MESA Science Working Group

Panel Report from the 1st MESA SWG Meeting (SWG-1)

Mexico City, Mexico, March 9-11, 2005.




Table of Contents


Executive Summary

1. MESA status and issues

2. Summary of the sessions

            2.1 Review of the scientific basis for MESA

               2.1.1 PRA-1: Diurnal and Mesoscale Variability in the SAMS region

               2.1.2 PRA-2: Intraseasonal variability in the SAMS region

               2.1.3 PRA-3: Interannual and longer-time variability in the SAMS region.

            2.2 MESA modeling issues

            2.3 MESA applications

            2.4 Enhanced observation activities and field campaigns

3. MESA SWG executive session

Appendix 1: List of Participants

Appendix 2: Agenda

Executive Summary

            The 1st MESA Scientific Working Group Meeting (SWG-1) was held 9-11 March, 2005 in Mexico City, Mexico.  The meeting brought together the members of the MESA SWG and a group of invited participants with expertise and interests relevant to the MESA objectives. The objectives of the meeting were: to review the MESA scientific questions, to review progress on MESA modeling and diagnostic studies, to discuss MESA applications on La Plata Basin, and future field campaigns. The MESA SWG was also asked to discuss, and organize the MESA implementation plan, and to review and update MESA milestones and time line.

            SWG-1 meeting then consisted in a series of presentations and plenary discussions highlighting the key scientific basis for MESA, observational needs, modeling strategy, phases of implementation as well as Program organization. This report summarized the presentations and the plenary discussions made on these items.

           
    Specific recommendations from SWG-1 are as follows:


Program structure:

MESA PRA-1:

MESA PRA-2:

MESA PRA-3:

Modeling issues:

MESA enhanced observation activities:

            The meeting Agenda and the list of participants are given in the appendixes. The Agenda and presentations are also available on the MESA webpage: www.eol.ucar.edu/projects/mesa/.

           
    The MESA SWG is very grateful to Carlos Ereño (ICPO), Michel Rosengaus and Miguel Cortez and SMN team, and UCAR/JOSS (Gus Emmanuel, José Meitín, Tara Jay and Gene Martin) for the excellent organization of the meeting. Special thanks are also extended to our sponsors [OGP / CPPA (Jin Huang, Mike Patterson), US CLIVAR (David Legler), WCRP (Valery Detemmerman, Howard Cattle)]. Finally, we wish to thank all the colleagues who contributed to this report.


1. MESA status and issues

            Carolina Vera reviewed the current status of the MESA program with special emphasis on MESA reorganization and raised the following issues and questions.

            Currently MESA has been organized in regional programs. It was proposed in the last VAMOS Panel meetings that MESA should start integrating the objectives of the different programs in South America (SALLJEX, PLATIN, LBA) in an unified program in order to facilitate the understanding, simulation and prediction of the different components of the South American Monsoon System (SAMS), their variations and connections with the extratropics.

            Therefore, the MESA implementation plan needs to be discussed in SWG-1 and written during the rest of 2005 in order to better address the main MESA goal related to the SAMS variability and the improvement of the SAMS monthly-to-seasonal prediction. It was suggested in VPM7 and in the informal MESA meeting held in the CLIVAR Conference, in Baltimore, MD, that that MESA should concentrate its efforts in the region extended from the Amazon Basin to the La Plata Basin, encompassing the Monsoon core region. Also, MESA will be organized in three main priority research areas (PRAs), PRA1: Diurnal and mesoscale variability, PRA2: Intraseasonal variability and PRA3: Interannual and longer timescale variability (including climate change).

            A better understanding and improved simulation is expected on the SAMS evolution and variability. Specific scientific issues for PRA-1 include: diurnal cycle of circulation and precipitation, mesoscale convective processes, role of aerosols from biomass burning in SAMS; for PRA-2: dynamics of the SA see-saw pattern; ITCZ-SACZ interaction; influence of MJO on SAMS; relative roles of internal vs forced low-frequency variability; and for PRA-3: land surface forcing, impacts of land use change, role of remote and local SST (South Atlantic), global response to SAMS forcing, Sources and limits of predictability on SAMS region.

            Finally, MESA future directions should also  be coordinated with the implementation of the GEWEX-CLIVAR/VAMOS/La Plata Basin continental scale experiment (LPB CSE).

           

2. Summary of the sessions


2.1 Session 1. Review of the scientific basis for MESA


2.1.1 PRA-1: Diurnal and Mesoscale Variability in the SAMS region


Rain-Producing Systems for the SAMS and their Moisture Sources

Edward J. Zipser. University of Utah, USA

Paola Salio and Matilde Nicolini, UBA, Argentina

            There are major, systematic, regional differences in the structure, intensity, and diurnal cycle of rainfall systems.  The La Plata Basin (LPB) has a particularly extreme domination by large and intense MCSs.  According to 5 years of TRMM data, the region of northeast Argentina, Uruguay, and extreme southern Brazil have almost 80 percent of their annual rainfall contributed by mesoscale convective systems (MCSs), the highest such percentage in the world. 

            Ever since the seminal work on mesoscale convective complexes (MCCs) by Velasco and Fritsch (1987), it has been well-known that MCCs are common in this region.  However, it is the TRMM database that permits quantification of this domination and comparison to other regions of the world. 

            MCSs and MCCs have a nocturnal rainfall peak not only in the LPB but almost everywhere they are found in the world, over land or oceans.  Specifically in the LPB, both TRMM and IR data confirm this nocturnal peak for the SALLJEX period.  It is established that during SALLJEX (and in other years) that the South American low level jet (SALLJ) has a similar nocturnal maximum, so we explore whether the connection is direct or indirect.  In spite of the time synchronization of the two phenomena, statistically speaking, we may speculate that the role of the SALLJ in MCSs is important, but mostly as a provider of high-CAPE air exported from the Amazon basin.

            Further research on this and related issues should be pursued as high priority.  When SALLJEX-period reanalyses are undertaken, it is recommended that they should assimilate high-resolution databases and also pay close attention to thermodynamic variables.  There is ample evidence that without SALLJEX data, the details of the high theta-e air and high moisture flux are not captured well.  These are an essential fuel for the LPB MCSs.   Lastly, satellite and radar databases can provide important statistical constraints on convective intensity as well as rainfall.


The Diurnal Cycle of Precipitation over South America based on CMORPH

Vernon E. Kousky

NOAA/ Climate Prediction Center

Camp Springs, MD 20746

            The diurnal cycle of precipitation over the region of South America is examined using high spatial and temporal resolution analyses that have been produced by NOAA's Climate Prediction Center morphing technique (CMORPH) (Joyce et al. 2004).  The 8-km spatial resolution (at the equator) and 30-minute temporal resolution of these analyses permit an in-depth look at the diurnal cycle of precipitation. 

