B31E-0332
Simulation for the expansion of the wildfire with numerical weather simulation MM5
1. Background Frequent occurrence of wildfires all over the world is considered as one of major resources of greenhouse gases. For example, a lot of wildfires in Alaska occur in summer. Now, the satellites of NOAA and Terra/Aqua are watching the earth and the wildfire are detected. Of course, to detection wildfire is very important, but the influence on inhabitants is more important. Our purpose is to make the numerical simulation of the wildfire spread in the small area with numerical weather simulation MM5. We think this will be useful to help fire fighting and global environment such as the replace of CO2. 2. Numerical Wildfire Spread Simulation There are many type of the numerical simulation of wildfire spread. In our simulation, the wildfire velocity is based on the Rhothermel equation and other parts are made of the cell automata. The area of the wildfire is the uniform vegetation consisted of the boreal forest (Picea mariana). The main factor of the expansion speed is wind velocity and speed. The continuous change of the weather is simulated with regional meteorological simulation MM5. The real spread of the Boundary Fire are observed by Alaska Fire Service. In this study, we validate the simulation result with the AFS data. 3. The Simulation Results We are constructing the simulation with Boundary Fire in 2004 in central Alaska. MM5 is very useful to reconstruct or forecast the distribution of local weather. We show the examples of the results in the poster. 4. Conclusion We constructed the numerical simulation model of wildfire spread with numerical weather simulation MM5. The result of simulation is being verified by the observed data by AFS .
B31E-0333
Wildfire Detection using by Multi Dimensional Histogram in Boreal Forest
Early detection of wildfires is an issue for reduction of damage to environment and human. There are some attempts to detect wildfires by using satellite imagery, which are mainly classified into three methods: Dozier Method(1981-), Threshold Method(1986-) and Contextual Method(1994-). However, the accuracy of these methods is not enough: some commission and omission errors are included in the detected results. In addition, it is not so easy to analyze satellite imagery with high accuracy because of insufficient ground truth data. Kudoh and Hosoi (2003) developed the detection method by using three-dimensional (3D) histogram from past fire data with the NOAA-AVHRR imagery. But their method is impractical because their method depends on their handworks to pick up past fire data from huge data. Therefore, the purpose of this study is to collect fire points as hot spots efficiently from satellite imagery and to improve the method to detect wildfires with the collected data. As our method, we collect past fire data with the Alaska Fire History data obtained by the Alaska Fire Service (AFS). We select points that are expected to be wildfires, and pick up the points inside the fire area of the AFS data. Next, we make 3D histogram with the past fire data. In this study, we use Bands 1, 21 and 32 of MODIS. We calculate the likelihood to detect wildfires with the three-dimensional histogram. As our result, we select wildfires with the 3D histogram effectively. We can detect the troidally spreading wildfire. This result shows the evidence of good wildfire detection. However, the area surrounding glacier tends to rise brightness temperature. It is a false alarm. Burnt area and bare ground are sometimes indicated as false alarms, so that it is necessary to improve this method. Additionally, we are trying various combinations of MODIS bands as the better method to detect wildfire effectively. So as to adjust our method in another area, we are applying our method to tropical forest in Kalimantan, Indonesia and around Chiang Mai, Thailand. But the ground truth data in these areas is lesser than the one in Alaska. Our method needs lots of accurate observed data to make multi-dimensional histogram in the same area. In this study, we can show the system to select wildfire data efficiently from satellite imagery. Furthermore, the development of multi-dimensional histogram from past fire data makes it possible to detect wildfires accurately.
