| Третья всероссийская открытая конференция
«Современные проблемы дистанционного зондирования Земли из космоса»
Москва, ИКИ РАН, 14-17 ноября 2005 г.
Сборник тезисов конференции
Remote Sensing of Fire Intensity and Burn Severity in Forests of Central Siberia
|Anatoly I. Sukhinin*, Susan G. Conard**, Douglas J. McRae***, Galina A. Ivanova*, Peter A. Tsvetkov*, Viacheslav A. Bychkov* and Olga A. Slinkina*|
|*Russian Academy of Sciences, V.N. Sukachev Forest Institute;
**USDA Forest Service, Washington, DC, USA;
***Natural Resources Canada, Canadian Forest Service, Sault Ste. Marie, Canada.
|Forests in central Siberia have a mixed-severity fire regime consisting of primarily low- to high-severity surface fires. Less frequent is the occurrence of crown fires. Increasingly sophisticated and comprehensive estimates of burned areas and of active fire areas are being produced using satellite imagery. However, accurate regional estimates of fire intensity, ecosystem fire effects, and fire emissions to the atmosphere require development of accurate remote-sensing based methods for estimating the intensity of active fires and the severity of fire impacts. Average fire return intervals are 25-45 years, and the bulk of fires are low- to moderate-severity surface fires. However, surface fires severe enough to kill large patches of overstory trees are a frequent occurrence, and in drought years, up to 50 percent of the burned area may burn in stand-replacement fires (high-severity surface fires or crown fires).
Validation data of active fires will be important in improving these estimates. As part of the FIRE BEAR project, multi-level research is being conducted to link ground measurements of fire intensity and severity on experimental fires and wildfires to aircraft and satellite observations of active fires and fire scars. Field studies were carried out on a series of experimental burns and wildfires of varying intensities, spread rates, and severities in the Krasnoyarsk Region during the 2000-2003 field seasons. Studies of flame height, stem and crown scorch, and insect damage are being combined with fire behavior and environmental data to provide a better basis for estimates of the effects of different patterns and intensities of fire on tree mortality. On a series of experimental burns of varying fireline intensities, spread rates, and severities carried out in P. sylvestris forests on two sites in the Krasnoyarsk Region, first-year mortality of canopy trees varied from 0 to 100 percent. Spatial patterns in mortality can be related directly to the effects of local variations in fire behavior as evidenced by measurements of fireline intensity, flame length, and stem bark char. Insect infestation is an additional predictor of potential for future mortality. Relationships between bark char and flame height also suggest that char height can be used as an effective proxy for flame height, which is a difficult fire behavior parameter to measure accurately.
Ground and aerial observations of forest stand structure and fire impacts showed that it is possible to estimate fire intensity and the type of fire (e.g., surface or crown fire) both from aircraft and from NOAA/AVHRR, BIRD, Terra/ASTER and other satellites. Ground observations from experimental burns and wildfires were then linked with intermediate-resolution satellite data (Landsat ETM) to help estimate fire severity, estimate emissions and carbon balance, monitor ecosystem effects and postfire recovery, and validate burned area estimates from AVHRR and MODIS. Further development and testing of these relationships will enable accurate, broad scale mapping of the intensity, severity, and extent of wildland fire in central Siberia and other regions. Earlier studies in the region have shown that it can take as long as three to four years for the full extent of mortality to be realized. Because of this fact, it is our plan to monitor survival on our sites for several years after the fires. Nonetheless preliminary data suggest that even first-year relationships can be useful in making early predictions of the effects of wildfires on tree survival and ecosystem recovery.