Revision of Fuzzified Fine-Kinney Method, an Adaptive Method for Natural Disaster Risk Management

Introduction

Turkey is vulnerable to multiple natural disasters because of its geological location. There are three major faults in Turkey, the Northern Anatolian Fault, the Western Anatolian Fault, and the Eastern Anatolian Fault. Because these faults are seismically very active and affect quite a large area, making 42% of Turkey a first-degree earthquake zone and 24% second-degree (Türkoğlu 2001). Turkey is also surrounded by the Mediterranean Sea, Black Sea, and Marmara Sea while hosting lots of streams and basins and having high rainfall in some regions, and therefore being vulnerable to floods. Turkey is also vulnerable to landslides, avalanches, and forest fires. Even though the vulnerability to these natural disasters except earthquakes is more area-specific, their impacts generally overlap with earthquake zones making a considerable amount of Turkey liable to multiple hazards.

Governments prepare risk maps (AFAD 2021), uniform non-governmental organizations (AFAD, T.C. İç İşleri Bakanlığı Afet Ve Acil Durum Yönetimi Başkanlığı), and manage disasters through policies, funds, and insurance (DASK, Doğal Afet Sigortaları Kurumu). They also set regulations as a precaution (Resmi Gazete 2007 as cited in DEPREM BÖLGELERİNDE YAPILACAK BİNALAR HAKKINDA YÖNETMELİK ). Municipalities apply every policy determined by governments and carry out local activities (Nilüfer İlçe Afet ve Acil Durum Yönetim Merkezi) to prevent disaster losses.

For some countries, even though they have plans and precautions for natural disaster situations, these may not be sufficient to prevent disaster losses. The purpose of the present research is to modify a risk-assessment method called “Fine-Kinney” which is used for industrial accidents. The main goal of the modification is to make the method applicable, easy to use, and reliable in natural disaster risk assessment for architects and civil engineers. The Fine-Kinney method is a risk assessment tool that needs very basic data to evaluate the risk level. By taking Fine-Kinney as a base and keeping its advantages, this method will be favorable not only for regions that have insufficient or no data on natural disasters but also for local and rural buildings, which may be constructed in the future.

In this research, the default version of Fine-Kinney is reviewed through each parameter in turn. It turns out that the frequency parameter of the default Fine-Kinney prevents results from being accurate when the method is applied to natural disasters. To solve this problem, the frequency parameter is adapted for natural disasters, taking a building’s industrial lifespan as a baseline for creating a new frequency parameter with new multipliers. To evaluate the method’s reliability and accuracy in natural disaster environments, an adapted version of Fine-Kinney was tested, along with its “fuzzy” version, which is created in MATLAB, a proprietary multi-paradigm programming language and numeric computing environment developed by MathWorks.

For the test, we assess the vulnerability against multiple disasters (avalanches, rockfall, landslide, earthquake, flood) in Karaköy Village in the Mustafakemalpaşa District (Bursa, Turkey). To evaluate frequency, severity, and probability parameters and to see the clarity of the results, we used risk maps and disaster records provided by the Disaster and Emergency Management Presidency (Aka. AFAD, a governmental disaster management agency operating under the Turkish Ministry of Interior). After comparison, it is shown that both methods produce results similar to the risk maps, but because the fuzzy results are more soft-edged compared to the standard Fine-Kinney, the Fuzzy Fine-Kinney results can be considered more applicable.

Literature Review

Researchers have used several methods to assess risks and vulnerability in different situations, including natural disasters (Du & Lin 2012). There are methods based on risk uncertainty, which include probability and statistics (Deng et al. 2001), the fuzzy mathematical method, which was introduced to natural disaster risk analysis by Huang et al. (2018), the gray system method, which is useful for minor data and poor information, used in meteorological disaster risk analysis by Gong & Forrest (2014), and the comentropy method, which is a way of describing “uncertainty” for information sources, was adapted to soil erosion and natural disasters by Ai (1987).

