EFFECTIVENESS OF WATER-RETAINING POLYMER AS FIRE RETARDANT IN INDIRECT USE

The use of fi re retardants increases effi ciency in fi ghting forest fi res, however, it still presents uncertainties regarding environmental contamination, recommendations for preparation, and it lack of regulation in Brazil. In this scenario, alternative products such as water-retaining polymers, that can reduce the rapid evaporation of water, can also have positive eff ects in terms of reducing fi re behavior. Effi ciency and ways of using the water-retaining polymer as a short-duration fi re retardant (indirect combat) in controlled burns in eucalyptus plantations were evaluated. Five concentrations (dilution in water), three volumes of spray solution, and two post-application times on the combustible material available in the area were evaluated. Controlled burns were conducted downwind, between 10 am and 2 pm, during dry season in the region, with micrometeorological and fi re behavior assessments (fi re propagation speed and length of fl ames). Increased spray volume and concentration of water-retaining polymer led to reductions in the spread of fi re. In eucalyptus combustible material, the water-retaining polymer can be used as a fi re retardant of short duration (eff ective up to two hours after application), considering a spray volume of 2.0 L m and concentration of 0.0060% (diluted


1.INTRODUCTION
Forest fi res are worrisome events on global, regional and local scales, due to the countless economic, social and environmental impacts they generate in the areas they aff ect. This scenario indicates the need to use operational management, prevention and combat systems that integrate the behavior of fi re and its circumstances such as meteorological conditions, characteristics of the combustible material and topographic conditions (SOARES and BATISTA, 2007;BATISTA et al., 2013).
Knowledge of fi re behavior can be used to establish procedures for training of fi refi ghters, activities to prevent ignitions of anthropic origin, defi nition of levels of readiness and pre-positioning of the means of suppression, design of tactics and strategies for suppressing fi re, and in the planning and execution of controlled fi res.
In this context, several methods of direct and / or indirect fi ghting of forest fi res have been developed; in this case, attention is given to chemical fi rebreaks composed of fi re retardants (RIBEIRO et al., 2006;FILHO et al., 2012). The retardants are products associated with water, which inhibit the preheating and ignition of combustible material by conserving moisture for a prolonged period (LIODAKIS et al., 2003;CANZIAN et al., 2016;PLUCINSKI et al., 2017).
In the indirect fi ght against forest fi res, the retardants available on the market are generally eff ective, but present drawbacks such as environmental impacts, high costs and lack of information regarding concentrations, spray volumes and duration of effi cacy after application (DIETRICH et al., 2014;CANZIAN et al., 2016), associated with the lack of legislation relevant to Brazilian conditions (IBAMA, 2018).
In the search for solutions to minimize the severity of forest fi res (in native or planted areas) and in crops (mainly in corn stover), alternative products have been studied as fi re retardants. The water-retaining polymers (hydrogels) used in the agroforestry sector to retain moisture in the soil have characteristics that indicate the potential to conserve moisture in combustible forest materials, thus being able to infl uence the variables of fi re behavior (BOURBIGOT and DUQUESNE, 2007;SOUZA et al., 2012).
The objective of this work was to evaluate the eff ectiveness and ways of using water-retaining polymer as a fi re retardant, in indirect use, in controlled burns of combustible material from hybrid clones of Eucalyptus grandis x Eucalyptus urophylla cl. H13, in the Cerrado-Amazon transition of the state of Mato Grosso.

