Chapter 1 | Chapter 2 | Chapter 3 | Chapter 4 | Chapter 5 | home
Arkansas's 1990 rural population was 52.8% by the Non-Metropolitan definition, used by several federal agencies such as the US Office of Management and Budget, or 46.5% by the Census Bureau's rural definition. The Census Bureau ranked Arkansas 11th in the nation in rural population. 'By virtually any reckoning, Arkansas has been and continues to be a rural state (Arkansas Rural Development Study Commission et.al, 1997). There are over one million persons living in rural areas. The States' annual population growth rate is 1.1% (90,107 persons annually, rural growth is 46,856 persons annually), the study area in Northwest Arkansas (Benton, Carroll, Crawford, Madison, Washington Counties) is growing at an 2.7% annual rate (9,601 persons annually, rural growth is 5,528 persons annually). In the study area there were 41,700 rural residences in the 1994 data set used (3.58 persons per residence on average), this category is growing annually at 1544 rural home per year or 2.7 % annually. Arkansas, like most southern states, has a wildland urban interface zone in jeopardy for a number of reasons: ongoing fire suppression force reduction (both funding and personnel), increasing rural population growth (Benton, Carroll, Crawford, Madison, and Washington Counties; 2.7% annual growth or 5000 persons annually), increasing forest fuel loads, and no clear fire-fighting responsibility, along with conflicting and compromising legal and government mandates.
Arkansas' 1990 - 2000 population growth (4.2% projected) is expected to be higher than the 1980 - 1990 (2.8%), the 1970 - 1980 saw a higher growth rate (18.9%) than anytime since world war. Most of this increase in growth is due to migration (3.5%), interestingly this growth will be primarily in the 65+ age group and 45 - 60 age group. The number of 45 - 60 aged individuals will increase 23%, as the nation overall matures. In Arkansas, the median age will increase from 33.8% to 38% by the end of the decade.
Eleven counties of the states' six metropolitan statistics areas will
absorb about 49% of the projected growth. Northwest Arkansas is expected
to absorbing 32% of the projected growth. One in six persons in Arkansas
will live in Northwest Arkansas by the year 2000. The annual growth rate
of population in District 6 is 2.9% and the annual growth rate in housing
is 2.96%. The total number of rural residences in the wildland urban interface
by the year 2000 is projected to be over 75 thousand (Table 9).
Table 9 . District 6 Rural Population and Housing Growth (Arkansas Rural Development Study Commission, 1997)
|
|
|
|
Table 10 shows Palmer Drought Severity Index (PDSI) ranges from 0.0 (normal) to -4 (extreme drought), actual range is from +7 to a -7. Palmer Drought Severity Index is used by the National Oceanographic and Atmospheric Administration's Climatic Diagnostic Center. Palmer Drought Severity Index is a monthly value (index) that indicates the severity of a wet or dry spell of weather. A PDSI value of 0.0 to -0.5 is considered normal.
Table 10. Palmer Drought Severity Index (PDSI) Values
Table 11 shows the relationship between the number of wildfires from 1992 to 1996 and the PDSI. A Pearson's Correlation (Downie and Heath, 1970) showed a strong negative correlation, -0.9 for wildfire statewide and -0.93 for wildfire in District 6. Not only does the correlation demonstrate a relationship between the number of wildfires but also shows them to be inversely related, the more negative the PDSI values the greater the number of wildfires. Examination of the occurrence of wildfires statewide from 1983 to 1987 (Table 12) showed a similar correlation of -0.84. The increase in wildfires statewide and in district 6 shows the effect of even mild drought upon the frequency of wildfires. Individual years with PDSI values approaching Incipient Drought (-0.5 to -1) levels had above average wildfire occurrence: 1986, -0.87 PDSI with 3,170 wildfires, 1987, -0.59 PDSI with 4,150 wildfires, and 1995, -0.54 PDSI with 4,204.
Table 11. Palmer Drought Severity Index Vs Number of Wildfires 92-96
|
| ||
Table 12. Palmer Drought Severity Index Vs Number of Wildfires 83-87
| Year | PDSI | Number of Fires |
Figure 3 demonstrates that drought as measured by PDSI can move quickly from a wet period to a dry period in one single growing season. In September 1995 3,247 acres burned costing $40,200 to suppress. Statewide wildfires increased 52% (4,204) above an average year of 2,761 wildfires. AFC District 6 experienced a 46% (247) increase in the number of wildfires, the annual average is 169 wildfires.
