Skip to main content

Advertisement

SpringerLink
Log in

Identifying wildland fire ignition factors through sensitivity analysis of a neural network

  • Original Paper
  • Published:
Natural Hazards Aims and scope Submit manuscript

Abstract

Artificial neural networks (ANNs) show a significant ability to discover patterns in data that are too obscure to go through standard statistical methods. Data of natural phenomena usually exhibit significantly unpredictable non-linearity, but the robust behavior of a neural network makes it perfectly adaptable to environmental models such as a wildland fire danger rating system. These systems have been adopted by many developed countries that have invested in wildland fire prevention, and thus civil protection agencies are able to identify areas with high probabilities of fire ignition and resort to necessary actions. Since one of the drawbacks of ANNs is the interpretation of the final model in terms of the importance of variables, this article presents the results of sensitivity analysis performed in a back-propagation neural network (BPN) to distinguish the influence of each variable in a fire ignition risk scheme developed for Lesvos Island in Greece. Four different methods were utilized to evaluate the three fire danger indices developed within the above scheme; three of the methods are based on network’s weights after the training procedure (i.e., the percentage of influence—PI, the weight product—WP, and the partial derivatives—PD methods), and one is based on the logistic regression (LR) model between BPN inputs and observed outputs. Results showed that the occurrence of rainfall, the 10-h fuel moisture content, and the month of the year parameter are the most significant variables of the Fire Weather, Fire Hazard, and Fire Risk Indices, respectively. Relative humidity, elevation, and day of the week have a small contribution to fire ignitions in the study area. The PD method showed the best performance in ranking variables’ importance, while performance of the rest of the methods was influenced by the number of input parameters and the magnitude of their importance. The results can be used by local forest managers and other decision makers dealing with wildland fires to take the appropriate preventive measures by emphasizing on the important factors of fire occurrence.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

References

  • Agee JK (1994) Fire and weather disturbances in terrestrial ecosystems of the eastern Cascades. General technical report PNW-GTR-320, USDA, Forest Service, Pacific Northwest Research Station, Portland

  • Alcázar J, Garcia CV, Grauet M, Pemán J, Fernández Á (1998) Human risk and fire danger estimation through multicriteria evaluation methods for forest fire prevention in Barcelona, Spain. In: Proceedings of the 3rd international conference on forest fire research and 14th conference on fire and forest meteorology, Luso, 16–20 Nov 1998, pp 2379–2387

  • Andersson FO, Aberg M, Jacobsson SP (2000) Algorithmic approaches for studies of variable influence, contribution and selection in neural networks. Chemom Intell Lab Syst 51:61–72. doi:10.1016/S0169-7439(00)00057-5

    Article  Google Scholar 

  • Andrews PL, Bevins CD, Seli RC (2003) BehavePlus fire modeling system user’s guide, v. 2.0. General technical report RMRS-GTR-106WWW, USDA, Forest Service, Rocky Mountain Research Station, Ogden

  • Bishop CM (1995) Neural networks for pattern recognition. Oxford University Press, Oxford

    Google Scholar 

  • Chen K, Blong R, Jacobson C (2003) Towards an integrated approach to natural hazards risk assessment using GIS: with reference to bushfires. Environ Manag 31(4):546–560. doi:10.1007/s00267-002-2747-y

    Article  Google Scholar 

  • Chou YH (1992a) Management of wildfires with a Geographical Information System. Int J Geogr Inf Syst 6(2):123–140. doi:10.1080/02693799208901900

    Article  Google Scholar 

  • Chou YH (1992b) Spatial autocorrelation and weighting functions in the distribution of wildland fires. Int J Wildland Fire 2(4):169–176. doi:10.1071/WF9920169

    Article  Google Scholar 

  • Chou YH, Minnich RA, Chase RA (1993) Mapping probability of fire occurrence in San Jacinto Mountains, California, USA. Environ Manag 17(1):129–140. doi:10.1007/BF02393801

    Article  Google Scholar 

  • Chuvieco E, Congalton RG (1989) Application of remote sensing and Geographic Information System to forest fire hazard mapping. Remote Sens Environ 29:147–159. doi:10.1016/0034-4257(89)90023-0

    Article  Google Scholar 

  • Chuvieco E, Salas J (1996) Mapping the spatial distribution of forest fire danger using GIS. Int J Geogr Inf Syst 10(3):333–345. doi:10.1080/02693799608902082

