Design and evaluation of helicopter landing variants for firefighting in Golestan National Park, Northeast of Iran

Document Type : Research Paper

Authors

Gorgan University

Abstract

Helicopter landing sites in proximity to the forest fire-risk zones are necessary for the delivery of supplies and fire emergency response teams. In this paper, we initially prepared forest fire risk map using Random Forest algorithm by overlaying the effective factors on fire occurring including vegetation types, physiographic, climatic and human factors. Then, three variants of natural candidate sites for helicopter landing were designed by analysis of terrain slope, site areas, canopy gap, and fire risk zone. The value of each variant was evaluated using proximity analysis. In this analysis, proximity to river, area covered by landing and time cost of response teams from landing to fire zone was estimated. The optimum variant was selected by Analytical Hierarchy Process. Based on results, it strongly recommends the use of the variant one whenever possible, since the time cost and proximity to the river were lower and the area covered was higher than other variants.

Keywords


[Research]

Design and evaluation of helicopter landing variants for firefighting in Golestan National Park, Northeast of Iran

 

A. Parsakhoo, M.A. Eshaghi*, Sh. Shataee Joybari

 

Department of Forestry, Faculty of Forest Sciences, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran.

* Corresponding author’s E-mail: amin.eshaghi@gmail.com

(Received: May. 21.2016 Accepted: Oct. 25.2016)

ABSTRACT

Helicopter landing sites in proximity to the forest fire-risk zones are necessary for the delivery of supplies and fire emergency response teams. In this paper, we initially prepared forest fire risk map using Random Forest algorithm by overlaying the effective factors on fire occurring including vegetation types, physiographic, climatic and human factors. Then, three variants of natural candidate sites for helicopter landing were designed by analysis of terrain slope, site areas, canopy gap, and fire risk zone. The value of each variant was evaluated using proximity analysis. In this analysis, proximity to river, area covered by landing and time cost of response teams from landing to fire zone was estimated. The optimum variant was selected by Analytical Hierarchy Process. Based on results, it strongly recommends the use of the variant one whenever possible, since the time cost and proximity to the river were lower and the area covered was higher than other variants.

 

Key words:Risk map; Firefighting; Helicopter landing; Time cost; Golestan National Park.


INTRODUCTION

Forest fires are frequency occurred in Golestan Province, especially in Golestan National Park Located in Alborz Mountains, the northeast of Iran. More than 52 cases of forest fire were recorded in this park in 2001 (Shataee et al. 2012). Forest fires caused air pollution and loss of ecological and socio-economic patterns (Kim et al. 2005). Therefore, it is important to prevent fires damages through the evaluation of the fire risk and prediction of fire behavior. Fire risk map refers the zones where the risk of the fire occurring and its spread is very high.

Thus, preparing this map can be useful in identifying high-risk zones and implementing firefighting plans to minimize the damages caused by fires (Kandya et al. 1998). Aerial firefighting by helicopter against forest fires is common practice throughout the world (Konishi et al. 2008).

Helicopters compete on economic and environmental terms particularly in stands with the absence of adequate road networks (Larry Mason 2005).

In a recent study Bordado & Gomes (2007) showed that firefighting polymers can be applied via helicopters. The landing site for helicopters is a level piece of ground free of trees, scrub, logs and boulders (Yu et al. 2007). Landing sites in proximity to the high-risk zones are necessary for the delivery of supplies, fire emergency response teams and construction of water pool. The proximity of helicopter landings to the river is extremely important to collect water from pools for firefighting in the forest (Stefanović et al. 2015).

For firefighting operations in unknown forest terrain, it is necessary to assess the safety and efficiency of landing sites. Landing site should be free from obstacles and vegetation with 50-75 m diameter. Besides, the slope gradient of landing should be less than 5% (Scherer et al. 2012). Ground trails are defined as extended linear features that are used for access to the fire emergency response teams from landing to fire zone (Chiou et al. 2010). The maximum walking speed of a fire emergency response team is evaluated on the base of physiological properties of terrain and it is used to find the best trail to target after landing (Knoblauch et al. 1996; Sanders & Mitchell 2000). Besides, time cost of walking on the trail is an important indicator to find the optimum route between points (Chiou et al. 2010). Dijkstra’s algorithm is often applied to find the least-cost trails (Rees 2004). In current study multi-criteria, decision- making process is used to determine the best landing variant. In this system, different alternatives of helicopter landing are identified and then the best one is selected based on the values and preferences of the decision maker. The purposes of this study were to prepare forest fire risk map and then evaluate the forest terrain conditions to find natural candidate sites for helicopter landing. Moreover, the optimum variant of landing sites was determined using GIS-based proximity analysis by estimating distances to the river, forest areas covered by helicopter flight as well as time cost of fire emergency response teams on trails from landing to the fire zone.

