COMPARISON OF IDF CURVES OF EXTREME RAINFALL

DOWNSCALED FROM DESIGN VALUES OF 1-DAY RAINFALL

ASSESSED BY DIFFERENT APPROACHES
Mrta Bara1, Silvia Kohnov2, Ladislav Gal2, Jn Szolgay2, Kamila Hlavov2
1) Institute of Hydrology, Slovak Academy of Sciences, Raianska 75, 831 02 Bratislava 3,
Slovak Republic,
bara@uh.savba.sk,
2) Department of Land and Water Resources Management, Faculty of Civil Engineering,
Slovak University of Technology, Radlinskho 11, 813 68 Bratislava 1, Slovak Republic,
silvia.kohnova@stuba.sk, ladislav.gaal@stuba.sk, jan.szolgay@stuba.sk, kamila.hlavcova@stuba.sk

Abstract
In this study, intensity-duration-frequency (IDF) curves of extreme rainfall in Slovakia were
examined. The IDF curves were derived from design values of 1-day rainfall amounts, which
were assessed by different approaches: by local and regional estimation methods. To derive
the design values of rainfall of higher temporal resolution, the simple scaling method was
applied. This method allows for an estimation of the design values of rainfall of selected
recurrence intervals and durations shorter than a day by using only the daily data from the
regular network of non-recording raingauges (or estimated/interpolated/modelled data). For
the analysis, three meteorological stations located in different regions of Slovakia were
selected, in which continuous measurements of rainfall amounts for a long period were
available. The IDF curves derived by different approaches were compared to one another and
to historical estimates.
Keywords: IDF-curves, extreme rainfall, simple scaling of rainfall, 1-day maxima of
precipitation totals, design values of rainfall intensity
Introduction
The intensity-duration-frequency (IDF) characteristics, or the design values of extreme
rainfall are of very great importance in engineering hydrology, such as input data for
hydrological modeling, for the prediction of flood events, or for planning and design in water
resources management. The establishment of IDF curves goes back to the 1930s (Bernard,
1932). Since then, different forms of relationships have been constructed for several regions
of the world. Since the 1960s, the regional properties of IDF relationships have been studied
in several countries, and maps have been constructed to provide the rainfall intensities or
precipitation totals for various annual frequencies and durations. IDF relationships were
mostly estimated by at-site statistical analysis of rainfall data for different durations of
intensive rainfall. IDF curves received considerable attention in engineering hydrology over
the past decades. Approaches based on statistical analysis of data were developed, e.g. Bell
(1969) and Chen (1983) derived the IDF formulae for the United States, Baghirathan and
Shaw (1978), Gert et al. (1987) and Niemczynowicz (1982) developed IDF formulas for
ungauged sites. More recently, Sivapalan and Bloeschl (1998) suggested a method of
constructing IDF curves based on the spatial correlation structure of rainfall. Koutsoyiannis et
al. (1998) proposed a novel approach for the formulation and construction of IDF curves,

which constitutes an efficient parameterization by facilitating the description of the

