Διερεύνηση της μεταβλητότητας του κλίματος σύμφωνα με την κατάταξη Koppen

Abstract

Η παρούσα μεταπτυχιακή εργασία σκοπό έχει να διερευνήσει τη μεταβλητότητα του κλίματος διαφόρων περιοχών του πλανήτη βάσει της κλιματικής κατάταξης Köppen. Για το σκοπό αυτό, συγκεντρώθηκαν δεδομένα μηνιαίων τιμών βροχόπτωσης και θερμοκρασίας για περιοχές της υφηλίου οι οποίες ανήκουν σε κάθε μία από τις 31 συνολικά κλιματικές υποκατηγορίες των 5 βασικών κλιματικών τύπων A, B, C, D και E. Για την κάθε κλιματική υποκατηγορία συγκεντρώθηκαν δεδομένα από δύο σταθμούς, έναν κοντινό (απόσταση έως 50 km) και έναν απομακρυσμένο από τη θάλασσα. Στη συνέχεια, υπολογίστηκαν οι παράμετροι της μεθόδου Köppen και υλοποιήθηκαν εκείνα τα κριτήρια τα οποία καθορίζουν τον κλιματικό τύπο στον οποίο ανήκει ο κάθε σταθμός. Υπολογίστηκαν οι στατιστικές παράμετροι των μετεωρολογικών δεδομένων. Αυτές με τη σειρά τους αποτέλεσαν αντικείμενο ανάλυσης έτσι ώστε να διερευνηθεί το μέγεθος της μεταξύ τους συσχέτισης για τον κάθε κλιματικό τύπο ξεχωριστά καθώς και μεταξύ των κοντινών και απομακρυσμένων από τη θάλασσα σταθμών. Συγκεντρώθηκαν, ακόμη, ημερήσιες τιμές βροχόπτωσης για τις πιο διαδεδομένες υποκατηγορίες των 5 κλιματικών τύπων. Υπολογίστηκαν οι μέγιστες ημερήσιες τιμές βροχόπτωσης στις οποίες και έγινε προσαρμογή της κατανομής Γενικών Ακραίων Τιμών. Εξετάστηκε, επίσης, ο χαρακτηριστικός λόγος της μέγιστης ημερήσιας προς την αντίστοιχη μέση ετήσια ποσότητα βροχόπτωσης. Το πηλίκο αυτό αποτελεί χαρακτηριστικό γνώρισμα της έντασης των ακραίων φαινομένων βροχόπτωσης για τον κάθε κλιματικό τύπο. Ακόμη, σημειώθηκαν εκείνοι οι μήνες του έτους στους οποίους παρουσιάζεται η μέγιστη μέση μηνιαία βροχόπτωση των σταθμών και εξετάστηκε η επίδραση της εγγύτητας των σταθμών στη θάλασσα. Τέλος, διερευνήθηκε η μακροχρόνια κλιματική μεταβλητότητα των σταθμών της κάθε κλιματικής υποκατηγορίας. Η ποσοτικοποίηση της κλιματικής μεταβλητότητας των σταθμών υλοποιήθηκε μέσω της στατιστικής μεθόδου της διαφοράς των μέσων τιμών και της κατανομής t-student.

The aim of this thesis was to investigate the climatic variability at various regions around the world by means of Köppen climate classification. For this purpose, monthly precipitation and temperature data were collected, which belong to each of the 31 subtypes of 5 basic climate types A, B, C, D and E. For each climate subtype, data were gathered from two stations, the first nearby the sea (distance up to 50 kms) and the other away from it. Data were collected from the Dutch Meteorological Institute website (http://climexp.knmi.nl). Only fully completed years were used for which there were both rainfall and temperature data. Köppen parameters were calculated and criteria which define the climate type in which each station belongs were implemented. Data availability of near and far from the sea stations showed that the coastal stations are being studied for a longer period of time. Basic statistical analysis of all data stations was conducted. Average, standard deviation, seasonality and variance of stations’ meteorological data were calculated. The difference of statistical parameters’ values between the near and far from the sea stations were analyzed. Then, statistical parameters of meteorological variables for all subcategories of each climate type A, B, C and D were correlated. Only those cases in which there was a high correlation were examined. Furthermore, changes in parameter values were commented. The temperature and precipitation seasonality between stations near and far from the sea for each climate category were compared. It was observed that temperature seasonality is more intense in type A coastal areas. In contrast, temperature and rainfall seasonality in B and D types appears greater in mainland areas. Daily rainfall data for the most common subtypes of 5 climate types were collected. These were Aw, Bwh, Cfa, Cfb, Csa and Dfc. Using the Hydrognomon program daily values were aggregated into monthly values. Ten stations from each of the six climate subtypes were examined. In turn, average, standard deviation, seasonality and variance of these monthly values were calculated. Through the Extremes evaluation selection of the hydrological data processing program ‘Hydrognomon’ maximum daily rainfall values were calculated. General Extreme Values distribution was fitted into these extreme value series. For the GEV adjustment two different methods were used. The first one was the method of moments and the other was the root mean square error minimization. The last one appeared to be the most effective. By the root mean square error minimization method it was also shown that for greater values of k shape parameter, GEV distribution fits better in daily maximum precipitation series (Figure 2). The characteristic ratio of daily maximum to the corresponding annual average rainfall amount of Aw, Bwh, Cfa, Cfb, Csa and Dfc stations was also examined. That quotient consists a hallmark of the extreme rainfall events’ intensity of each climate type. The ratios of the coastal and mainland areas were compared. It was observed that in the Aw, Bwh and Csa coastal stations there are more frequent extreme rainfall events. The opposite was observed for Cfb type, whereas the proximity to the sea factor appeared not to influence the occurrence of extreme rainfall events in Cfa and Dfc regions. The months in which maximum mean monthly rainfall occurs during the three climate periods 1900-1935, 1936-1970 and 1971-2005 were noted. Months were grouped into the four seasons of the year and the results were summarized for the whole stations of A, B C and D climate types. The year season in which maximum mean monthly precipitation occurs for all stations as well as between the coastal and mainland stations were compared. Finally, long-term climate variability of each climate subcategory’s station was investigated. Mean monthly rainfall and temperature change between the climate periods 1900-1935 and 1971-2005 was examined. The quantification of the climatic variability of the stations was determined by t-student distribution. The difference of average values statistical method was used for applying of t-student distribution into the sample data. The variation values of meteorological parameters was examined for significance levels of 1% and 0,001%. According to studies describing Hurst and Kolmogorov dynamics (long term persistence) 1/100.000 percentage is being considered as a statistically significant incidence of climate change in temperature and precipitation values. An increase in monthly average temperature of some climate subtypes was observed, mainly in coastal stations. Indicatively, a statistically significant (0,001%) increase in the temperature of the whole months of the year occurred in the Am tropical monsoon nearby the sea station.