Content uploaded by Miltiadis Athanasiou
Author content
All content in this area was uploaded by Miltiadis Athanasiou on Nov 02, 2022
Content may be subject to copyright.
Edited by
DOMINGOS XAVIER VIEGAS
LUÍS MÁRIO RIBEIRO
2022
Advances in Forest Fire Research 2022 - D. X. Viegas & L.M. Ribeiro (Ed.)
Chapter 1 - Decision Support Systems and Tools
https://doi.org/10.14195/978-989-26-2298-9_57
Advances in Forest Fire Research 2022 – Page 361
Simple firefighting demand modelling and its use for estimation of the
potential influence of fuel treatment scenarios on the number of required
firetrucks on the island of Kythira, Greece
Gavriil Xanthopoulos1*; Miltiadis Athanasiou1; Vassiliki Varela2, Konstantinos Kaoukis1,
Panagiotis Xanthopoulos1
1*Hellenic Agricultural Organization “Demeter”, Institute of Mediterranean Forest Ecosystems.
Terma Alkmanos, 11528, Athens, Greece, {gxnrtc@fria.gr, info@m-athanasiou.gr, kako@fria.gr,
panosxant@hotmail.com}
2 Center for Security Studies (KEMEA). 4, P. Kanellopoulou str., 101 77, Αthens, Greece,
{v.varela@kemea-research.gr}
*Corresponding author
Keywords
forest fire, fire suppression, firefighting demand, fuel management
Abstract
The island of Kythira, in Greece, suffered a devastating forest fire that started on August 4, 2017. After that, it became
evident that the location of the island, away from aerial fire suppression resources bases and with limited capacity for
quick arrival of significant ground firefighting reinforcements, necessitates careful fire prevention and presuppression
planning to avoid repetition of the disaster. The study presented here aimed to examine the adequacy of the available 13
firetrucks on the island to successfully carry out initial attack under similar conditions to those of 2017, and to evaluate
what could be the effect of four alternative fuel treatment scenarios on reducing the potential of a future disaster. A map
of the forest fuels on the island, a weather scenario similar to the conditions at the start of the 2017 fire, and the Digital
Elevation Model (DEM) of the island were used with a fire spread simulator (G-FMIS) first to simulate the actual fire
and to examine if it matches the observed fire spread in 2017. Once good agreement was verified, four fuel treatment
scenarios were applied on the fuels. The accordingly adjusted fuel map was used for further simulations. The resulting
fire perimeter growth, taking flame length into consideration, was examined against the capacity of ground forces
(firetrucks) to control lengths of the perimeter using a simple but effective fire suppression model, that is based on an
equation developed earlier for assessment of the effectiveness of such forces in Greece. The results showed that under
broadcast grazing on the island the risk of escaped fires can be minimized, reducing the need for heavy aerial support in
case of a fire.
1. Introduction
In Greece, the island of Kythira, which lies south of Peloponnese, suffered a devastating forest fire that started
on August 4, 2017, and spread actively for three days. It finally burned 2,471 ha (8.91% of the island) creating
significant problems to the environment and the economy of Kythira including its important tourism sector. In
the frame of a post-fire effort to reduce the chance of repetition of such a disaster (Xanthopoulos et al. 2022) it
was needed to estimate the requirement of firefighting resources for suppression of potential fires and compare
it with those currently available on the island. The island’s location, away from the bases of aerial resources,
and with a long time requirement (4-6 hours) before significant ground reinforcements can arrive by boat, means
that the initial attack must be handled by the locally stationed 13 firetrucks with their crews. An analysis of the
firefighting requirement was identified as important for highlighting the need for effective fire prevention to the
local authorities and population, both in regard to mitigation of fire starts and to fuel treatment for biomass
reduction. Furthermore, it can demonstrate the importance of quick and strong mobilization of resources in case
of a fire start under weather conditions that could lead to a failure of initial attack and repetition of the 2017
disaster. The effort for estimation of the firefighting requirement, using fire spread simulation both with the
actual fuels and those under different fuel management scenarios, is described here. The study uses the area of
the August 4, 2017 burn as an example, because it spans the most typical fuel types on the island.