 

MATERIALS AND METHODS

Study site

Golestan National Park with an area of 91890 ha is located in Golestan Province. It ranges from 37° 16' 43"N to 37° 31' 35"N and 55° 43' 25"E to 56° 17' 48"E. There are 1350 plant species and 302 animal species with variety of habitats like temperate broadleaf forests, grasslands, shrubs and rocky areas in this forest. Altitude was 1000 to 1400 m above sea level (Fig. 1).

The annual rainfalls are ranging from 150 mm to 750 mm and decrease from west to east. Annual temperature is from 11.5 to 17. 5°C.Wildfire has been a constant event in Golestan National Park for a long time (Safaian et al. 2005).

 

 

 

Fig. 1. The geographical position of the study area.

 

 

Data collection and analysis

The maps of effective factors on fire occurring including vegetation types, physiographic, climatic and human factors were prepared from different sources. The fire risk modeling and zoning were done in Random Forest (RF) using 70% of the fire points as training samples.

 

The obtained fire risk map was zoned into five categories as very low-risk, low risk,

Moderate, high and very high-risk zone (Eshaghi & Shataee 2014).

Three variants of natural candidate sites for helicopter landing were designed by consideration of terrain slope, site areas, canopy gap, and fire risk zone. Digital elevation model (DEM) and triangulated irregular network (TIN) was used to extract slope data less than 5% and area with at least 50 meters in diameter (more than 2000 m2). Canopy gap was analyzed using high-resolution imagery on Google Earth.

 Suitable natural sites for landing selection were the sites without canopy cover, slope less than 5%, an area more than 2000 m2 and close to fire risk zone. Buffers with a width of 0-1000, 1000-2000, 2000-3000, 3000-4000, 4000-5000 meter as helicopter landing coverage were designed in the study area. Moreover, buffers with widths of 0-600, 600-1200, 1200-1800 & >1800 were designed at both sides of rivers.

Each landing position can be accessed from parameters such as walking distance, time required to walk the trail to fire zone and walking speed, especially in mountainous forests with uphill and downhill slopes. Time cost of walking on each trail in the study was built on the evaluation of the maximum walking speed a walker weighting 60 kg can reach in an area (Chiou et al. 2010).

Walking speed (V) is calculated using Equation 1.

Time cost (K) can be modelled as Equation 2 (Rees 2004):

 

 

(1)

 

(2)

 

Where V is walking speed in km.h-1, m is mean of slope in % (m is defined as dh.dx-1 where h was height and x was the horizontal distance), K is time cost in a minute, d is mean of walking distance in the kilometer.

The optimal helicopter landing variant based on the proximity to river, area covered by landing and time cost of response teams were identified using AHP in Expert Choice software.

 

RESULTS

Locating candidate sites for helicopter landing

In the current research, the relative frequency percent of risk classes in fire risk maps prepared in very low-risk, low risk, moderate-risk, high-risk and very high with RF algorithm were 28%, 16%, 23%, 14% & 19% (Fig. 2). In this study slope, less than 5% of area more than 2000 m2 and without canopy cover were selected as candidate sites for helicopter landing preparation (Fig. 3).


 

Fig. 2. Fire risk map.

 

 

Fig. 3. The position of candidate sites for landing preparation.


Analysis of the time cost, areas covered by landing and proximity to river

We focused on trial determination in the case of forest fire. Our goal here is to provide natural, safe, cost effective, fast and consequently optimal helicopter landing variants for managing forest fire. If the speed of the helicopter is 200 km.h-1, then the time traveled every buffer of 1000, 2000, 3000, 4000, 5000 are the 0.3, 0.6, 0.9, 1.2 & 1.5 minutes (Figs. 4 - 6). In this study, landings are the start node, while the risk grid is the end node. The trails based on the nearest distances to seven nodes were identified in GIS. Mean slope and distances of trails for each landing are shown in Tables 1 - 2, respectively.

 

 

Fig. 4. Forest area covered by helicopter landing variant 1.

 

 

Fig. 5. Forest area covered by helicopter landing variant 2.

