Διερεύνηση χωρικής κατανομής μετεωρολογικής ξηρασίας

Καμίτσου, Χάρις-Αθηνά; Kamitsou, Charis-Athina

dc.contributor.author	Καμίτσου, Χάρις-Αθηνά	el
dc.contributor.author	Kamitsou, Charis-Athina	en
dc.date.accessioned	2016-06-21T10:36:05Z
dc.date.available	2016-06-21T10:36:05Z
dc.date.issued	2016-06-21
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/42778
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.5408
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) “Επιστήμη και Τεχνολογία Υδατικών Πόρων”	el
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.subject	Ξηρασία	el
dc.subject	Δείκτης τυποποιημένης βροχόπτωσης	el
dc.subject	Χρονοσειρές	el
dc.subject	Κατακρήμνιση	el
dc.subject	Μεσόγειος	el
dc.subject	Drought	en
dc.subject	Standardized Precipitation Index (SPI)	en
dc.subject	Time scales	en
dc.subject	Precipitation	en
dc.subject	Mediterranean	en
dc.title	Διερεύνηση χωρικής κατανομής μετεωρολογικής ξηρασίας	el
dc.title	Investigation of the spatial distribution of meteorological drought	en
dc.contributor.department	Επιστήμη και Τεχνολογία Υδατικών Πόρων	el
heal.type	masterThesis
heal.classification	Υδρολογία	el
heal.language	el
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2015-10-19
heal.abstract	Η ξηρασία και η έγκαιρη αναγνώριση της, αποτελούν μείζον θέμα της σημερινής κοινωνίας, εξαιτίας των δυσμενών επιπτώσεων που αυτή επιφέρει. Η ανάγκη για έγκαιρη πρόγνωση οφείλεται στην δυνατότητα που δίνει για να οργανωθεί μια κοινότητα και να την αντιμετωπίσει. Προκειμένου να επιτευχθεί επομένως η έγκαιρη αναγνώριση ενός φαινομένου ξηρασίας, έχουν αναπτυχθεί διάφοροι δείκτες. Μέσω αυτής της μελέτης, πραγματοποιείται μια σύντομη επισκόπηση της ξηρασίας και μερικών από τους πιο διαδεδομένους δείκτες που χρησιμοποιούνται προκειμένου αυτή να ανιχνευθεί για να μπορέσει να αντιμετωπιστεί. Παρόλα αυτά, μελετάται συγκεκριμένα ο Δείκτης τυποποιημένης βροχόπτωσης (Standardized Precipitation Index - SPI), και κατατάσσονται οι εντάσεις των ξηρασιών των υπό μελέτη περιοχών σύμφωνα με την κατάταξη του McKee. Έπειτα, μελετώνται ιδιαίτερα οι ξηρασίες που σύμφωνα με την παραπάνω κατάταξη, χαρακτηρίζονται ως ακραίες. Επιπλέον, διερευνάται η δυνατότητα μελέτης της ξηρασίας σε επίπεδο χωρικής κατανομής. Πιο συγκεκριμένα, στην παρούσα εργασία, μελετήθηκαν οι ξηρασίες στη Μεσόγειο για το χρονικό διάστημα 1950-2014. Συνολικά, 64 χρονοσειρές μελετήθηκαν σε μηνιαία βάση και αναλύθηκαν σε ετήσια βάση (υδρολογικού έτους), για 56 σταθμούς κατανεμημένους στη Μεσόγειο που επιλέχθηκαν τυχαία. Η ανάλυση της ξηρασίας, πραγματοποιήθηκε με την βοήθεια του Δείκτη τυποποιημένης βροχόπτωσης (Standardized Precipitation Index - SPI). Τα ακραία επεισόδια ξηρασίας σε βάθος 64 ετών στη Μεσόγειο, που καταγράφονται ταυτόχρονα σε πέντε ή και περισσότερους σταθμούς, εμφανίζονται τα έτη 1989-90, 1999-00, 1960-61 και 1981-82, με αύξουσα σειρά εμφάνισης. Από την άλλη, οι σταθμοί με περισσότερα από τέσσερα επεισόδια ακραίων ξηρασιών, σε βάθος χρόνου 64 ετών είναι με αύξουσα σειρά εμφάνισης οι σταθμοί Ain Sefra της Αλγερίας και Antalya της Τουρκίας. Τέλος, εξάγονται κάποια συνολικά συμπεράσματα από την εργασία και παρατίθενται προτάσεις για μελλοντική έρευνα.	el
heal.abstract	A drought occurs when water availability reduces temporarily for example because of insufficient rainfall (Makropoulos, 2015). That complicated phenomenon not only occurs in areas with dry - desert climate, but also in areas with significant rainfall amounts, mainly due to the connection of drought with more than one type of environmental conditions (Karabourniotis, 2014; Vicente-Serrano, 2006). Each drought episode is different from the rest thanks to three key features that it has: intensity, duration and its spatial distribution (Nalbantis (a), 2012; Paparas, 2011; Skarantonakis, 2010). Drought indicators are tools that identify the characteristics of drought (Nalbantis (a), 2012). So they help identify and record the intensity and extent of drought episodes. Additional, from the study of indicators, they may indicate the possibility of reoccurrence of drought events (Paparas, 2011; Skarantonakis, 2010). There are many types of indicators, but some indicators compared to others have better results for specific situations or areas. Most of the water managers, before taking a decision consult one or more indicators (Karabourniotis, 2012; Paparas, 2011; Skarantonakis, 2010). In this study the index chosen was Standardized Precipitation Index (SPI). Standardized