B31E-0334
A study of algorithm to detect wildfire with edge of smoke plumes
Recent years, huge wildfires occur in many part of the world. And some researches have proceeded to improve wildfire detection with satellite imagery. Dozier (1981) developed the method that detects hotspot pixel by comparing the pixel with adjacent pixels. After that, Threshold method based on Dozier's approach and Contextual Method using relationship among neighbor pixels were appeared. But each of these algorithms needs more improvement in accuracy. In this study, we formulate a new algorithm with the edges of smoke plumes based on the rule of fire pixels match the origin of smoke plumes, and validate with the truth data. In this algorithm, MODIS band 1 (visible red) is extracted and smoke plumes are accented by histogram stretching. The edges of smoke plumes are extracted. Edge pixels that consist of fire smoke plumes are approximated by least squares method. Finally, the origins of the smoke plumes are determined and fire pixels are detected by the threshold approach. Our method, however, contain a problem that hotspot area shapes often a rectangle under the condition of not so high threshold temperature. In the results of this algorithm applied, it is found that it is easy to detect fire when clouds are not so thick and when smoke shape is visible clearly. On the other hand, false alarms along are detected along coast line and at the high refraction areas on a glacier, cirrocumulus clouds and so on. In addition, excessive detections increase in the low latitude because brightness temperature is raised by sunlight reflection. The wildfires in Alaska were detected well with our method. To validate this result, it is compared with the observational data and the common detection method. The Alaska Fire History Data (AFHD) is observed by Alaska Fire Service frequently, and the AFHD is offered as GIS data. On the other hand, MOD14 is one of the most famous and common methods to detect wildfire. It is calculated easily by MODIS data. Its accuracy rate to detect fire is high. So our results are compared with the observed AFHD and the MOD14 results. We compared the ratios of the accurate detections with both methods. Because our detected resolution is 250m and the resolution of MOD14 is 1000m, the pixels detected as wildfires with our method are almost four times of the ones with MOD14. Although they are quite different methods, the accurate ratios are similar: the correct detection with MOD14 is 87.2% and ours is 78.4%. Our method can detect some more wildfires that MOD14 cannot detect. In this result, our method must be able to supply the MOD14.
B31E-0335
An improvement and validation of wild fire detection algorithm of wild fire using 2- dimensional stochastic model.
Performance of algorithm to detect wild fire was remarkably improved as switching from AVHRR to MODIS
with MOD14 algorithm. However, we still have many false alarm and omission errors in boreal forest fire and
tundra fire. One of the reasons is that algorithm is that the essence of fire detection is defined as 1
dimensional stochastic test. However, the variable for the stochastic test is not efficiently chosen. Therefore,
we will propose a improved algorithm modified from MOD14 algorithm and validate wild fire detection
algorithms for boreal forest. To improve the algorithm, we used stochastic test based on 2-dimensional
distribution. To validate the wild fire detection algorithm, hotspot pixel perimeters dataset is compared with
observed wild fire by pilots of passenger flights. This wild fire cooperative observation was established in
2003 and observation method has been improved year by year. Based on comparison, author found one of
bottle necks of wild fire detection as 1 dimensional contextual threshold. Therefore, author modified MOD14
using stochastic test based on 2-dimensional distribution. As a result of this improvement, authors found
that the proposed algorithm detects 16% more hotspots without any more false alarms comparing to existing
MOD14 algorithm in preliminary validation. This also means 14% less false alarm rate comparing to existing
MOD14. More precise validation result will be presented. The proposed algorithm is operationally used in
IJIS fire monitor website.
http://www.ijis.iarc.uaf.edu/cgi-bin/fire-monitor.cgi
B31E-0336
Human amplification of drought-related biomass burning in Indonesia since 1960
Biomass burning in Indonesia is a singularly large source of greenhouse gas emissions at a global scale, with
pronounced regional impacts on air quality. Although some fire events have been documented on a case-
by-case basis, no continuous record exists prior to 1996, due to the absence of satellite estimates or ground-
truthed records of fire extent. Here, we provide a continuous record of severe haze in Indonesia from 1960 to
2006 using the visibility reported at airports, which was found to be an excellent proxy for particulate matter
emissions.
We used the visibility proxy to show that the haze events in Indonesia were worse, by a factor of five, than
extreme periods in cities with the world's worst air quality, and to better understand the underlying climatic
and anthropogenic causes of the fire. Large fire events have occurred in Sumatra at least since the 1960s,
but in Kalimantan only since the 1980s, despite the occurrence of several severe droughts during 1960-
1980. This difference can be attributed to different patterns of deforestation and population growth, which
intensified in Kalimantan only in the 1980s during Indonesia's official program of transmigration.
