Διερεύνηση της επίδρασης των δασικών πυρκαγιών στην υδρολογική απόκριση λεκανών απορροής της ανατολικής Αττικής

Abstract

Περίληψη Αντικείμενο της παρούσας μεταπτυχιακής εργασίας είναι η ανάπτυξη ενός ολοκληρωμένου πλαισίου εκτίμησης της μεταβολής της υδρολογικής απόκρισης των λεκανών απορροής μετά την επίδραση πυρκαγιάς. Η μελέτη πραγματοποιήθηκε στη λεκάνη απορροής της Ραφήνας, η οποία τοποθετείται ανατολικά του Νομού Αττικής και έχει έκταση 127 km2. Η περιοχή παρουσιάζει μεγάλο ενδιαφέρον ως προς την εκτίμηση της πλημμυρικής επικινδυνότητας λόγω της ραγδαίας αύξησης του πληθυσμού στους κύριους οικισμούς της λεκάνης. Επίσης, η περιοχή θεωρήθηκε κατάλληλη για την συγκεκριμένη έρευνα αφού διαθέτει μετρητικούς σταθμούς σε όλο το μήκος της. Τέλος η περιοχή επλήγη από το φαινόμενο της πυρκαγιάς τον Αύγουστο του 2009 δημιουργώντας με αυτόν τον τρόπο κατάλληλες συνθήκες για διερεύνηση των αλλαγών στο υδρολογικό καθεστώς της λεκάνης μετά την επίδραση πυρκαγιάς. Εκτός από τη διερεύνηση της πραγματικής πυρκαγιάς, για τις ανάγκες της παρούσας μελέτης διαμορφώθηκαν σενάρια πυρκαγιάς με τη χρήση του συστήματος διαχείρισης δασικών πυρκαγιών G.FMIS. Το σύστημα αυτό είναι το πρώτο ολοκληρωμένο σύστημα διαχείρισης δασικών πυρκαγιών στην Ελλάδα και διαθέτει μοντέλα καύσιμης ύλης που αντιπροσωπεύουν τις συνθήκες των μεσογειακών οικοσυστημάτων. Το σύστημα αυτό έχει επιχειρησιακό χαρακτήρα και χρήστης του συγκεκριμένου λογισμικού είναι και η Πυροσβεστική Υπηρεσία Ελλάδος. Επιπλέον στο πλαίσιο της εργασίας αναπτύχθηκε μεθοδολογία επαναπροσδιορισμού των υδρολογικών παραμέτρων (αριθμού καμπύλης, αρχικής κατακράτησης και χρόνου υστέρησης) μετά την επίδραση πυρκαγιάς σύμφωνα με αντίστοιχα αποτελέσματα μελετών στη διεθνή βιβλιογραφία αλλά και συμπεριλαμβάνοντας τις συνθήκες που επικρατούν στην συγκεκριμένη περιοχή μελέτης. Τα αποτελέσματα της μεθοδολογίας που αναπτύχθηκε εισήχθησαν ως δεδομένου εισόδου στο υδρολογικό μοντέλο HEC-HMS, με την βοήθεια του οποίου προσομοιώθηκε η απορροή γεγονότων βροχόπτωσης διαφορετικών χαρακτηριστικών. Από τη γενική διερεύνηση των αποτελεσμάτων και ειδικότερα από τη σύγκριση των παρατηρημένων και εκτιμημένων υδρογραφημάτων στην περιοχή μελέτης, για συνθήκες μετά την πυρκαγιά, αποδεικνύεται η αποτελεσματικότητα και ορθότητα της μεθοδολογίας που αναπτύχθηκε στην παρούσα εργασία.