The literature also includes methods based on risk disadvantage. Qualitative risk analysis is a method where experience and theoretical knowledge are used. This method was first used for disaster risk and vulnerability by Eldeen (1980). GIS (geographic information system) is a series of computer-spatial analysis tools that can analyze information to create map data, and these methods were used in the risk analysis of geological disasters by Zhu et al. (2002). Modeling is another effective method for quantifying disaster risk (Peduzzi et al. 2009). Analysis methods are based on risk future, and they consider future hazard effects too. Risk characterization is a method that presents disaster risk results in the form of risk mapping or risk curves or risk index construction (Ma 2015). Ensemble forecast technology is based on weather forecasting and was used to forecast water-based disasters by Deng et al. (2006). Scenario analysis is a method that was first used in the military, and it tries to predict the future by finding and analyzing possible scenarios. Ranger et al. (2010) evaluated future flood risk for Mumbai with this method.

Some of these methodologies have been improved with the help of fuzzy logic systems (Pamučar et al. 2016), while some research has relied solely on fuzzy logic to assess risk (Karimi & Hüllermeier 2007).

Few studies adapted the Fine-Kinney risk assessment method for natural disasters (Derse 2021; Işık et al. 2022), which was originally a method for assessing industrial accidents. Some research has applied fuzzy logic on Fine-Kinney for dam risks (Daneshvar Rouyendegh & Gür 2020), which include several natural disasters.

When these methods are implemented in an environment at the scale of a building, settlement, or region for a natural disaster, however, the frequency parameter starts to cause inaccuracy in results. For example, in Daneshvar  Rouyendegh & Gür’s (2020) method, the lowest multiplier of the frequency parameter is 0.5, corresponding to “very sparse” and is defined as “once in a year”. Defining the lowest multiplier as “once in a year” can cause problems because earthquakes occur at multiple magnitudes and severities. Even though bigger and more destructive earthquakes usually happen less than once a year in a region, “big and destructive” or “major” is a very broad definition of an earthquake. All earthquakes at a Richter scale value of 5 or higher can be considered major. But if two earthquakes with a Richter scale of 5 and 8 respectively are multiplied with the same frequency multiplier, their risk scores can come out close and this can provide a false sense of security. Derse (2021) found that the least frequent multiplier is defined as “very rare” and “yearly”. However, because evaluations are not limited to a region, the quantity and frequency of natural disaster is greatly increased.  Derse (2021) showed that the size of the measured area can vastly change the reliability of the frequency scale. For example, while there have been only 34 earthquakes with a magnitude higher than 6 in the last 600 years in the Marmara Region (AFAD 2021), an average of 128 earthquakes occur around the world each year with a magnitude higher than 6.  

For these reasons and the lack of research on the adaptation of the Fuzzy Fine-Kinney method to a pure building-oriented environment for natural disasters in a building, settlement, or region scale, the need for a new method was established.

Method

Fine-Kinney is a quantitative risk assessment method that is derived from a safety system (Department of Defense Standard Practıce) whose purpose is to minimize and eliminate hazards in the work field. This method was developed by Wiruth and Kinney (1976). In this method, each hazard is associated with three parameters (likelihood, exposure, and possible consequences), and the “Risk Score” is calculated by multiplying these three parameters (Kokangül et al. 2017) (Kinney & Wiruth 1976).

Equation of risk score in Fine-Kinney

“Risk Score = Possibility x Frequency x Severity”

According to the risk score, the risk is evaluated at five levels. These are acceptable risk, possible risk, substantial risk, high risk, and very high risk. Acceptable risk demands no action, possible risk needs to be taken into account and must be corrected in the long term, substantial risk needs correction in the mid-term, high risk needs to be corrected as soon as possible, and very high risk demands the stoppage of the process until the correction is made (Oturakçı 2017).

Probability: In the original study by Kinney and Wiruth (1976), the probability scale includes ten divisions, and “might well be expected” which is used for events that happened before and can reoccur in the future, taken as a reference point and given the value of 10. “Only remotely possible” is given the value of 1 and “virtually impossible” is given the value of 0.1. The scale is completed with intermediate values given depending on experience, as displayed in Table 1.