In the presence of intensive land use, there is a non-linear relationship between rainfall and fire, whereby fire
events occur only during drought years when rainfall falls below a certain threshold, which we estimated
using change-point analysis. Whereas recent fire events have been linked to exclusively El Niño, our long-
term record suggests that the Indian Ocean Dipole is equally as, if not more, important a factor. Better
understanding of these controls may help to assess future fire risk in Indonesia, which recent studies suggest
could increase due to reduced precipitation and accelerated deforestation.
http://www.atmosp.physics.utoronto.ca/people/rfield/index_files/IndonFire.AGU.pdf
B31E-0337
Prognostic Fire Algorithm in CLM-CN
Fire is an important Earth System process. It constitutes one of the most important feedbacks between global carbon and the hydrological cycle. Despite of the importance of fire in the climate system it is yet not well understood nor it is well represented within coupled carbon-climate models used to study climate change. To improve our understanding of this feedback linked to changes in the hydrological cycle we aim for a detailed consideration of fires within the fully coupled CCSM3 model. The current version of the land model within CCSM (CLM-CN) includes a prognostic fire algorithm based on Thonicke et al. 2001 in which fire emissions occur when climatic conditions are sufficiently dry and there is sufficient litter available. This current submodel of CLM-CN does not represent agricultural or land use change and severely underestimates contemporary fire emissions in magnitude and interannual variability by more than a factor of 2. Making use of advanced satellite data on area burnt and land use change data an improved prognostic fire module for CLM-CN will be developed driven by both human and climate parameters that: (i) reproduces regional and interannual variability in fire emissions during the satellite era (ii) is consistent with past and future trajectories of land use change (iii) provides emissions of CO2, CH4, reactive trace gases, and aerosols (iv) takes advantage from insights gained from a comparison of contemporary land use change and fire dataset.
B31E-0338
Study of biomass burning plume heights using combined satellite measurements
Vertical distribution of carbonaceous aerosols, especially those with strong absorbing and scattering properties, are of a great importance for current climate studies. Vertical aerosol profiles, however, still need further investigation to improve our understanding, and ability to simulate them in global aerosol and climate models. Satellite observations present a good dataset for model validation due to their global and regular nature. We present the preliminary results of a study on the height of biomass burning (BB) aerosol plumes as observed by Multiangle Imaging Spectroradiometer (MISR) on Terra satellite, and Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) on CALIPSO satellite. BB plume heights detected by these instruments are compared for several fire cases, where these measurements are close in space and time. Additionally, study shows how the fire events detected by Moderate Resolution Imaging Spectroradiometer (MODIS) are correlated with the MISR and CALIOP measurements. Several fire cases in different locations on the Earth during the years 2006-2008 have been chosen for analysis. Aerosol heights observed by the satellite instruments will further be compared with the aerosol profiles produced by the global Goddard Chemistry Aerosol Radiation and Transport (GOCART) model.
B31E-0339
Prescribed Burning as a Means of Reducing Emissions From Fires?
Terrestrial ecosystems, particularly forest ecosystems, have been identified for their carbon sequestration potential. However, many of the world's terrestrial systems experience periodic fire events, which emit a significant amount of carbon to the atmosphere, in the form of carbon dioxide, particles, and other trace gases. Recent studies have highlighted the importance of fire emissions throughout North America and demonstrated how these emissions impact regional climate and air quality. For example, emissions of volatile organic compounds and nitrogen oxides from fires can have a detrimental impact on air quality in seasons not typically prone to photochemical smog. Fire emissions are also a critical component of the carbon cycle and need to be considered when evaluating regional sources and sinks of carbon dioxide. Although the increasing extent and severity of wildfires can potentially result in different, and larger, emissions than those from prescribed burns, there has not been a continental-scale examination of the benefit of prescribed fire as compared to wildfire emissions of carbon dioxide, particles, and trace gases to the atmosphere. Here we use a continental-scale fire emissions model to investigate the potential changes in fire emissions when prescribed burns are applied more widely, and wildfires are assumed to be prevented. We will evaluate the impact of fire management practices on the overall emissions of carbon dioxide, particles, and other trace gases and how these emissions compare to those from anthropogenic sources. Quantifying the emissions of wildfire versus prescribed fire will aid in identifying the potential climate and air quality implications, allowing for further refinement of carbon accounting policy aimed at reducing atmospheric carbon concentrations.
