Examining the Road Networks and Locations of Firefighting Teams by Using GIS Techniques 8. ЛІСОВА ІНЖЕНЕРІЯ : ТЕХНІКА , ТЕХНОЛОГІЯ , ДОВКІЛЛЯ

are located. There are three ﬁ re ﬁ ghting teams located in the boundaries of the study area. The sites in the study area where previously occurred forest ﬁ res (15), which burned 1 ha or more forest areas, were evaluated as potential ﬁ re sites. The analysis results showed that 64,12% of the forest areas in the study area was reached by the ﬁ re ﬁ ghting teams within 20 minutes, which is the critical response time for ﬁ rst degree ﬁ re sensitive forests. It was found that the teams could reach 12 potential ﬁ re sites within the critical response time. This result revealed the necessity to establish new ﬁ re ﬁ ghting teams in the study area. In addition, it is thought that improving the road network density in the study area by building new roads or increasing the truck travel speed by improving the conditions of existing roads will help to solve the problem

of the Mediterranean region are fi re sensitive at the fi rst degree (Akay, Wing, Zengin, & Köse, 2017). In order to combat forest fi res effectively, fi re must be suppressed as soon as possible. For this reason, it is of great importance to identify detect and locate forest fi res as soon as they start and to report them to the fi refi ghting teams without delay (Akay, Karaş, & Kahraman, 2018).
In order to combat forest fi res effectively, the time to reach the fi re site, especially in the fi rst degree fi re sensitive forest areas, should not exceed the critical response time in which the fi re is more likely to be brought under control. For this reason, the fi refi ghting teams should be placed in a suitable location from which most of forest area can be reached in critical response time (Sakar, 2010). After the fi re alarm is received, the optimum route that allows the fi refi ghting team to reach the fi re site in the shortest time should be determined as soon as possible.
The fi re sensitivity rate of a zone is determined based on the number of fi res in that zone, the ratio of the burning area to the forest area of the forest enterprise and the fi re constant (Erdoğan, 2019). In order to combat forest fi res effectively, the concept of fi re sensitivity has been developed to rank the sensitivity of the forest enterprise and to show the status of a forest enterprise with a known sensitivity rate compared to other forest enterprises. The most important factor in determining the degree of sensitivity is the archive information about the forest fi res that occur (Küçük & Ünal, 2005).
The degree of fi re sensitivity, which represents enterprises with a certain degree of sensitivity, actually constitutes the fi re sensitivity class. The degree of fi re sensitivity varies according to the size of burning area and the number of fi res per unit area. As the number of annual fi res in the burning area and unit area decreases, the sensitivity decreases, and as it increases, the sensitivity to the fi re increases (Akay et al., 2017). Fire constant values are used to determine the fi re sensitivity levels of the forest enterprises. Figure 1 shows the fi re sensitivity map of the forest enterprise directorates in Turkey.
In order to effectively fi ght forest fi res, the fi refi ghting team must reach the fi re site within the critical response time. Table 1 indicates the critical response times according to the degree of fi re sensitivity (GDF, 2008). For effective fi ght against forest fi res, the arrival time of the fi refi ghting team should not exceed these time limitations. In this study, the locations of the existing road networks and the fi refi ghting team in a sample forest enterprise were evaluated using GIS techniques in order to determine the optimum route that will provide the fastest access to the fi re site. Material and Methods. 1. Study Area. The study was carried out within the boundaries of the Adana Forestry Regional Directorate (FRD), Adana Forestry Enterprise Directorate (FED), Sarıçam Enterprise Chief (FEC) (Fig. 2). The tree species that dominate the area are Brutian pine, Stone pine and other coniferous species. The area of the Sarıçam FEC is approximately 70,000 hectares, and approximately 9,000 hectares are covered with forests. There are three fi refi ghting teams within the boundaries of the study area. The sites in the study area where previously occurred forest fi res (15), which burned 1 ha or more forest areas, were evaluated as fi re-prone sites.
2. GIS Database. Within the scope of the study, GIS database was produced to apply GIS techniques in ArcGIS 10.4 program. To generate this database, road network map, forest stand map and location information about potential forest fi res and fi refi ghting teams were used (Fig. 3). Information about the fi res that occurred at the working area boundaries and where the burned area is 1 ha and above is given in Tab. 2. A data layer showing forest areas was produced using the stand map to evaluate the transportation of the fi refi ghting teams to the forest areas in the study area.  3. Network Analysis. Network analysis method is widely used in the solution of transportation problems requiring the optimum route (Chiang et al., 2016). In the network analysis method, links (arc) and nodes where the links intersect form a network system. In order to determine the optimum route that provides access to the fi re site in the shortest time possible, the average transportation time of fi re truck to be spent on each road section was determined. Transportation time was calculated based on the length of the road and the average speed of the fi re truck. The average vehicle speed varies depending on the type of road (asphalt, gravel, forest road) and condition (good, average, poor). The average vehicle speed assigned for each road section according to the type and condition of the road is given in Tab. 3 (Bilici, 2009).
Network Analyst Application. The advances in GIS and computer technology enable the use of network analysis method within GIS software for solving transportation problems (Yıldırım & Bediroğlu, 2019). In this study, Network Analysis (Network Analyst) plugin available in ArcGIS 10.4 program was used to determine the optimum route that will provide the fastest access to a possible fi re site from fi refi ghting headquarter.
In order to apply the network analysis method with the Network Analysis plugin, a Personal Geodatabase was fi rst produced in the ArcCatalog module of ArcGIS 10.4 program. Later, Network Data (Network Dataset) was developed by making use of the road network data layer containing the values of the links (travel time) that constitute the road network in the study area. Finally, link (ND_Edges) and node (ND_Junctions) fi les were produced using Network Dataset.
After completing the Network Data set, the «New Service Area» and «New Closest Facility» methods under the Network Analysis plugin were applied. Firstly, how much of the forest areas on the site can be reached during the critical response period was evaluated by the New Service Area method. In the New Service Area method, which is similar to the Buffer Analysis (Buffer) method, a service point determined on the network system is accepted as the start and the regions within a total link value (access time) determined by the user are determined on the network system. In this study, with the help of the «New Service Area» method, it is aimed to determine the forest areas that can be reached within 20 minutes, with critical response times for the fi rst degree fi re sensitive forests, by centering the locations where the fi rst fi refi ghting teams are located. Using the «New Nearest Facility» method, it was aimed to determine the optimum route between the potential fi re areas in the study area and the fi rst fi refi ghting teams. In the study, the fi rst fi refi ghting teams that could reach each fi re site as soon as possible were determined. In addition, optimum routes that will allow all the fi rst fi refi ghting teams in the study area to reach each fi re site have been determined.
Results and Discussion. 1. GIS Database Results. According to the results obtained, the total length of the road network in the study area was calculated as 592.00 km. A large part of these roads is forest road (57,57%), followed by gravel road (22,47%) and asphalt paved road (19,96%) (Tab. 4). Considering the condition of road networks in the study area, 65,28% of roads are classifi ed as good, 26,74% as average and 7,98% as poor (Fig. 4). All asphalt paved roads are considered to be in good condition in terms of traffi c fl ow. When the gravel roads are taken into consideration, it is determined that a large part of the roads are in good condition (93%), 5% are in the average and the remaining roads are in poor condition (2%). As to the forest roads, 45% of them are average, 42% -good, and 13% -poor.