geographical variability and rationalization of IDF curves and allowing for the incorporation
of data from non-recording stations for constructing IDF curves at ungauged sites. Willems
(2000) analyzed the IDF properties of extreme precipitation for two different storm types. He
estimated different mixtures of these two types of storms both for summer and winter
conditions, which brought new elements into the current understanding of what determines
IDF curves and their scaling properties. Minh Nhat et al. (2007) developed a regional IDF
relationship for ungauged locations based on the scaling theory. Acar et al. (2008) used
a multilayer perceptron artificial neural network model to assess IDF relationships for short
duration rainfalls.
The first attempts to construct regional IDF curves for Slovakia were made by Dub (1950).
amaj and Valovi (1973) presented a comprehensive IDF study based on 68 stations
covering the area of Slovakia using data mostly from the period 19311960. Their results
were re-evaluated by Urcikn and Horvth (Urcikn and Imrika, 1986); however, the
analysis mostly concentrated on the different formal presentations of IDF curves. They also
proposed a method for the spatial interpolation of IDF curves for sites with no direct
observations. In the last years, the properties of extreme rainfall in Slovakia were studied e.g.
by Fako et al. (2000), Cebulak et al. (2000), Gal and Lapin (2000) and Kohnov et al.
(2005). In these studies only smaller regions were examined, but since the 1980s the shortterm rainfall and its IDF characteristics has not been globally processed for the whole teritorry
of Slovakia.
Precipitation data with sufficient temporal resolution necessary for estimation of design
precipitation totals are available from a limited number of raingauges with continuous
recording. The most commonly available data are daily rainfall amounts, but for engineering
purposes, data with higher time-resolution are often required. The simple scaling method
allows for estimating design values of rainfall at selected recurrence intervals and durations
shorter than a day by using only daily precipitation data, available from a denser network of
non-recording raingauges.
In this paper, the possibility of using the simple scaling method for the estimation of the
IDF curves of extreme rainfall was tested. For the analysis, three meteorological stations
located in different regions of Slovakia were selected, in which continuous measurements of
rainfall amounts for a long period were available. In the three test stations the T-year quantiles
of 1-day precipitation maxima were assessed by different approaches: by local at-site
estimation, by the region-of-influence (ROI) method and by a conventional regionalization
approach of Hosking and Wallis (1997). The IDF curves for different durations and
recurrence intervals were derived by downscaling the quantiles of the 1-day precipitation
maxima using the simple scaling method. The IDF curves derived by different approaches
were compared to one another and to historical estimates of amaj and Valovi (1973).
Input data
Two different datasets were used for the various sub analyses in this study; one for assessing
the scaling exponents and another for the regional frequency analysis of rainfall.
In previous studies (Bara, 2010; Bara et al, 2010), the scaling properties of rainfall in
Slovakia were analyzed and the scaling exponents for 56 meteorological stations were
derived. The data used for assessing the scaling exponents were quantiles of rainfall
intensities of 8 durations ranging from 5 to 180 minutes estimated empirically by amaj and
Valovi (1973). The historical quantiles were preprocessed and completed by daily rainfall
amounts, taken from the archive database of the Slovak Hydrometeorological Institute. The
return periods of the quantiles of the 1-day (1440 min) precipitation duration were estimated

using the GEV distribution function. Generally, the periods of observation range from the
1930s to the 1960s, and the average length of observation is approximately 17 years.
For the region of influence (ROI) method and for the purposes of the regionalization
approach according to Hosking and Wallis (1997) 1-day precipitation totals were used (Gal,
2006). Data from 56 meteorological stations from the whole territory of Slovakia were
selected (with an average density of 1 station per 900 km2). The observation periods range
from 1961 to 2003 at the majority of the stations; moreover, data from the period 19511960
that is available at some stations were also included in the basic data set. The shortest
observation records are 35 years long (in 2 cases); on the other hand, the longest ones consist
of 53 years (at 13 sites). The most common value of the record lengths is 43 years (15
occurrences). For the analysis, maxima of 1-day precipitation amounts for three different
seasons: cold, warm and the calendar year were used (warm season: AprilSeptember).
For the comparison of design values of rainfall estimated by local and regional approaches
three test stations from Slovakia were chosen: Hurbanovo, Liptovsk Hrdok and Koice. The
test sites were treated as there were no direct observations, although daily precipitation
maxima were available at these stations. The stations were selected according to the following
criteria: i) they characterize different geographical locations; ii) they more or less represent
the mean, minimum and maximum value of the scaling exponents; iii) they are located in
different regions according to the regionalization of Gal (2006). The selected test stations are
displayed on Fig.1.

Fig.1 Schematic map of the analysed stations (source: www.oslovensku.sk).