B31E-0340
The Interannual Variability of Biomass Burning in North America using MODIS Data: Observations and Meteorological Influences
Meteorological impacts on the interannual variability of wildfires in North
America including Alaska are investigated using six years of the MODIS fire and
Aerosol Optical Depth (AOD) products, the meteorological data from North American
Regional Reanalysis (NARR), and the lightning data collected by the National Lightning
Detection Network (NLDN). The relationships of MODIS fire counts, fire radiative power, and AOD with over
13 meteorological variables were investigated in four sub-regions of the North American continent (Alaska,
western U.S., Québec, and the rest of Canada). Atmospheric instability and anomalies in the 500 hPa
geopotential height field explain more than 60% of the interannual variability in wildfires in Alaska and
Quebec; while in the western Unites States, pre-season precipitation is a dominant factor. Lightning strike
data show little correlation with fire counts in the western United States, suggesting the importance of
anthropogenic cause of fires in this region. Relationships between fire occurrence, atmospheric instability,
and smoke production were also investigated. It is revealed that although the Haines Index is widely used for
fire forecasting, it is not sufficient to interpret the interannual variability of fires in Boreal North America, but
its performance improves when used with 500mb geopotential height anomalies. Continuing work will focus
on the meteorological impact and interannual variability of smoke production and subsequent transport
between regions. In addition, analysis using lightning strike data may also be preformed for the Canada and
Alaska regions via Environment Canada and the Bureau of Land Management (BLM) respectively.
http://www.geosciences.unl.edu/~jwang
B31E-0341
Modeling the Global Emissions From Vegetation Fire for the Scenario Data of Integrated Assessment Model
Biomass burning is one of the important source of anthropogenic emissions of trace gases and aerosols. Recent developments of global fire emission inventories based on various satellite observations have improved the quality and seasonality of its emissions, however, the division of the causes into natural fires and anthropogenic fires usually cannot be determined easily by the remote sensing measurements. In this study, we evaluated emissions based on possible anthropogenic vegetation fires by comparing the emissions of simulated results of vegetation fire using a biogeochemical process model (VISIT: Vegetation Integrative SImulation Tool), the Global Fire Emissions Dataset (GFED), and the historical land use change dataset for years around 2000, to get relevant estimates of the anthropogenic emissions supplemented for the Integrated Assessment Models for the global climate change. Estimating long term trends in fire emissions was conducted using historical climate data of CRU TS 2.1 and the climate projections by a coupled atmosphere–ocean general circulation model (AOGCM).
B31E-0342
Effects of Siberian forest fires on regional climate in spring 2003
Forest fires are one of important sources for carbonaceous aerosols which are mostly comprised of organic carbon (OC) and black carbon (BC) aerosols. They have important climatic implications because of their extinction of solar radiation: OC scatters and BC absorbs solar radiation. These contrasting radiative properties add another complexity to our understanding the effects of those aerosols on climate. In spring 2003, the record-breaking intense forest fires occurred over Siberia, which emitted huge amount of aerosols in the atmosphere. We here examine the effect of these Siberian forest fires aerosols on regional climate in East Asia using a combination of numerical models and observations. First a global chemical transport model (CTM) with a biomass burning emission inventory constrained by satellite was used to simulate the enhancements of the aerosol concentrations due to the Siberian fires over East Asia. Our simulated aerosols were evaluated against the observations from the MODIS satellite and at the EANET sites. We then applied the simulated aerosols concentrations to climate simulations using the National Center for Atmospheric Research (NCAR) coupled global climate model, Community Climate System Model version 3.0 (CCSM3) to examine the impact of Siberian fire aerosols on regional climate. The difference in the model between with and without simulated Siberian fire aerosols defines the impact of fires on regional climate. The results indicated that fire aerosols resulted in a strong cooling at the surface and a general warming in the free troposphere and thus increased atmospheric stability. We also found significant decreases in geopotential heights over Siberia and decreases in cloud cover and precipitation in both Japan and the western North Pacific due to fire aerosols. Such changes were consistent with the observations based on the NCEP/DOE reanalysis II data, indicating the importance of fire impacts for regional climate simulations.