Fig. 4. Road network map of the study area
According to the land use type map developed for the study area, a total of 10 different land use classes have been identifi ed (Fig. 5). It was found that the type of land use with the largest area was agriculture (71,90%) followed by forest areas (16,39%). Burned forest areas, forest soil and forest depots are considered under forest areas, total forest areas were calculated as 17,32% of the total area. The spatial distribution of land use types can be seen in Tab. 5.

Network Analysis
Results. Network analysis application was carried out by using two basic data layers produced in Network Database. These are data layers; the link data layer (Network_ND_Edges) representing the road sections on the road network and the node data layer (Network_ND_Junction) representing the points where these links intersect (Fig. 6). Using the «New Service Area» method, the locations of the fi refi ghting teams were centered and the forest areas that can be reached in the study area within the critical response times were determined. The study analyzes fi re response time for 3 fi refi ghting teams to reach 15 fi re sites. Table 6 shows that the fi refi ghting teams were not able to reach all 15 fi re sites within the optimal period of time (no more than 20 minutes).  Table 6 Fire response time (min) of fi refi ghting teams to reach each fi re site Team Name  Fires  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  Y_Ekip_1  23  13  25  10  17  21  10  16  26  23  17  22  8  19  6  Y_Ekip_2  49  38  23  35  28  47  19  32  52  20  43  12  27  44  27  Y_Ekip_3  5  34  31  31  32  42  31  22  47  38  38  39  29  13  27 Since forests are fi rst degree fi re sensitive in the study area, areas which fi refi ghting teams can reach on the road network within 0-20 minutes have been identifi ed (Fig. 7). Later, forest areas that could be reached within 0-20 minutes were found. The results showed that 64,11% of forest areas in the study area were reached by fi re crews within 20 minutes. The fi refi ghting teams that reach potential fi re site in the shortest time were identifi ed by using «New Closest Facility» method. Figure 8 shows the optimum routes that provide access to each fi re site in the shortest time. The optimum routes that allow all the fi refi ghting teams in the study area to reach each fi re site were determined (Tab. 7). It was found that the teams responded 12 fi res within the critical response time (20 minutes) while 3 fi res (3, 6 and 9) could not be reached on time (See Tab. 7). The results showed that there is a close relationship between the distance to fi re sites, road types, and the transportation time.  Conclusions. In this study, the fi refi ghting teams were evaluated by using GIS techniques in order to determine the optimum route that provides the fastest access to the fi re sites. Sarıçam FEC was selected within the borders of Adana FED. The forest areas in the Directorate are classifi ed as fi rst degree fi re sensitive forests. Within the scope of the study, three fi refi ghting teams available in the area and 15 fi re incidences happened between 2009-2019 that affected 1 ha or more forests were taken into consideration. It was determined that the existing fi refi ghting teams in the study area could not reach three of the potential fi re sites within the critical response time of 20 minutes. In the study, forest areas that the fi refi ghting teams could not reach within the critical response time were also identifi ed. According to the results, it was found that 35,99% of the forest areas could not be reached in 20 minutes. These results reveal the necessity of locating new fi refi ghting teams in addition to the existing fi refi ghting ones. Moreover, it is thought that improving the road network in the study area by building new roads or increasing the travel speed by improving the conditions of existing roads will also contribute to the solution of the problem.