Acta geographica Slovenica, 57-1, 2017, 31–49
Drought AnAlysis using
the stAnDArDizeD
PreciPitAtion inDex (sPi)
AnAlizA suŠnih rAzMer
s PoMoČJo stAnDArDizirAnegA
PADAVinsKegA inDeKsA (sPi)
Urša Šebenik, Mitja Brilly, Mojca Šraj
Effects of drought on agricultural land.
Posledice suše na kmetijskih površinah.
M
O
JC
A
 Š
R
A
J
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 31
Urša Šebenik, Mitja Brilly, Mojca Šraj, Drought Analysis Using the Standardized Precipitation Index (SPI)
Drought Analysis Using the Standardized Precipitation Index (SPI)
DOI: http://dx.doi.org/10.3986/AGS.729
UDC: 556.167(497.4)
551.577.38(497.4)
COBISS: 1.01
ABSTRACT: Drought indices are commonly used for detection, monitoring and evaluation of drought
events. One of the most commonly used drought indices is the Standardized Precipitation Index (SPI).
This paper presents the effect of theoretical distribution selection on SPI values, and the analysis of drought
events for five selected meteorological stations in Slovenia. We found that the SPI on the annual time scale
shows a similar pattern of occurrence of dry and wet periods at Ljubljana-Bežigrad, Novo mesto, and Trieste
meteorological stations; something similar can be said for the Celje and Maribor-Tabor stations. The analy-
sis of the correlations between the standardized data river discharge and precipitation data for the selected
river basin of the River Pesnica shows the strongest correlation between the SPI-2 and standardized dis-
charges.
KEY WORDS: geography, drought, precipitation, probability analysis, Standardized Precipitation Index
(SPI), standardized river discharge data, the River Pesnica, Slovenia
The article was submitted for publication on 12th February, 2014.
ADDRESSES:
Urša Šebenik
University of Ljubljana
Faculty of Civil and Geodetic Engineering
Jamova 2, 1000 Ljubljana, Slovenia
E-mail: ursa.sebenik@gmail.com
Mitja Brilly, Ph.D.
University of Ljubljana
Faculty of Civil and Geodetic Engineering
Jamova 2, 1000 Ljubljana, Slovenia
E-mail: mitja.brilly@fgg.uni-lj.si
Mojca Šraj, Ph.D.
University of Ljubljana
Faculty of Civil and Geodetic Engineering
Jamova 2, 1000 Ljubljana, Slovenia
E-mail: mojca.sraj@fgg.uni-lj.si
32
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 32
1 Introduction
Drought results from a combination of meteorological, physical, and human factors (Natek 1983; Sustainable
Water Use 2001; Sušnik 2006). Drought events, in comparison with other natural disasters, differ in sev-
eral aspects (Wilhite 2003; Wilhite and Buchanan-Smith 2005):
• There is no accurate, universal and objective definition of drought. Consequently, this leads to doubts
about whether or not drought conditions are present in a given period, and if it is established that they
are present, what is their intensity. This leads to indecision and lack of action from the competent authorities.
• It is difficult to determine when a drought event began and when it ended. Usually, its consequences
accumulate slowly throughout a long period of time, and can remain present in an area for several years.
• Drought impacts do not have a one-off effect and are spread over a large geographical area. These char-
acteristics of drought have hindered the development of accurate, reliable, and timely estimates of severity
and impacts and, ultimately, the formation of drought preparedness plans.
• Problems in the quantification of drought impacts and providing disaster relief. Drought must be con-
sidered a relative, rather than an absolute condition, since it reflects a deviation from the long-term average
over a long period of time.
Drought events differ in the following aspects: intensity, duration, and spatial coverage (Wilhite 2003;
Wilhite and Buchanan-Smith 2005). The intensity of a drought event refers to the degree of precipitation
deficit and/or the severity of impacts. The spatial extent and impact of a drought event depend mostly on
the time of the onset of precipitation deficit, its intensity, and duration. The impacts and consequences of
drought can be direct and indirect. For example, loss of crops due to drought is a direct impact. The con-
sequences of this impact (i.e. loss of crops) include loss of income, damage claims from farmers; these are
indirect impacts, i.e. secondary or tertiary impacts. The impacts of drought can be economic (energy indus-
try, tourism industry, fishery production, water supplies), environmental (loss of biodiversity, degradation
of environment, erosion of soils, water quality and quantity effects) and social (food shortages, increased
groundwater depletion, loss of natural and cultural heritage, decreased quality of life; Wilhite 2003).
In order to implement adequate and timely measures, it is necessary to know the characteristics of drought
and how it affects the different levels of society and its functioning. Today, drought indices are indispensable
tools to detect, monitor and evaluate drought events (Niemeyer 2008). One of the most commonly used
indices is the Standardized Precipitation Index (SPI) (Guttman 1999), distinguished by simplicity and tem-
poral flexibility, due to which the index can be used over different time scales.
The purpose of this paper is to identify drought conditions, i.e. to analyse and compare drought peri-
ods using the SPI for the five selected sites, and try to describe hydrological drought events in the selected
river basin using standardized monthly river discharge data and the SPI.
2 Methods
2.1 Data
The only input data for calculating the SPI are monthly precipitation data. We selected four meteorolog-
ical stations in Slovenia (Ljubljana-Bežigrad, Maribor-Tabor, Celje, and Novo mesto) and one station in
Italy (Trieste), which are evenly spaced and for which long-term data series are available (ARSO 2011a;
UL FGG 2012) (Table 1).
Table 1: Features of the selected meteorological stations (ARSO 2009).
Meteorological station Elevation (AMSL) Latitude Longitude Considered period
Ljubljana-Bežigrad 299 46° 04' 14° 31' 1853–2010
Maribor-Tabor 275 46° 32' 14° 39' 1876–2010
Celje 240 46° 15' 15° 15' 1853–2010
Novo mesto 220 45° 48' 15° 11' 1951–2010
Trieste 32 45° 38' 13° 45' 1851–2004
Acta geographica Slovenica, 57-1, 2017
33
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 33
Urša Šebenik, Mitja Brilly, Mojca Šraj, Drought Analysis Using the Standardized Precipitation Index (SPI)
SPI values were calculated for six different time scales: one-month (SPI-1), two-month (SPI-2),
three-month (SPI-3), six-month (SPI-6), nine-month (SPI-9), and twelve-month, i.e. annual, (SPI-12) time
scales for the entire observation period of the selected meteorological stations as well as for the cross-sec-
tional period (1951–2004).
We chose the River Pesnica with a rain-snow regime for the comparison between the SPI and river
discharges. The Maribor-Tabor station was used for SPI calculation. The comparison was made using mean
monthly river discharge data from the Gočova gauging station for the longest available period (1970–2009)
(ARSO 2011b).
2.2 Standardized Precipitation Index (SPI)
SPI was developed by McKee et al. (1993) as a relatively simple index to be used for determining precipitation
deficit or excess. Through SPI, we can also determine the frequency of extremely dry or wet events for
a certain time scale for any location where precipitation data series are available (Gregorič and Ceglar 2007).
The standardized nature of the index allows us to obtain comparable data on drought frequency for any
location (Guttman 1999).
The first step in calculating the SPI index is to determine the probability density function for select-
ed precipitation series. The distribution most commonly used in calculating the SPI is the gamma distribution
(McKee et al. 1993; Hayes et al. 1999; Guttman 1999; Hayes 2000; Lloyd-Hughes and Saunders 2002; Ceglar
and Kajfež-Bogataj 2008). Guttman 1999, Vicente-Serrano and Lopez-Moreno (2005) as well as Blain
(2011) used Pearson III distribution in their analysis. Guttman (1999) compared the SPI values calculat-
ed with different distributions and found that the gamma and Pearson III distributions fitted data the
best. The distribution function of each monthly amount of precipitation for the given time scale is then
computed. Distribution function is then normalized into a standard normal random variable Z, which
represents the value of SPI index (Lloyd-Hughes and Saunders 2002); this quantifies the drought inten-
sity (Table 2).
