• google scholor
  • Views: 6485

  • PDF Downloads: 393

Forest Fire Risk Assessment Using Fuzzy Analytic Hierarchy Process

Divya Mehta1 * , Parminder Kaur Baweja1 and R. K. Aggarwal1

1 Department of Environmental Science,, Dr. Y. S. Parmar University of Horticulture and Forestry Nauni, College of Forestry, Solan, 173230 India

Corresponding author Email: divya.mehta1726@gmail.com

DOI: http://dx.doi.org/10.12944/CWE.13.3.05

Forest fires in the mid hills of Himachal Pradesh are mostly related to human activities. More than 90% of fires are originated from either deliberate or involuntary causes. The purpose of study is linked to identification of forest fire risk factors in 19 villages under Nauni and Oachhghat Panchayats. The methodology paradigm applied here is based on knowledge and fuzzy analytic hierarchy process (FAHP) techniques. Knowledge-based criteria involve socio-economic and biophysical themes for risk assessment. The risk factors are identified according to past occurrence of fire. Fuel type scores highest weight (0.3109) followed by aspect (0.2487), agricultural workers (0.1865), nutritional density (0.1244), population density (0.0622), elevation (0.0311), literacy rate (0.0207) and distance from road (0.0155) in descending order. In the study area applying FAHP, 24.96% of total area was classified under high-risk prone area, 21.69% area classified under high-risk, 34.63% area under moderate risk, while 18.61% area under low risk. The results were in accordance with actual fire occurrences in the past years.

Cumulative Fire Risk Index; Forests Fire; Index Modeling; MCDA; Risk Management

Copy the following to cite this article:

Mehta D, Baweja P. K, Aggarwal R. K. Forest Fire Risk Assessment Using Fuzzy Analytic Hierarchy Process. Curr World Environ 2018;13(3). DOI:http://dx.doi.org/10.12944/CWE.13.3.05

Copy the following to cite this URL:

Mehta D, Baweja P. K, Aggarwal R. K. Forest Fire Risk Assessment Using Fuzzy Analytic Hierarchy Process. Curr World Environ 2018;13(3). Available from: https://bit.ly/2TF0f4r


Download article (pdf)
Citation Manager
Publish History


Article Publishing History

Received: 2018-07-17
Accepted: 2018-11-21
Reviewed by: Orcid Orcid Dr. M. Yuvaraj
Second Review by: Orcid Orcid Wilson Engelmann
Final Approval by: Dr. Kostas Chronopoulos

Introduction

Forest fires are extensive and critical facet of the world. The annual global area burned due to forest fire ranges from 300 and 450 Mha.1 Over 80 percent of the global area burned occurs in grasslands and savannahs, primarily in South Asia, Africa, Australia and South America. Globally fires are frequent over most of the earth except in areas of scant vegetation and near the poles.2

India witnesses most of severe forest fires during the summer season in the hills of Himachal Pradesh.Forest fires have caused extensive damage in recent years leading to loss ofwildlife habitat and biodiversity, change in micro-climate, adverse effect on livelihood of people, addition of green-house gases etc.. Average estimated loss due to forest fire in Himachal Pradesh is INR 113 million per annum.The forests of Himachal Pradesh are mainly comprised of Chir, Oak, Deodara, Khair, Saal, Bamboo and other broad-leaved tree species. Out of above species area occupied by Chir is highly prone to forest fires due to shedding of highly inflammable chir needles.5 The forests of the Solan district are occupied by pure and mixed stands of chir pine and mostly conform to lower Shiwalik chir pine (9C1a) forest type and covers 7.68 per cent of total area of district 5-6. There was need to generate forest fire risk zone for the study area in order to carry out prevention and management measures.

Common practice of Forest Fire Risk Zones has been delineated by assigning knowledge base weights to the risk factor classes according to their sensitivity to fire. Fuzzy Analytic Hierarchy Process (FAHP) has been used as multi-criteria decision analysis (MCDA) tool for weight estimation.7-9

Study Area

The study was carried out in Solan district of Himachal Pradesh, India. Solan occupied 10 percent area of the state i.e. 1,93,600 hectares. The area was primarily occupied by Pinus roxburghii, Quercus leucotric hophora, Acacia catechu, bamboos and other broad leaved tree species. Average daily mean temperature, relative humidity and annual rainfall were 18.4 °C, 1038.2 mm and 51.2 %, respectively.

