Northern Prairie Wildlife Research Center
Larson, Diane L., and Wesley E. Newton. 1996. Effects of fire retardant chemicals and fire suppressant foam on North Akota prairie vegetation. Proceedings of the North Dakota Academy of Science. 50:137-144.This resource should be cited as:
Larson, Diane L., and Wesley E. Newton. 1996. Effects of fire retardant chemicals and fire suppressant foam on North Akota prairie vegetation. Proceedings of the North Dakota Academy of Science. 50:137-144. Northern Prairie Wildlife Research Center Online. http://www.npwrc.usgs.gov/resource/othrdata/fireweb/larson.htm (Version 02MAR98).
Fire suppressant foams and fire retardant chemicals are used in wildland fire control and in prescribed burns for habitat management. In 1988 alone, more than 5 million gallons of retardant chemicals were dropped from aircraft on wildland fires in the western United States (1). Class A foams, the type typically used in wildland fire suppression, also receive widespread use in constructing fire lines for prescribed burns. Some 210,000 gallons of concentrate -- enough to make 42 million gallons of foam -- were sold in 1992 (C. Johnson, pers. comm.). Despite their relatively widespread use, little is known about their potential effects on terrestrial and aquatic ecosystems. A literature search at the beginning of this study revealed only two published scientific articles on ecological effects of fire retardant chemicals used for wildfire suppression: one on aquatic toxicity (2) and one on annual grassland response (3). No studies have been published on ecological effects of Class A foams.
Because these chemicals most often are used in natural areas and areas set aside for wildlife, the fire suppression community has identified a need to determine their potential effects on ecosystems. In response to this need, three National Biological Service Science Centers have undertaken research to determine ways in which fire retardants and foams affect both terrestrial and aquatic ecosystems. Patuxent Environmental Science Center took the lead on terrestrial bird, small mammal, and invertebrate responses; Columbia Environmental Research Center has evaluated effects on aquatic organisms; and Northern Prairie Science Center has examined responses in terrestrial plant communities. Taken together, these studies provide land managers and fire control professionals with a starting point in establishing guidelines for safe use of fire retardant chemicals and fire suppressant foams.
The purpose of this portion of the study was to examine experimentally the effect of retardant and foam application on vegetation. We studied the effects alone and in combination with fire. In addition, we examined the effects of the chemicals and fire on insect herbivory, which provides a link to higher levels in the food chain. The first year's work was conducted in the structurally simple mixed-grass prairie, so that general patterns could be identified. Subsequent studies are being done in more complex habitat.
Our objectives were to estimate effects of fire suppressant foam and fire retardant chemical application on growth and species diversity of burned and unburned prairie vegetation, and to assess the response of herbivorous insects, in terms of number of insects and their effects on plants, to burning and application of foam and retardant to their host plants.
The study was conducted at the Woodworth Study Area, a research site of the Northern Prairie Science Center, Jamestown, N.D. The 65-ha field containing the study site has never been plowed. Biologists burned the field in 1969, 1970, 1971, 1972, 1976, 1979, 1981, and 1990; it has not been grazed since 1974. Currently, vegetation in the study area is dominated by Poa pratensis, an exotic cool-season grass. Other grass species found during previous studies on the site include Stipa viridula, S. comata, Agropyron repens, Muhlenbergia cuspidata, and Bromus inermis. Rosa arkansana, Elaeagnus commutata, and Symphoricarpos occidentalis are common woody plants.
We used one Class A foam -- Silv-Ex -- and one fire retardant -- Phos-Chek G75-F -- in our field tests. Silv-Ex concentrate is a proprietary mixture of sodium and ammonium salts of fatty alcohol ether sulfates, higher alcohols, and water, as well as butyl carbitol and ethyl alcohol (4). It functions as a surfactant, allowing water to penetrate and expand over the surface of fuels so that they retain moisture longer. Silv-Ex, like other Class A foams, is applied operationally either from ground tankers or helicopters.
Phos-Chek G75-F is a formulation composed of monoammonium phosphate and ammonium sulfate, fugitive coloring agent, and small amounts of gum-thickener, bactericide and corrosion inhibitor (5). It is typically applied from helicopter bucket or ground tanker in advance of a fire; other retardants with higher viscosity are applied from fixed-wing aircraft. The ammonium salts retard fire by chemically combining with cellulose as fuels are heated, as well as through evaporative cooling of the fuels (6).