Figure 3. Arkansas PDSI Values for Each Division (1995)
Table 13 shows the statewide incident report records show number of fires, total acreage burned, forested acreage burned, and non-forested acreage burned by year. Analysis of recent fire incident reports from 1992 to 1996 show a slight increase in the yearly average of fires statewide from 2,140 to 2,900. Kluender et.al (1988) found a general increase from 2,185 wildfires to 4,150 wildfires, this may have been due to variable weather conditions, particularly drought. The largest fire in the 1983 to 1987 data set was 900 acres, the largest forested fire was 900 acres, and the largest non-forested fire was 370 acres. There were 11 fires that burned 500+ forested acres and 214 fires that burned 100+ forested acres. The largest non-forested fire was 370 acres with 28 fires larger than 100 acres. The largest fire in the 1992 to 1996 data set was 1835 acres, there were 10 wildfires larger than 500 acres. There were 215 wildfires over 100 acres, 11,248 wildfires were less than 100 acre in size. The largest forested fire was 1835 acres with 180 fires over 100 acres. The largest non-forested fire was 370 acres with 27 non-forested fires over 100 acres.
A decrease in all categories can be seen between the 92-96 data and the 83-87 data, whether this is from improved suppression techniques or due to some environmental variables cannot be determined from these small data sets.
TABLE 13. Summary of Wildfires Statewide by Year
|
|
|
| |
| TOTAL | ||||
| TOTAL | ||||
Table 14 shows the number of fires by fuel type, total area burned by fuel type, the area of non-forested burns, the number of structures damaged or destroyed by fire, and the number of vehicles destroyed by wildfire. Table 15 shows the acreage for each of Anderson's 13 fire behavior fuel model statewide. The FM 0 category indicates the number of fire responses by either the AFC or rural volunteer fire departments for which a fuel model was not reported in the incident report, normally because it was a false alarm or a fuel model was simply not reported. Oral communication with various fire fighting agencies around the state confirm that only about 80% of all fires are reported through the Fire Incident Reports to the National Fire Administration (NFA) a branch of the Federal Emergency Management Agency (FEMA).
Fuel Model 9 (Hardwood Litter) is the largest by area statewide at 35.84% (4,928,710 Ac/12,109,840 Ha) and shows the highest frequency of fires, second highest in number of acres burned, second highest in the number of structures either damaged or lost, and third highest in the number of vehicles destroyed. Fuel Model 10 (Timber - Litter with Understory) has the second highest wildfire frequency, highest number of burned acres, fourth highest number of structures damaged or destroyed, and second highest in the number of vehicles destroyed. Fuel Model 10 is not officially recognized in Arkansas but is used to reflect fuels which have had some type of mechanical treatment which resulted in residual debris left on the forest floor, e.g. thinning or harvesting operations. Additionally, there was no Fuel Model 10 developed from the reclassification of the Arkansas vegetation data set so no available acreage were generated. Fuel Model 8 (Closed Timber Litter) is the fourth by area statewide at 16.34% (2,246,478 Ac/5,519,596 Ha) and has the third highest wildfire frequency, third highest total acres burned, and third highest destroyed or damaged structures. The fuel map of District 6 (Figure 4) is a re-sampled into a smaller region from the data set of the fuel map of the state of Arkansas (Figure 5).
Table 14. Statewide Fire Incident Occurrence by Fuel Type (92-96)
|
|
TBurned Ac. (Percent) |
|
|
Fuel Model 1 (Short Grasses) is second in area statewide at 24.57% (3,378,436 Ac/8,300,817 Ha) and has the fourth highest wildfire frequency, fourth in total acres burned, most damaged or destroyed structures, and most damaged or destroyed vehicles. Fuel Model 3 (Tall Grasses) ranked third in area statewide at 18.65% (2,565,299 Ac/6,302,940 Ha) and ranks second in number of structures destroyed or damaged, and ranks third in the number of vehicles damaged or destroyed. Fuel model 3, 8, and 9 represent 70.83% of the total vegetated landcover in Arkansas.
