Αντικείμενο της εργασίας αποτελεί η επισκόπηση των πιο σημαντικών δεικτών ξηρασίας, με παράλληλη ανάλυση και εφαρμογή του δείκτη δριμύτητας ξηρασίας Palmer για δεδομένα που είχαν συλλεχθεί και αφορούν στην περιοχή της λίμνης Υλίκης. Η μέθοδος αποτυπώνεται αναλυτικά σε φύλλο εργασίας excel με στόχο να μπορεί εύκολα να χρησιμοποιηθεί για τον υπολογισμό του δείκτη ξηρασίας Palmer και σε άλλες περιοχές. Απώτερος στόχος της εργασίας ήταν να δοθεί η δυνατότητα πειραματισμού και βελτίωσης του δείκτη, δεδομένου ότι όλα τα στάδια υπολογισμού του είναι φανερά στον χρήστη. Η δομή της εργασίας βασίζεται στα εξής 8 κεφάλαια:
Κεφάλαιο 1 - Επιχειρείται η παρουσίαση και κατανόηση της ξηρασίας με παράλληλη αναφορά στις διάφορες και διαφορετικές μορφές ξηρασίας, των βασικών χαρακτηριστικών της, των παραμέτρων εκτίμησης της, των αιτιών εμφάνισης της καθώς και των επιπτώσεων της.
Κεφάλαιο 2 – Εισάγεται η έννοια του δείκτη ξηρασίας ως μέτρο ποσοτικοποίησης της ξηρασίας και επιχειρείται η ιστορική καταγραφή των σημαντικότερων δεικτών ξηρασίας. Παρουσιάζονται πιο αναλυτικά οι πιο διαδεδομένοι δείκτες ξηρασίες.
Κεφάλαιο 3 – Πραγματοποιείται μία σύντομη παρουσίαση της μεθοδολογίας εκτέλεσης υπολογισμών του δείκτη δριμύτητας ξηρασίας Palmer.
Κεφάλαιο 4 – Παρουσιάζεται αναλυτική η μεθοδολογία εκτέλεσης υπολογισμών του δείκτη δριμύτητας ξηρασίας Palmer, με παράλληλο σχολιασμό και παράθεση παρατηρήσεων επί της μεθοδολογίας.
Κεφάλαιο 5 – Εφαρμόζεται η μεθοδολογία εκτέλεσης υπολογισμών του δείκτη δριμύτητας ξηρασίας Palmer και αναλύονται τα αποτελέσματα αυτής. Παράλληλα, παρουσιάζεται η περιοχή μελέτης και οι παραδοχές που λήφθηκαν υπόψη κατά το στάδιο εφαρμογής της μεθοδολογίας.
Κεφάλαιο 6 – Παρατίθενται τα συμπεράσματα της εργασίας μετά την εφαρμογή της μεθοδολογίας εκτέλεσης υπολογισμών του δείκτη δριμύτητας ξηρασίας Palmer.
Κεφάλαιο 7 – Παρατίθενται οι ελληνόγλωσσες και ξενόγλωσσες πηγές της μεταπτυχιακής εργασίας.
Κεφάλαιο 8 – Αναλυτική παράθεση όλων των δεδομένων εισόδου και αποτελεσμάτων σε πίνακες και διαγράμματα.

Introduction and Structure
Even though Greece is considered as having adequate water resources, drought is becoming an increasingly often phenomenon that requires on time identification and application of appropriate management practices. This postgraduate thesis was elaborated as part of the "Water Resources Science and Technology" Interdisciplinary – Interdepartmental Postgraduate Programme of the National and Technical University of Athens (NTUA) and specifically the "Hydrology and Environmental Management of Water Resources" course direction. This postgraduate thesis provides an overview of the most important drought and scarcity indices. It also attempts to analyze and implement the Palmer Drought Severity Index (PDSI) for data collected from the study area of Lake Yliki. The method has been "built" from scratch and is depicted in detail in an excel spreadsheet in order to be easily implemented in every possible area. The ultimate goal of this work was to enable experimentation and improvement of the PDSI index, since all stages of the method are apparent to the user. The structure of this postgraduate thesis is the following:
Chapter 1 – It is attempted to present and comprehend drought while referring to the different forms of drought, key features, drought estimation parameters, the causes of its occurrence and its impacts.
Chapter 2 - Introduction of the drought index as a measure for quantifying drought. Elaboration of a historical record of major drought indices, providing specific details for the most commonly used.
Chapter 3 - Brief presentation of the PDSI methodology of drought severity quantification.
Chapter 4 - Detailed presentation of the PDSI methodology, while commenting on the methodology.
Chapter 5 – Implementation of the PDSI methodology on Lake Yliki’s data. Presentation of the study area and determination of the assumptions considered during the implementation of the methodology.
Chapter 6 – Presentation of the conclusions after implementing the PDSI methodology.