The station of Hurbanovo is situated in the south-western part of Slovakia, on the

Danubian Lowland. The altitude of the station is 116 m a.s.l. The Liptovsk Hrdok station is
located in the northern part of the country, in the Liptov Basin at an altitude of 640 m a.s.l.,
and is bordered by the Low Tatra Mountains to the south and partially bordered by the
Western Tatras (to the west) and the High Tatra Mountains (to the north). The Koice station
is positioned in the south-eastern part of Slovakia, in the Koice Basin at the root of the
Slovak Ore Mountains, at an altitude of 230 m a.s.l.
Assessing the 1-day maximum precipitation totals
In the test stations the T-year quantiles of 1-day maximum precipitation totals were assessed
by three different approaches: by a local at-site estimation without regional approach, by a

conventional regionalization approach of Hosking and Wallis (1997) (HW) and by the
region-of-influence (ROI) method.
The Hosking and Wallis (1997) regionalization approach. The regional frequency analysis
proposed by Hosking and Wallis (1997) is a method of probability modeling of heavy
precipitation within regions with a fixed composition, and according to an index value
methodology. In general, this method consists in four main steps: screening of the data,
identification of the homogeneous regions, choice of a frequency distribution and estimation
of the regional dimensionless quantiles and the actual at-site quantiles. Based on the
methodology of Hosking and Wallis (1997) a regional frequency analysis of heavy
precipitation in Slovakia has been carried out in Gal (2006). Gal (2006) in his
regionalization process delineated three homogenous regions within Slovakia, as a results of a
combination of cluster analysis (objective method) and a process-based regionalization
(subjective method). The homogeneity of the regions was assessed by the H1 measure of
Hosking and Wallis (1997). The H1 values indicated the regional homogeneity of all the
regions. The regional dimensionless growth curves of 1-day precipitation maxima in each
region were derived. The Z-test by Hosking and Wallis (1997) was applied to select an
appropriate regional frequency distribution. Generally, the goodness-of-fit of the proposed
distribution functions for rainfall maxima in warm season (Gal, 2006) showed that the
generalized extreme value (GEV) distribution was a suitable regional distribution function.
The region of influence (ROI) method. In this method, there is no need to delineate fixed
boundaries between the regions. Instead of this, pooling groups (regions) are defined in a
flexible way (Burn, 1990). It means that each analyzed station has its own, unique region that
consists of a sufficient number of similar stations. Regional information is transferred to the
site of interest from all the available sites; the difference is in their relative contribution (i.e.
the pooled weighting factors) for more details, see Burn (1990). The similarity of the sites is
evaluated by a properly chosen set of site attributes. In this study, three climatological site
characteristics were selected:
- the mean annual precipitation,
- the ratio of the precipitation amounts for the warm/cold season, and
- the index of the Mediterranean effect - a special descriptor of the intra-annual variability
of monthly precipitation totals in Slovakia (see e.g., Lapin and Melo, 2005; Gal, 2005).
Thus, ROI for a given site consists of several further sites with similar climatological
conditions that may be, to some extent, related also to mechanisms generating heavy
precipitation.
The N-year quantiles of daily precipitation maxima were assessed using the GEV distribution
both in local (at-site estimation) and the regional (region of influence method, regionalization
approach of Hosking and Wallis, 1997) estimation.
The simple scaling method
The basic theoretical development of scaling has been investigated by many authors and
considerable amount of studies were devoted to extreme rainfall and its scaling properties, see
e.g., Waymire and Gupta (1981), Rodriguez-Iturbe et al. (1984), Gupta and Waymire (1990),
Koutsoyiannis and Foulfoula-Georgiu (1993), Burlando and Rosso (1996), Veneziano et al.
(1996, 2007), Hubert et al. (2002), De Michele et al. (2002), Molnr and Burlando (2005,
2008). In this study, a simple scaling hypothesis was adopted in order to derive IDF curves of
extreme rainfall, following the methodology described, e.g., in Menabde et al. (1999) and Yu
et al. (2004).

