Correlation between wildfire statistical datas, weather parameters, and climate

The object of this oral presentation is to show the inter-correlations existing between statistics of wildfires (occurrences: N, areas burned: A) and climatic parameters (precipitation: P, temperature: T).We will then show you how obtained correlations can be representative of the climate of the represented regions.

To this purpose, statistics of wildfires have been studied in several regions of the world, focusing on temperature and precipitation. The present analysis has been performed on French Mediterranean Departments (Bouches-du-Rhône, Hérault, and Var), Northern Ontario (Canada), Alicante Region (Spain), Yellowstone National Park –YNP-(USA), San Diego (California), Savana Forest (Senegal). Concerning the temperature, the monthly analysis of fire occurrence shows two distinct periods with different fire production laws, both being of the Arrhenius type: the first one corresponding to the vegetation during the rest period, the second one during the vegetative production period. In this last period, the rate of fire production is controlled by photosynthetic step (a limiting step of plant production?), the activation energy being comprised between PAR blue radiation (257kj/mol -> 4639 Å, Bouches-du-Rhône case) and PAR red radiation (186 KJ/mol -> 6410 Å, YNP Case).

Annual Wildfires in Yellowstone National Park have been extensively analysed as much on area burnt as on wildfire occurrences, and as much on lightening caused fire as on human ones (Douglas 1975 - Romme 1989 - Despain 1998) – (Bernard 1998: interpretation of fire intensity in Yellowstone National Park for the 1700-1990 period, 1998). Most of the wildfires in Y.N.P. are class 3 fires. We observe the same Arrhenius law with the lightening fires on the annual occurrence criterion. For the Human caused fires, the Arrhenius law is verified for both annual occurrences and area burnt criterion. Thus, we can conclude that the rate of photosynthetic radiations regulates the monthly occurrence of fires during the vegetation production period and yearly occurrences of fires and area burnt at the inter-annual scale.

Concerning the precipitation, the monthly analysis of fire occurrence led to two distinct laws depending on the region climatic type (or eventually the seasonal characteristics). Thus, the monthly variation of fire occurrence (in log scale) is a linear increasing or decreasing function according to precipitation. In French Mediterranean Regions we have shown that these expressions are bound to the rate of soil water extraction by the plant roots (a limiting step of plant production?). This point has been carefully analysed in the Var Department, not only on a monthly scale, but also on an – June n) is expressed by the equation:

Log N(annual) = 2.90 – 0.000365 P(annual) R2 = 0.98

Moreover, we will show you how volumes of burnt wood may be concurrently used to occurrence of fires and burnt area so as to obtain the preceding fundamental relations.

Wildfire and Climate:

During the vegetation period we define the monthly occurrence of fires as function of variable P and T.

Log N = a + b * P Log N = c - d / T

with a : a > 0, b : b > 0 (NW Ontario Climate) b < 0 (Mediterranean type Climate)

In the equation 1, N is univocal of P, in the number 2, N is univocal of T. P and T are measured data or data taken from meteorological stations and are inter-dependant in the vegetation production period. Identifying the 2 equations, we get:

P = (c-a) / b + d / (b *T) (1)

We now need to expose this relation (1) to real temperature and precipitation data of different regions.

Precipitation and temperature data of meteorological stations in N.W. Ontario and Languedoc-Roussillon regions have been tested and verified with accuracy the equation (1) as a method of reconstruction of climate in the regions. That leads us to the concept of climatologic chart.

Other reconstructions of climate from wildfire occurrence have been stated. This reconstruction has been extended to volumes of destroyed wood (m3/ha) in Senegal savana forest. That confirms the generality of the method. So we introduced the concept of climatologic chart which is based on the law of interaction between the vegetation, the soil, the water, the atmosphere and the solar radiations. Climatologic charts could be more useful than historic representations of climate characteristics (climatogram, Emberger or Gaussen ombrothermal diagrams) to build a new classification of climate,and in order to perceive any possible climate changes.

Reference: Wildfire, weather and productivity (M.L.Bernard, N.Nimour) In: Butler, Bret W.; Cook, Wayne, comps. 2007. The fire environment--innovations, management, and policy: conference proceedings. 26-30 March 2007; Destin, FL. Proceedings RMRS-P-XX. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. XXX p. CD-ROM and DVD.