NOTICE TO ALL USERS:
The fire perimeters in this directory are DRAFT data.  Efforts are ongoing 
to locate and automate additional fire perimeters, and to perform quality 
checking on the data.  The data is being made available in draft form due 
to a very high level of demand.  Because of this, we ask that you 
DO NOT REDISTRIBUTE THIS DATA TO OTHER USERS, but refer all 
requests for this data to the Alaska Fire Service.  Contact information is 
included at the end of this file, as well as in the metadata document included 
with the data (metadata.htm).  This will allow us to keep an accurate record of 
who has received the data so that we can notify them when a production version 
is available.

NOTE:  Even if you choose not to read the metadata, please read the file called
faq.txt.  It contains important information that will probably save you a
phone call to ask questions for which you already have the answers.

Perimeters of fires greater than 1000 acres between 1950 and 1992, inclusive, 
were automated by the Environmental Resources Institute of Michigan from Alaska 
Fire Service records.  Perimeters of fires greater than 100 acres between 1993 
and 1999, inclusive, were automated by the Alaska Fire Service. Perimeters of
fires between 100 and 1000 acres between 1988 and 1993 were also automated by 
the Alaska Fire Service.  Thus, the current state of the data set includes
fires greter than 1000 acres from 1950 to 1987 and fires greater than 100 acres
from 1988 to 1999. Please refer to the enclosed metadata document, metadata.htm,
for more detailed descriptions of the data. 

All files included in this package were produced on an RS6000 computer running 
AIX version 4.1.  The metadata file is a Hyper-text Markup Language (HTML) 
formatted text file, and should be readable by standard web browsers (Netscape, 
Mosaic, Internet Explorer, etc.) on any machine. 

The Alaska Fire Service and the Environmental Resources Institute of Michigan 
provide no warranty, express or implied, with regard to the content, format,
quality, or accuracy of this data.  The Alaska Fire Service and the 
Environmental Resources Institute of Michigan accept no responsibility for the 
inappropriate or uninformed use of this data.

Contact information:
PLEASE REFER ALL REQUESTS FOR THIS DATA TO:

Tim Hammond, GIS coordinator
Bureau of Land Management, Alaska Fire Service
P.O. Box 35005
Ft. Wainwright, Alaska 99703 

Phone: (907)356-5581 or (907)474-2210
Fax: (907)356-5583
E-mail: tim_hammond@ak.blm.gov


The following is a brief description of the data received from ERIM.
Note that the dates represent the data provided to the Alaska Fire Service
by ERIM, not the entire database.
*****************************************************

This coverage is the Alaska Firescar location database.  It contains point 
and boundary location information for fires in Alaska between 1950 and 1996.
The accuracy and completeness of the database has not yet been assessed, 
but, in general newer (> 1970) fires have more accurate, more complete 
information.  

The database has been built using the ArcINFO Regions geographic feature 
model.  The region subclass containg the firescar information is 
region.firescars.  This feature class is particularly useful in delimiting 
fire boundaries for several reasons.  It allows firescars to overlap, which 
may occur not only between years but also within a single fire season.  It 
can also allow several spot fires to be mapped and attributes assigned to a 
single record entry.  Also, unburned island polygons within firescar 
boundaries need not have a record entry in the attribute table.  This allows 
each fire to be described by a single record entry and simplifies generation 
of summary statistics.  One drawback to the region feature class is that 
point locations may not be converted to regions directly.  I thought it was 
important to have all existing fire information incorporated into a single 
database.  Therefor, I came up with a work-around for this problem.  I 
buffered each point location to 10m.  This created a polygon that could then 
be added to the region subclass, but could be differentiated because it was 
much smaller (< 500m^2) than any other true fire boundary.  Also, after 
creating the database, an attribute was added (BOUNDARY) that indicates 
whether the feature is a point (BOUNDARY = 'N') or a boundary (BOUNDARY = 'Y'). 
Thus, using this work-around, summary statistics on fire burn area should 
not be calculated using the default AREA item, but should be done using 
either the ha_burned or the acres_burned items. 

The following is a brief description of the non-default region subclass 
attribute items, folllowed by projection information.

From ak_fires_a.patfirescars:

COLUMN   ITEM NAME
   17  BOUNDARY		- Indicates whether boundary (Y) or point (N) location information (See text above)
   19  COUNTRY		- Country in which fire occurred (will all be AK in this case)
   23  YEAR		- Year of burn
   27  SITENAME		- Fire id (in most cases Fire Service fire number)
   35  ACRES_BURNED	- Area burned in acres
   43  HA_BURNED	- Area burned in hectares