NERC Centre for Global Atmospheric Modelling Department of

NERC Centre for Global Atmospheric Modelling Department of Meteorology, University of Reading Scale Interactions on Diurnal to Seasonal Timescales: Their Relevance to Seasonal Model Systematic Error Julia Slingo, Peter Inness, Richard Neale, Steve Woolnough and Gui-Ying Yang

CGAM Tropical Group TOWARDS TROPICAL CLIMATE PREDICTION Teleconnection ge Diabatic Heating Response rid Translation Global Circulation Anomalies Scale Interactions e. g. MJO Lagged Ocean/Land Response Primary route Secondary route Atm os ph eri c. B SST Variability e. g. El Nino Final Impact on Statistics of Local Weather Applications e. g. Crop Models

‘THE TALE OF TWO ERRORS!’ – – The Maritime Continent The Madden-Julian Oscillation

TOOLS • Integrations of the Met Office Unified Model -Had. AM 3 AMIP II (observed SST, 1979 -95) -Aquaplanet version of Had. AM 3 -Had. CM 3 Control • CMAP Precipitation data • High resolution (0. 50, 3 hourly) window brightness temperature data from the EU Cloud Archive User Service (CLAUS)

Typical window brightness (K) image from the CLAUS dataset 12 z 1 January 1992

Annual Mean Precipitation (CMAP) and AMIP II Mean Model Errors Climatology NCAR ECMWF NCEP JMA Met Office

Topography of the Maritime Continent

Resolving the Maritime Continent in GCMs

Annual Mean Precipitation Errors in Had. AM 3: Sensitivity to Horizontal Resolution

Had. AM 3 Sensitivity Experiments: Impact of removing the islands of the Maritime Continent (Neale and Slingo, 2001: Submitted to J. Clim. ) • Land grid-points removed and replaced by ocean gridpoints. • Increased moisture availability from the sea surface leads to enhanced convection and partial correction of the model dry bias. • Note also corrections to model’s wet bias in adjacent areas.

Global Impacts of Improved Maritime Continent Heat Source DJF: 500 h. Pa height (m) and Surface Temperature (K) • Potential improvements in the Maritime Continent heat source can have significant remote effects. • Related to the generation of Rossby waves by the enhanced divergent outflow from the Maritime Continent heat source. • Substantially reduces model systematic error over the extratropics of the winter hemisphere. • Emphasizes the importance of considering the global context of model systematic error in which biases in the tropics may be a key factor.

The Diurnal Cycle in the Tropics (Yang and Slingo, 2001: MWR, 129, 784 -801) Amplitude (K) of the diurnal harmonic DJF JJA

Phase of diurnal harmonic: Local time of maximum brightness temperature DJF JJA

Phase of diurnal cycle showing systematic propagation of convective signal away from the coast Bay of Bengal, JJA: Implied propagation speed ~15 -20 ms-1 ? Deep gravity wave Mexico, JJA: Implied propagation speed ~ 10 ms-1 ? Shallower gravity wave associated with land/sea breeze

Rapidly propagating squall lines down Bay of Bengal as observed in JASMINE (Webster et al. 2001)

Maritime Continent, DJF: Evidence of complex land/sea breezes which organize convection for several 100 km Are sub-gridscale land/sea breezes a crucial component of the energy and hydrological budgets of the Maritime Continent?

Schematic of a sea breeze Sea breeze has two major impacts: • Convergence along sea breeze front provides additional convective mass flux • Winds associated with land sea breezes enhance surface fluxes leading to increased moisture supply

Investigating Maritime Continent sea breezes using a mesoscale model (MM 5) Embedded MM 5 simulations with Kain-Fritsch convection scheme. Morning versus evening precipitation differences show signal over ocean, indicative of land -sea breezes. System of propagating land-sea breezes evident in model. Precipitation is generated over the ocean during the early morning by the convergence initiated by the land breeze. Evidence that orographic effects enhance the land breeze.

CONCLUSION I • Maritime Continent heat source is a key component of the global climate. Improvements in its simulation may have significant impacts on remote systematic errors. • Specifically, land/sea breezes may be a crucial part of the energy and hydrological budgets of coastal regions and especially around large island complexes. • In general, horizontally propagating gravity waves, generated by convection, may be important for organizing convection on larger scales?

Typical window brightness (K) images showing scales of convective organization Note tendency for cloud clusters to congregate together to form super-clusters with multi-day life cycles e. g. Madden Julian Oscillation Self organization

Temporal behaviour of convection around the equator from window brightness temperature for Jan. -Feb. 1992 Note evidence of coherent propagation.

