the maritime continent and its role in the global climate: a gcm...

834 VOLUME 16J O U R N A L O F C L I M A T E

q 2003 American Meteorological Society

The Maritime Continent and Its Role in the Global Climate: A GCM Study

RICHARD NEALE* AND JULIA SLINGO

Centre for Global Atmospheric Modelling, Department of Meteorology, University of Reading, Reading, United Kingdom

(Manuscript received 16 July 2001, in final form 5 September 2002)

ABSTRACT

The Maritime Continent, with its complex system of islands and shallow seas, presents a major challenge tomodels, which tend to systematically underestimate the precipitation in this region. Experiments with a climateversion of the Met Office model (HadAM3) show that even with a threefold increase in horizontal resolutionthere is no improvement in the dry bias. It is argued that the diurnal cycle over the islands and the complexcirculation patterns generated by land–sea contrasts are crucial for the energy and hydrological cycles of theMaritime Continent and for determining the mean climate. It is shown that the model has substantial errors inits simulation of the diurnal cycle over the islands, which can rectify onto the seasonal mean climate.

It is further argued that deficient rainfall over the Maritime Continent could be a driver for other systematicerrors, such as the excess precipitation over the western Indian Ocean. To demonstrate the sensitivity of globalsystematic model errors to the heating in this region, two experiments have been performed, one with the existingdistribution of islands and a second where the island grid points are replaced by ocean grid points. In the absenceof the islands of the Maritime Continent, the local precipitation increases by 15%, reducing the existing drybias and bringing the model closer to observations. In response to this improved heating distribution, precipitationdecreases over the west Indian Ocean and South Pacific convergence zone, reducing the systematic wet bias inthese regions. This supports the hypothesis that tropical systematic errors are often related through vertical(Walker) circulations.

The extratropical response to changes in the Maritime Continent heat source is also well demonstrated bythese experiments. The enhanced heating and, hence, divergent outflow generates Rossby waves, which have asignificant impact on the winter circulation and surface temperatures across much of North America and thenortheast Eurasian region. These changes are such as to substantially reduce model systematic error in theseregions. These results reinforce the critical role played by the Maritime Continent in the global circulation. Itemphasizes the need for better representation of convective organization over regions of complex land–seaterrains and the importance of considering the global context of model systematic errors in which biases in theTropics may be a key factor.

1. Introduction

The tropical Maritime Continent has a unique envi-ronment where convective activity responds to forcingon many timescales and space scales, the net result ofwhich is able to influence climate on the global scale.This region, so named because of the complex distri-bution of several large islands with elevated orographyoccupying the domain 108S–208N and 908–1508E, alsoencompasses some of the warmest ocean temperaturesof the world and is known as the ‘‘boiler box’’ of theTropics (Ramage 1968). The Maritime Continent re-ceives much of its rainfall from convective activity as-

* Current affiliation: NOAA–CIRES Climate Diagnostics Center,Boulder, Colorado.

Corresponding author address: Dr. Richard Neale, NOAA–CIRESClimate Diagnostics Center, R/CDC1, 325 Broadway, Boulder, CO80305-3328.E-mail: [email protected]

sociated with localized thunderstorms. The Island Thun-derstorm Experiment (ITEX; Keenan et al. 1989) andthe Maritime Continent Thunderstorm Experiment(MCTEX; Keenan et al. 2000) were established in orderto diagnose and model the importance of the thunder-storm-scale convection for the climate of the region.

The islands play an important part in the meteorologyof the Maritime Continent. Larger-scale organization ofthunderstorm activity is strongly influenced by the orog-raphy of the region as well as the sea-breeze circula-tions, which show strong diurnal variations. Satelliteimages show this diurnal variation over the islands tobe a striking feature of this region (Holland and Keenan1980). The nature of these sea-breeze circulations hasbeen studied extensively. The convection related to sea-breeze convergence is able to aggregate into mesoscaleconvective complexes (MCCs) later in the day, whichmove offland to give the greatest precipitation duringthe local morning time over the oceans (Williams andHouze 1987).

The timing and magnitude of this convective activity

1 MARCH 2003 835N E A L E A N D S L I N G O

on diurnal timescales is known to vary significantly be-tween land and ocean regions. Using a climatology ofwindow brightness derived from high temporally sam-pled satellite data, Yang and Slingo (2001) show thatthe surface inhomogeneity of the Maritime Continentmakes for complex diurnal forcing of convection. Dur-ing Northern Hemisphere winter, in particular, the di-urnal amplitude of convective rainfall over the islandscan be three times as great as that over the adjacentocean. This is in response to the smaller thermal heatcapacity of the land surface leading to large diurnalvariations in low-level instability. However, there arefurther complications to the response, since the strongvariability over the islands is able to propagate out overthe oceans as gravity waves leading to coherent varia-tions in the phase of the convective peak. Much of thisvariability may be related to the standard lifetime ofconvective systems that start over land, build, then moveaway over the oceans (Saito et al. 2001). Such an evo-lution suggests that the atmosphere has some memoryof a particular convective disturbance at some previoustime.

The Maritime Continent also exhibits an influence onlarger-scale intraseasonal activity. The Madden–Julianoscillation (MJO) propagates over the region in its ma-ture phase and is modulated by the presence of the un-derlying surface properties and the diurnal cycle beforemoving out over the West Pacific. This makes for acomplex response whereby the diurnal cycle of con-vective activity becomes suppressed during the activephase of the MJO and enhanced during the break phase(Sui and Lau 1992). The presence of the Maritime Con-tinent also modulates the strength and phase speed ofthe MJO. In particular, it is able to both weaken andsplit the active phase of the oscillation before it rein-tensifies in the South Pacific convergence zone (SPCZ;Zhu and Wang 1993).

The southern part of the Maritime Continent regionis strongly influenced by the winter monsoon circulationinvolving significant transport of warm moist air fromnorth of the equator that meets the trade wind flow southof the equator to generate the austral monsoon trough.The first strong MJO of the winter season usually sig-nifies the onset of the monsoon regime (Sui and Lau1992). The winter monsoon associated with convectionover the Maritime Continent has also been shown toinfluence the midlatitude circulation through short-termteleconnections (Lau et al. 1983). For example, duringan active phase of the monsoon the tropical and extra-tropical circulations vary in a coherent way, intensifyingthe local Hadley and Walker circulations and strength-ening the east Asia subtropical jet (Chang and Lau1982). The action of cold surges, which strengthen theAsian winter monsoon northeasterlies, also results inperiods of enhanced convective activity throughout theseason, particularly north of Borneo (Houze et al. 1981).These cold surges can typically last from 5 to 14 days(Zhang et al. 1997) and so may account for a large

proportion of climate variability over the Maritime Con-tinent during winter.

