The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad,

FOREST FIRE SCENARIO AND CHALLENGES OF MITIGATION DURING FIRE SEASON IN NORTH EAST INDIA

Kasturi Chakraborty1, Partho Protim Mondal, Mayuri Chabukdhara and S. Sudhakar North Eastern Space Applications Centre, Umiam, Meghalaya-793103,India [email protected]

KEY WORDS: fire, North Eastern Region, Fire hazard map, fire mitigation, fire spread, DEM, , MODIS

ABSTRACT

Forest fires are a major environmental problem in North East Region (NER) with large tracts of forest areas being affected in every season. Forest fires have become a major threat to the forest ecosystems in the region, leading to loss of timber, biodiversity, wildlife habitat and loss to other natural resources. Studies on forest fire have reported that about 50% of forest fire in the country takes place in NE region. The forest fire in NER is anthropogenic in nature. The forest fire hazard map generated based on appropriate weightage given to the factors affecting fire behavior like topography, fuel characteristic and proximity to roads, settlements and also historical fire locations helped to demarcate the fire prone zones. Whereas, during fire season the weather pattern also governs the fire spread in the given area. Therefore, various data on fuel characteristics (land use/land cover, forest type map, forest density map), topography (DEM, slope, aspect) proximity to settlement, road, waterbodies, meteorological data from AWS on wind speed, wind direction, dew point have been used for each fire point to rank its possible hazard level. Near real time fire location data obtained from MODIS/FIRMSwere used to generate the fire alerts. This work demonstrates dissemination of information in the form of maps and tables containing information of latitude and longitude of fire location, fire occurrence date, state and district name, LULC, road connectivity, slope and aspect, settlements/water bodies and meteorological data and the corresponding rating of possibility of fire spread to the respective fire control authorities during fire season.

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 27 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad, India

INTRODUCTION decision on solutions can only be satisfactory when a fire risk zone Forest fire is considered a frequent and mapping is available (Jaiswal et constant natural disaster in the forest al.,2002). Geospatial technology, ecosystems. The slash and burn shifting including Remote sensing (RS) and cultivation, or locally known as jhum the Geographic Information Systems (GIS), predominant form of agriculture in the provides the information and the tools hill tracts of North East India is found to necessary to develop a forest fire be the major cause of forest fire in the susceptibility map in order to identify, region. The fires in the northeast India classify and map fire hazard area mostly pertain to slash and burn (Sowmya and Somashekar,2010). Geo- agriculture (Ramakrishnan,1988 and spatial modeling included a lot of Majumder et al., 2011). In northeast information about the settlements, road, India, indigenous institutions play an elevation and all the other parameters important role in . which are very useful when forest fire About 9.8 million ha of forestland is preparedness and management under community control in the seven strategies are designed (Roy and north-eastern states. Based on the Porwal,2005). This study aims to information collected by Forest Survey analyse the forest fire incidents in the of India officials in consultation with field past and its association with land officials of the State Forest use/land cover, topography and Departments, the main reasons for the infrastructure and to generate a forest change in the forest cover has been fire vulnerability map showing the summarized to be due to the shortening proneness of the to forest fire of the shifting cultivation and biotic hazard. With the help of such pressure besides other reasons of information forest fire mitigation encroachment and illicit strategies. (SFR,2011). These regions are also characterized by the high fuel loads Study area (Baishya et al.,2009]. The history of incidence of forest fires in the state of Arunachal Pradesh has been studied The study has been carried out in the between 1985 to 2005 and concluded North Eastern Region (NER) of India. that incidences of forest fire is related to The study area covers the seven sister increasing interference of human states of NER viz., Arunachal Pradesh, activities inside natural forest covers Assam, Manipur, Meghalaya, Mizoram, (Palit,2007).Increased incidences of forest fire have prompted government intervention and schemes aimed at preventing and controlling forest fire in Mizoram (Darlong,2001). In North Eastern India, satellite data observation reveal numerous fires in the Mizoram and surrounding Indian states of Tripura, Assam, Manipur, and along the border with Myanmar. A precise evaluation of forest fire problems and Fig1. Study area

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 28 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad, India

