While relatively more congestus were observed during days with DWL formation, there were nevertheless a nonnegligible amount of deep convective clouds.

b. Air–sea interaction

The clear distinction between the two diurnal cycles of convection that is depicted in the previous section strongly suggests that air–sea interaction may be important. Figure 3 shows, for both days with (solid) and without (dotted) DWL, the daily evolutions of the averaged (for R/V Mirai and m-TRITON buoys observations) horizontal wind, DSA, and surface heat fluxes deduced from surface meteorological data with (thick lines) and without (thin lines) considering the observed SST diurnal variation in the flux computation. This is done in order to quantify the impact of the DWL in the daily variations of surface heat fluxes. We thus compute these fluxes (i) with the actual skin SST and (ii) with the SST being linearly interpolated during the day between 0600 and 0200 LT. Figure 3 also shows the composite daily evolutions of mixed-layer height, R/V Mirai precipitation rate, and the upper boundary layer vertical wind averaged between 300- and 800-m height from the SB-LTR. The atmospheric mixed-layer height (H) is defined, using specific humidity profiles (smoothed over 50 m), as the height at which the humidity vertical gradient exceeds an ad hoc threshold of −0.06 g kg−1 mb−1. The observations at R/V Mirai position may provide some insights on processes that could explain the diurnal variation of convection as depicted by the radar (Fig. 2). However, because of sampling issues, the R/V Mirai observations presented in Fig. 3 (in particular, precipitation) may not be representative of the whole area that is observed by the radar. This is particularly true for days without DWL that were observed 11 times (cf. the 19 days with DWL).

For days with DWL, DSA increases from 0800 LT and reaches its maximum around 1500 LT (Fig. 3d). The numbers of deep and shallow convective features increase rapidly from 1000 to 1500 LT, then shallow convection decreases and the number of deep convective features increases more slowly (Fig. 2b). During this increase in DSA, the atmospheric mixed layer deepens, reaching a maximum around 1400–1700 LT (Fig. 3a). This suggests that the SST diurnal variations may act to (i) destabilize the atmosphere through surface fluxes, (ii) deepen the atmospheric mixed layer, and (iii) help air parcels to overcome the convective inhibition. This would explain the relatively strong increase in the number of convective features around noon (Figs. 2a,b). The average diurnal variation of SST (Fig. 3d) is indeed associated with a maximum in both surface latent and sensible heat fluxes (Figs. 3f,g, thick lines). Around 1500 LT, the surface latent heat flux increases of about 50%, and the sensible heat flux is roughly multiplied by 4 compared to the values that would be obtained without taking DWL into account.

Around 2000 LT, the number of shallow convective features reaches a local minimum while the number of deep convective features increases toward its maximum (Fig. 2b). Local observations at R/V Mirai position may provide some information on the processes that can explain such an evolution. Around 2000 LT, downdrafts are observed, which cause the average decrease in mixed-layer height (Fig. 3a) and the increase of downward vertical wind in the upper boundary layer (Fig. 3b). These downdrafts are locally associated with an increase of precipitation (Fig. 3c). These convectively induced downdrafts may act to increase and maintain convection during the late afternoon and early evening through the mechanical lifting of boundary layer parcels (see also Johnson et al. 2001). The most effective downdrafts are presumably linked with the deeper convective features. This would then explain how deep convection is maintained during the evening while shallow convection decreases. It is noted that the surface heat fluxes increase after 1500 LT even when the DWL-related SST increase is not taken into account. This is in fact associated with the increase of surface wind (Fig. 3e) that can be partly linked with the convective gusts. However, this increase is only significant at 90% for the latent heat flux.

