SOUTHERN HEMISPHERE THORPEX SCIENCE PLAN
John L McBride, Kamal Puri, Peter Steinle
Bureau of Meteorology, Australia
Ian Simmonds
University of Melbourne, Australia
Neil Gordon
MetService New Zealand
Michael J. Uddstrom
National Institute of Water and Atmospheric Research, New Zealand
Warren J. Tennant, Eugene Poolman
South African Weather Service
Manoel A. Gan
National Institute for Space research (INPE), Brazil
Christian P. Rousseau
Direccion Meteorologica de Chile, Chile
Arona Angari
Meteorological Service, Cook Island
The current document is a draft Science Plan for Southern Hemisphere THORPEX. It
was initially written by scientists from the Australian Bureau of Meteorology (lead
authors Kamal Puri, John McBride) and reflected the views and experience of
Australian scientists. Subsequently contributions were obtained from scientists from
South America, South Africa, New Zealand and Cook Islands. The document in its
current form covers a large number of ideas and phenomena from a Southern
Hemisphere perspective. The next stage is to have a detailed discussion on the
document that should lead to the development of a science plan which provides an
agreed focus on a subset of these strategies and phenomena.
1. Introduction and Background.
THORPEX is an international research programme to accelerate improvements in the
accuracy of 1-day to 2-week high impact weather forecasts. It was established in May
2003 by the Fourteenth World Meteorological Congress (Resolution 12) as a ten-year
international global atmospheric research and development programme under the
auspices of the WMO Commission for Atmospheric Sciences (CAS). THORPEX is a
component programme of the WMO World Weather Research Programme (WWRP).
The primary scientific aims are to extend the skill that has been achieved in recent
decades by global operational Numerical Weather Prediction modelling out to the
two-week time frame, and to shift the focus of weather forecasting and verification
procedures towards high impact forecasts such that the benefit to society is optimised.
1
As with the collection and dissemination of real-time meteorological and
oceanographic data, and as with the development of numerical weather prediction
itself, the task of meeting these challenges is considered as being beyond the
capabilities of any single nation or continent. The focus of THORPEX is therefore to
establish an international framework specifically to address global weather research
and forecast problems whose solutions require international collaboration among
academic institutions, operational forecast centres, and users of forecast information.
This includes engagement with other international programmes within the World
Meteorological Organisation (WMO), the International Council of Science (ICSU)
and the Intergovernmental Oceanographic Commission (IOC).
Research Objectives for THORPEX are developed under four sub-programmes:
Predictability and Dynamical Processes
Observing Systems
Data Assimilation and Observing Strategies
Societal and Economic Applications.
A detailed International Science Plan has been developed which expands on the
scientific challenges associated with each sub-programme. The specific research
objectives (taken verbatim from the International Science Plan) are reproduced in
Appendix I of this document. The four sub-programmes reflect an infrastructure
dividing into subcomponents the challenge of improving the understanding, prediction
and mitigation strategies associated with high impact weather events on the 1-day to
2-week timescale. The inclusion of the fourth subprogram reflects the importance of
societal impacts and the development of end-to-end systems in the forward
development of weather forecasting research.
Throughout the Science Plan, the scientific methods described emphasise advances in
scientific understanding of the processes affecting the development of high impact
weather events as well as the development of advanced data assimilation and
ensemble prediction systems. Another underlying theme is the development of
Interactive Forecast Systems, which are tuned for end-users through the strategy of
targeted observations as input to the analysis -prediction systems. A pdf or Word file
copy of the International Implementation Plan is available at the THORPEX website
http://www.wmo.int/thorpex/
The Organisational Structure of THORPEX includes a WMO-CAS International Core
Steering Committee (ICSC), and a THORPEX International Programme Office
located at WMO in Geneva Switzerland. The implementation of demonstration
projects, is developed through a number of THORPEX Regional Committees. To
date these projects come under the description of either THORPEX ObservingSystem Tests (TOSTs) or THORPEX Regional field Campaigns (TReCs).
There are three Regional Committees for the North American, European and Asian
components of THORPEX. Each of these Regional Committees is in the process of
developing a Science Plan and an Implementation Plan. The way THORPEX is
evolving at this initial stage, it appears that the majority of THORPEX research will
be carried out under the auspices of Regional Committees. There has been an
expectation from the ICSC that a Southern Hemisphere THORPEX Regional
Committee be established and that a Southern Hemisphere Coordinated Plan be
developed. The purpose of the current document is to provide input to the
2
development of an Implementation Plan for Southern Hemisphere THORPEX,
involving scientists from South America, Australia, New Zealand, Southern
Hemisphere Africa and Southern Hemisphere nations in the Pacific, Atlantic and
Indian Oceans.
A two-day Workshop was held in Melbourne during 28-29 November 2005. The
Workshop’s aim was to initiate discussions towards the establishment of a Southern
Hemisphere THORPEX Committee and to determine initial research foci and plans
for international cooperation within the Southern Hemisphere on conducting research
and observing system test projects. The outcomes and documents from this
Workshop were presented to a Meeting of the ICSC immediately following the
Workshop and have been incorporated in this plan.
The following Section of this document presents a brief rationale for establishing a
Regional THORPEX Committee based on the Southern Hemisphere. After that are
four Subsections presenting ideas and plans for research under the four subprograms
of THORPEX. The description in each subsection covers a large number of ideas and
phenomena from a Southern Hemisphere perspective.
2. The Rational for a Southern Hemisphere regional focus for THORPEX.
Because a large percentage of the Southern Hemisphere is covered by oceans, the
various countries of the hemisphere have strongly overlapping problems associated
with the monitoring and forecasting of weather and climate. For the same reason
(greater area of ocean) the scientific questions and the technological challenges tend
to be similar around the Hemisphere. This is reflected in the fact that the American
Meteorological Society has Southern Hemisphere Meteorology as one of its specialist
subcommittees and holds a regular series of conferences with Southern Hemisphere
Meteorology as the theme. Thus, by default, Southern Hemisphere Meteorology is
treated as a research subfield of Meteorology.
Besides the logistical challenges of atmospheric data collection, assimilation and
analysis over a hemisphere mainly covered by ocean (although these have been
reduced by increasing use of satellite data), there are also large differences from the
Northern Hemisphere in terms of the meteorology on the 1-day to 2 weeks timescale.
This is partly due to the weaker orographic and continental forcing of the Southern
Hemisphere flow so that stationary long waves have a very much weaker influence in
terms of contribution to eddy transport of heat and momentum (see for example,
Karoly et al., 1998). It also manifests itself in the way in which baroclinic instability
operates to govern the development of storm tracks and the locations of midlatitude
cyclonic activity.
This is illustrated in Fig. 1 which contrasts the Eady growth rate (a localised measure
of the potential for baroclinic instability (Hoskins and Valdes, 1990) for the winter of
the two (Northern versus Southern) Hemispheres. The upper panel shows the Eady
growth rate averaged over 40 years (1961-2000 from NCEP reanalyses) for the
Northern Hemisphere winter months of December through January. As is well known
the areas of maximum potential for baroclinic instability are concentrated into two
localised regions on the western side of the two ma jor oceans. The instability is
concentrated here primarily due to the large north-south low-level temperature
gradient in those locations between the north-eastern peninsulas of the cold
wintertime continents and the warm waters of the oceanic western boundary currents.
3
In the northern hemisphere baroclinic disturbances tend to develop in these two
regions and then grow as they move northward and eastward along the Northern
Hemisphere storm tracks (Blackmon, 1976, Wallace et al., 1988).
In contrast the lower panel shows the same field for the Southern Hemisphere winter
months of June through August. Here the potential for instability is not localised at
the western edge of continents. Rather it follows a spiral pattern beginning off the
eastern side of Australia and circling the earth such that it has a double structure
across the Pacific Ocean where it has a subtropical and high-latitude branch.
Synoptically, this is consistent with the tendency for a split jet structure and a high
incidence of blocking in the Pacific Ocean in the winter months.
Figure 1: Eady growth rate at the 780 hPa level, averaged over the period 1961 2000 from the NCEP reanalysis data set. The upper panel is for the northern
hemisphere winter (December-February). Lower is for southern hemisphere
winter (June -August).
Comparing the two panels of Fig. 1, it is seen that the spatial arrangement of the
development of baroclinic systems is fundamentally different between the two
hemispheres. Thus it would be expected that the spatial (and temporal) organisation
should also be quite different for singular vectors (Molteni and Palmer, 1993) and for
sensitive regions (or regions where the quality of the analysis potentially has the
greatest influence on the expected outcome of the forecast, Shapiro and Thorpe,
2004). Consistent with this difference, it is appropriate that predictability studies and
4
data impact studies in the region have a hemispheric bias; and this presents a scientific
reason for the regional focus of THORPEX to be hemispheric.
Further evidence is found through inspection of the core objectives of the
Predictability and Dynamical Processes Research sub-programme in the THORPEX
International Science Plan. The plan discusses the important role played by such
phenomena as excitation of Rossby wave trains and the influence of large -scale
phenomena such as the Mei-Yu front and longer time scale phenomena like the El
Nino Southern Oscillation (ENSO), the Pacific -North Atlantic Oscillation (PNA) and
the North Atlantic Oscillation (NAO). Very little mention or emphasis, however, is
given to the dynamical phenomena and large scale influences that would first spring
to mind for operational forecasters in the Southern Hemisphere such as cut-off lows,
summertime cold fronts, low-latitude incursions of winter-time fronts, interactions
between fronts and the South Pacific Convergence Zone, the influence of the northSouth-running Andes mountains, tropical-extratropical interaction, the Madden Julian
Oscillation and the Southern Annular Mode (Limpasuvam and Hartmann, 1999).
Problems associated with data impact studies also have a hemispheric nature to them.
For example, the Southern Hemisphere countries have traditionally been responsible
for the installation and maintenance of radiosonde networks with the ground stations
located on remote islands distributed across the southern hemisphere oceans. These
stations are expensive to run and maintain such that a Southern Hemisphere
THORPEX priority should be to quantify the impact and required density of
observations across this oceanic islands network. Similarly there is a hemispheric
nature to the large data void regions, such that it would be appropriate for Southern
Hemisphere countries to pool resources to evaluate the potential impact of advanced
sounding systems such as drift sondes.
3. Southern Hemisphere Predictability and Dynamical Processes Research.
The purpose of the dynamical processes research in THORPEX is to “advance
knowledge of the global-to-regional influences on the evolution and predictability of
high impact weather” (Shapiro and Thorpe, 2004).
In the Australian context the major weather phenomena having a high socioeconomic
impact on the 1-day to 2-week time scale are:
•
•
•
•
•
•
•
Tropical cyclones
Summertime cold fronts, responsible for extreme weather leading to
widespread fires
Cut-off lows, responsible for both flash flooding and catchment-scale
flooding of major river systems
Severe convective weather
Fog and low cloud
Frost
Transient blocking situations whereby a sequence of short lived blocking
highs form in the same location such that areas west of the block undergo
extended continuous periods of high temperature and low rainfall while
locations east of the block undergo extended perio ds of anomalously cold
temperature and high rainfall.
5
In southern Africa similar weather phenomena have a high geographically dependent
socio-economic impact, and include:
•
•
•
•
•
•
•
•
Cut-off lows, causing widespread rain and flooding during summer and
extreme cold and snow during winter
Rapid cyclogenesis near the South African coast, causing gale-force winds and
heavy sea swells
Tropical cyclones in the southwest Indian Ocean causing flooding in
Mozambique, Madagascar, Mauritius, South Africa and Comoros
Berg winds causing hot and dry conditions conducive to the spread of
wildfires (grassland and forests)
Severe thunderstorms (hail, wind gusts, tornadoes)
Semi-stationary mid-tropospheric anticyclone leading to heatwave conditions
during summer
Frost
Fog at busy air ports and along national roads
Since the South American continent extends from 10 oN to 55oS, it is affected by both
tropical and extratropical regions weather systems. In the western part there are the
Andes Mountains with north-south orientation that extends continuously from the
southern tip of the continent, around 55°S, to about 10°N with a quasi-meridional
orientation. The most important weather systems in South America are:
•
•
•
•
•
•
•
Cold fronts
Intense extratropical cyclones near the east coast causing intense winds
Upper level cyclonic vortices (UTCV) (cut-off lows), in some cases
responsible for cyclogenesis and frontogenesis
South Atlantic Convergence Zone (SACZ)
Squall lines
Mesoscale convective complexes
The Low Level Jet (LLJ)
Suggested research foci under the Southern Hemisphere Predictability and Dynamical
Processes sub-programme would be the following:
3.1 Tropical cyclones
Tropical cyclones represent the most regular major natural meteorological disaster
affecting the tropical regions of the Southern Hemisphere. As shown in the
climatology in Fig. 2, they affect the longitudinal span from 35E to 140 W, which in
longitudinal extent represents approximately half the hemisphere. The socioeconomic
impact of tropical cyclones is major. For an industria lised country like Australia,
much of the impact is the effect on industry. For example, the mining industries on
north western Australia depend on constant access to the three ports of Port Hedland,
Karratha and Cape Lambert (Fig.3). The dependence is such that industry estimates
put the cost of closure of any one of these ports as being of the order of $3 million
Australian per day. In the event of a cyclone the Port Authority automatically closes
any port immediately it is affected by gale force winds . The climatology of cyclones
numbers along that coastline in Fig. 2 would indicate this occurs at least annually for
each port. Fig. 3 illustrates the vulnerability of these ports to cyclones. The figure
6
Figure 2: The average number of tropical cyclones per year occurring in each 2
degree by 2 degree grid box across the Southern Hemisphere
(http://www.bom.gov.au/climate/averages/climatology/tropical_cyclones/tropical
_cycl.shtml).
Figure 3: The tracks of category 4 and 5 tropical cyclones affecting the mining
ports of Port Hedland, Karratha and Cape Lambert (red squares) over a ten
year period.
7
shows the tracks of all category 4 and 5 cyclones (10 minute sustained winds greater
than 45 m/sec) in that region over a ten year period. The red squares delineate the
location of the three ports.