Extended Abstract Over the past few years, Mediterranean countries including Greece suffer from frequent wildfire events. This has led to increased awareness not only in terms of the effects of forest fires on vegetation, but also in order to prevent possible loss of life and property, changes in the hydrological response of river basins and increase in erosion/sedimentation process at the once vegetated areas. Moreover, according to the European Flood Directive (EU 2007/60), published on 26/11/2007, which intends to reduce the risks and adverse consequences of floods in European Union and is a complement to EU legislation for integrated water resources management (EU WFD 2000/60), it is very important to develop an integrated framework for assessing the change in the hydrological response in watersheds. More specifically, the Directive aims to establish a framework for evaluation and management of risks associated with floods in particular on human health and life, the environment, cultural heritage, economic activity and infrastructure. The purpose of this thesis is to investigate the change of the hydrological response in watersheds following a wildfire event and develop a methodological framework that will support the redefinition of the hydrological parameters after the influence of the wildfire. For that reason, a systematic literature review has been accomplished in order to investigate all aspects related with the way the wildfires affect the soil properties, the vegetation cover and in general the hydrological response. In addition to that, the results about the increase of runoff volume and peak discharges of similar studies are included in the thesis. Specifically, the direct and indirect effects of wildfires on hydrological response are mentioned below: a. changes in the physical and chemical soil properties: organic ground cover is converted to soluble ash and gives rise to phenomena such as water repellency. During combustion hydrophobic organic compounds in litter and topsoil are volatized and released upwards to the atmosphere and downwards into the soil profile along a temperature gradient. Downward translocated hydrophobic compounds condense on cooler soil particles at or below the soil surface forming water repellent conditions. b. runoff and erosion rates increase: the formation of the water repellency at the soil acts as a barrier to water infiltration and as a consequence the runoff volume and peak discharges increase. Furthermore, there seems to be an increase to the sediment yield. The Rafina catchment has been selected as the case study of the current research on the way the hydrological response changed. The area is located in the eastern region of the Prefecture of Attica, covers 127.17 km2 and includes different land uses. This area has been chosen for the following reasons: • the catchment is considered to be suitable for this research since it had been monitored during the last years for academic and research purposes and therefore exist reliable pre-fire data records • a significant increase in the population of the area at the main settlements has been recorded. • according to several studies the area is vulnerable to erosion at great extent in case of forest destruction. In the framework of this thesis the Geographic Fire Management Information System – G.FMIS– has been chosen for the fire simulations of several scenarios. It is considered appropriate because it provides fuel models that consist samples of mediterrenean ecosystems. In particular, three scenarios have been considered changing the wind and fuel moisture, thus affecting in a different way the catchement. A real fire case which occurred in the area in August 2009 has been studied, as well. After the study of the fire a methodology is developed to assess the changes in the hydrological parameters: curve number, initial abstraction and lag time. The values of the curve numbers were defined in proportion with the land uses and hydrology soil group, in accordance with Goodrich’s results. The initial abstraction was defined according to the curve numbers’ values that resulted from the methodology above using the SCS method. The value of lag time after the fire was defined based on literature review and was reduced by 40% compared with its pre-fire values (Elliot, 2005, Cydzik, 2010) Finally, the semi-distributed hydrological model HEC-HMS has been used to assess the hydrological response of the catchment for the different scenarios. The catchement was divided in five subbasins, each one with different pre-fire and post-fire values of hydrological parameters. Conclusions The main conclusions of the thesis: From the Geographic Fire Management Information System (G.FMIS): • The burnt area grows simultaneously with the speed of the wind and at the same time the humidity of fuels reduces. The first factor seems to affect most the final results. • An uneven expansion of the fire in different parts of the watershed is observed, which results mostly from the variety of fuels. • Unevenness in the shape of the burnt areas is also observed. The main reason of this result is that the model should not be applied in urban areas • The final percentage of the burnt area for the scenarios 1,2,3 comes up to 4%, 23% and 48% respectively. From the study of the real fire of 2009: • According to the digitization of the burnt area of the real fire of 2009 from satellite image, the fire affected 35.000 acres, that is 28% of the total watershed. • The event affected the forest lands by 20,2%, while the percentages of the affected urban and cultivable area come up to 5,0% and 2,4% respectively. From the methodology of redefining hydrological parameters: • The values of the CN change proportionately to the surface of the area affected by each fire. More specifically, the subbasins 2 and 3, which actually are the experimental basin (X-Basin), appear to have a significant difference in the value of CN in the fire of 2009, with percentage changes 137% and 73% respectively. Remarkable changes of CN are shown in the subbasin 4 in the second and third fire scenarios, with percentage changes 34% and 56% respectively. • Due to the extremely limited area that is affected by the first fire scenario, the slightest differentiation in the values of CN are observed in this case for all subbasins. • The value of initial abstraction is significantly reduced in the subbasins where a significant CN increase was estimated. The reason of this reduction is the inversely proportional relation of the initial abstraction with the CN. • The lag time which was estimated according to literature review tends to represent in more detail the hydrological conditions of the experimental basin. From the hydrological model HEC-HMS: • A significant influence of the surface of the burnt area on its hydrological response both for the subbasins and the entire catchement can be observed form the simulated hydrographs. • It has also been detected that as the surface of the burnt area increases, the peak discharge in the subbasins’ outputs increases as well • The smaller the subbasin surface is, the more the peak discharge is increased From the correlation between the burnt area and peak discharges for fires which were studies: • A linear correlation between the burnt area and the peak discharges is also detected, with a high correlation coefficient for all the storm events which were studied. • The most important characteristic of the rainfall is its depth, which increased, it increases the peak discharges as well. • As the percetange of the burnt area increases, the variation of peak discharge before and after the fire, increases as well. In particular, according to the redefining method for the two parameters (CN, Iα) for the first scenario the values of the peak discharge vary up to 5%, while for the second, third scenario and the real fire of 2009, the variation is estimated up to 32%, 100% and 44% respectively. As far as the redefining method for the three parameters (CN, Iα, tp) is concerned, the variation is estimated up to 6%, 37%, 122% and 55% respectively. • As the burnt area percentage increases, the range of variation increases as well. Particularly, according to the redefining method for the two parameters (CN, Iα) the range in the first scenario rises up to 0,8%, in the second 8,6%, while it increases up to 23,5% and up to 17,5% for the third scenario and the real fire of 2009. As far as the redefining method for the three parameters (CN, Iα, tp) the range rises up to 1,5%, 11%, 32% and 23,7% respectively. From the comparison between simulated and observed hydrographs after the fire of 2009: • The results indicate the necessity of redefining the model parameters for the simulation of the conditions after the fire event, because the simulation under the existing conditions before the fire, for the same storm event, underestimates significantly the hydrological response of the watershed. • Generally, both approaches overestimate the runoff volume, but at the same time they approach in a better way the peak discharge, with the approach- estimated(-40%)- in which the three parameters (CN, Iα, tp) are redefining and they give better result almost in every observation point • In terms of flood risk the most important observation point is the Fladar point (before Rafina’s settlement). From the comparison of the hydrographs for a specific storm event, both in peak discharge and runoff volume of the observed hydrograph are approached well with the redefining method of the two parameters (CN, Iα). To sum up, according to the previous conclusions, radical change in the hydrological response of the basin after the fire event is observed. Finally, the effectiveness and the validity of the methodology developed and presented in the present thesis are proved and a possible expansion of their application is proposed.