For days without DWL, convection is strong in the night and morning (Fig. 2a) and does not appear to be linked with surface fluxes variations (Figs. 3f,g). Strong boundary layer subsidence is triggered just after the maximum number of deep convective features is reached (Fig. 2b). This subsidence is still observed during the daytime (Fig. 3b). This boundary layer subsidence is thus roughly out of phase with the convective activity (Fig. 2a). However, it may be mostly linked with local convection (Fig. 3c) and may not be representative of the convective activity observed by the Doppler radar. This daily evolution of vertical wind is, however, associated with an increase of surface horizontal wind (averaged for R/V Mirai and the two buoys, Fig. 3e) during the day that is responsible for the observed diurnal cycle of the surface heat fluxes (Figs. 3f,g).

c. Static stability analysis

To gain insight into the role of DWL in triggering the convection in the early afternoon (Fig. 2), the diurnal evolution of stability is investigated. Following Zhang (2002), convective available potential energy (CAPE), convective inhibition (CIN), and their generation rate from changes in the characteristics of the parcel (here, the average boundary layer) and of its environment (free troposphere) are computed and shown in Fig. 4. The CAPE maxima are comparable for both days with and without DWL. But because of the weaker convection, the CAPE varies less for days with DWL and the resulting daily average CAPE is then higher than for days without DWL (Fig. 4a). The CIN is almost always higher for days without DWL suggesting that processes different from static instability (e.g., low-level convergence) may cause the nighttime maximum in convection. Note that we will here refer to an increase of CIN if its absolute value increases. For days with DWL, CAPE increases and CIN decreases from 1000 to 1500–1700 LT in association with the DWL-induced increase of surface heat flux and in association with the number of convective features (Figs. 2a,b). Then, the CAPE remains high during the night while the CIN increases in association with triggering of downdrafts. On the other hand, for days without DWL, CAPE and CIN evolve in phase opposition with a CAPE (CIN) maximum (minimum) around 0300 LT (0500 LT) before the maximum convection (Fig. 2a) and a minimum (maximum) around 1200–1500 LT.

The CAPE and CIN changes are primarily due to boundary layer parcel moist static energy changes (Figs. 4b,c). Yet, the environmental CAPE increases during the night is the same for both days with strong DWL and for perturbed days without DWL. Note that for days with DWL, the CAPE increase due to environment during the night counterbalances the decrease of CAPE due to boundary layer processes (Fig. 4b). Then CAPE remains high throughout the night and this could explain the maintenance of deep convection during the night (Fig. 2b) as inhibition may be overcome by gusts inducing mechanical lifting of air parcels (Fig. 3b).

For days without DWL the decrease of inhibition and the increase of CAPE due to changes in the boundary layer are not directly linked with the observed diurnal variation in surface fluxes (Figs. 3f,g). CAPE (CIN) increases (decreases) in the evening and decreases (increases) in the early morning. On the other hand, for days with strong DWL, the decrease of inhibition and the increase of CAPE due to changes in boundary layer show different behavior (Fig. 4b): From 0900 to 1500 LT, the CIN decreases and CAPE increases in association with the increase of surface fluxes due to DWL formation (Figs. 3d–f,g). After 1500 LT, CIN rises probably in association with injection of cold and dry air by ΔTp downdrafts. The CIN variations show the role of local destabilization in the intensity of the convective activity. We can hypothesize that the increase of convective activity in the afternoon is due to static destabilization. Then, it is possible to verify whether change in surface fluxes due to DWL might account for the convection increase. Following Crook (1996) we can evaluate the changes of CIN due to changes of temperature and humidity in the boundary layer. The variation of CIN due to variations in the rising parcel temperature (ΔTp) and humidity (Δqp) can then be written as
i1520-0493-138-6-2426-e1
where the saturation specific humidity qs and the temperature T are averaged between the lifting condensation level (lcl) and the level of free convection (lfc), Pi is the pressure at the top of the boundary layer and the overbars denote an average between the two levels, which are indicated below them. We make the assumption that, for days with DWL, the surface heat fluxes mainly act to heat and humidify the boundary layer in the beginning of the afternoon (before downdrafts are observed). We then compute the changes in the parcel temperature and humidity due to surface sensible and latent heat fluxes (SHFsurf and LHFsurf) following:
i1520-0493-138-6-2426-e2
We compute the decrease of CIN due to surface heat fluxes induced changes in temperature and humidity for the average boundary layer air parcel. We find that the decrease in CIN around 1100–1500 LT due to surface fluxes is only of 2.5 J kg−1 h−1 if the fluxes are computed without a DWL-related SST increase. If we now consider the DSA impact on increasing fluxes, the rate of CIN decrease is ~4 J kg−1 h−1. Thus, in the early afternoon, the increase in surface heat fluxes due to DWL formation is of the same order of magnitude as the CIN decrease due to boundary layer processes (~2 J kg−1 h−1). This increase in surface fluxes may thus largely account for the CIN decrease (Fig. 4b). DWL may thus play a role in the triggering of convection in the afternoon (deep and shallow, Fig. 2b). It has to be noted that CIN is sensitive to both surface latent and sensible heat flux. The DWL-related change in sensible heat flux can indeed account for roughly 40% of the CIN decrease in the early afternoon in spite of it being relatively weak compared to the increase of the latent heat flux (Figs. 3f,g).

