ANALYSIS OF EFFICIENCY OF FIRE DANGER INDICES IN FOREST FIRE PREDICTION<sup>1</sup>

Torres, Fillipe Tamiozzo Pereira; Lima, Gumercindo Souza; Martins, Sebastião Venâncio; Valverde, Sebastião Renato

doi:10.1590/1806-90882017000200009

Acessibilidade / Reportar erro

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

Articles • Rev. Árvore 41 (2) • 2017 • https://doi.org/10.1590/1806-90882017000200009 copy

ANALYSIS OF EFFICIENCY OF FIRE DANGER INDICES IN FOREST FIRE PREDICTION¹

ANÁLISE DA EFICIÊNCIA DE ÍNDICES DE PERIGO NA PREDIÇÃO DE INCÊNDIOS FLORESTAIS

Authorship SCIMAGO INSTITUTIONS RANKINGS

ABSTRACT

Despite the existence of different fire danger indices, the use of an inefficient index can lead to making wrong decisions on the appropriate procedures for preventing and fighting forest fires, while a trusted prediction index can help the most quantification and allocation of resources for prevention. Thereat, the objective of this study is to analyze the efficiency of Fire Weather Index (FWI), Logarithmic of Telicyn Index, Nesterov Index, cumulative indexes of precipitation - evaporation (P-EVAP) and evaporation / precipitation (EVAP/P), Monte Alegre Index (FMA) and Monte Alegre Changed Index (FMA⁺) in the prediction of forest fires for the city of Viçosa (MG). The indices were compared using the method known as Skill Score (SS) taking into account the days that the indexes pointed to the risk of events with focus fire identified by satellite images on the 01/01/2005 to 31/12/2014 period. According to the results, the Logarithm of Telicyn Index (0.53257) as the most efficient for the study area, followed by the indices EVAP/P (0.46553), P-EVAP (0.43724), Nesterov (0.40445), FWI (0.39213), FMA⁺(0.34595) and FMA (0.28982).

Keywords:
Calculation; Danger; Fire

Table 1 Restrictions of the indexes according the rain.

Index	Rain (mm)	Change in calculation
Telicyn	>2,5	Leave the sum and resume the calculation the next day, or when the rain
		ceases. On rainy days the index is zero.
EVAP/P	< 11 a 15>15	Does not enter the calculation.
		Divide EVAP/P from the previous day’s rain.
		Stop the calculation, restarting the next day or when the rain ceases. On
		rainy day EVAP/P = 0
Nesterov	≤2,02,1 a 5,05,1 a	None
	8,08,1 a 10,0>10,0	Decrease 25% in the amount of G calculated the day before and add (d.t)
		of the day.
		Decrease 50% in the value of G calculated the day before and add (d.t) of
		the day.
		Leave the previous summation and start a new calculation, that is, G = (d.t)
		of the day.
		Interrupt the calculation (G = 0), resuming the summation the next day
		or when the rain ceases.
P-EVAP		Same as Nesterov
FMA	≤2,42,5 a 4,95,0 a	None
and FMA+	9,910,0 a 12,9>12,9	Decrease 30% in FMA calculated the day before and add (100 / H) of the
		day.
		Decrease 60% in FMA calculated on the previous day and add (100 / H)
		of the day.
		Decrease 80% in FMA calculated the day before and add (100 / H) of the
		day.
		Interrupt the calculation (FMA = 0), resuming the summation the next
		day or when the rain ceases

Table 3 Fire hazard class according to the value obtained for each index.

Index	Null	Small	Average	High	Very high
Telicyn	< 2	2,1 a 3,5	3,6 a 5	5 a 15	> 15
EVAP/P	< 5	5 a 20	20,1 a 50	50,1 a 100	> 100
P-EVAP	> -5	-5 a -30	-30,1 a -55	-55,1 a -125	<-125
Nesterov	d" 300	301 a 500	501 a 1000	1001 a 4000	> 4000
FWI	0 a 1,9	2 a 4,9	5 a 8,9	9 a 16,9	> 17
FMA+	d" 3	3,1 a 8	8,1 a 14	14,1 a 24	> 24
FMA	d" 1	1,1 a 3	3,1 a 8	8,1 a 20	> 20

Table 4 Percentage of expected number of days and observed fires according to the classes of indexes.

	Null		Small		Average		High		Very high
Index	Days	Fires	Days	Fires	Days	Fires	Days	Fires	Days	Fires
Telicyn	48,2%	5,1%	10,8%	1,3%	6,3%	5,1%	21,3%	31,6%	13,4%	57,0%
EVAP/P	29,0%	1,3%	22,8%	5,1%	20,0%	13,9%	16,5%	40,5%	11,7%	39,2%
P-EVAP	30,6%	1,3%	27,3%	13,9%	12,1%	16,5%	18,3%	12,7%	11,6%	55,7%
Nesterov	35,6%	2,5%	9,1%	2,5%	15,8%	6,3%	32,3%	38,0%	7,2%	50,6%
FWI	29,1%	1,3%	14,2%	3,8%	15,9%	0,0%	23,2%	1,3%	17,7%	93,7%
FMA+	21,3%	0,0%	17,6%	5,1%	12,0%	1,3%	11,6%	1,3%	37,5%	92,4%
FMA	14,9%	0,0%	16,0%	2,5%	18,9%	2,5%	19,8%	8,9%	30,4%	86,1%

Table 5 Skill Score and Success Percentage of each index.

Index	SS	PS
Telicyn	0,53257	76,63%
EVAP/P	0,46553	73,28%
P-EVAP	0,43724	71,86%
Nesterov	0,40445	70,22%
FWI	0,39213	69,61%
FMA+	0,34595	67,30%
FMA	0,28982	64,49%

Sociedade de Investigações Florestais Universidade Federal de Viçosa, CEP: 36570-900 - Viçosa - Minas Gerais - Brazil, Tel: (55 31) 3612-3959 - Viçosa - MG - Brazil
E-mail: rarvore@sif.org.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] *Corresponding author.

Estimated	Observed		Total estimated
	Fires	No Fires	Total estimated
Fires	a	b	N2 = a + b
No Fires	c	d	N4 = c + d
Total Observed	N1 = a + c	N3 = b + d	N = a + b + c + d
Contingency table calculations
Estimated	Observed		Total estimated
Estimated	Fires	No Fires	Total estimated
Fires	a / (a + c)	b / (b + d)
No Fires	c / (a + c)	d / (b + d)
The variables for performing the calculations are:
"G" - Number of hits in the forecast. G = a + d;
"P" - Probability of having at least one event per day. P = N1 / N;
"Q" - Probability of exceeding the index limit value. Q = N 2 / N;
"H" - Expected number of hits. H = N * (1-p) * (1-q) + N * p * q;
"SS" - Skill Score. SS = (G-H) / (N-H);
"PS" -Percentage of success. PS = G / N

Brasil

Brasil

ANALYSIS OF EFFICIENCY OF FIRE DANGER INDICES IN FOREST FIRE PREDICTION1

ANÁLISE DA EFICIÊNCIA DE ÍNDICES DE PERIGO NA PREDIÇÃO DE INCÊNDIOS FLORESTAIS

ABSTRACT

ANALYSIS OF EFFICIENCY OF FIRE DANGER INDICES IN FOREST FIRE PREDICTION¹