Modelagem do comportamento do fogo com base em experimentos laboratoriais e de campo

Abstract

This research was carried through with the purpose to develop and to adjust fire behavior prediction models in laboratory experiments and burns under pine stands as well as to propose an artificial plot assembly methodology in field to monitor fire behavior. The "Laboratório de Incêndios Florestais da Universidade Federal do Paraná", located in Curitiba County, Paraná State and the "Laboratório de Pesquisa e Controle de Incêndios" of Rio Sagrado Industrial Química Ltda enterprise, located in Quatro Barras County, Paraná State, were both sites at which experiments were performed to a) evaluate the relationship between fuel weight and land slope considering fire spread variables b) define patterns of fire behavior based on fire burns on a flat landscape, with predetermined weight and fuel depth. The "Estação Experimental de Rio Negro", located in Rio Negro County, Paraná State, was the location chosen for the development and adjustment of fire behavior models based on burns under pine stands and the "Fazenda Experimental do Canguiri", located in Pinhais County, Paraná State, was used for the development of an artificial plot assembly methodology to monitor fire behavior based on burns with known variables. For the lab experiments, fuel weight, fuel depth and slope figures were predetermined. During the burns conduction in these environments, fire behavior variables such as fire spread and flame height, temperature and relative humidity were monitored. For burns under pine stands in Rio Negro, the characterization of the area was complemented by a superficial forest fuel sampling and the burns were conducted in plots with 50.0 m². At the Canguiri farm, the development of the new methodology and burns were carried out in plots with dimensions of 2x6 meters, with fuel weight, fuel depth and further fuel arrangements predetermined and with the monitoring of wind, temperature and relative humidity. It was verified during the lab experiments that in inclinations less than 15 degrees, the fuel weight was the principal factor in the fire spread and that above 15 degrees, the slope was the principal factor in the fire spread. The second order polynomials models presented the best adjustment in the fire spread prediction according the slope, with R² above 0.76. Fire spread prediction models were adjusted according to flame height (R² = 0.8063) and fire intensity prediction according to flame height (R² = 0.8272). A model comprising of variables such as flame height, temperature and relative humidity was also adjusted, providing R² equal to 0.8436. Regarding burns under pine stands, of which its total fuel load provided an average of 32.23 Mg ha-1, the model to provide the most accurate estimate for fire spread was r = 0.104716 + 0.0124587T - 0.00400681UR + 0.0478819Vv (R² = 82.73) and for fire intensity was I = -718.028 + 1275.87hc + 8.33757UR (R² = 85.23). The burns of artificial plots produced fire intensities between 30.19 and 582.22 kcal m-1 s-1, flame height between 0.75 and 5.63 meters and fire spread between 0.19 and 3.66 m min-1. The results of this methodology allowed fire intensity prediction model adjustments based on flame height (R² = 0.7999), wind speed (R² = 0.7090) and fire spread estimate models based on flame height (R² = 0.7995), wind speed (R² = 0.7101) and relative humidity (R² = 0.5661).

This research was carried through with the purpose to develop and to adjust fire behavior prediction models in laboratory experiments and burns under pine stands as well as to propose an artificial plot assembly methodology in field to monitor fire behavior. The "Laboratório de Incêndios Florestais da Universidade Federal do Paraná", located in Curitiba County, Paraná State and the "Laboratório de Pesquisa e Controle de Incêndios" of Rio Sagrado Industrial Química Ltda enterprise, located in Quatro Barras County, Paraná State, were both sites at which experiments were performed to a) evaluate the relationship between fuel weight and land slope considering fire spread variables b) define patterns of fire behavior based on fire burns on a flat landscape, with predetermined weight and fuel depth. The "Estação Experimental de Rio Negro", located in Rio Negro County, Paraná State, was the location chosen for the development and adjustment of fire behavior models based on burns under pine stands and the "Fazenda Experimental do Canguiri", located in Pinhais County, Paraná State, was used for the development of an artificial plot assembly methodology to monitor fire behavior based on burns with known variables. For the lab experiments, fuel weight, fuel depth and slope figures were predetermined. During the burns conduction in these environments, fire behavior variables such as fire spread and flame height, temperature and relative humidity were monitored. For burns under pine stands in Rio Negro, the characterization of the area was complemented by a superficial forest fuel sampling and the burns were conducted in plots with 50.0 m². At the Canguiri farm, the development of the new methodology and burns were carried out in plots with dimensions of 2x6 meters, with fuel weight, fuel depth and further fuel arrangements predetermined and with the monitoring of wind, temperature and relative humidity. It was verified during the lab experiments that in inclinations less than 15 degrees, the fuel weight was the principal factor in the fire spread and that above 15 degrees, the slope was the principal factor in the fire spread. The second order polynomials models presented the best adjustment in the fire spread prediction according the slope, with R² above 0.76. Fire spread prediction models were adjusted according to flame height (R² = 0.8063) and fire intensity prediction according to flame height (R² = 0.8272). A model comprising of variables such as flame height, temperature and relative humidity was also adjusted, providing R² equal to 0.8436. Regarding burns under pine stands, of which its total fuel load provided an average of 32.23 Mg ha-1, the model to provide the most accurate estimate for fire spread was r = 0.104716 + 0.0124587T - 0.00400681UR + 0.0478819Vv (R² = 82.73) and for fire intensity was I = -718.028 + 1275.87hc + 8.33757UR (R² = 85.23). The burns of artificial plots produced fire intensities between 30.19 and 582.22 kcal m-1 s-1, flame height between 0.75 and 5.63 meters and fire spread between 0.19 and 3.66 m min-1. The results of this methodology allowed fire intensity prediction model adjustments based on flame height (R² = 0.7999), wind speed (R² = 0.7090) and fire spread estimate models based on flame height (R² = 0.7995), wind speed (R² = 0.7101) and relative humidity (R² = 0.5661).