Precipitation Index (SPI) SPI was developed by McKee et al. in 1993, to quantify the precipitation deficit for different time scales and can be effectively used for either wet or dry periods. (Mamasis et al., 2012; Paparas, 2011; McKee et al., 1993; WMO, 2012). When the SPI has positive values, the present precipitation is greater than the average, while it has negative values, indicate precipitation less of the average price. In this analysis classification of McKee et al. was adopted due to identify and evaluate droughts. McKee et al. classified and described the intensity of drought in the following categories according to the value of the SPI as shown in Table 1 (Karabourniotis, 2012; Mamasis et al, 2012; McKee et al., 1993; WMO, 2012): Table 1 Characterization of droughts according to SPI index and colour scale that has been used in this study (McKee et al., 1993). Price range of SPI index Drought Characterization 0,0  SPI > –0,99 Mild –1,0  SPI > –1,49 Moderate –1,5  SPI > –1,99 Severe –2,0  SPI Extreme Each drought episode is characterized, with the help of the indicator, by a beginning, an end and a tension for each month of emergence episode. This episode begins when the SPI index takes a negative value and is completed when it takes a positive value (Karabourniotis, 2012; Papalexis, 2013; Skarantonakis, 2010; McKee et al., 1993). After testing the robustness of 18 different drought indicators and the use of statistical methods, Keyantash & Dracup (2002), concluded that the SPI index is the best climate indicator to identify droughts, as well as to quantify the severity, the duration and the spatial distribution of drought. Methodology Study area The study of time series on drought was held in the Mediterranean region. All countries bordering the Mediterranean have similar climatic conditions, but more as a general rule, droughts occur to countries in the southern section (Africa). In order to study the climate and particularly the drought in the Mediterranean was used as database the Dutch website KNMI Climate Explorer which is available at: http://climexp.knmi.nl/start.cgi?id=someone@somewhere. The KNMI Climate Explorer is a research tool for the statistical analysis of climate data and contributes to understanding climate through the time series (KNMI, 2015). Data collection Specifically there are held eight different combinations of coordinates in order to obtain adequate data. Basically we gathered data to study the rainfall and then drought for countries - stations of the Mediterranean which are shown in Table 2. Table 2 Countries - stations of drought study Countries - Stations Greece Hiraklion Libya Tripoli Hellenikon Misurata Athens Sirte Kalamata France Ajaccio Kerkyra Nice Araxos Marseille Cyprus Larnaka Perpignan Egypt Mersa Matruh Nimes Turkey Mugla Toulouse Antalya Spain Barcelona Isparta Palma de Mallorca Usak Tortosa IZMIR Alicante Afyon Ibiza airport Konya Almeria Italy Messina Albacete Trapani,Birgi Algeria Oran_es Senia Brindisi Alger_dar el beida Cagliari Elmas Miliana Trieste Ain Sefra Pisa san Giusto Orleans ville Malta Luqa Biskra Tunisia Tunis,Carthage Constantine Kairouan Croatia Zagreb Jendouba Montenegro Titograd Gabes Hungary Pecs Gafsa Austria Klagenfurt Graz Sonnblick During the collection of data occurred some gaps, which were supplemented by calculating the overall average of the available values from the database in order to have a more reliable result. The most common reasons to create gaps in measurement values are the possible malfunctioning of an instrument tolerance as well as the absence of an observer (Mimikou, 2006). Climatic Data Calculation procedure To calculate the SPI drought index in the Mediterranean region, there were used 64 time series of precipitation for the period of 1950-2014 and they were analyzed based on the gamma distribution. In any gaps that existed, there was calculated an average price data for a longer period of time without many shortcomings in all years for each month, in order to estimate the missing value approaching as possible a theoretical actual value. Then there was created a table with all the data and additional their values were calculated per three months, six months, nine months and twelve months, but always according to the hydrological year, which starts from October. Then, for each of these columns and for each station, the average value was calculated, as also the standard deviation and the Alpha and