Space-time spectra showing the organization of convection in association with theoretical equatorial waves. Anti-symmetric Symmetric Inertio-gravity Mixed Rossby-gravity Inertio-gravity Kelvin Rossby MJO From Wheeler and Kiladis 1999: J. Atmos. Sci.

Space-time spectra from R 30 version of GFDL model Note lack of organization, an error common to many GCMs. Lack of self-organization mechanism?

Spectra of the zonal and meridional wind in the upper troposphere. Data points show actual observations from commercial aircraft flights. Solid curve is for the N 270 L 40 SKYHI model along the 45°N latitude circle and at 211 h. Pa, monthly averaged for a single July. For clarity the results for the meridional wind have been shifted one decade to the right. From Koshyk and Hamilton 2001: J. Atmos. Sci.

July mean spectra as a function of total horizontal wave-number of (a) the total KE spectrum, (b) the rotational part of the total KE spectrum, (c) the divergent part of the total KE spectrum. Presence of strong divergent component at meso-scales consistent with presence of resolved gravity waves? From Koshyk and Hamilton 2001: J. Atmos. Sci.

GOES Visible Images for 30 September 2001 20: 00 UTC 22: 45 UTC 21: 30 UTC 23: 45 UTC

Florida sea breezes and mesoscale organised convection

Super-parametrizations Results from T 21 simulation with an embedded 2 -D CRM (1 km resolution) in place of convective parametrization. Standard AGCM + CRM Note that the cloud-resolving models from neighbouring columns interact only through the large-scale dynamics. Therefore limits the propagation of gravity waves from one GCM column to another but does allow gravity waves to organise convection within the GCM column. Note dramatic improvement in MJO (upper panels) and synoptic waves (lower panels). Courtesy: David Randall, CSU

Why the MJO is important • Intimately related to active/break cycles of the Australian and Asian Monsoons • Offers potential to provide extended predictability up to 15 -20 days in tropics • Affects weather over the western US and possibly elsewhere • Associated westerly wind events generate ocean Kelvin waves which may significantly modify the evolution and amplitude of El Nino (e. g. 1997) • Large interannual variability in the activity of the MJO has implications for the predictability of the coupled ocean-atmosphere system

Sensitivity of the MJO to AGCM vertical resolution (Inness et al. , 2001: Clim. Dyn. , 17, 777 -793. )

Model Levels: L 19 vs. L 30 Note additional levels in free troposphere

Exploring sensitivity of convective organization to vertical resolution • A water-covered or “aqua-planet” version of the UM is used to investigate the behaviour of tropical convection when the vertical resolution is doubled. • Aqua-planet version of the UM chosen because: - Homogeneity of the model allows us to obtain a large sample of convective events over warm SSTs - Removal of the land areas excludes circulations forced by land-sea contrasts. Convective events in different geographical locations are subject to the same large scale forcings, giving a cleaner comparison. - Aqua-planet provides a more realistic test of the convection scheme than using a single-column model with idealized boundary and large scale forcing functions. • Aqua-planet setup: - Zonally symmetric SST distribution, typical of equatorial Indian Ocean/West Pacific warm pool values. - Incoming solar radiation fixed at zonally symmetric, equinoctial (March) values. - Aqua-planet model integrated for 15 months with both 19 and 30 levels in the vertical. First 3 months of each integration were discarded.

Time-height evolution of convective cloud over 3 x 3 model grid boxes (7. 50 lat. x 11. 250 long. ) centred on the equator L 30 L 19

Time-height sections of specific humidity increment (g/kg/day) by the convection scheme L 30 L 19 Note periods of moistening in the L 30 case – convection is generally a moisture sink.

TOGA-COARE IFA apparent heat source (Q 1) and moisture sink (Q 2) for suppressed (A) and active (B) periods (Lin and Johnson, 1996: J. Atmos. Sci. , 53, 3367 -3383) A: Suppressed B: Active Q 1 Q 2 Note periods of moistening (negative Q 2) during suppressed period

Time-height sections of potential temperature lapse rate L 30 Observations from TOGA-COARE (Johnson et al. 1999, J. Clim. ) L 19 Note presence of stronger stable layer between 600 and 400 h. Pa in L 30, and similarity with observations.

Inferences from aqua-planet model results • When the vertical resolution is increased, the spectrum of tropical cloud types changes from a bimodal to a tri-modal distribution with a third peak in mid-troposphere near the melting level. Associated with periods when these mid-level congestus clouds are dominant, the detrainment from these clouds significantly moistens the midtroposphere. • The appearance of these congestus clouds is shown to be partly due to improved resolution of the freezing level and the convective processes occurring at this level. • The resulting cloud distribution more closely resembles observations, particularly during the suppressed phase of the MJO when cumulus congestus is the dominant cloud type. • The moistening of the free troposphere by cumulus congestus clouds acts to precondition the atmosphere for deep convection. This preconditioning may set the timescale for the next active phase of the MJO and thus influence the intraseasonal organization of convection.