            Major features of the summertime diurnal cycle, as depicted by CMORPH for the South American Monsoon, include an afternoon maximum in precipitation over the Andes and the high terrain in central and eastern Brazil, a nocturnal maximum in precipitation over areas just east of the Andes (western Argentina, central Bolivia and western Paraguay), and a nocturnal maximum over the Atlantic Ocean in the vicinity of the South Atlantic Convergence Zone (Fig. 1). A remarkable diurnal cycle in precipitation occurs in coastal areas of northern and northeastern South America. With daytime heating, precipitation rapidly develops along and just inland from the coast (Fig. 1, lower left panel), probably related to the sea breeze (Kousky 1980). This precipitation advances westward and southward, producing a nocturnal maximum in areas approximately 500 km inland from the coast (Fig. 1 upper left panel). The inland propagation of sea-breeze-induced rainfall systems is a feature most frequently found during late SH summer (December-February) and fall (March-May).

            The seasonal average diurnal cycle for equatorial South America (Eq. - 5¡N) for March-May 2003 indicates that sea-breeze-induced precipitation systems propagate westward, reaching the western Amazon Basin in about two days (see dashed lines in Fig. 2). As these systems propagate inland they contribute to a nocturnal precipitation maximum in some areas and a diurnal precipitation maximum in other areas. A nocturnal or early morning precipitation maximum also occurs along the immediate coast and offshore in the vicinity of the Atlantic ITCZ and over the Pacific near the west coast of South America.

References: Joyce, R. J., J. E. Janowiak, P. A. Arkin, and P. Xie, 2004: CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. J. Hydromet., 5, 487-503.

Kousky, V. E., 1980: Diurnal rainfall variation in Northeast Brazil.  Mon. Wea. Rev., 108, 488-498.

Figure 1. Mean percent of daily total precipitation for 03-06 UTC, 09-12 UTC, 15-18 UTC, and 21-00 UTC. The mean is computed for the combined December-February periods for 2002-03 and 2003-04. Note: if rainfall were distributed equally throughout the 24-h period then 12.5% would be the expected percentage of the daily total for each 3-h interval. Percentages have been masked out in regions where rainfall average less than 1 mm d-1.


On the role of soil moisture in the precipitation processes of the South American Monsoon.

Estela Collini and E. Hugo Berbery

            Ensemble simulations with the regional mesoscale Eta model are employed to investigate the feedbacks between soil moisture and precipitation over the South American Monsoon region. Earlier work has shown that moisture is supplied to the monsoon region in large part by advective processes through a Low-level Jet east of the Andes, although local soil conditions may modulate the intensity of precipitation.  Recent studies have shown that soil moisture may have significant links to regional climate during summer, and that it increases the persistence of precipitation events. Other studies have shown that the predictability of precipitation over a given region can be increased if soil moisture effects are properly taken into account.

            The ensemble simulations are performed for Octobers of different years, and with varying conditions of initial soil moisture.  October is chosen because it is the month when the monsoon gets established. Changes in soil moisture affect the latent heat, sensible heat (and consequently the Bowen ratio), as well as the vertical structure of the troposphere.  Our results suggest more sensitivity to drier conditions than to wetter states. This would suggest that the land surface in the monsoon area is near saturation; hence increasing soil moisture does not have a notorious effect. On the other hand, drier conditions at the surface favor less precipitation (which feeds back to lower soil moisture).

            Notably, southward of the core monsoon the soil moisture effects tend to be more difficult to recognize, probably because El Niño has a strong influence in south eastern South America that overshadows the local effects. An important caveat is that the anomalies in soil moisture are assumed to be proportional to the total field. Therefore, local changes in soil moisture might have better defined effects. We have plans to investigate this last aspect.


Coupling between precipitation and the large scale flow on the diurnal cycle in the Amazon: sensitivity to ground wetness

Rafael Terra  IMFIA, Universidad de la Repœblica, Uruguay

            The influence of ground wetness on precipitation on the Central Amazon region during January was analyzed through tailored GCM simulations, emphasizing the role of the diurnal cycle.

            The model's diurnal cycle in the selected region shows a quite uniform behavior. The diurnal variation of the large-scale ascending motion is out of phase with precipitation whose initiation seems to be dominated by surface forcing, unlike in the less continental regions where land-sea type of circulations play a role.

            Sensitivity was studied with respect to ground wetness, which directly controls flux partitioning thus bypassing the vegetation dynamics. Surface roughness and albedo effects usually associated with deforestation are also precluded in the model setting in order to isolate the effect of ground wetness.

            For a broad range (0.4 to 1.0) of the parameter multiplying ground wetness, the sensitivity of precipitation is quite small. Moreover, the precipitation grows as the sensitivity parameter is brought down from 1.0 to around 0.8 due to a compensating effect of enhanced mass (moisture) convergence that overcomes the reduced evaporation. The precipitation decreases sharply with the sensitivity parameter only for values smaller than 0.4 as convection in the box decreases.

            As we look into the diurnal cycle, we notice that the evolution of ground and surface temperature is gradual through the entire range of the parameter. However, this is not the case for PBL humidity that clearly shows a change in the diurnal cycle at a value of 0.3 in the parameter. This feature is also evident in the precipitation distribution during the day and reflects a drastic change in the PBL diurnal evolution into a situation that is too dry to sustain convection, which collapses as the ground wetness gets smaller.

            In view of the initial results, it is hypothesized that this sensitivity of precipitation to ground wetness is bound to depend on the region and the dominating diurnal destabilization mechanisms.


Diurnal variability during SALLJEX
Celeste Saulo (UBA - Argentina),
Julia Paegle (U of Utah - USA), Ben Kirtman (COLA - USA)
            The diurnal cycle over South America is discussed based on current analysis and observations.  Analyses are: 1) the NCEP/NWS/NOAA Global Data Assimilation System available during SALLJEX and 2) COLA climate model output forced with observed SSTs.  Observations are those obtained from the South American Low Level Jet Experiment and precipitation estimates from satellites as obtained with the NOAA CPC Morphing Technique (CMORPH). The latter produce estimates of precipitation at a high horizontal and temporal resolution. CMORPH shows an apparent eastward propagation of precipitation from the subtropical Andes foothills eastward towards the Argentine Pampas, with a well-defined nocturnal maximum over south-east South America (SESA). The COLA model does not capture the diurnal cycle of temperature and moisture over SESA, consistent with the lack of nocturnal precipitation over this region. Analyses agree with observations in depicting an anti-cyclonic turning of the wind with time of day within the planetary boundary layer.


Conclusions of the plenary discussion for PRA-1:  Diurnal and Mesoscale Variability in the SAMS Region

E.J. Zipser, Chair; Rafael Terra, Rapporteur

MCSs:

  • Mechanisms forcing and organizing MCS in addition to the SALLJ include eastward propagating gravity waves excited by convection on the Andes slopes, and it is noted that group velocities of the waves may be dependent on latent heat profiles which in turn are dependent upon the precipitation system (nonlinear problem).

  • Resolution is crucial for the ability to simulate MCSs.  Speculation is that without explicit convection in models with grid sizes less than 3 km, we should not expect accurate simulations.

  • Another reason why MCSs are absent in many models may be the weakness of transient forcing from the large-scale model.

    Soil  moisture:

  • Soil hydrology in pasture in the COLA model is bad and generates the warm bias in SESA (land model too coarse).   Soil hydrology is as uncertain as convection and should have close attention in data assimilation for the SALLJEX period.  It is a difficult issue but a very important one.