    Article  Google Scholar 

  • Chuvieco E, Deshayes M, Stach N, Cocero D, Riaño D (1999a) Short-term fire risk: foliage moisture content estimation from satellite data. In: Chuvieco E (ed) Remote sensing of large wildfires in the European Mediterranean Basin. Springer-Verlag, Berlin

    Google Scholar 

  • Chuvieco E, Salas J, Carvacho L, Rodríguez-Silva F (1999b) Integrated fire risk mapping. In: Chuvieco E (ed) Remote sensing of large wildfires in the European Mediterranean Basin. Springer-Verlag, Berlin

    Google Scholar 

  • Dedecker AP, Goethals PL, D’heygere T, Gevrey M, Lek S, De Pauw N (2005) Application of artificial neural network models to analyse the relationships between Gammarus pulex L. (Crustacea, Amphipoda) and river characteristics. Environ Monit Assess 111(1–3):223–241. doi:10.1007/s10661-005-8221-6

    Article  Google Scholar 

  • Deeming JE, Burgan RE, Cohen JD (1977) The national fire-danger rating system—1978. General technical report INT-39, USDA, Forest Service, Intermountain Forest and Range Experiment Station, Ogden

  • Dimopoulos I, Bourret P, Lek S (1995) Use of some sensitivity criteria for choosing networks with good generalization ability. Neural Process Lett 2(6):1–4. doi:10.1007/BF02309007

    Article  Google Scholar 

  • Dimopoulos I, Chronopoulos J, Chronopoulou-Sereli A, Lek S (1999) Neural network models to study relationships between lead concentration in grasses and permanent urban descriptors in Athens city (Greece). Ecol Model 120:157–165. doi:10.1016/S0304-3800(99)00099-X

    Article  Google Scholar 

  • Fausett L (1994) Fundamentals of neural networks: architecture, algorithms and applications. Prentice-Hall Inc, USA

    Google Scholar 

  • Finklin AI, Fischer WC (1990) Weather station handbook—an interagency guide for wildland mangers. National Wildfire Coordinating Group, NFES no. 2140, Boise

  • Garson GD (1991) Interpreting neural network connection weights. Artif Intell Expert 6:47–51

    Google Scholar 

  • Gevrey M, Dimopoulos I, Lek S (2003) Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecol Model 160(3):249–264. doi:10.1016/S0304-3800(02)00257-0

    Article  Google Scholar 

  • Goh ATC (1995) Back-propagation neural networks for modeling complex systems. Artif Intell Eng 9:143–151. doi:10.1016/0954-1810(94)00011-S

    Article  Google Scholar 

  • Hampshire JBII, Pearlmutter BA (1990) Equivalence proofs for multi-layer perceptron classifiers and the Bayesian discriminant function. In: Touretzky D, Elman J, Sejnowski T, Hinton G (eds) Proceedings of the 1990 Connectionist Models Summer School. Morgan Kaufmann, CA

    Google Scholar 

  • Henderson M, Kalabokidis K, Marmaras E, Konstantinidis P, Marangudakis M (2005) Fire and society: a comparative analysis of wildfire in Greece and the United States. Hum Ecol Rev 12(2):169–182

    Google Scholar 

  • Hoffmann AA, Schindler L, Goldammer JG (1999) Aspects of a fire information system for East Kalimantan, Indonesia. In: Proceedings of the 3rd international symposium on Asian tropical forest management, Samarinda, 20–23 Sept 1999

  • Hosmer D, Lemeshow S (1989) Applied logistic regression. A much-cited recent treatment utilized in SPSS routines. Wiley, NY

    Google Scholar 

  • Howes P, Crook N (1999) Using input parameter influences to support the decisions of feedforward neural networks. Neurocomputing 24:191–206. doi:10.1016/S0925-2312(98)00102-7

    Article  Google Scholar 

  • Ibarra AA, Gevrey M, Park Y, Lim P, Lek S (2003) Modelling the factors that influence fish guilds composition using a back-propagation network: assessment of metrics for indices of biotic integrity. Ecol Model 160:281–290. doi:10.1016/S0304-3800(02)00259-4

    Article  Google Scholar 

  • Jordan MI (1995) Why the logistic function? A tutorial discussion on probabilities and neural networks. MIT computational cognitive science report 9503. http://www.cs.berkeley.edu/~jordan