Figure 1
click here to view figure

Materials and Methods

In this investigation Saaty’s (1998)7Fuzzy Analytic Hierarchy Process (FAHP) was used. FAHP is Multi-criteria Decision Making methodology which involves decision-making framework to rank and prioritize the forest fire risk factors. Table 1 summarizes the related work done over the world.

Table 1: Summary of related works.
 

First Author

Year

Place of Study

Title of Research Work

Variable Studied and Fire Risk Model

Aumedes10

2017

Global

Human-caused fire occurrence modelling in perspective: a review

Distance from roads, railways, urban areas and settlements

Model: HCF model (Human Caused Fire)

Ruffault11

2017

France

Contribution of human and biophysical factors to the spatial distribution of forest fire ignitions and large wildfires in a French Mediterranean region

Human activities and settlements

Ajin12

2016

Kerala, India

Forest Fire Risk Zone Mapping Using RS and GIS Techniques: A Study in Achankovil Forest Division, Kerala, India

Distance from roads and distance from settlements

Model: FRI (Fire Risk Index)

Baweja13

2014

Himachal Pradesh, India

Perceptions of communities exposed to forest fires in western Himalayan region of India

Family size and literacy rate

Vilar14

2014

Europe

Modelling socio-economic drivers of forest fires in the Mediterranean Europe.

Population density, road networks, wildland-urban interface, railway network, protected area, landscape fragmentation etc.

Model: Logistic Regression Model

Spies15

2014

Oregon

Examining fire-prone forest landscapes as coupled human and natural systems

Ownership of land

Ganteaume16

2013

Europe

A review of the main driving factors of forest fire ignition over Europe

Unemployment rate, transport networks and distance to urban areas

Lafragueta17

2013

Spain

GIS and MCE-based forest fire assessment and mapping- A case study in Huesca, Aragon,Spain

Distance from roads, railway track, camping and settlements

Model: FRI (Fire Risk Index)

Sharma18

2012

Himachal Pradesh, India

Fuzzy AHP for forest fire risk modeling

Distance from road and distance from settlement

Model: CFRISK (Cumulative Fire Risk Index)

Gai19

2011

China

GIS-based Forest Fire Risk Assessment and Mapping

Population density and value of forest resources

Model: FRI (Fire Risk Index)

Hoyo20

2011

Spain

Logistic regression models for human-caused wildfire risk estimation: analysing the effect of the spatial accuracy in fire occurrence data

Road infrastructure, recreational and natural protected areas, Cattle-grazing pressure, Buffer of electric lines etc.

Model: Logistic Regression Models

Archibald21

2010

South Africa

Southern African fire regimes as revealed by remote sensing

Population density

Model: FRP (Fire Radiative Power Index)

Calcerrada22

2010

Spain

Spatial modelling of socioeconomic data to understand patterns of human-caused wildfire ignition risk in the SW of Madrid (central Spain)

Population, secondary housing, cattle, sheep and goats

Vadrevu23

2010

Andhra Pradesh, India

Fire Risk Evaluation using multi-criteria analysis- A case study

Population density, agricultural workers, nutritional density and literacy rate

Model: Analytical Hierarchy Process (AHP)

Leone24

2009

Mediterranean

region

Human factors of fire occurrence in the Mediterranean

Agricultural burning, bonfires, power line, engines, machines etc.

Martinez25

2008

Spain

Human-caused wildfire risk rating for prevention planning in Spain

Rural exodus, forest

lands, rural population aging or declining, fuel accumulation in

abandoned agricultural lands, lack of interest in conservation etc.

Model: Logistic Regression Model

Maingi26

2007

United States of America

Factors influencing wildfire occurrence and distribution in eastern Kentucky, USA

Unemployment rates, distance to roads and distance to populated places

Yang27

2007

United States

Spatial Patterns of Modern Period Human-Caused Fire Occurrence in the Missouri Ozark Highlands

Roads, municipalities, ownership, and population density

Rawat28

2003

Uttarakhand, India

Fire Risk Assessment for Forest Fire Control Management in Chilla Forest Range of Rajaji National Park, Uttaranchal, India

Road index and settlement index

Model: CFRISK (Cumulative Fire Risk Index)

 

Hierarchical Structure Development of Fire Risk Criteria

We used population density (PD), agricultural workers (AGRI-W), literacy rate (LR) nutritional density (ND), distance from road (DR), fuel type (FT), aspect (A), slope (S) and elevation (E) for evaluating the fire risk in the study area (Fig. 2).Fuzzy Analytic Hierarchy model was followed in order to construct the hierarchical structure, for reckoning fire risk (Fig. 2).