Variability on seasonal timescales modulates the rain-fall over the Maritime Continent particularly during thewarm phase of ENSO (Philander 1985). As the warmestSSTs move out into the central Pacific, the strongestconvection follows and generates an anomalous longi-tudinal circulation, leading to suppression over the Mar-itime Continent. This is a particularly large effect sinceENSO events typically reach a maximum in NorthernHemisphere winter when the precipitation totals usuallyreach a maximum over the Maritime Continent.

On the planetary scale, convection over the MaritimeContinent represents a dominant heat source for the at-mospheric circulation. Convection is most intense andthe tropopause at its highest over this region. Upper-tropospheric divergent outflow from this convective re-gion has been shown, in an idealized model, to be amajor source of wave activity due to the generation ofglobal rotational flow (Sardeshmukh and Hoskins 1988).Such a response establishes the stationary wave patternsclearly observed in time-mean fields. So it is clear thatthe Maritime Continent has an important role to playboth in the variability of the tropical climate and for theglobal circulation as a whole.

The aim of this paper is to gain an insight into theclimate of the Maritime Continent region and to high-light its role in the tropical and global mean climate.The following section assesses the skill of the Met Of-fice Unified Model (UM) at reproducing the mean cli-mate over the Maritime Continent region and identifiessome possible shortcomings of the model. Specific er-rors relating to the presence of the Maritime Continentislands are identified and results from idealized sensi-tivity experiments, which attempt to remove these er-rors, are described. Further shortcomings of the UM’srepresentation of the diurnal cycle are then highlighted.Investigation of a possible sensitivity to the details ofthe UM’s radiation scheme is performed with a furthershort 1-yr integration of the model. Significant large-scale influences of changes in the Maritime Continentheat source region are revealed in a section on the globalimpacts; the implications of this study and conclusionsare summarized in the final section.

2. Description of the model and its systematicerrors

A suite of experiments using the Met Office UnifiedModel (UM), performed as part of the second phase ofthe Atmospheric Model Intercomparison Project (AMIPII; Gates 1992), has been analyzed and the performanceof the UM assessed over our region of interest, the Mar-itime Continent. The experimental period extends over17 years (1979–95) and the model is forced by monthlymean SSTs and sea ice. The climate version of the UM(HadAM3; Pope et al. 2000) is used at a resolution of2.58 in latitude, 3.758 in longitude, and 19 levels in the


FIG. 1. Seasonal mean errors from a six-member standard AMIP II ensemble of the UM when compared with the Xie–Arkin climatology(precipitation) and ERA (winds): (a) DJF precipitation (mm day21); (b) as in (a) but for JJA; (c) DJF 850-hPa wind vectors (m s21); (d) asin (c) but for JJA. Precipitation values less the 22 mm day21 are shaded. The zero contour is omitted.

vertical. This model version has significant changescompared to earlier versions, namely, the inclusion ofmomentum transports by subgrid-scale convective pro-cesses (Kershaw and Gregory 1997), an updated radi-ation scheme (Edwards and Slingo 1996), and a newland surface and vegetation scheme [Met Office SurfaceExchange Scheme (MOSES); Cox et al. 1999]. For com-parison, observational data are provided by the Euro-pean Centre for Medium-Range Weather Forecasts(ECMWF) Re-Analysis (ERA; Gibson et al. 1997) andthe Climate Prediction Center Merged Analysis of Pre-cipitation (CMAP) rainfall climatology (Xie and Arkin1996) covering the same analysis period as the AMIPII experiments.

In the UM, as with many climate models, the Mari-time Continent exhibits a deficit of precipitation in allseasons (see the AMIP II Web site online at http://www-pcmdi.llnl.gov/amip). Figure 1 shows the precipitationand lower-tropospheric wind errors in the model duringNorthern Hemisphere winter [December–January–Feb-ruary (DJF)] and summer [June–July–August (JJA)].During DJF precipitation deficits in excess of 2 mmday21 cover most of the Maritime Continent north ofthe equator and the western part south of the equator.Only to the east and west of Guinea and just north ofthe Australian coast is there a significant excess of pre-cipitation in the model.

In JJA the errors in precipitation appear less coherent,but in general the same pattern as in DJF occurs with adry bias centered mostly off the coasts of the large is-lands. The pattern of errors over the Asian monsoon re-

gion show that the rainfall is overestimated across north-ern and central India as well as into the far southeast ofAsia. In addition, the equatorial maximum over the IndianOcean, which characterizes the break phase of the mon-soon, is positioned too far to the west. Due to the presenceof large-scale tropical vertical circulations it is possiblethat shortcomings over the Maritime Continent may havesome role to play in the model’s simulation of the mon-soon during JJA also. Certain very local errors persistthroughout the whole year. For instance, the negativeerrors over the Philippines and off the southwest coastof Sumatra suggest that some of the problems may liein the model’s response to the presence of these islands.

Associated with the errors in the seasonal mean pre-cipitation are errors in the lower-tropospheric flow. InDJF (Fig. 1c), the zonal flow has a strong easterly biasfrom the West Pacific through to the Indian Ocean northof the equator. It is conceivable that the large-scale na-ture of this error is related to the lack of a strong or-ganized MJO, which would have periods of significantwesterly wind activity projecting onto the mean basicstate. Indeed the AMIP I study by Slingo et al. (1996)showed that a poor model MJO and errors in the Mar-itime Continent basic state could be closely linked. Em-bedded within the large-scale flow bias are errors on thescale of the Maritime Continent islands. In general, theeasterly bias tends to be increased on the west side ofthe islands and points toward problems in simulatingthe local island-scale circulation associated with sea-breeze effects and the diurnal cycle. The problems withsimulating the coupling between the diurnal cycle and


FIG. 2. Annual mean precipitation error (mm day21) from four UM AMIP II experiments with different horizontal resolution: (a) climateresolution (2.58 3 3.758); (b) 1.5 3 climate resolution (1.678 3 2.58); (c) 2 3 climate resolution (1.258 3 1.8758); (d) 3 3 climate resolution(0.838 3 1.258). Values less than 22 mm day21 are shaded. The zero contour is omitted.

sea-breeze circulation and its implications for the meanclimate of the model are addressed in a later section.