Advances in Forest Fire Research 2022 - D. X. Viegas & L.M. Ribeiro (Ed.)
Chapter 1 - Decision Support Systems and Tools
https://doi.org/10.14195/978-989-26-2298-9_57
Advances in Forest Fire Research 2022 – Page 362
2. Methodology
In the frame of the post-fire fire prevention effort on Kythira, a forest fuels map was developed for the island
(Xanthopoulos et al. 2022). The fire behavior and the growth of the August 2017 fire was documented in detail
and it was possible to reconstruct fire evolution with good accuracy. Next, the G-FMIS fire spread simulator,
was used for simulating fire spread (Xanthopoulos et al. 2022). Inputs included a weather scenario similar to
the conditions at the start of the 2017 fire, the forest fuels map and the DEM of Kythira. The outcome was a
realistic simulation, matching well the true evolution of the fire (Xanthopoulos et al. 2022). Once good
agreement was verified, G-FMIS was used to examine the effect on fire spread, that four potential fuel
treatments could have if they had been applied, prior to the fire event, to the area that burned. The four scenarios
were:
• Scenario 1: Mechanical treatment (tractor) only in agricultural areas
• Scenario 2: Mechanical treatment (hand tools) only in agricultural areas
• Scenario 3: Grazing everywhere (in all types of vegetation)
• Scenario 4: Intense grazing everywhere (in all types of vegetation)
Modified post-treatment fuel models were estimated for the four scenarios, based on scientific literature review,
the forest fuel map was adapted accordingly, and new fire spread simulations, using the same weather conditions
as those at the time of the 2017 fire, tested the effect of the fuel treatment on fire perimeter growth and flame
length under the four scenarios. Fire spread was simulated for six hours in time steps of 1 hour. The fire
perimeter of the simulated fires under the four scenarios, for each time step, were traced manually and the
perimeter length for each of five classes of flame length (m) was calculated. The classes were a) 0-1.5 m, b)
1.5-2.5 m, c) 2.5-3.5 m d) 3.5-10 m e) >10 m.
The effect of the fuel treatments on the firefighting demand (firetrucks) was assessed using a published formula
for calculating the length of the flank of a fire that can be extinguished by a firetruck having a water carrying
capacity of 2500 l, as a function of flame length (Simos and Xanthopoulos 2014). The formula is as follows:
EXT2500L_Flank = 20.756 + 57.493 / FLflank
Where the extinction length EXT2500L_Flank and the flank flame length FLflank are expressed in meters.
Table 1 is based on the equation above.
Table 1- Length of the fire perimeter of a fire (EXT2500L_Flank ) that can be extinguished by a 2500 l firetruck, as a
function of flame length
Flame length class
FL value used in simulation
Extinction of perimeter (m)
per 2500l firetruck load
1: up to 1.5 m
1.2
68.7
2: up to 2.5 m
2.2
46.9
3. up to 3.5 m
3.5
37.2
4. up to 10 m
10.0
26.5
5. more than 10m
20.0
23.6
In order to develop a more realistic firefighting requirement estimation, the reduction in the effectiveness of the
firetrucks due to the need for refilling with water needs to be taken into consideration, as well as the average
time for emptying the load of a firetruck to the fire (Joerscke 1999). Assuming a typical average distance of 4
km to water sources at Kythera, a fire truck travel speed of 50 km h-1 on the winding and narrow road network
of the island, and a time of 15 minutes for using-up one water load on the fireline, the truckloads that can be
achieved by a firetruck per hour were calculated as shown in Table 2.
Advances in Forest Fire Research 2022 - D. X. Viegas & L.M. Ribeiro (Ed.)
Chapter 1 - Decision Support Systems and Tools
https://doi.org/10.14195/978-989-26-2298-9_57
Advances in Forest Fire Research 2022 – Page 363
Table 2- Estimated fight and reload rounds per hour (LOADS_PER_HOUR) that can be achieved by a firetruck with
a 2500 l water carrying capacity on Kythira island.