Beta variables (as shown below) to calculate the gamma (for three months, six months, nine months and hydrological year): Alpha=((Average value )/(Standard deviation ))^2 Beta=(Standard deviation )^2/(Average value) Next, two new tables were created. In the first of them, for each original data table calculated, respectively, the equation of the probability density function of gamma in accordance with the following equation: f(x;α,β)=1/(β^α Γ(α)) x^(α-1) e^(-x/β) Eventually, a table, with calculated value of SPI index was created using the command Normsinv(probability), which performs the inverse of the standard normal cumulative distribution for distribution with arithmetic mean 0 and standard deviation 1. So if 100% of a chance removed the corresponding value of the table where the distribution gamma applied, the result is the possibility to overcome this case the normal standard distribution. The probability obtained in the end, gives the value which corresponds to a normal distribution, that possibility overcomes the normal standard distribution. This final value obtained is the value of SPI. In order to immediate and quickly discern droughts and their intensity, in each one of these final tables stained the cells with corresponding intensity droughts (as at Table 3). Finally, for even greater convenience of calculations and comparisons, all the values in the SPI based hydrological year were placed in a new table and a plot was built on each channel for the entire time series. Moreover, the multitude of each drought’s intensity was estimated, as long as the multitude of each drought’s intensity per station and per hydrological year. Results and discussion After the study and analysis performed on the drought in the Mediterranean, we can discern droughts occurring at each station according to SPI index for three-month, six-month, nine-month basis and for hydrological year basis for the length of the time series (64 years). According to Table 3, we can discern that the total number of droughts is 1654 for all 64 years. Of these, only 94 are considered extreme droughts, which is just 5.68% of total drought. Mild droughts amounting to 1101 and appear to far outweigh the others by a percentage of 66.57% overall droughts (ie more than half). Then moderate droughts follow amounting to 320 and 19.35% percentage and severe to 129 and 7.80% percentage. So it is observed that fortunately the greater drought, the smaller area occupies. Below is shown Chart 1, which depicts the above percentages for each drought category on the total droughts and, for convenience, each category has the same colour as Table 3. Chart 1 Percentages in each category drought in all of them In Table 3 below are depicted values of SPI for the hydrological years of the time series in the Mediterranean stations. The degree of intensity for each drought is classified according to McKee et al. (1993), in all cases and for ease of monitoring the events, there is common colouring for each intensity in all figures and tables of that study. Specifically, all the droughts that have been characterized as mild are somon (light pink) coloured, moderate droughts coloured with lila (light purple), severe droughts coloured with mustard (dark orange-yellow) and finally the extreme droughts coloured with bright red colour, as shown in Table 1 with the colour classification of drought intensities. Table 3 Spatial distribution of drought So by analyzing the 94 extreme drought that occurred during the 64 years studied (1950-2014) and Table 3, it was found that in just 4 hydrological years during the time series studied recorded five or more stations extreme droughts in the study area. The most adverse hydrologic year was 1989-1990 where extreme droughts recorded in 11 stations. The hydrological year of 1999-2000 follows with extreme droughts in 7 stations, 1960-1961 with 6 stations and finally 1981-1982 with 5 stations having extreme droughts. So we believe that in most cases throughout the 64 years for the study area, extreme droughts occurred were isolated events. Further, observed that extreme droughts occurring in more than 5 hydrological stations for 4 years, while serious for 5 years. In particular, greater variety of serious droughts displays hydrologic year 1992-93 with 8 stations, followed by 1989-90 with 7 stations, 1988-89 and 1999-00 with 6 stations and finally, 1996-97 with five stations. As observed for moderate droughts occur in 