Frequency: Kinney and Wiruth (1976) used a scale of ten for frequency, and the reference points are 1 to 10. In the scale, risks are rated by their occurrence frequency. The value “10” is given to events that happened on an hourly basis and are considered “continuous”. The reference point “1” is given to events that happen a few times per year and are considered “rare”. Intermediate values are given according to experience (Table 2).

Severity: Severity is evaluated based on casualties and the damage they cause in dollars ($). The scale (Table 3) sets the loss of life and property against a scale of 100. The endpoints are “catastrophe” with a value of “100” and “noticeable” with a value of “1”.

Risk Score: When these three parameters are multiplied, the risk score is determined (Table 4).

Problems with Fine-Kinney: When the Fine-Kinney method is applied to natural disasters without any changes, several errors can be expected. The probability parameter is rather qualitative and can be evaluated with common knowledge when it comes to natural disasters. Severity requires only a minor change: economic damage depends on the scale of the environmental impact of the disaster, but not every disaster loss is calculated as “tangible damage” in developing or underdeveloped countries. Environmental impact is a scale that is far easier to measure than cost and can be made in the field. The parameter frequency is where a major error occurs because natural disasters are unpredictable, and their frequency is very different from industrial accidents. For example, if we take a destructive earthquake as an example, which happens once every few years or more often, its frequency multiplier will be 0.5. These earthquakes can generate a disaster and are quite possible, which makes their severity and possibility score 40 and 10 respectively. This makes a total risk score of 120 (40 x 0.5 x 6 = 120) which means they are a substantial risk and will be very inaccurate. Destructive earthquakes can easily be high risk and can go as far as a very high risk depending on conditions. Depending on the result of these calculations, a revision of the frequency scale becomes mandatory. In the frequency scale, a building’s rough lifespan, assumed to be 50 years, is taken as a reference point, and given the value of “1,” which means a building will definitely face a disaster with a frequency value of “1”. Another reference point set for value “10” is “more than once in a year” to evaluate frequent hazards like earthquakes. Therefore, intermediate values are given to generate a revised scale (Tables 5, 6).

Fuzzy Fine-Kinney: To prevent uncertainties in the evaluation of parameters, frequency, possibility, and severity multipliers are converted to fuzzy numbers and a new set of rules has been set. As shown in Figure 1, frequency, severity, and possibility are set as input, while risk is set as output. Here, we use the Mamdani inference process method and is coded in MATLAB, a proprietary multi-paradigm programming language and numeric computing environment developed by MathWorks, Fuzzy Logic Designer (Gul & Celik 2018).

Figure 1. Flowchart of the method.

The membership functions of probability, frequency, and severity are given in Tables 7, 8, and 9 respectively. When determining this membership function, adjacent values of each value have been used. For example, in the fuzzification process of “Once in 25 years”, “Once in 50 years” and “Once in a decade” values are added, and the membership function is set as (1,2,3). Fuzzification values are given in Tables 7–10.

Figure 2. Fuzzy fine-kinney design.

Figure 3. Fuzzy diagram of Probability.

Figure 4. Fuzzy diagram of frequency.

Figure 5. Fuzzy diagram of severity.

Every member’s membership function is a domain starting with the value of the member below and ending with the value of the member above (Figs. 2–5; Tables 7–10). The first member’s domain starts with its value and the last member’s domain ends with its own because they have no other member below and above respectively.

In this study, both Adapted Fuzzy Fine Kinney and Adapted Fine Kinney methods are applied to Karaköy Village in Mustafakemalpaşa/Bursa district. Bursa is one of the most developed cities in Turkey. Bursa is very important for Turkey both economically and historically, being one of the biggest industrial cities of Turkey and the first capital of the Ottoman Empire. These assets make Bursa suitable for the case study.

Results for both methods are evaluated individually and comparatively, to see if they are reliable for natural disaster risk assessment. The reasons behind choosing Karaköy Village for this case study were the exposure of Karaköy to multiple natural disasters and the existence of casualty and date data of five disasters provided by AFAD, a national organization in Turkey that carries out the necessary studies for the effective management of disaster and emergency processes while ensuring coordination between relevant institutions and organizations and to producing policies in this field. Locations of Mustafakemalpaşa, Karaköy and Bursa are shown in Figure 6.