4. Discussion and conclusions

Large diurnal SST variations were observed two-thirds of the time during MISMO as an intraseasonal convective event developed in the tropical Indian Ocean. These variations were linked to the formation of DWLs that were organized on a spatial scale of a typical GCM grid. The diurnal cycle of convection for the whole period exhibits a primary maximum in the early morning and a secondary one in the afternoon. By separating days with and without formation of DWL, we showed that these two maxima actually correspond to two types of days with a maximum in the afternoon (morning) for days with (without) DWL. For both days with and without DWL, boundary layer processes seem to control the daily evolution of CIN and CAPE. In particular, when DWL forms, the DWL-induced surface heat fluxes increase explains a large part (75% on the average) of the decrease in CIN during the afternoon. As pointed out by Crook (1996), CIN is particularly sensitive to sensible heat flux perturbation. It indeed explains roughly 40% of the CIN decrease related to surface heat fluxes in the early afternoon. Because of this, the DWL-related sensible heat flux increase that is weak (~6 W m−2) compared to the increase in latent heat flux (~30 W m−2) still has to be taken into account to understand the observed diurnal cycle of convection. In the early evening, convective downdrafts are observed that inject cold, dry air into the boundary layer. The CIN then increases but the convection can be maintained through the associated mechanical lift of boundary layer parcels. At the same time, CAPE remains high during the night because of changes in environmental air characteristics probably due to radiative cooling aloft.

Using simple DWL models forced by the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) data between 1979 and 2002, Bellenger and Duvel (2009) showed that, in the tropics, DWLs are often organized on a large scale: a DWL with an equivalent radius larger than 1000 km would occur around 2–3 times a week for a DSA of more than 0.7 K. Moreover, DWLs associated with a DSA of more than 0.7 K can be observed for periods that are characteristic of the intraseasonal time scale (20–90 days). In particular, they show that on the average there are strong DWLs in the Indian Ocean during the suppressed and developing stage of the wintertime ISO. Then the convection that is triggered by DWLs can have different impacts on the ISO and MJO activity: (i) it may induce a diurnal variation in the boundary layer circulation, (ii) it may participate to precondition the atmosphere for further development of convection by moistening the lower troposphere, and (iii) it can also simply trigger an intraseasonal convective event by triggering and organizing the convection on a large scale. Moreover, DWLs change the upper-ocean energy budget through increasing surface heat and longwave radiation fluxes. By triggering convection, DWLs also indirectly induce a decrease of the shortwave radiation that is absorbed by the ocean. DWLs may also play an important role in climate variability from intraseasonal to interannual time scales. Finally, the effect of DWLs should be taken into account in climate models as they appear to be an important part of the interaction between the ocean and the atmosphere. This could be accomplished by using simple DWL parameterizations such as the one proposed by Fairall et al. (1996, see Duvel et al. 2008) or the one developed at the ECMWF (Zeng and Beljaars 2005).

