No abstract

The aim of this study was to analyse the effects of duff thickness and moisture content, and of soil moisture content on the transfer of heat in the soil. The experimental design used intact soil blocks with their duff layer, subjected to controlled fires of variable very low intensities of up to 100 kW m-1. The fuel on the surface was composed of needles and twigs of Pinus pinaster. The maximum temperatures measured within the fuel were of the order of 650 degrees C and were independent of the fireline intensities. For fires with fireline intensity of the order of 30 kW m-1, the presence of the duff layer reduced from 330 degrees C the temperature rise at the soil surface. Duff thickness played only a secondary role, but increasing moisture content reinforced its insulating effect, so that the temperature rise was 2.5 times less at 1 cm depth in the duff when the moisture content exceeded 70% dry weight, than when the moisture content was less than 30%. For more intense fires (> 50 kW m-1) that produced longer-lasting surface heating, duff thickness and moisture content played an important role in significantly reducing the temperature rise at the soil surface (range 140 degrees C to 28 degrees C). Because of low soil thermal conductivity, temperature attenuation with increasing depth was noticed. In the case of low intensity fires (< 30 kW m-1) in the absence of a duff layer, the maximum temperatures were reduced from 350 degrees C at the surface to 7 degrees C at 3.5 cm. The temperature rise in the soil decreased with depth according to a negative exponential relation. The rate constant of this relation was greater when the initial surface temperature and the soil moisture content were higher. For the soil studied, and under the moisture conditions encountered (between 7 and 19% of dry weight), the rate constant could be predicted with acceptable precision (r2 = 0.67), if the surface soil temperature rise and the soil moisture content were known. In these experimental fires, which were carried out when the air temperature did not exceed 20 degrees C, lethal temperatures (> 60 degrees C) were measured in the upper few centimetres of the duff layer in very low-intensity fires, and in the upper few centimetres of the soil (where nutrients are most concentrated and biological activity most intense) in the slightly more intense fires. The fire intensities were always very moderate, and of the order of magnitude df those encountered in the prescribed burns conducted on fuel-breaks of the french Mediterranean area. Their impact on the surface of the forest soil, in terms of lethal temperatures transmitted to the horizon rich in organic matter, are not negligible. In contrast, below 3 to 5 cm depth, prescribed burns, conducted under the conditions of the experiments, would not lead to significant change to nutrients or microfaunal or microfloral activity; in particular, root tips would not be subjected to heat stress sufficient to kill them.

A set of 109 laboratory fires in Pinus halepensis fuel beds (1 kg m–2) was used to test the effects of slope (0°, 10°, 20°, 30°) and fuel bed width (1, 2, 3 m) on fire behaviour variables such as rate of spread, fuel consumption, flame residence time, temperatures and flame geometry. The qualitative behaviour of the fires is also reported. The 20° and 30° upslope fires are pointed in shape and fire whirls moving along the fire flanks in the direction of the fire head are systematically observed in 30° upslope fires. Flame residence time increases with increasing slope angle, and both slope angle and fuel bed width affect rate of spread. The slope effects observed in 10° and 20° slope angles and in the narrowest fuel beds (1 and 2 m) are similar to those predicted by operational models. However, the observed slope effect at the 30° slope angle is underestimated by these models, in particular in 3 m-wide fuel beds. Flame temperatures correlate closely with dimensionless height and flame lengths correlate closely with fire line intensity. Mechanisms that could explain the different effects observed are suggested and discussed.

The aim of this study was to assess the effects on combustion characteristics, and their consequences on nutrient losses, of (1) the change in load and packing ratio of the fuel bed, and (2) the change in fuel moisture content. Eighty-one experimental burns were carried out, on a test bench in the laboratory; the fuel was composed of needles and twigs of Pinus pinaster. Two levels of fuel load an dpacking ratio (8t ha-1 needles, packing ratio of 0.040; and 16t ha-1 twigs and needles, packing ratio of 0.066) were compared at constant moisture content (6%); and four levels of moisture content(6%, 12%, 24% and 30% dry weight) were compared at constant fuel load (8t ha-1 needles). At constant moisture content, an increase in the load and packing ratio of the fuel bed led to an increase in the height of flames and in the maximum temperature 25 cm above the fuel bed, in the duration of the rise in temperatures within the fuel, and in the fireline intensity. Conversely, the rate of fire spread decreased. At constant fuel load, an increase in the moisture content of the fuel led to a decrease in the rate of fire spread, in the flame height and the maximum temperature 25 cm above the fuel bed, and in the fireline intensity. In contrast, the maximum temperatures reached within the fuel, when the flaming front was continuous, did not significantly change with varying fuel loads or fuel moisture contents. The percentage fuel consumption was always high, more than 80%, but it significantly decreased with increasing fuel load and packing ratio and with increasing moisture content. Total losses of N, S, and K significantly decreased with increasing fuel load and packing ratio, with increasing moisture content and with decreasing percentage fuel consumption. Losses in P only significantly decreased with increasing fuel load and packing ratio. Losses in Mg and Ca were not significantly affected by fuel load, moisture content. or percentage consumption. An attempt was made to separate volatile from particulate losses, based on the assumption that all the losses of Ca were in particulate form. Whereas losses in particulate form remained relatively constant, losses of nutrients in volatile form seem to have been related to the percentage fuel consumption. Even if these experimental burns were of low intensity (40 to 56 kW m-1), their impact, in terms of lethal temperatures and nutrient losses, was not negligible, particularly for N and P. The increasing fireline intensity with increasing fuel load was not accompanied by an enhancement in the proportion of nutrient losses. In the same way, the strong decrease in fireline intensity with increasing fuel moisture content led only to a slight decrease in some nutrient losses. It was through their effect on the percentage fuel consumption that fuel load or moisture content modified the nutrient losses, particularly volatile losses.

We describe emission–transmission measurements performed at different heights in a flame from a cylindrical forest fuel burner, using a camera operating in the thermal infrared (7.5–13 µm). The forest fuel burner was made of a cylindrical wire mesh basket filled with a forest fuel (Pinus pinaster needles), which was ignited at the base of the basket. Three diameters of basket were used (20, 28 and 40 cm). Heat release rates, as calculated from weighing of the basket and heat of combustion of the fuel, ranged between 50 and 170 kW and flame heights ranged between 1 and 2 m. The emission–transmission device allows the determination of the transmittance of the flame and of a radiometric temperature. We show that radiation was dominated by soot in the spectral range of the camera, but that radiation from gaseous products of the combustion was not negligible. Using the Mie theory in its Rayleigh limit, we deduced some average volume fractions of soot from the measurements, which peaked at 6.8 × 10−6 in the persistent region of the flame. Then the total extinction coefficient and the total emissivity of the flame due to soot were calculated according to a standard method. Measured transmittance, soot volume fraction, total extinction coefficient and total emissivity were found to scale with the normalised height of measurement Z, defined as the ratio of the height of measurement to the height of the flame (0.25 < Z < 1.6).