Let Id and Id denote the annual maximum rainfall intensity series for the time durations d
and d, respectively. The two random variables Id and Id are related through the scale factor
(which is the ratio between the two time durations) and have the same scaling property as (Yu
et al., 2004):
I d I d

(1)

where the equality is meant in the sense of the equality of the probability distributions of both
variables, and represents the scaling exponent. Such a behaviour is denoted as simple
scaling in the strict sense (Gupta and Waymire, 1990).
This type of scaling implies that both variables have the same probability distribution
function if the finite moments of an order q exist for both. The relationship between the qth
moments of rainfall intensity can be obtained after raising both sides of Eq. (1) to power q and
taking the ensembles average (Yu et al., 2004):

E I qd q E I dq

(2)

where q represents the scaling exponent of order q. The scaling exponent can be estimated
from the slope of the linear regression relationships between the log-transformed values of the
moments and the scale exponents for the various orders of moments. The case where the
relationship between the scaling exponent and order of moment is linear is referred to as
wide sense simple scaling (Gupta and Waymire, 1990).
The scaling behavior can also be found for the parameters of a fitted cumulative
distribution function (CDF) if the CDF has a standardized form (Menabde et al., 1999; Yu et
al., 2004)
i d
Fd i F
d

(3)

where function F is independent of d. For the simple scaling process it can be shown that the
statistical properties of the CDF for the two different timescales d and d are related as
follows (Menabde et al., 1999; Yu et al., 2004):

d d

(4)

d d

(5)

and

where d (d) is the location (scale) parameter of the annual precipitation maxima of duration
d, denotes a scale factor, and is the scaling exponent. Based on these assumptions, the
intensities I dP of events with the same annual frequency P, but a different duration d, will
satisfy the scaling relation

I Pd I dP .

(6)

On the basis of Eq. (6), it is possible to estimate the scaling exponent even if the data record is
short (Menabde, 1999).
Results and discussion
The design values of the 1-day precipitation maxima in the warm season were estimated for
the three analysed stations using the above mentioned approaches (at-site estimation, ROI,

HW). The N-year quantiles of daily precipitation maxima were assessed using the GEV
distribution function. Using the simple scaling methodology, the design values of the rainfall
of selected recurrence intervals and durations shorter than a day were estimated by
downscaling the design values of 1-day rainfall maxima. In previous studies (e.g., Bara et al.,
2010), the scaling exponents for 56 raingauges from the whole territory of Slovakia were
derived. In this study the scaling exponents of moments of rainfall intensities were used for
downscaling 1-day precipitation maxima and assessing design values of rainfall for the
durations of 5, 10, 15, 20, 30, 40, 50, 60, 90, 120 and 180 minutes in the test stations. The
scaling exponents used for each test station are listed in Table 1.
Table 1: Scaling exponents of the test stations

Station
Hurbanovo
Liptovsk Hrdok
Koice

Scaling exponent
0,7545
0,7151
0,7352

Figs. 2-4 show a comparison of design rainfall of the sub-daily duration for the periodicity P
= 0.05 and assessed by the three different estimation approaches and historical estimates of
amaj and Valovi (1973) at the three verification stations. The design values of rainfall
intensities are expressed in l.s-1.ha-1, because the historical data of amaj and Valovi (1973)
are expressed in these units.

Fig. 2 Comparison of the design rainfall values for the periodicity P = 0.05 estimated by different
approaches at the Hurbanovo station. Abbreviations in the legend: S&V method of amaj and
Valovi, H&W regionalization based on Hosking and Wallis approach, ROI regionalization based
on region-of-influence approach.

Fig. 3 The same as in Fig. 2, but for the Liptovsk Hrdok station.

Fig. 4 The same as in Fig. 2, but for the Koice station.