Table 15. Anderson's Fuel Models in Arkansas by Area
| 1 Short Grasses (1 foot) | ||
| 3 Tall Grasses (2 1/2 foot) | ||
| 4 Brush mixed with Cedar | ||
| 6 IntermediateBrush/Hardwood | ||
| 8 Closed Timber Litter | ||
| 9 Hardwood Litter | ||
| 14 Water | ||
| TOTAL |
AFC District 6 Fire Incident Report
AFC Fire District 6 incident report records (Table 16) show the number of fires, total acreage burned, forested acreage burned, and non-forested acreage burned by year. The largest fire in District 6 was 700 acres with 8 fires greater than 250 acres and 17 fires greater than 100 acres. The largest forested acreage burned was 700 acres and the largest non-forested acreage burned was 100 acres. Of the 847 fires in district 6 171 of these fires occurred in non-forested areas.
Analysis of recent fire incident reports for District 6 from 1992 to
1996 shows a increase similar to the statewide trends, from 163 fires annually
to 199 in 1996. The average fire size in District 6 is 6.11 hectares (15.10
acres), and was similar to the statewide average of 4.89 hectares (12.09
acres). District 6 makes up 7.34% of all fire districts and has 6.13% of
the fires.
Table 16. Summary of Wildfires 1992 - 1996 in AFC Fire District 6
|
|
|
| |
Table 17 shows the frequency of fires in Fire District 6 by Fuel Models (FM) as was reported in the incident report. The FM 0 category indicates the number of fire response by either the AFC or rural volunteer fire department for which a fuel model was not reported in the incident report, normally because it was a false alarm or a fuel model was not reported. Normal prescribed or control fires were not reported. Informal communication with various fire fighting agencies around the state confirm that only about 80% of all fires are reported through the Fire Incident Reports to the National Fire Administration (NFA) a branch of the Federal Emergency Management Agency (FEMA).
Fuel Model 9 ((Hardwood Litter) ranks first in area in District 6 at
51.14% (517,988 Ac/1,272,697 Ha) and ranks first in wildfire occurrence
in District 6, tied for first for the most structures and vehicles damaged
or destroyed. Fuel Model 9 has the second highest number of occurrence of
structures at 25.51% (11,573). Fuel Model 3 (Tall Grasses) ranked second
by area in District 6 at 38.35% (388,459 Ac/954,454 Ha) and is tied for
first in the number of structures damaged or destroyed. Fuel Model 3 has
the highest occurrences of structures at 64.81% (27,267). These two fuel
models make up 89.5% (906,447 Ac/2,227,151 Ha) of the total vegetated landcover
in District 6 and contains 92.32% (38,840) residences.
Table 17. Fire District 6 Incident Occurrence by Fuel Type
|
|
TBurned Ac. (Percent) |
|
|
In District 6, Anderson's Fuel Model 9 (Hardwood Litter) is the largest at just over 1/2 million hectares (51.14%). The second most abundant fuel model is Fuel Model 3 (Tall Grasses), at under 1/2 million hectares (38.36%) and 2.6 million hectares (18.65%). These two fuel types makeup 89.5% of the landcover in District 6 (Table 18).
Table 18. Anderson's Fuel Models in AFC District 6 by Area
| 1 Short Grasses (1 foot) | ||
| 3 Tall Grasses (2 1/2 foot) | ||
| 4 Brush mixed with Cedar | ||
| 6 IntermediateBrush/Hardwood | ||
| 8 Closed Timber Litter | ||
| 9 Hardwood Litter | ||
| 14 Water | ||
| TOTAL |
A. Ignition Potential Component:
Sources of ignition for the 1992 to 1996 data (Table 19) was: Incendiary 50.3%, (59.8% by area), and Debris Burning at 25.6%, (21.4% by area). These results were similar to those found by Kluender et al. in a 1988 study (Table 20). In that study incendiary ranked 1st at 47%, (64% in area burned), slightly above the 5-year average for the SE United States was 44%. Kluender (1988) found debris burning was second at 27% (18% in area burned), the 5-year average for SE United States is 30% (US Forest Service 88). Table 21 indicates that almost 3/4 of all fires are a direct result of human activity, not only in Arkansas' two separate studies but also in the National averages.