Nagaland, and Tripura (Fig.1). The state Table 1: Themes and parameter used in of Sikkim has not been considered in the study the study considering the low fire incidents in the region. NER covers Sl. Parameters Classes about 8% of total India’s size. The No. geographical coordinates of the study Deciduous o o o Evergreen/Semi- area is 90 26’ 0’’ to 97 33’ 0” and 21 1. Vegetation o evergreen 50’ 0” to 29 27’ 30”.The forests of NER Pine/ Bamboo/Temperate are rich in biodiversity and timber. Open According to the India State of Forest 2. Vegetation Density Moderate Report, 2011North Eastern states that Dense make up one-fourth of the country’s 15-25° forest cover. NER is one of 25-35° the“biodiversity hotspots” in India. 3. Slope (degree) 8-15° 3-8° Various survey in this region reveal that 0-3° the land is under heavy pressure due to N / NE/ E / SE / SW / S 4. Aspect jhum (shifting cultivation) for agriculture, /W / NW 500-1000 the major land-use practice and facing 1000-2000 5. Elevation (m) frequent forest fire events. 200-500 0-200/>2000 DATA USED AND METHODOLOGY 1000-1500 500-1000/1500-2000 6. Distance to settlement (m) Influencing factors like vegetation, 0-500/2000-2500 topography, wind direction, proximity to 2500-3000 village and road etc., have to be taken 50-100/100-150 200-250 into consideration in identification of fire 7. Distance to road (m) prone areas and their prioritization in 0-50/250-300 forest fire risk zone mapping. The 150-200 >1000 distribution of fire events derived from 500-1000 8. Distance to water bodies (m) Forest Survey of India (FSI) was clipped 300-500 to obtain the fire counts in the region. 0-100/100-200/200-300 The forest fire risk map has been From January 2001 to prepared with the following inputs: Historical data and recent fire April 2013. 9. events (number of fire points) Fire incidents of 2014 used for validation of the map. 10. Land use/land cover map For removal of non-forest (2005-06) area

Vegetation map (Forest / Fuel type) has been generated taking in account the vegetation characteristics like type and density. It is generated by reclassification of vegetation map into risk classes depending upon fire sensitivity. Elevation information is obtained from Digital Elevation Model (DEM). Slope is one of the important

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 29 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad, India parameters that influence fire behaviour. generated using multi-criteria decision Fire moves most quickly upward slope analysis (MCDA) in GIS environment. and least quickly downward slope Fire risk zones were classified into five (Jaiswal et al., 2002). With increase in categories, viz., very high, high, angle of slope, fire risk increases. moderate, low and very low. The map Hence, slope risk zone map has been shows the different zones with different generated from DEM. Aspect generated level of proneness to forest fire. The using DEM is another important definition of risk level (high, moderate parameter to be considered in fire and low) is a crucial subject to indicate mapping. Distance from roads and the zonation showing high, moderate urban areas are important consideration and low risk zones with regard to forest because forest regions are more fire fire in the final fire risk map. prone where they are located near to roads and high road density. We cannot RESULTS use same weights and same variables indifferent regions because forest fire in Based on the history of fire occurrence each part of earth has its own (from January 2001 to April2013), forest characteristics. The fire points fire scenario in NER has been downloaded from NRSC,Bhuvan and studied.Fig. 1 shows Mizoram is highly FSI website are actually the fire points affected by forest fire. Fig2 and Fig.3 generated from MODIS data which uses shows the fire prone zones of NER. brightness temperature derived MODIS 4µm and 11 µm channels. The FSI, Dissemination of forest fire risk alerts 2011 vegetation type and density map In addition to the ongoing forest fire have been used for fuel characteristics, alerts being provided everyday to the topographic information is generated respective forest departments from from DEM (30m ASTER DEM) while authorized agencies, NESAC under its available maps of road and settlement ongoing program of Disaster Risk were used for proximity analysis. Reduction in North East Region is Considering the uniqueness of location, issuing fire risk alert of each location in vegetation type etc., of North Eastern NER obtained from MODIS website Region, the given parameters has been everyday during fire season (Fig.4). analyzed separately for each state. Each of the theme/parameter has been Most of the fire incidents in NER are classified into different classes as anthropogenic in nature. To meet the shown in table1. Appropriate challenges of forest fire mitigation weightages were applied to elevation, activities in NER it is observed that slope and aspect, vegetation type and during fire season it is very important to density, the distance to settlement, road build a strong communication network and water bodies. For the forest fire between the forest departments and the vulnerability mapping a decision making community and the respective table was generated with the given departments engaged in providing parameters and historical data analysis alerts. It is very important that the forest of forest fire occurrence in each state. department and the local community are With the help of the above parameters aware of the fire risk zones where forest and weights fire hazard zonation map is