34
J A  D  R A  N  S  K O
M  O  R  J  E
Reka
Mirna
P
ivka
Idrijca
S
ot
la
S
a
v
in
ja
Dra
vinja
Dragonja
Vipava
Me až
P
e sn
ica
Ščavnica
Ledava
Sora
Lju
blj
an
ica
Savinja
Kolpa
Kr
ka
So ač
Sava
Sava
Drava
Mura
Author of content/avtor vsebine:
š Š
Author of map/avtor zemljevida:
Source/vir:
in
Ur a ebenik
Urša Šebenik
Agencija Republike
Slovenije za okolje Geografski
in titut  AM ZRC SAZUš
0 10 20 30 40 50 km
S r ace water gauging station/
vodomerna postaja
u f
Meteorogical station/
meteorolo ka postajaš
Celje
Gočova,
reka Pesnica
Murska
Sobota
Novo
mesto
Ljubljana ž- Be igrad
Trst
Maribor -
tabor
Figure 1: Locations of the selected meteorological stations and the Gočova gauging station located on the River Pesnica.
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 34
Table 2: Drought classification by SPI value and corresponding event probabilities (Lloyd-Hughes 2002, 67).
SPI value Category Probability (%)
2.00 or more Extremely wet 2.3
1.50 to 1.99 Severely wet 4.4
1.00 to 1.49 Moderately wet 9.2
0.00 to 0.99 Mildly wet 34.1
0.00 to –0.99 Mildly drought 34.1
–1.00 to –1.49 Moderate drought 9.2
–1.50 to –1.99 Severe drought 4.4
–2 or less Extreme drought 2.3
McKee et al. (1993) established the criteria for determining the beginning and the end of a drought
event. A drought event begins when the SPI is continuously negative and reaches the value of –1 or less.
The event ends when the SPI value becomes positive.
2.3 Standardized river discharge data
Water resources, such as watercourses, groundwater, snow cover, etc., are highly dependent on the amount
of precipitation. The response of individual components of the hydrological cycle to the time period for
which the SPI is calculated varies. In order to determine the relationship between precipitation and river
discharges, we have to use a normal distribution to standardize mean monthly discharge data for each gaug-
ing station (Vicente-Serrano and Lopez-Moreno 2005; Gregorič and Ceglar 2007).
3 Results and analysis
3.1 Effects of probability distribution selection on SPI values
We calculated the index values for the Ljubljana-Bežigrad meteorological station using Gumbel distrib-
ution (G) and Pearson III distribution (P3) in addition to the two-parameter gamma distribution (G2).
Results are compared using Pearson's correlation coefficient (Table 3).
Table3: Correlation coefficients between selected distributions, for SPI-1 to SPI-12 (Šebenik 2012).
SPI–1 SPI–1 SPI–2 SPI–2 SPI–3 SPI–3 SPI-6 SPI-6 SPI-9 SPI-9 SPI-12 SPI-12
G P3 G P3 G P3 G P3 G P3 G P3
SPI-1 G2 0.992 0.987
SPI-2 G2 0.997 0.997
SPI-3 G2 0.994 0.988
SPI-6 G2 0.986 0.9961
SPI-9 G2 0.988 0.876
SPI-12 G2 0.993 0.539
Unlike Pearson III distribution, Gumbel distribution closely correlates with the gamma distribution
on all time scales. All correlation coefficients reached at least 0.98. Pearson III distribution has higher vari-
ability. It correlates better on longer time scales than on shorter ones (Table 3). All SPI calculations below
referred to the gamma probability distribution.
3.2 SPI values for individual meteorological stations for the entire measurement period
Annual SPI values for the Ljubljana-Bežigrad meteorological station show (Figure 2) three severe drought
events before 1900, i.e. in 1858, 1865, and 1877. Between 1900 and 1950, the SPI-12 shows four extreme
Acta geographica Slovenica, 57-1, 2017
35
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 35
Urša Šebenik, Mitja Brilly, Mojca Šraj, Drought Analysis Using the Standardized Precipitation Index (SPI)
36
drought events. The first extreme drought event was detected between 1920 and 1922, as confirmed also
by archival drought records in Slovenia (Trontelj 1997). Drought events were followed by wet periods, but
then again dry periods were detected in 1943, 1947, and 1949. Only shorter time scales show a slightly
higher frequency of extreme drought in Ljubljana in the second half of the 20th century, which occur more
frequently after 1990. The year 2003 definitely stands out after 2000 and is detected on all time scales. 2006
and 2007 are also identified as years with negative deviations, as was noted also by Sušnik and Gregorič
(2008), and by Zorn and Komac (2011).
Data analysis for the Maribor-Tabor meteorological station (Table 1) shows that extreme SPI values
appear only on shorter time scales (Šebenik 2012). Before 1900, the SPI-12 scale shows two severe drought
events with the lowest value (–1.52) in 1877. In the first half of the 20th century, annual index values indi-
cate three moderate drought events with the minimum value of the SPI-12 (–1.64) in December 1921. The
total annual precipitation for the same year was only 725 mm, which is lower than the long-term average
(i.e. 1032 mm) (Trontelj 1997). After 1950, drought events occurred more frequently and reached the high-
est frequency of occurrence in the last decade of observation (2000–2010). Annual index values for these
years do not significantly exceed the limits specified for moderate drought, with the exception of
December 1971 (–1.75) and December 2003 (–1.68). Index values for the year 2003 differ significantly less
on shorter time scales. Since only short-period precipitation totals are taken into account in the calculation
of SPI values at shorter time scales, such index values do not reflect past long-term drought conditions,
which began already in 2000 and continued in 2001 and 2002, as confirmed also by Kobold (2003).
The annual time scale for the Celje meteorological station (Table 1) shows a long period of negative
deviation between 1854 and 1859. A longer period of negative deviation repeated between 1861 and 1864
and also from 1865 to 1866, and from 1883 to 1885. Longer drought periods with constant negative index
values occurred again during 1920–1922, in 1924, and 1925. Shorter negative deviations were followed
by wet periods, which reached extreme index values in 1937 and 1938. Wet periods were again followed
by two long drought periods lasting from 1941 to 1944, and from 1945 to 1948. The year 1946 stands out,
when virtually all months of the year had negative index values. The exception after 2000 was the year 2003,
when the SPI reached values indicative of severe drought.
The lowest values for the Novo mesto meteorological station (Table 1) on the annual time scale were
within the limits of moderate or severe drought (Table 2). However, short periods of negative deviation
occurred quite frequently (Šebenik 2012). Longer periods of precipitation deficit were more common in
the last three decades. In 2007, negative deviation persisted throughout the year.
The calculations of the SPI-12 for the Trieste meteorological stations (Table 1) show that drought
events were not particularly severe, since the lowest SPI value in the whole observation period is –1.01.
Several long periods of negative deficits appear on the 12-month time scale before 1900, alternating with
distinctively wet periods with extreme index values. It continued in a similar way in the 20th century,
reaching the lowest index values in 1946. A similar pattern can also be observed in the second half of
the 20th century. 2003 stands out from the last analyzed years, as it has extreme index values on all shorter
time scales.
We can see that the year 2003 definitely stood out among all observed meteorological stations in the last
observed decade. The 2003 extreme drought event in Europe caused EUR 8.7 billion in losses (Commission
of the European Communities 2007). The substantial damage caused by drought relative to the total dam-
age caused by natural disasters in 2003 in Slovenia, was as high as 83.3% (Zorn and Komac 2011).