B31E-0343
Quantitative Assessment of biomass burning from satellite measurements of Fire Radiative Power
Satellite measurement of fire radiative energy (FRE) release rate or power (FRP) provides a vital means of distinguishing fires of different strengths. Analysis of 1-km resolution fire data, acquired globally by the MODerate-resolution Imaging Spectro-radiometer (MODIS) sensor aboard the Terra and Aqua satellites from 2000 to 2006, showed instantaneous FRP values ranging between 0.02 MW and 1866 MW. By applying a set of simple thresholds to this wide range of FRP values, it has been possible to classify fires into five categories, thereby facilitating simplified fire rating by strength, in a similar manner as the scales used for categorizing the strengths of earthquakes and hurricanes. Over 90% of all fires occurring in most regions of the world fall into category 1, while only less than 1% fall into each of categories 3 to 5, although these proportions may differ significantly from day to day and by season. The frequency of occurrence of the larger fires is region specific, and could not be explained by ecosystem type alone. Time-series analyses of the proportions of higher category fires based on MODIS-measured FRP from 2002 to 2006 do not show any noticeable trend because of the short time period. The global daily mean FRP ranges between 20 and 40 MW, but regionally the range is wider and varies by local time of day, with the Canadian eastern region or Quebec having the overall highest value of 85 MW at the Aqua-MODIS afternoon local overpass time of around 1:30 PM. Analysis of regional mean FRP per unit area of land (FRP flux) shows that at peak fire season in certain regions, fires can be responsible for up to 0.2 W/m2 at peak time of day. Zambia has the highest regional monthly mean FRP flux of ~0.045 W/m2 at peak time of day and season, while the Middle East has the lowest value of ~0.0005 W/m2.
B31E-0344
Mapping wildfire susceptibility in Southern California using live and dead fractions of vegetation derived from Multiple Endmember Spectral Mixture Analysis of MODIS imagery
Wildfire is a significant natural disturbance mechanism in Southern California. Assessing spatial patterns of wildfire susceptibility requires estimates of the live and dead fractions of vegetation. The Fire Potential Index (FPI), which is currently the only operationally computed fire susceptibility index incorporating remote sensing data, estimates such fractions using a relative greenness measure based on time series of vegetation index images. This contribution assesses the potential of Multiple Endmember Spectral Mixture Analysis (MESMA) for deriving such fractions from single MODIS images without the need for a long remote sensing time series, and investigates the applicability of such MESMA-derived fractions for mapping dynamic fire susceptibility in Southern California. Endmembers for MESMA were selected from a library of reference endmembers using Constrained Reference Endmember Selection (CRES), which uses field estimates of fractions to guide the selection process. Fraction images of green vegetation, non-photosynthetic vegetation, soil, and shade were then computed for all available 16-day MODIS composites between 2000 and 2006 using MESMA. Initial results indicate that MESMA of MODIS imagery is capable of providing reliable estimates of live and dead vegetation fraction. Validation against in situ observations in the Santa Ynez Mountains near Santa Barbara, California, shows that the average fraction error for two tested species was around 10%. Further validation of MODIS-derived fractions was performed against fractions from high-resolution hyperspectral data. It was shown that the fractions derived from data of both sensors correlate with R2 values greater than 0.95. MESMA-derived live and dead vegetation fractions were subsequently tested as a substitute to relative greenness in the FPI algorithm. FPI was computed for every day between 2000 and 2006 using the derived fractions. Model performance was then tested by extracting FPI values for historical fire events and random no-fire events in Southern California for the same period and developing a logistic regression model. Preliminary results show that an FPI based on MESMA-derived fractions has the potential to deliver similar performance as the traditional FPI but requiring a greatly reduced data volume and using an approach based on physical rather than empirical relationships.
B31E-0345
Tropical fire emissions injection heights and their impact on climate
We estimated biomass burning smoke plume heights in equatorial Asia using data from the Multi-angle Imaging Spectro-Radiometer (MISR). Smoke plumes were processed with the MISR Interactive eXplorer (MINX). We analyzed 110 plumes on the island of Borneo from 31 MISR scenes during April through November of 2006 (during a moderate intensity El Nino). The average regional plume height was 800 ±132 m (mean ± SD), substantially lower than boreal Alaskan plumes found using similar techniques. Kahn et al. (2008) linked smoke plume heights with atmospheric boundary layer (ABL) heights, finding that between 10% and 30% of Alaskan plumes in 2004 made it to the free atmosphere. We found a smaller fraction of plumes extending above the ABL. The average distance the plume traveled before it was undetectable was 35 km. We did not find evidence that plume heights varied as a function of fires in different vegetation types. Cross-sectional analysis of vertical smoke distribution suggests a link between smoke height and wind speed along the plume. To assess the impact varying plume heights have on climate, we performed simulations using the Community Atmosphere Model, version 3.1 (CAM). We injected smoke at 500m, 1000m, and 1500m in three separate experiments to assess impacts on the regional radiation budget and climate. References: Kahn, R. A., Y. Chen, D. L. Nelson, F-Y Leung, Q. Li, D. J. Diner, and J. A. Logan (2008), Wildfire smoke injection heights: Two perspectives from space. Geophysical Research Letters, 35, L04809, doi:10.1029/2007GL032165.