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 30 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad, India fire can spread rapidly by taking up states have significant proportion of awareness programs. Since in this area under high risk zones. Tripura has region forests are community controlled approx. 52% followed by Nagaland and fire are mostly occurring due to the 42%, Mizoram 36%, Manipur 35% and shifting cultivation practice therefore, Meghalaya 32% area under highly appropriate precaution needs to be prone to forest fire hazard. Whereas, taken up during fire season. There is a Arunachal Pradesh has maximum area need for bringing up proper scientific under moderate hazard zone (44%) and knowledge. The forest departments and Assam has maximum low hazard zones. fire control authorities are required to be Mizoram has above 40% area under equipped with modern fire fighting moderate vulnerability followed by instruments. The topography of the Arunachal Pradesh, Manipur, region is an important consideration in Meghalaya and Nagaland. Contrary to the fire mitigation. Therefore, the forest the given facts, the observation of fire controlling authorities and historical data of fire occurrence and the communities should be aware of the present scenario shows that Mizoram is alternate roads, nearest waterbody etc. facing maximum fire incidents (31.38%) It is very important that the fire in NER in the region in comparison to Manipur should be tackled with the community (approx.16%), Assam (approx 15%), participation. The various information Meghalaya (approx 12%),Nagaland generated under scientific programs has (10.15%), Tripura (9.16%) and to reach to the community level for Arunachal Pradesh (6.23%).The number better forest fire management. of forest fire points were compared with current shifting cultivation (r = 0.5) and abandoned shifting cultivation (r = 0.99) DISCUSSION in Mizoram. The result indicates that The historical data analysis of the forest most fire incidents are occurring in fire incidents in the region helped to abandoned shifting patches along with understand the severity of the forest fire other forest area. The forest fire incidents in each north eastern state of incidents that occurred during 2014 India. The classification of the have been compared with the variable/parameter for zonation of fire vulnerability map that shows that most proneness has been done depending on of the fire incidents have occurred in the the history of fire incidents. Past studies high to moderate vulnerable areas. on forest fire helped to form a The forest fire in the region has to be knowledge base in defining the tackled with the help of community parameters. The climate data are participation. important factors to be considered but it is found to play more important role during fire season. The fire incidents are varying with fuel characteristics, topography and proximity to road and settlements with each region. The forest fire vulnerability map shows that large areas in NER are prone to forest fire. Five of the seven

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 31 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad, India

35 30 Fire incidents in NER 25 20 15 10 5 0

Forest fire incidence ( %) fireForest incidence North Eastern States

Fig1 . Fire incidences (%) in Northeastern states Fig.g.2 Fire vulnerability zones of NER in India (January 2001-April 2013) Source: NRSC (January 2001-December 2010) and FSI (January 2011-April 2013)

60 Status of Fire vulnerability zones in NER 50 Very low low 40 moderate 30 high 20 very high 10 0

Percentage of(%) area Arunachal Assam Manipur Meghalaya Mizoram Nagaland Tripura Pradesh North Eastern States

Fig3. Area-wise distribution of different fire hazard zones in NER

Fig.4 Sample table format of daily forest fire hazard alerts

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 32 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-8, 2014 ISPRS Technical Commission VIII Symposium, 09 – 12 December 2014, Hyderabad, India

References Palit G.,2007,Forest cover and Baishya R., Bari S. K., and forest fires in eastern Himalaya: Upadhaya K., 2009, “Distribution A study on ecological disaster pattern of aboveground with special reference to in natural and forests Arunachal Pradesh, PhD thesis. of humid tropics in Northeast Ramakrishnan P. S., 1988, India,” Tropical E cology, vol. 50, “Sustainable development, no. 2, pp. 295–304,. climate change and tropical rain Darlong V T., 2001,Traditional forest landscape,” Climatic community-based fire Change,vol . 39, pp. 583–600. management among the Mizo Roy N. and Porwal M C.,2005, shifting cultivators of Mizoram in Forest fire risk zonation using northeast India, FAO corporate geo-spatial modeling in part of document repository. Rajaji National park, India, Asian FSI, State of Forest Report, Journal of Geoinformatics, Vol.5, 2011.Forest Survey of India, No.3, p-40-46 , India. Sowmya S V. and Jaiswal R.K., Mukherjee S. , Somashekar,2010, Application of Raju K.D. , Saxena R. 2002, Remote sensing and Forest Fire Risk Zone Mapping geographical information system through Satellite Imagery & inmapping forest fire zone at Geographical Information Bhadra wildlife sanctuary, India, System. International Journal of Journal of Environmental Biology, Applied Earth Observation and 2010, p-969-974 Geoinformation, Elsevier http://bhuvan- Publications, 4:1 -10. noeda.nrsc.gov.in/disaster/disast Majumder M., Shukla A. K., and er/disaster.php Arunachalam A., 2011 http://www.fsi.nic.in/forest- “Agricultural practices in fire.php Northeast India and options for sustainable ACKNOWLEDGEMENT management.Biodiversity, biofuels, and The authors are thankful to North Eastern Regional node for Disaster Risk Reduction (NER-DRR) conservation agriculture,” program of NESAC sponsored by Dept. of Space, SustainableAgriculture Reviews, ISRO HQ, Banglore under which the study has been vol. 5, pp. 287–315, 10.1007/978- carried out. We are also thankful to the reviewers for valuable suggestions. 90481-9513-8-10.

This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-8-27-2014 33