  • Kalabokidis KD, Koutsias N (2000) Geographic multivariate analysis of spatial fire occurrence. Geotech Sci Issues 11(1):37–47

    Google Scholar 

  • Kalabokidis KD, Gatzojannis S, Galatsidas S (2002a) Introducing wildfire into forest management planning: towards a conceptual approach. For Ecol Manag 158(1–3):41–50. doi:10.1016/S0378-1127(00)00715-5

    Article  Google Scholar 

  • Kalabokidis KD, Konstantinidis P, Vasilakos C (2002b) GIS analysis of physical and human impact on wildfire patterns. In: Viegas DX (ed) Proceedings of the IV international conference on forest fire research & wildland fire safety, Coimbra. Millpress, Rotterdam

  • Kalabokidis KD, Koutsias N, Konstantinidis P, Vasilakos C (2007) Multivariate analysis of landscape wildfire dynamics in a Mediterranean ecosystem of Greece. Area 39(3):392–402. doi:10.1111/j.1475-4762.2007.00756.x

    Article  Google Scholar 

  • Larjavaara M, Kuuluvainen T, Tanskanen H, Venäläinen A (2004) Variation in forest fire ignition probability in Finland. Silva Fenn 38(3):253–266

    Google Scholar 

  • Masters T (1993) Practical neural networks recipes in C++. Academic Press, USA

    Google Scholar 

  • Mendenhall W, Sincich T (1996) A second course in statistics: regression analysis. Prentice-Hall Inc, USA

    Google Scholar 

  • Montano JJ, Palmer A (2003) Numeric sensitivity analysis applied to feedforward neural networks. Neural Comput Appl 12:119–125. doi:10.1007/s00521-003-0377-9

    Article  Google Scholar 

  • Nord LI, Jacobsson SP (1998) A novel method for examination of the variable contribution to computational neural network models. Chemom Intell Lab Syst 44:153–160. doi:10.1016/S0169-7439(98)00118-X

    Article  Google Scholar 

  • Norusis MJ (1990) SPSS/PC + Advanced Statistics™ 4.0 for the IBM PC/XT/AT and PS/2. SPSS Inc, Chicago

    Google Scholar 

  • Olden JD, Jackson DA (2002) Illuminating the “black box”: a randomization approach for understanding variable contributions in artificial neural networks. Ecol Model 154:135–150. doi:10.1016/S0304-3800(02)00064-9

    Article  Google Scholar 

  • Olden JD, Joy MK, Death RG (2004) An accurate comparison of methods for quantifying variable importance in artificial neural networks using simulated data. Ecol Model 178:389–397. doi:10.1016/j.ecolmodel.2004.03.013

    Article  Google Scholar 

  • Özesmi SL, Özesmi U (1999) An artificial neural network approach to spatial habitat modeling with interspecific interaction. Ecol Model 116:15–31. doi:10.1016/S0304-3800(98)00149-5

    Article  Google Scholar 

  • Papadokonstantakis A, Lygeros A, Jacobsson SP (2006) Comparison of recent methods for inference of variable influence in neural networks. Neural Netw 19(4):500–513. doi:10.1016/j.neunet.2005.09.002

    Article  Google Scholar 

  • Park YS, Rabinovich J, Lek S (2007) Sensitivity analysis and stability patterns of two-species pest models using artificial neural networks. Ecol Model 204(3–4):427–438. doi:10.1016/j.ecolmodel.2007.01.021

    Article  Google Scholar 

  • Preisler HK, Brillinger DR, Burgan RE, Benoit JW (2004) Probability based models for estimation of wildfire risk. Int J Wildland Fire 13(2):133–142. doi:10.1071/WF02061

    Article  Google Scholar 

  • Pyne SJ, Andrews PL, Laven RD (1996) Introduction to wildland fire, 2nd edn. Wiley, NY

    Google Scholar 

  • Rumelhart DE, McClelland JL (1986) Parallel distributed processing: explorations in the microstructure of cognition. The MIT Press, USA

    Google Scholar 

  • Salas FJ, Chuvieco E (1994) GIS applications to forest fire risk mapping. Wildfire 3:7–13

    Google Scholar 

  • Sarle WS (1997) Neural network FAQ. Periodic posting to the Usenet newsgroup comp.ai.neural-nets. ftp://ftp.sas.com/pub/neural/FAQ.html