Much of the important geographic detail of the Mar-itime Continent is not resolved at the coarse climateresolution of the model, and at first sight, it could beconcluded that the errors described above might be cor-rected with higher horizontal resolution. A number ofAMIP II experiments have been performed to addressthe sensitivity of model errors to horizontal resolution(Stratton 1999). Many improvements are seen in theseexperiments, such as a stronger midlatitude circulationdue to better resolved storm tracks. However, even witha threefold increase in horizontal resolution there is nosystematic improvement of the dry bias in the model(Fig. 2). The pattern of the errors in the tropical pre-cipitation persists and, if anything, is enhanced withincreasing resolution.

The results in Fig. 2 demonstrate that computationallyachievable increases in resolution do not necessarilylead to improvements in model error, and indicate thatit is probably deficiencies in the representation of thephysical system that are the primary source of theseerrors. The fact that many models show similar prob-lems over the Maritime Continent suggests that they alllack some key ingredient. In section 4 it is hypothesizedthat the diurnal cycle and the generation of land–seabreezes around this complex system of islands may bepossible factors.

The precipitation errors shown in Fig. 1 represent asubstantial fraction of the mean precipitation, which se-verely compromises the tropical heating pattern in the

model (not shown). The global scale of this problem isdemonstrated in section 5. In the following sections, asensitivity experiment is described, which aims to im-prove the heating distribution over the Maritime Con-tinent and thereby show the importance of an accuratesimulation of the climate of the Maritime Continent forthe global circulation.

3. Role of the islands in the climate of theMaritime Continent

As shown in the AMIP II experiments, the UM isunable to simulate well the mean climate over the Mar-itime Continent in terms of both the precipitation dis-tribution and the circulation. It is clear, therefore, thatthe UM has great trouble in representing the observedimpact of the islands of the region on the resolvableportion of the flow. Closer investigation over the indi-vidual islands reveals errors in their surface propertiesthat may be having a detrimental impact on the modeledconvection, possibly leading to the mean negative bias.The correct mean surface air temperature is importantto the boundary layer stability, which is the closuremethod for the convection scheme in the UM (Gregoryand Rowntree 1990). Figure 3a shows the mean surfaceair temperature error, when compared with the clima-tology of Legates and Willmott (1990), from the stan-dard AMIP II ensemble experiments at climate reso-lution. Over the island grid points a cold bias predom-inates (see Fig. 4a for exact location of land grid points).This is in excess of 28C over Borneo, the Philippines,


FIG. 3. Annual mean surface air temperature error (8C) for the UM AMIP II experiments over the Maritime Continent region at(a) climate resolution and (b) 3 3 climate resolution.

FIG. 4. Elevation of land grid points (m) and annual mean SST (8C) used in (a) control and (b) no islands experiments.

and New Guinea. The problem is not simply a resolutionissue, since even at 3 times climate resolution there isstill a mean cold bias in surface air temperature overthe islands (Fig. 3b).

A further hindrance to convection may be the pres-ence of layer cloud in the lowest layer of the model,essentially a fog layer. Mean values can reach in excessof 30% over the islands (not shown). This also masksthe fact that there exists a strong diurnal signal wherecloud amounts in the lowest model layer can build toalmost 100% cover during the night. Therefore aftersunrise, insolation has to burn off this excessive lowcloud before it can begin to heat the surface in order togenerate low-level instability. This could conceivably

be a hindrance to the correct evolution of convection,the consequences of which are discussed in the nextsection.

In an attempt to investigate the influence of the coldbias over the islands, and to artificially improve theclimate of the region, two further AMIP II–type exper-iments have been carried out. One of the aims of theseexperiments is to identify how much of the error in theglobal Tropics seen in Fig. 1 can be attributed to aremote response to errors over the Maritime Continent.The first experiment is equivalent to the standard AMIPII integration for the full 17 years of the AMIP II periodbut with additional diagnostics. This will be referred toas the ‘‘control’’ experiment. The second is an identical


FIG. 5. Mean differences between the no islands and control experiments: (a) DJF precipitation (mm day 21); (b) as in (a) but for JJA; (c)DJF 850-hPa wind vectors (m s21); (d) as in (c) but for JJA. Precipitation changes less than 21 mm day21 are shaded and the zero contouris omitted.

experiment except that the land grid points composingthe Maritime Continent islands have been removed andreplaced by ocean grid points with SST bilinearly in-terpolated from the surrounding existing ocean gridpoints (see Fig. 4). This will be referred to as the ‘‘noislands’’ experiment. In fact, two realizations of the noislands experiment were performed to ensure that theresults were significant.

The rationale behind the removal of the land gridpoints is to determine how detrimental the incorrect rep-resentation of the Maritime Continent islands may beto the generation of strong convection throughout theyear. By replacing these land grid points with warmerSSTs, the cold bias noted in Fig. 3 is eliminated, whichwith the enhanced moisture availability from the seasurface should lead to greater convective activity. There-fore, it is hoped that the mean bias may be correctedby providing boundary forcing, which is more condu-cive to strong convection than in the control experiment.While this experiment is a useful excerise in determiningthe impacts of improving model error, it would not rep-resent a viable solution given that any improvementswould essentially be for the wrong reason.

The mean response to the removal of the islands inthe Maritime Continent is a net increase in precipitationover and around the region. Figure 5 shows the changesin precipitation and lower-tropospheric flow associatedwith the removal of the Maritime Continent islands. Thisfigure should be compared with Fig. 1, showing theerrors from observations of the original AMIP II ex-periments. In DJF there is an increase in precipitation

off the coast of the western islands of the MaritimeContinent and a decrease in the SPCZ region and northof the Australian continent, reducing the strength of thewinter monsoon. These changes largely correct for thecomplex pattern of precipitation error seen in the stan-dard version of the UM. Nonlocal changes in precipi-tation are also evident in the western Indian Ocean,which again partially correct for errors there.