The tourism industry and population centres based around coastal areas, particularly
on the east and west coasts of Australia, is another example where the socio-economic
impact of tropical cyclones is major. To give an example, cyclone Althea that hit the
east coast city of Townsville in 1971 was accompanied by a 3.6 m storm surge and
caused damage estimated at 147 million dollars (Emergency Management Australia
database: http://www.ema.gov.au/ema/emadisasters.nsf/ ). Across the Pacific Ocean
are a large number of island based nations. The impact of a tropical cyclone landfall
on major islands can totally wipe out infrastructure for the nation and it can take years
(up to decades) to recover. Examples in recent times include the direct hit of
Category 5 cyclone Heta on the small independent state of Niue in January 2004 and
Tropical Cyclone Waka that hit Tonga in December 2001.
The impact of tropical cyclones in southern Africa is predominantly related to loss of
life and damage to infrastructure that isolates impoverished and remote communities.
Tropical Cyclone Eline (Fig. 4) struck southern Africa in February 2000 and flooded
large parts of Mozambique, northern South Africa, Zimbabwe and Botswana (see
http://www.weathersa.co.za/Pressroom/2000/2000FebTCEline.jsp). Several stations
received over 1000mm during February 2000, more than 10 times their climatological
mean. Fortunately, systems like these are relatively rare in southern Africa, but the
scale of the damage nonetheless requires that we improve predictions of such systems.
Figure 4: Tropical Cyclone Eline making landfall in southern Africa in February
2000.
Internationally, tropical cyclones are already an area of research focus under both the
World Weather Research Programme (WWRP) and the Working Group on Tropical
Meteorology Research Programme (WTMRP). The emphasis under THORPEX
should be on predictability studies and on extending the range of prediction out to a
week and longer. Though not yet an operational system, the Australian Bureau of
8
Figure 5: Dynamical ensemble forecasts (from the Australian Bureau of
Meteorology’s LAPS model) of the positions of Tropical Cyclone Ingrid from a
starting time at 1200 UTC 11 March 2005. The clusters of dots represent
forecast positions for each member of the ensemble, out to a forecast of 3 days.
Figure 6: The forecasts of the central pressure of cyclone Ingrid for the
ensembles shown in Fig.5. Each orange dot represents the pressure forecast for
one member of the ensemble. The blue line is the pressure for the control
(unperturbed forecast) while the green line is the mean of all ensemble members.
9
Meteorology (through BMRC) have been conducting real time tests on dynamical
ensemble forecasts for tropical cyclones. An example is shown in Fig. 5 which gives
the forecast positions of each member of the ensemble out to a lead time of three days.
This forecast is a particularly critical one in terms of impact as most members of the
ensemble make landfall, though a large portion retain a westward track and do not
cross the coast line. The spread of central pressure forecasts for each member of the
ensemble is shown in Fig. 6, with a wide spread of central pressure revealing the
uncertain impact of landfall and consequent cyclone filling.
3.2 Cut-off lows
Cut-off lows are responsible for some of the major flooding events as well as severe
wind storm and high sea events in Australia. In the mature stage, a cutoff low is
separate d at the surface from the mid-latitude westerlies by a high pressure system
immediately poleward. Also in this mature stage, the cutoff is characterised by
existing within an upper level (200 hPa) split jet configuration. The prevalence of
cut-off lows in the Southern Hemisphere, particularly the South West Pacific, is
associated with the existence of the climatological split jet stream in winter across that
region, though this association should be explored theoretically and numerically under
THORPEX.
In southern Africa cut-off lows develop over the continent when the surface
anticyclone ridges around the coast, first deepening a mid -tropospheric trough through
cold-air advection and then cuts the system off through warm-air advection. The high
topography of South Africa, with accompanying steep escarpment adjacent to the east
coast, and the warm, southward flowing Agulhas Current play an important role in the
synoptic development of these systems. From 24 to 28 August 1970, one of the
heaviest rainfall events associated with a cut-off low in South Africa was recorded in
East London on the southeast coast, when 855mm fell in just 5 days (see Fig. 7).
Destruction to roads and bridges along this coastline was tremendous.
Over South America two kinds of cut -off lows are present. The first type is usually of
tropical origin and develops over the South Atlantic Ocean and moves the west into
Northeast of Brazil (NEB). The second type develops over subtropical latitudes of
South Pacific, displaces to the east, crosses the South American coast and acts in
south of Peru, north of Chile and Argentina, Uruguay and South Brazil. Normally,
when a cut-off low is close to the NEB coast in the north and west regions of NEB
and Southeast of Brazil intense precipitation occurs, and clear weather occurs in the
southern region of NEB. When a subtropical cut-off low crosses the Andes,
cyclogenesis and frontogenesis can occur over north of Argentina associated with the
Andes effect (Gan and Rao, 1999; Funatsu et al. 2004; Jusevicius, 1999). In some
cases intense rainfall can develop and in some places of South Brazil flooding occurs.
Recent examples of major rain events associated with cut-offs include the following:
20 January 2005 : Localised heavy thundershowers occurred over large parts of the
interior of South Africa and Botswana. The NCEP ensemble system was able to
capture the high probability of this event some days ahead.
29-30 June 2005 : Localised heavy falls over coastal Queensland and New South
Wales. Unrestricted 24 hour totals exceeded 550 mm in several locations. A number
of stations registered their highest ever daily June rainfall totals (24 hours to 9am)
10
Figure 7: Satellite image, observed rainfall and forecast probabilities of heavy
rain for 20 January 2005 over South Africa.
11
over more than 100 years of record. These include Nerang (346 mm) and Oxenford
(290 mm). The town of Tweed Heads recorded its highest daily rainfall (24 hours to
9am) for any month (382 mm).
2-4 February 2005: This event brought severe storms, record rains and record low
temperatures to South Eastern Australia, with approximately 100 weather records
broken. A total of 13 locations (with 50 years or more of weather information) set alltime daily rainfall records. This included Melbourne, with 120.2 millimetres in the
24 hours to 9 am, Thursday 3 February. A further 25 locations in Victoria and
southern NSW had their wettest February day on record. Among them was Dookie
which broke its February record on two successive days, wit h 90.0 mm on 3 February
and 103.4 mm on 4 February. In addition a total of 43 locations in northern Victoria,
eastern South Australia and southern NSW had their coldest February day on record
on either 2 or 3 February.
Closely associated events are rapidly developing mesoscale cyclones that also, when
formed, have the cut-off characteristics of a poleward high at the surface and exist
within a split-jet configuration at upper levels. One of these systems was responsible
for the disaster in the Sydney to Hobart yacht race of December 1998. Of the 115
yachts that contested the race, 71 retired, 6 were abandoned and five sank, fifty five
crew were rescued and six lives were lost. The destruction caused by the storm
encountered by the fleet triggered a ma ssive search and rescue operation involving
numerous personnel from organisations such as the Australian Maritime Safety
Authority, the Royal Australian Navy, the Royal Australian Air Force and Police.
Figure 8: An intense cyclone off the coast of Brazil in March 2004. The cloud
circulation bears spiral configuration and an eye characteristic of a tropical
cyclone. The image is from the MODIS satellite for 26 March 2004.
(http://rapidfire.sci.gsfc.nasa.gov/).
12
Turning to other parts of the hemisphere, in the South Atlantic, an intense cyclone off
the coast of Brazil in March 2004 caused much damage when it made landfall.
Catarina affected the states of Santa Catarina and Rio Grande do Sul between 27 and
28 March 2004, producing intense winds, up to 180 km/h. About 1500 houses were
destroyed and more than 40000 damaged, at a cost of more than US$350 million.
Three people died and seven are still missing. As shown in Fig. 8, from satellite
imagery this system looked like a mature tropical cyclone; many researchers believe
that it was a tropical cyclone with many of the structural features of a tropical cyclone
(e.g. Pezza and Simmonds, 2005), while others regard it as a cut -off low whose
development occurred as a cut-off from a pre-existing baroclinic zone and remained
within the upper level split-jet configuration throughout its lifetime. Regardless of its
classification, it is clearly the sort of system of direct relevance to the aim of SH
THORPEX.
There are examples of cut-off lows being predicted by operational NWP models
several days ahead. This can be seen in the 7-day forecast (left panel) and verifying
analysis (right panel) in Fig. 9. Despite the importance of the systems in the Southern
Hemisphere, however, as yet there have been no systematic predictability studies
carried out for these systems, and many aspects of the dynamics are still not well
understood.
Figure 9: Mean Sea Level Pressure (Solid) and 1000-500 hPa thickness (dashed)
for a 7-day prognosis by the ECMWF operational NWP model (left panel) and
the verifying analysis (right panel) for 1200 UTC 9 July 2005.
3.3 Intense extratropical cyclones
The eastern coast of South America is known as a cyclogenesis zone (Gan and Rao,
1991; Sinclair, 1995). These studies have shown that there are two regions of
preference for the development of cyclogenesis, namely: San Matias Gulf (Argentina)
and Uruguay. When these cyclones have a rapid intensification they can have
significant impact on the population, with the accompanying floods causing loss of
life and property. Recently Dal Piva (2005) has shown that both baroclinic instability
and downstream development can have an important role in the cyclone
intensification. The Andes Cordillera also has an important influence in the initial
formation of some extratropical cyclone cases over the South-American continent
(Gan and Rao, 1996).
3.4 Summertime cold fronts
13
Some of the major high impact weather events in South Eastern Australia have been
bush-fires associated with the passage of cold fronts across the region. Besides the
impact in terms of loss of property, infrastructure and lives, these events also have an
immense social impact in terms of anxiety and trauma for the inhabitants of this
region. This is evidenced by the fact that two of the major events (the Ash
Wednesday fires of 1983, the Black Friday fires of 1939) are part of the history and
culture of the region, and are referred to as such in daily speech and in the press.
As illustrated in Fig. 10, during summer the enhanced northerly (poleward) flow
ahead of the cold front is dry and warm having a trajectory from over the continent
and thus is conducive to fire. In addition at this time of year the thermal contrast
across the front is high, being between hot continental air and cool maritime air; so
the pressure gradients and consequently the wind speeds ahead of the front are
extreme.
Figure 10: Schematic illustration of the conditions conducive to widespread fire
outbreaks ahead of the approach of a cold front. The figure is taken from the
website
of
the
International
Fire
Information
Network
(http://sres.anu.edu.au/associated/fire/index.html).
According to their
description, the figure is adapted from a Report on the Meteorological Aspects of
the Ash Wednesday fires (Bureau of Meteorology, 1984).
The research of Mills (2005) reveals that the majority of widespread fires and deaths
occur on days of the passage of a deep strong cold front. Mills quantified the strength
of the front through the parameter T G, which is the magnitude of the temperature
gradient at 850 hPa calculated over a rectangle over the state of Victoria (the rectangle
coordinates being 41oS – 33oS, 136 oE – 150oE). Using NCEP reanalysis data he
calculated twice-daily values TG for the three summer months. The 30 highest values
are shown in Table 1. As reported by Mills, over the 40 years analysed 188 deaths
occurred associated with bushfires in the states of South Australia, Victoria and
Tasmania. Of these deaths, 80 % occurred on days of the 30 entries in Table 1.
A similar situation (see Fig. 11 for example) exists in South Africa, but is more
common to the winter months when the vegetation is dry. Runaway veldfires
(grassland), fuelled by strong winds continue unchecked and destroy millions of
14
hectares of grazing, killing livestock and gutting farm buildings. These events have a
high impact because strong wind in these areas is uncommon.
15
Figure 11: Strong winds over the interior of South Africa ahead of an
approaching cold-front.
In South America, incursions of cold air to the east of the subtropical Andes are a
typical feature of the synoptic climatology. During winter, cold air incursions may
produce frost events as far north as southeastern Brazil and central Bolivia (Fortune
and Kousky 1983, Hamilton and Tarifa 1977, Marengo et al. 1997, Muller 2004) with
major social and economic impacts. In contrast, summertime episodes, while
exhibiting a similar structure and evolution like wintertime cases, are accompanied by
weaker fluctuations in temperature and pressure, due partly to the smaller seasonal
temperature gradient between mid - and low latitudes and to the stronger incoming
radiation flux. However those fr ontal epsodes are usually accompanied by a band of
enhanced convection and rainfall at the leading edge of the cool air (Ratisbona 1976;
Parmenter 1976; Kousky 1979; Kousky and Ferreira 1981; Garreaud and Wallace
1998). These synoptic-scale bands of organized deep convection produce some of the
heaviest rainfall episodes during summer, mainly over the Central-Western and
Southern Brazilian regions. In some cases this situation has been responsible for loss
of life and property in the Rio de Janeiro State and eastern region of São Paulo State
associated with mudslides.
3.5 South American Low level Jet (SALLJ)
One of the major features of the summertime circulation in South America is the
South American Low Level Jet (SALLJ) East of the Andes. The SALLJ constitutes
one of the components of the South American Monsoon System (SAMS) because it
16
represents a relevant feature of the warm season low-level circulation. The SALLJ has
been related to moisture transport from the Amazon region into the fertile lands of
southern Brazil-Northern Argentina. Recent studies have identified some of the
circulation and moisture transport features of the low level circulation east of the
Andes suggesting an active role of the SALLJ in the position and intensity of the
South Atlantic Convergence Zone (SACZ), and the rainfall and convection at the exit
region of the jet in Southeastern South America (Nogues-Paegle and Mo 1997,
Berbery and Collini 2001, Mo and Nogues-Paegle 2001, Berbery and Barros 2002,
Nogues-Paegle et al. 2000, Cazes -Boezio et al. 2003).
Fig. 12 shows a conceptual model of the SALLJ as a component of the South
American monsoon. It illustrates the moisture transport from the Amazon-Tropical
North Atlantic region towards the southern Brazil-Northern Argentina region at the
exit region of the jet, and the change of direction of the flow from northeast to
southeast once it encounters the Andes. Figure 13 also shows the moisture transport
from the tropical North Atlantic into the Amazon Basin and from the tropical South
Atlantic both converging into the SALLJ mean stream, as well as the development of
convective activity and rain at the exit region of the jet in southeastern South
America. Evapotranspiration from the Amazon forest, the latent and sensible heat
release of the Altiplano and the SALLJ play an important role in the functioning of
the South American monsoon.