3.3 SPI comparison between selected meteorological stations for the common period
of measurement 1951–2004
SPI values for all selected meteorological stations and all time scales were also compared for the com-
mon measurement period. On longer time scales values for all stations have a similar distribution of major
dry and wet periods (Figure 3). A major difference between stations occurred in 2002 when Trieste stood
out with a distinctively wet year, while data for the other four stations already indicated extreme drought
conditions, which later affected all the selected sites in 2003. If we examine the data for this period more
Figure 2: SPI-2, SPI-6, SPI-9 and SPI-12 for the Ljubljana-Bežigrad meteorological station for the 1853–2010 period (Šebenik 2012). p
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 36
Acta geographica Slovenica, 57-1, 2017
37
S
P
I
-
2
–
4
–
3
–
2
–
101234
1853
1858
1864
1870
1876
1882
1888
1893
1899
1905
1911
1917
1923
1928
1934
1940
1946
1952
1958
1963
1969
1975
1981
1987
1993
1998
2004
2010
Y
e
a
r/
le
to
S
P
I
-
9
–
4
–
3
–
2
–
101234
S
P
I
-
6
–
4
–
3
–
2
–
101234
S
P
I
-
1
2
–
3
–
2
–
101234
SPI SPI
1853
1858
1864
1870
1876
1882
1888
1893
1899
1905
1911
1917
1923
1928
1934
1940
1946
1952
1958
1963
1969
1975
1981
1987
1993
1998
2004
2010
Y
e
a
r/
le
to
1853
1858
1864
1870
1876
1882
1888
1893
1899
1905
1911
1917
1923
1928
1934
1940
1946
1952
1958
1963
1969
1975
1981
1987
1993
1998
2004
2010
Y
e
a
r/
le
to
1853
1858
1864
1870
1876
1882
1888
1893
1899
1905
1911
1917
1923
1928
1934
1940
1946
1952
1958
1963
1969
1975
1981
1987
1993
1998
2004
2010
Y
e
a
r/
le
to
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 37
Urša Šebenik, Mitja Brilly, Mojca Šraj, Drought Analysis Using the Standardized Precipitation Index (SPI)
LJUBLJANA-BEŽIGRAD meteorological station/meteorološka postaja
–3
–2
–1
0
1
2
3
4
1
9
5
2
1
9
5
4
1
9
5
6
1
9
5
8
1
9
5
9
1
9
6
1
1
9
6
3
1
9
6
5
1
9
6
6
1
9
6
8
1
9
7
0
1
9
7
2
1
9
7
3
1
9
7
5
1
9
7
7
1
9
7
9
1
9
8
0
1
9
8
2
1
9
8
4
1
9
8
6
1
9
8
7
1
9
8
9
1
9
9
1
1
9
9
3
1
9
9
4
1
9
9
6
1
9
9
8
2
0
0
0
2
0
0
1
2
0
0
3
Year/leto
S
P
I
MARIBOR-TABOR meteorological station/meteorološka postaja
–3
–2
–1
0
1
2
3
4
Year/leto
S
P
I
CELJE meteorological station/meteorološka postaja
–3
–2
–1
0
1
2
3
4
Year/leto
S
P
I
1
9
5
2
1
9
5
4
1
9
5
6
1
9
5
8
1
9
5
9
1
9
6
1
1
9
6
3
1
9
6
5
1
9
6
6
1
9
6
8
1
9
7
0
1
9
7
2
1
9
7
3
1
9
7
5
1
9
7
7
1
9
7
9
1
9
8
0
1
9
8
2
1
9
8
4
1
9
8
6
1
9
8
7
1
9
8
9
1
9
9
1
1
9
9
3
1
9
9
4
1
9
9
6
1
9
9
8
2
0
0
0
2
0
0
1
2
0
0
3
1
9
5
2
1
9
5
4
1
9
5
6
1
9
5
8
1
9
5
9
1
9
6
1
1
9
6
3
1
9
6
5
1
9
6
6
1
9
6
8
1
9
7
0
1
9
7
2
1
9
7
3
1
9
7
5
1
9
7
7
1
9
7
9
1
9
8
0
1
9
8
2
1
9
8
4
1
9
8
6
1
9
8
7
1
9
8
9
1
9
9
1
1
9
9
3
1
9
9
4
1
9
9
6
1
9
9
8
2
0
0
0
2
0
0
1
2
0
0
3
1
9
5
1
1
9
5
1
1
9
5
1
38
p
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 38
closely, we can see that, in most cases, the Ljubljana-Bežigrad, Novo mesto and Trieste meteorological
stations share a similar pattern of occurrence of dry and wet periods, and something similar can be said
for the Celje and the Maribor-Tabor stations (Šebenik 2012). Differences within each group lie in drought
severity (SPI values occasionally differ by more than one classification scale) and in the duration and
the onset of a drought event, which differ by one or two months between the stations, in each group.
Drought never affects the whole Slovenian territory evenly, which confirms the claim that drought is
a regional phenomenon (Kobold 2003). During the last period the frequency and intensity of extreme events
increased.
The results of the SPI-12 calculations for the entire period of observation for each station and select-
ed common period show that the values of the correlation coefficient for all stations and all periods calculated
are higher than 0.95, which means that, as regards the selected meteorological stations, the length of data
series does not have a significant effect on SPI values (Šebenik 2012).
39
NOVO MESTO meteorological station/meteorološka postaja
–3
–2
–1
0
1
2
3
4
Year/leto
S
P
I
TRST meteorological station/meteorološka postaja
–3
–2
–1
0
1
2
3
4
Year/leto
S
P
I
1
9
5
2
1
9
5
4
1
9
5
6
1
9
5
8
1
9
5
9
1
9
6
1
1
9
6
3
1
9
6
5
1
9
6
6
1
9
6
8
1
9
7
0
1
9
7
2
1
9
7
3
1
9
7
5
1
9
7
7
1
9
7
9
1
9
8
0
1
9
8
2
1
9
8
4
1
9
8
6
1
9
8
7
1
9
8
9
1
9
9
1
1
9
9
3
1
9
9
4
1
9
9
6
1
9
9
8
2
0
0
0
2
0
0
1
2
0
0
3
1
9
5
2
1
9
5
4
1
9
5
6
1
9
5
8
1
9
5
9
1
9
6
1
1
9
6
3
1
9
6
5
1
9
6
6
1
9
6
8
1
9
7
0
1
9
7
2
1
9
7
3
1
9
7
5
1
9
7
7
1
9
7
9
1
9
8
0
1
9
8
2
1
9
8
4
1
9
8
6
1
9
8
7
1
9
8
9
1
9
9
1
1
9
9
3
1
9
9
4
1
9
9
6
1
9
9
8
2
0
0
0
2
0
0
1
2
0
0
3
1
9
5
1
1
9
5
1
Figure 3: SPI-12 for the selected meteorological stations and the common measurement period 1951–2004.
Acta geographica Slovenica, 57-1, 2017
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 39
Urša Šebenik, Mitja Brilly, Mojca Šraj, Drought Analysis Using the Standardized Precipitation Index (SPI)
40
3.4 The relationship between the SPI and the standardized mean monthly discharge
of the Pesnica river basin
The analysis of the results for the 1970–2009 period showed that the correlation between the standard-
ized series of river discharge data and the SPI for the River Pesnica is positive for all time scales, but the
value of Pearson correlation coefficient varies between different time scales. It is also evident that higher
correlation coefficients were obtained on shorter time scales in late spring, summer (July and August), and
autumn (September, November) (Figure 4).
The September SPI-2 and the September standardized discharge had the strongest correlation (= 0.754)
(Figure 5). The results show that mean monthly discharges of the River Pesnica depend highly on pre-
cipitation amounts of the current and the past month, which means that the river's watercourse or basin
responds quickly to rainfall. The primary water surplus of the River Pesnica occurs in April (Kolbezen 1998).
It means that the River Pesnica responds quickly to increased amounts of water resulting from snowmelt
or abundant precipitation. The secondary water surplus occurs in November (Kolbezen 1998), which also
has high correlation with the index values at shorter time scales. Summer months have higher correlation
0.0 0.2 0.4 0.6 0.8
SPI-1
SPI-2
SPI-3
SPI-6
SPI-9
SPI-12
Correlation coefficient/korelacijski koeficient
January
januar
July/
julij
February/
februar
August/
avgust
March/
marec
September/
september
April/
april
October/
oktober
May/
maj
November/
november
June/
junij
December/
december
Figure 4: Representation of monthly correlations between standardized discharge data and the SPI.
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 40
Acta geographica Slovenica, 57-1, 2017
41
–5
–4
–3
–2
–1
0
1
2
3
4
5
1970 1975 1980 1985 1990 1995 2000 2005
Year/leto
Standardized river discharge/standardizirani pretok SPI-2
S
P
I
 a
n
d
 s
ta
n
d
a
r
d
iz
e
d
 d
is
c
h
a
r
g
e
S
P
I
 i
n
 s
ta
n
d
a
r
d
iz
ir
a
n
 p
r
e
to
k
/
Figure 5: Standardized discharge and the SPI at two-month time scale in September.
coefficient values on longer time scales, where precipitation totals include early spring and winter
months, which have higher precipitation levels. Extreme standardized discharge values coincide with extreme
SPI-2 values, but the former are slightly higher than the latter.