In JJA, changes in precipitation are more dramaticand extend to a greater part of the adjacent regions northand east of the equator. Over the Maritime Continentregion there is a coherent increase in precipitation cor-recting for the existing bias. In response to these in-creases there is a general reduction in precipitation onthe periphery of the Maritime Continent. Over the equa-torial Indian Ocean this leads to an improved, morezonally oriented precipitation distribution. To the northand northeast of the region, the changes lead to a morerealistic strength for the Asian summer monsoon andthe removal of the spurious extension of the precipi-tation pattern into the western Pacific, east of the Phil-ippines. As in DJF the SPCZ wet bias is reduced.

Although Fig. 5 shows a consistent improvement inthe precipitation field, the lower-tropospheric windchanges tend to reinforce existing errors in DJF, withenhanced easterlies in the region, particularly north ofthe equator. In JJA the wind changes also add to theexisting errors in the west Pacific, but in the IndianOcean and over southern Asia the changes are somewhatmore favorable, consistent with the precipitation chang-es. The low-level flow into the Indian and Asian mon-


FIG. 6. Mean monthly variation of precipitation (mm day21) av-eraged over the region 208S–108N and 908–1508E for the 17 yearsof the standard AMIP II integrations for the control and the no islandsexperiments.

TABLE 1. Surface energy components (W m22) [latent heating (LH); sensible heating (SH); shortwave heating (SW); Longwave heating(LW)] and surface moisture components of precipitation and evaporation (mm day21) averaged over the Maritime Continent region (158S–158N; 908E–1508E) during the AMIP II experiment period (1979–95) for the control and no islands experiments, and the ERA/Xie–Arkinclimatologies. Positive energy budget values indicate fluxes out of the surface.

Source LH SH SW LW Precipitation Evaporation

ControlNo islandsERA/Xie–Arkin

130.9136.9118.6

12.2810.25

9.09

2226.42230.32188.8

56.356.348.3

5.836.646.43

4.534.724.10

soon region is reduced and the flow is diverted moretoward the Maritime Continent after crossing the equa-tor in the western Indian Ocean. Such a change in theflow is consistent with a stationary Rossby wave re-sponse to the enhanced heating over the Maritime Con-tinent. However, considerable errors remain, particularlyin the low-level flow over the Indian subcontinent. Thissuggests that these errors may be due to more localproblems associated with the interaction of the modelphysics in the monsoon region (Martin 1999).

The changes seen in the no islands experiment aresignificant when compared to both the standard AMIPII ensemble and the control integration. Figure 6 showsthis clearly, with the mean annual variation in precip-itation averaged over the Maritime Continent area lyingoutside the ensemble spread of the standard AMIP IIintegrations and being significantly different from thecontrol experiment, particularly during northern sum-mer. More importantly the mean annual evolution is inmuch closer agreement with observations than any ofthe standard AMIP II or control experiments.

The overall increase in precipitation over the Mari-time Continent region must be accompanied by an in-crease in the supply of moisture. Since the island gridpoints are being replaced by warmer, saturated oceangrid points, one possibility is that the low-level buoy-ancy is increased, due to temperature and humidity ef-fects, with the moisture supply for the increased con-vection being provided locally by the enhanced evap-oration from the sea surface. However, this is not en-tirely the case. Table 1 summarizes the surface energy

and moisture budgets for the Maritime Continent region.There is a net increase in precipitation in the no islandsexperiment, which brings the total into closer agreementwith estimates from the Xie–Arkin climatology. How-ever, local evaporation within the Maritime Continentregion accounts for only a quarter of the precipitationchange between the two experiments. Therefore, the re-maining 75% must be provided by nonlocal moistureconvergence. This also raises the issue of whether theincrease in surface temperature, due to the reduction oforographic elevation to sea level, or the increase in sur-face moisture availability, is important for the increasein precipitation in the no islands experiment.

To address this issue a second much shorter sensitivityexperiment was performed where the land grid pointswere retained but the orography was reduced to zero.Briefly, the results show a local and remote responsethat is quantitatively similar to the no islands experi-ment, but of much reduced magnitude. Therefore, wecan conclude that the absence of an interactive landsurface and its replacement by a fixed SST boundaryforcing provides a much greater contribution toward theenhanced mean precipitation in the no islands experi-ment than the removal of the orographic forcing.

As well as a change in the hydrological cycle, Table1 also shows that the surface shortwave radiation in-creases despite the enhanced convective activity. Thisis because the diurnal cycle of cloud in the lowest modellayer is eliminated over the island grid points. Table 1suggests that although the desired effects of increasedprecipitation and some improved aspects of the clima-tology have been achieved, this has been at the expenseof a deterioration in the surface energy budget with thecaveat that the ERA fluxes are themselves not alwaysreliable.

The no islands experiment has demonstrated that in-adequate treatment of the islands of the Maritime Con-tinent in global climate models may be responsible forthe deficiencies in precipitation in this region and henceto errors in the tropical heating distribution throughoutthe warm pool region of the west Pacific and IndianOceans. Despite the unrealistic nature of the no islandsexperiment, the improved rainfall climatology over theMaritime Continent in this integration has proved keyto answering questions relating to the global effects oferrors in the atmospheric heat source of the MaritimeContinent, as discussed in section 5. It is clear that themeteorology of the islands is a key factor in determining


FIG. 7. Local time of the maximum in the diurnal cycle of precipitation over the Maritime Continent: (a) mean of two winters derivedfrom the global window brightness temperature of the Cloud Archive User Service (CLAUS; see Yang and Slingo 2001); (b) a single winterfrom the control experiment.

the heat and moisture budgets. In the next section aparticular aspect of that meteorology, the diurnal cycle,will be examined.

4. Role of the diurnal cycle in the MaritimeContinent

On shorter timescales the characteristics of the diurnalcycle, its phase and amplitude, are important for theorganization of precipitation in the Tropics. The studyof Yang and Slingo (2001) reveals that the UM has lowskill in reproducing the observed diurnal variation ofrainfall over the Maritime Continent, and indeed overall tropical land areas. Throughout the convectively ac-tive Tropics, the model systematically maximizes pre-cipitation too early during the day. Over land the peakoccurs predominantly near local noon, too soon afterthe solar maximum, while over the ocean the peak isaround local midnight.