Figure 12: Conceptual model of the SALLJ east of the Andes (Marengo et al.,
2004)
While the reanalyses show some success in depicting the struc ture and characteristics
of the SALLJ as well as its variability in time, they have limitations in reproducing
the diurnal cycle or even the interannual variability. The discrepancy between the
diurnal cycle derived from the surface/upper-air observations and the NCEP
reanalyses, indicates the need for a denser network of observations with increased
17
observational frequency so that a more accurate evaluation of the diurnal cycle can be
made. This has been one of the major objectives of the South American Low Level Jet
Experiment (SALLJEX) that took place during the austral summer 2002-2003, as part
of the Variability of South American Monsoon Systems (VAMOS), from the
CLIVAR program.
3.6 Mesoscale Convective Systems (MCS)
Several authors have focused on the South American mesoscale convective
complexes (MCC) and MCS. The classical work by Velasco and Fristch (1987)
presented the first climatology of occurrence of MCC in South America. After that,
Conforte (1997) determined the occurrence of MCC by following the Maddox criteria
while others like Torres e Nicolini (1999, 2002) have used more flexible criteria but
focused on the larger systems. Figure 13 shows a compilation of all these results for
the location of MCS at the time of maximum extent.
Figure 13: Compilation of the MCC in the work of Velasco and Fritsh (1987),
Conforte (1997) and Torres and Nicolini (2002). Prepared by J.C.Conforte.
Typically the initiation of large MCCs has an afternoon and nocturnal character. In
subtropical South America, east of the Andes formation typically occurs at night,
evolving in the following morning until maturity and decay. The causes for the night
formation are attributed to (i) the large scale mountain-valley circulation during nighttime, between the Andes Mountains and the Paraná River valley, providing low level
convergence in the valley (e.g. Nicolini et al, 1987); (ii) the intensification of the low
level jet from the north, during the night, due to the decoupling from the convective
boundary layer, thermal differential heating over sloping terrain or (iii) the
intensification of the thermal low (Chaco or NOA) during Chaco jet events when the
LLJ diurnal cycle has a larger amplitude.
3.7 The Impact of Longer Timescale phenomena
18
The strongest (in terms of contribution to variance) known large -scale coherent
phenomenon acting on the intraseasonal time scale is the Madden-Julian Oscillation
(MJO). This phenomenon has major impact on the Southern Hemisphere tropics at the
2-week time scale, particularly during the Southern Summer. The influence of the
MJO for 2-week forecasts in the tropics is analogous to that of the El Nino Southern
Oscillation (ENSO) for 3-month or seasonal forecasts. Thus even though the time
Figure 14a: Climatological probability based on thirty years data of a tropical
cyclone forming in the region 0 o – 20 o S, 90 o – 135 o E in the next seven days.
Figure 14b: Same as 14a, except the two probability curves are for years when
there is an El Nino event (solid line) and years where there is a La Nina event
(dashed line). Adapted from Leroy (2004).
19
scale of the MJO (ENSO) is 50 days (2 -3 years), its influence is dominant for
forecasts on the shorter time scale of 2-weeks (3 months). The MJO also drives
teleconnections to the extra-tropics (through the propagation of Rossby waves) that
may affect the location of storm tracks and weather systems in the mid -latitudes.
As stated, the focus of THORPEX is on the short to medium term (1 day to 2 weeks).
A demonstration that there is a strong modulation of weather on this time scale by the
current state (or by anomalies) on the inter -annual, seasonal and sub-seasonal time
scales can be shown through simple statistical analysis. As the simplest and most
trivial example to make this point: in a monsoon regime there is clearly a higher
probability in the middle of the wet season of there being a major rainfall event in the
next seven days than there is in the middle of the dry season. A slightly more
complex example is shown in Fig. 14, adapted from Leroy (2004). Part a of the figure
shows the daily probability of a tropical cyclone forming in the next seven days for
the Southern Indian Ocean between 90oE and 135oE, derived from a climatology of all
cyclone tracks in the region over the period 1969-1999. Part b shows the same daily
probability but for El Nino conditions (full line) and for La Nina conditions (dashed
line).
This clear empirical link between the probability of an event on the one week time
scale and conditions on the interannual timescale has not yet been extended to
predictability. That will be one of the themes of Southern Hemispher e THORPEX
dynamical research. Besides ENSO, other large scale slower phenomena that can
affect the 1 to 2-week forecasts include the Madden Julian Oscillation (MJO) and the
Southern Annular Mode (Limpasuvam and Hartmann, 1999). In a recent study Rao et
a l. (2003) have shown that the Southern Annual Mode can influence substantially the
Southern Hemisphere storm tracks. An appropriate emphasis for Southern
Hemisphere THORPEX research would be to investigate the influence of these
background flow regimes on predictive skill.
3.8 The Madden Julian Oscillation
As discussed above, predictions and high impact events in the tropical parts of the
Southern Hemisphere are strongly affected by the Madden Julian Oscillation. An
example of the social impact of the MJO comes from the Southern summer of 20032004. This summer was marked by an active or high amplitude MJO. As a
consequence of this the nation of Indonesia, situated on the equator in the maritime
continent experienced a greatly enhanced active -break cycle in their “musim hujan” or
wet season. This can be seen in Fig. 15 which is a time series for 12 months of
Outgoing Longwave Radiation (OLR) averaged over a rectangular region located over
Indonesia. The only time smoothing that has been applied is that of a 3-day running
mean. Superimposed upon the curve is the climatological seasonal cycle (dashed
curve) with the differences between the 3-day mean curve and the seasonal cycle
coloured purple for negative excursions and yellow for positive. In the tropical
research literature OLR is used as a proxy for atmospheric convection, with the OLR
value of 220 w/m2 marking an approximate threshold, with values below that
representing regions of organised deep tropical convection (see for example Wheeler
and McBride, 2004).
Looking at the OLR scale on the ordinate, using this threshold the Indonesian region
is under dry (non-convective) conditions from approximately April through October
and in convective conditions over the summer months of November through March.
20
The presence of the strong MJO event in this particular summer is manifested in a
negative (convective burst) excursion throughout the month of December, followed
by an almost complete shutdown of the monsoon (positive excursion) through
January, followed by another convective burst through the month of February.
Following the Indonesian press through that season, the lack of rainfall during
January caused major problems for health and agriculture. In particular the period
from January to March was characterised by major outbreak of Dengue fever, with
more than 40000 documented cases of the disease including more than 500 deaths
(ASEAN Disease Surveillance Net, http://www.asean-disease-surveillance.net).
Though links between climate, weather and dengue are poorly understood (Kovats et
al, 2003) it is known that the fever is usually transmitted by mosquitoes. Thus in a
humid climate a period of unusual drought (Aron and Patz, 2002), such as during the
MJO suppressed event of January 2004 would lead to sufficient decrease in river flow
to allow mosquito breeding, and hence could be responsible for the dengue outbreak.
Figure 15: Three-day running mean time series of NOAA satellite observed
outgoing longwave radiation (OLR) averaged for the box (2.5 ON – 10 O S, 100O E –
135 OE) for the period July 1 2003 to June 31 2004. The dashed line shows the
climatological seasonal cycle created by taking the mean and first three
harmonics of the 1979 – 2001 climatology. (Adapted from the web site of
Matthew Wheeler
http://www.bom.gov.au/bmrc/clfor/cfstaff/matw/maproom/OLR/ts.html)
An impediment in using the state of a slow period phenomenon (such as the MJO or
ENSO) in forecasting is that nowcasting can be difficult when some form of time
filtering is necessary to isolate the state of the phenomenon. With ENSO this has
been avoided through the us e of real-time indices such as the SOI and Nino-3 to
quantify the state of the system based on present and past data. A similar
breakthrough has been made for the MJO through the work of Wheeler and Hendon
(2004) whereby a method was developed for diagnosing the state of the MJO in realtime, without using temporal filtering. This result effectively saves a half-period of
the oscillation in terms of lead time for prediction and so makes prediction practical
on the 2-week time scale.
The Australian Bureau of Meteorology carries out real-time monitoring and
diagnostics of MJO as well as experimental statistical predictions for tropical activity
using indices of the MJO as predictors. An example prediction is given in Fig.16
which shows MJO associated anomalies of OLR and 850 hPa wind using lagged
linear regression, with the Wheeler -Hendon MJO indices as predictors.
Despite the major importance of the MJO in terms of high impact events on the 1-day
to 2-weeks time scale, there is still little predictive skill from dynamical weather
prediction models. This is shown in Fig. 17 (from Hendon et al., 2000). Hendon et al.
21
studied systematic errors associated with the MJO by considering numerical weather
prediction models initialised during active phases of the MJO. The forecast model
used was the T62L28 reanalysis version of the NCEP medium range forecast model;
and strong MJO activity was defined as existing whenever either of the two leading
principle components (PC1, PC2) of band-pass filtered OLR exceeded 1. 5 standard
deviations over the five years for which forecast were available. Fig. 17 shows
Hovmöller plots of equatorial average analyses and forecasts for the 850 mb zonal
wind, composited relative to times of maximum positive values of PC2. The
composite analyses running from 14 days prior to 14 days after the time of maximum
positive PC2 are shown in the upper panel while the lower panel shows the composite
forecast initialised at the time of maximum PC2. The analyses depict coherent
eastward propagation at about 5 m/s, which typifies the MJO. In contrast the
composite forecast exhibit no indication of eastward propagation beyond 2 to 3 days.
While being only one result for one model, this figure is indicative of the current skill
Figure 16: Forecasts of OLR anomalies (coloured shading) and 850 hPa wind
anomalies associated with the MJO. The forecasts are carried out by lagged
linear regression with the RMM1 and RMM2 MJO indices of Wheeler and
Hendon (2004) as predictors. The upper panel is the initial state based on
regression of the anomalies with RMM1, RMM2, following (lower) panels are
forecasts for 5, 10 15 and 20 days forward.
22
at maintaining the MJO in most operational models. Given the importance of the
phe nomenon in 2-week forecasts, addressing the dynamics and predictability of the
MJO has to be part of Southern Hemisphere THORPEX.
Figure 17: Composite anomalies (annual cycle removed) of 850 hPa zonal wind
(contour interval 1 m/s) relative to maximum positive value of one of the leading
EOFs of band-pass filtered OLR, averaged from 5oN to 15 oS. The upper panel is
the observed anomalies from time -14 days to +14 days relative to the time of
maximum of the EOF (day 0). The lower panel is forecast anomalies (mean
model error and annual cycle removed) initialised at day 0. From Hendon et al
(2000)
Important modulations of the MJO on extreme precipitation events in South America
have been identified as well. Carvalho et al. (2004) determined that the MJO
modulates intense South Atlantic Convergence Zone (SACZ) episodes with
persistence longer than 3 days. They additionally found that the MJO phase
characterized by suppression of convective activity over Indonesia and enhancement
over the central Pacific increases the 95th daily precipitation percentile over
north/northeastern Brazil, whereas the opposite features are observed during
enhancement of convection over Indonesia and suppression over the central Pacific.
Liebmann et al. (2004) investigated the variability of extreme precipitation events and
associations with the SACZ and the South American low-level jet. They performed
composite analysis of the MJO relative to anomalous precipitation events and
obtained statistically significant variations associated with precipitation downstream
of the jet and in the SACZ.
23
A study of the southern hemisphere energy cycle shows that regional rainfall in
southern Africa and Australia are linked to hemispheric changes in the energy
exchange. Daily characteristics of the barotropic and baroclinic energy states show
how the circulation archetypes vary between wet and dry years (Tennant and Reason,
2005). Changes in the position of the mid -latitude and sub-tropical jets, tropicaltemperate cloudbands and the strength of the Hadley Cells all seem to form part of a
hemispheric teleconnection mechanism. Similar work on the Antarctic Oscillation
(Reason and Rouault, 2005) also shows how weather patterns around the southern
hemisphere may be linked.
3.9 The Antarctic and Southern Hemisphere polar latitudes
The Antarctic and SH polar latitude regions host frequent intense weather events. As a
typical example Fig. 18 shows a synoptic ma p (from ERA-40), with open and closed
and strong and weak cyclones identified (by the Melbourne University cyclone
tracking scheme (Simmonds and Keay, 2000)). What is immediately apparent is the
large number of intense, mobile systems in the very baroclinic subantarctic
environment, and how the influence of these systems can be seen to extend into the
mid latitudes.
Figure. 18: ‘Satellite’ view of the ERA -40 mean sea level pressure analysis at
1200 UTC on 1 July 2002. the contour interval is 5 hPa, and the centre of intense,
closed cyclones is indicated with a red dot.
The complexity of the tracks of cyclones is indicated in Fig. 19 which shows every
fifth track (for cyclones which last at least 24 hours) from the December – February
1999 - 2002 from ERA-40.
24
Figure. 19: ‘Satellite’ view of the ERA-40 mean sea level pressure cyclone tracks
from the ERA -40 reanalysis. Only every fifth track for cyclones which last at
least 24 hours is shown for the period December – February 1999-2002. The
tracks in each of the four years are shown in different colours.
These cyclonic systems impact in a variety of ways on activities in the Antarctic, on
shipping in the southern oceans, and on significant portions of the southern
continents. Improvements in the skill of forecasts of these would be of immense
benefit, although a number of regionally specific characteristics make difficult the
forecasting problem. These aspects include sharp topographic and thermal contrasts,
and the complexity of the interactions between the atmosphere, ocean, continental ice
and sea ice. These regional aspects also make difficult the task of obtaining
trustworthy analyses which adequately reflect the complexity of the thermal
properties and circulation of the region. Until recent times the region has been one of
very poor data coverage, and hence the initial conditions for forecasting have been
known wit h a high degree of uncertainty.
One of the specific regions focused on in the THORPEX International Research
Implementation Plan (February 2005) are the Polar regions. The Plan has drawn
attention (Section 13.1) to the synergies of proposed THORPEX activities and those
planned for the International Polar Year (IPY) (2007-2008). The case has been made
that the polar regions are host to high impact weather events, including spring thaws,
sea ice movement, and severe winter cyclones resulting in strong winds, high seas,
and heavy precipitation. As indicated above these impact on safety, fisheries and
fishery management, and transportation. In addition, climate change in the Antarctic
region has the potential to change the distribution of ice cover, with consequences for
the frequency, location and intensity of cyclonic storms. Our knowledge of the
complex interactions involved with this, and specifically the prediction of intense
storms, is in urgent need of enhancement.
25
The IPY has identified five core obje ctives. These dovetail very nicely with the goals
of THORPEX. However, it should be made clear that while there are many synergies,
the timescales for completion (two years cf 10 years), and the compass, of the two
international programs are quite different. Accordingly, there are a number of
important Antarctic region activities which should be perused under THORPEX, but
which would not be feasible under the IPY. Below we discuss some directions which
should form part of the southern polar region THORPEX activities.
