 |
 |
|
||||||||||||
| Home | George Mason | COS | Email Us |
||||||||||||||
|
|
ObjectivesAlthough wildfires east of the Mississippi are less frequent and less intense, the proximity of a huge population closer to fire-prone areas, a vast expanse of wildland-urban interface and the importance of prescribed burns makes fire-weather monitoring a necessity in this region. Direct broadcast or real time transmission of satellite data to the ground from MODIS abroad the Terra and Aqua satellites allows opportunities for real time fire weather monitoring of the eastern US. The Direct Readout Lab (DRL) station at NASA Goddard Space Flight Center, Greenbelt, Maryland has the technological capability of acquiring this freely transmitted data with coverage across the entire eastern United States. Four satellite revisits two each of Terra and Aqua provides opportunities for observing the diurnal variations in fire weather. The availability of such data in real time in DRL’s ftp site facilitates our endeavor. Our objective is to build up a one stop portal for fire weather monitoring which in addition to the relevant data and images would provide the tools to visualize, analyze and simulate the wildfire conditions.
| |||||||||||||
|
|
Soil MoistureSoil moisture is one of the key parameters controlling the exchange of water and heat energy between land surfaces and atmosphere. It is also an important indicator for monitoring flood, drought and fire danger potential. Combining optical/IR observations with accurate ground measurements of soil moisture can produce daily, fine resolution soil moisture information. It is known that the daily Normalized Difference Vegetation Index (NDVI) data product and Land Surface Temperature (LST) are related to surface soil moisture, therefore, relationships among ground measurements of soil moisture and land surface NDVI and LST products have been developed. These regression relations, in conjunction with MODIS NDVI and LST, are then used to generate soil moisture information in areas where there are no ground station data. The following figure is a fuel moisture map derived from NDVI and LST measurements of the Moderate Resolution Imaging Spectroradiometer (MODIS), flown on the NASA Aqua spacecraft.
Soil Moisture Map (% by weight), Eastern USA; MODIS Aqua 15th April 2005: 1825-1835 UTC;; Generated by Eastfire Group,CEOSR, GMU EastFireWatchEastfireWatch would be the wildfire weather observing portal for eastern USA covering the 25 states east of the Mississippi. EastFireWatch would provide real time monitoring of fire weather in this region by extracting relevant fire weather parameters from processing of level 1B data acquired through MODIS direct broadcast. These parameters are True color images, Vegetation Indices (NDVI, EVI) Fuel Moisture, Fuel Temperature, Fuel Temperature, Fire Danger Indices, Thermal anomaly, Burn Scar Index, Fuel Loading. Real time monitoring would be achieved by providing mosaiced images of the aforementioned parameters for the region as soon as the satellite overpass occurs over the region and the Level 1B data products become available on the NASA directreadout portal. Additionally the system will generate daily products by integrating the Aqua and Terra data. The integration provides a better assimilation of the data, since cloud obscured pixels during one pass can be replaced by values in that pixel under clear conditions in another pass. The integration scheme would help in minimizing errors due to sun/satellite geometry such as BRDF effects. EastFireWatch would maintain the images and data products in its archives. The corresponding metadata will be maintained in a separate metadata archive. A metadata server built on the metadata archive will provide support to the web applications. The web pages will be dynamically built based on the query answers from the metadata server. The metadata server thus supports and facilitates easy searching and downloading of generated images and products. In future the EastFireWatch portal would be adding several sophisticated tools for better visualization and analysis of the generated data and images. The proposed ImageVisualizer would allow for better visualization of the images by including facilities for zoom in, zoom out, panning, and overlaying of GIS layers. This system would be implemented using ESRI ArcIMS. EastFireWatch also intends to provide in the future Fire simulators and smoke simulators by integration of FARSITE and BlueSkyRain respectively into the system. Farsite integration would allow forest managers to select a region and simulate a fire using the current fuel and weather conditions. They would also be allowed to run the simulations using their own conditions. On the other hand BlueSkyRain integration would serve as a tool for air quality assessment in events of wildfire. A DataAnalyzer section would be adding data mining and knowledge discovery tools. True Color imagesWe build true color images using MODIS band 1 (250m), band 4 (500m) and band 3 (500m). Standard atmospheric corrections are performed for generating the true color image.
Mosaiced True Color Image: Eastern USA; MODIS Terra 18th April 2005 :1540-1550 UTC; Generated by Eastfire Group,CEOSR, GMU Vegetation indicesWe are currently using the MODIS Rapid Response version of the Vegetation Index algorithm which can be used to generate Vegetation indices for a single swath. Vegetation indices include the Normalized Difference Index and the Enhanced Vegetation Index. The algorithms are simple and are applied on a corrected reflectance product. The corrected reflectance product is generated by a simple atmospheric correction using MODIS visible and near-infrared bands (bands 1 to 7). Corrections for molecular (Rayleigh) scattering and gaseous absorption (water vapor, ozone) are also made using climatological values for gas contents. Thus no real time input or ancillary data is necessary. However no aerosol correction is made. Since the reflectance product is not corrected for aerosols the Rapid Response version of the products may be of poorer quality in presence of aerosols compared to the official MOD13 16 day composite vegetation indices product. Moreover the Rapid response version of NDVI/EVI currently performs no correction for directional effect and can be generated for a single swath as opposed to MOD13 16 day composite product. We are currently investigating techniques to perform reasonable BRDF corrections on single swath data.