    Reanalysis

  • Moisture reanalysis is not good, with large imbalances in water budgets.  Probable causes include errors in evapotranspiration, and poor resolution in both models and data.

    Model intercomparisons and Diurnal cycles

  • There is a consensus that devising a sound but workable strategy for intercomparisons is a high priority for MESA research.  We should begin with models that are already operational in the region and perform seasonal SALLJEX simulations, in spite of the major effort required.

  • After the mean, the diurnal cycle is basic for all other scales (i.e. resonance with intraseasonal).  We may get the annual mean right for the wrong reasons.  We should document the current status of the diurnal cycle in existing models.  A suggestion is to undertake a small diagnostic project, with runs already available, and then perform new runs. (WGCIP can help on that.)

  • An important metric of validation of the diurnal cycle for LBA has been flux towers and satellite data.  If we are to do the diurnal cycle intercomparison, we should check with GCSS groups (they did an intercomparison for convection in the Amazon) and LBA experience.  More controlled intercomparisons similar to those done by GCSS could use single column test cases with single column versions of the GCM.  Radar experts should be in the group, as they may provide the information needed to test sub-grid scale aspects and parameterizations.


    2.1.2 PRA-2: Intraseasonal variability in the SAMS region


    Intraseasonal to interannual variability of extreme precipitation over southeastern South America during the Austral Summer

    Leila M. V. Carvalho1, Michel Nobre Muza1& Brant Liebmann2

        1 University of Sao Paulo

        2 NOAA-CIRES Climate Diagnostics Center

                The occurrences of extreme precipitation on intraseasonal to interannual timescales are investigated over Southern (S) and Southeastern (SE) Brazil and Subtropical Western Atlantic (SA) during the austral summer. Pentad precipitation based on gauge observations and satellite estimates from 1979 to 2002 are used to identify extremes. Extreme precipitation and drought are investigated over regions with domain size 10°x10° latitude-longitude and 2.5o spatial resolution. Precipitation time-series are filtered on intraseasonal timescales (20-90 days) and interannual timescales (periods > 370 days).  Extreme wet and dry conditions are analyzed in each band-passed precipitation time-series separately and are determined based on the quartiles of the seasonal rainfall distribution, on each grid point for each timescale. Persistence of extreme events is also examined in this study.

                The results indicate that extreme precipitation and drought on interannual timescale are modulated by El Niño-Southern Oscillation (ENSO) phases over S Brazil, Uruguay and Northeastern Argentina. Extreme precipitation over SE Brazil is related to phases of the South Atlantic TSM dipole and is more frequent during Neutral/La Nina episodes. Extremes over Subtropical Atlantic are clearly decoupled from those over SE Brazil. Extreme precipitation on intraseasonal timescales does not show a clear relationship with ENSO phases, except when persistence of the events is considered.

                On intraseasonal timescales, the Madden-Julian Oscillation phases modulate part of the variability of extreme rainfall and drought events. Nevertheless, lag composites of circulation during extreme events on intraseasonal timescales show the importance of an atmospheric mechanism characterized by the propagation of midlatitudes wave trains that originate from tropical disturbances over the Pacific. In addition, lag-composites suggest that extreme precipitation events in the South Atlantic convergence zone are associated with positive intraseasonal precipitation anomalies over the South Pacific convergence zone, east of the date line, 2 pentads before the event. Moreover, an enhancement of the zonal wind component in the upper-troposphere appeared to be an important dynamical forcing for the occurrence of extreme precipitation events. Extreme wet events over southern Brazil are related to a strengthening of the northerly low-level jet (eastern of the Andes) that corresponds to the 70th percentile of its seasonal (DJF) magnitude.  On the other hand, extreme precipitation events over SE Brazil are associated with the strengthening of the northerly wind component over the tropical Atlantic near the northeast coast of Brazil, which corresponds to the 80th percentile of its seasonal magnitude.

                This study shows that the precipitation seesaw between southern and southeastern Brazil appears more evident on intraseasonal time-scales. Nevertheless, the opposition in the signal of the extremes in the two regions was observed in less than 50% of the cases. The seesaw feature is much less evident on interannual timescale. Furthermore, less then degrees of the extremes occurred simultaneously on intraseasonal and interannual timescales.


    An Observational Analysis of Decadal Variations in the Madden-Julian Oscillation

    Charles Jones 1 and Leila M. V. Carvalho 1, 2

    1 Institute for Computational Earth System Science

    University of California, Santa Barbara, California, USA

    2 Dept. of Atmospheric Sciences

    Institute of Astronomy, Geophysics and Atmospheric Sciences

    University of São Paulo, São Paulo, Brazil

               
        The Madden-Julian oscillation (MJO) is the most prominent mode of tropical intraseasonal variability. The MJO influences the variability of the monsoons in Asia-Australia and Americas. This modulation has been shown to affect rainfall as well as extreme events in many locations around the world. Furthermore, since the MJO involves intense tropical convective heating anomalies, tropical-extratropical interactions are known to be significant during its life cycle. Therefore, some studies have detected noticeable impacts on the skills of weather forecasts on medium-to-extended ranges as well as in potential predictability in the tropics and extra-tropics.

                In a previous study, Slingo et al. (1999) investigated interannual variations in the MJO and relationships with ENSO. Using indices of equatorial intraseasonal activity in the zonal winds (200 h-Pa) derived from NCEP/NCAR reanalysis and Hadley Centre climate model simulations forced with observed SST, Slingo et al. (1999) found evidence of a positive trend in intraseasonal activity since the mid 1970's. This led Slingo et al. (1999) to speculate that the MJO may become more active as tropical oceans experience prolonged warming.

                In this study, we propose an index that specifically characterizes the occurrences of MJO events and their amplitudes from 1958 to 2003. Such index does not have any apparent spurious trends or discontinuities due to changes in observational sampling in the NCEP/NCAR reanalysis. More importantly, contrary to Slingo et al. (1999) results, we have not found statistically significant evidence that the MJO has experienced a steady increase in activity since the mid 1970's (i.e. a linear trend). In contrast, our results do indicate that the MJO exhibits considerable decadal variations but point to a different and much more complex behavior than the one suggested by Slingo et al. (1999). The MJO shows substantial changes in its occurrence and intensity and fluctuates between regimes of high and low activity. These changes in regimes exhibit characteristic time scales of about 8 yrs and 14 yrs.


    Intraseasonal variability during SALLJEX

     J. Nogues-Paegle (University of Utah), C. Saulo and C. Vera (University of Buenos Aires)
    B. Kirtman and V. Misra (George Mason University and Center for Ocean-Land-Atmosphere Studies)

              
             Intra-seasonal (IS) variability during the SALLJEX is examined to quantify the extent to which numerical integrations are able to sustain it with a focus on the source region of the western Pacific and its manifestation over South-America. Seasonal forecasts of the  COLA model with prescribed SSTs, 15 day integrations of the MRF model operational in 1998 NCEP-NCAR  (CDC Reforecast experiment),  NCEP-NCAR Reanalysis and GDAS ( NCEP operational Global Data Assimilation System) as well as  GPCP (Global Precipitation Climatology Project)  daily merged precipitation data from NASA Goddard Space Flight Center are used in the discussion. The SALLJEX period presents a well defined easterly propagating event from Dec 1 2002 - Jan 15 2003. The COLA model weakly reproduces the easterly propagation. The Re-forecasts lose it by day 5 and damp the IS signal by day 15. The two models fail to capture the IS variability over SESA, which is clearly present in the GPCP data. The evolution of low level jet events in central South America is poorly represented in the model integrations.