4 Discussion
Analysis showed that the SPI at shorter time scales has high variability and shows more short-term drought
events. Drought events occur less frequently, but last longer on longer time scales. Longer SPI time scales
do not necessarily detect all the negative deviations that are evident on shorter time scales. It is also evi-
dent that SPI values at shorter time scales show slight increases in precipitation during dry periods, which
do not necessarily reflect an improvement in drought situation on a longer time scale. When analyzing
past periods, we have to keep in mind that several consecutive months of negative index values do not
necessarily indicate drought. Negative index values actually identify the months with less precipitation com-
pared against the long-term comparative period.
Precipitation deficit is one of the main causes of drought onset, but not the only one (Vicente-
-Serrano et al. 2010), since evapotranspiration, temperature, wind speed, water retention capacity of soil
and human impacts also significantly influence the development of drought. Precipitation deficit in win-
ter months is problematic with regard to groundwater recharge and recharge of other water resources, which
are among the important factors affecting the status of drinking water supply in Slovenia. The SPI is based
mainly on precipitation data, therefore, in order to analyze individual types of drought in more detail, we
have to use other instruments: drought indices which include other variables in addition to precipitation,
water balance models, low-flow analysis, etc. In particular, the SPI provides the first important informa-
tion regarding drought conditions (Hayes et al. 1999).
In order to identify drought events, we also have to analyze long-time scales of SPI, which are also indi-
cators of hydrological drought conditions of surface and groundwater sources (McKee et al.  1993;
Hayes et al. 1999). The index value calculated at a specific time scale must be representative of the drought
status in a hydrological system to be operative for water resources management purposes. The strongest
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 41
Urša Šebenik, Mitja Brilly, Mojca Šraj, Drought Analysis Using the Standardized Precipitation Index (SPI)
correlations between standardized discharge data and the SPI were detected on the two-month time scale
for the River Pesnica. This case shows that it is necessary to identify the most suitable scale for calcula-
tion, since hydrological, meteorological and terrain characteristics differ significantly between river basins.
For the same reason, the results could not be generalized to the whole territory of Slovenia. To date, there
have not been many studies conducted in this area and not many definite relationships were found between
different drought monitoring periods and water resources.
We standardized discharge data using a normal distribution to achieve greater comparability and more
accurate evaluation of correlations between the SPI and standardized discharge data, and thus facilitate
the comparison between meteorological and hydrological variables. It would be possible to obtain even
more accurate results if discharge data had been standardized using any other distribution function.
5 Conclusion
Droughts and associated water shortages are a global challenge, and Slovenia is no exception. Nevertheless,
Slovenia is relatively abundant in water resources. However, despite the high total amount of rainfall, the
timing of precipitation is often unfavourable for various activities (high-quality crop production, drink-
ing water supply, hydroelectric power generation) (Gregorič and Sušnik 2008). In recent years, drought
losses have reached extremely high levels in Slovenia also (Zorn and Komac 2011). The results show that
the largest share (48.6%) of total losses in the period 2000–2005 was caused by drought (2007 Audit Report:
Performance Audit of Drought Preventing and Drought Recovery in Agriculture by the Republic of Slovenia).
The data therefore suggest that Slovenia, too, should seriously tackle drought-related problems.
6 References
ARSO, 2009. Meteorološki letopis 2009. Internet: http://www.arso.si/vreme/podnebje/meteorolo%c5%a1ki%
20letopis/2009mreza.pdf (26. 1. 2012).
ARSO, 2011a. Podatki o mesečnih padavinah. Ljubljana.
ARSO,  2011b. Arhiv površinskih voda. Internet: http://vode.arso.gov.si/hidarhiv/pov_arhiv_tab.php
(13. 12. 2011).
Blain, G.C. 2011: Standardized Precipitation Index based on Pearson Type III Distribution. Revista Brasiliera
de meteorologia 26. DOI: http://dx.doi.org/10.1590/S0102-77862011000200001
Ceglar, A., Kajfež-Bogataj, L. 2008: Obravnava meteorološke suše z različnimi indikatorji. Acta agricul-
turae Slovenica 91.
Commission of the European Communities, 2007. Communication from the Commission of the European
Communities to the European Parliament and the Council: addressing the challenge of water scarcity
and droughts in the European Union. Brussels.
Gregorič, G., Ceglar, A. 2007: Monitoring suše – regionalni aspekt. 18. Mišičev vodarski dan: zbornik refer-
atov. Maribor.
Gregorič, G., Sušnik, A. 2008: Center za upravljanje suše v jugovzhodni Evropi. 60 let slovenske meteo-
rološke in hidrološke službe. Naše okolje 15.
Guttman, N. B. 1999: Accepting the standardized precipitation index: a calculation algorithm. Journal of
the American water resources association 35-2. DOI: http://dx.doi.org/10.1111/j.1752-1688.1999.tb03592.x
Hayes, M. J., Svoboda, M. D., Wilhite, D. A., Vanyarkho, O. V. 1999: Monitoring the 1996 drought using the
standardized precipitation index. Bulletin of the American meteorological society 80. DOI: http://dx.doi.org/
10.1175/1520-0477(1999)080<0429:MTDUTS>2.0.CO;2
Hayes, M. J. 2000: Revisiting the SPI: clarifying the process. Drought network news 12. Lincoln.
Kobold, M. 2003: Hidrološka suša slovenskih vodotokov v obdobju 2000–2002. Ujma 17–18.
Kolbezen, M. 1998: Rečni režimi. Površinski vodotoki in vodna bilanca Slovenije. 50 let organizirane hidrom-
eteorološke službe na Slovenskem 1947–1997. Ljubljana.
Lloyd-Hughes, B., Saunders, M. A. 2002: A drought climatology for Europe. International journal of cli-
matology 22. DOI: http://dx.doi.org/10.1002/joc.846
42
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 42
Lloyd-Hughes, B. 2002: The long-range predictability of European drought. Ph. D. thesis, University college
London, Department of space and climate physics. London.
McKee, T. B., Nolan, D. J., Kleist, J. 1993: The relationship of drought frequency and duration to time scales.
Preprints of 8th Conference on Applied Climatology. Anaheim.
Natek, K. 1983: Ogroženost Slovenije zaradi suše. Naravne nesreče v Sloveniji kot naša ogroženost. Ljubljana.
Niemeyer, S. 2008: New drought indices. Drought management: scientific and technological innovations
(Options Méditerranéennes: Série A. Séminaires Méditerranéens 80). Zaragoza.
Revizijsko poročilo o smotrnosti ravnanja Republike Slovenije pri preprečevanju in odpravi posledic suše
v kmetijstvu, 2007. Internet: http://www.rs-rs.si/rsrs/rsrs.nsf/I/K99638A13FF506FB3C1257322003D2E6B/
$file/Susa_RSP00-06.pdf (9. 4. 2011).
Sustainable water use 3: Extreme hydrological events: floods and droughts, 2001. Internet: http://www.eea.europa.eu/
publications/Environmental_Issues_No_21 (19. 10. 2012).
Sušnik, A. 2006: Vodni primanjkljaj v Sloveniji in možni vplivi podnebnih sprememb. Magistrsko delo,
Biotehniška fakulteta Univerze v Ljubljani. Ljubljana.
Sušnik, A., Gregorič, G. 2008: Trendi ranljivosti na kmetijsko sušo. 19. Mišičev vodarski dan: zbornik refer-
atov. Maribor.
Šebenik, U. 2012: Analiza suše s pomočjo standardiziranega padavinskega indeksa. Diplomsko delo, Fakulteta
za gradbeništvo in geodezijo Univerze v Ljubljani. Ljubljana.
Trontelj, M. 1997: Kronika izrednih vremenskih dogodkov XX. stoletja. Pomembni vremenski dogodki
v zgodovini: vreme ob pomembnih dogodkih. Ljubljana.