More specifically, over the Maritime Continent regionthe model demonstrates particularly poor performance(cf. Figs. 7a,b). In the satellite observations, convectionis seen to maximize in the late evening over the islands,whereas the model shows a maximum near local noon.This implies that the life cycle of convection in themodel, from initial buoyancy excess of air parcels,through shallow cumulus and cumulus congestus, to or-ganized thunderstorms, is much too short. Strong con-vection develops too soon during the day, which thencuts off the solar radiation to the surface, leading totime-average surface temperatures that are consistentlylower than observations over all the Maritime Continentislands (Fig. 3). Such shortcomings of the model maynot be surprising, since the observed diurnal cycle overthe Maritime Continent islands is known to involve sub-

tle interactions with small horizontal scale (tens of ki-lometers) sea-breeze circulations (Keenan et al. 2000),as well as a smooth transition through a multistage con-vective life cycle (Saito et al. 2001).

As well as the problems noted above, other short-comings in the simulated diurnal cycle have been high-lighted, most acutely over land and just off the coastswhere the diurnal cycle accounts for a large proportionof the precipitation variability. As an example, Fig. 8demonstrates the contrasting characteristics in the di-urnal cycle averaged separately over land and oceanareas within the Maritime Continent. Here the diurnalcycle is calculated using the first three diurnal harmonicsderived from instantaneous data at three hourly inter-vals.

The convective heating in Figs. 8a,b shows that, av-eraged over the land grid points, there is a strong max-imum just before noon in the midtroposphere. What isalso evident is the dominance of deep convection in themodel, with no apparent buildup of convection throughthe morning. Over ocean grid points, the weaker am-plitude shows that the diurnal cycle is a less importantcomponent of precipitation variability. The convectivemoistening in Figs. 8c,d reveals that the action of theconvection is predominantly to dry the atmosphere, im-plying that the convection rapidly develops to precipi-tating convection in the model without the moistening,preconditioning phase associated with cumulus conges-tus. Idealized experiments with an aquaplanet versionof the UM have confirmed that the drying of the at-mosphere by precipitating convection is an overdomi-nant process in the model (Inness et al. 2001), in thatcase with implications for the intraseasonal organizationof convection associated with the MJO.


FIG. 8. Mean diurnal cycle derived from one year of a 3-hourly sampled integration of the control ex-periment: (a) convective heating (8C day21) over the land grid points of the Maritime Continent; (b) convectiveheating (8C day21) over the ocean grid points adjacent to the Maritime Continent; (c) as in (a) [and (d) asin (b)] but for convective moistening (g kg21 day21); (e) as in (a) [and (f ) as in (b)] but for total cloudcover fraction.

The mean diurnal evolution of the total cloud fractionis shown in Figs. 8e,f. First, the diurnal cycle of cloudfraction in the lowest model layer is very marked overland, ranging from 0% at the time of the convectivemaximum to an islands-wide maximum of 70% a coupleof hours prior to sunrise. This undoubtedly has an im-pact on the evolution of convection since clouds willdelay the time at which insolation can start to heat the

land surface. As already noted, it also affects the overallenergy budget of the island grid points.

It is clear from the above results that the model hasserious difficulties in simulating the diurnal cycle overthe islands of the Maritime Continent and it is importantto investigate whether this failing has implications forthe simulated mean climate of the region. In commonwith many climate models (e.g., ECMWF; Morcrette


2002), the UM does not perform a full radiation cal-culation at every time step. For reasons of computationalcost, the standard practice is to perform a full radiationcalculation every 3 h or six model time steps. At in-termediate time steps, for the purpose of the radiationcalculation, the cloud amounts and surface and atmo-spheric temperatures remain unchanged, although thesolar fluxes and heating rates are scaled by the correctzenith angle. This means, first, that the longwave fluxescannot respond to changes in solar forcing, and second,that there is a time lag in the cloud forcing, particularlyfor the surface fluxes.

It is possible that this approximation may be impor-tant for the evolution of the diurnal cycle. For example,a strong surface energy imbalance may occur at the startand end of the solar day, when the insolation is changingrapidly but the atmospheric and surface properties areonly updated every 3 h. The combination of the short-wave and longwave approximations could contribute toa too-rapid increase in surface temperature and the gen-eration of strong lower-tropospheric buoyancy excesstoo early in the day. This could lead to a maximum inconvection over the tropical islands also too early in theday, which would in turn cut off the solar radiation andlimit the net energy input to the land surface (Yang andSlingo 2001).

To investigate the effect of such approximations inthe model’s radiation calculation, a sensitivity experi-ment was performed with a full radiation calculationcarried out every time step, hereafter referred to as the‘‘full radiation’’ experiment. Due to the high compu-tational cost, the integration was run for only 15 months,with the final 12 months retained from March 1979–February 1980 to provide data for each season. Figure9a shows the mean diurnal evolution of surface air tem-perature in the control and full radiation experiments.Although the results from the full radiation experimentshow a less rapid rise in surface temperature early inthe day, changes in the diurnal cycle of convective rain-fall (Fig. 9b) are slight. There is a systematic increasein precipitation of between 1 and 2 mm day21 between0600 and 1800 local time but no significant shift in thephase of the maximum precipitation. There does, how-ever, appear to be a shift in the phase of the maximumoutgoing longwave radiation (OLR) by about 2 h from1100 to 1300 local time (Fig. 9c). This can be relatedto an increase in the terminal detrainment from the en-hanced convection (Fig. 9d), evident in the increase inupper-tropospheric cloud of almost 10% over the landgrid points (Fig. 9f). Over the ocean grid points of theMaritime Continent, the characteristics of the diurnalcycle are mainly unchanged in the full radiation exper-iment.

Although an accurate treatment of the diurnal cyclein the radiation calculation has not had a major impacton the phase of the diurnal cycle, nevertheless thechanges in strength of the diurnal cycle appear to affectthe mean climate of the model. Figure 10a shows the

precipitation difference between the full radiation andcontrol experiments. In general, over the islands thereis a net increase in precipitation of around 1 mm day21,consistent with the 1–2 mm day21 increase seen duringthe daytime in the diurnal cycle. In response to theselocalized changes in heating there are larger-scalechanges in the circulation. These lead to a general en-hancement of precipitation on and south of the equator,coupled with a decrease north of the equator. They gosome way to correcting for the geographical distributionof the systematic precipitation errors shown in Fig. 1,although, unlike the no islands experiment, the changesare not significant in the context of the ensemble ofAMIP II integrations described earlier (see Fig. 6).