Chapter 7 – Greek and Foreign Bibliography.
Chapter 8 - Detailed presentation of all input data and results in tables and charts.
Analytical Purpose
The purpose of this postgraduate thesis was to create a spreadsheet, "accessible" to users, that will allow them to enter Palmer method’s inputs (rainfall and temperature) and with simple adjustments (due to the different sample size of their data) to calculate the Palmer Drought Severity Index (PDSI) for the given area which is being studied. Moreover, it aimed at analyzing the methodology used to export the results through the implementation of the Palmer’s method, due to the fact that many applications of the PDSI were available on the internet in a “black box” form, depriving the ability to understand the computational procedure and steps that were being followed for the calculation of the index. Another goal of this thesis was to review the most widely used drought indices and finally, apply the methodology developed by Palmer (Palmer, 1965) and studied, reviewed and commented by Alley (Alley, 1984) in Yliki’s Lake area. Drought
Drought is an extreme hydrological phenomenon particularly complex to understand, study, prevent and treat, mainly due to the plurality of parallel actions and parameters involved in its emergence (Apostolou, 2010).
A general definition of drought is that is an extreme hydrological condition characterized by a deficiency of water. Furthermore, a more specific definition is presenting drought as a natural, random, temporary and irregular situation of a natural irregular process of continuous reduction of precipitation and water availability, with respect to normal values, which extends over a considerable period of time and covers a wide area. However, there is not an axiomatically, objective and universally accepted definition of drought, as it is defined only from the standpoint of a given scientific side or study area. Drought is considered the most complex of all natural hazards (Wilhite and Glantz, 1985; Wilhite et al., 2000). Its specific characteristic is that it can occur many years after the termination of the phenomenon (Mimikou & Fotopoulos, 2004). Drought indices Drought is quantified through drought indices. Drought index is an index that is associated with some of the cumulative effects of a prolonged and abnormal moisture deficiency (World Meteorological Organisation, 1992). The ultimate characteristics of a complete drought index are the following:
Simplicity - ease of use and understanding.
Correct definition - with scientifically accepted physical sense.
Sensitivity - response to a large range of values.
Timely response to climate fluctuations.
Transferability - ability to use the index in different regions.
Data availability - access to long time series and good data.
Low cost to produce the index (Tsakiris & Pangalou, 2009).
The Palmer Drought Severity Index (PDSI) The Palmer Drought Severity index was developed in 1965 by Wayne C. Palmer. It is a hydrological balance index which is based on the application of a water balance concept considering water supply and demand. It is a soil moisture algorithm which is calibrated for relatively homogeneous regions and its ultimate objective is to provide standardized measurements of moisture comparisons between different locations and months. It is widely used by U.S. Government agencies enabling them to address drought. A specific characteristic of the PDSI is the fact that its results are being recorded in a monthly time step. Moreover, as shown in scheme 1 below, the soil moisture is being stored in two soil layers that are being arbitrarily defined by Palmer (upper soil layer) and by each researcher (underlying layer) that is implementing the Palmer’s method on a given study area.
Furthermore, Palmer’s method includes assumptions related to the way that water is being transferred between these two soil layers. For instance, evapotranspiration losses take place only if the potential evapotranspiration is higher that the precipitation and the losses from the top layer are being utilized to the maximum possible extend. On the other hand, losses from the underlying soil layer are subject to the initially stored moisture in the underlying layer, the potential evapotranspiration and the combined available moisture capacity (AWC - Available Moisture Capacity) of two soil layers.
Advantages of the PDSI
The PDSI is the first comprehensive drought index that is widely used, measuring the climate anomaly in an area, placing the current conditions in historical samples, providing spatial and temporal representations of historical droughts and measuring the soil moisture conditions (Willeke et al., 1994). It is very effective in large areas of uniform topography and readily available and standardized, facilitating direct spatial and temporal comparisons.
Disadvantages of the PDSI
Some indicative disadvantages of the PDSI are that it delays to highlight the ongoing drought for several months and it is unsuitable for very mountainous areas or areas with extreme climatic events (Smith et al., 1993). Furthermore, it does not calculate the snow and ice and does not take into account the runoff until the capacity of the top and bottom soil layers are complete in water, thus leading to underestimation of runoff. Finally, it is widely used in the United States but has little acceptance in other areas (Kogan, 1995). Brief presentation of the Palmer Drought Severity Index method In brief, the Palmer Drought severity index method, consists of the following steps:
1. Carry out a hydrologic accounting by months for a long series of years.
2. Summarize the results to obtain certain constants or coefficients which are dependent on the climate of the area being analyzed.
3. Reanalyze the series using the derived coefficients to determine the amount of moisture required for "normal" weather during each month.
4. Convert the departures to indices of moisture anomaly.
5. Analyze the index series to develop:
a. Criteria for determining the beginning and ending of drought periods. b. A formula for determining drought severity.