Figure 6. Fuzzy diagram of risk score

To make the analysis more specific and pinpoint the results, Karaköy quarter in Mustafakemalpaşa was selected. The reason for choosing this quarter is because Karaköy has been exposed to several landslides in recent years and it caused disturbed daily life and damaged many buildings in the region, which makes these landslides a disaster. Its geolocation makes Karaköy vulnerable to earthquakes which makes it a multihazard area. Even though the region is not vulnerable to all of them, all five disasters (avalanche, rockfall, landslide, earthquake, and flood) risk scores will be evaluated for authentication of the method. After risk scores are calculated they will be compared to risk maps produced by AFAD.

Avalanche: There isn’t any snow mass to cause an avalanche at any time of year near the Karaköy. That makes the probability of an avalanche disaster next to impossible which is equal to a multiplier of “0.1”. There are no records of avalanches in Karaköy which makes frequency “more than once in a century” with a multiplier of “0.5”. Because there is no significant snow mass, even if an avalanche occurs, it will hardly cause any damage. And that makes the severity “barely miss, no environmental impact” with a score of “1”.

Flood: As with avalanches, no flood disasters happened at Karaköy in recent history. So, we can use the same multipliers we used for avalanches.

Rockfall: As seen in Table 11, there is only one rockfall event recorded in M. Kemalpaşa which was not in Karaköy. And this means the same multipliers as avalanches can be used again. In both rockfall and flood situations, possibility and severity can be changed depending on conditions. However, if there is no drastic change in either parameter, as the multipliers for other parameters are at a minimum, the Risk Score will still be very low.

Landslide: Even though it was stated only once in AFAD’s records, according to local journals (Milliyet 2015, Ensonhaber 2015, and Dostmedya n.d 2015 ), Karaköy seems to be exposed to landslides quite often because of groundwater under the soil and it disturbs local people’s daily life while damaging several buildings. This turns these landslides into disasters. Groundwater under the soil means that landslides in Karaköy can be considered “Quite possible.” This makes the possibility multiplier “6”. Local journals state that Karaköy was exposed to landslides in winter multiple times. This makes the frequency “more than once in a year” with a multiplier of “10”.  Disturbing daily life and damaging buildings to the extent that making them “unusable” can be considered as “considerable environmental impact” and that makes the severity multiplier “15”.

Earthquake: When it comes to earthquakes, they have a wide variation in frequency, possibility, and severity, and they cannot be specified for districts because of their very wide effect range. Considering these facts, only significant earthquakes in Marmara’s history (>Magnitude 6) have been evaluated. Recorded seismic activities which have a magnitude greater than 6 in Anatolia’s history since the 15^th century are shown in Table 13. There were 34 major earthquakes in Marmara’s 600-year history, which can be considered as “quite possible” with a multiplier of “6”. If we divide 34 earthquakes by 600 years, the outcome is 17.6 which is closest to the term “once in 25 years”. This makes the frequency multiplier “2”. Earthquakes bigger than Magnitude 6 often result in disasters in Turkey. That makes severity “disaster” with a multiplier of “40”.

Results

The results from Mustafakemalpaşa’s risk analysis show that both methods produce quite reliable and accurate results when compared with risk maps (Tables 14, 15).

Discussion

According to the results of both methods, Mustafakemalpaşa is vulnerable to earthquakes and landslides but with virtually no risk against rockfall, flood, and avalanche. When results are superimposed on risk maps, both results seem to be very accurate. However, the difference between the two methods can be seen in their maximum and minimum values. With the standard Fine-Kinney method, landslides and earthquakes possess such a high threat that stoppage of other actions is recommended. The Fuzzy Fine-Kinney method identifies these disasters as high risk, so they still need immediate precautions but daily life does not need to be suspended. For avalanches, rockfalls, and floods, according to the standard Fine-Kinney, they possess practically no threat and therefore require no precaution. However, when parameters are fuzzified, these disasters are identified as not associated with substantial threat but should be considered nonetheless.