    Local forcing and intraseasonal modulation of the South American summer monsoon: soil Moisture, sea surface temperature, and topography

    Alice M. Grimm

    Federal University of Paraná, Brazil

            
               The El Niño/La Niña impact on the summer monsoon in Brazil is not adequately assessed through seasonal analysis because it shows significant subseasonal variations. Grimm (2003, 2004) shows that     there are abrupt changes of anomalies within the summer monsoon season, suggesting the prevalence of regional processes over remote influences during part of the season.

                El Niño events impact on precipitation by enhancing or suppressing the mechanisms that produce rainfall. In the case of the summer monsoon, the driving mechanism is the establishment of a continental heat low and a thermal contrast between the continent and ocean, that brings about circulation anomalies. These anomalies provide enhanced moisture convergence, which leads to moistening and destabilization of the troposphere and thus, to enhanced convection. It seems that the surface temperature anomalies brought about by drought in spring (0) of El Niño events in the highlands of southeast Brazil set up the conditions for circulation anomalies that enhance convection in central-east Brazil during part of the monsoon season, especially January (+). This is the only month within the summer season of El Niño events in which monsoon-type circulation anomalies, associated with a regional anomalous heating, are enhanced. Topographic lifting effect in southeast Brazil may also be important in enhancing ascending motion, low-level convergence, and therefore, cyclonic anomaly with moisture convergence in this region. The more vigorous latent heat release can also reinforce the cyclonic circulation in the lower troposphere over that region.

                Influence function analysis shows that while the anomalies of circulation over southeast Brazil in spring of El Niño years are probably due to remote influences from the tropical East Pacific, those in January are probably due to local influence.

                During La Niña events the circulation, precipitation and temperature anomalies show opposite sign to those observed during El Niño episodes, but the intraseasonal variation is also present during the summer.

                The probable role of regional processes is assessed through several simulations with the regional climate model RegCM3, in which the influence of given conditions for soil moisture, topography, changes in parameterizations, and even short term changes in the Atlantic SSTs are tested.

                The results indicate a significant role of the soil moisture in setting up temperature anomalies and circulation anomalies that might explain the observed intraseasonal changes, as hypothesized by Grimm (2003, 2004). Intraseasonal changes in the Atlantic SST, off the southeast coast of Brazil, do also exert significant influence on regional rainfall. A very interesting effect of the orography in central-east Brazil on the monsoon circulation and precipitation is disclosed by sensitivity experiments with flat terrain in this region. This is a new aspect, for up to the moment the studies on the influence of orography on the South American climate have been focused on the role of the Andes Mountains.

                  

    Intraseasonal variability and model forecasting skill in the monsoon area of S. America

    Pedro L. Silva Dias  (1)

    J. Aravequia (2)

    M. Schneider (1)

    (1) Institute of Astronomy, Geophysics and Atmospheric Sciences

    University of São Paulo

    (2) Center for Weather Forecasting and Climate Studies

    National Institute of Space Research

                The motivation for this work is based on the operational experience that model drift (bias) shows persistence and goes through cycles. Operational forecasters frequently ask the question:  why are there periods with much higher model forecast skill? The hypothesis to be explored is that the Intraseasonal Oscillation may have some impact on model forecast skill through changes in the basic state and therefore on the Òerrors of the dayÓ.  This work has been motivated by the goals of THORPEX (www.wmo.int/thorpex) which is an international research programme of WMO aimed at extending the limits of predictability, and at increasing the accuracy of high-impact weather forecasts from day 1 to day 14. In this report it is shown that the CPTEC and NCEP global models forecast skill, measured in terms of bias and mean square error at selected levels, show significant modulation on the time scale of 20-60 days.  Both CPTEC and NCEP global models show similar behaviour.   The prevailing period of the horizontal structure of the model bias is of the order of 5 to 15 days but there are longer periods.  The propagation speed is of the order of 6 to 12 degrees/day. Both CPTEC and NCEP models underestimate/overestimate the geopotential height when and where precipitation is intense in low latitudes.  Both CPTEC and NCEP models underestimate the initial growth of individual convective systems and the decay phase is excessively slow in the models. Both models have a negative bias in precipitation in the Amazonia, Bolivia and southern Brazil and present anticyclonic bias to the east of the NE coast of Brazil. NCEP underestimates the SACZ precipitation and CPTEC overestimates beyond 72 hr.  The wavelet analysis of the model bias indicates significant intraseasonal control (SACZ, NE Brazil Upper Level Low, Bolivian High). The dynamical control of the intraseasonal variability on the model forecasting skill is possibly related to the changes in the basic flow and its effect on the dynamical instabilities.   The fact that significant intraseasonal signal is found in the model forecasting skill provides some clues on how to predict predictability.


    Conclusions of the plenary discussion for PRA-2: Intraseasonal variability in the SAMS region 

    Julia N-Paegle (moderator) Leila Vespoli de Carvahlo (rapporteur)

                The discussion emphasized:


    1) Observational studies:

    i) to evaluate existing gridded and field data (e.g  SALLJEX) to obtain measures of analysis fit to observations, inter-comparison of existing and future reanalysis with and without  field data to obtain best data sets to study IS variability

    ii) to describe remote and local forcing as well as the internal dynamics of ISO. Easy access to vegetation cover data, as well as other measured quantities of fluxes (including carbon ) from towers, best cloud climatologies, precipitation estimates from space (CMORPH) Radar, gages and other data from remote platforms (including the extent of the Antarctic ice) would accelerate progress in this topic. This can be accomplished by establishing electronic pointers to such data sets.

    2) Diagnostic analyses need to pursue subjects related to:

    i) complex scale interactions that modulate  intraseasonal variations from diurnal through inter-annual and decadal, and from mesoscale systems through planetary scale and inter-hemispheric interactions.

    ii)  local versus remote forcing through observational studies that include special surface data and simplified numerical (and theoretical) models .

    iii)  remote contributions to the SA ISO from the tropics and extra-tropics

    iv) evaluation of climate and extended numerical forecasts. Do the GCM models we use have intraseasonal oscillations? Which models do? How does the ISO over South America affect regions outside South America?

    v) there is a need to develop a model of the ISO  and its manifestation over SA to test the veracity of its simulation by different models.