Mosiaced NDVI Image: Eastern USA; MODIS Terra 16th April 2005 :1555-1605 UTC; Generated by Eastfire Group,CEOSR, GMU
Mosiaced EVI Image: Eastern USA; MODIS Terra 16th April 2005 :1555-1605 UTC; Generated by Eastfire Group,CEOSR, GMU Fuel moistureBurgan’s algorithm for live vegetation moisture estimation is based on
current observation and historical records of NDVI, and it is easy to be
implemented for operational use. Currently it is used in Wildland Fire
Assessment System (WFAS). With Burgan’s approach, live vegetation moisture (MC)
is computed as follows:
Fuel TemperatureFuel temperatures can be derived using Surface Temperature Products. The Surface Temperature product is generated using the split-window algorithm originally developed by Wan et al., 1996. A new implementation of the algorithm for near real time applications was developed by Ana Pinheiro of the Rapid Response team. This version does not require ancillary datasets. It uses monthly climatological Tiros Operational Vertical Sounder (TOVS) data sets to get surface air temperature and water vapor and a land cover based emissivity. We are currently awaiting the availibility of the stand-alone version of the algorithm. Fuel loadsA preliminary analysis indicates that there is a correlation between Short-Wave Vegetation Index (SWVI) and total forest biomass, and the vegetation index can explain 60-66% of the variation in post-fire forest re-growth age, an indirect measure of biomass content (Fraser and Li, 2002). For the estimation of the fuel load we will use the National Fire Danger Rating System (NFDRS) fuel model.The NFDRS fuel model defines parameters like the heat content, fuel particle density, total mineral content, effective mineral content, total fuel load, fuel bed depth, surface area to volume ratio and moisture of extinction for each class. The fuel load for each fuel model class is already computed by extensive field measurement, a lookup table will be used to get the 1 hour, 10 hour, 100 hour and 1000 hour dead fuel load corresponding to each fuel model class. In order to better represent the current fuel load we will take the weekly relative greenness map from MODIS NDVI data.
Active Fire LocationsThe Active Fire Detection Product alternatively called the Thermal anomaly product is generated using the algorithm used by the MODIS Rapid Response program. The algorithm uses brightness temperatures derived from the MODIS 4 and 11 micrometer channels (Giglio et al 2003). The fire detection strategy is based on absolute detection of the fire, if the fire is strong enough i.e radiating at a brightness temperature of 360K. Temperatures less than that could be natural to the surface, for eg in deserts. To filter out naturally hot surfaces and detect weaker fires, the algorithm uses a contextual test. This latter test identifies pixels with values elevated above a background thermal emission obtained from the surrounding pixels. This method accounts for variability of the surface temperature and reflection by sunlight. The contextual approach was a significant improvement in the original algorithm and results in less false detection than traditional threshold-based algorithms. The algorithm is also sensitive to small fires, which makes it even more suited to wildfire applications.  Real-time Active Fire Detection using MOD14 algorithm. MODIS Aqua 15th April 2005 :1825-1835 UTC. Red dots represent fire locations detected by the MOD14 algorithm; Generated by Eastfire Group,CEOSR, GMU Thermal anomaly RGBThe thermal anomaly RGB is generated with MODIS band 6 corrected reflectance in the green, band 7 corrected reflectance in the blue and band 20 at 3.75 micrometer in the red (Roger and Vermonte,1998). Qualitatively, the RGB thermal composite is a very useful way to locate fire activity, where the thermal energy released by fires is interpretated as an extra relectance contribution at 3.75 micrometer and produces "red spot" on the image. The relative intensity of red is directly proportional to the size and temperature of of the fire. Fire Danger IndexWe are currently evaluating a normalized difference index using MODIS bands 7
and 2 for fire danger estimation. Band 7 in SWIR, being a water absorption band
is sensitive to fuel moisture content. Band 2 in NIR is not very sensitive to
fuel moisture. The following index may be useful for fire danger
rates
Burn scar IndexComing up
Air QualityCarbon Monoxide: Biomass burning is an important source of global CO concentration which, when influenced by the atmospheric circulation system, can affect areas lying far from the actual source of the fire. CO concentration is computer for the eastern Unites States using CO data from MOPITT (Measurements of Pollution in the Troposphere) measurements.
Monthly Carbon Monoxide Concentration for May 2004
|