Figure 7. Locations: A) Mustafakemalpaşa (Turkiyebilgi n.d.). B) Bursa (Pekel S 2010). C) Karaköy (highlighted with red, reteived from Turkiyebiligi).

Figure 8. Risk maps. A) Bursa Landslide Risk Map (AFAD n.d.), Green-Low Risk, Red-High Risk. B) Bursa Rockfall Risk Map (AFAD n.d.), Yellow-Low Risk, Green-High Risk. C) Bursa Avalanche Risk Map (AFAD n.d.), Purple-Low Risk, Red-High Risk. D) Bursa Earthquake Risk Map (AFAD n.d.), Blue-Low Risk, Brown-High Risk (Afet Haritaları.2023).

Figure 9. Mustafakemalpaşa Flood Risk Map (AFAD n.d.), Light Blue-Low Risk, Dark Blue-High Risk

The standard Fine-Kinney has sharper edges than the Fuzzy Fine-Kinney when it comes to very low and very high risks (Figs. 7, 8; Table 16). However, “taking precautions without disturbing daily life” and “it should be on consideration even though the risk is very low” can be considered more rational and safe than “suspending daily life to take measures” and “no need to take any precaution” in disaster situations.

It is important to note that, when using these methods, extreme results should be evaluated by architects and civil engineers. For example, in a region that is exposed to earthquakes, if the risk score is very high, this does not mean that there should not be any settling at all in the region. The very high-risk score should be interpreted as “If there are any existing buildings which are at very high risk, precautions should be taken in that manner, and if there will be a new settlement in the area, the earthquake must be one of the biggest considerations in the design process.”. This highlights the need to recall that this method is only an evaluation tool for the assessment of the risks of natural disasters, and with the results of the method, civil engineers’/architects’ knowledge, scientific facts, and local regulations must be taken into consideration too. For example, in a tropical climate, even if the fuzzy risk score is 30 for an avalanche, which is the result of minimum parameters, even though the method recommends considering the disaster, taking precautions for an avalanche in a zone where there is no snow will be a waste of both money and time.

 Conclusion

We aimed to produce a fast and reliable natural disaster risk-assessment method for underdeveloped and developing countries that could produce rather precise results even with limited data. For this reason, the severity parameter’s economic loss aspect has been replaced with the environmental damage scale. Another problem was the drastic gap between the periods of recurrence of industrial accidents and natural disasters. To adapt the frequency parameter to natural disasters, “once in 50 years” has been taken as a reference point because a building’s life span is considered as roughly 50 years. Another decision was to fuzzify the parameters, which allowed us to set our inputs not as definitive numbers but as a range and compensate for the unpredictability and diversity of natural disasters.

The new method made results less sharp but more accurate. While landslides and earthquakes are “very high risk” in Standard Fine-Kinney they become “high risk” in Fuzzy Fine-Kinney which means their risk score decreased, while avalanche, flood, and rockfalls risk scores increased and they became “possible risks”. This means in the Standard Fine-Kinney avalanche, rockfall, and flood risks were so insignificant that they will probably be ignored, and earthquakes and landslides had such a risk score that they required stoppage of current activity. In the Fuzzy Fine-Kinney model, on the other hand, avalanches, rockfalls and floods can be ignored, and earthquakes and landslides are still high risks but there is no need to stop any events. When the two methods’ results are compared with the risk maps of the region, they both produce accurate results, but the Fuzzy Fine-Kinney results are more rational and reliable.

Conflict of Interest

The authors declare that they have no competing interests. The view expressed in this short paper does not necessarily reflect those of their institutions

Tables

Table 1. Fine-Kinney Probability Scale.

Table 2. Fine-Kinney Frequency Scale.

Table 3. Fine-Kinney Severity Scale.

Table 4. Fine-Kinney Risk Score Scale.

Table 5. Revised Fine-Kinney Severity Scale.

Table 6. Revised Fine-Kinney Frequency Scale.

Table 7. Fuzzy Fine-Kinney Probability Scale.

Table 8. Fuzzy Fine-Kinney Frequency Scale.

Table 9. Fuzzy Fine-Kinney Severity Scale.