Relevant socio-economic data for sub-districts of Solan were collected from District Census Handbook. Road maps, Terrain maps and fuel type maps were generated using Shuttle RADAR Topographic Mission (90m), GLOBECOVER (300m) and GLCF, respectively.

Figure 2: Hierarchical data organization for quantifying fire in the study area.
click here to view figure

Forest Fire Risk Index

All factors were classified into five classes, where higher value represented more risk as compared to the lower values (Table 2).

Table 2: Index value and fire rating classes for forest fire risk parameters.
 

Parameter

Class

Index Value

Fire rating class

Population Density

(Peoplekm-2)

0-150

1

Very low

150-300

2

Low

300-450

3

Moderate

450-600

4

High

≥600

5

Very high

Literacy Rate

(%)

0-20

5

Very high

20-40

4

High

40-60

3

Moderate

60-80

2

Low

80-100

1

Very low

Agricultural Workers

(people)

0-5000

1

Very low

5000-10000

2

Low

10000-15000

3

Moderate

15000-20000

4

High

≥20000

5

Very high

Nutritional Density

(People km-2)

0-100

1

Very low

100-200

2

Low

200-300

3

Moderate

300-400

4

High

≥400

5

Very high

Distance from Road Network

(km)

0-1.00

5

Very high

1.00-2.00

4

High

2.00-3.00

3

Moderate

3.00-4.00

2

Low

≥ 4.00

1

Very low

Fuel Type

Conifer Forest

5

Very high

Broad-leaved Forest

4

High

Mixed Forest

3

Moderate

Scrub Lands

2

Low

Cultivated Areas

2

low

 

Urban Areas

1

Very low

 

Bare Areas

1

Very low

Aspect

North

1

Very low

Northeast

1

Very low

Northwest

2

Low

West

2

Low

East

3

Moderate

 

Southeast

4

High

 

Southwest

5

Very high

 

South

5

Very high

Elevation

(m)

≤500

5

Very high

500-1000

4

High

1000-1500

3

Moderate

1500-2000

2

Low

≥2000

1

Very low

Slope

0-10

1

Very low

(degree)

10-20

2

Low

 

20-30

3

Moderate

 

30-40

4

High

 

≥40

5

Very high

 
 

Figure 3: Index map for socio-economic and bio-physical factors.
click here to view figure


Fuzzy Analytic Hierarchy Process (FAHP)

FAHP was used for determining weights for the parameters. A judgmental pair wise comparison matrix ‘A’, was formed using the comparison scales (Table 3). Each entry aij of the matrix ‘A’ was formed comparing the row element ai with the column element aj .29

A=(a_ij ) (i,j…n=1,2…n;n= number of criteria)

The entries a_ij in matrix ‘A’ were done following rules given below:

a_ij>0;a_ij=â–¡(1/a_ji );a_ii=1 for all i

Standardized matrix ‘W’ was formed by using following equation:

Final weights were derived by taking row average of matrix ‘W’.

Consistency of comparisons

The value of ÊŽmax was required to calculate the consistency ratio (CR).24

Consistency index (CI) = (ÊŽmax–n) / (n-1).

Where; ÊŽmax = largest eigen value and n = number of criteria.

The final consistency ratio was calculated by dividing the consistency index with the random index.

CR = CI / RI

Where; RI = Random index and CI = Consistency index.

Consistency ratio was designed such a way that shows a reasonable level of consistency in the pair wise comparisons if CR < 0.10 and CR ≥ 0.10 indicated inconsistent.

Table 3: Scale used in Fuzzy Analytical Hierarchy Process30
 

Intensity of scale

Linguistic variable

1

Equally important

3

Weakly important

5

Essentially important

7

Very strongly important

9

Absolutely important

2,4,6,8

intermediate values between two adjacent judgments


Results and Discussion

Results pertaining to estimated weights of selected fire risk factors revealed highest weight for fuel type (0.3109) followed by aspect (0.2487), agricultural workers (0.1341), nutritional density (0.1244), population density (0.0622), slope (0.0524), elevation (0.0311), literacy rate (0.0207) and distance from road (0.0155), respectively (Table 4).