B31E-0346
Developing and evaluating a synoptic and diurnally varying time series of global fire emissions
To assess how recent changes in fire activity and burned area influence climate and human health, we need to improve our understanding of fire emissions. Information on synoptic and diurnal time scales may be particularly relevant for capturing interactions between climate drivers of fire activity and atmospheric transport of particulate and trace gas emissions. Developing emissions time series at this high temporal resolution requires combining active fire products with burned area information that is typically available only at monthly or annual intervals. Here we used several different active fire products from MODIS and GOES to distribute monthly fire emissions from Global Fire Emissions Database version 2 (GFEDv2) at a 3-hour time step. This process required several steps. First, we used the 8-day overpass corrected climate modeling grid active fires from MODIS AQUA and TERRA to distribute emissions during 8-day intervals within a year. In a second step we distributed emissions day by day within each 8-day interval using raw active fire detections from MODIS. In a final step we used GOES active fire observations to construct mean diurnal cycles for different vegetation types and latitude zones. These mean cycles were then applied to the daily emissions. The results show reasonable agreement of daily fire activity as compared with GOES, VIRS, and ATSR active fire products. We constructed mean diurnal cycles from GOES observations for three different land types: forest, shrublands and savannas, and grasses and crops in different latitude zones. We found that maximum fire activity occurred in the early afternoon 12:00-3:00 PM and minimum fire activity occurred at night from 00:00 to 6:00 AM. The diurnal cycles varied substantially in different land cover types and regions with boreal forest fires, for example, showing substantially more burning in late afternoon and night than fires in grassland or cropland areas. Fires in shrublands and savannas and grasslands and croplands showed a narrower mid-day peak than fires in forests. These features may reflect the diurnal cycle of fire controlled by the local background climate, especially air temperature and relative humidity as well as different roles of humans in regulating ignition. In a final step, we used the GEOS-Chem model to investigate whether the higher temporal resolution product improved agreement with atmospheric observations of CO.
B31E-0347
Hydrochemical Leaching of Wildfire Ash
A century of fire suppression, combined with recent droughts has provoked some of the worst wildfire seasons in the western US. Although wild and prescribed fires are known to supply nutrients to grassland, shrubland and forest ecosystems, when ash and combustion byproducts are leached into surface waters the nutrients and other materials can affect aquatic ecosystems and pose a considerable risk to water quality. This ash may be persistent for periods as short as a storm or snowmelt event or up to several years, as suggested by periodic increases in dissolved nutrients and suspended solids. Here I present results from field sampling and bench scale experiments that examine the rate of change and chemical quality of leachate from ash samples collected from two wildfires that burned in Colorado in 2003 and 2006. Bench scale- experiments suggest that the conductivity of ash leachate increases in a continuous and modelable manner. Stream grab samples collected in burned and unburned areas within two weeks of the 2006 Mato Vega fire suggest an initial increase in pH, and conductivity, as well as an increase in solutes including dissolved organic carbon and manganese; however the results were spatially variable.
B31E-0348
Nitrogen Subsidies and Fire Prone Ecosystems: the Case of African Savannas
Savannas are the most frequently burned ecosystems on earth and fires are known to have significant effects on community composition and species interactions across all trophic levels. Fire also plays a significant role in the biogeochemical cycles of carbon and various nutrients in savannas as they are emitted or chemically transformed into more available, or more recalcitrant, forms during fires. In Africa, savanna fires occur more frequently in the unstable (mean annual precipitation (MAP) > 516mm), dystrophic (nutrient poor) savannas than they do in the relatively more stable (MAP < 516mm), eutrophic (nutrient rich) savannas. Are these dystrophic savannas dystrophic because they burn? Are the eutrophic savannas eutrophic because they are subsidized by nutrient inputs from spatially remote fires? Africa is an ideal location to study ecosystem sensitivity to atmospheric inputs of nitrogen from biomass burning for several reasons. A large proportion of the continent is very fire prone and approximately 40% of global emissions from biomass burning originate in Africa. In contrast, emissions from fossil fuel sources are low compared to highly industrialized regions in the Americas, Europe and Asia and agricultural input of nitrogen is limited in Africa. As a result, the ecosystem impacts of pyrogenic loss and gain of nitrogen are likely to be more important in Africa than elsewhere. We combine data from recent analyses of fire frequency (GFED) with an atmospheric transport model (HYSPLIT4) to simulate pyrogenic nitrogen deposition fields and analyze the fate of fire-related nitrogen in the atmosphere, gross and net N-deposition fields and thus the role fires may play in displacing nitrogen and depleting or enriching adjacent regions.
