Supervisors info:
Εμμανουήλ Ανδρεαδάκης, ΕΤΕΠ, Τμήμα Γεωλογίας και Γεωπεριβάλλοντος, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Summary:
The subject analyzed in the present thesis is the series of wildfires that occurred in Neos Voutzas and Mati, Attica in Greece during the summer of 2018. This big disaster, which is the second deadliest wildfire event to have occurred worldwide in the 21st century, took place in a Wildland Urban Interface (WUI) where urban areas and forestlands interdigitate . The impact of this disaster was huge in terms of human casualties (100 fatalities and 172 non – fatal injuries) as well as property damage (thousand of cars and houses were damaged or destroyed). The circumstances under which the event occurred and unfolded are described in detail and the preliminary results of the initial investigation in the area are included as well as photographic material depicting the impact of the disaster. Furthermore, a retrospection of previous big WUI wildfire events in Greece, Portugal, Russia, USA and Israel is carried out, as well as a report of the development of WUI wildfire events in Greece from 2000 to 2009. The goal of this study is to assess and compare the presumed wildfire vulnerability of the buildings in these areas to the final impact on the buildings as well as analyze and explain the results in order to determine the efficiency of this vulnerability calculation method so that it can potentially be utilized in the future for fire risk assessment analyses in different areas in Greece, better decision making, better planning and ultimately minimization of the wildfire impact in terms of casualties and property loss. The method we followed was to firstly carry out a building disaster risk analysis in order to calculate the scores A (Wildfire Features Risk Score), T (Building Vulnerability Score) and K (Total Building Disaster Risk Score) based on features such as a) the kind and density of tall trees and plants in the area around the buildings, b) the distance between the tall trees, the plants and the building, c) the accessibility of the building and d) the fire resistance of the walls, roof and windows of the building. The next step was to assess and analyze the wildfire impact to the buildings and the surrounding vegetation using the photographic material from the autopsy that was conducted by the team of the National and Kapodistrian University of Athens instantly after the disaster as well as videos and photos from drone flights above the area. Using the Geographic Information Systems (G.I.S.) software ArcMap 10.1 we constructed fifteen (15) building wildfire vulnerability maps, twelve (12) for each vulnerability factor individually and three (3) for every vulnerability score A, T and K individually. Furthermore, three (3) post – wildfire impact maps were constructed depicting the state and the damage on the buildings as well as the state of the burnt vegetation in the nearby area of the building. Combining the aforementioned maps we created six (6) kernel density maps showcasing the vulnerability attributes A, T, K individually as background in contrast with distinct points on each building to represent the impact on the building (state and damage) and the impact on the vegetation in the surrounding area of the building. Finally, using the Microsoft Excel software we created three (3) graphs illustrating the vulnerability scores K – A, K – T and K – State of the building (using a number for each individual state). Observing and analyzing the kernel density maps and the Microsoft Excel Graphs we come to the conclusion that: 1) there is not absolute correlation between the Total Building Disaster Risk Score (K) with the damage done to the building 2) the correlation between the Total Building Disaster Risk Score (K) and the state of the building seems pretty decent 3) the indicator Α (Wildfire Features Risk Score) is not totally correlated to either building damage or building state 4) the T indicator (Building Vulnerability Score) is very well correlated to the state of the building 5) buildings with very high scores of vulnerability seem to affect and raise the vulnerability of the nearby buildings, even if their initial vulnerability is low. This fact lead us to the conclusion that the vulnerability of the nearby buildings should affect the way we calculate the vulnerability of each individual building. 6) The buildings in the Mati area suffered more serious damage than the ones in Neos Voutzas due to the generally poorer building material quality which is also highlighted through the higher values in the T indicator scores 7) the Excel Graphs indicate that the values are more evenly spread in K – A graphs than in the K – T graphs so we conclude that the indicator A has a bigger effect than the indicator T to the total vulnerability K 8) as far as Neos Voutzas is concerned we noticed that the R2 value in the graph K – A is approximately 1 while in K – T is close to 0. This is because of the steep incline in Neos Voutzas area which greatly affects the values of A and consequently, those of the K indicator, while the values of T are generally low in the area. 9) The graph illustrating the relation between K value and the state of the building indicates that although there is a decent number of buildings with low vulnerability that were burned, there are no buildings with vulnerability scores over 30 unaffected by the fire. This could be important because it leads us to the preliminary conclusion that we should try to reduce the values of the building vulnerability to at least below 30. Overall, we can safely assume that the method needs some additions in order to be more accurate but in general terms the unpredictability of the WUI wildfires due to the countless factors that affect them make it nearly impossible to design a fully accurate method of calculating the vulnerability in order to nullify the wildfire impacts. In the final part of the thesis, we propose actions addressed to the home owners and the state in order to reduce the overall building vulnerability while analyzing the vagueness in Greek Law and the difficulties in taking actions and necessary measures.
Keywords:
Wildfires, Mati, Neos Voutzas, Attica, Greece, Geographic Information Systems, Vulnerability, Impact, Mixed forest – urban zones, Natural Disasters, Environment, Wildfire Risk