  • Schroeder MJ, Buck CC (1970) Fire weather. Agriculture handbook 360. USDA Forest Service, Washington, DC

    Google Scholar 

  • Stocks BJ, Lawson BD, Alexander ME, Van Wagner CE, McAlpine RS, Lynham TJ, Dubé DE (1989) The Canadian forest fire danger rating system: an overview. For Chron 65:258–265

    Google Scholar 

  • Taylor SW, Alexander ME (2006) Science, technology and human factors in fire danger rating: the Canadian experience. Int J Wildland Fire 15:121–135. doi:10.1071/WF05021

    Article  Google Scholar 

  • Tchaban T, Taylor MJ, Griffin A (1998) Establishing impacts of the inputs in a feedforward network. Neural Comput Appl 7:309–317. doi:10.1007/BF01428122

    Article  Google Scholar 

  • Thompson WA (2000) Using forest fire hazard modeling in multiple use forest management planning. For Ecol Manag 134:163–176. doi:10.1016/S0378-1127(99)00255-8

    Article  Google Scholar 

  • Trewartha GT, Horn LH (1980) An introduction to climate, 5th edn. McGraw-Hill, NY

    Google Scholar 

  • Van Wagner CE (1987) Development and structure of the Canadian Forest Fire Weather Index System. Forestry technical report 35, Canadian Forestry Service, Ottawa

  • Vasconcelos MJP, Silva S, Tome M, Alvim M, Pereira JMC (2001) Spatial prediction of fire ignition probabilities: comparing logistic regression and neural networks. Photogramm Eng Remote Sens 67(1):73–81

    Google Scholar 

  • Vasilakos C, Kalabokidis K, Hatzopoulos J, Kallos G, Matsinos J (2007) Integrating new methods and tools in fire danger rating. Int J Wildland Fire 16(3):306–316. doi:10.1071/WF05091

    Article  Google Scholar 

  • Vega-García C, Lee B, Wooddard PM, Titus SJ (1996) Applying neural network technology to human-caused wildfire occurrence prediction. Artif Intell Appl 10(3):9–18

    Google Scholar 

  • Viegas X, Bovio G, Ferreira A, Nosenzo A, Bernard S (1999) Comparative study of various methods of fire danger evaluation in Southern Europe. Int J Wildland Fire 9(4):235–246. doi:10.1071/WF00015

    Article  Google Scholar 

  • Yoon Y, Guimaraes T, Swales G (1994) Integrating artificial neural networks with rule-based expert systems. Decis Support Syst 11:497–507. doi:10.1016/0167-9236(94)90021-3

    Article  Google Scholar 

  • Yuan M (1997) Use of knowledge acquisition to build wildfire representation in Geographical Information Systems. Int J Geogr Inf Sci 11(8):723–745. doi:10.1080/136588197242059

    Article  Google Scholar 

Download references

Acknowledgments

This research was funded by the European Union within the RTD project “Automated Fire and Flood Hazard Protection System/AUTO-HAZARD PRO” (EVG1-CT-2001-00057). The authors would like to thank their colleagues at the Geography of Natural Disasters Laboratory, University of the Aegean, and the Greek Fire Brigade and Forest Service authorities for their cooperation. Dr. Philip N. Omi of Colorado State University, USA; Dr. Peter F. Moore of Sustainable Forestry Management Ltd, Australasia; and two anonymous referees are acknowledged for their review comments and suggestions in an earlier version of the manuscript.

Author information

Authors and Affiliations

  1. Department of Environmental Studies, University of the Aegean, University Hill, 81100, Mytilene, Greece

    Christos Vasilakos, John Hatzopoulos & Ioannis Matsinos

  2. Department of Geography, University of the Aegean, University Hill, 81100, Mytilene, Greece

    Kostas Kalabokidis

Authors
  1. Christos Vasilakos
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Kostas Kalabokidis
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. John Hatzopoulos
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Ioannis Matsinos
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Christos Vasilakos.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Vasilakos, C., Kalabokidis, K., Hatzopoulos, J. et al. Identifying wildland fire ignition factors through sensitivity analysis of a neural network. Nat Hazards 50, 125–143 (2009). https://doi.org/10.1007/s11069-008-9326-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11069-008-9326-3

Keywords

Search

Navigation