Acknowledgments

H. Bellenger was funded by the Japanese Society for the Promotion of Science and the University of Tokyo. The authors thank the two anonymous reviewers for their useful comments.

REFERENCES

  • Bellenger, H., , and J. P. Duvel, 2009: An analysis of ocean diurnal warm layers over tropical oceans. J. Climate, 22 , 36293646.

  • Chen, S. S., , and R. A. Houze, 1997: Diurnal variation and life cycle of deep convective systems over the tropical Pacific warm pool. Quart. J. Roy. Meteor. Soc., 123 , 357388.

    • Search Google Scholar
    • Export Citation
  • Crook, A., 1996: Sensitivity of moist convection forced by boundary layer processes to low-level thermodynamic fields. Mon. Wea. Rev., 124 , 17671785.

    • Search Google Scholar
    • Export Citation
  • Duvel, J. P., , H. Bellenger, , P. Xavier, , and F. Doblas-Reyer, 2008: Air-sea processes in the Indian Ocean and the intraseasonal oscillation. Proc. ECMWF Workshop on Atmosphere-Ocean Interaction, Reading, United Kingdom, ECMWF, 83–92.

    • Search Google Scholar
    • Export Citation
  • Fairall, C. W., , E. F. Bradley, , J. S. Godfrey, , G. A. Wick, , J. B. Edson, , and G. S. Young, 1996: Cool-skin and warm-layer effects on sea surface temperature. J. Geophys. Res., 101 , 12951308.

    • Search Google Scholar
    • Export Citation
  • Fairall, C. W., , E. F. Bradley, , J. E. Hare, , A. A. Grachev, , and J. B. Edson, 2003: Bulk parameterization of air–sea fluxes: Updates and verification for the COARE algorithm. J. Climate, 16 , 571591.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., , and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105 , 11711188.

  • Hashiguchi, H., , S. Fukao, , Y. Moritani, , T. Wakayama, , and S. Watanabe, 2004: A lower troposphere radar: 1.3 GHz active phased-array type wind profiler with RASS. J. Meteor. Soc. Japan, 82 , 915931.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., , P. A. Arkin, , and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122 , 22962311.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., , T. M. Rickenbach, , S. A. Rutledge, , P. E. Ciesielski, , and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12 , 23972418.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., , P. E. Ciesielski, , and J. A. Cotturone, 2001: Multiscale variability of the atmospheric mixed layer over the western Pacific warm pool. J. Atmos. Sci., 58 , 27292750.

    • Search Google Scholar
    • Export Citation
  • Kawai, Y., , and A. Wada, 2007: Diurnal sea surface temperature variation and its impact on the atmosphere and ocean: A review. J. Oceanogr., 63 , 721744.

    • Search Google Scholar
    • Export Citation
  • Parsons, D. B., , K. Yoneyama, , and J-L. Redelsperger, 2000: The evolution of the tropical western Pacific atmosphere-ocean system following the arrival of a dry intrusion. Quart. J. Roy. Meteor. Soc., 126 , 517548.

    • Search Google Scholar
    • Export Citation
  • Randall, D. A., , Harshvardhan, , and D. A. Dazlich, 1991: Diurnal variability of the hydrologic cycle in a general circulation model. J. Atmos. Sci., 48 , 4062.

    • Search Google Scholar
    • Export Citation
  • Steiner, M., , R. A. Houze, , and S. E. Yuter, 1995: Climatological characterization of three-dimensional storm structure from operational radar and rain gauge data. J. Appl. Meteor., 34 , 19782007.

    • Search Google Scholar
    • Export Citation
  • Sui, C. H., , K. M. Lau, , Y. N. Takayabu, , and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54 , 639655.