As can be seen on Figures 2-4, the three estimation methods at-site estimation, ROI and
HW give similiar results, The largest differences can be seen between the results of the
above mentioned three estimation approaches and the estimations of amaj and Valovi
(1973). Generally, the IDF curves downscaled from 1-day precipitation maxima assessed by
different approaches are underestimated in comparison to values of amaj and Valovi,
except for 5 and 180 minutes duration in Hurbanovo station and 5, 120 and 180 minutes
duration in Koice station.
To compare the IDF curve values assessed by amaj and Valovi (1973) and the similar
statistics assessed by downscaling the 1-day maxima estimated by at-site, ROI and HW
methods, the relative differences (RD) in percentage were calculated using the formula:

x y
(7)
100 [%],
y
where x are the design rainfall values determined by downscaling and y are the design values
assessed by amaj and Valovi (1973).
We can summarize that for a duration between 15 and 90 minutes and for periodicities P =
0.5, 0.2, 0.1, 0.05 the values of the relative differences [%] between the design rainfall
intensities estimated by simple scaling are lower compared to the IDFs by amaj and Valovi
(1973). The differences rise up to about 30% (in an absolute sense) for the test stations and
the absolute average is about 13%. In Table 2, the absolute averaged maximum, minimum and
RD

average relative differences of the various methods (at-site, ROI, HW) comparing to amaj
and Valovi (1973) are listed.
Table 2 Comparison of absolute averaged maximum, minimum and average relative differences [%]
between the design values downscaled from 1-day precipitation maxima assessed by different
approaches and the historical IDF curves derived by amaj and Valovi (1973).

Station
Hurbanovo
Liptovsk
Hrdok
Koice

Statistics
maximum
minimum
average
maximum
minimum
average
maximum
minimum
average

Relative deviation [%]

At-site
ROIcli
HW
16,07
18,17
15,38
0,11
0,11
0,36
6,38
7,15
7,84
29,65
27,29
24,76
8,07
7,40
8,34
21,81
20,00
19,05
21,90
18,70
19,34
0,32
1,33
1,58
12,12
11,80
11,33

The results presented in Table 2 indicate that the values of the design rainfall estimated by
downscaling the design values of the 1-day precipitation maxima estimated by different
approaches (at-site, ROI, HW) are underestimated on average by 13% in comparison to the
values derived by amaj and Valovi (1973). The largest differences are observed for the
shorter rainfall durations, while the results of all the methods tested are very similar up to the
durations of about 90 minutes.
The comparison of the downscaled design rainfall, based either on the regional (ROI, HW)
or the at-site estimates of the design values of the 1-day precipitation maxima, reveals that the
regionally estimated values can be lower as well as higher than the at-site estimates. This can
be caused by the different geographical locations of the test stations and by the smoothing
effect of the regional approach to the estimation of the 1-day quantiles.
Conclusions
In this study the applicability of the simple scaling method for the estimation of intensityduration-frequency (IDF) curves of extreme rainfall in Slovakia was tested. Three
meteorological stations located in different regions of Slovakia were chosen as test stations,
where the T-year quantiles of 1-day precipitation maxima were assessed by different
approaches: by local at-site estimation, by the Hosking and Wallis (1997) regionalization
approach and by the region-of-influence (ROI) method. The local IDF curves at the test
stations were estimated for different durations and recurrence intervals by downscaling the
quantiles of the 1-day precipitation maxima estimated by the before mentioned approaches
using the simple scaling method. The derived IDF curves were compared to one another and
to historical estimates of amaj and Valovi (1973). From the results it is notable, that the
values of the design rainfall estimated by the simple scaling method are similar to each other
and are underestimated with a comparison to the values derived by amaj and Valovi (1973).
The average relative differences are about 13%, in the analysed stations.