Cyclonic systems and their role in high southern latitude interactions
Cyclonic systems in the subantarctic region play a central role in determining the
weather and climate of the high southern latitudes. In recent times our understanding
of these systems, in what had been a data sparse region of the globe, has progressively
increased, as our theoretical understanding and the quality of subantarctic analyses
has improved.
Explosively developing cyclones (or ‘bombs’) are characterised by rapid central
pressure reduction and dramatic increase in intensity. Such characteristics are
associated with difficulty of prediction and also with serious threats to human life and
property when these cyclones occur off coastal regions, or in shipping lanes. Until
recent years most of the research on these weather phenomena has been conducted on
NH events. However, the availability of reanalysis products and the improvement of
analysis quality in the high southern latitudes has meant that these features can be
now be realistically studied in the subantarctic region. For example, Lim and
Simmonds (2002) and Simmonds et al. (2003) explored the distribution and
seasonality of these features. They also indicated how the criteria for a bomb must be
generalized to account for the strong westerly environment at high southern latitudes.
Leslie et al. (2005) used this methodology to discuss two bombs. One of their cases
was a “supercyclone” bomb that developed to the southwest of New Zealand region
during May 29 to 31, 2004, and was well-captured by the QS scatterometer
instrument. Improvements in analysis quality and forecast accuracy are a key aspect in
predicting the likelihood and evolution of ‘bombs’, and the potential they represent
for harm and destruction.
Polar lows
It has been known since the earliest times by seafarers in the high southern latitudes
that violent small storms could arrive with little warning. Only when the products of
the polar orbiting satellites became available in the 1960s was it appreciated that these
features are actually quite common and one could make sense of their presence in
terms of the broader polar synoptic environment (Rasmussen and Turner, 2003). Our
knowledge of polar lows and mesocyclones has, hence, come almost entirely during
the period for which we have satellite data, as the resolution of 'conventional' weather
charts prior to that time was insufficient to represent them. These features represent a
dramatic example of polar region extreme weather (often assuming ‘hurricane’
strength) and present a considerable challenge to modellers. High resolution and the
representation of non-hydrostatic processes appear to be important in capturing the
characteristics of these features (Yanase and Niino, 2005). Southern polar model
forecasts undertaken within THORPEX should give due c ognizance of the importance
of appropriate spatial resolution.
26
Teleconnections and complex air-sea -ice interactions
The high southern latitudes appear to communicate (in a two-way sense) with the
mid- and low latitudes through a very diverse range of processes and connections.
Among these should be mentioned the El Niño-Southern Oscillation (through the
Pacific -South America pattern), the quasi zonally symmetric Southern Annular Mode
(SAM), and Antarctic sea ice. Past research has highlighted the myriad of aspects in
which the southern polar regions interact with those to the north. For example, we can
mention the works of Karoly (1989), Watkins and Simmonds (1995), Simmonds and
Jacka (1995), Hall and Visbeck (2002), Carleton (2003), Simmonds (2003), Raphael
(2003), Simmonds and King (2004), Parkinson (2004), Yuan (2004), and Turner
(2004). However, new and consistent data sets which are now, or will soon be,
available provide the opportunity to study these teleconnections in more detail and in
integrated ways. One key role which THORPEX can play in these advances is the
appropriate inclusion of state -of-the-art sea ice (and boundary layer) schemes into
high resolution forecast models, to allow expression of the complex atmosphere -ice
interactions which occur on 1 and 2 week time scales.
Cold outbreaks
The Antarctic continent and its surroundings influence greatly the weather and climate
over the mid latitudes of the SH continents. One of the most dramatic examples of this
influence is the occurrence of ‘cold outbreaks’. The most severe of these involve the
meridional or near-meridional transport of cold air from the subantarctic or Antarctica
itself. These are usually of short duration (2-3 days) and tend to be associated with
specific synoptic patterns. They impact significantly on many economic and social
aspects in the affected regions.
Cold outbreaks are relatively frequent over southern parts of the Australian and South
American continents. Given this frequency, it is perhaps surprising that the
mechanisms associated with these events (and their remote forcing) have been
investigated only in a handful of studies. For Australia we can mention, among the
more the recent investigations, those of Perrin and Simmonds (1995), Simmonds and
Rashid (2001), and Jones (2003). For the South American case the papers of Marengo
et al. (1997, 2002), Garreaud (1999, 2000), Vera and Vigliarolo (2000), Lupo et al.
(2001), and Pezza and Ambrizzi (2005a, b) are particularly worthy of mention.
Overall, these papers reflect a steady increase in the understanding and ability to
model these extreme SH events. The events have great impacts on many aspects of
human activities and are an important focus for SH THORPEX. It has become clear
that these events are associated with significant hemispheric organization, but their
regional consequences differ (e.g., through the presence of the Andes). An
enhancement of international collaboration in this area of meteorology, and
specifically the role Antarctica and its environs play in these very influential episodes,
would be an ideal deliverable from THORPEX on the diagnosis and forecasting of
these extreme events.
The Southern Annular Mode and cyclones
Recent research has identified significant trends in the geopotential differences
between mid and high southern latitudes (Thompson et al., 2000). These trends have
been attributed to, variously, natural variability, enhanced greenhouse gas
concentrations, and reductions in Antarctic stratospheric ozone. Regardless of the
27
causes, the index of Southern Annular Mode (SAM) assumes at present a value higher
than it has exhibited for some time, with implications for baroclinicity and the
westerlies in the subantarctic region. This state of affairs is intimately tied up with
cyclone behaviour, as it is known there are strong feedbacks between the SAM and
the activity of the transients (Hartmann and Lo (1998), Limpasuvan and Hartmann
(2000), Lorenz and Hartmann (2001), Rashid and Simmonds (2004), Rashid and
Simmonds (2005)). Given that the SAM sets the environment in which cyclonic
developments occur in the high southern latitude, the further exploration of this is a
most appropriate topic for THORPEX.
Ocean waves in the southern polar region
A new appreciation of the intensity of storm activity in the southern polar regions is
also leading to new perspectives on the consequent wave fields. The disposition of
these have broad implications for marine activities and safety, as well as coastal
locations which face the Southern Ocean. Young (1999) discussed the important role
played by the intense wave generation systems of the Southern Ocean and confirmed
it to be consistently the roughest ocean on earth. Campbell et al. (1994) used Geosat
radar altimeter data to diagnose winter wave fields in the extratropical SH. Their
compilation indicated large average wave heights as far south as the sea ice edge.
(They found at the ice edge in the Indian sector wave height exceed 5 m almost 50%
of the time.) It is an appropriate task for THORPEX to undertake analyses of wave
setup in the southern polar region using the most recent scatterometer data (QS) and
the most sophisticated wave models.
3.10 Real-time monitoring of dynamics and predictability
A suggestion is that under the banner of THORPEX, the Southern Hemisphere
countries set up a publicly accessible web-page to monitor predictability and
dynamical processes in real-time. Contributions to the web-page would include:
•
•
•
•
Storm tracks over the past 4-weeks, calculated by both the system-tracking
and the upper tropospheric eddy variance and eddy covariance methods,
Areas of large model errors (prognosis – analysis) fields for each day over the
hemisphere, for each lead time,
Any results of backward calculations using an adjoint model or other means to
determine the initial source of model errors,
Some form of chart-discussion or synoptic overview describing predictability
aspects, dynamical features and any high impact forecast or events in the past
week.
4. Southern Hemisphere Observing Systems
Observing systems research under THORPEX is carried out in the context that each of
the meteorological services of the Southern Hemisphere countries has an ongoing
need to maintain an observing system such as is necessary to deliver its basic
forecasting and climate service as required by its government. Despite the major
impact that satellite sounding data has on analyses and forecasting in the Southern
Hemisphere, there still remains the tradition of maintaining observing systems on
remote outposts (for example Gough Island, Marion Island, Macquarie Island,
Kerguelen Island, Cocos Island) due to the large regions of ocean that are devoid of
conventional (non-satellite) observational platforms.
28
It is proposed that a cooperative Observing Systems research for Southern
Hemisphere THORPEX take advantage of the ongoing review processes undertaken
by the individual countries in the maintenance and upgrade of their basic observing
systems. In the case of Australia, for example, changing priorities and technologies
have seen major increases in the volume and quality of data from land based radar
systems, whereas there has been a reduction in the volume of data from radiosondes.
It is often difficult to quantify the impacts of these changes in the character of the
observational data base.
It is proposed that the Southern Hemisphere countries work together to carry out
impact studies of the effects of changes in the background observing system on a
hemispheric scale. A collaborative group could be established to regularly discuss and
review available publications and research findings on observational data impacts, to
interpret that information in a SH context, to propose priorities and recommendations
for national networks in the SH and for satellite operators covering the SH, and to
propose further specific questions and studies that might usefully be pursued.
One of the main problems on the southern African continent is the lack of
observations. South Africa is currently conducting impact studies of local in-situ
observations. Results from these studies will be used to develop a strategy for
additional observations in the region. The strategy must be as cost effective as
possible. The optimal placing of observations will be determined using Ensemble
Transform Kalman Filter (see Bishop et al., 2001 and Majumdar et al., 2002). It is
proposed that the strategy form part of the Southern African Observation Experiment
(SAFROBEX). The experiment will consist of a series of field experiments once the
ETKF code has been implemented.
Simple methods have been developed to quantify the impact of individual
observations in data assimilation systems (for example, Seaman 1994); and these
could be extended to assess the impact of various key observational platforms across
the hemisphere. An example of the upper data base accepted by an operational limited
area NWP model over Australia is shown in Fig. 20. Both panels are for the
observation/analysis time of 1200 UTC. The upper panel is for a day early in the
month of July 2005 and shows the normal 1200 UTC conventional upper air (sonde)
data base. The lower panel is for two days prior to the cutoff low development of Fig.
9 (above) and so has additional sondes, launched in response to the impending
development. In addition to these changes from day to day in the 1200 UC data base,
the data volume at other times is different again, with the highest volume of
conventional data over Australia being at 0000UTC, and very few observations other
than at the times of 0000, 0600, 1200 and 1800 UTC.
The absence of upper air data in some parts of South America is also a problem. The
Herdies et al. (2003) study shows that the 6-hour simulation with inclusion of the
South America Low -Level Jet Experiment (SALLJEX) data represents the SALLJ
better and leads to an improvement in the moisture transport from the Amazon region
to southeast of South America (Figure 21).
.
29
Figure 20: Sonde data accepted by the operational numerical weather prediction
system LAPS at 1200 UTC on two days during the month of July 2005. For each
data point a circle shows that sonde -derived wind data were available at that
observation time. When the circle is filled with blue it shows that simultaneous
temperature and moisture data were also available. The upper panel is for a day
early in the month showing the normal upper air data base at this observation
time. The lower panel is for two days prior to the cutoff low development of Fig.
10.
The quality of analyses over the high southern latitudes has in recent years
dramatically improved (Bromwich and Fogt, 2004; Uppala et al., 2005). An important
milestone in this development was the FROST (First Regional Obse rving Study of the
30
Figure 21: Meridional wind profiles (ms -1) at Santa Cruz de La Sierra (17.5 oS 63.5 oW) on 21 January 2003. Radiosonde data (red line), 6h simulation with
SALLJEX data (black line) and without SALLJEX data (green line). From
Herdies et al. (2003)
Troposphere) project (Turner et al., 1996). This was organized by the Physics and
Chemistry of the Atmosphere Group of the Scientific Committee on Antarctic
Research. The goals of FROST were ‘… to study the meteorology of the Antarctic, to
determine the strengths and weaknesses of operational analyses and forecasts over the
continent and in the surrounding ocean areas, and to assess the value of new forms of
satellite data that are becoming available.’ The project may be seen as a forerunner of
the Antarctic component of SH THORPEX.
Much of this improvement has come as a result of the innovative and optimum use of
available observations, and of profile information from a range of satellites. It should
be pointed out, however, that profile information is of most use when data at a
reference level is available. High quality sea level information are now being provided
by the latest generation of scatterometer data. These scatterometer data have also been
used to explore the near-surface structure of intense midlatitude cyclones in southern
waters. Buckley and Leslie (2004) developed a preliminary climatology of such
systems for the notoriously data -sparse region 20-60°S, 30-130°E. One of their key
findings was that, historically, the frequency and intensity of the cyclones in this
domain have been significantly underestimated, and resulted in analyses that have
serious flaws.
The Sea Winds -on-QuikSCAT (QS) data represents an extraordinarily rich set for a
range of applications to understanding the workings, diagnosis and predictions of
extreme synoptic events in the southern polar regions. A range of creative uses of
these data should be followed up under THORPEX.
The strategies to be taken in research on observing systems for Southern Hemisphere
THORPEX still need to be thought about and developed. A guiding principle could
31
Figure 22: Example of the impact of AMDAR observations on the 12-h 250 hPa
SAWS Eta model forecasts of the horizontal wind components. Positive
(negative) impact are shaded red (blue).
be to determine methods of combining both scientific/research resources as well as
logistical and financial resources for mutual benefit. While repeating the caveat this
subprogramme needs a lot of further development, some initial ideas proposed for
discussion are as follows:
•
•
•
•
As the Southern Hemisphere is largely covered by ocean, it is necessary to
identify a ship on a suitable route to install equipment to collect daily data.
One suggestion for this equipment is the Ship Radiosonde Deployments
(European E-ASAP automatic aerological programme and Australian ASAP).
Investigate the feasibility of a sonde program for Antarctic Supply ships South
Africa, for example, has one ship). We believe this has been carried out at
times in the past, though the details, investigation of data accessibility, and the
experiences and findings still need to be researched.
The Aircraft Meteorological Data Relay (AMDAR) can be another source of
atmospheric data. Collaborative effort to better utilise data from aircraft is
necessary to develop a Southern Hemisphere AMDAR coverage, define route
responsibility, shared payment, solicit commitment from countries and assess
new humidity data impact. A further enhancement is to use the TAMDAR
(Tropospheric Airborne Meteorological Data Reporting) instrument which
measures a broad range of atmospheric variables from commuter aircraft
flying to small and medium size cities. A three-year study over southern
Africa to determine the impact of AMDAR observations on the South African
Weather Service (SAWS) 32km regional Eta model showed generally positive
impacts in this region that is characterised by a low areal coverage of
radiosonde data (Figure 22).
