Fuel Classifications - Wildland Fuel Fundamentals and Applications - page 108

7.2 Classification Approaches

Several fuel classifications are currently used by land management agencies across
the globe, and most of these systems appear quite similar because they have com-
parable categories, components, and description variables (Anderson 1982 ; Keane
2013 ; Sandberg et al. 2001 ; Weise and Wright 2014 ). The main distinction between
most existing fuel classification systems is in the approaches used to create them
(Keane 2013 ). Although it would be much easier if there was only one fuel classifi-
cation for all fire science and management applications, multiple fuel classification
systems exist today because each fire modeling system requires a specific set of fuel
inputs and its own unique classification input scheme. Fire behavior fuel classifica-
tions, for example, include fuel component attributes, such as fuel depth, that may
not be needed in fire effects prediction systems.
Effective biological classifications are designed to be systematic (well organized),
practical (easily identified using a key), singular (uniquely identifies a class), and
comprehensive (the key can be used across a broad range of fuelbeds). This usually
implies that the classes that comprise them are mutually exclusive, and a change in
the value of an attribute of one class usually affects the values of the same attribute
in other classes (Gauch and Whittaker 1981 ). However, many of today's fuel clas-
sifications were not created using systematic classification procedures that group
fuelbeds based on statistical and ecological differences. Because of this, the fuel
classifications in this chapter will be summarized by the four broad approaches used
to create them: (1) association, (2) opportunistic, (3) classification, and (4) abstrac-
tion (Table 7.1 ). Of course, some of the fuel classifications presented as examples
were created using a combination of approaches.

7.2.1 Association

Many have associated or linked fuel component information, such as loading, to
the categories of other extant classifications commonly used in natural resource
management (Keane 2013 ). This is often accomplished by summarizing field-col-
lected fuels data by extant classification categories. For example, Reinhardt et al.
( 1997 ) average field-measured fuel loadings for eight fuel components across the
vegetation-based categories of both the Eyre ( 1980 ) forest cover type classification
and the Shiflet ( 1994 ) range cover type classification to facilitate input to the First-
Order Fire Effects Model (FOFEM). In Canada, Hawkes et al. ( 1995 ) assigned fuel
loadings to various categories of vegetation and timber type classifications, and the
Canadian Fire Behavior Prediction System contains fuel input types that are associ-
ated with major forest vegetation types (FCFDG 1992 ). Poulos et al. ( 2007 ) created
vegetation composition and structure layers from environmental gradients, satellite
imagery, and forest inventory data, then scaled fuels information to the resultant
biophysical classification for Texas fuelbeds. The fuel type group classification was
created by summarized Forest Inventory and Analysis georeferenced fuels data by
forest type groups (Keane et al. 2013 ).