UL FGG, 2012. Arhiv padavinskih podatkov. Ljubljana.
Vicente-Serrano, S. M., Lopez-Moreno, J. I. 2005: Hydrological response to different time scales of clima-
tological drought: an evaluation of the standardized precipitation index in a mountainous Mediterranean
basin. Hydrology and earth system sciences 9. DOI: http://dx.doi.org/10.5194/hess-9-523-2005
Vicente-Serrano, S. M., Begueria, S., Lopez-Moreno, J. I. 2010: A multiscalar drought index sensitive to
global warming: the standardized precipitation evapotranspiration index. Journal of climate 23. DOI:
http://dx.doi.org/10.1175/2009JCLI2909.1
Zorn, M., Komac, B. 2011: Damage caused by natural disasters in Slovenia and globally between 1995
and 2010. Acta geographica Slovenica 51-1. DOI: http://dx.doi.org/10.3986/AGS51101
Wilhite, D. A. 2003: Drought. Encyclopedia of atmospheric sciences. Amsterdam.
Wilhite, D. A., Buchanan-Smith, M. 2005: Drought as hazard: understanding the natural and social context.
Drought and water crises, science, technology and management issues. Boca Raton.
Acta geographica Slovenica, 57-1, 2017
43
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 43
Urša Šebenik, Mitja Brilly, Mojca Šraj, Ana li za sušnih raz mer s po moč jo stan dar di zi ra ne ga pada vin ske ga indek sa (SPI)
Ana li za sušnih raz mer s po moč jo stan dar di zi ra ne ga pada vin ske ga
indek sa (SPI)
DOI: http://dx.doi.org/10.3986/AGS.729
UDK: 556.167(497.4)
551.577.38(497.4)
COBISS: 1.01
IZVLEČEK: Za zaz na va nje, sprem lja nje in oce no sušnih raz mer se danes pogo sto upo rab lja jo sušni indeksi.
Eden izmed naj po go ste je upo rab lje nih je stan dar di zi ra ni pada vin ski indeks (SPI). V pris pev ku je pred -
stav ljen vpliv izbi re teo re tič ne poraz de li tve na vred no sti SPI ter ana li za sušnih obdo bij za pet izbra nih
meteo ro loš kih postaj v Slo ve ni ji. Ugo to vi li smo, da SPI na let ni rav ni kaže podo ben vzo rec pojav lja nja sušnih
in mokrih obdo bij za meteo ro loš ke posta je Ljub lja na-Be ži grad, Novo mesto in Trst. Podob no lah ko reče -
mo tudi za meteo ro loš ki posta ji Celje in Mari bor-Ta bor. Ana li za pove za no sti stan dar di zi ra nih pre to kov
in pada vin za izbra no poreč je reke Pesni ce kaže naj viš jo kore la ci jo med stan dar di zi ra nim pre to kom in SPI-2.
KLJUČNE BESEDE: geo gra fi ja, suša, pada vi ne, ver jet nost na ana li za, stan dar di zi ra ni pada vin ski indeks
(SPI), stan dar di zi ra ni pre tok, Pesni ca, Slo ve ni ja
Ured niš tvo je pre je lo pris pe vek 12. fe bruar ja 2014.
NASLOVI:
Ur ša Šebe nik
Uni ver za v Ljub lja ni
Fa kul te ta za grad be niš tvo in geo de zi jo
Ja mo va 2, 1000 Ljub lja na, Slo ve ni ja
E-po šta: ursa.se be nik@gmail.com
dr. Mit ja Brilly
Uni ver za v Ljub lja ni
Fa kul te ta za grad be niš tvo in geo de zi jo
Ja mo va 2, 1000 Ljub lja na, Slo ve ni ja
E-po šta: mit ja.brilly@fgg.uni-lj.si
dr. Moj ca Šraj
Uni ver za v Ljub lja ni
Fa kul te ta za grad be niš tvo in geo de zi jo
Ja mo va 2, 1000 Ljub lja na, Slo ve ni ja
E-po šta: moj ca.sraj@fgg.uni-lj.si
44
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 44
1 Uvod
Su ša je rezul tat zdru že va nja vre men skih, narav nih in člo veš kih dejav ni kov (Na tek 1983; Sustai nab le Water
Use 2001; Sušnik 2006). Suša se od dru gih narav nih nesreč raz li ku je v več vidi kih (Wil hi te 2003; Wil hi te
in Buc ha nan-Smith 2005):
• Ne poz na mo uni ver zal ne in objek tiv ne opre de li tve suše. Posle dič no nasta ne dvom, ali suša v da nem obdobju
sploh obsta ja in kak šna je nje na inten ziv nost, kar po nava di vodi v neod loč nost in neu kre pa nje.
• Zače tek in konec suše sta tež ko določ lji va dogod ka. Posle di ce se obi čaj no kopi či jo sko zi dalj še časov no
obdob je in lah ko obsta ja jo več let.
• Vpli vi suše nima jo enkrat ne ga učin ka in so raz šir je ni prek več je ga območ ja. To ovi ra raz voj zanes lji ve
in pra vo ča sne oce ne inten ziv no sti in vpli vov suše ter tudi pri pra vo načr ta pri prav lje no sti na sušo.
• Teža ve so pri koli čin ski opre de li tvi vpli vov suše in zago tav lja nju pomo či. Sušo upo šte va mo v re la tiv nem
in ne abso lut nem smi slu, saj je izra že na na pod la gi odklo na od dol go let ne ga pov preč ja v dalj šem časovnem
obdob ju.
Po sa mez ne suše se med seboj raz li ku je jo po: inten ziv no sti, tra ja nju in pro stor ski raz sež no sti (Wil hi -
te 2003; Wil hi te in Buc ha nan-Smith 2005). Inten ziv nost sušne ga dogod ka se nana ša na stop njo pri manj klja ja
pada vin in/ali resnost učin kov. Kak šen obseg in vpliv ima suša, je odvi sno pred vsem od časa nasto pa primanj -
klja ja pada vin, nje go ve inten zi te te in tra ja nja. Vpli vi in posle di ce suše so lah ko nepo sred ni in posred ni. Izgu ba
pri del ka je pri mer nepo sred ne ga vpli va, kate re ga posle di ce so: izgu ba v do hod ku, odškod nin ski zah tevki
kme tov. To so posred ni ozi ro ma sekun dar ni ali ter ciar ni vpli vi. Govo ri mo tudi o vpli vih suše na gos podars -
tvo (ener ge ti ka, turi zem, ribiš tvo, oskr ba z vodo), oko lje (zmanj ša nje biot ske pestro sti, degra da ci ja oko lja,
ero zi ja prsti, kako vost in koli či na vod nih virov) in druž bo (po manj ka nje hra ne, izčr pa va nje pod zem ne
vode, izgu ba narav ne in kul tur ne dediš či ne, zmanj ša na kva li te ta biva nja; Wil hi te 2003).
Za ustrez no in pra vo ča sno ukre pa nje je nuj no poz na va nje zna čil no sti suše ter nje nih vpli vov na različ -
ne rav ni delo va nja druž be. Nepo greš lji vo orod je za zaz na va nje, sprem lja nje in oce no sušnih raz mer so sušni
indek si (Nie me yer 2008). Eden izmed naj po go ste je upo rab lje nih je stan dar di zi ra ni pada vin ski indeks (SPI)
(Gutt man 1999), ki ga odli ku je pred vsem pre pro stost in časov na pri la go dlji vost. To omo go ča nje go vo upo -
ra bo na raz lič nih časov nih les tvi cah.
Na men član ka je opre de li ti sušne raz me re ozi ro ma nare di ti ana li zo in pri mer ja vo sušnih obdo bij s po -
moč jo SPI za pet izbra nih loka cij in posku ša ti opre de li ti tudi hidro loš ko sušo na izbra nem poreč ju s po moč jo
stan dar di zi ra ne ga meseč ne ga pre to ka in SPI.