The meteorological satellite observing system comprising geosynchronous and
polar-orbiting satellites has good international collaboration and is used
32
•
•
extensively in NWP; however it is necessary to improve the link to terrestrial
systems to get optimal networks and to make informed decisions about how
satellite and terrestrial systems complement each other.
Regional field experiments can be very useful in testing observing and forecast
systems. Southern Hemisphere regional field experiments such as SALLJEX,
TWP -ICE to be conducted around Darwin in January/February 2006 and
SAFROBEX (Moisture transport in southern Africa) planned for 2006 are
examples of field experiments that could be used for such testing.
Look at launching drift sondes along key paths during the International Polar
Year (2007 - 2008). The proposal at this stage is that Southern Hemisphere
countries combine resources to fund this initiative, and that it be carried as a
Southern Hemisphere THORPEX Observing Systems Test (TOST). Funding
would of course be of a voluntary nature and the expected results of the
experiments and benefits to the National Meteorological Services would need
to be clearly demonstrated.
In this context it is noted the French meteorological community has proposed an
investigation during the IPY of whether advanced sounders onboard polar orbiting
satellites could be used to complement routine radiosonde observations over
Antarctica to provide a 4-dimensional picture of the atmosphere over the region. To
this end, they have proposed a campaign of driftsondes during the IPY as a platform
for targeted launches of drop sondes. It would be useful to carry out a coordinated
programme of impact studies across the hemisphere of this enhanced data platform.
(Information concerning the French programme available from Florence Rabier of
Meteo-France).
•
Take an inventory of the Islands across the Southern Hemisphere Oceans, in
terms of those currently launching sondes, previously launched sondes, and
never launched sondes. During the International Polar Year carry out an
extensive sonde launching programme from as many Islands as possible to
determine their importance for predictability and pre diction.
In order to use new observation types effectively, the South Hemisphere Committee
needs to coordinate efforts to:
1. Provide input from assessment and impact studies,
2. Assess information from studies that assimilate routinely available observation
network reports and make appropriate recommendations for adjusting networks
and for further investigation,
3. Determine what is the impact of different/new systems on a research basis, e.g.
AMDAR moisture,
4. Look into cost against usefulness of certain stations and set priorities,
5. Consider resources available to perform the various tasks.
Some further issues that arise are:
1. How to assimilate this data into NWP models?
2. Is it beneficial to divert funds from existing observation programmes to areas
more useful for the southern hemisphere in general?
3. Would national governments get greater value from more international
collaboration operationally or on research issues?
33
4. Advise existing programmes (e.g. AMDAR) on what the observing systems
should look like (spatial and temporal) and who would fund it?
5. Southern Hemisphere Data Assimilation and Observing Strategies Research
A significant component of forecast error originates from uncertainty in the initial
condition. This uncertainty arises from uncertainty in the observations, the
background forecast (used as a first guess analysis field in the data assimilation
process) and from approximations in the assimilation scheme (Shapiro and Thorpe,
2004). As described in the International Science Plan, there have be en a number of
recent advances in data observing systems including a greatly increased volume and
quality of atmospheric observations particularly from satellites and the development
of adaptive observational techniques or targeting.
The differences between the hemispheres in terms of ocean areas, population densities
and sizes of national economies introduces differences between the hemispheres as to
the some of the priorities of data assimilation and observing strategies. The
dominance of ocean areas has meant that the Southern Hemisphere is generally
ideally suited for observation by remotely sensing from satellites. This is seen in any
number of data impact tests for numerical weather prediction (e.g. the series of WMO
workshops on the impact of various observing systems on NWP, and references
therein).
The interiors of the southern land masses on the other hand seem destined to remain
relative data voids. The cost of maintaining remote stations and the relatively small
economies mitigate against any major in situ observing systems that would be
comparable to North America, Europe or East Asia. With the growth of satellite data,
and the economic constraints expected to continue, the disparity between the
observations over the oceans and over the land will increase.
Another issue is that the size of the economies implies that it is unlikely for a
Southern Hemisphere medium range prediction effort to deliver significantly more
information than is available from elsewhere. This has meant that Southern
Hemisphere NWP centres have tended to focus on 1 to 3 day prediction of subsynoptic meso-scales and the accuracy of these predictions. This focus on local highimpact weather means that their activities are well aligned within the THORPEX
objectives. The differences between the hemispheres also provide scope for the
Southern Hemisphere centres to make significant scientific contributions in areas that
are distinct from other regional THORPEX activities. This is not to imply that
Southern Hemisphere institutions cannot contribute to other areas within THORPEX,
or that only Southern Hemisphere institutions can contribute to the areas discussed.
The intention is rather to direct attention to the issues that are of particular importance
to the Southern Hemisphere.
The performance of assimilation systems is directly influenced by the accuracy and
resolution of the observing network and forecast models, and how well the accuracy is
specified. This immediately introduces differences in the performance of assimilation
schemes between the hemispheres as differences in state of the atmosphere and its
variability effect the performance of forecast models (this is the underlying premise of
THORPEX). The non-linear relationship between model performance and
atmospheric state can produce significantly different responses when techniques are
transferred to other regions and/or models. This is particularly true of high resolution
34
prediction of high impact weather where quite different processes can dominate local
weather and therefore the local forecast accuracy. This regional variation in
assimilation performance is one of the main reasons for requiring a regional data
assimilation and observing strategy focus group for the Southern Hemisphere.
This section will discuss some areas where a Southern Hemisphere regional focus
within THORPEX can both provide scientific advances within the larger THORPEX
program, as well as meeting the aims and objectives of the participating institutions.
5.1 Collaboration
It is recognized within THORPEX that data assimilation is one of the key topics for
improving predictability and forecast utility of many types of high impact weather.
The provision of accurate and timely high resolution initial conditions is often best
obtained via the use of local limited area NWP systems. The increase in complexity
and data volumes associated with these systems has meant that development and
maintenance has become a major issue for Southern Hemisphere countries. This has
been addressed for some time by various levels of collaboration between NWP
centres.
Recently, these collaborative agreements have become more formalized, and involved
entire NWP systems. This raises the opportunity for heightened collaboration between
the different Southern Hemisphere centres that use similar NWP systems. Such
collaboration will be of significant direct benefit to those involved, as any
developments at one centre will immediately be applicable and relevant to the other
centres. This coordinated attention to issues associated with the Southern Hemisphere,
along with the interest of other centres and consortia provides further impetus for the
establishment of a Southern Hemisphere regional assimilation and observing systems
focus group within a larger international forum. Such a focus group would also be a
suitable forum for enhancing interest in data assimilation etc. within universities and
other research organizations. This focus group would be best initiated under the
umbrella of THORPEX, by interested organizations and consortia.
5.2 Error covariance modelling
Background error covariance specification is an area that has long been a problem
within data assimilation and is being studied widely, but the topic is so fundamental to
data assimilation that it must be included in any discussion on data assimilation. In
particular, research into flow dependent background error covariances has long been a
concern, and its importance was established at the first THORPEX executive board
meeting (Geneva, 1-2 September, 2005).
There are several approaches to estimating, modelling and evolving background error
covariances. The decision as to which approach is the most appropriate depends not
only on scientific issues, but also the capabilities of the individual centres, and how
systems related to NWP can benefit from any by-products of these approaches. An
example of this is the use of ensemble representations of the background error
covariance also being used to initialize ensemble prediction systems.
Differences in the relative importance of some scientific issues will also have a
bearing on which approaches are most suited for use within the Southern Hemisphere.
The importance of tropical and extra-tropical features amongst the list of high impact
35
weather events will require greater weight being given to approaches that perform
well near areas of organized or large scale tropical convection. The highly non-linear
relationships between variables, the importance of the distribution of moisture
species, and the dependence on satellite da ta means that the best methods for the
Northern Hemisphere mid-latitudes are not necessarily the best for the Southern
Hemisphere. Similarly differences in atmospheric variability and observing systems
will also have some impact, as they both effect the accuracy of the initial state. These
differences between the hemispheres and the resources of the NWP centres imply that
a Southern Hemisphere focus on background error covariance estimation, modelling
and prediction is warranted.
5.3 Performance metrics
Regional and meso-scale data assimilation schemes have generally (but not entirely)
evolved from global assimilation systems. This is understandable as the performance
of global systems is not complicated by lateral boundary conditions, and the larger
scale relationships between model variables are better understood. The performance
of these systems must be routinely monitored according to some metrics to both
assess the value of proposed changes and as part of general quality assurance. These
metrics are mainly sensitive to large scale information, and give confusing or
misleading results when applied to high resolution systems. The most obvious
problem with many of these metrics is that a forecast of an intense weather system
with a small phase error is penalized more than a forecast with a weather system that
is too weak. Given that the costs associated with responses to a forecast are often very
non-linear, it is often more important to know the potential for any intense weather
systems and the performance metrics should reflect this. The development of skill
measures that are suited to high resolution forecasts is therefore important in assessing
where improvements are required. This is as important for data assimilation and
observing systems as it is for other aspects for forecasting.
This is an area that clearly must have strong links with the Social and Economic
Applications Working Group, and is a problem being studied around the world. There
is not however, likely to be a single measure that suits all applications and weather
events. The Southern Hemisphere regional committee should be encouraging the
development and assessment of performance metrics suited to Southern Hemisphere
needs and conditions.
5.4 Assimilation of satellite data
As mentione d previously, satellite data is the primary source of observations for NWP
in the Southern Hemisphere. The use of high volume datasets (both high spectral
resolution soundings and high horizontal resolution imagery) was also identified as an
important task at the first THORPEX executive board meeting. Again some of the
issues related to satellite data assimilation are being pursued within various
organizations around the world, but the importance of this topic to Southern
Hemisphere NWP requires it to be included within the regional science plan.
Current assimilation systems discard the majority of high density satellite data –
particularly that from polar orbiting satellites. The prime scientific justification for
this is the poor understanding and characterization of observation errors. These errors
are correlated via the observation operators (i.e. forward models or fast radiative
transfer codes). The correlations are not well known, and for computational ease they
36
are generally ignored, but this is equivalent to over-estimating the information
contained within the data. To overcome these problems, the data are thinned so that
the observations used are close to independent. This thinning is usually rather
simplistic and often discards important information in areas of meteorological
importance, i.e. near strong gradients.
To avoid this discarding of information requires either the explicit allowance for
correlated observation errors within the analysis scheme or the use of intelligent
thinning or data compression algorithms. Thinning and compression techniques have
advantages in that the volume of data being processed by the assimilation scheme is
reduced. This is a significant consideration given the amount of data that will be used
during the next decade. These algorithms are however only approximations built
around assessments of the information content of the observations, and therefore still
depend to some extent on knowledge of both the background and observation error
covariances. The accuracy of the approximations is therefore related to how well the
error covariances are specified, and uncertainties in this specification may negate any
other advantages. The problem of thinning and compression techniques becomes more
complex when dealing with high resolution observations that are sensitive to both
temperature and moisture (e.g. AMSU -B) as these variables can potentially have
different horizontal observation error correlations. There are many techniques for
dealing with correlated observation error covariances – and assessing which are best
suited for various Southern Hemisphere applications is an important task that can be
undertaken within THORPEX.
The specification of the observation error covariances requires knowledge of the
behaviour of both the instrument and the observation operators. Work in this area is
also being undertaken around the world, although it will be important to determine
suitable characterization of the error covariances for important areas for the Southern
Hemisphere such as tropical convective areas, the southern ice edge, the Antarctic
plateau etc. as well as for more general areas and phenomena. There is considerable
capability in satellite data assimilation within the Southern Hemisphere community,
and by providing coordination and a focus for this expertise, the Southern Hemisphere
regional THORPEX committee can facilitate significant developments in this area.
There are several other topics within satellite data assimilation that are highly relevant
to high impact weather in the Southern Hemisphere that would benefit from greater
attention. One such area is the assimilation of sounding (particularly microwave) data
from areas of tropical and extra-tropical convection. The high impact of some tropical
and extra-tropical convective systems means that the accurate specification of initial
conditions in these areas is a matter of high importance. The technology to use
soundings of convective systems requires accurate and sophisticated linear and nonlinear forecast models. Such models have recently been shown to be feasible, and are
expected to be available to some Southern Hemisphere centres in the next few years.
Another important topic within satellite data assimilation is the use of data over land.
The analysis of the mid- to lower-troposphere over the southern land masses currently
relies mostly on in situ observations. As mentioned earlier this makes these regions
relatively data sparse and for this reason alone it is crucial to enhance the use of
satellite data over land. The land surfaces of the Southern Hemisphere are
substantially different to those of the northern mid-latitudes, and so again, techniques
developed for one area need not necessarily be the best choice for use in another.
Analyses of the types of high impact weather events that occur in the Southern
37
Hemisphere show that areas of both very high and very low relative humidity are a
concern. The accurate depiction of areas of high relative humidity is obviously
important for predicting major rain and convective events, however the accurate
analysis of very low relative humidities is also crucial for predicting high impact
weather related to fires. Fire weather and heavy rain events are both of such
importance to many Southern Hemisphere countries that the accurate specification of
both extremes of mid- and lower- tropospheric moisture is a matter of high
importance.
5.5 Assimilation of other remotely sensed data
Relative to the amount of data produced by land-based radars, virtually no data is
routinely used in data assimilation schemes, and none in the Southern Hemisphere.
This is despite radar data being one of the major sources of high resolution data
through the lower- to mid-troposphere. The coverage of radar data lends itself more to
short term (order of a day) prediction, but nonetheless can be decisive in planning the
response to forecasts of high impact weather.
With the introduction of Doppler radars to many countries during the next few years,
high resolution data of both wind and reflectivity can be used to initialize NWP
forecasts. Of the two, the wind data is easiest to use, although the quality control is a
major issue. Reflectivity data is very complex to use due to limitations of forward
models, and the strong non-linear dependence on the distribution of water species.
The non-linearity poses difficulties in itself, but it is compounded by the lack of
understanding of the background error covariances of the relevant variables, with
neither the relationships between the variables or the variances being well
characterized. There is a significant amount of work that needs to be undertaken by
the global community, but there will also need to be a section of this community
ensuring that the developments are appropriate for the Southern Hemisphere. This is
where the Southern Hemisphere regional THORPEX focus group can make a
substantial contribution.