    3)  There is a recognition that ISO evolve on a time varying basic state and this offers opportunities and challenges to develop a numerical simulation strategy. The importance of IS variability in monsoonal rains has been established, and thus simulations of SAM onset need to consider which initial conditions containing IS signals offer best forecasts. Ensemble techniques based on initial states that span different phases of the ISO may not be suitable, and optional methods should be considered. There was general agreement that simulation of the ISO over SA requires a coupled model to properly simulate the ocean-air interaction in the SACZ, sophisticated regional models with adequate land-air interactions to downscale the signal over the highly complex surface conditions of SA, and a variable resolution modeling approach that allows for local processes to feedback in the large scale circulation.

    4)  Extended numerical integrations (from CPTEC and four different model runs) are available and could be diagnosed to examine the veracity of ISO simulations.  For this to be efficiently accomplished, it is desirable to place such runs in an archive that can be easily accessed (such as the UCAR data portal or the NCEP NOMAD archive)


    2.1.3 PRA-3: Interannual and longer-time variability in the SAMS region.


    The main moisture flux sources in SAMS and evolution of the system

    Iracema F.A. Cavalcanti

    CPTEC/INPE

                Results of climate simulations using the CPTEC/COLA AGCM and Eta regional model are analysed to show some features of the South America Monsoon System. The   vertically integrated moisture flux has a seasonal variability, and in the summer season, the sources of humidity to the SAMS are the North and South Atlantic and the Amazonia region. The orientation of the moisture flux changes with seasons, and there is a very distinct flux between summer and winter, that is captured by the models. However, the CPTEC/COLA AGCM shows stronger humidity flux to southeastern Brazil, than the reanalysis, that could explain the excess of precipitation in the southern part of SACZ. Changes during ENSO years are noticed in model results, consistent with observations. The Eta model also shows features of summer moisture flux to the monsoon region and the associated precipitation, and displays also a diurnal cycle of precipitation, temperature and wind flow. Recent experiments using the Grell ensemble convection parameterization on CPTEC/COLA AGCM, improved the climatological precipitation field over Amazonia. Other analysis and implementations are in progress to improve the results over SAMS region.


    Dominant large-scale patterns influencing the seasonal predictability of
    precipitation over South America

    C. Vera (CIMA/UBA-CONICET, Argentina)

    G. Silvestri (CONICET, Argentina)

    B. Liebmann (NOAA/CDC, USA)

            
               The presence of the dominant modes of the coupled atmosphere-ocean system will obviously determine the predictability of any climate forecast on interannual timescales. In the case of southern South America, the El Niño-Southern Oscillation (ENSO) phenomenon provides some skill to forecast seasonal precipitation anomalies particular from spring to early fall (Grimm et al. 2000). Recently Silvestri and Vera (2003) showed that the Antarctic Oscillation, which is largely independent of ENSO, influences interannual precipitation changes over subtropical South America (especially during austral winter and spring). Nevertheless, the response of the precipitation variability over southeastern South America to the activity of the main leading modes of variability of the SH atmospheric circulation and SST anomalies, has not been systematically analyzed yet. The reanalysis datasets like that of the NCEP/NCAR (Kalnay et al. 1996) are instrumental for diagnostic studies of the physical origins of climate variability Higgins et al. 2000). Therefore, the first objective of this work is to describe the relative contributions of the leading modes of variability of the atmospheric circulation and the sea surface temperature in the SH to the precipitation variance over southeastern South America using NCEP/NCAR reanalyses.

                How modes of variability will change under anthropogenic forcing and whether the response of the climate system to anthropogenic forcing will project onto modes of internal variability are some of the climate change key questions. However, in order to address those issues it is essential to assess whether the observed modes of variability are reasonably simulated by climate GCMs. Very recently, climate modeling groups around the world have been charged with performing an unprecedented set of coordinated 20th and 21st century climate change experiments, for the IPCC Fourth Assessment Report (AR4). Therefore, an additional objective of this work is to investigate the ability of the IPCC/AR4 climate models in reproducing the main features of the leading modes of variability SH circulation and their impact on South America precipitation. The variations of the activity of such leading modes on climate change simulations, and the assessment of climate change scenarios of precipitation over South America based on such variations will be subject of future works.

                It is well known that the first leading pattern of circulation variability in the SH is related with zonally symmetric mass transfers between mid- and high latitudes, and it is known among other names as the Antarctic Oscillation pattern (Thompson and Wallace 1998). The second and third leading patterns correspond to the "Pacific-South America" (PSA) patterns (Mo 2000, and references therein). They are characterized by centers of anomalies extended from the tropics and a wavenumber 3 structure at middle latitudes in quadrature with each other. While the source region of the second leading pattern (PSA1), that it can be considered as the response of the SH circulation to the ENSO forcing, is located to the east of the dateline, that of the third leading pattern (PSA2) is in the vicinity of eastern Australia.

                The analysis of the influence of the leading modes of Southern Hemisphere circulation onto precipitation variability over southeastern South America showed that ENSO produces large, reasonably reproducible spatial and temporal shifts in precipitation over tropical and subtropical South America during spring and summer, in agreement with previous works. The AAO has a significant contribution in spring while the PSA2 has it in summer and particularly during fall, when the contribution of ENSO and AAO is negligible. Results indicate that a significant portion of the skill of climate forecast models will not only arise from the ability to forecast the temporal and spatial variability of ENSO and the associated teleconnection patterns into midlatitudes but also from the ability to reproduce the circulation patterns associated with the AAO and the PSA2. It is worth to mention, however, that most of the AAO variability and a significant portion of that of PSA2 are due to atmospheric internal dynamics.

                The analysis of the IPCC/AR4 climate model outputs of the climate of the 20th Century experiment (20C3M) made so far for the period (1969-1999) show that model are able to reproduce some of the features of the SH leading modes although the associated anomalies are weaker and misplaced. The ability of the models in representing the SH leading modes is related with their ability in reproducing ENSO features and the circulation along the subpolar regions of the SH. NASA/GISS model is not able to reproduce a significant response of SH circulation to ENSO while the GFDL model exhibit similar signatures than those observed. The variance explained by the AAO in both NASA/GISS (24%) and GFDL (21%) models is larger than observed (18%), and both models exhibit stronger air-sea interaction at subpolar latitudes associated with the AAO than that observed. Similar analysis are being performed for the other IPCC models trying to identify which are those that better reproduce the SH circulation variability and their influence over South American climate.


    Interannual Variability during summer (DJF)  in Observations and in the COLA model
    J. Nogues-Paegle (University of Utah), C. Saulo and C. Vera (University of Buenos Aires)
    B. Kirtman and V. Misra (GeorgeMason University and Center for Ocean-Land-Atmosphere Studies)
                An assessment is made of the summer  (DJF)  inter-annual variability  of rainfall and low-level winds over South America (SA)  based on the following data sets:  the Climate Prediction Center Merged Analysis of Precipitation (CMAP), Reanalysis 1 and 2 (R-1 and 2) and seasonal runs  of the COLA model forced with observed SSTs. DJF precipitation averages from R-1 and 2 shift southward the Equatorial  center of precipitation over the northern coast of SA and do not reproduce the continental maximum at 50-65W in the deep tropics. The COLA model reproduces to a certain extent this continental maximum. All three estimates have spurious topographic effects over the Andes.  The ENSO dipole is poorly reproduced in the R-1 and 2 rainfall estimates in that the SESA pole is not apparent (except to a certain extent in the COLA model). Instead, all three estimates show an unrealistic maximum over the Andes. The ENSO response is overestimated in the EOF variance partition in the COLA model at the expense of other IA signals.  SSTs anomalies in the South Atlantic (30-50W 20-30S) evoke a credible response in all analyses, except for the SST driven COLA simulations over the SST anomalies. Warm SSTs in this region of the South Atlantic are associated with a belt of 500 mb high pressure (25S) that extends through the Pacific and Atlantic oceans. This is partly a manifestation of a circulation response to ENSO and indicative of   a contemporaneous correlation with Pacific SSTs.