Table 19. Source
of Ignition in Arkansas (92 - 96 Fire Incident Reports)
| Ignition Source |
| |
| Incendiary | ||
| Debris Burning | ||
| Equipment Use | ||
| Smokers | ||
| Lightning | ||
| Railroads | ||
| Children | ||
| Camp Fires | ||
| Miscellaneous |
Table 20. Source of Ignition in Arkansas (Kluender et.al 88)
| Ignition Source |
| |
| Incendiary | ||
| Debris Burning | ||
| Smokers | ||
| Equipment Use | ||
| Children | ||
| Railroads | ||
| Lightning | ||
| Camp Fires | ||
| Miscellaneous |
Table 21. Human Activity as a Source of Ignition
| Ignition Source | Kluender (83-87) | Analysis (92-96) | National (83-87) |
| Incendiary | 47% | 50.3% | 44% |
| Debris Burning | 27% | 25.6% | 30% |
| TOTAL | 74% | 75.9% | 74% |
The Ignition Potential Component data set comprised of historical fires, each were buffered to 20 acre in size, this represents the most common sized fire occurring in District 6. The Ignition Potential area in District 6 is 1.69% (43,535 Ac/17,214 Ha) of the total area and contains 705 residences. Since residences are a source of ignition you should expect that all residences are at risk from fire. For the purposes of this study, which is to develop predictive models of residences at risk, including residences in the Ignition Potential Component model would be of little value. See Figure 6 for a map of Ignition Potential in District 6.
The Fuel Potential Component was reduced in the number of categories by looking at the rate-of-spread described by Anderson (Table 2) for each fuel type. This volatility class contained three categories, low, moderate, and high.
Of the 41,700 rural residences located in District 6, Table 22 shows the distribution of these residences among the different volatility classes. The None category are residences which do not reside in a volatility class, mainly due to the fact that the edges of different data sets do not exactly match. The None category represents less than 0.2% of the total number of residences. The Water category may represent those residences which overlap areas with water, the minimum mapping unit for each residence is 30m x 30m. Again the Water category represents less than 0.2%. The Moderate volatility class contains the majority of residences, 39,916 (95.72%). During the 1992 to 1996 fire seasons, 453 (53.23%) wildfire occurred in the Moderate volatility class. All of the structures and all the vehicles destroyed by wildfire during the same time period were located in the Moderate volatility class. It should be noted that while the Moderate class has the highest incident of wildfire it also is the largest of all volatility classes (Tables 22 and 23). See Figure 7 for a map of District 6 Fuel Potential Components.
Table 22. Residences
vs. Volatility Classes
| Volatility Class |
|
|
| NONE | ||
| LOW | ||
| MODERATE | ||
| HIGH | ||
| WATER | ||
| TOTAL |
Table 23. Volatility Classes in AFC District 6 by Acreage
| Volatility Class | ||
| LOW | ||
| MODERATE | ||
| HIGH | ||
| WATER | ||
| TOTAL |
Once a fire is in progress any attempt to save rural residences falls mainly to the local rural fire departments. These departments carry water to the fire and have specialized equipment necessary to suppress a structure fire. The AFC on the other hand uses different techniques to combat a wildfire, primarily by establishing fire breaks and then setting backfires to remove fuels from the wildfire path. The amount of risk to a rural residence is in part the ability for equipment and firefighters to access the property. Four different categories make up the Access Potential and are based upon the road type: permanent (paved and all-weather gravel roads), seasonal (graded roads), limited (non-maintained roads), and none (residences more than 300 feet from a road) (Table 24). Note: All roads were buffered 300 feet to account for driveways, those residences that are more than 300 feet are considered remote and do not receive fire protection. See Figure 8 for a map of District 6 Access Potential Component.
Table 24.