5.6 Observing strategies
There are many reasons to suspect that targeted observing systems may not be
appropriate for the Southern Hemisphere: inconclusive results in the Northern
Hemisphere, the variability of target regions, the sparseness of the existing in situ
observing network, financial considerations etc. This does not however preclude
investigations into observing strategies for the Southern Hemisphere. There are still
some topics related to targeting that are very important and where coordination and
promotion by the Southern Hemisphere regional THORPEX committee would be of
considerable benefit.
Firstly there is the issue of targeting satellite data, particularly from geostationary
satellites. With the latest generation of satellites being capable of providing high
resolution scans of target areas, there is a need to further investigate how this can be
best utilized. This will require not only studies on determining appropriate target areas
and sampling strategies, but also improvements in satellite data assimilation to better
use this data. There is also the prospect of targeting the processing of data from polar
orbiting satellites, so that expensive calculations involving observation error
covariances and/or more sophisticated radiative transfer models are used in some
38
areas, with simpler processing on other regions. This area is closely related to the
topic of data thinning/compression discussed above.
A second issue is the extension of the AMDAR program to include smaller aircraft
that fly at lower levels. This is of particular interest to Southern Hemisphere
countries, where there a few large cities, and large areas where small aircraft are a
major form of transport. This means that there is a limited number of airports
frequently dealing with AMDAR enabled aircraft, but the coverage from a system
such as TAMDAR could be significant. The transmission costs are still an issue and
so studies are needed to assess the best sampling strategies - an area that is directly
covered by THORPEX.
Finally, monitoring the performance and quality of the in situ observing network to
provide estimates of both the accuracy of the analysed fields, and the benefit relative
to the cost is becoming increasingly important as accountability demands increase.
Studies of this nature have been initiated in Europe and North America, and the
process is no less important for the Southern Hemisphere. This is also an issue that
falls directly under the objectives of THORPEX. The importance of this being done
on regional to hemispheric scales rather than country-by-country has been stressed at
many meetings, such as the WMO workshops on the impact of various observing
systems in NWP. As such it would be highly appropriate for the regional THORPEX
committee to take a leading role in ensuring that thorough assessments are made of
the importance of the in situ observing network.
One of Africa’s most serious meteorological problems is that of sparse or non-existent
observation networks. Adaptive observation strategies offer promise to mitigate some
of these problems. The Southern African Observation Experiment (SAFROBEX) has
been proposed to investigate the utility of the approach. SAFROBEX will mainly
comprise of adaptive sampling experiments beginning with Botswana, Namibia and
Angola moving on further afield to the rest of the Southern African Development
Community (SADC) region, which comprises only of Southern Hemispheric African
countries. Areas that are prone to tropical cyclones will also be considered. At present
observation impact studies are being conducted in South Africa to lay a solid
foundation for adaptive sampling in the above-mentioned areas. These methods could
als o be used in buoy deployment in South Africa’s area of responsibility in the
southern Atlantic Ocean. SAFROBEX is envisaged to begin in June 2006.
Combining adaptive observing networks and predictability studies has not been tested
within the tropics, unlike the northern mid -latitudes where campaigns such as
FASTEX and NORPEX have provided valuable insight. The international tropical
experiment TWP-ICE that will occur in Darwin January/February 2006 would be a
suitable candidate for a preliminary campaign, before possibly a more major
experiment at a later date. Thus the extra sondes to be deployed during TWP-ICE may
provide some useful information towards refining the numerical system before the
later experiment. Similarly the use of the ARM (Atmospheric Radiation
Measurement) site at Darwin to validate and assess the accuracy/performance of
observing systems in the tropics during the 2006 experiment could provide useful
information.
5.7 Summary
39
In summary there is a clear need for a group such as the Southern Hemisphere
regional THORPEX committee to take a lead role in coordinating and promoting
work on data assimilation and observing systems that is relevant to the region. The
topics to be addressed include:
•
•
•
•
•
•
Collaboration between centres interested in Southern Hemisphere NWP in
general and data assimilation in particular.
Characterization of background errors in the tropics and southern hemisphere.
This should include their estimation, specification and evolution for high
resolution NWP.
The development of performance metrics suited to assessing southern
hemisphere high impact weather predictions and the impact of these forecasts
on the major users.
Methods for processing and assimilating high volume satellite data. This
should include observation error characterization, forward modelling, data
thinning and compression and for targeting both instruments and processing
of the observations. This should encompass all sources of satellite data.
The use of new observing systems that are suited for use wit hin southern
hemisphere countries.
The assessment and monitoring of the southern hemisphere observing
network.
Some of these tasks will have strong links to other regional committees and/or other
experiments, but within each of these there are aspects that are of special concern to
the southern hemisphere community. Differences in observing systems, modelling
systems, atmospheric conditions and the economic impact of different weather events
mean that there is sufficient justification for supporting a Sout hern Hemisphere focus
on data assimilation and observing strategies within THORPEX.
6. Southern Hemisphere Societal and Economic Applications Research.
“THORPEX Societal and Economic Applications research (SEA) will contribute to
the development of forecast systems that are responsive to the needs of users of
weather forecast information, with an emphasis on high -impact weather forecasts”.
(Shapiro and Thorpe, 2004). The International Science Plan includes an earlier
version of Fig. 23 (provided by A. Simmons, ECMWF), which demonstrates the very
large increase in model skill over the past 25 years, with the skill in the Southern
Hemisphere now matching that in the Northern Hemisphere. Despite the rapid
advances shown, it is not clear to what extent these advances have been applied to
bring direct benefit to society, the economy and the environment. Thus the purpose of
THORPEX SEA is to investigate means of optimising the benefit to society of the
advanced numerical weather prediction forecast-modelling s ystems.
The International Science Plan also states “High-impact weather forecasts are defined
by their effect on society, the economy and the environment. They are typically
associated with forecasts of cyclones of extratropical and tropical origin that contain
significant embedded mesoscale weather and its impacts. These include localized
flooding by convective and orographic precipitation; blizzard snows; dust-storms;
destructive surface winds. They also encompass forecasts of meteorological
conditions affecting air quality, periods of anomalous high/low temperature, drought,
and non -extreme weather with high societal/economic impact. Many of these events
40
are characterised as low probability, but with high risk, in that the event is unlikely,
but the consequences of occurrence may be catastrophic. Improving the skill of high impact weather forecasts is one of the great scientific and societal challenges of the
21 st century. THORPEX responds to this challenge.”
Figure 23: Evolution of forecast skill for the northern and southern hemispheres:
1980-2004. Anomaly correlation coefficients of 3, 5, 7 and 10 -day ECMWF 500 mb height forecasts for the extratropical northern and southern hemispheres,
plotted in the form of running means for the period of January 1980-October
2005. Shading shows differences in scores between hemispheres at the forecast
ranges indicated (provided by A. Simmons, ECMWF).
In both the global context and for the Southern Hemisphere specifically, it is
important to note that “high-impact weather forecast” does not equate solely to
forecasts of severe weather.
The International Science Plan further states, in relation to Societal/Economic
Applications: “A key objective is to develop a framework within which researchers in
the meteorological research, economic, policy and social science communities will
interact with operational forecast centres and users of weather forecast information.
This interaction will contribute to the development of improved forecast systems
designed for the diverse geographic regions involved in THORPEX. Research
findings will be made available for training and educational material to all nations.
“THORPEX Societal/Economic Applications research will: i) define and identify
high-impact weather forecasts; ii) assess the impact of improved forecast systems; iii)
develop advanced forecast verification measures; iv) estimate the cost and benefits of
improved forecast systems; v) contribute to the development of user-specific products;
vi) facilitate the transfer of THORPEX advances to forecast centres throughout the
41
world. This research will be conducted through collaboration with forecast providers
(operational forecast centres and private -sector forecast offices) and forecast users
(energy producers and distributors, transportation industries, agriculture producers,
emergency management agencies and health care providers). It will provide wideranging information on what constitutes high -impact weather forecasts for individual
sectors and users and assist in establishing research priorities within all elements of
THORPEX.”
The application of weather forecast information to the benefit of society and economy
at large is quite diverse and complicated in the case of Southern Africa, for example,
due to its mixture of mostly developing and few developed areas. Most rural regions
in Southern Africa are particularly vulnerable to high impact weather due to their
limited coping and adaptive capacity as a result of widespread poverty, HIV/AIDS,
population growth and the general vulnerability of rural subsistence communities to
particularly natural disasters. The significant impact that tropical cyclone Eline had
on the rural communities in Mozambique is a telling example. In many areas
application of weather forecasts to meet the needs of society is non-existent. On the
other hand, in some well developed regions industries and communities benefit
significantly through the use of weather forecast products. Important in this context is
also the fact that the technical capabilities of weather services in the region (14
countries in the region) are quite diverse. South Africa is fortunate enough to have a
small but technically well developed weather service, however, the majority of the
other weather services lack financial resources to establish sophisticated systems like
weather radars, numerical models, or even proper data communication links in a few
cases.
Any application to enhance societal and economic use of weather information must
take these realities into account. The challenge lies in the application of sophisticated
new technologies, developed in first world environments, to the benefit of
communities at risk in developing environments allowing them to take efficient
advantage of weather forecasts, reducing their vulnerabilities and increasing their
coping capacities.
The enhancement of the early warning systems of hazardous weather as an end-to-end
system should be high on the agenda of weather services. This implies that the
technical abilities of weather services to identify hazards well in time and issue
warnings to vulnerable communities must be enhanced. This in turn needs a proper
understanding of the hazards and vulnerabilities of communities at risk and the
consequences of high impact weather on the various activities of these communities.
It also implies development of effective dissemination systems of forecasts and
warnings to disaster management structures and local communities in a way that they
understand and can react on. Communication of forecast information will be enhanced
by taking a true end-to-end approach, and providing predictions of the impacts in user
terms, rather than just predictions of weather elements alone.
For the six objectives of global SEA research noted above, these are the issues that
need to be addressed in the Southern hemisphere context:
(i)
Identify high-impact weather forecasts: In Section 3 a number of
meteorological phenomena were described in the context that they had a high
economic impact and they were phenomena whose dynamics and predictability were
not well understood. The type of information described in Section 3 was based on
42
various sources including the meteorological research literature, knowledge gained
from experience working in na tional meteorological services and from insurance
industry and newspaper reports of damage estimates from individual events.
There has not been a systematic inventory carried out, however, of high-impact
weather forecast opportunities for the Southern Hemisphere countries. This is
necessary as a first step, prior to carrying out predictability studies for related weather
phenomena and prior to developing user specific weather products relevant for these
opportunities. Thus a key objective for the Southern Hemisphere THORPEX SEA
research would be to develop a catalogue/inventory of high-impact weather forecast
opportunities for the Southern Hemisphere. The purpose for such an inventory
should be not simply to list and catalogue past events, but to identify opportunities
where, even with the existing skill level of numerical prediction and weather
forecasting, users and communities could be provided with improved specific weather
products to assist in decision making. So the focus should primarily be on user
impacts and the opportunities for the marginal value - i.e., the added value to users
and society in mitigating losses, increasing gains, or improving the management of
resources.
As an example, NIWA (in collaboration with Macquarie University, and the PerilAus
project) has begun building a database in which such event information can be stored.
The aim is to search-out and collate information on historical events of extreme
weather conditions that have affected New Zealand, including storms, snow/hail,
tornados, flooding, wildfire, heat-waves, droughts, cyclones and landslides – but this
could be extended to include other high-impact weather events that are not associated
with physical damage (e.g. long-duration fog events). The type of information that can
be incorporated in the event-database includes: date and duration, regions affected,
return period, number of casualties, number of injuries, number of evacuees, lifeline
utilities affected, property damage (dollar amount), crop damage (dollar amount),
stock damage (dollar amount), insurance claims and payouts, and emergency
classification. Additional information may be associated with each event – such as
weather analyses, images, rainfall, wind, river and wave statistics, bibliographic
references etc. Using this Historical Events database it may be possible to determine
whether there are any trends in the impacts. The information in, and design of this
impacts database could inform the wider goal of achieving a catalogue/inventory of
high impact weather events for the Southern Hemisphere. It would be highly
desirable, under the Southern Hemisphere THORPEX plan, to collate such
information from all Southern Hemisphere countries and develop a readily accessible
unified database.
Some examples of high-impact weather forecast opportunities on the 1 – 14 day time
scale are:
•
•
•
•
•
•
•
Major flooding events
Major windstorm events
Major snowstorm events
Rain-triggered landslides
Tropical cyclones (including landfall, marine impacts, and ex-tropical
cyclone impacts in higher latitudes)
Major coastal damage or inundation
Fire weather events
43
•
•
•
•
•
•
•
•
Severe thunderstorms and major hailstorms
Unusually warm or cold weather
Drought cessation
Monsoon onset. (For example, at the beginning of the period of
continuous heavy monsoon rains over northern Australia, the roads
become impassable and remain that way for 3 to 5 months. Thus on the
day of monsoon onset, fuel and transformation costs rise in many northern
communities.)
Agricultural decision making, such as critical forecasts for sowing or
harvesting crops or for haymaking, or instances where crops are lost due to
frost, or potential stock losses due to inclement weather. (For example, the
wheat crop over most of Southern and Eastern Australia requires a
threshold amount of rain during planting time, which is during the month
of April. Farmers must make a decision as to when to plant. The planted
crop needs to go into wet soil and the soil must remain wet for about two
weeks. Thus there are highly critical short-term rain forecasts that must be
made at this time of year.)
Planning of energy demand and production, including hydroelectric
storage, and wind-power production.
Fog and low cloud with associated transport disruption
Periods of poor air quality
(ii)
Assess the impact of improved forecast systems: The catalogue of
identified high-impact weather forecast opportunities for countries and regions of the
southern hemisphere will provide the basis for studies estimating the marginal value
of improvements to forecast systems. The source of such improvements includes not
only the “back-end” observing, data assimilation and numerical modelling systems,
but also the provision of forecast information and products in ways that can assist
decision making. Such research will require working closely with users and decisionmakers (including, for example, emergency management agencies, water resource
agencies, insurance companies, customers for specialised services) to better
understand the impacts and their decision making processes. It will also benefit from
the involvement of academia – in particular in the social sciences.