Acknowledgements
This work was supported under the VEGA Projects No. 2/0123/11 and No. 1/0103/10. This
support is gratefully acknowledged.
References
Acar, R., Celik, S., Senocak, S., 2008: Rainfall intensity-duration-frequency model using an artifical
neural network approach. Journal of Scientific and Industrial Research, 67, 198202.
Baghirathan, V.R., Shaw, E.M., 1978: Rainfall depthdurationfrequency studies for Sri Lanka.
Journal of Hydrology, 37, 223239.
Bara, M., 2010: Analza intenzt krtkodobch daov na Slovensku. [PhD-thesis]. Faculty of Civil
Engineering, Slovak University of Technology, Bratislava. 122 p. [in Slovak, with English
abstract] (Analysis of short-term rainfall intensities in Slovakia)
Bara, M., Kohnov, S., Gal, L., Szolgay, J., Hlavov, K. 2010: klovanie intenzt krtkodobch
daov na Slovensku. Key Publishing, Ostrava, 74 p. [in Slovak] (Scaling of short-term rainfall
intensities in Slovakia)
Bell, F.C., 1969: Generalized rainfalldurationfrequency relationship. Journal of the Hydraulics
Division ASCE, 95(1), 331327.
Bernard, M.M., 1932: Formulas for rainfall intensities of long durations. Transactions of the American
Society of Civil Engineers, 96, 592624.
Burlando, P., Rosso, R., 1996: Scaling and multiscaling models of depth-duration-frequency curves
for storm precipitation, Journal of Hydrology, 187, 4565.
Burn, D.H., 1990: Evaluation of regional flood frequency analysis with a region of influence
approach. Water Resources Research, 26(10), 22572265.
Cebulak, E., Fako, P., Lapin, M., astn, P., 2000: Extreme precipitation events in the Western
Carpathians. In Images of Weather and Climate. Prace Geograficzne, fasc. 108. Institute of
Geography of the Jagellonian University, Cracow, 117124.
Chen, C.L., 1983: Rainfall intensitydurationfrequency formulas. Journal of Hydraulic Engineering
ASCE, 109, 16031621.
De Michele, C., Kottegoda, N.T., Rosso, R., 2002: IDAF (intensitydurationareafrequency) curves
of extreme storm rainfall: a scaling approach. Water Science and Technology, 25(2), 8390.
Dub, O., 1950: Hydrologick podklady pre vpoet kanalizanch siet. Voda, 5-6. [in Slovak]
(Hydrological data for sewage system design)
Fako, P., Lapin, M., astn, P., Vivoda, J. (2000): Maximum daily sums of precipitation in Slovakia
in the second half of the 20th century. In Images of Weather and Climate. Prace Geograficzne, fasc.
108. Institute of Geography of the Jagellonian University, Cracow, 131138.
Gal, L., 2005: Introduction of Lapin's indices into the cluster analysis of maximum
k-day precipitation totals in Slovakia. Meteorological Journal, 8(2), 8594.
Gal, L., 2006: Metdy vpotu tatistickch charakteristk nvrhovch hodnt krtkodobch a
viacdennch hrnov zrok. [PhD-thesis]. Faculty of Mathematics, Physics and Informatics,
Comenius University in Bratislava. 228 p. [in Slovak, with English abstract] (Estimation methods
of statistical properties of short-term to several-day design precipitation in Slovakia)
Gal, L., Lapin, M., 2002: Extreme several day precipitation totals at the Hurbanovo observatory
(Slovakia) during the 20th century. Contributions to Geophysics and Geodesy, 32(3), 197213.
Gert, A., Wall, D.J., White, E.L., Dunn, C.N., 1987: Regional rainfall intensitydurationfrequency
curves for Pennsylvania. Water Resources Bulletin, 23, 479486.
Gupta, V.K., Waymire, E., 1990: Multiscaling properties of spatial and river flow distributions.
Journal of Geophysical Research, 95, D3, 19992009.
Hosking, J.R.M., Wallis, J.R., 1997: Regional frequency analysis: an approach based on L-moments.
Cambridge University Press, Cambridge; New York; Oakleigh. 224 p.
Hubert, P., Bendjoudi, H., Schertzer, D., Lovejoy, S., 2002: Multifractal taming of extreme
hydrometerological events. In: The Extremes of the Extremes (Proceedings of Reykjavik
Symposium, July 2000). IAHS Publication No. 271, 5156.