Table 10. Fuzzy Fine-Kinney Risk Score Scale.

Table 11. AFAD’s rockfall event records for Bursa (AFAD, 2021).

Table 12. AFAD’s Landslide records for M. Kemalpaşa (AFAD, 2021).

Table 13. AFAD’s record for major earthquakes in the Marmara region (AFAD, 2021).

Table 14. Fine Kinney Risk Scores.

Table 15. Fuzzy Fine Kinney Risk Scores.

Table 16. Comparison among Risk Maps and Risk Scores of Fuzzy and Standard Fine-Kinney.

References

AFAD Bursa İl Afet ve Acil Durum Müdürlüğü İl Afet Risk Azaltma Planı (IRAP) (2021) Retrieved March 2, 2023, from https://bursa.afad.gov.tr/kurumlar/bursa.afad/Bursa-IRAP.pdf

 Afet Haritaları. (“Risk Map” in English) (n.d.). Retrieved February 5, 2023, from https://www.afad.gov.tr/afet-haritalari

Ai NS (1987). Comentropy in erosional drainage-system. Journal of Soil and Water Conservation. 1(2): 1–7.

 Bursa. (n.d.) Retrieved February 10, 2023 from http://turkiyebilgi.byethost8.com/mahalle_yerlesim/marmara/bursa/bursa.htm?i=1

 Daneshvar Rouyendegh B  & Gür L (2020). Bulanık Fine–Kinney Yöntemiyle risk Değerlendirmesi Uygulaması. Journal of Industrial Engineering (Turkish Chamber of Mechanical Engineers). 31(1): 75–86.

 Deng G, Chen, HL, Zhou YS (2006). Application of ensemble forecast technique to rainstorm risk analysis. Ziran Zaihai Xuebao/ Journal of Natural Disasters. 15(1): 115–122.

 Deng G, Wang AS, Li SK, Zhou YS (2001). Risk theory and method and its initial application in grain yield. Journal of Natural Resources. 16(3): 221–226. https://doi.org/10.11849/zrzyxb.2001.03.005

 Deprem Bölgelerinde Yapilacak Binalar Hakkinda Yönetmelik. (2007, March 6) Resmi Gazete, 1-159.

 Derse O (2021). A new approach to the Fine Kinney method with AHP-based ELECTRE I and math model on risk assessment for natural disasters. Journal of Geography. 42: 155–164. https://doi.org/10.26650/JGEOG2021–875427

 Du X, Lin X (2012). Conceptual model on regional natural disaster risk assessment. Procedia Engineering. 45: 96–100. https://doi.org/10.1016/j.proeng.2012.08.127

 Eldeen MT (1980). Predisaster physical planning: Integration of disaster risk analysis into physical planning: A case study in Tunisia. Disasters. 4(2): 211–222.

https://doi.org/10.1111/j.1467–7717.1980.tb00274.x

 Gong Z & Forrest JYL (2014). Special issue on meteorological disaster risk analysis and assessment: on basis of grey systems theory. Natural Hazards. 71(2): 995–1000. https://doi.org/10.1007/s11069–013–0864–y

 Gul M, Celik E (2018). Fuzzy rule-based Fine–Kinney risk assessment approach for rail transportation systems. Human and Ecological Risk Assessment: An International Journal. 24(7): 1786-1812.

 Heyelan bir köyü yok ediyor. (2015, April 13). ensonhaber. https://www.ensonhaber.com/emlak/heyelan-bir-koyu-yok-ediyor-2015-04-13

Huang CF, Liu XL, Zhou GX, Li XJ (1998). Agricultural natural disaster risk assessment method according to the historic disaster data. Journal of Natural Disasters. 7(2): 1–9.