We delineated a grid of 30 0.4-ha blocks in the study field (Figure 1). Each block was separated from adjacent blocks by a mowed, 5-m-wide fire break. Four treatments for the foam study [burning (B), foam application (F), burning and foam application (BF), and no manipulation (C)] were each assigned at random to six blocks. We established a 10 m × 10 m plot in the center of each of these 24 blocks for vegetation sampling (Figure 1). For the retardant study, we established five 10 m × 10 m vegetation plots in each of the remaining six 0.4 ha blocks. Four treatments [burning (B), retardant application (R), retardant application plus burning (RB), and no manipulation (C)] were assigned at random to one of the five plots within 0.4 ha blocks; each 0.4 ha block had each of the four treatments. Vegetation plots for which the treatment included burning were surrounded by mowed 1.5 m wide fire breaks.
|Figure 1. Study area, showing 0.4-ha blocks, 10 m × 10 m vegetation plots, and randomly located 1-m² permanent vegetation subplots and 0.25 m × 0.25 m biomass plots.|
Inside each 10 m × 10 m vegetation plot we randomly selected five 1-m² permanent vegetation subplots and four 0.25-m² biomass subplots (Figure 1). Prior to treatment, we counted stems of S. occidentalis and R. arkansana, counted the total number of plant species, and measured litter depth in each permanent vegetation subplot. We made all pretreatment measurements during 17 - 28 May 1993.
Retardant application and burning of retardant plots was accomplished on May 26. Representatives of Monsanto applied Phos-Chek G75-F at the rate of 1 gallon per 100 ft², resulting in approximately 12 pounds of retardant per plot, to the R and RB plots within each 0.4 ha block. This is the application rate recommended for grassland vegetation. The retardant was applied by hand using a hose from a slip-on pumping unit. We burned the RB plots after retardant had been applied and allowed to dry for 0.5 - 1 hour. B plots were also burned at this time.
On 1 June the 0.4 ha B and BF blocks were burned. All fires were allowed to burn to completion; vegetation was reduced to ash. On 10 June we applied Silv-Ex in 0.5%-solution maintained by a proportioner to F and BF blocks. The rate of application was approximately 189 liters per 10 m × 10 m plot, resulting in approximately 1 liter of Silv-Ex concentrate on each 10 m × 10 m vegetation plot. Only the vegetation plots were treated on BF blocks. The foam was applied from a 3.66-m boom mounted on bicycle tires and pushed by two people. Nozzles mounted on the boom every 30 cm each produced approximately a 1:10 expansion.
|Application of Silv-Ex fire suppressant foam at mixed-grass prairie study site in North Dakota.|
We measured the length of two fully expanded leaves on each species except P. pratensis. We measured the total length and counted the number of galls, leaf miners, aphids, chewed leaves, and flowers on each of the five shoots. Galls, leaf miners, aphids, and chewed leaves were recorded on a per-leaf basis. In each permanent subplot, we counted the total number of plant species and measured litter depth at four locations. Total stems of S. occidentalis, R. arkansana, and S. rigida were also recorded in each plot at each sample period.
We conducted post-treatment vegetation sampling at 2-week intervals, beginning June 16 and ending August 27. We concentrated on four species: P. pratensis, S. occidentalis, R. arkansana, and Solidago rigida. Height of P. pratensis was measured at four locations on each subplot at each sampling period. For the other three species, we marked individual plants in each permanent vegetation subplot as follows: two S. occidentalis, two R. arkansana, and ten S. rigida. If fewer individuals were found in a subplot, we marked all found individuals. Plants were marked near the base with either blue or red flagging (R. arkansana and S. occidentalis), or numbered metal tags (S. rigida). Current year's growth was followed on five shoots through the three sampling periods on each S. occidentalis and R. arkansana plant.
Two of the 0.25 m × 0.25 m biomass subplots were clipped to ground level on 23-29 June and 7-8 September, 1993 and 14-15 July, 1994 (retardant); and 7-8 July and 9-10 September, 1993 and 11-12 July, 1994 (foam). Dead and woody vegetation was removed and discarded. Live non-woody vegetation was oven dried to constant weight and weighed.
We used analysis of variance (ANOVA) techniques in a repeated-measures type design with subsampling to assess the effects of the burn-foam treatments, time since treatment, and their interaction on all measured variables. Mean separations of significant effects in the ANOVAs were done with Fisher's protected least significant difference value (7). Analyses were made in the original scale of measurement and with a log(y+1) transformation (8), but only results in the original scale of measurements are reported because only slight differences were observed in ANOVA results. ANOVAs were done using the General Linear Models procedure of SAS (9). Significance was set at the 0.05 level.