The full radiation experiment has enabled two im-portant conclusions to be drawn. The first is that errorsin the phase of the diurnal cycle over land are relatedto more fundamental errors in the physical processes ofthe model, such as the evolution of the convection field,as discussed by Yang and Slingo (2001). Second, it hasdemonstrated that even quite small but systematicchanges to the diurnal cycle can rectify onto the meanclimate, suggesting that significant improvements to thediurnal cycle over the Maritime Continent could havethe potential to make a major impact on the model’slarge-scale systematic errors.

5. Global effects of systematic errors over theMaritime Continent

In both the no islands and full radiation experiments,there is an enhancement of the heat source over theMaritime Continent region. On seasonal to interannualtimescales this will have an impact away from this re-gion through the generation of Rossby waves. There-fore, it is appropriate to assess the global effects of thisenhanced heating, and to determine the possible impactson model error if improvements in the mean climate ofthe Maritime Continent were achieved. The oppositeside of this argument is that we can also assess wherethe model may be in error remotely, because of the errorsassociated with the Maritime Continent heat source.

As identified in the analysis of the standard AMIP IIexperiments, the model underestimates the tropical con-vective heat source over the Maritime Continent, com-pared to observations, while creating local maxima overthe west Indian Ocean and west Pacific. The increaseddeep convection in the no islands experiment enhancesthe divergent outflow and, depending on the ambientconditions, has the potential to influence remote areasof the globe through planetary wave propagation, asproposed by Sardeshmukh and Hoskins (1988). Figure11 shows that this is certainly the case. The change inthe tropical heat source in the no islands experiment hassignificant remote effects, which give rise to large-scalesurface temperature increases over western Russia andScandinavia, as well as to other significant changes overNorth America. The extratropical surface temperature


FIG. 9. Comparison of the mean diurnal cycle over the Maritime Continent land grid points from 12-month integrations of the control and full radiation experiments: (a) surface air temperature (8C) using 3-hourly radiation time steps (solid line) and 0.5-hourly radiation time steps (dashed line); (b) as in (a) butfor precipitation (mm day21); (c) as in (a) but for OLR (W m22); (d) change in the mean diurnal cycle ofconvective heating (8C day21), 0.5-hourly radiation time steps minus 3-hourly radiation time steps; (e) asin (d) but for convective moistening (g kg21 day21); (f ) as in (d) but for total cloud cover fraction.

changes in Figs. 11a,b are a consequence of changes inthe global circulation, which can be described by the500-hPa geopotential height anomalies in Figs. 11c,d,respectively. Only the winter hemisphere is shown ineach case since it is in this season that the ambientconditions are most conducive to tropical–extratropicalinteraction.

The changes in surface temperature and 500-hPa geo-potential height can be compared with the UM’s sys-tematic errors in these fields, as shown in Fig. 12. This

reveals the importance of an improved representation ofthe Maritime Continent heat source. In DJF the NorthernHemisphere is mainly dominated by a cold bias in sur-face temperature, although over North America themodel has a warm bias over Canada and a cold biasover the United States and Mexico. This general large-scale pattern of systematic errors is substantially cor-rected in the no islands experiment by changes of 0.58–18C (Fig. 11a). Similarly over Scandinavia and northeastRussia, the systematic cold bias of over 58C is partially


FIG. 10. Change in the mean climate over the Maritime Continent from the full radiation experiment compared with the controlintegration: (a) precipitation (mm day21); (b) OLR (W m22).

FIG. 11. Seasonal mean differences averaged for 1979–95 between the no islands and the control experiments: (a) DJF Northern Hemispheresurface air temperature (8C); (b) JJA Southern Hemisphere surface air temperature (8C); (c) DJF Northern Hemisphere 500-hPa geopotentialheight (m); (d) JJA Southern Hemisphere 500-hPa geopotential height (m).

corrected by the changes in the no islands experimentwith an increase in surface temperature of over 28C.Errors in Southern Hemisphere surface temperature dur-ing JJA are complicated by the problems with modelsea ice on the periphery of the Antarctic continent (Fig.12b).

The 500-hPa height changes during both DJF and JJA(Figs. 11c,d) in the no islands experiment lead to moreobvious improvements in the model’s systematic errors(Figs. 12c,d). Again the effects are wide ranging, with

substantial changes in the stationary waves over theEuro–Atlantic sector. The potential for heating anom-alies over the warm pool to influence the Euro–Atlanticsector is consistent with the response of the global cir-culation to La Nina. The pattern of 500-hPa geopotentialheight anomalies over the Euro-Atlantic sector in Fig.11c is reminiscent of that associated with the 1998/99La Nina. This was also characterized by enhanced con-vection over the Maritime Continent and west Pacificregions (Dong et al. 2000), similar to the changes in the


FIG. 12. Seasonal mean errors in the standard AMIP II ensemble experiments: (a) DJF Northern Hemisphere surface air temperature (8C);(b) JJA Southern Hemisphere surface air temperature (8C); (c) DJF Northern Hemisphere 500-hPa geopotential height (m); (d) JJA SouthernHemisphere 500-hPa geopotential height (m).

heating pattern between the control and no islands ex-periments.

In JJA, the 500-hPa geopotential height anomaliesover the southern oceans are clearly linked to an equiv-alent barotropic planetary wave response emanatingfrom the enhanced convection over the Maritime Con-tinent region. However, in DJF the wave propagationpatterns are less coherent and, if anything, the surfacetemperature anomalies over western Russia appear tobe the result of upstream development from the Mari-time Continent. During DJF, although there is a net in-crease in precipitation over the Maritime Continent, thechange in the heating pattern is less coherent spatially.This leads to a smaller-scale increase in upper-level di-vergence over the region, which does not excite thecoherent wave activity seen in JJA, where there is alarger and more expansive region of enhanced upper-level divergence.