Kohnov, S., Gal, L., Szolgay, J., Hlavov, K., 2005Analza maximlnych hrnov zrok v povod
hornho Hrona. Faculty of Civil Engineering, Slovak University of Technology, Bratislava, 120 p.
[in Slovak] (Analysis of maximum precipitation totals in the Upper-Hron river basin)
Koutsoyiannis, D., Foufoula-Georgiu, E., 1993: A scaling model of a storm hyetograph. Water
Resources Research, 29(7), 23452361.
Koutsoyiannis, D., Kozonis, D., Manetas, A., 1998: A mathematical framework for studying rainfall
intensity-duration-frequency relationships. Journal of Hydrology, 206, 118135.
Lapin, M., Melo, M., 2005Priestorov interpretcia vstupov klimatickch scenrov v povod Hrona a
Vhu geotatistickmi metdami. In. Pekrov, P. and Szolgay J. (Eds) 2005: Scenre zmien
vybranch zloiek hydrosfry a biosfry v povod Hrona a Vhu v dsledku klimatickej zmeny.
VEDA, Bratislava, 4980. [in Slovak] (Spatial interpretation of climate scenarios outputs in
catchments of the Hron and Vh rivers by means of geostatistical methods. In: Scenarios of
selected components of the hydrosphere and biosphere in catchments of the Hron and Vh rivers
due to climatic change)
Menabde, M., Seed, A., Pegram, G., 1999: A simple scaling model for extreme rainfall. Water
Resources Research, 35(1), 335339.
Minh Nhat, L., Tachikawa, Y., Sayama T., Takara, K., 2007: Regional rainfall intensity-durationfrequency relationships for ungauged catchments based on scaling properties. Annuals of Disaster
Prevention Research Institute, Kyoto University 50B: 3343.
Molnr, P., Burlando, P., 2005: Preservation of rainfall properties in stochastic disaggregation by a
simple random cascade model. Atmospheric Research, 77, 137151.
Molnr, P., Burlando, P., 2008: Variability in the scaling properties of high-resolution precipitation
data in the Alpine climate of Switzerland. Water Resources Research, 44, W10404,
doi:10.1029/2007WR006142.
Niemczynowicz, J., 1982: Areal intensitydurationfrequency curves for short rainfall events in Lund.
Nordic Hydrology, 13, 928934.
Rodriguez-Iturbe, I., Gupta, V.K., Waymire, E.C., 1984: Scale considerations in the modeling of
temporal rainfall. Water Resources Research, 20(11), 16111619.
amaj, F., Valovi, S., 1973: Intenzity krtkodobch daov na Slovensku. Proceedings of works of
HMI, Nr. 5, SPN Bratislava. [in Slovak] (Intensities of short-term rainfall in Slovakia)
Sivapalan, M., Bloeschl, G., 1998: Transformation of point rainfall to areal rainfall: intensity
durationfrequency curves. Journal of Hydrology, 204(1-2), 150167.
Urcikn, P., Imrika, L., 1986. Stokovanie a istenie odpadovch vd. Tabuky na vpoet stk.
ALFA-SNTL Bratislava. [in Slovak] (Sewage and water treatment. Tables for sewage system
design)
Veneziano, D., Bras, R.L., Niemann, J.D., 1996: Nonlinearity and self-similarity of rainfall in time
and a stochastic model. Journal of Geophysical Research, 101, D21, 26,37126,392.
Veneziano, D., Lepore, Ch., Langousis, A., Furcolo, P., 2007: Marginal methods of IDF estimation in
scaling and non-scaling rainfall. Water Resources Research. 43 p.
Waymire, E.C., Gupta, V.K., 1981: The mathematical structure of rainfall representation. 1. A review
of stochastic rainfall models, 2. A review of the point processes theory, 3. Some applications of the
point process theory to rainfall processes. Water Resources Research, 17(5), 12611294.
Yu, P.Sh., Yang, T.Ch. Lin, Ch.Sh., 2004: Regional rainfall intensity formulas based on scaling
property of rainfall. Journal of Hydrology, 295(1-4), 108123.

http://www.oslovensku.sk/download/projekty/plra/mapy%20SR/mapa_vseobecnozemepisna_
SR.jpg