Because vegetation plots differed significantly in number of plant species at pre-treatment in the foam study, this difference was taken into account in subsequent analysis by using the change in number of species between pre- and post-treatment as the response variable. Plots were similar in all other pre-treatment measurements for both the retardant and foam studies.
Overall, Silv-Ex application had little affect on the vegetation characteristics we measured. Effects we detected were subtle. Of the 24 response variables, only five showed significant (P< 0.05) effects of Silv-Ex application (Table 1). Change in number of species, ratio of chewed to total leaves per shoot in S. occidentalis and R. arkansana, and mean shoot length and leaf length in S. occidentalis were affected by treatment.
The number of plant species increased between pre- and post- treatment in all plots, but the increase was smaller in plots treated with Silv-Ex (_x = 1.53 + 0.26) than in untreated plots (_x = 2.34 + 0.26). Burning did not influence this difference.
Because the summer of 1993 was exceptionally cool and wet, insect abundances were low at our study site (D. Larson, personal observation). However, Silv-Ex application influenced herbivory, as evidenced by the proportion of chewed leaves on S. occidentalis and R. arkansana (Figure 2). Silv-Ex treated plants of both species experienced greater herbivory late in the season; more untreated R. arkansana leaves were chewed early in the season. Herbivory on burned Rosa was not affected by Silv-Ex.
Silv-Ex application had little affect on herbaceous plant growth, as evidenced by the lack of difference in herbaceous biomass accumulation between treated and untreated plots, irrespective of burning (Figure 3). Growth characteristics of S. occidentalis were affected, however. Leaf length increased more rapidly on plants treated with Silv-Ex than on untreated plants (Figure 4). Burning significantly enhanced the rate of shoot growth compared with other treatments; Silv-Ex tended to depress shoot growth. The decline in shoot length between June and July for Silv-Ex treated plants suggests senescence, shoot damage and subsequent breakage, or vertebrate herbivory.
Of the 24 response variables, five showed a significant effect involving Phos-Chek treatment (Table 1). Phos-Chek G75-F application resulted in increased biomass, whether or not the plots were burned (two-way nested ANOVA; F = 18.61; df = 1, 15; P = 0.0006). No interaction between retardant and burning was evident (F = 0.84; df = 1, 15; P = 0.3726). The effect was transitory, however; biomass did not differ among treatments the following year (Figure 3).
Larson and Duncan (10) documented similar changes in herbaceous biomass in a California oak-savanna rangeland after a diammonium phosphate (DAP) retardant was applied to extinguish an October fire. Herbage yield the season after application was significantly higher on plots to which DAP had been applied, whether burned or unburned. By the second season, DAP plots were statistically indistinguishable from burned, untreated plots.
The fertilization effect in our study seemed to be concentrated in P. pratensis. Grass not only was longer on plots treated with retardant, but the effect was enhanced over the course of the growing season (Figure 5). Measures of shoots and leaves on woody species, and of stem length on Solidago, did not indicate any affect on growth of these species.
Phos-Chek influenced the number of leaves per shoot in S. occidentalis (Figure 6). Early in the season, retardant produced similar changes in leaf production on burned and unburned plants. However, between July and August the relationship changed: only burned, untreated plants were still producing leaves.
Of concern to land managers is the potential depression in species richness associated with both Silv-Ex (Table 1) and Phos-Chek (Figure 7) application. The change in number of species per plot was significantly lower after Silv-Ex application, regardless of whether or not the plot was burned. The change in number of species per plot was depressed, especially between July and August, on Phos-Chek plots.
All plots were dominated by P. pratensis, which clearly benefitted from retardant fertilization, and also may have increased in response to the general disturbance, crowding out other species. Greater than average precipitation during the first growing season after treatment also may have influenced P. pratensis growth. Work in Wisconsin has suggested a larger positive response in P. pratensis to burning in mesic compared with xeric sites (10). Further work in areas not dominated by P. pratensis will help define this relation.
Implications of this research depend on the objectives of the land manager. If the objective is to halt an uncontrolled fire, subtle changes caused by Silv-Ex and Phos-Chek may be of little importance. On the other hand, if the objective is to aid in the control of prescribed burns, the potential effect on species richness should be considered. In particular, if the control of exotic, robust grasses such as P. pratensis is important, these results suggest that use of these chemicals should be avoided.