2 Meto de
2.1 Podat ki
Edi ni vhod ni poda tek za izra čun SPI so meseč ne pada vi ne. Za ana li zo smo izbra li šti ri meteo ro loš ke postaje
v Slo ve ni ji (Ljub lja na-Be ži grad, Mari bor-Ta bor, Celje in Novo mesto) in posta jo iz sosed nje Ita li je (Trst),
ki so pro stor sko ena ko mer no raz po re je ne in za kate re so na voljo dalj ši časov ni nizi pada vin skih podatkov
(ARSO 2011a; UL FGG 2012) (pre gled ni ca 1).
Sli ka 1: Lega izbra nih meteo ro loš kih postaj ter vodo mer ne posta je Gočo va na reki Pesni ci.
Glej angleš ki del pris pev ka.
Pre gled ni ca 1: Zna čil no sti izbra nih meteo ro loš kih postaj (ARSO 2009).
me teo ro loš ka posta ja nad mor ska viši na [m] zem lje pi sna širi na zem lje pi sna dol ži na obrav na va no obdob je
Ljub lja na-Be ži grad 299 46° 04' 14° 31' 1853–2010
Ma ri bor-Ta bor 275 46° 32' 14° 39' 1876–2010
Ce lje 240 46° 15' 15° 15' 1853–2010
Novo mesto 220 45° 48' 15° 11' 1951–2010
Trst 32 45° 38' 13° 45' 1851–2004
Acta geographica Slovenica, 57-1, 2017
45
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 45
Urša Šebenik, Mitja Brilly, Mojca Šraj, Ana li za sušnih raz mer s po moč jo stan dar di zi ra ne ga pada vin ske ga indek sa (SPI)
Izra čun SPI smo izved li na šestih časov nih les tvi cah: eno me seč ni (SPI-1), dvo me seč ni (SPI-2), tri me -
seč ni (SPI-3), šest me seč ni (SPI-6), devet me seč ni (SPI-9) in dva najst me seč ni les tvi ci (SPI-12) za celot no
opa zo va no obdob je posa mez nih meteo ro loš kih postaj ter nji ho vo pre seč no obdob je (1951–2004).
Za pri mer ja vo med SPI in pre to ki smo izbra li reko Pesni co z dež no-snež nim reč nim reži mom. Za izračun
SPI smo upo ra bi li podat ke meteo ro loš ke posta je Mari bor-Ta bor. Pri mer ja va je bila nare je na s po dat ki sred -
nje ga meseč ne ga pre to ka za vodo mer no posta jo Gočo va za naj dalj še dostop no obdob je meri tev (1970–2009)
(ARSO 2011b).
2.2 Stan dar di zi ra ni pada vin ski indeks (SPI)
SPI je raz vil McKee s so de lav ci (1993) kot raz me ro ma pre prost indeks za ugo tav lja nje pri manj klja ja oziroma pre -
sež ka pada vin. Omo go ča dolo ča nje pogo sto sti ekstrem no suhih ozi ro ma ekstrem no mokrih obdo bij na določe ni
časov ni les tvi ci za kate ro ko li loka ci jo, za kate ro obsta ja niz pada vin skih podat kov (Gre go rič in Ceglar 2007). Stan -
dar di zi ra na nara va indek sa omo go ča pri mer lji vost frek venc sušnih dogod kov na kateri ko li loka ci ji (Guttman 1999).
V pr vem kora ku izra ču na SPI dolo či mo gosto to ver jet no sti izbra ne ga vzor ca pada vin. Naj po go ste je
upo rab lja mo gama poraz de li tev (Mc Kee in osta li 1993; Hayes in osta li 1999; Gutt man 1999; Hayes 2000;
Lloyd-Hug hes in Saun ders 2002; Ceglar in Kaj fež-Bo ga taj 2008). Gutt man (1999), Vicen te-Ser ra no in Lopez-Mo -
re no (2005) ter Blain (2011) pa so upo ra bi li Pear so no vo III poraz de li tev. Gutt man (1999) je pri mer jal vred no sti
SPI več poraz de li tev, in ugo to vil, da se podat kom naj bo lje pri la ga ja ta gama in Pear so no va III poraz de li -
tev. V na sled njem kora ku za meseč no vso to pada vin in izbra no časov no les tvi co izra ču na mo poraz de li tve no
funk ci jo. To nato nor ma li zi ra mo v stan dar di zi ra no nor mal no slu čaj no spre men ljiv ko, kar pred stav lja vred -
nost indek sa SPI (Lloyd-Hug hes in Saun ders 2002), s ka te rim ovred no ti mo inten zi te to suše (pre gled ni ca 2).
Pre gled ni ca 2: Kla si fi ka ci ja suše ter pri pa da jo ča ver jet nost poja va sušne ga dogod ka pri dolo če nem SPI (Lloyd-Hug hes 2002, 67).
SPI kla si fi ka ci ja ver jet nost [%]
2,00 ali več ek strem no mokro 2,3
1,50 do 1,99 zelo mokro 4,4
1,00 do 1,49 zmer no mokro 9,2
0,00 do 0,99 nor mal no 34,1
0,00 do –0,99 nor mal no 34,1
–1,00 do –1,49 zmer na suša 9,2
–1,50 do –1,99 huda suša 4,4
–2 ali manj ek strem na suša 2,3
Mc Kee in sode lav ci (1993) so dolo či li tudi kri te rij za dolo či tev začet ka in kon ca sušne ga dogod ka. Ko
je indeks SPI dalj časa nega ti ven in dose že vred nost –1 ali manj, govo ri mo o za čet ku sušne ga dogod ka, ki
se kon ča, ko vred nost indek sa posta ne pozi tiv na.
2.3 Stan dar di zi ra ni pre tok
Vod ni viri, kot so voda v vo do to kih, pod zem na voda, snež na ode ja, so ključ no pove za ni s ko li či no pada -
vin. Odziv posa mez nih kom po nent hidro loš ke ga kro ga na časov na obdob ja izra ču na indek sa SPI je raz li čen.
Če želi mo ugo to vi ti pove za vo med pada vi na mi in pre to ki, mora mo tudi podat ke sred nje ga meseč ne ga pretoka
za posa mez no vodo mer no posta jo stan dar di zi ra ti z nor mal no poraz de li tvi jo (Vi cen te-Ser ra no in Lopez-Mo -
re no 2005; Gre go rič in Ceglar 2007).
3 Rezul ta ti in ana li za
3.1 Vpliv izbi re ver jet nost ne poraz de li tve na vred nost SPI
Za meteo ro loš ko posta jo Ljub lja na-Be ži grad smo poleg dvo pa ra me tr ske gama poraz de li tve (G2), upo ra -
bi li še Gum be lo vo (G) in Pear so no vo III (P3) poraz de li tev ter rezul ta te pri mer ja li s po moč jo Pear so no ve ga
kore la cij ske ga koe fi cien ta (pre gled ni ca 3).
46
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 46
Pre gled ni ca 3: Kore la cij ski koe fi cien ti izbra nih poraz de li tev za SPI-1 do SPI-12 (Še be nik 2012).
SPI–1 SPI–1 SPI–2 SPI–2 SPI–3 SPI–3 SPI-6 SPI-6 SPI-9 SPI-9 SPI-12 SPI-12
G P3 G P3 G P3 G P3 G P3 G P3
SPI-1 G2 0,992 0,987
SPI-2 G2 0,997 0,997
SPI-3 G2 0,994 0,988
SPI-6 G2 0,986 0,996
SPI-9 G2 0,988 0,876
SPI-12 G2 0,993 0,539
Gum be lo va poraz de li tev se v nas prot ju s Pear so no vo III poraz de li tvi jo na vseh časov nih les tvi cah dobro
uje ma z gama poraz de li tvi jo, saj kore la cij ski koe fi cien ti dose že jo vred nost vsaj 0,98. Pear so no va III poraz -
de li tev kaže več jo varia bil nost. Bolje kore li ra na kraj ših kot na dalj ših časov nih les tvi cah (pre gled ni ca 3).
V na da lje va nju so vsi izra ču ni SPI nare je ni z upo ra bo gama ver jet nost ne poraz de li tve.