Average distance (km)
ADIST
4.00
Firetruck speed (km/h)
SPEED
50.00
Water loading time with delays
WLT
5.00
Travel time (min)
TRAVEL=WLT+((2*ADIST/SPEED)*60)
14.60
Time for using-up one water load (min)
WUT
15.00
Total time needed per truckload for each round (min)
TT=TRAVEL+WUT
29.60
Fight & reload rounds per hour
LOADS_PER_HOUR=60/TT
2.03
The required firetruck loads for extinguishing a length of the perimeter of each flame length class at each time
step, was estimated by dividing the length of the perimeter by the corresponding EXT2500L_Flank for that
flame length. The number of hours for each time step (1 to 6 hours), divided by the LOADS_PER_HOUR,
provided an estimate of the truckloads that can be delivered by a firetruck in this step's hours. Thus, the required
number of firetrucks to deliver the needed truckloads at each time step was calculated by dividing the required
firetruck loads for extinguishing the perimeter until then, with the truckloads that can be delivered by a firetruck.
3. Results
Figure 1 shows the simulated growth of the August 4, 2017 fire, in hourly steps, indicating through different
colours the flame length along the perimeter. The flame length classes correspond to the broadly accepted limits
for firefighting (Tedim et al. 2018). It is noted that at no point across the perimeter, a flame length over 10 m is
predicted. Figure 2 illustrates the effect that the four fuel treatment scenarios would have on the burned area
and flame length along the perimeter
Figure 1- Simulation of the August 4, 2017 fire on Kythira based on the existing fuels at the time, showing the flame
length classes along the perimeter at each (hourly) simulation step.
Figure 3 presents the growth of the perimeter with time along the 6-hour simulation, including the simulation
with the “real” fuel situation (i.e. the fuels that actually burned in the 2017 fire) and the simulations with the
four fuel treatment scenarios. The significant effect of grazing (treatments 3 and 4) is obvious. On the other
hand, treatments 1 and 2 seemingly have little effect. This is explained by the relatively small percentage
(7.19%) that agricultural areas that receive these treatments, occupy within the specific burned area.
Advances in Forest Fire Research 2022 - D. X. Viegas & L.M. Ribeiro (Ed.)
Chapter 1 - Decision Support Systems and Tools
https://doi.org/10.14195/978-989-26-2298-9_57
Advances in Forest Fire Research 2022 – Page 364
`
Figure 2- Influence of the four fuel treatment scenarios on burned area and flame length along the perimeter
(simulation of 6 hours)
Figure 3- Fire perimeter length growth with time for the simulations with the real fuels and under four fuel treatment
scenarios
The treatments also have an effect on flame length along the perimeter, which influences the firefighting
requirements, as shown in table 4 which presents the required of firetruck loads (T) with time. As an example,
the perimeter length after 3 hours of simulation with the real fuels, indicated as PL3 in table 4, is 2271.4 m. It
consists of 1924.5 m of FL class 1 (i.e. 1.2 m), 85.5 of FL class 2 (2.2 m), 121.2 m of FL class 3 (3.5 m) and
140.2 of FL class 4 (10.0 m). Taking table 1 into consideration, the 1924.5 m of FL class 1 require 1924.5/68.7=
28.0 truckloads. The other parts of the perimeter require 1.8, 3.3, and 5.3 truckloads respectively for a total
requirement of 38.4 truckloads in 3 hours.
Advances in Forest Fire Research 2022 - D. X. Viegas & L.M. Ribeiro (Ed.)