(iii)
Develop advanced forecast verification measures: Working with users of
the weather forecast information will lead to the development of user -relevant
verification to supplement the widely-used scientific verification measures such as
500 hPa anomaly correlations and root-mean-square errors of temperature forecasts.
(iv) Estimate net benefits of improved forecast systems : Few Southern
Hemisphere countries are involved in advanced global-scale numerical weather
prediction modelling and data assimilation, and none directly fund operational
weather satellite systems. However, all contribute to the global database of real-time
observations, particularly through conventional surface and upper air (radiosonde)
observations, which are utilised by all Members of WMO, and in particular the major
modelling centres. The Southern Hemisphere focus on the net benefits of improved
forecast systems is therefore expected to be primarily on net benefits taking into
account changes to observing systems as well as the costs associated with
implementation of improved forecast products drawing on global model information
(at global resolution as well as with limited area model downscaling).
44
(v)
Develop new user-specific weather products: In broad terms there are four
groupings of types of user-specific weather products, each of which has its own
challenges in the development of new products to ensure that the benefits of improved
forecasting systems can be captured.
(1) Public safety, where the products on the 1-14 day range will be intended
primarily for emergency management organizations and other government
agencies, as well as the public in general. An example of an existing product
is the severe weather outlook provided by Meteorological Service of NZ (see
http://www.metservice.co.nz/default/index.php?pkey=190599&ckey=190600
or http://tinyurl.com/azetq), which uses probabilistic terms to communicate
the relative risk of severe weather events in the 2 to 5 day outlook period.
Other products can include tropical cyclone cones of risk. Full use needs to
be made of existing ensemble -based information such as that made available
to all WMO Members by ECMWF; particularly the tropical cyclone track and
intensity forecasts.
Development of such products generally involves consultation involving the
national weather service (including forecasters, researchers, modellers),
emergency management organisations, other government agencies,
community representatives, and the media. Work in this area should be
coordinated, inter alia, with any demonstration projects undertaken by the
CBS OPAG on Data Processing and Forecasting Systems.
(2) Public good information for use by the general public for planning their own
activities. This is not specifically safety-related, although it can help in
avoiding the public taking part in activities under unsafe conditions (e.g., poor
weather in mountainous areas). Such information is widely available, but
much work is needed on how to communicate probabilistic information in an
understandable and useful way. It would be useful for Southern Hemisphere
NMSs to prototype and share experiences on how to to do this.
(3) Public good information intended for specific user groups for planning
activities – e.g., agricultural information, tourism, and marine. For example,
see the 3-day marine forecast issued by the South African Weather Service at
http://www.weathersa.co.za/FcastProducts/Special/WesternCape.jsp?fcast=fis
hing-surfing or http://tinyurl.com/aq6my . Products of this type can be very
specific to the activities conducted by the user group, and to support decisions
made in planning, so communicating probabilistic information is more
tractable. Given the importance of agriculture to many Southern Hemisphere
countries, this could be a target or demonstration sector for new services. The
marine and tourism sectors could also be candidate demonstration sectors for
Small Island Developing States.
(4) Specialised information generally provided on a commercial basis to a
specific user, usually for a specific location. This is generally the domain of
either a commercial arm of a national weather service, or the value-added
commercial weather business. Products include services to the energy
industry, and for farmers at specific locations (see, for example, the 7 day
meteograms available on a commercial basis from the Australian Bureau of
Meteorology at http://www.bom.gov.au/silo/products/Meteograms.shtml or
http://tinyurl.com/chssc ; a sample is available) . Product development in this
area will take place in association with the value-added providers, and the role
of the THORPEX programme should be to foster such developments, to
45
promote “proof of value” studies which explicitly demonstrate the potential
benefits from new products, and to facilitate the availability of information
and research results to maximise the overall benefit to the economy as a
whole.
(vi)
Facilitate transfer of THORPEX advances to forecast centres throughout
the world: As noted earlier, most Southern Hemisphere meteorological services do
not run advanc ed global models, so will be particularly reliant on making use of
products from the major centres. This will include making use of results of progress
on, and use of near-real-time data from, the THORPEX Interactive Grand Global
Ensemble (TIGGE). There are also communications bandwidth issues in the Southern
Hemisphere, so strategies such as subsetting of TIGGE data for limited geographical
regions or provision of web-based products will be essential to enable its use by
smaller Southern Hemisphere countries.
Societal and economic benefits will only be realised if users are able to take actions
that make a difference, and they are provided with the right information at the right
time to optimise their decision making. A workshop or series of workshops could be
held in Southern Hemisphere countries which bring sectoral representatives together
with meteorologists, service providers and academia to interact on decision making
and identifies the opportunities that should be exploited to bring the most immediate
benefit.
7. Summary
The purpose of this document is as a preliminary draft that provides a rationale for a
Southern Hemisphere THORPEX campaign and lists an appropriate subset of the
scientific problems that could be addressed based on contributions from scientists in
Southern Hemisphere countries. The document in its current form covers a wide range
of ideas and phenomena from a Southern Hemisphere perspective.
From the various contributions it is clear there is a commonality in forecast problems
across the hemisphere. Examples are as follows:
•
•
•
•
•
•
Fire weather is common to Australia and South Africa, and in both cases is
associated with synoptic scale conditions leading to strong pressure gradients
with a cross continental trajectory.
The Madden Julian Oscillation is a major modifier of weather on the 1-2 week
timescale (and longer) for the tropical portions of Southern Hemisphere
Africa, South America and Australia-New Zealand.
Cut-off lows are producers of major wide -spread flooding events off the east
coast of all three southern hemisphere continents.
Rapid cyclogenesis causing gale -force winds and rapid sea swells have
brought about major boating disasters off the east coasts of South Africa,
South America and Australia.
Wide spread flooding and loss of life associated with tropical cyclone landfall
is the major high impact phenomenon for the Australian tropical coastline, the
South Pacific countries and the region of Mozambique, Madagascar and
Mauritius.
Semi-stationary mid-tropospheric (blocking) anticyclones lead to extended
heat-wave conditions over southern Africa and Australia.
46
Problems associated with data impact studies also have a hemispheric nature to them.
For example, the Southern Hemisphere countries have traditionally been responsible
for the installation and maintenance of radiosonde networks with the ground stations
located on remote islands distributed across the southern hemisphere oceans. These
stations are expensive to run and maintain such that a Southern Hemisphere
THORPEX priority should be to quantify the impact and required density of
observations across this oceanic islands network. Similarly there is a hemispheric
nature to the large data void regions, such that it would be appropriate for Southern
Hemisphere countries to pool resources to evaluate the potential impact of advanced
sounding systems such as drift sondes.
It is hoped that, through the interaction between Southern Hemisphere countries
arising from collaboration through THORPEX, a shift in paradigm will occur such
that we take an integrated hemispheric view on advancing the science of predictability
and understanding of the above weather phenomena.
47
8. References
Aron, J., Patz, J.A. (eds) 2002: Ecosystem change and public health. Baltimore:
John Hopkins University Press.
Bishop, C.H. B.J. Etherton and S.J. Majumdar, 2001: Adapative sampling with the
Ensemble Transform Kalman Filter. Part 1: Theoretical aspects. Mon. Wea . Rev.,129,
420-129.
Blackmon, M.L., 1976: A climatological spectral study of the 500 mb geopotential
height of the Northern Hemisphere. J. Atmos., Sci., 33, 1607-1623.
Berbery EH, and Barros V (2002) The hydrological cycle of the La Plata Basin in
South America. J Hydromet 3: 630-645.
Berbery, E.H. and E.A. Collini 2001: Springtime precipitation and water vapor flux
convergence over southeastern South America, Mon. Wea. Rev., 128, 1328-1346.
Bond , H.G., Mackinnon, K., and Noar, P.F. 1967. Report on the meteorological
aspects of the catastrophic bushfires in south-eastern Tasmania on 7 February 1967.
Bureau of Meteorology, Australia. 54pp
Buckley, B. W., and L. M. Leslie, 2004: Preliminary climatology and improved
modelling of south Indian Ocean and Southern Ocean mid-latitude cyclones. Int. J.
Climatol., 24, 1211-1230.
Bureau of Meteorology, 1984: Report on Meteorological Aspects of the Ash
Wednesday Fires - 16th February 1983. Australian Government Publishing Service,
Canberra.
Bureau of Meteorology, 1985. Summary report of the fires in central and north-central
Victoria on 14 January 1985. Bureau of Meteorology, Australia. 21pp.
Campbell, W. J., E. G. Josberger and N. M. Mognard, 1994: Southern Ocean wave
fields during the austral winters, 1985-1988, by Geosat radar altimeter. The Polar
Oceans and Their Role in Shaping the Global Environment. O. M. Johannessen, R. D.
Muench and J. E. Overland, Eds., American Geophysical Union, 421-434.
Carleton, A. M., 2003: Atmospheric teleconnections involving the Southern Ocean. J.
Geo. Res., 108, 8080, doi:10.1029/2000JC000379.
Cazes-Boezio GH., A.W. Robertson, C. R. Mechoso, 2003: Seasonal dependence of
ENSO teleconnections over South America and relations with precipitation in
Uruguay. J. Climate, 16, 1159-1176.
Cheney, N. P. 1979. Bushfire disasters in Australia 1945-1966. In (eds)
R.L.Heathcote and B.G.Thom - Natural Hazards in Australia. Aust. Acad.of Sci.,
Canberra, 72-93.
Conforte J.C., 1997: Um estudo de complexos convectivos de mesoescala sobre a
América do Sul (A study of the Mesoescale convective complexes over South
America). Dr. Thesis INPE.
48
Dal Piva, E., 2005: Cavados em médios níveis no Hemisfério Sul: climatologia,
energética e ciclogênese em superfície. PhD. Dissertation.
Fortune, M. A., and V. E. Kousky, 1983: Two severe freezes in Brazil: Precursors and
synoptic evolution. Mon. Wea. Rev., 111, 181–196.
Funatsu, B.M., Gan, M.A., Neto, E.S.C., 2004: A case study of orographic
cyclogenesis over South America. Atmosfera, 91-113.
Gan, M.A. and Rao, V.B., 1991: Surface cyclogenesis over South America. Mon.
Wea. Rev., 119(5), 1293-1302.
Gan, M.A.; Rao, V.B. 1999: Energetics of high frequency disturbances over South
America. Revista Brasileira de Geofísica, , vol 17(1), 21-28
Garreaud, R. D., 1999: Cold air incursions over subtropical and tropical South
America: A numerical case study. Mon. Wea. Rev., 127, 2823-2853.
Garreaud, R. D., 2000: Cold air incursions over subtropical South America: Mean
structure and dynamics. Mon. Wea. Rev., 128, 2544-2559 Part 2.
Geddes, D.J., and E.R. Pfeiffer: 1981. The Caroline forest fire, 2nd February 1979.
Woods and Forests Department, South Australia, 1981. 52pp.
Garreaud, and J. M.Wallace, 1998: Summertime incursions of midlatitude air into
tropical and subtropical South America. Mon. Wea. Rev., 126, 2713–2733.
Hall, A., and M. Visbeck, 2002: Synchronous variability in the southern hemisphere
atmosphere, sea ice, and ocean resulting from the annular mode. J. Climate., 15,
3043-3057.
Hamilton, M. G., and J. R. Tarifa, 1978: Synoptic aspects of a polar outbreak leading
to frost in tropical Brazil, July 1972. Mon. Wea.Rev., 106, 1545–1556.
Hartmann, D. L., and F. Lo, 1998: Wave -driven zonal flow vacillation in the Southern
Hemisphere. J. Atmos. Sci., 55, 1303-1315.
Hendon, H.H., B. Liebmann, M. Newman, J.D. Glick, and J.E. Schemm, 2000:
Medium -range forecast errors associated with active episodes of the Madden-Julian
Oscillation. Mon. Wea. Rev., 128, 69-86.
Herdies, D.L., I.F.A.Cavalcanti, J.A. Aravéquia, J. Marengo, R. Cintra, 3002.
SALLEX data assimilation. SALLJEX Data Wo rkshop, Buenos Aires, Argentina.
Hoskins, B.J., and P.J. Valdes, 1990: On the existence of storm tracks. J Atmos. Sci.,
47, 1854-1864.
Jones, M. C., 2003: Climatology of cold outbreaks with snow over Tasmania. Aust.
Met. Mag., 52, 157-169.
49
Jusevicius, M. A. R., 1999:
Ciclogênese a sotavento dos Andes: estudo
observationa le simulação numérica de um evento (Andine Lee cyclogenesis:
observational study an numerical simulation of a case). Master theses - INPE.
Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El
Niño-Southern Oscillation events. J. Climate, 2, 1239-1252.
Karoly, D.J., D. G. Vincent, and J. M. Schrage, 1998: General Circulation. Chapter 2
of Meteorology of the Southern Hemisphere, Meteor. Monograph. Vol. 27, No. 49.
American Meteorological Society, Boston, 410 pp.
Kousky, V. E., 1979: Frontal influences on northeast Brazil. Mon. Wea. Rev., 107,
1140–1153.
Kousky, V. E., and J. Ferreira, 1981: Frontal influences on northeast Brazil: Their
spatial distributions, origins and effects. Mon. Wea. Rev., 109, 1999–2008.
Kovats, R.S. , M.J., Bouma, S. Hajat, E. Worrall, and A. Haines, 2003. The
Lancet, published online, http://image.thelancet.com/extras/02art5336web.pdf
Leroy, A., 2004: Statistical predic tion of the weekly tropical cyclone activity in the
Southern hemisphere. Rapport de Stage de fin d’etude No. 939. Ecole Nationale de
la Météorlogie, Meteo France.
Langland, R., (2005) “Current issues in targeted observing”, invited talk at the
Fourth WMO International Symposium on Assimilation of Observations in
Meteorology and Oceanography. Prague, Czech Republic.
Leslie, L. M., M. Leplastrier, B. W. Buckley and L. Qi, 2005: Climatology of
meteorological "bombs" in the New Zealand region. Met. and Atmos. Phy., 89, 207214.
Lim, E.-P., and I. Simmonds, 2002: Explosive cyclone development in the Southern
Hemisphere and a comparison with Northern Hemisphere events. Mon. Wea. Rev.,
130, 2188-2209.
Limpasuvam V., and D.L. Hartmann, 1999: Eddies and the annular modes of climate
variability. Geophys. Res. Let., 26, No 20, 3133-3136.