 

 

Fig. 6. Forest area covered by helicopter landing variant 3.

 

Table 1. Mean slope of trails for each landing to center of risk grid (%).

Variant

1

2

3

4

5

6

7

Variant 1

27

29

23

24

36.5

37

15.5

Variant 2

24.5

11.8

12

25.1

26.3

41

20.5

Variant 3

24.1

15

15

25.7

37.7

24.7

20

 


A comparison between these three variants reveals that the area covered by landing variant 1 (69.92%) was much greater than that of the other variants. The difference between the landing variants is that the time cost on trails chosen for the variant 1 was much lower than the trails chosen for the other landing variants (Table 3). The distance of landing variants 2 from the river was shorter than the other landing variants (Fig. 7).

 

 

Table 2. Mean theoretical distance of each landing to center of risk grid.

Variant

1

2

3

4

5

6

7

Variant 1

3480

4850

4752

4942

4961

4915

4410

Variant 2

5111

4580

3328

4571

4820

6543

4564

Variant 3

5079

4337

4600

4589

4814

4898

5301

 

Table 3. Basic information for optimal landing variant.

Variant

Area covered by landing

 (%)

Proximity to river

(m)

Walking speed

(km.h-1)

Time cost

(minute)

Variant 1

69.92

366.7

1.77

157.7

Variant 2

66.98

228.9

1.51

190.1

Variant 3

67.20

348.9

1.50

194.8

 

Fig. 7. The distances of different helicopter landing variants from rivers.

 

 

Selection of the Best variant

Based on AHP, the greatest importance (0.42) (Rees 2004) was given to a time - cost criterion relating to spent time for fire emergency response teams to access to fire zone (Table 4).

 

 

In AHP process it was selected variant 1 as the most suitable option (Fig. 8). The overhead image of the optimum variant is shown in Fig. 9.

 

 

Table 4. Weights and ratings assigned to landing variants optimization criteria (Scherer et al. 2012).

Variant

Rating

Proximity to river (0.37)

Area covered by landing (0.21)

Time cost (0.42)

Variant 1

1

3

3

Variant 2

3

1

2

Variant 3

2

2

1

 

Fig. 8. Overall weights of different variants.

 

 

Fig. 9. The overhead image of selected variant.

 

ISCUSSION

Over the years, fire occurrence and severity in Iranian forests have been increasing and therefore more efforts are needed to detect and control forest fire. In the present study, GIS plays a vital role on mapping and predicting fire as well as analyzing alternative fire-fighting strategies, and directing those in the field (Borisov & Tsipenko 2012; Furdu et al. 2013). Five factors including physiographic, vegetation, ecological and human factors, as well as distance from the river were used as the independent input variable of RF data-mining model, which were used for predicting fire risk in Golestan National Park.

In this study, slope, less than 5 % of the area more than 2000 m2 and without canopy cover were selected as candidate sites for helicopter landing preparation.

Westcott & Cleary (1950) selected a natural clearing measuring 55 by 28 meters as a landing site for the helicopter. Scherer et al. (2012) used Light Detection and Ranging-based perception system to select landing sites and approach paths by considering factors such as plane fitting, terrain condition, load bearing capability of the contact surface, rotor clearance, ground paths and wind direction. In this study, landings are the start node, while the risk grid is the end node. The trails based on the nearest distances to seven nodes were identified in GIS. Wang et al. (2014) proposed a model and algorithm which were effective in planning routes to avoid one or more fire-affected areas. Helicopter landing sites may be adjacent to a river or a lake, a railroad, a freeway, or a highway, all of which offer the potential for multi-functional land usage (Singh et al. 2014). The method of AHP is one of the many methods of Multiple Attribute Decision Making (MADM) that can be applied when selecting the best variant of helicopter landing site. The greatest importance (0.42) was given to a time - cost criterion relating to spent time for fire emergency response teams to access to the fire zone. In AHP process variant 1 was selected as the most suitable option. The selected variant proves that the investment is economically justified.

 

CONCLUSIONS

Our approach has extended the state of the process in finding natural helicopter landing by incorporating not only forest fire risk zone but also by considering factors such as terrain slope, site area, canopy gap and ground trails. We also present results from three successful landing variants with varying pattern in proximity to the fire zone. The selection of the best landing variant was made on the basis of four criteria of proximity to river, area covered by landing, the walking speed, and time cost. Based on our findings, we strongly recommend the use of the variant 1whenever possible, since the time cost and proximity to the river were lower and the area covered was higher than other variants. Fire-fighting in Golestan National Forest Park requires creation and improvement of fire-fighting helicopters.