Residences vs. Access Potential Component
| Access Potential | ||
| All Weather Roads | ||
| Seasonal Roads | ||
| Limited Roads | ||
| No Access > 300 ft. | ||
| TOTAL |
D. Response Potential Component
Access Potential only looks at the type of road used by rural fire departments to access the fire scene, the other issue is how long it might take the crew to reach the residence from the fire station. Response Potential Component contains locations of all rural fire departments in AFC District 6. Each rural fire department was buffered in 5 and 10 mile radius out from the station location. The 5 mile criteria is based upon agreements between the local rural volunteer fire department and the state and federal surplus equipment programs. Of the approximately 1,000 fire departments in Arkansas fully 80% are volunteer, these fire departments typically have small operating budgets. The AFC uses various federal grant and surplus programs to place surplus fire trucks and hand equipment at these fire stations. The local rural fire department must respond to any fire within 5 miles of their station. Some fire departments actively solicit membership for a small annual fee to provide protection to those members. In many cases if a volunteer fire department responds to a non-member fire they may submit a suppression bill to the homeowners insurance company or to the homeowner. Many different kinds of agreements exist between neighboring fire departments and state agencies.
Table 25 shows the number of residences within 5 miles of a fire department and those residences that are outside the 5 mile response area. The primary issue is will the fire truck reach the residence in time to extinguish the fire. NOTE: One primary limitation on response are inadequate or unsafe bridges. This aspect is of note but beyond the scope of this project. See Figure 9 for a map of District 6 Response Potential Component.
Table 25. Residences vs. Response Potential Component
| Response Potential | ||
| Within 5 Miles | ||
| Greater than 5 Miles | ||
| TOTAL |
Table 26 shows the matrix used to fuzzy a large number of variables into three, high risk, moderate risk, and low risk. A map (Figure 10) was built using 'if statements' in GRASS (Geographic Resources Analysis Support System) 4.1 GIS Software. Table 27 shows the number of residences and the area of each risk level. The no data numbers occurred when some of the residences (30x30 meter pixel size) plotted within incorporated areas and along edges of the map.
Table 26. Threat Matrix Values
IGNITION POTENTIAL |
|
|
|
| HIGH | |||
| LOW |
Table 27. Threat Matrix Vs. Area and Residences
| Risk Level | |||
| No Data | |||
| Low Risk | |||
| Moderate Risk | |||
| High Risk | |||
| Total |
Table 28 shows the different levels of risk within the Suppression Matrix. The Suppression matrix was used to fuzzy a large number of variables into three, high risk, moderate risk, and low risk. Table 29 shows the number of residences and the area of each risk level. The no data numbers occurred when some of the residences (30x30 meter pixel size) plotted within incorporated areas and along edges of the map. A map (Figure 11) was built using 'if statements' in GRASS (Geographic Resources Analysis Support System) 4.1 GIS Software.
Table 28. Suppression Matrix Values
RESPONSE POTENTIAL |
|
|
|
|
WITHIN 5 MILES |
||||
OUTSIDE 5 MILES |
Table 29. Suppression Matrix Vs Area and Residences
| Risk Level | Area | Percent | Residences | Percent |
| No Data | 123 | 0.01% | 14 | 0.03% |
| Low Risk | 462,185 | 20.23% | 34,327 | 82.32% |
| Moderate Risk | 41,104 | 1.8% | 1,796 | 4.31% |
| High Risk | 1,781,767 | 77.97% | 5,563 | 13.34% |
| Total | 2,285,178 | 100.00% | 41,700 | 100.00% |
Table 30 shows the different levels of risk within the Risk Potential Matrix. The Suppression matrix was used to fuzzy a large number of variables into three, high risk, moderate risk, and low risk. Table 31 shows the number of residences and the area of each risk level. The no data numbers occurred when some of the residences (30x30 meter pixel size) plotted within incorporated areas and along edges of the map. A map (Figure 12) was built using 'if statements' in GRASS (Geographic Resources Analysis Support System) 4.1 GIS Software.
Table 30. Risk Potential
Matrix Values
IGNITION/ FUEL POTENTIAL |
|
|
|
| HIGH | |||
| MODERATE | |||
| LOW |
Table 31. Risk Potential Matrix Vs Area and Residences
| Risk Level | ||||
| No Data | ||||
| Low Risk | ||||
| Moderate Risk | ||||
| High Risk | ||||
| Total |
Chapter 1 | Chapter 2 | Chapter 3 | Chapter 4 | Chapter 5 | home