IRIS (2011). How Often Do Earthquakes Occur? Education and Outreach Series No. 3 https://www.iris.edu/hq/inclass/fact-sheet/how_often_do_earthquakes_occur

 Işık E, Aydın B, Yunus LE, Mehmet CA (2022). Determination of rockfall risk using Fine–Kinney methodology for Bitlis downtown area, eastern Türkiye. Disaster Science and Engineering. 8(2): 1–17. https://dergipark.org.tr/en/pub/dse/issue/74703/1129764

 Karaköy’de heyelan korkusu. (n.d.) (2015). Dostmedya. https://www.dostmedya.com/karakoyde-heyelan-korkusu

 Karimi I , Hüllermeier E (2007). Risk assessment system of natural hazards: A new approach based on fuzzy probability. Fuzzy Sets and Systems. 158(9): 987–999. https://doi.org/10.1016/j.fss.2006.12.013

  Kinney GF,Wiruth AD (1976). Practical risk analysis for safety management. Naval Weapons Center China Lake CA. https://apps.dtic.mil/sti/pdfs/ADA027189.pdf

 Kokangül A, Polat U, Dağsuyu C (2017). A new approximation for risk assessment using the AHP and Fine Kinney methodologies. Safety Science. 91:24–32. https://doi.org/10.1016/j.ssci.2016.07.015

Ma BC (2015). Definition and expression methods for natural disaster risk. Journal of Catastrophology. 30(3): 16–20.

MIL-STD-882E, Department of Defense Standard Practice: System safety (2012, May 11). EverySpec. http://everyspec.com/MIL-STD/MIL-STD-0800-0899/MIL-STD-882E_41682/#:~:text=-MAY-2012)-,MIL-STD-882E%2C%20DEPARTMENT%20OF%20DEFENSE%20STANDARD%20PRACTICE%3A,those%20hazards%20cannot%20be%20eliminated. Retrieved February 10, 2023 

Nilüfer İlçe Afet ve Acil Durum Yönetim Merkezi. (n.d.). Retrieved February 5, 2023, from https://www.nilufer.bel.tr/kategoriler/tesisler/idari-binalar/nilufer-ilce-afet-ve-acil-durum-yonetim-merkeziN12BR

 Oturakçı M, Dağsuyu C (2017). Risk Değerlendirmesinde Bulanık Fine‐Kinney Yöntemi ve Uygulaması. Karaelmas Journal of Occupational Health and Safety. 17–25. https://web.archive.org/web/20200310022619id_/https://dergipark.org.tr/tr/download/article–file/395102

(Özel Haber) Heyelan bir köyü yok ediyor. (2015, April 13). Milliyet. https://www.milliyet.com.tr/yerel-haberler/bursa/ozel-haber-heyelan-bir-koyu-yok-ediyor-10725663

Pamučar DS, Božanić D, Komazec N (2017). Risk assessment of natural disasters using fuzzy logic system of type 2. Management: Journal of Sustainable Business and Management Solutions in Emerging Economies. 21(80): 23–34.https://doi.org/10.7595/management.fon.2016.0016

 Peduzzi P, Dao H, Herold C, Mouton F (2009). Assessing global exposure and vulnerability towards natural hazards: the Disaster Risk Index. Natural Hazards and Earth System Sciences. 9(4): 1149–1159. https://doi.org/10.5194/nhess–9–1149–2009

 Pekel S (2010). Bursa Metropol İlçelerindeki Köy Yerleşmelerinin Kasabalaşma Süreci. Dissertation, Sakarya University.

 Ranger N, Hallegatte S, Bhattacharya S, Bachu M, Priya S, Dhore K, Rafique F, Mathur P, Naville N, Henriet F, Herweijer C, Pohit S, Corfee–Morlot J (2010). An assessment of the potential impact of climate change on flood risk in Mumbai. Climatic Change. 104(1): 139–167. https://doi.org/10.1007/s10584–010–9979–2

 Türkoğlu N (2001). Türkiye’nin yüzölçümü ve nüfusunun deprem bölgelerine dağılışı. Ankara Üniversitesi Türkiye Coğrafyası Araştırma ve Uygulama Merkezi Dergisi. 8:133–148. http://tucaum.ankara.edu.tr/wp–content/uploads/sites/280/2015/08/tucaum8_7.pdf  

 Zhu L, Yin K, Zhang L, Li M (2002). Risk Analysis System Of Geo–Hazards By Using GIS Technique. Journal Of Engıneering Geology. 10(4): 428–433.