 

ACKNOWLEDGEMENTS

Thanks to the staff and local team of Golestan National Park for their help in transportation and data collection in the experimental areas.

Bordado, JC, Gomes, FP 2007, New technologies for effective forest fire fighting. International Journal of Environmental Studies. 64: 243–251.
Borisov, IV & Tsipenko, AV 2012, Functioning simulation of Ka-32 helicopter with helicopter buckets when fire extinguishing. Science Bulletin MSTUCA, 177: 59-64.
Chiou, C, Tsai, W, Leung, Y 2010, A GIS-dynamic segmentation approach to planning travel routes on forest trail networks in Central Taiwan. Lands Urban Plan, 97: 221–228.
Eshaghi, MA & Shataee, Sh 2014, Golestan National Park Fire Risk Zoning Using Tree-Based Random Forest Algorithm. International Forest Fire Conference in Black Sea Region, 6-8 November 2014, Kastamonu, Turkey.
Furdu, I, Tomozei, C, Pandel, I 2013, Improving management of risks and natural disasters by regional GIS distributed application. Environmental Engineering and Management Journal 12: 11-16.
Kandya, AK, Kimothi, MM, Jadhav, RN, Agarwal, JP 1998, Application of GIS in identification of fire prone areas. a feasibility study in parts of Junagarh Gujrat, India. The Indian Forester. 124: 531-535.
Kim, C, Koo, K, Byun, J, Jeong, J 2005, Post‐fire effects on soil properties in red pine(pinus densiflora) stands, al Forest Science and Technology. 2005, 1, 1–7.
Knoblauch, RL, Pietrucha, MT, Nitzburg, M 1996, Field studies of pedestrian walking speed and start-up time. Transportation Research Record. 1538: 27-38.
Konishi, T, Kikugawa, H, Iwata, Y, Koseki, H, Sagae, K, Ito, A, Kato, K 2008, Aerial firefighting against urban fire. Mock-up house experiments of fire suppression by helicopters, Fire Safe Journal, 43: 363–375.
Larry, MC 2005, An Examination of the Washington Department of Natural Resources Timber Sale Program Against a Backdrop of Changing Regional Infrastructure and a Growing Forest Health Crisis. Rural Technology Initiative College of Forest Resources University of Washington, P. 2.
Rees, WG 2004, Least-cost paths in mountainous terrain. Computers & Geosciences, 30: 203–209.
Safaian, N, Shokri, M, Ahmadi, MZ, Atrachali, A, Tavili, A 2005, Fire influence on the grassland vegetation in Golestan National Park (Alborz MTS. Iran). Polish Journal of Ecology, 53: 435-443.
Sanders, JE, Mitchell, SB 2000, An accurate inexpensive system for the assessment of walking speed. Journal Prosthetic Orthotic, 12: 117-119.
Scherer, S, Chamberlain, L, Singh, S 2012, Autonomous landing at unprepared sites by a full-scale helicopter. Robotics and Autonomous Systems, 60: 1545-1562.
Shataee, Sh, Kalbi, S, Falah, A, Pelz, DR 2012, Forest attributes imputation using machine-learning methods and ASTER data: comparison of K-NN, SVR and random forest regression algorithms. International Journal of Remote Sensing, 33:6254-6280.
Singh, Y, Sharma, MP, Sharma, SD, Prawasi, R, Yadav, K, Hooda, RS 2014, Application of GIS technique to select suitable sites for erecting watch towers in forest areas of mountainous tract. International Journal of Computer Applications in Technology, 5: 462-468.
Stefanović, B, Stojnić, D, Danilović, M 2015, Multi-criteria forest road network planning in fire-prone environment: a case study in Serbia. Journal of Environmental Planning and Management, In press
Wang, Z, Zlatanova S, Moreno, A, Oosterom, P, Toro, C 2014, A data model for route planning in the case of forest fires. Computers & Geosciences, 68: 1–10.
Westcott, JC, Cleary, VP 1950, An account of the use of the helicopter in fire suppression work in the Victorian mountain forests. Australian Forestry, 14: 11-16.
Yu, Z, Nonami, K, Shin, J, Celestino, D 2007, 3D Vision Based Landing Control of a Small Scale Autonomous Helicopter. International Journal of Advanced Robotic Systems. 4: 51–56.