Table 4: Estimated weights of forest fire risk parameters
 

Socio-economic Parameter

Weight

Bio-physical Parameter

Weight

Population density (person km-2)

0.0622

Fuel type

0.3109

Literacy rate (%)

0.0207

Aspect

0.2487

Agricultural workers (person)

0.1341

Slope (degree)

0.0524

Nutritional density (person km-2)

0.1244

Elevation (m)

0.0311

Distance from road (m)

0.0155

 

The resulting weights from Fuzzy Analytic Hierarchy Process were applied in the Cumulative Forest Fire Risk Index model. Table 5 demonstrated the fire risk for five classes of CFRISK index value. CFRISK model had been shown in the following equation:-

CFRISK = 0.0622*PDI +0.0207*LRI + 0.1341*AWI + 0.1244*NDI + 0.0155*DRI+ 0.3109*FTI + 0.0524*SI+ 0.2487*AI + 0.0311*EI.

Where;

CFRISK = Cumulative Fire Risk Index

PDI = Population density index

LRI = Literacy rate index

AWI = Agricultural worker index

NDI = Nutritional density index

DRI = Distance from road index

FTI = Fuel type index

SI = Slope index

AI = Aspect index

EI = Elevation Index

Table 5: Cumulative Forest Fire Risk (CFRISK) Index potential scale 24.

Index

Forest Fire Risk

0-1

Very low

1-2

Low

2-3

Moderate

3-4

High

4-5

Very high


Out of total geographical area of Solan district, 4.15% area was classified under very high risk, 40.63% area under high risk, 54.00% area under moderate risk, 0.84% area under low risk and 0.37% under very low risk (Fig. 4a).Accuracy of the Forest Fire Risk map was tested using NASA FIRMS forest fire dataset for the year 2018 (Fig. 4b). The Forest Fire Risk map for the three classes alone viz. moderate, high and very high predicted 99.4% of the total fire pixels (1012).The moderate class predictive capability was highest (60.77%), followed by high (33.99%) and very high (4.64%) fire risk class.
 

Figure 4: (a) Forest Fire Risk Map for Solan district and (b) Forest Fire hot spot derived from NASA FIRMS datasets for the year 2018.
click here to view figure 

  
Abbreviations
 

Mha

million hectares

%

Percent

°C

degree Celsius

CFRISK

cumulative fire risk index

FAHP

Fuzzy Analytic Hierarchy Process

FFRZ

forest fire risk zone

GIS

geographical information system

km2

kilometer square

MCDA

multi-criteria decision analysis

m

meter

mm

Millimeter

PCM

pairwise comparison matrix

ÊŽmax

lambda maximum

α

Alpha

β

Beta

γ

Gamma

δ

Delta

 

Acknowledgements

The assistance provided by Dr SK Bhardwaj Prof. &Head, Department of Environmental Science, and Dr IK Thakur, Principal Scientist, Department of Tree Improvement and Genetic Resources, Dr YS Parmar University of Horticulture and Forestry, Nauni, HP-India in the present study is highly acknowledged.

