Fuel Classifications - Wildland Fuel Fundamentals and Applications - page 116

the most commonly used FBFM classifications are the (1) 13 FBFMs described
by Anderson ( 1982 ), (2) 40 + models of Scott and Burgan ( 2005 ), and (3) 20 fire
danger fuel models used in the National Fire Danger Rating System (Deeming et al.
1977 ). Others have created new sets of custom FBFMs to these classifications.
Reich et al. ( 2004 ), for example, created several new BEHAVE custom fuel models
using field loading data that were then mapped to a South Dakota US landscape, and
Cheyette et al. ( 2008 ) created custom fuel models for the wildland urban interface
lands around Anchorage, Alaska, using a supervised vegetation-based classification
of 13 cover types. In Greece, Dimitrakopoulos ( 2002 ) created seven FBFMs by
synthesizing fuel data from 181 natural fuel complexes described by vegetation. In
Corsica, Santoni et al. ( 2011 ) developed two fuel models for a spatially explicit fire
model built to simulate fire behavior for maquis and juniper shrublands. To evalu-
ate fire hazard in Portugal, Fernandes ( 2009 ) developed a suite of 19 fuel models
based on the dominant vegetation structures and complexes in mainland Portuguese
forests.
The main advantage in creating abstract fuel description systems is that, ideally,
the resolution of fuel classes (FBFMs) match the resolution of the fire models for
which the classes will be used as inputs. Another words, each FBFM represents a
major change in predicted fire behavior in the Rothermel ( 1972 ) model. This means
that the uncertainty and error in model predictions may be minimized from inaccu-
rate and inappropriate fuel inputs because the fuel models were calibrated to actual
fire behavior observations (Burgan 1987 ). Another advantage is that new custom
fuel models can be developed for unique local situations or for broad use across
large regions (Burgan and Hardy 1994 ).
The biggest drawback to the abstraction classification approach and their prod-
ucts, such as FBFMs, is that without prior knowledge of fire behavior in local fuel
conditions, it is nearly impossible to accurately and consistently identify, use, and
interpret most of the abstract classes. Identification of FBFMs in the field, for exam-
ple, is highly subjective because it is based on an individual's perception of how fire
will burn the fuelbed under severe weather conditions, rather than on actual mea-
surements of fuel loadings. There are no standardized keys to consistently identify
FBFMs for either the Anderson ( 1982 ) or Scott and Burgan ( 2005 ) FBFM classifica-
tion systems. Because abstract classifications are inherently subjective and difficult
to use, most fuel mapping efforts based on abstract classification products must rely
on expert knowledge and past experience (Keane and Reeves 2011 ). FBFMs are also
difficult to create because their development requires a delicate balance of parameter
adjustments to match observed fire behavior with fire weather and fuel properties
that should only be done by experienced analysts and fire managers (Burgan 1987 ).
These limitations may preclude the use of FBFMs in the future as new fire behav-
ior simulation models are developed, as novel fuelbeds are created from innovative
fuel treatments, and as abundant fuel input data become available for describing
fuelbeds.
Abstract fuel classifications can only be used for fire behavior prediction and are
rarely used in other areas of fire and land management. FBFMs, for example, don't
include loadings for some major fuel components, such as logs and duff, which are
critical for computing smoke emissions, simulating post-frontal combustion, and