    • Search Google Scholar
    • Export Citation
  • Takayabu, Y., 2002: Spectral representation of rain profiles and diurnal variations observed with TRMM PR over the equatorial area. Geophys. Res. Lett., 29 , 1584. doi:10.1029/2001GL014113.

    • Search Google Scholar
    • Export Citation
  • Yasunaga, K., , M. Fujita, , T. Ushiyama, , K. Yoneyama, , Y. N. Takayabu, , and M. Yoshizaki, 2008: Diurnal variations in precipitable water observed by shipborne GPS over the tropical Indian Ocean. SOLA, 4 , 97100.

    • Search Google Scholar
    • Export Citation
  • Yoneyama, K., , M. Fujita, , N. Sato, , M. Fujiwara, , Y. Inai, , and F. Hasebe, 2008a: Correction for radiation dry bias found in RS92 radiosonde data during the MISMO experiment. SOLA, 4 , 1316.

    • Search Google Scholar
    • Export Citation
  • Yoneyama, K., and Coauthors, 2008b: MISMO field experiment in the equatorial Indian Ocean. Bull. Amer. Meteor. Soc., 89 , 18891903.

  • Zeng, X., , and A. Beljaars, 2005: A prognostic scheme of sea surface skin temperature for modelling and data assimilation. Geophys. Res. Lett., 32 , L14605. doi:10.1029/2005GL023030.

    • Search Google Scholar
    • Export Citation
  • Zhang, G. J., 2002: Convective quasi-equilibrium in midlatitude continental environment and its effect on convective parameterization. J. Geophys. Res., 107 , 4220. doi:10.1029/2001JD001005.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

SST evolution from R/V Mirai, the two m-TRITON buoys, and the corresponding average (black solid line). For m-TRITON, SSTs are computed using the Fairall et al. (2003) algorithm. Days with strong DWL corresponding to average DSA of more than 0.5 K are spotted with thick black lines.

Citation: Monthly Weather Review 138, 6; 10.1175/2010MWR3249.1

Fig. 2.
Fig. 2.

Total cumulative hourly distributions (6 radar time steps) of convective features at 3 km from the R/V Mirai Doppler radar for (a) (left axis) the whole period (thick line) and (right axis) for the 19 days with strong DWL corresponding to DSA of more than 0.5 K (thin solid line) and the 11 days without strong DWL (thin dashed line); and (b) convective features with an echo-top (left axis) higher than 9 km (thick lines) and (right axis) lower than 9 km (thin lines) for days with (solid) and without DWL (dashed).

Citation: Monthly Weather Review 138, 6; 10.1175/2010MWR3249.1

Fig. 3.
Fig. 3.

Average diurnal evolution for days with (solid) and without (dashed) strong DWL of (a) the mixed-layer height (m, see text for details) from R/V Mirai soundings, (b) the 300–800-m height average vertical wind (m s−1) from SB-LTR, (c) R/V Mirai precipitation rate (mm h−1), and the average from R/V Mirai and the two m-TRITON buoys of (d) DSA (K) due to DWL formation, (e) horizontal wind (m s−1), (f) surface latent, and (g) sensible heat fluxes (W m−2). For surface fluxes thick lines represent heat fluxes computed with DWL-induced SST variations and thin lines represent heat fluxes computed with measured bulk SST from which the DWL effect has been removed; during the day SST is simply linearly interpolated.

Citation: Monthly Weather Review 138, 6; 10.1175/2010MWR3249.1

Fig. 4.
Fig. 4.

Average diurnal evolution of (a) CAPE (solid) and CIN (dashed) for days with (crosses) and without strong DWL; (b) as in (a), but for CAPE and CIN generation rate from changes in boundary layer parcel (averaged between 100 and 400 m) characteristics and (c) from changes in environment characteristics. The time series are smoothed over 3 consecutive time steps (one sounding every 3 h).

Citation: Monthly Weather Review 138, 6; 10.1175/2010MWR3249.1

Save