3.2 SPI za posa mez ne posta je za celot no obdob je meri tev
Vred no sti SPI na let ni rav ni za meteo ro loš ko posta jo Ljub lja na-Be ži grad kaže jo (sli ka 2) pred letom 1900
tri pomemb nej ša sušna obdob ja in sicer v le tih 1858, 1865 in 1877 (Še be nik 2012). Med leto ma 1900 in 1950
SPI-12 kaže šti ri ekstrem na sušna obdob ja. Prve ga je zaz na ti med leto ma 1920 in 1922, kar potr ju je jo tudi
arhiv ski zapi si o su ši v Slo ve ni ji (Tron telj 1997). Sle di jo kraj ša mokra obdob ja, tem pa zopet sušnej ša v le -
tih 1943, 1947 ter 1949. V dru gi polo vi ci dvaj se te ga sto let ja so bila v Ljub lja ni ekstrem na sušna obdob ja
le na kraj ših časov nih les tvi cah, ki so pogo stej ša po letu 1990. Po letu 2000 po sušnih raz me rah izsto pa
leto 2003, ki ga zaz na jo vse časov ne les tvi ce. Tudi v le tih 2006 in 2007 SPI kaže nega tiv no odsto pa nje, kar
ugo tav lja jo tudi Sušnik in Gre go rič (2008) ter Zorn in Komac (2011).
Sli ka 2: SPI-2, SPI-6, SPI-9 in SPI-12 za meteo ro loš ko posta jo Ljub lja na-Be ži grad za obdob je 1853–2010 (Še be nik 2012).
Glej angleš ki del pris pev ka.
Ana li za podat kov za meteo ro loš ko posta jo Mari bor-Ta bor (pre gled ni ca 1) je poka za la, da se ekstrem -
ne vred no sti SPI pojav lja jo le na kraj ših časov nih les tvi cah (Še be nik 2012). Pred letom 1900 nam SPI-12
kaže dve sušni obdob ji z mi ni mal no vred nost jo (–1,52) leta 1877. V prvi polo vi ci dvaj se te ga sto let ja letni
indeks kaže tri zmer na sušna obdob ja z mi ni mal no vred nost jo SPI-12 (–1,64) decem bra 1921. V tem
letu je pad lo le 725 mm pada vin, kar je pre cej manj od dol go let ne ga pov preč ja, ki je 1032 mm (Tron -
telj 1997). Po letu 1950 sle di več je šte vi lo sušnih obdo bij, z naj več jo pogo stost jo v zad njem deset let ju
(2000–2010). Vred no sti SPI-12 po kla si fi ka ci ji v teh letih bis tve no ne pre se ga jo meje zmer ne suše, razen
decem bra 1971 (–1,75) in decem bra 2003 (–1,68). Indek si na kraj ših časov nih les tvi cah ima jo za leto 2003
bis tve no manj še odklo ne. Ker upo šte va jo le kraj še obdob je vsot pada vin, se v njih ne odra ža jo dalj še
pre te kle sušne raz me re, ki so se zače le že leta 2000 in nada lje va le v leto 2001 in 2002, kar potr ju je tudi
Kobold (2003).
Za meteo ro loš ko posta jo Celje (pre gled ni ca 1) je na let ni časov ni les tvi ci opa zen dalj ši nega ti ven odklon
med leto ma 1854 in 1859. Dalj še obdob je nega tiv ne ga odklo na se ponov no poja vi med leto ma 1861 in 1864
ter se pono vi v ob dob jih od leta 1865 do 1866 ter od leta 1883 do 1885. Sušna obdob ja smo zaz na li še v le -
tih 1920 do 1922, 1924 in 1925. Kraj šim nega tiv nim odklo nom sle di jo namo če na obdob ja, ki dose že jo
ekstrem ne vred no sti v le tih 1937 in 1938. Sle di ta dalj ši sušni obdob ji med leto ma 1941 in 1944 ter med leto -
ma 1945 in 1948. Izsto pa leto 1946, ko so prak tič no vsi mese ci ime li nega tiv ni indeks. Po letu 2000 izsto pa
leto 2003, ko vred nost indek sa po kla si fi ka ci ji dose že mejo hude suše.
Naj niž ji indek si za meteo ro loš ko posta jo Novo mesto (pre gled ni ca 1) se na let ni časov ni les tvi ci po kla -
si fi ka ci ji gib lje jo v me jah zmer ne do hude suše (pre gled ni ca 2). Kraj ša obdob ja nega tiv ne ga odklo na so
pre cej pogo sta (Še be nik 2012). V zad njih treh deset let jih so pogo stej ša tudi dalj ša obdob ja pri manj klja ja
pada vin. V letu 2007 se nega ti ven odklon kaže sko zi celo leto.
Acta geographica Slovenica, 57-1, 2017
47
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 47
Urša Šebenik, Mitja Brilly, Mojca Šraj, Ana li za sušnih raz mer s po moč jo stan dar di zi ra ne ga pada vin ske ga indek sa (SPI)
Izra ču ni SPI-12 za meteo ro loš ko posta jo Trst (pre gled ni ca 1) kaže jo, da obdob ja s pri manj klja jem ne
dose ga jo veli ke inten ziv no sti, saj je mini mal na vred nost indek sa v ce lot nem ana li zi ra nem obdob ju ena -
ka –1,01. Do leta 1900 se na let ni časov ni les tvi ci kaže pred vsem izme nja va dalj ših ekstrem no mokrih obdo bij
s kraj ši mi sušnej ši mi obdob ji. Podob no se nada lju je tudi v dvaj se tem sto let ju z naj niž ji mi vred nost mi
v letu 1946. Tudi v dru gi polo vi ci 20. sto let ja se kaže podo ben vzo rec. V zad njih ana li zi ra nih letih izstopa
leto 2003, ki ga zaz na jo vse les tvi ce kraj še ga tra ja nja.
Ugo to vi mo lah ko, da v zad njem ana li zi ra nem deset let ju na vseh obrav na va nih posta jah izsto pa leto 2003.
V tem letu je ekstrem na suša v Evro pi dose gla enorm ne stroš ke v vi ši ni 8,7 mi li jar de evrov (Com mis sion
of the Euro pean Com mu ni ties 2007). V Slo ve ni ji je ško da zara di suše gle de na celot no ško do zara di narav -
nih nesreč v letu 2003 zna ša la kar 83,3% (Zorn in Komac 2011).
3.3 Pri mer ja va SPI med izbra ni mi posta ja mi za enot no obdob je meri tev 1951–2004
Izra ču na ne vred no sti SPI za vse izbra ne posta je in vse časov ne les tvi ce smo pri mer ja li tudi za enot no obdobje
meri tev. Na dalj ših časov nih les tvi cah vse posta je kaže jo podob no raz po re di tev glav nih suhih in mokrih
obdo bij (sli ka 3). Do več je raz li ke pri de v letu 2002, kjer izsto pa Trst z izra zi to mokrim letom, na osta lih
meteo ro loš kih posta jah pa se v tem času že naka zu je jo ekstrem ne sušne raz me re, ki so v letu 2003 pri zadele
vse obrav na va ne loka ci je. Podrob nej ša ana li za je poka za la, da meteo ro loš ke posta je Ljub lja na-Be ži grad, Novo
mesto in Trst kaže jo podo ben vzo rec pojav lja nja sušnih in mokrih obdo bij, podob no pa bi lah ko rekli tudi
za meteo ro loš ki posta ji Celje in Mari bor-Ta bor (Še be nik 2012). Raz li ke zno traj vsa ke sku pi ne se kažejo v inten -
ziv no sti suše, ki se lah ko raz li ku je za cel raz red ter tra ja nju in začet ku sušne ga obdob ja, ki se lah ko raz li ku je za
mesec ali dva. Suša niko li ne zaja me ena ko mer no celot ne Slo ve ni je, kar potr ju je trdi tev, da je suša regiona -
len pojav (Ko bold 2003). V zad njem obdob ju se šte vi lo ekstrem nih dogod kov pove ču je in hkra ti intenzivi ra.
Pri mer ja va rezul ta tov SPI-12 za celot no obdob je meri tev posa mez ne posta je in za izbra no enot no obdob -
je je poka za la, da so vred no sti kore la cij ske ga koe fi cien ta za vse posta je nad 0,95, kar pome ni, da časov no
obdob je v pri me ru izbra nih meteo ro loš kih postaj ne vpli va v ve liki meri na vred no sti SPI (Še be nik 2012).
