BORA
ON THE NORTHERN ADRIATIC COAST DURING THE ALPEX SOP
20-25 MARCH 1982
BRANKA IVANČAN-PICEK, VIŠNJA VUČETIĆ
Hydrometeorological Institute of Croatia, Zagreb
Abstract: Results from
investigation of the bora during the period of 20-25 March 1982 are characterized
by a large variability od bora strenght, direction and duration. The two
distinct wind speed maxima indicate two different origins of bora genesis.
The first bora
(21 March) was caused by a cold air stream, followed by the Genova cyclogenesis
process. The second bora situation (23-25 March) is connected with the central
Europe anticyclone and the cold NE flow – upstream of the Dinaric Alps.
Assuming that the bora has a certain similarity with the
supercritical flows described by hydraulic theory, the parameters of the real
flow could be successfully predicted on the entire northern Adriatic in the
case with the strongest bora.
Key words: bora layer, hydraulic theory, ALPEX SOP,
temperature inversion
SPATIAL
BORA VARIATIONS IN RELATION TO COLD AIR OUTBREAK AND SURFACE PRESSURE GRADIENT
ALICA BAJIĆ
Hydrometeorological Institute of Croatia, Zagreb
Abstract: Bora wind on the northern Adriatic is associated with an
upstream cold air outbreak although the most pronounced outbreaks are not
always connected with the strongest bora storms.
During the cold
season between 1973 – 1982 in Zagreb only 3 strong cold air outbreaks were
associated with severe bora in the northern Adriatic. In these cases bora onset
is almost simultaneous along the entire northern Adriatic coast, it lasts 1 – 2
days and its strength is nearly equal at all considered meteorological
stations.
The cut-off processes and blocking circulation pattern
connected with the less pronounced long lasting cold air outbreak resulted with
the longest bora duration.
In 12 considered severe bora situations a drop of
temperature due to cold air outbreaks is followed by the characteristic
mesoscale pressure field with low pressure area along the coast and islands and
higher pressure in the upstream bora region.
The estimated correlation coefficient between the cross
mountain sea level pressure difference Δp and mean hourly bora velocity Vs
show the stohastic significancy in all 12 severe bora cases. The relationship
between ∆p and Vs is stronger using Senj data than data
measured at Omišalj and Pula what is a consequence of Senj′s specific
location in relation to the mountain barrier. In 3 most pronounced cold air
outbreak situations the constants in the correlation expression obtained using
Senj data are almost the same. They enable us to estimate the mean hourly bora
wind speed in Senj for predicted Zagreb – Senj pressure differences in similar
weather situations.
Key words: severe bora wind, the northern Adriatic, surface
pressure difference
MOUNTAIN
DRAG AND SURFACE PRESSURE VARIATIONS
DURING SEVERE
ADRIATIC BORA STORMS
VESNA JURČEC
Hydrometeorological Institute of Croatia, Zagreb
Abstract: The observed sea-level pressure variations in Senj and
pressure difference Zagreb-Senj are presented for 197 observations in 17 case
studies considering only the most severe bora storms in Senj from the period
1957 – 1986. Using Smith′s (1985) hydraulic theory with the continuous
stratification model, mountain drag is calculated predicting the pressure
difference across the mountain in these bora events.
It is shown
that the theory can predict well the pressure difference in cases when the
atmospheric state is close to the idealized modelling structure. However, due
to constraints in the presented model there is a relatively small sampling of
such cases, although the real atmosphere allows some reasonable modifications
of hydraulic parameters which would be unnecessary if the flow were truly
two-dimensional and steady.
Key words: severe bora, hydraulics, mountain drag, surface
pressure variations, hydrostatic bora layer, Adriatic storms.
ISENTROPIC
HIGH RESOLUTION TIME CROSS-SECTION BASED
OD POLYNOMIAL
HYDROSTATIS ADJUSTMENT
DRAŽEN GLASNOVIĆ
Abstract: The objective routine which includes the polynomial
hydrostatic adjustment technique has been developed and applied to construct a
high resolution time cross-section in the isentropic coordinate system. The
mechanism to insure vertical consistency is based on the reciprocal of thermal
stability parameter ∂p/∂Θ and its derivatives of a higher
order. The Hermitian scheme is used to interpolate, but there it is essentially
modified and applied with polynomials up to the fifth order. It enables the
elimination of any computation in finite difference approximations over the
layers of various thicknesses and uses only the observed values in the vertical
interpolation points even in computing their vertical changes.
