Volume 8 [2006]

The El Nino-Southern Oscillation is the dominant pattern of short-term climate variation, and
is therefore of great importance in climate studies. Some recent studies showed the
teleconnection between stream flow and the El-Nino Southern Oscillation (ENSO) of the
equatorial Pacific Ocean. This paper presents an overview of the relationship between ENSO
and stream flow in the Brahmaputra-Jamuna and the potential for wet season flow
forecasting. This seasonal forecast of stream flow is very invaluable to the management of
land and water resources, particularly in Bangladesh to improve the predictability of severe
flooding. Over the years, large investments have been made to build physical infrastructure
for flood protection, but it has been proved that it is not feasible, both economically and
technically, to adopt solely structural mitigation approach. The choice of non-structural
measures in this country focused mainly on flood forecasting because many of the nonstructural
measures including flood plain zoning, compulsory acquisition of flood prone land,
relocation etc have also been proved inappropriate for Bangladesh.
The aim of this research is to find out an effective and long-lead flow forecasting method with
lead time greater than hydrological time scale, using El Nino-Southern Oscillation index.
Some studies indicate that SST can be predicted one to two years in advance using several
ocean/ coupled ocean atmosphere models, therefore the ability to predict flow patterns in
rivers will be highly enhanced if a strong relationship between river discharge and ENSO
exists, and is quantified. With this view, to assess the strength of teleconnection between river
flow and ENSO, at first correlation analyses between ENSO indices of any year and wet
season flow of that year have been done. Here sea surface temperature (SST) has been
used as ENSO index. This correlation analysis demonstrates a noteworthy relationship
between natural variability of average flow of the months July-August-September (JAS) of the
Brahmaputra-Jamuna River with SST of the corresponding months. Then discriminant
prediction approach, also known as “Categoric Prediction” has been used here for the
assessment of long range flood forecasting possibilities. This approach will be able to forecast
the category of flow (high, average or low) using the category of predictor (predicted SST) at
a sufficient lead time. In order to judge the forecast skill, a synoptic parameter “Forecasting
Index” has also been used. This discriminant approach will improve the forecasting lead-time
while the hydrologic forecast through rainfall-runoff modeling could provide a lead time on the
order of the basin response time, which is several days or so. As the Ganges–Brahmaputra
river basin is one of the most populous river basins of the world and is occupied by some
developing countries of the world like Bangladesh, any reduction in the uncertainty about the
flood in the Brahmaputra-Jamuna River would contribute a lot to the improvement in flow
forecasting as well as to the economic development of the country.

A teleconnection pattern has been identified between Eastern Mediterranean and Northwestern
Europe in the geopotential fields of the upper troposphere (hereafter will be referred
to as EMP) during winter and it seems to play significant role in the large-scale atmospheric
circulation over Mediterranean region with distinctive temporal and spatial characteristics.
Since teleconnections are significant components of the natural climatic variability, any future
climatic changes are expected to influence their structure and intensity and subsequently their
implications on the regional climate.
Taking into account these findings, the aim of this study is to investigate the existence of EMP
and its potential spatial variations in the future, in connection with the IPCC Emission
Scenarios of greenhouse gases. For this purpose, gridded daily data of geopotential height at
500 hPa have been used. For the present climate, the datasets derive from NCEP/NCAR
archive and cover the period 1958 to 2003 on 2.5°x 2.5° latitude by longitude grid. For the
future, the datasets derive from HadAM3P model for the period 2070-2100, on the basis of
two different IPCC scenarios. Two different approaches were employed on a seasonal basis
in order to identify the centres of action for this pattern: Rotated Principal Component Analysis
and Correlation analysis.
First, the validation of the model results (control run) against the gridded data of NCEP/NCAR
for the present period 1960-1990 demonstrated that the two datasets agree quite well. Then,
it was found that the EMP exists in the future (2070-2100) under different scenarios of
greenhouse gases emissions and subsequent global warming but not in its present form. For
the scenario that predicts global temperature increase by 2.3°C, the pattern forms with a
northeastward shift of the two poles relative to location in the present climate. If the global
temperature increase is higher (of 3.3°C), the whole EMP is shifted to the southeast by almost
10°, as compared to its present form. These results are collaborated by both approaches
employed.

A method for automated circulation patterns (CPs) definition and classification is presented,
based on optimized fuzzy rules. The target of the method is to provide a basis (daily classified
CPs) for downscaling of the most common climate data, i.e. precipitation and temperature.
Therefore the presented classification method is objective providing CPs that explain the
dependency between the large-scale atmospheric circulation and the surface climate. Thus
the downscaling can be done by means of downscaling models with parameters depending
on the CPs. The CPs are defined using 700 hPa geo-potential field anomalies. Fuzzy rules
are described by the position of high- and low-pressure anomalies. The fuzzy rules are
obtained automatically, using an optimization for the performance of the classification. For the
precipitation, the performance of the classification is measured by rainfall frequencies and
rainfall amounts conditioned on the CPs. Thus, the task is to define wet or dry CPs. For the
temperature, the deviation from the average long-term annual cycle is used. In this way, warm
or cold CPs are identified. The presented method produces physically realistic CP definitions.
With the help of these definitions, the observed (historical) pressure fields can be classified
for twelve precipitation stations and four temperature stations inside and around the
Mesochora catchment in Greece.

Several epidemiological studies have shown an association between particulate air pollution
and health effects. Suspended particulates and more specifically the inhalable PM10 fraction
appear to cause respiratory health effects and heart diseases. Furthermore, particulate
pollution is of paramount importance in areas with open-pit mines and especially when it is
combined with raw lignite transfer and combustion in power stations through the suspension
of particles and stack emissions, respectively. The penetration of particles into respiratory
track is a function of the size of the particles and thus, it is more likely for the finer PM2.5
fraction to reach the deepest of the lugs.
The fast economic growth the last decades has resulted in an increase of the sources of
pollution not only in large metropolitan areas but also in medium-sized urban areas like the
city of Kozani, Greece. It is the most densely populated city in the area of West Macedonia
affected by urban particulate matter originated from local and stationary sources, from
regional and long-range transport, and from street dust resuspension. Kozani is located a few
kilometers away from lignite power stations that contribute to about 70% of the total electrical
energy produced in Greece. Dust emissions seem to be the most serious problem in the area,
as the measured ambient concentrations of suspended particles are at high levels and
exceed local and international standards.
In this study PM10 and PM2.5 concentrations are presented. The measurements have been
carried out, from April to December 2002, by the Lab of Atmospheric Pollution and
Environmental Physics (LAPEP) of Technological Education Institute of West Macedonia in
the commercial centre of the city of Kozani. The temporal variation of PM10 and PM2.5
concentrations was studied and allowed a further insight on the factors affecting the
measured ambient particulate levels. PM2.5 – PM10 correlation and PM2.5/PM10 ratios were
investigated and compared to those in the literature together with the factors affecting their
diurnal variation. The pollution levels were also detected in process of the experimental time
series data by fractal dimension. Generally, fractal analysis is able to detect the data set
complexity by scaling empirical data using threshold values. These values define the levels of
air pollution episodes. The method presented in this study, is the transformation of PM10 and
PM2.5 concentrations into a set of points whose dimension was estimated by box counting.
This technique has estimated the fractal dimension of both the time series by the relationship
between data variance and time scale.