    Inter - El Niño variability and its impact on the South American Low-Level Jet east of the Andes during the austral summer.

    Tércio Ambrizzi and Gyrlene A. M. Silva

    Dept. of Atmospheric Sciences

    University of São Paulo

                The impact of the inter El Niño-Southern Oscillation (ENSO) variability on the South American Low-Level Jet east of the Andes (SALLJ) during the austral summer (defined as the period between December and February) was analyzed. The Bonner criterium 1 was applied to the NCEP-NCAR circulation fields for the period from 1981 to 2003 to identify the SALLJ episodes. The impact of the location of the maximum convection in the Eastern Central Pacific over the intensity and location of the SALLJ was investigated. The composites of the atmospheric circulation over the South American (SA) showed that the moisture transport from the tropics to the extratropics by the SALLJ is influenced by small displacements of the position of the quasi-stationary Rossby waves. The most important contribution of the large scale flow to generate the SALLJ occurred when the atmosphere was influenced by ENSO events. The composites showed that during strong El Niño events the SALLJ is maintained by the eastern trade winds, while the northern trade winds maintain the SALLJ during weak El Niño events. In strong La Niña events and in neutral conditions there is a secondary western flux from the Equatorial Eastern Pacific crossing the SA and contributing for the maintenance of the SALLJ. When baroclinic transients interact with the SALLJ in the region between 20¡S and 35¡S, they modify the localization and intensity of the jet. The displacement of the jet to the south or southeast influences the convective activity over the South Atlantic Convergence Zone. A dry baroclinic model was used in the research. The simulations suggested that changes in the position of the convective forcing over the equatorial center- eastern Pacific can influence the extratropical atmospheric response. The SALLJ was better configured when the forcing was located in the Niño 3.4 region.


    Problems in Diagnosing Precipitation Trends in South America

    Brant Liebmann.

    NOAA/CDC

                Problems that arise in the diagnosis of precipitation trends are discussed, with examples from South America. It is noted that trend calculations are much more sensitive to small errors than are calculations of interannual or subseasonal variability. There are two types of problems. The first comes from incomplete, or sparse data. Although these can cause quantitative errors in the analysis, it is shown that, for the examples considered, the tendencies are qualitatively correct despite the errors. Errors in the collection and recording of data, which add to the uncertainty, are not considered. The second, more serious problem, is that of the choice of record length from which to calculate the trend. Most studies are done by using the longest record available. In many cases, however, the tendency can change drastically when a different length of record is used, suggesting that the "tendency" is actually decadal scale variability. It is suggested that the calculation of trend is actually rather meaningless unless it can be associated with environmental change (e.g., land use change, etc.).


    Relative role of Pacific and Atlantic SST anomalies in streamflow variability in the São Francisco River.

    Andrea Cardoso

    P. L. Silva Dias

    (1) Institute of Astronomy, Geophysics and Atmospheric Sciences

    University of São Paulo

                The objective of this work is to explore the possibility of improving medium range forecast on river discharge in reservoirs in the São Francisco Basin located in the eastern region of Brazil.  The hypothesis is that there is a significant SST anomaly control on the precipitation regime.  The river data spans the period 1950 to 2003.  Selected SST data is used:  Atlantic Ocean (OA) - 40¡S to 15¡N, 60¡W to 15¡E; Pacific Ocean  (OP)  - 40¡S to  15¡N, 150¡E to 70¡W . Spatial means of the order of (3¡X3¡) in the Atlantic Ocean and (5°X5°) in the Pacific Ocean where determined in order to decrease the dimensionality of the statistical problem. Separate PC's for the Atlantic and Pacific oceans were determined in order to explore the skill of a Linear Regressions Model based on the time series of the PC's as predictors of the river discharge. As a consequence, the fitted forecasting model - and perhaps the model structure itself - may also need to be modified with each successive month that passes. Updating forecasts of future flows requires that the SST scores be updated every month, even if model parameters and model structure are not.   The main conclusions are:


    (1) PC scores of SST anomalies in the Atlantic Ocean can be used to predict natural mean monthly inflows to the Sobradinho Reservoir;

    (2) Correlations between natural monthly inflows and Atlantic Ocean SST anomalies declined slowly as the forecast lead time increased;

    (3) Correlations between Sobradinho inflows and PC scores derived from Pacific Ocean SST anomalies were much smaller, but were still useful when used together with Atlantic Ocean SSTs to obtain predictor variables;

    (4)    The better correlation between reservoir inflows and Atlantic SSTs is to be expected because of the relative proximity of that ocean as a source of heat and water vapour entering atmospheric circulation.


    Conclusions of the plenary discussion for PRA-3: Interannual and longer-time variability in the SAMS region

    P. Silva Dias (Chair), R. Terra (Rapporteur)

  • Need to define metrics for evaluation of the beginning of the monsoon season. Is it possible to define monsoon start from the moisture flux point of view  ( a box in Central Brazil is a reliable measure or the flux along the northern boundary? Or is the moisture flux at 10S? Charge: to define the metrics and explore its use in diagnostic and operational work.

  • Regional moisture recycling needs to be explored in the context of the memory of the system through soil moisture and the vegetation control on surface processes;

  • There are indications that coupling the meteorological model with other components of the climate system (ocean, biosphere) are important to understand the interannual to longer time scales in the SAMS region;

  • Some SH teleconnection patters are well studies such as the AAO, PSA1 and PSA2. However, there are other SH modes  (e.g., summer connection between SACZ, SPCZ, SICZ) which are much less understood;

  • Important issue on predictability:  models reproduce the signal associates to major teleconnection patterns but the predictability is very low.

  • What is the role of the interdecadal variability in the interannual and shorter time scales?  E.g.: are the weak/strong ENSO cases uniformly distributed in time? Is the probability of occurrence of extreme events associated to the longer time scales? Proxy data such as glacier ice cores, tree rings, speleothemes should be used to identify the decadal and longer time scale control.

    There is a strong concern about the metrics for classification of the ENSO events (CDC) and the potential impact of the metrics on the analysis on long term series in order to explore the longer time variability.

    2.2 MESA modeling issues

    Summary of MESA modeling related activities discussed in plenary session of VMP8

    Iracema F. A. Cavalcanti

    INPE/CPTEC

                In the plenary session, the discussions were mainly related to the choice of common topics to be explored on modeling the North and South America Monsoon systems. The main idea is to integrate field activities, modeling research and operational forecasting, in MESA, NAME and VOCALS.  A list of topics with subjects presented on the workshop was proposed to start the discussion. The list should be reduced to a few topics, in order to focus the activities, and should be uniform for MESA, NAME and VOCALS.