B31E-0349
Landscape Estimates of Total Heat Release and Fuel Consumption From Prescribed Fires: Analysis and Calibration of Sequential Infrared Images From Aircraft
We used infrared images acquired from aircraft and ground-based calibration to produce sequential maps of fire radiative power (FRP, kW m-2) over the period of active flaming during prescribed fires. Ground sensors were calibrated for total radiant heat flux. Because the residence time of flaming combustion is short relative to the aircraft return time (3-5 min), integrating FRP over time to estimate fire radiative energy (FRE, kJ m-2) is not straightforward. Integration can be done for pixels where active flaming fronts can be identified and if the typical time course of FRP can be estimated independently. Heat flux data from ground sensors within the fires and from experimental fires were used to describe the time course of heat release. Experimental data and the literature were used to estimate the proportionality between energy release and total and rate of fuel consumption. Analysis showed considerable variability in heat release within prescribed fires because of topographic effects and ignition pattern and among fires because of differences in fire weather and ignition.
B31E-0350
Tundra Fires in the Noatak National Preserve, Northwestern Alaska, Since 6000 yr BP
Over 1.7 million hectares of Alaskan tundra have burned over the past 50 years, including the record-setting Anaktuvuk River fire in 2007. Despite this evidence indicating the flammable nature of these ecosystems under warm and dry conditions, land managers and global change scientists lack critical information concerning long-term relationships among fire, climate and tundra vegetation. This knowledge gap limits the ability to assess the response of the tundra fire regime to ongoing and predicted climate warming and potential feedbacks with Earth systems. We utilize macroscopic charcoal from lake-sediment cores to characterize the frequency component of fire regimes in shrub-dominated and herb-dominated tundra ecosystems in northwestern Alaska over the past 6000 years. Here we present the first long-term records of tundra fire regimes from the Noatak National Preserve, a region encompassing some of the most flammable tundra in the state. Results from three lakes indicate that fire has been a consistent process in the region, with fire return intervals (FRIs) ranging from 70 to 800+ years since 6000 yr BP. FRIs were similar between herb- and shrub-dominated tundra sites before ~2000 yr BP, with a mean FRI of 167 yr (95% CI 145-195) Over the past ~2000 years, however, herb- dominated sites burned more frequently (mean FRI 112 yr [95% CI 80-151]) than shrub-dominated sites (mean FRI 247 yr [95% CI 141-377]). At millennial time scales, shifts in historic FRIs were likely related to regional climate changes and/or associated vegetation changes. These results provide a context for resource management and serve to refine the tundra component of an ecosystem model designed to aid land managers in assessing fuels and fire hazards in the context of climatic change.
B31E-0351
Burn Severity Assessment in the Okanogan-Wenatchee Forest Using NASA Satellite Missions
Fire severity is an increasingly critical issue for forest managers. A long history of fire suppression has led to millions of acres of dry western forests and a buildup of hazardous fuels. Satellite imagery offers a cost- effective and feasible tool for fire severity assessment and can provide near real-time data for mitigation measures. This study focused on the Tripod Complex Fire that burned more than 175,000 acres of the Okanogan-Wenatchee Forest in Washington in 2006. Field data were collected in order to calculate the Composite Burn Index (CBI), a ground-based measurement of burn severity which can directly correlate with satellite measurements. These in-situ data were used to calibrate the satellite data from the Landsat TM5 sensor and the MODIS sensor on the NASA TERRA satellite. The satellite data were used to calculate the differenced Normalized Burn Ratio (dNBR) and the relative rNBR. These algorithms use the relationship between the near infrared and the shortwave infrared pixel values to quantify burn severity. After comparing these two algorithms, it was determined that there was no significant difference between dNBR and rNBR. Using the burn severity map created with the dNBR data, an analysis was performed to examine the relationship between burn severity and variables such as slope, aspect, and vegetation type. The results showed that there was no relationship between burn severity and any of the variables in the Tripod Complex Fire.