Chapter 1 - Decision Support Systems and Tools
https://doi.org/10.14195/978-989-26-2298-9_57
Advances in Forest Fire Research 2022 – Page 365
Table 4- Estimated requirement of 2500 l firetruck loads (T) for extinguishing the fire Perimeter at each time step
(i=1-6), taking the Flame Length of all the parts of the perimeter (in classes) into consideration (PLi)(m)
Fuels
PL1
T
PL2
T
PL3
T
PL4
T
PL5
T
PL6
T
Kythira fire fuels
505.8
7.4
864.3
14.2
2271.4
38.4
4572.6
88.2
7117.3
204.0
8302.3
202.4
Scenario 1
504.3
7.3
867.4
14.1
2516.8
40.8
4118.6
76.7
6654.9
162.3
8153.4
175.9
Scenario 2
438.1
6.4
841.2
12.3
2093.8
36.2
4703.6
102.8
7084.1
183.5
8275.3
189.6
Scenario 3
394.0
5.7
875.9
12.8
1292.2
19.3
1839.0
29.0
2060.6
30.0
2629.0
41.4
Scenario 4
388.6
5.7
917.7
13.4
1292.7
19.7
2328.8
36.2
2379.6
34.7
3190.1
51.2
Continuing on the previous example, using LOADS_PER_HOUR = 2.03 from table 2, it is calculated that a
firetruck can transport and use 3 x 2.03= 6.1 truckloads in 3 hours. The number of firetrucks that can achieve
the required 38.4 truckloads, are 38.4/6.1 = 6.3 firetrucks. Figures 4 and 5, present the estimated evolution of
required truckloads as the fire spreads, and ultimately the number of the required firetrucks for the 6 hours of
the simulation.
Figure 4- Number of required truckloads of 2500 l for extinguishing the fire perimeter with time, for the five
simulations
Figure 5- Evolution of the required number of firetrucks for controlling the perimeter of the fire as a function of time,
for the five simulations
Advances in Forest Fire Research 2022 - D. X. Viegas & L.M. Ribeiro (Ed.)
Chapter 1 - Decision Support Systems and Tools
https://doi.org/10.14195/978-989-26-2298-9_57
Advances in Forest Fire Research 2022 – Page 366
4. Conclusions
In conclusion, the simple method developed here for estimating ground firefighting demand has produced some
very interesting results. It demonstrated the need for increased prevention and better presuppression planning
and the value of systematic broadcast grazing as a fuel treatment across the island. Under such fuel treatment
the 13 available firetrucks could be able to control starting fires even under difficult conditions. Aerial resources,
if dispatched, could have an auxiliary role. The concept presented here can be used for similar analyses
elsewhere.
5. Acknowledgements
The study was supported by the project “PREVAIL PREVention Action Increases Large fire response
preparedness” (826400 — PREVAIL — UCPM-2018-PP-AG), funded by the European Union Humanitarian
Aid and Civil Protection (DG-ECHO).
6. References
Joerscke, J. D. (1999). Pumping Apparatus: Driver Operator's Handbook. International Fire Service Training
Assn.
Simos, M., & Xanthopoulos, G. (2014). Assessment of the effectiveness of the forest fire fighting ground forces
in Greece. In proceedings of the 7th International Conference on Forest Fire Research on “Advances in
Forest Fire Research”, November 17-20, 2014. Coimbra, Portugal. Viegas D. X., (editor). ADAI/CEIF,
University of Coimbra, Portugal. pp. 665-672.
Tedim, F., Leone, V., Amraoui, M., Bouillon, C., Coughlan, M. R., Delogu, G. M., Fernandes, P., Ferreira, C.,
McCaffrey, S., McGee, T., Parente, J., Paton, D., Pereira, M., Ribeiro, L.M., Viegas, D. X., & Xanthopoulos,
G. (2018). Defining extreme wildfire events: difficulties, challenges, and impacts. Fire, 1(1), 9.
Xanthopoulos, G., Athanasiou, M., Nikiforaki, A., Kaoukis, K., Mantakas, G., Xanthopoulos, P., Papoutsakis,
C., Christopoulou, A., Sofronas, S., Gletsos, M., & Varela, V. (2022). Innovative Action for Forest Fire
Prevention in Kythira Island, Greece, through Mobilization and Cooperation of the Population: Methodology
and Challenges. Sustainability, 14(2), 594.