Although the mechanism for the remote, extratropicalresponse in DJF is unclear, it is certainly the case thatthese are significant changes. The same patterns werereproduced in both no islands integrations, and theanomalies were statistically different from the intraen-semble variability of the climate in the standard AMIPII experiments. Such changes to the global circulationraise two points. First, they show that it is vital that thehydrological cycle over the Maritime Continent is sim-ulated accurately since the effects of correcting the localheat source are indeed global. Also, if tropical changesin heat sources can have such dramatic effects on the

surface temperatures of the continents of the mid-lati-tudes and high latitudes, then they serve as a warningwhen trying to correct locally for biases in the model.For example, over western Eurasia, the significant coldbias of up to 88C has been attributed to problems withfreezing of soil moisture. However, it has been shownhere that even a modest improvement in the heat sourceover the Maritime Continent can account for up to 28Cof this bias.

6. Discussion and conclusions

This paper has investigated the errors in the simu-lation of the climate of the Maritime Continent in theMet Office Unified Model. The lack of precipitationover the region is an ubiquitous feature of the modelresults during all seasons, as shown in standard AMIPII experiments. These errors persist even in AMIP IIintegrations at much higher resolution, implying that,at these resolutions, deficiencies in the representation ofthe physical system are primarily responsible.

The importance of the Maritime Continent climatehas been demonstrated using AMIP II sensitivity ex-periments. By removing the island grid points of theregion and replacing them with oceanic grid points,some aspects of the mean climate are improved. Pre-cipitation increases to be closer to observed amountsand the flow patterns over the Indian Ocean during theAsian monsoon season are improved, such that the ex-cessive circulation strength is reduced. However, sig-


nificant errors remain. In particular the excessive lower-tropospheric easterlies in the west Pacific have beenmade worse by the removal of the island grid points.

Following the results of Slingo et al. (1996), whichsuggested that the MJO was sensitive to the mean cli-mate in the warm pool region, it was hypothesized thatthe no islands experiment might lead to an improvedMJO. Diagnosis of the MJO shows no significant changein this experiment suggesting that, like the diurnal cycle,it is more fundamentally dependent on aspects of themodel’s formulation, such as the vertical resolution(e.g., Inness et al. 2001), interaction with the oceansurface (Inness and Slingo 2003), and the physical pa-rametrizations.

The influence of the Maritime Continent has beenshown to be global. A robust teleconnection pattern isseen, particularly in the Southern Hemisphere winterseason, which is consistent with an equivalent baro-tropic Rossby wave response to an enhanced tropo-spheric heat source located over the Maritime Continent.Such a response raises important questions concerningthe methodology for tackling systematic errors in cli-mate models. It is possible that model errors in a par-ticular region, for example, Eurasia, may be attributedto local problems such as soil moisture freezing (e.g.,Viterbo et al. 1999). However, given that improvementsover the Maritime Continent have been shown to leadto significant changes in remote locations, then care hasto be taken when addressing systematic errors basedentirely on local processes.

Although no final solution to the systematic errors inthe model’s tropical heating distribution has been pre-sented here, this paper has highlighted the importanceof the diurnal cycle for the climate of the MaritimeContinent. Recent analysis of high-resolution satellitedata by Yang and Slingo (2001) has demonstrated thecoherent propagation of convection away from the is-lands, indicative of gravity waves. In particular, coastalregions with a large dry bias are strongly collocatedwith regions where these propagating diurnal signals areabsent in the model. It is hypothesized that capturingor better representing the effects of these diurnallyforced gravity waves and land–sea breezes may lead toimprovements in the mean precipitation field in twoways. First, they may increase the wind variability nearthe surface leading to enhanced surface fluxes, partic-ularly of moisture. Second, they may act as triggers forconvection.

It is clear that models with coarse-grid domains areincapable of capturing the smaller-scale land–sea breezecirculations and diurnal variability, which are likely tobe key processes in the regions of largest dry biasaround the Maritime Continent in the UM. At the sametime, experiments that have addressed island-scale con-vection at smaller scales (e.g., Saito et al. 2001) do notattempt to reproduce the large-scale organization in theoff-coastal regions by the diurnal forcing initiated overland, essentially the interaction of convection with the

larger scales. Therefore, to understand relevant pro-cesses an intermediate regional modeling approach isneeded. This requires the domain of the Maritime Con-tinent to be at fine enough resolution to capture thediurnal variability and sea breezes initiated over the is-land regions, but with a domain large enough to simulatethe large-scale forcing and organization. Such a mod-eling activity is currently being undertaken with the goalof developing a parameterization of the subgrid meso-scale organization in the context of a fractional tilingof coastal grid cells, which takes into account the dif-ferent surface fluxes from land and sea.

Much work is still needed to improve the subgrid-scale organization aspect of convective parameterizationin course-grid-scale general circulation models, boththrough collaborative programs aimed at better under-standing the physics of convection [e.g., European Pro-ject on Cloud Systems in Climate Models (EUROCS),online at http://www.cnrm.meteo.fr/gcss/EUROCS/EU-ROCS.html] and the use of novel techniques such asgrid-box cloud resolving model sampling (e.g., Gra-bowski 2001). In the intermediate term the use of aparameterization that is able to represent the knowndominant organization processes related to sea breezesis considered to be the best option for providing im-provements to the simulation of the Maritime Continentclimate.

Acknowledgments. The authors acknowledge manyuseful contributions from the CGAM Tropical Groupand the much appreciated input from three anonymousreviewers. Richard Neale is supported by the Met Officethrough grant Met 1b/2601. J. Slingo acknowledges sup-port through the Natural Environment Research Council(NERC)–funded UK Universities’ Global AtmosphericModeling Programme.

REFERENCES

Chang, C. P., and K. M. Lau, 1982: Short term planetary-scale in-teractions over the tropics and mid-latitudes during northern win-ter. Part I: Contrasts between active and inactive periods. Mon.Wea. Rev., 110, 933–946.

Cox, P. M., R. A. Betts, C. B. Bunton, R. L. H. Essery, P. R. Rowntree,and J. Smith, 1999: The impact of new land surface physics onthe GCM simulation of climate sensitivity. Climate Dyn., 15,183–203.

Dong, B. W., R. T. Sutton, S. P. Jewson, A. O’Neill, and J. M. Slingo,2000: Predictable winter climate in the North Atlantic sectorduring the 1997–1999 ENSO cycle. Geophys. Res. Lett., 27, 985–988.