Analytical computational procedure of Palmer’s method
Below are presented in brief, the analytical steps that are being implemented in Palmer’s computational procedure.
1. Definition of the study area and definition of the necessary assumptions.
2. Finding, processing and tabulation of the monthly rainfall (P) and potential evapotranspiration (PE) data.
3. Calculation of the difference between potential evapotranspiration and precipitation in a monthly step.
4. Definition the available water capacity of the soil (upper layer and the underlying layer) (AWC).
5. Calculation of moisture loss from the upper and underlying soil layer in a monthly step.
6. Calculation of the total losses.
7. Calculation of the stored moisture in the upper and underlying soil layer.
8. Calculation of the total stored moisture.
10. Estimation of the potential recharge (PR), potential loss (PL) and Potential Runoff (PRO).
11. Calculate the average monthly values of all variables of the studied sample.
12. Calculate the following constants Evapotranspiration (a) Recharging (b) Runoff (c) Loss (d).
13. Calculating CAFEC values of Evapotranspiration (ET) Recharging (R), Runoff (RO) and loss (L).
14. Calculation of the CAFEC Precipitation (Pcafec).
15. Calculation of the rainfall excess and deficits deficit (d).
16. Calculation of the climate characteristic K
17. Calculation of the index disorder moisture (z).
18. Calculation of drought severity for the initial month of the studied sample.
19. Calculation of the Drought Severity Index Palmer.
Furthermore, the analytical steps of Palmer’s computational procedure, including basic functions, are being illustrated in scheme 2 below.
Implementation of the Palmer methodology and results
The implementation of Palmer’s methodology in Yliki’s area data has resulted in the conclusion that the applied methodology successfully records the two-year droughts that were recorded during 1988-90 and 2000-2002 in the area. Taking under consideration diagram 1, which is presented below, the results are grouped in the following four classes.
First class 1977 – 1985 - Positive PDSI values - Slightly wet climate conditions with an average index value of 1.96.
Second class 1985 – 1994 - Negative PDSI values - Mild drought climatic conditions with an average index value of -1.36.
Third class 1994 – 1999 - Positive PDSI values - Near normal climatic conditions with an average index value of 0.16.
Fourth class 1999 – 2002 - only negative PDSI values - Moderate drought climatic conditions with an average index value of - 2.78.
According to the extracted results, evapotranspiration shows a periodicity, both in terms of annual and monthly changes.
However, the relation between rainfall intensity and the PDSI is very close. It appears that low values of rainfall are followed by low values (negative) of the PDSI. The four largest PDSI values are being recorded during January, June and May 2002 and during July 1990 with values of -6.48, - 6.45, -6.15 and -6, respectively. Moreover, it results that 50 items (months) are assigned to near normal climatic conditions (-0,5 <PDSI <0,5). Furthermore, 18 months show values smaller than -4, while 16 months are recording extremely wet climatic conditions. The periods during which there is an extreme drought with negative values greater than PDSI = -4 are:
From March until August 1989.
From June until October 1993.
From November 2001 until June 2002.
Conclusions
The general conclusions that have been extracted, after the application of Palmer’s methodology to Lake Yliki’s area, are the following.
The process and computational steps of the index which are reflected in the excel spreadsheet that has been created, record successfully the biennial extreme droughts that have historically occurred in 1988-89 and 1989-90 as well as during 2000-01 and 2001-02.
The Palmer index successfully captures the drought conditions in the study area. The droughts recorded were conducted by indistinct years and does not seem to exhibit any periodicity.
The calculated values of drought severity index show reasonable results qualitatively classifying the droughts in the extremely dry periods, with values exceeding the benchmark of the very dry season (PDSI <-4).
There is a trend toward decreasing values of the Palmer index, defining a shift to drier conditions.
The cases where extreme droughts are recorded are not isolated events that occur sporadically for only one month, but instead are periods lasting at least 5 months.
Even assuming different initial conditions (initial values) where the two soil layers, upper and underlying, were completely empty, the results of the index are not affected either qualitatively or quantitatively.
Suggestions for further research The creation of the excel spreadsheet that integrates Palmer’s computational procedure and the implementation of Palmer’s methodology in Yliki Lake, has triggered the need for further investigation of some aspects of the method that can be optimised. These are the following:
Recalibration of the PDSI using data from EU regions or from corresponding regions from Greece.
Integration of a process that will take into account the following:
o Constructions that regulate the supply of water resources (eg. Dams).
o The irrigation needs of the studied area.
Integration of a tool that can calculate the potential evapotranspiration during the first stage of the water balance calculation.
Creation of a list that provide the thickness of the underlying soil layer through objective criteria (geological characteristics / underground conditions, soil characteristics, etc.).
Codification of the created excel spreadsheet and creation of an online tool for the determination of drought in Greece.