To obtain a
more detailed insight into the prevailing processes the vertical profiles of
Brunt-Vaisala frequency, Scorer parameter, vertical acceleration,
nonhydrostatic coefficient and the wind are computed. The successfulness and
applicability of presented methods are tested in two events of severe weather
development.
Key words: objective analysis, time cross-section,
polynomial, isentropic, hydrostatic adjustment, vertical initialization,
successive interpolation
WIND
PERSISTENCE AT THE ADRIATIC
DRAŽEN POJE
Zagreb, Savska c. 1
Key words: Wind
persistence, Adriatic
UPPER EXTREMES
OF THE RAINLESS PERIODS IN
BANAT AND SREM
VERICA NEŠKOVIĆ-ZDRAVIĆ, EMIR ZELENHASIĆ, MILICA BERIĆ
University of Novi Sad – Faculty of Agriculture
Institut for Water Resources Development and Management
Abstract: The Zelenhasić-Todorović method of describing and analysing the extremes in hydrology, which is modified for extreme dry weather intervals, is presented. The basic assumption of the method is that the extreme dry weather intervals are independent, identically distributed random variables and that their occurrence is subject to the Poisson probability law. According to the theory of supreme of random number on random variables, the explanation and analyses of the largest rainfall deficit and the largest drought is obtained by the Zelenhasić-Todorović method for given location. Application of the methods is performed on the records of 13 meteorological stations in Banat and Srem, Yugoslavia, for period 1949-1985. The relations between dry weather intervals and field crops yield, as well as between dry-weather intervals and mean daily air temperature and relative air humidity during these intervals are investigated.
Key words: rainless period, dry weather intervai, drought, yield
DETERMINANTION
OF BORA TURBULENCE LOCATION AND INTENSITY
KATARINA STANKOVIĆ
Savezna uprava za kontrolu letenja
Zrakoplovni meteorološki centar, aerodrom Zagreb
Abstract: This paper considers the possibility of determining bora turbulence intensity from the fields empirical index derived by Dutton as the “best” CAT indicator. The reason for such an analysis lies in the fact that bora turbulence has probably the same or similar physical mechanism of generation as clear air turbulence. The attempt to locate bora turbulence shows good coincidence with observed bora turbulence and turbulence located by the empirical index.
Key words: bora turbulence, Kelvin-Helmholtz instability, gravity shear waves.
A METHOD OF VERY –
SHORT – RANGE LOCAL FORECAST
OF AIR TEMPERATURE
BORIVOJ ČAPKA, ZORAN REBAC
Republički hidrometeorološki zavod SR Hrvatske, Zagreb
Abstract: This paper presents method of very-short-range local forecast of air temperature. The method is based on the correction of the actual temperature values by their interhourly differences, dependent on mean daily cloudiness as well as its amount at the moment of forecasting. Data from the 1960-1988 period at the Zagreb-Maksimir observatory are used. This method is tested by means of the root mean square error. Verification has shown the applicability of the method for clear days and for days with changing cloudiness. For cloudy days the method doesn't give sufficient reasonable.
Key words: Very-short-range local forecasts, nowcasting, air temperature
CLOUD SYSTEM IDENTIFICATION
BASED ON DIGITAL
SATELLITE DATA
BOJAN LIPOVŠĆAK
D.Sc. Thesis-Zagreb University, Faculty of Natural Sciences
Abstract: The thesis shows the basis and results of automatic cloud and underlaying surface classification of NOAA-7 AVHRR (Advanced Very High Resolution Radiometer) satellite data. This paper demonstrates the initial classifier learning and the procedure of data base renovation.
Cloud radiation depends on physical properties of various cloud types, density and size of cloud particles and cloud structure (ice and water). The physical characteristics are measured in five spectral bands: visible (channel 1 and 2), near infrared (channel 3) and infrared (channel 4 and 5), which make the construction of cloud classifier possible.
The results from satellite measurement of cloud types in one spectral area can be represented by normal distribution. The multivariate normal density function is used as a discriminant function for cloud separation and identification of Earth surface from clouds. Three channels data are used and therefore a three-dimensional normal density function is used. In this case, for each discrimination function it is necessary to define 9 elements: 3 means and 6 covariances. The number of discrimination function depends on the number of classes into which data are classified. A maximum likelihood classifier is applied. The classification is made with two algorithms using the Bayes rule of maximum likelihood and the threshold method. Threshold is defined in advance, prior to the classification, and all the values of discrimination functions which are minor to the threshold value are classified in the rejact class.