                The list of suggested topics was: Diurnal Cycle; Low Level Jets; PBL Processes; Orographic Effects: Coastal Winds, Precipitation; Mixed-Layer Processes; Low Level Cloudiness; Air-Sea Interactions; Air-Land Interactions; Warm Season Precipitation: Remote vs. Local Impacts; Regional Climate Change Assessment; Resolution Issues: Interactions with SSTs other than ENSO; Extra-Tropical Interactions; Predictability Assessment; Seasonality in Predictability; Low Frequency Modes; Prediction Assessment; Fresh Water Flux/Salinity Issues; Hemispheric Interactions; Scale Interactions: Temporal and Spatial (TransientsÉ); Convergence Zone Processes (SACZ, ITCZ); Impact of Land Use Changes; Seasonal Predictability in a Changing Climate; Extreme Events; Intraseasonal Variability; Interactions with ENSO

                There were several suggestions of themes:

    1. Drought prediction in the Americas, which was changed to droughts and floods in the Americas.

    2. Diurnal cycle, and the hydrological component, Cumulus convection and surface heating budgets. Simulation of nocturnal precipitation related to diurnal cycle.

    3. Life cycle of NAMS and SAMS. Prediction of onset of monsoon. There is a need to identify metrics to the Monsoon onset.

    4. Prediction of South Atlantic SST. There are social impacts in the prediction of droughts or floods in the region. The model responses are related to the interaction of air-sea. A more complete theme could be Prediction of SST surrounding Americas (air sea interaction).

    5. There was a suggestion of a theme in environmental problems (droughts, fire, aerosols). Aerosol is important in the transition season.  There is a need to look at surface process, deforestation.

    6. The climate change issue can be included on other themes.

              
             In the final discussion, the 5 selected themes were:


    Diurnal cycle, Droughts and Floods, Life cycle of Monsoon, Predicting SST surrounding Americas, Data assimilation.

    Improvement in model predictability in the monsoon area of S. America:  impact of a simple super-model ensemble

    Pedro L. Silva Dias

    Demerval S. Moreira.

    Institute of Astronomy, Geophysics and Atmospheric Sciences

    University of São Paulo

                Motivated by the SALLJEX Intercomparison Program in 2003 and the THORPEX goals to improve predictability through the proper combination of numerical weather forecasts produced by a large set of models, we have explored the potential predictability associated with the numerical products available in S. America.  There are several models outputs currently available of regular basis in S. America. The global models are: CPTEC, NCEP, JMA, ECMWF, UKMO, CMS) and the regional models are: CPTEC(ETA), INMET (DWD), MASTER (BRAMS), SIMEPAR (ARPS, BRAMS), UFRJ (MM5, RAMS), UFSC(ARPS), FURGS (BRAMS), CEMIG (MM5),  LNCC (ETA),  UBA (ETA, LMD, RAMS), Univ. Chile (MM5). Approximately 14 models outputs are available on a daily basis.   The question is: how can we combine several forecasts in an optimal way? A possible solution is based on basic concepts of data assimilation. The objective it to  combine the several forecast through the optimization problem based on the cost function: T=    · (Ti-Bi)/MSEi, where  Ti  is the forecast provided by the ith model Bi is the ith model bias and MSEi is the ith model mean square error. However, the model bias and MSE need an averaging period and how long is this period? Two years is the typical length for MOS. From the point of view of practical applications one should consider much shorter periods: 10, 15, 20, 30 days?  Given that a strong intraseasonal signal has been detected in the model bias it is suggested that a shorter period may lead to stable results.  The preliminary experience with the models available in S. America indicates that 15 days provides a rather stable statistical measure.  It is concludes that the simple procedure based on data assimilation principles was quite successful and the results are routinely available at the MASTER Laboratory homepage (www.master.iag.usp.br). Future implementations are based on the optimal choice of the averaging period for computing bias and MSE. One possibility is to include longer time scales impact on model error (e.g., interannual). However, about 70% of the potential result has already been achieved and we need to improve 30%.  This experience has been quite successful not only in terms of providing a realistic statistical estimate of the optimal forecast up to 7 days but also in terms of the exchange of  experience among participating groups.


    Evaluation of Nested Regional Model Ensemble Climatology for South America:     Annual Cycle, Interannual variability, and Rainy Season Onset

    Angie Seth, S. Rauscher and B. Liebmann

                Multi-year ensemble integrations performed with RegCM3 nested in the ECHAM global atmospheric model and in NCEP/NCAR Reanalyses are analysed with emphasis on simulated annual cycle, interannual variability and rainy season onset.  The regional model captures the gross features of the evolution of warm season rains and is quite good in the transition and dry seasons.  However, during the primary warm season rains, the Amazon region exhibits a dry bias ( similar to that seen in many models) and the subtropical continent shows excess precipitation. This bias results in a clear semi-annual cycle in the Amazon basin, which is not observed.  The error in the annual cycle also affects the inter-annual variability in the southeast.  Wetter than observed, the region does not respond as observed to inter-annual variations in tropical Pacific SST.  Rainy season onset is well simulated in the Northeast and the Southeast, and improves upon the driving GCM, however the length of the season is not improved.  

    A sensitivity test is performed to the choice of convection scheme.  The standard RegCM employs the Grell mass-flux, with single cloud updraft/downdraft and no lateral mixing with the stability based closure.  The sensitivity test employs the Emanuel scheme, which is based on a bouancy mixing hypothesis. The results from the sensitivity test show substantial improvements in the large scale circulation including SLP, low level winds, and moisture transport into the continent from the Atlantic Ocean.  The annual cycle of precipitation is also much improved in the Amazon, and also in the Northeast and Southeast regions of the continent. 


    The South Atlantic Circulation and Climate
    R. P. Matano

    College of Oceanic & Atmos. Sc., Oregon State University, Corvallis, OR 97331-5503

                This presentation discusses the low frequency variability of the South Atlantic circulation from observational and modeling results and links it to climate variability. Analysis of altimeter data indicates that the interannual variability of the basin-scale circulation is characterized by a dipole mode with peaks of opposing phases in the subtropics and the western subpolar region. The temporal structure of this mode was characterized by a strengthening of the oceanic circulation in the subtropical gyres from 1992 to 1996 and a weakening from 1996 to 2003. The subpolar expression of this dipole is characterized by a weakening of the recirculation cell over the Argentinean basin from 1992 to 1996 followed by a strengthening from 1996 to 2003. The sea level pressure and the wind stress curl show basin scale variations similar to those of the sea surface height anomaly (SSHA). The lowest mode of sea level pressure variability represents changes of the anticyclonic circulation over the subpolar basin. The structure of the wind stress curl shows changes similar to those of the SSHA. The good correspondence between the dominant modes of variability of sea surface heights and sea surface temperatures (SSTs) indicates that temporal and spatial structure of the South Atlantic SST dipole reported in previous studies, might be highly influenced by the inter-decadal variations of the basin scale, geostrophic circulation of the ocean.


