Methods

Methods selection for this project was fundamentally driven by two requirements: (1) The need to develop consistent information across all lands within the project extent, and (2) the need to develop consistent information spanning a significant historical period. Based on these requirements, remotely sensed data were considered to be the only cost-effective and spatially resolved source to consistently delineate and measure the response of thousands of individual fires across a continental extent and multidecade time frame. A significant body of literature exists evaluating the effectiveness of various scales of remotely sensed data to characterize fire severity (Brewer and others 2005, Chuvieco and Congalton 1988, Diaz-Delgado and others 2003, Fernandez and others 1997, Justice and others 1993, Kasischke and French 1995, Key 2005, Milne 1986, Patterson and Yool 1998, Pereira 1999, Roy and Landmann 2005, Sa and others 2003, Smith and others 2005, Sunar and Ozkan 2001, Wagtendonk and others 2004, White and others 1996). Scientific and operational precedent exists for the use of a remote sensing-based approach.

Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM) data provide the longest record of relatively high spatial and spectral resolution data for mapping fire severity. Not only does this enable the mapping of historical fire severity, it also facilitates the use of multitemporal approaches for characterizing postfire effects. Landsat data have been shown to be responsive to relative changes in aboveground biomass as a result of fire (Epting and others 2005, Kushla and others 1998, Lopez-Garcia and Caselles 1991, Miller and Yool 2001). More specifically, multitemporal change detection approaches based on prefire and postfire Landsat data have proven to be a cost effective and relatively accurate means of mapping fire severity (Brewer and others 2005). Furthermore, the availability and low cost of Landsat data were additional factors supporting their use for a project of this geographic and temporal extent.

Multitemporal approaches applying image ratios and image differencing techniques to Landsat data have been developed for a variety of assessment objectives. Imagery is commonly transformed mathematically into indices by ratioing a spectral component(s) or band with another spectral component(s) for each pixel. The transformation of Landsat data into vegetation indices, (e.g., Normalized Difference Vegetation Index) to strengthen the relationship between spectral response and vegetation characteristics has been widely used, and a number of indices exist (Lyon and others 1998). Lopez-Garcia and Caselles (1991) published the first index specifically derived to enhance the relationship between Landsat spectral response and burned vegetation. This Normalized Difference index was subsequently adapted and operationally implemented by Key and Benson (2002) and was used to develop historical fire severity data and atlases on numerous national parks. The approach has been named the Normalized Burn Ratio (NBR) and, combined with multitemporal differencing, has been utilized in fire severity mapping efforts by the U.S. Geological Survey and the USDA Forest Service since 2002. The Normalized Burn Ratio is calculated as: (TM4-TM7)/(TM4+TM7) where TM4 represents the near infrared spectral range and TM7 represents the short wave infrared spectral range. A differenced NBR image (dNBR) is created by subtracting the prefire NBR image from a postfire NBR image.

dNBR data have been used operationally for both rapid response and longer term assessment and monitoring (Bobbe and others 2003, Gmellin and Brewer 2002, Key and Benson 2002). Rapid response needs require the use of immediate postfire imagery to show first order fire effects on vegetation and soils and to prioritize rehabilitation resources. Longer term assessments have relied on image data typically acquired during the growing season following the fire in order to include delayed first order effects, (e.g., delayed tree mortality) and dominant second order effects that are ecologically significant, (e.g., initial site response and early secondary effects). Extended assessments are intended to provide a more comprehensive ecological indication of fire severity. Both immediate and extended assessments have uncertainty associated with the dNBR-based fire severity characterizations. Prefire vegetation conditions and postfire management activity influence the nature and magnitude of this uncertainty. The sensitivity of a given set of analysis objectives to the uncertainties associated with immediate and extended assessment dNBR data should be considered when using these data.

Based on the scientific foundation in the literature and operational precedent, the dNBR approach was selected to characterize fire severity and delineate fire perimeters for this project. Extended assessments will be conducted on forest and shrub ecosystems, and initial assessments will be conducted on grasslands and other specific vegetation communities known to recover from fire within a single growing season. A simple production model was developed around this approach to ensure timely and consistent products. We recognize that application of a simple production model across the ecological extent and variation covered by the project may itself be a source of uncertainty. Indeed, limitations associated with this approach to characterizing postfire effects are not yet known in many ecosystems, and we expect the MTBS project to expand our understanding of how consistently and precisely Landsat data can map burn severity. The MTBS product suite is a reflection of the need for a range of data to suit specific analysis objectives. The following steps outline the process of identifying fire locations through summarization of the results:

• Fire history database compilation
  • Data acquisition
  • Data standardization and aggregation
• Image data selection and preprocessing
  • Scene selection
  • Preprocessing
  • Delivery and archiving
• Fire severity interpretation and perimeter delineation
  • Normalized Burn Ration calculation and differencing
  • Interpretation and thresholding into severity classes
  • dNBR partitioning
  • dNBR fire perimeter delineation
• Stratification and summarization of severity information.