The data received during passes number 9514, 12522 and 13323 of the NOAA-7 satellite are separated for the initial classifier learning. The data received during pass No. 9514 consist of the visible spectral part (channel 1), while passes 12522 and 13323 consist of visible, near infrared and infrared spectral parts (channels 1, 3 and 4). Discrimination function parameters are computed from the grey level values in different channels for the same pixel.
Analysis of one-dimensional Gaussian distributions indicates the difficulties in classification of one-channel spectral data. Therefore, bi-dimensional histograms are made, on which the clustering of data is evident. The most distinct group is the classes sea and Cumulonimbus clouds. The class Cirrus is easily distinguished on all the three pictures, and the greatest overlapping is in case of channels 1 and 3 for Altostratus and Cumulus. The separation of classes is better distinguished in bi-dimensional, than in one-dimensional histograms. The location of mean values of different classes can be presented in three-dimensional space defined by grey scale values in three channels.
The normalized and the Jeffries – Matusita (J-M) distances are used for the statistical separability evaluation among classes. The separability among classes is calculated by use of a theoretical J-M / normalized distance ratio. The J-M values which are minor to the theoretical value are in the area “correct classification questionable”. J-M values for Cumulonimbus and Cumuluscongestus, Stratocumulus and Cirrus, as well as for Altostratus and Cumulus suggest the possibility of erroneous classification.
The cloud classifier is checked in three steps. The first step is the testing of clasification algorithms, for one – dimensional normal distribution. The identification is made for clouds and for underlaying surface only. The next step is the use of the three-dimensional normal distribution as a discriminant function for the classification of the training area. The final step is the use of the classifier in the identification of all the pictures of NOAA-7 passes No. 9514 and 12522.
The cloud classifier orders the clouds in 12 classes. Following the classification each pixel is presented by a defined colour.
The presented algorithm is the key part of the application software developed with the INTERACT project Hyderabad – India.
LONG GRAVITY WAVES CAUSED
BY CYCLONES AND FREE
OSCILLATIONS (SEICHES) ON THE ADRIATIC
MILAN HODŽIĆ
D.Sc.Thesis – Belgrade University, Faculty of Physics. Beograd, 1986.
Abstract: The aim of this thesis has been to study and mathematically determine the connection between storm surges and atmospheric disturbances on the Adiratic. A statistical procedure of multiple regression and variance regression has been used here. The procedure is applied to the series of all registered maximum heights of storm surges and to cyclonic-meteorological parameters for a 5-year period of analysis of synoptic charts and meteorological data (1976-1980). We have therefore calculated the forecast equations of storm surges in the form of regression equations with different combinations from five to nine cyclonic-meteorological parameters.
In this thesis the tracks of 428 cyclones have been analysed with the use of Radinović΄s scheme of tracks over the west Mediterranean, enhanced by three new tracks (crossing the Adriatic Sea). Consideration of 463 cold fronts passing over the Adriatic has led to a new classification into ten particular types.
The basic periods of bay oscillations in the 19 bays and ports on the Adriatic have been calculated and analysed. Finally, the problem of bay oscillations has been solved for Vela Luka where amplitudes greater than one meter have caused much damage.
FACTOR ANALYSIS OF LOCAL
CLIMATIC FIELDS WITHIN THE
GLOBAL CIRCULATION
KREŠO PANDŽIĆ
D.Sc.Thesis – Zagreb University, Faculty of Natural Sciences
Abstract: In addition to consideration of the Principal Component Analysis (PCA) technique and annual temperature and precipitation courses in Croatia, the main objective of the paper was to investigate anomalies of considered parameters within circulation on the broader European-Northern Atlantic region. In order to reach these goals, the Principal Component technique was applied. Advantages of this method are in the fact of explaining more observed variables in terms of a small number of latent variables (factors).
Here the considered problem of a comparison of real weather conditions on the territory of Croatia and global circulation anomalies above Europe-Northern Atlantic region is very important for understanding and forecasting atmospheric processes. The results can be divided into two groups.
More important is the explanation of a relation between local weather anomalies and global circulation after they are typed. The results in this group are new, and until now, unknown. They will directly contribute to the long range weather forecast problem in Croatia, because anomalies of global circulation can be calculated nowadys.
The second group consists of results analysing the annual course of considered fields, about which some knowledge thorough not as reliable did exist previously. New here is the number of physical factors which mainly generate the annual course and an objective division of Croatian territory into quite reasonable subregions.