    Interannual and interdecadal variability observed in the tropical glacier studies

    Edson  Ramirez

    Instituto de Hidráulica e Hidrología-IRD, CP. 699, Universidad Mayor de San Andrés, La Paz, Bolivia,

                Three new ice-cores were drilled in the tropical and subtropical part of the Andes. The first one, 137 meters long, was extracted from the ice sheet of the Illimani (16°37'S, 67°46'W), a 6350m high mountain in the southern part of the Cordillera Real, Bolivia. The second one, 40 meters long, stems from the Sajama (6542 m, 18°06'S, 68°53'W), situated in the western part of the Bolivian Highlands. The third one, 56 meters long, was obtained at the summit of Chimborazo  (6280 m, 1.6¡S, 78.6¡W) in Ecuador close to the ENSO influence region (Pacific Ocean). The isotopic composition of the ice as well as a number of various tracers (dust, calcium ions, and electric conductivity) were analyzed in the laboratory and used for a year-by-year dating of the ice. A detailed comparison between the isotope signal of five different Andean high altitude sites (Huascarán, Quelccaya, Illimani, Sajama and Chimborazo) shows a nearly identical inter-annual variability in the 20th century. An Andean Isotope Index (AII) constructed from these Andean sites has been compared with general circulation model simulations [Hoffmann et al., 2003]. The results suggest that the Andean high altitude records are primarily controlled by precipitation variability over the Amazon basin. The Illimani ice core record covers approximately the last 18 000 years BP and the analysis suggest that Illimani/Huascaran  water isotope history can be explained in terms of a wetter/cooler conditions during glacial times to drier/warmer conditions in the Early Holocene [Ramirez et al., 2003].

                Chacaltaya Glacier is located in the Cordillera Real of Bolivia (5370 m, 16°S). Extending on an area of 0.06 km2, this glacier is representative of small glaciers of the outer part of the Tropical Andes. A strong correlation exists between Chacaltaya mass balance and the reanalyzed NCEP-NCAR temperature at 500 hPa. Since temperature integrates all the fluxes at the glacier surface, it appears to explain the best part of the mass balance variance at year scale. Nevertheless, at local scale and at short-time scale, as demonstrated by the energy balance analysis conducted on the nearby Zongo Glacier, radiative fluxes and precipitation are the most important factors controlling the glacier evolution. Chacaltaya glacier accurately reflects the climate change occurring in the tropical Pacific-South American domain. The higher frequency and the changed spatiotemporal evolution of El Niño since the mid 1970's together with a generally warming troposphere over the tropical Andes, explains the recent dramatic shrinkage of glaciers in this part of the world.

    Conclusions of the plenary discussion for MESA modeling issues

    Leila Carvalho (Chair) Anjie Seth, (rapporteur)

                This discussion of MESA modeling issues was organized first by timescales (diurnal, intraseasonal, interannual, decadal and climate change)  then by modeling strategy (assessment and hypothesis testing), and finally by specific modeling activities (methodological improvements, data assimilation, and parameterization development).

                Appropriate simulation of the diurnal cycle of convection is critical for the production of user relevant prediction information including subseasonal climate statistics such as extreme temperatures and rainfall.  Errors in the simulation of the diurnal cycle can lead to biases in the annual cycle, in the regional response to remote SST forcing, and in local response to anthropogenic climate forcing.  Thus, careful examination of the diurnal cycle of convection and related surface hydrology in a number of climate models is an important task for MESA. A range of models should be examined including models used for short-term weather forecasts and those employed for seasonal prediction and climate change studies.  Evaluation of the simulated diurnal cycle could be performed with as few a 4 times daily data (6 hourly), however, if shifts are being examined it is possible that hourly data may be needed.  There are also questions regarding in what terms the diurnal cycle should be defined, and the metrics used for evaluation.  Consideration of specific metrics will have implications for the extension of monitoring networks (e.g., radar or automatic stations for the provision of pdfs).  SALLJEX field observations are a critically important source of verification data for evaluations of simulated diurnal cycle as are TRMM and CMORPH.

                It has been shown that on intra-seasonal timescales local land surface feedbacks can be important and may be associated with some predictability.  Thus in addition to remote forcing of IS variability (e.g., MJO and PSA), it is important to examine local land surface feedbacks and both regional and global models can be used to test such hypotheses. For example, can an IS signal seen in surface flux tower data be replicated by the models? Does local variability have global effects? Can we use regional models to understand IS fluctuations in surface data and improve the model simulation? By moving beyond the mean to examining higher order statistics in the models our understanding of both problems and potential for improvement will be expanded.   These issues should be examined in a number of models (including simple models) with analysis performed in a probabilistic manner.  SALLJEX observations will provide the reference for understanding and evaluating tropical-extratropical variability in the models.

                MESA modeling interest on inter-annual timescales begins with the simulation of ENSO teleconnections in South America. While many models perform well in the region of Nordeste, when forced with observed SST, the predictability of tropical Atlantic SST is poor, and fundamentally, the two-tier approach (prescribing SST in an atmospheric model) is flawed due to resulting inconsistencies in oceanic surface fluxes. Thus it is recognized that coupled ocean-atmosphere models are the appropriate tools for seasonal prediction (and climate change) studies, though much evaluation and improvement is needed of these models, including transient activity, horizontal and vertical resolution, and evaluation of derived variables. Finally, and crucially important, is the issue of developing clear metrics for the evaluation of the models on inter-annual time scales.  How is the onset of the monsoon to be defined? Is monitoring needed of moisture flux, or other elements of the surface water balance?  Are there other metrics which are more appropriate for the South American monsoon system?

                On decadal and century time scales, long term trends provide a critical background state upon which interannual variability is manifested.  Thus examination of trends and projections of future climate scenarios is important not only for planning but also for understanding the context of present day variability.  There exist numerous global model and multi-model data repositories for present day (historical: CMIP, AMIP) climate integrations and future scenarios (IPCC) which are available in standardized data formats. These global models should be examined in detail for the South American region (and its influence globally).  Analysis of these model data will provide a baseline which would guide directions of further model development and experimentation.

                Given consideration of these timescale related issues, the MESA Modeling Strategy should include both Model Assessment and Hypothesis Testing.  The model assessments should be comprised of (1) regional model assessment during SALLJEX, (2) seasonal simulation assessment, (3) IPCC AR-4 model simulation assessment, (4) Predictability studies for SALLJEX, (5) Diurnal cycle assessment (6) assessment of extremes.  The hypothesis testing should include but may not be limited to (1) Synergy between SALLJ and MCS, (2) Mechanism for the NW Argentina heat low, (3) Sensitivity to soil moisture, (4)  Coupled simulation for SACZ, (5) Basic mechanisms of diurnal, (6)  Coupled simulations to understand multi-scale framework for SAMS variability    [Not necessarily hypothesis driven, are models good enough], (7) the nature of convection on diurnal timescales, (8) changes in the basic state related to decadal variability, and (9) local and remote influence of SAMS globally.


    2.3. MESA Applications

    La Plata Basin, Climate and Hydrology Report

    Pedro L.  Silva Dias