Liu, Z.-Q. and F. Rabier, (2003). “The potential for high density observations for
numerical weather prediction: A study with simulated observations”. Quart. J. Royal
Met. Soc .129, pp. 3013-35.
Long, M. 2002. A climatology of extreme fire weather days in Victoria. Extended
abstracts, 2002 Weather Services Scientific Conference, Bureau of Meteorology,
Australia, May 2002.
Lorenz, D. J., and D. L. Hartmann, 2001: Eddy-zonal flow feedback in the Southern
Hemisphere. J. Atmos. Sci., 58, 3312-3327.
50
Lupo, A. R., J. J. Nocera, L. F. Bosart, E. G. Hoffman and D. J. Knight, 2001: South
American cold surges: Types, composites, and case studies. Mon. Wea. Rev., 129,
1021-1041.
Majumdar S.J., C.H. Bishop, B.J. Etherton and Z. Toth, 2002: Adpative sampling
with the Ensemble Transform Kalman Filter. Part II. Field Program Implementation.
Mon. Wea. Rev., 130, 1356 – 1369.
Marengo, J., A. Cornejo, P. Satymurty, C. Nobre, and W. Sea, 1997: C old surges in
tropical and extratropical South America: The strong event in June 1994. Mon. Wea.
Rev., 125, 2759–2786.
Marengo, J. A., T. Ambrizzi, G. Kiladis and B. Liebmann, 2002: Upper-air wave
trains over the Pacific Ocean and wintertime cold surges in tropical-subtropical South
America leading to freezes in southern and southeastern Brazil. Theoretical and
Applied Climatology, 73, 223-242.
McArthur, A.G., Cheney, N.P., and Barber, J. 1982. The fires of 12 February 1977 in
the western district of Victoria. CSIRO Division of Forest Research,
Canberra/Country Fire Austhority, Melbourne. 73pp.
Mills, G.A., 2005: A re-examination of the synoptic and mesoscale meteorology of
Ash Wednesday 1983. Aust. Met. Mag., 54, 35-55.
Mo, K. C., J. M. Paegle, 2001: The Pacific-South American modes and their
downstream effects. Int. J. Climatol, 21, 1211-1229.
Molteni, F., and T.N. Palmer, 1993: Predictability and finite -time instability of the
northern winter circulation. Quart. J Roy. Met. Soc., 119, 269-298.
Muller G. V. 2004: Variabilidad Interanual en la frecuencia de ocurrencia de heladas
en la Pampa Húmeda y su relacion con la circulación atmosférica asociada. PhD
Dissertation. Buenos Aires University.
Nicolini, M., J. Paegle and M.L. Altinger, 1987 : Numerical simulation of convection
and boundary layer convergence. Preprints, Second Int. Congress of Meteorology,
Buenos Aires, Argentina, Latin Amer. Meteor. Fed. and Am .Met. Soc., 8.5.1-8.5.7.
Nogués-Paegle, J. and K. C. Mo (1997). Alternating wet and dr y conditions over
South America during summer, Mon. Wea. Rev., 125, 2 , 279-291.
Nogues-Paegle, J., L. A. Byerle, and K. C. Mo, 2000: Intraseasonal modulation of
South American summer precipitation. Mon. Wea. Rev., 128, 837-850.
Parkinson, C. L., 2004: Southern Ocean sea ice and its wider linkages: Insights
revealed from models and observations. Antarctic Science, 16, 387-400.
Parmenter, F. C., 1976: A Southern Hemisphere cold front passage at the equator.
Bull. Amer. Met. Soc., 57, 1435–1440.
Perrin, G. , and I. Simmonds, 1995: The origin and characteristics of cold air outbreaks
over Melbourne. Aust. Met. Mag., 44, 41-59.
51
Pezza, A. B., and T. Ambrizzi, 2005a: Cold waves in South America and freezing
temperatures in São Paulo: Historical background (1888-2003) and case studies of
cyclone and anticyclone tracks. Brazilian J. Met., 20, 141-158.
Pezza, A. B., and T. Ambrizzi, 2005b: Dynamical conditions and synoptic tracks
associated with different types of cold surge over tropical South America. Int. J.
Climatol., 25, 215-241.
Pezza A. B., and I. Simmonds, 2005: The first South Atlantic hurricane:
unprecedented blocking, low shear and climate change. Geophys. Res. Let., 32,
L15712.
Rao, V.B.; S. Do Carmo, A. M. C. and S. H. Franchito. 2003: Interannua l variation of
storm tracks in the Southern Hemisphere and their connection with the Antartic
oscillation. Int. J. Climatol., 23, 1537-1545.
Raphael, M. N., 2003: Impact of observed sea-ice concentration on the Southern
Hemisphere extratropical atmospheric circulation in summer. J. Geophys. Res., 108,
4687, doi:10.1029/2002JD003308.
Rashid, H. A., and I. Simmonds, 2004: Eddy-zonal flow interactions associated with
the Southern Hemisphere annular mode: Results from NCEP -DOE reanalysis and a
quasi-linear model. J. Atmos. Sci., 61, 873-888.
Rashid, H. A., and I. Simmonds, 2005: Southern Hemisphere annular mode variability
and the role of optimal nonmodal growth, J. Atmos. Sci., 62, 1947-1961.
Rasmussen, E. A., and J. Turner, Eds., 2003: Polar Lows: Mesoscale Weather
Systems in the Polar Regions. Cambridge University Press, 622 pp.
Ratisbona, C. R., 1976: The climate of Brazil. Vol. 12, Climate of Central and South
America, W. Schwerdtfeger and H. E. Landsberg, Eds., World Survey of
Climatology, Elsevier, 219– 293.
Reason C.J.C., and M. Rouault, 2005: Links between the Antarctic Oscillation and
winter rainfall over western South Africa. Geophys. Res. Let., 32, L07705.
Seaman, R.S., 1994: Monitoring a data assimilation system for the impact of
observations. Aust Met. Mag., 43, 41-48.
Simmonds, I., 2003: Modes of atmospheric variability over the Southern Ocean. J.
Geophys. Res., 108, 8078, doi: 10.1029/2000JC000542.
Simmonds, I., and T. H. Jacka, 1995: Relationships between the interannual
variability of Antarctic sea ice and the Southern Oscillation. J. Climate, 8, 637-647.
Simmonds, I., and K. Keay, 2000: Mean Southern Hemisphere extratropical cyclone
behavior in the 40-year NCEP -NCAR reanalysis. J. Climate, 13, 873-885.
Simmonds, I., K. Keay and E.-P. Lim, 2003: Synoptic activity in the seas around
Antarctica. Mon. Wea. Rev., 131, 272-288.
52
Simmonds, I., and J. C. King, 2004: Global and hemispheric climate variations
affecting the Southern Ocean. Antarctic Sci., 16, 401-413.
Simmonds, I., and H. A. Rashid, 2001: An investigation of a dramatic cold outbreak
over southeast Australia. Aust. Met. Mag., 50, 249-261.
Sinclair, M. R., 1995: A climatology of cyclogenesis for the southern hemisphere.
Mon. Wea. Rev., 123 (6), 1601-1619.
Shapiro, M. A., and A.J. Thorpe, 2004: THORPEX International Science Plan,
Version III: 2 November 2004, WMO/TD-No.. 1246, WWRP/THORPEX No. 2,
World Meteorological Organization, Geneva, 51pp.
Tennant, W.J. and C.J.C. Reason, 2005: Associations between the Global Energy
Cycle and Regional Rainfall in South Africa and Southwest Australia. J. Climate, 18,
3032-3047.
Torres, J. C. and M. Nicolini, 1999: "Analysis of a mesoscale convective system
centered over the Río de la Plata". Aust. Met. Mag., Vol. 48, 261-272.
Thompson, D. W. J., J. M. Wallace and G. C. Hegerl, 2000: Annular modes in the
extratropical circulation. Part II: Trends. J. Climate, 13, 1018-1036.
Turner, J., 2004: The El Niño-Southern Oscillation and Antarctica. Int. J. Climatol.,
24, 1-31.
Turner, J., D. Bromwich, S. Colwell, S. Dixon, T. Gibson, T. Hart, G. Heinemann, H.
Hutchinson, K. Jacka, S. Leonard, M. Lieder, L. Marsh, S. Pendlebury, H. Phillpot,
M. Pook and I. Simmonds, 1996: The Antarctic First Regional Observing study of the
Troposphere (FROST) project. Bull. Amer. Met. Soc., 77, 2007-2032.
Velasco, I., and J. M. Fritsch, Mesoscale convective complexes in the Americas. J.
Geophys. Res., 92, D8, 9591-9613, 1987.
Vera, C. S., and P. K. Vigliarolo, 2000: A diagnostic study of cold-air outbreaks over
South America. Mon. Wea. Rev., 128, 3-24.
Wallace, J.M., G-H Lim and M.L. Blackmon, 1988: Relationship between cyclone
tracks, anticyclone tracks and baroclinic waveguides. J. Atmos. Sci., 45, 439-462.
Watkins, A. B., and I. Simmonds, 1995: Sensitivity of numerical prognoses to
Antarctic sea ice distribution. J. Geophys. Res., 100, 22,681-22,696.
Webster, P., and J. Holton, (1982). “Cross –Equatorial Response to Middle-latitude
Forcing in a Zonally Varying Basic State”. J. Atmos. Sci. 53, pp.2365-2378.
Wheele r, M. C, and H. H. Hendon, 2004: An all-season real-time multivariate MJP
index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132 ,
1917-1932.
53
Wheeler M.C., and J.L. McBride, 2004: Australian – Indonesian Monsoon. Chapter
5 of Intraseasonal Variability in the Atmopshere-Ocean Climate System. W.K.M.
Lau and D. Waliser (eds). Praxis. Springer Berlin Heidelberg.
Yanase, W., and H. Niino, 2005: Effects of baroclinicity on the cloud pattern and
structure of polar lows: A high-res olution numerical experiment. Geophys. Res. Let.,
32, L02806, doi: 10.1029/2004GL020469.
Young, I. R., 1999: Seasonal variability of the global ocean wind and wave climate.
Int. J. Climatol., 19, 931-950.
Yuan, X., 2004: ENSO -related impacts on Antarctic sea ice: A synthesis of
phenomenon and mechanisms. Antarctic Sci., 16, 415-425.
54
THORPEX core objectives (from the THORPEX
International Science Plan, WMO/TD – No. 1246, WWRP/THORPEX No.2)
APPENDIX
I:
•
Increase knowledge of global-to-regional influences on the initiation,
evolution and predictability of high-impact weather. This objective includes
research on: i) the excitation of Rossby wave trains by extratropical cyclone
development, large-scale topography, continent/ocean interfaces, organised
tropical and extratropical convective flare-ups, and the role of these processes
in the consequent development of high-impact weather; ii) the dependence of
predictive skill on inter-annual and sub-seasonal climate variability, e.g., El
Niño Southern Oscillation (ENSO); Pacific North-Atlantic oscillation (PNA);
North-Atlantic Oscillation (NAO); Arctic Oscillation (AO); monsoon
circulations; iii) the relative contribution to the limits of predictive skill by
uncertainty in observations, data assimilation, model formulation and
ensemble prediction system design.
•
Contribute to the design and demonstration of interactive forecast systems that
allow information to flow interactively among forecast users, numerical
forecast models, data-assimilation systems and observations to maximise
forecast skill. As an example, targeted observing strategies incorporate
dynamical information from the numerical forecast model itself to identify
when, where, and what types of observations would provide the greatest
improvement to specific weather forecasts of societal, economic, and
environmental interest.
•
Contribute to the development of advanced data assimilation and ensemble
prediction systems. This effort will include: i) improving the assimilation of
existing and experimental observations, including observations of water in its
three phases and atmospheric composition (e.g., ozone; aerosols); ii)
developing adaptive data assimilation and targeted-observing strategies; iii)
incorporating model uncertainty into data-assimilation systems and in the
design of ensemble prediction systems; iv) evaluating the utility of multimodel ensemble prediction systems.
•
Develop and apply new methods that enhance the utility and value of weather
forecasts to society, economies and environmenta l stewardship through: i)
user-specific probabilistic forecast products; ii) the introduction of interactive
procedures that make the forecast system more responsive to user needs; iii)
the design of and training in the use of user-specific forecast products.
Research will identify and assess the societal/economic costs and benefits of
THORPEX recommendations for implementing interactive forecast systems
and improvements in the global observing system.
•
Carry out THORPEX Observing-System Tests (TOSTs) and THORPEX
Regional field Campaigns (TReCs). TOSTs: i) test and evaluate experimental
remote sensing and in-situ observing systems, and when feasible, demonstrate
their impact on weather forecasts; ii) explore innovative uses (e.g., targeting)
of operational observing systems.
TReCs are operational forecast
demonstrations contributing to the design, testing and evaluation of all
components of interactive forecast systems. They are organised and
coordinated by regional consortia of nations under THORPEX Regional
55
Committees, e.g., Europe; Asia; North and South America; Southern
Hemisphere. TReCs address forecasts of regional weather systems, e.g., arctic
storms and cold-air outbreaks; extratropical cyclones over Europe, Asia, and
North America; warm-season heavy precipitation over Asia; organized
equatorial
convection
flare-ups;
tropical-to-extratropical
cyclone
transformations. TReCs require collaboration between Regional Committees.
THORPEX will explore the opportunities to carry out TReCs in conjunction
with major international programmes such as the International Polar Year
(IPY) and the African Monsoon Multi-disciplinary Analysis (AMMA).
•
Demonstrate all aspects of THORPEX interactive forecast systems, over the
globe for a season to one year to assess the utility of improved weather
forecasts and user products. This includes the THORPEX Interactive Grand
Global Ensemble (TIGGE) that integrates developments in observing systems,
targeting, adaptive data assimilation, model improvements, forecast user
requirements, and a multi-model/multi-analysis ensemble prediction system.
•
Coordinate THORPEX research with the World Climate Research
Programme: Coordinated Observation and Prediction of the Earth System
(WCRP/COPES) and the mesoscale/microscale community to address the
observational and modelling requirements for the prediction of weather and
climate for two weeks and beyond.
•
Facilitate the transfer of the results of THORPEX weather prediction research
and its operational applications to developing countries through the WMO by
means of appropriate training programmes.
56