Trimodal distribution of convection and cumulus congestus TOGA COARE results emphasize the dominance of cumulus congestus and point to a TRIMODAL cloud distribution in which the freezing level inversion is the key Many conceptual models of tropical convection are based on a BIMODALcloud distribution, emphasizing shallow “tradewind” or boundary layer cumuli and deep cumulonimbi. From Johnson et al. 1999, J. Clim.

CONCLUSION II • Vertical resolution in the free troposphere must be adequate to resolve the formation of the freezing level inversion and the cooling associated with melting precipitation • Convective parametrizations need to represent a TRIMODAL rather than bimodal cloud distribution.

Coupling with the upper ocean: Bringing together the diurnal cycle and the MJO TOGA-COARE buoy data showed pronounced diurnal variations in skin temperature in excess of 1 K are evident, as well as slower variations related to the MJO. Note that the diurnal variations occur only during break (B) periods. Active (A) periods are preceded by a warming on sub-seasonal timescales. (From Anderson et al. 1996, J. Clim. ) B Nov. A B March

Cumulus congestus and the diurnal cycle TOGA-COARE observations also suggest that cumulus congestus clouds are most prevalent during light wind conditions in the presence of a strong diurnal cycle in SST. Further, these cloud occur most frequently in the late afternoon suggesting that they are triggered by the diurnal cycle in SST. Coupling with the upper ocean is important on diurnal timescales

The MJO and coupling with the ocean: Observations (Woolnough et al. , 2000: J. Clim. , 13, 2086 -2104) Observations show a coherent relationship between convection and SST. Warm SSTs precede convection by 510 days and are the result of weaker winds, reduced LH flux and increased SW flux during suppressed phases of the MJO.

The MJO and coupling with the ocean: Modelling (Inness, personal communication) CGCM has a propagating convective signal compared with standing oscillation in AGCM. Coherent variations in SST in CGCM Coupling with the upper ocean is important for the MJO

BUT intraseasonal SST variations in CGCMs are too small and the MJO signal is still weak: Is the representation of the upper ocean adequate? Schematic showing formation of salt barrier layer Large freshwater flux sets up a salt stratified barrier layer so that a shallow mixed layer forms which can respond rapidly to flux variations, such as the diurnal cycle in solar radiation. The presence of this barrier layer can potentially provide much stronger local coupling in the warm pool region than is currently found in coupled models which do not resolve the detailed structure of the warm pool upper ocean. (From Anderson et al. , 1996: J. Clim)

Temperature cross -section from the TOGA-COARE WHOI mooring Note complex temperature structure in top 40 meters during periods of light winds, associated with suppressed phase of the MJO and a strong diurnal cycle.

CONCLUSION III • Good evidence that MJO and diurnal cycle of cumulus congestus involve coupling with the upper ocean • To simulate diurnal and intraseasonal variations in SST requires detailed representation of salinity and temperature gradients in the mixed layer • Need to consider an upper ocean/atmosphere system in which the structure of the upper ocean is adequately resolved

Probability distribution functions (PDF) of monthly mean SST and precipitation over the tropical Pacific: DJF (upper panels), MAM (lower panels) CMAP Had. AM 3 -CMAP Note tendency for Had. AM 3 to overestimate precipitation over warm SSTs. PDF is also too tight, following closely the exponential relationship implied by the Clausius-Clapeyron equation for saturated vapour pressure.

Summary of Key Issues (1) • Maritime Continent heat source is a major driver of the global circulation yet is poorly represented in GCMs. This complex system of islands gives rise to extensive sea/land breeze circulations which may critically influence the energy budget and hydrological cycle of the region. • Moistening of the free troposphere by cumulus congestus clouds, which form during the suppressed phase of the MJO, may be crucial for convective preconditioning. This dominant cloud type is not represented in models which generally fail to capture the observed tri-modal distribution of convection.

Summary of Key Issues (2) • Diurnal cycle in SST is large during suppressed or light wind conditions in the tropics and may be a trigger for cumulus congestus. It may therefore be a crucial part of the energy budget and hydrological cycle of the Warm Pool • SSTs vary coherently with the MJO in such a manner as to suggest that they are an important component of the eastward propagation and timescale of the MJO. • Both the diurnal and intraseasonal variations in SST involve detailed changes in the salinity and temperature structure of the mixed layer which cannot be adequately represented in current coupled models.