Sli ka 3: SPI-12 za obrav na va ne meteo ro loš ke posta je za enot no obdob je meri tev 1951–2004.
Glej angleš ki del pris pev ka.
3.4 Raz mer je med SPI in stan dar di zi ra nim sred njim meseč nim pre to kom za poreč je
Pesni ce
Ana li za rezul ta tov za obdob je 1970–2009 je poka za la, da je med se boj na pove za nost stan dar di zi ra nih preto -
kov za reko Pesni co in SPI za vse časov ne les tvi ce pozi tiv na, ven dar se vred no sti Pear so no ve ga koe fi cien ta
kore la ci je spre mi nja jo gle de na dol ži no časov ne les tvi ce. Ugo to vi mo lah ko, da so kore la cij ski koe fi cien ti
viš ji na kraj ših časov nih les tvi cah in da se pojav lja jo poz no spom la di, pole ti (ju lij in avgust) in jese ni (sep -
tem ber, novem ber) (sli ka 4).
Sli ka 4: Pri kaz meseč nih kore la cij skih koe fi cien tov med stan dar di zi ra ni mi pre to ki in SPI.
Glej angleš ki del pris pev ka.
Naj viš ja kore la ci ja (= 0,754) je med sep tem br skim SPI-2 in sep tem br skim pre to kom (sli ka 5). Rezul -
ta ti kaže jo, da na sred nje meseč ne pre to ke reke Pesni ce v več ji meri vpli va jo pada vi ne teko če ga in pre te kle ga
mese ca, kar kaže na hiter odziv vodo to ka ozi ro ma poreč ja na pada vi ne. Pri mar ni višek vode reke Pesnice
pra vi lo ma nasta ne v me se cu apri lu (Kol be zen 1998). Takrat se reka Pesni ca hitro odzo ve na več jo koli či -
no vode zara di talje nja sne ga ali obil nej ših pada vin. Sekun dar ni višek nasta ne v no vem bru (Kol be zen 1998),
kar se rav no tako dobro uje ma z in dek som SPI na kraj ših časov nih les tvi cah. Polet ni mese ci kaže jo boljše
uje ma nje na dalj ših časov nih les tvi cah, ko so v vso tah pada vin všte ti tudi mese ci zgod nje pom la di in zime,
ko je koli či na pada vin več ja. Ekstrem ne vred no sti stan dar di zi ra ne ga pre to ka se časov no dobro uje ma jo
z ek strem ni mi vred nost mi SPI-2, so pa neko li ko viš je.
Sli ka 5: Stan dar di zi ra ni pre tok in SPI na dvo me seč ni časov ni les tvi ci v sep tem bru.
Glej angleš ki del pris pev ka.
48
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 48
4 Raz pra va
SPI na kraj ših časov nih les tvi cah kaže veli ko varia bil nost in več je šte vi lo kraj ših sušnih dogod kov. Sušne
raz me re na dalj ših časov nih les tvi cah so manj pogo ste, ven dar tra ja jo dlje. Dalj še časov na les tvi ce ne pre -
poz na jo nuj no vseh nega tiv nih odklo nov, ki so vid ni na kraj ših časov nih les tvi cah. Prav tako kraj ši pada vin ski
sko ki na kraj ših časov nih les tvi cah ne pome ni jo nuj no izbolj ša nja sušnih raz mer na dalj ši les tvi ci. Pri anali -
zi pre te klih obdo bij se je tre ba zave da ti, da več zapo red nih mese cev z ne ga tiv ni mi vred nost mi indek sa ne
pome ni nuj no sušne ga obdob ja. Nega tiv na vred nost indek sa namreč pred stav lja mese ce, ko je pad la manj -
ša koli či na pada vin v pri mer ja vi z dol go let nim pri mer jal nim obdob jem.
Po manj ka nje pada vin je eden od glav nih vzro kov nastan ka suše, ven dar ne edi ni (Vi cen te-Ser ra no in
osta li 2010), saj so pomemb ni vpliv ni dejav ni ki za raz voj suše tudi eva po trans pi ra ci ja, tem pe ra tu ra, hitrost
vetra, vodo za dr že val na spo sob nost tal ter vpli vi člo ve ka. Pomanj ka nje pada vin v zim skih mese cih je prob -
le ma tič no z gle diš ča boga te nja pod tal ni ce in dru gih vod nih virov, ki so pomemb ni dejav ni ki pri oskr bi
s pit no vodo v Slo ve ni ji. SPI upo šte va samo pada vi ne, zato je tre ba za podrob nej šo ana li zo posa mez ne vrste
suše upo ra bi ti še dru ga orod ja: sušne indek se, ki poleg pada vin vklju ču je jo tudi dru ge spre men ljiv ke, vod -
no bi lanč ne mode le, ana li zo niz kih pre to kov rek ipd. SPI zato pred stav lja pred vsem prvo infor ma ci jo o su šnih
raz me rah (Ha yes in osta li 1999).
Za iden ti fi ka ci jo sušnih raz mer smo ana li zi ra li tudi dalj ša obdob ja, ki so hkra ti kazal ci hidro loš kih
sušnih raz mer na povr šin skih in pod zem nih vod nih virih (Mc Kee in osta li 1993; Hayes in osta li 1999).
Indeks, pri me ren za ope ra tiv no rabo pri uprav lja nju z vod ni mi viri, mora biti repre zen ta ti ven za sušne
raz me re v hi dro loš kem siste mu na dolo če ni časov ni les tvi ci izra ču na. Za reko Pesni co smo naj viš jo kore -
la ci jo med stan dar di zi ra ni mi pre to ki in SPI zaz na li na dvo me seč ni časov ni ska li. Pri mer kaže, da je tre ba
za vsa ko poreč je pose bej dolo či ti naj pri mer nej šo les tvi co izra ču na, saj se hidro loš ke, meteo ro loš ke in relief -
ne zna čil no sti bis tve no raz li ku je jo. Iz iste ga raz lo ga rezul ta ta ne more mo pos plo ši ti za celo Slo ve ni jo. Na
tem področ ju do sedaj še ni bilo veli ko razi skav in ugo tov lje nih goto vih pove zav med raz lič ni mi časovni -
mi obdob ji sprem lja nja sušnih raz mer in vod ni mi viri.
V štu di ji smo pre to ke stan dar di zi ra li po nor mal ni poraz de li tvi zara di več je pri mer lji vo sti ter bolj še
oce ne med se boj ne pove za no sti SPI in stan dar di zi ra ne ga pre to ka, kar omo go ča laž jo pri mer ja vo meteo -
ro loš kih in hidro loš kih spre men ljivk.
5 Sklep
Su ša in z njo pove za no pomanj ka nje vode se kaže ta kot izziv za celo ten svet, pri tem pa tudi Slo ve ni ja ni
izje ma. Slo ve ni ja se sicer uvrš ča med drža ve, ki so z vi di ka vod na to sti rela tiv no boga te. Ven dar pa je kljub
viso kim skup nim koli či nam dež ja za raz lič ne dejav no sti (ka ko vost na kme tij ska pri de la va, oskr ba s pit no vodo,
proi zvod nja elek trič ne ener gi je) časov na raz po re di tev pada vin pogo sto neu god na (Gre go rič in Sušnik 2008).
V pre te klih letih je tudi v Slo ve ni ji ško da zara di suše dose gla viso ke zne ske (Zorn in Komac 2011). Rezulta ti
kaže jo, da je daleč naj več ji delež (48,6 %) v ce lot nem obse gu oce nje ne ško de v le tih od 2000 do 2005 pov -
zro či la prav suša (Re vi zij sko poro či lo … 2007). Podat ki nam torej kaže jo, da mora mo tudi v Slo ve ni ji na
sušo resno raču na ti.
6 Lite ra tu ra
Glej angleš ki del pris pev ka.
Acta geographica Slovenica, 57-1, 2017
49
57-1_02_729-Ursa Sebenik_acta49-1.qxd  5.5.2017  9:29  Page 49