33 of the 64 hydrological years, while mild droughts appear in all hydrological years for more than 5 stations. Additionally, it is worth noting that only 2 of the 56 stations studied, the number of the extreme droughts exceeds 4, for the time series of 64 years. Also, during the studied period, it should be noted that in 14 of the 56 stations, the number of severe droughts exceeds or is equal to four and the number of stations that have four or more moderate droughts greatly increased. Finally, the number of mild droughts overcomes or equals to 10 for all stations. So, it seems that the smaller intensity a drought has, the more of them appears at each station over time. Furthermore, Mediterranean region is not greatly affected by frequent high intensity droughts. Moreover, in Figures 1-4 it is observed the spatial distribution of the intensities of droughts (as also the existence of rainfall) for the hydrological years of 1989-90, 1999-00, 1960-61, 1981-82 and 1988-89. These maps were created for the years with the most extreme droughts and have been placed in descending order of extreme droughts. Observing the maps, noteworthy is that, there is not a certain spatial distribution pattern of droughts in the Mediterranean. It is also remarkable that there is a slight tendency in concentrating of extreme droughts in nearby areas. Figure 2 shows that the hydrological year 1999-00 has the least display of rainfall (from 56 stations studied only the 11 note precipitation instead drought). On the other hand, the year 1962-63 constitutes the year with the most rainfalls (50) and consequently the year with the smallest plurality of droughts. Figure 1 displays the hydrological year 1989-90 having the next lower incidence of rainfalls (16), so the next higher occurrence of droughts. Moreover, the years 1960-61 and 1981-82 (Figure 3 and 4) are the other two years with increased extreme droughts and so they occupy few rainfalls (23 and 20 respectively). So, it is observed the inverse correlation between rainfall and droughts. Finally, Figure 5 shows the hydrological year 1988-89 having a low incidence of rainfalls (just 16), and therefore a bigger of drought. In detail, there are 4 extreme and 6 severe droughts, but also 14 moderate and 16 mild droughts. Thus, it is noticed a uniform spatial distribution of all intensities of droughts in the area, with small local increased concentrations of some of these. Figure 1 Map for hydrologic year 1989-90 (11extreme droughts-most of) Figure 2 Map for hydrologic year 1999-00 (7extreme droughts) Figure 3 Map for hydrologic year 1960-61 (6extreme droughts) Figure 4 Map for hydrologic year 1981-82 (5extreme droughts) Figure 5 Map for hydrologic year 1988-89 (4 extreme droughts) Conclusions After the study of the SPI index and the literature review that was conducted for drought, it was found that the drought is not an easy phenomenon to be predicted, especially concerning the duration period. Eventually, as it was observed by that study, drought is not a spatial phenomenon since it affects either the entire country or a small part of it or even many areas-countries. So, droughts affect an area in different period of time and simultaneously a neighbor area can have either drought or plenty of rainfalls. Specifically, we believe that in most cases the entire time series for the study area as extreme droughts occurred were isolated events. But the year 1989-90 repeatedly observed droughts, with occurrence of extreme droughts in 11 stations and severe droughts in 7 stations. So 1989-90 could be described as the worst drought year. Αlso analyzing droughts per station we conclude that none of the stations has not simultaneously been a large number of extreme and severe droughts. Finally, it is also interesting that the Mediterranean region is not greatly affected by frequent high intensity droughts, but mainly from small. Moreover, droughts do not seem to follow any particular spatial distribution pattern, except that usually affect nearby areas. So, the smaller the drought’s tension is, the more appears at each station and at each hydrological year.	en
heal.advisorName	Μαμάσης, Νίκος	el
heal.committeeMemberName	Μαμάσης, Νίκος	el
heal.committeeMemberName	Μπαλτάς, Ευάγγελος	el
heal.committeeMemberName	Παναγούλια, Διονυσία	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Πολιτικών Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	151 σ.	el
heal.fullTextAvailability	true