References
 

  1. Van der Werf G.R., Randerson J.T., Giglio L., Collatz G. J., Kasibhatla P. S., Arellano Jr. A. F. Interannual variability in global biomass burning emissions from 1997 to 2004. Atmosphere Chemistry Physics. 2006;6(11):3423–3441.
    CrossRef
  2. Mouillot F., Field C.B. Fire history and the global carbon budget: A fire history reconstruction for the 20th century. Global Change Biology. 2005;11(3):398-420.
    CrossRef
  3. Forest Survey of India. State forest report, Dehradun, Uttarakhand. 2003.
  4. Anonymous. Forest fires and its effect on environment, forests, bio-diversity and remedial/preventive measures. Directory of Forest Fire Control and Management, Himachal Pradesh. 2016.
  5. Shah S., Sharma D. P. Land use change detection in Solan Forest Division, Himachal Pradesh, India. Forest Ecosystems. 2015;2:26-38.
    CrossRef
  6. Champion H. G., Seth S. K. A Revised Survey of Forest Types of India. Management of Publications, Govt. of India, Delhi. 1968;34.
  7. Saaty L. T. Multicriteria Decision Making: The Analytic Hierarchy Process, RWS Publications, Pittsburgh, PA. 1988.
  8. Vaidya O. S., Kumar S. Analytic hierarchy process: An overview of applications. European Journal of Operational Research. 2006;169(1):1-29.
    CrossRef
  9. Vahidniaa M. H., Alesheikhb A., Alimohammadic A., Bassirid A. Fuzzy Analytical Hierarchy Process in GIS application.The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. 2008;37:593-596.
  10. Aumedes S.C., Comas C., Garcia V.C. Human-caused fire occurrence modelling in perspective: a review. International Journal of Wildland Fire. 2018;26(12):983-998.
    CrossRef
  11. Ruffault J., Mouillot F. Contribution of human and biophysical factors to the spatial distribution of forest fire ignitions and large wildfires in a French Mediterranean region. International Journal of Wildland Fire. 2017;26(6):498-508.
    CrossRef
  12. Ajin R. S., Loghin A. M., Vinod P. G., Jacob M. K. Forest Fire Risk Zone Mapping Using RS and GIS Techniques: A Study in Achankovil Forest Division, Kerala, India. Journal of Earth, Environment and Health Sciences. 2016;2(3):109.
    CrossRef
  13. Baweja P. K., Kundu S. Perceptions of communities exposed to forest fires in western Himalayan region of India. Online International Journal of Biosolution. 2014;2(4):94-99.
  14. Vilar L., Camia A., Ayanz J. S. M. Modelling socio-economic drivers of forest fires in the Mediterranean Europe. In: Social and Economic issues, Coimbra University Press, Spain. 2014;1874-1882.
    CrossRef
  15. Spies T. A., White E. M., Kline J. D., Fischer A. P., Ager A., Bailey J., Bolte J., Koch J., Platt E., Olsen C. S., Jacobs D. Examining fire-prone forest landscapes as coupled human and natural systems. Ecology and Society. 2014;19(3):9.
    CrossRef
  16. Ganteaume A., Camia A., Jappiot M., San-Miguel-Ayanz J., Long-Fournel M., Lampin C. A review of the main driving factors of forest fire ignition over Europe. Environmental management. 2013;51(3):651-662.
    CrossRef
  17. Lafragueta J. F. GIS and MCE-based forest fire risk assessment and Mapping:A case study in Huesca, Aragon, Spain. M Sc Thesis. School of Engineering and the Built Environment, Spain. 2013.
  18. Sharma L. K., Kanga S., Singh M. N., Sinha S., Chandra P. P. Fuzzy AHP for forest fire risk modeling. Disaster Prevention and Management: An International Journal. 2012;21(2):160-171.
    CrossRef
  19. Gai C., Weng W., Yuan H. GIS-based forest fire risk assessment and mapping. Paper presented at: Computational Sciences and Optimization (CSO), 2011 Fourth International Joint Conference. 2011;1240-1244.
    CrossRef
  20. Hoyo L. V., Isabel M. P. M., Vega F. J. M. Logistic regression models for human-caused wildfire risk estimation: analysing the effect of the spatial accuracy in fire occurrence data. European Journal of Forest Research. 2011;130:983-996.
    CrossRef
  21. Archibald S., Scholes R. J., Roy D. P., Roberts G., Boschetti L. Southern African fire regimes as revealed by remote sensing. International Journal of Wildland Fire. 2010;19(7):861-878.
    CrossRef
  22. Calcerrada R., Barrio-Parra F., Millington J. D. A., Novillo C. J. Spatial modelling of socioeconomic data to understand patterns of human-caused wildfire ignition risk in the SW of Madrid (central Spain). Ecological Modelling. 2010;221(1):34-45.
    CrossRef
  23. Vadrevu K. P., Eaturu A., Badrinath K. V. S. Fire Risk Evaluation using multicritiria analysis- A case study. Environmental monitoring and assessment. 2010;166:223-239.
    CrossRef
  24. Leone V., Lovreglio R., Martín M.P., Martínez J., Vilar L. Human factors of fire occurrence in the Mediterranean. In: Earth observation of wildland fires in Mediterranean ecosystems. 2009;149-170.
    CrossRef
  25. Martinez J., Garcia C. V., Chuvieco E. Human-caused wildfire risk rating for prevention planning in Spain. Journal of environmental management. 2009;90(2):1241-1252.
    CrossRef
  26. Maingi J. K., Henry M. C. Factors influencing wildfire occurrence and distribution in eastern Kentucky, USA. International Journal of Wildland Fire. 2007;16(1):23-33.
    CrossRef
  27. Yang J., He H.S., Shifley S.R., Gustafson E. J. Spatial patterns of modern period human-caused fire occurrence in the Missouri Ozark Highlands. Forest science. 2007;53(1):1-15.
  28. Rawat G.S. Fire Risk Assessment for Forest Fire Control Management in Chilla Forest Range of Rajaji National Park, Uttaranchal, India. M Sc Thesis. International Institute for Geo-information Science and Earth Observation Enscheda. 2003.
  29. Ramanathan A. A note on the use of the analytic hierarchy process for environmental impact assessment. Journal of Environmental Management. 2001;63(1):27–35.
    CrossRef