Edwards, J. M., and A. Slingo, 1996: Studies with a flexible newradiation code. I. Choosing a configuration for a large-scale mod-el. Quart. J. Roy. Meteor. Soc., 122, 689–719.

Gates, W. L., 1992: AMIP: The atmospheric model intercomparisonproject. Bull. Amer. Meteor. Soc., 73, 1962–1970.

Gibson, J. K., P. Kallberg, S. Uppsala, A. Hernandez, A. Nomura,and E. Serrano, 1997: ERA Description. ECMWF Re-analysis,Project Report Series 1, Tech. Rep., ECMWF, Reading, UnitedKingdom, 72 pp.

Grabowski, W. W., 2001: Coupling cloud processes with the large-


scale dynamics using the Cloud-Resolving Convection Param-eterization (CRCP). J. Atmos. Sci., 58, 978–997.

Gregory, D., and P. R. Rowntree, 1990: A mass flux convectionscheme with representation of cloud ensemble characteristics andstability dependent closure. Mon. Wea. Rev., 118, 1483–1506.

Holland, G. J., and T. D. Keenan, 1980: Diurnal variations of con-vection over the ‘‘maritime continent.’’ Mon. Wea. Rev., 108,223–225.

Houze, R. A., S. G. Geotis, F. D. Marks, and A. K. West, 1981:Winter monsoon convection in the vicinity of North Borneo.Part I: Structure and time variation of the clouds and precipi-tation. Mon. Wea. Rev., 109, 1595–1614.

Inness, P. M., and J. M. Slingo, 2003: Simulation of the Madden–Julian oscillation in a coupled general circulation model. Part I:Comparison with observations and an atmospheric-only GCM.J. Climate, 16, 345–364.

——, ——, S. J. Woolnough, R. B. Neale, and V. D. Pope, 2001:Organization of tropical convection in a GCM with varying hor-izontal resolution: Implications for the simulation of the MaddenJulian Oscillation. Climate Dyn., 17, 777–793.

Keenan, T. D., B. R. Morton, M. J. Manton, and G. J. Holland, 1989:The island thunderstorm experiment (ITEX)—a study of tropicalthunderstorms in the maritime continent. Bull. Amer. Meteor.Soc., 70, 152–159.

——, and Coauthors, 2000: The Maritime Continent ThunderstormExperiment (MCTEX): Overview and some results. Bull. Amer.Meteor. Soc., 81, 2433–2455.

Kershaw, R., and D. Gregory, 1997: Parameterization of momentumtransport by convection. I: Theory and cloud modelling results.Quart. J. Roy. Meteor. Soc., 123, 1133–1151.

Lau, K. M., C. P. Chang, and P. H. Chan, 1983: Short term planetary-scale interactions over the Tropics and mid-latitudes duringnorthern winter. Part II: Winter-MONEX period. Mon. Wea. Rev.,111, 1372–1388.

Legates, D. R., and C. J. Willmott, 1990: Mean seasonal and spatialvariability in global surface air temperature. Theor. Appl. Cli-matol., 41, 11–21.

Martin, G. M., 1999: The simulation of the Asian summer monsoon,and its sensitivity to horizontal resolution, in the UK Meteo-rological Office Unified Model. Quart. J. Roy. Meteor. Soc., 125,1499–1525.

Morcrette, J. J., 2002: Assessment of the ECMWF model cloudiness

and surface radiation fields at the ARM SGP site. Mon. Wea.Rev., 130, 257–277.

Philander, S. G. H., 1985: El Nino and La Nina. J. Atmos. Sci., 42,652–662.

Pope, V. D., M. L. Gallani, P. R. Rowntree, and R. A. Stratton, 2000:The impact of new physical parameterizations in the HadleyCentre climate model: HadAM3. Climate Dyn., 16, 123–146.

Ramage, C. S., 1968: Role of a tropical ‘‘maritime continent’’ in theatmospheric circulation. Mon. Wea. Rev., 96, 365–370.

Saito, K., T. Keenan, G. Holland, and K. Puri, 2001: Numerical sim-ulation of the diurnal evolution of tropical convection over theMaritime Continent. Mon. Wea. Rev., 129, 378–400.

Sardeshmukh, P. D., and B. J. Hoskins, 1988: The generation of globalrotational flow by steady idealized tropical divergence. J. Atmos.Sci., 45, 1228–1251.

Slingo, J. M., and Coauthors, 1996: Intraseasonal oscillations in 15atmospheric general circulation models: Results from an AMIPdiagnostic subproject. Climate Dyn., 12, 325–357.

Stratton, R. A., 1999: The impact of increasing horizontal resolutionon the HADAM3 climate simulation. Hadley Centre Tech. Note13, Hadley Centre, Met Office, United Kingdom.

Sui, C. H., and K. M. Lau, 1992: Multiscale phenomena in the tropicalatmosphere over the western Pacific. Mon. Wea. Rev., 120, 407–430.

Viterbo, P., A. Beljaars, J. F. Mahfouf, and J. Teixeira, 1999: Therepresentation of soil moisture freezing and its impact on thestable boundary layer. Quart. J. Roy. Meteor. Soc., 125, 2401–2426.

Williams, M., and R. A. Houze, 1987: Satellite-observed character-istics of winter monsoon cloud clusters. Mon. Wea. Rev., 115,505–519.

Xie, P., and P. A. Arkin, 1996: Analyses of global monthly precipi-tation using gauge observations, satellite estimates, and numer-ical model predictions. J. Climate, 9, 840–858.

Yang, G. Y., and J. M. Slingo, 2001: The diurnal cycle in the Tropics.Mon. Wea. Rev., 129, 784–801.

Zhang, Y., K. R. Sperber, and J. S. Boyle, 1997: Climatology andinterannual variability of the East Asian winter monsoon: Resultsfrom the 1979–95 NCEP/NCAR reanalysis. Mon. Wea. Rev.,125, 2605–2619.

Zhu, B., and B. Wang, 1993: The 30–60 day convection seesawbetween the tropical Indian and western Pacific Oceans. J. Atmos.Sci., 50, 184–199.

the maritime continent and its role in the global climate: a gcm...

Documents