2.MATERIALS & METHODS
The study was carried out in the municipality of Sorriso, MT (12 ° 51'35.04" S and 55 ° 52'33.54" W, fl at relief and 365 m altitude), in July and August 2017. The region is located in the Cerrado-Amazon biome transition. According to the Köppen classifi cation, the climate is of the hot and humid (Aw) tropical type, with two well-defi ned seasons: dry (May to September) and rainy (October to April). The average annual precipitation is 1.940 mm and average monthly air temperature ranges from 22.0 to 25.0 ° C (SOUZA et al., 2013).
The experiment was carried out in the center of stands (minimizing edge eff ect) of hybrid clones of Eucalyptus grandis x Eucalyptus urophylla cl. H13, 6 years old and with 3.0 x 3.0 m spacing (line x interplant). The trees were subjected to pruning, had an average total height of 26.0 m and a percentage of crown occupation of 62.0%. To the east of the experimental area was a fragment of native riparian forest and to the west farmland with successive soybean, corn and cotton crops (ALVES et al., 2017).
The experimental design was a three-level factorial: 6 x 3 x 2 (concentrations x spray volumes x time after application), with three repetitions (subplots) per treatment. In the random distribution (drawing) of the experimental units (subplots), interactions between concentrations and spray volumes were considered at the fi rst level, and at the second level the post-application time (1.0 and 2.0 h) of the retardant over combustible material (Figure 1).
Plots of 25.0 x 3.0 m (length x width) were installed, composed of subplots of 3.0 x 3.0 m, interspersed "without" and "with" the application of the fi re retardant (water-retaining polymer). The subplots were divided into three repetitions of 1.0 x 3.0 m (length x width), where points of observation of the behavior of the fi re were fi xed (ALVES et al., 2017). Between the subplots, transitions were installed (zones of fi re extinguishing / fi re line) of 1.0 x 3.0 m (length x width) to eliminate residues from the application of the retardant and start of another line of fi re (another subplot).
The eff ectiveness of water-retaining polymer (commercial Nutrigel ® ) as a fi re retardant was evaluated. It is composed of methylcellulose and 27.80; 49.70; 8.70 and 18.10% of CaO, CaCO 3 , MgO and MgCO 3 , respectively, with a neutralization power of 67.50% (classifi cation similar to dolomitic limestone).
Six concentrations of the water-retaining polymer diluted in water (0 -water; 0.0010; 0.0025; 0.0050; 0.0075 and 0.0100%) in three spray volumes (0.5; 1.0 and 2.0 L m -²) were evaluated. The concentrations were determined (between 0.1 and 1.0 g L -1 ) in tests that aimed to avoid clogging of the fan-type nozzles used in fi reproof backpack sprayers. Subplots without retardant / water were considered controls.
The available fuel material was characterized by random collection of samples of 1.0 m², at most 3.0 m from the subplots, on the same planting line and on the same day as the controlled burning, with one sample per subplot. The average thickness of the combustible material layer (litter) was obtained, with subsequent separation of the plant partitions in the classes: i) dead (dry) combustible material: leaves, barks, thin branches with diameter (d) <0.7 cm and branches mean 0.7 ≤ d ≤ 2.5 cm; and ii) live (moist) combustible material composed of herbaceous and grassy plants (ALVES et al., 2017;ALVES et al., 2018). The classes of the combustible material were weighed in the fi eld to obtain the fresh wet mass. Subsequently, they were subjected to drying in an oven with forced air circulation at a temperature of 65 ° C (± 2 ° C), until reaching constant mass, to determine the moisture content of the combustible material.
Micrometeorological variables (air temperature, relative air humidity, wind speed and direction) were monitored every minute during burns controlled with a portable automatic weather station (Instrutemp Weather Station, model ITH1080), suspended at a height of 2.0 m in the center of the planting. Controlled burns occurred downwind between 10 am and 2 pm (local solar time), times with greater fi re intensity and few variations in zenith angles (LIMA et al., 2017).
Fire behavior was assessed using the following variables: i) fi re propagation speed (PS: m min -1 ), timing the fi re line travel time between two consecutive observation points; ii) length of the fl ames (L: cm), given by the visual estimate with a ruler attached to the observation point at the time of the fi re. Subsequently, the rates of reduction in fi re propagation speed (RRPS) and fl ame length (RRL) were defi ned, considering as reference the subplots that did not receive retardant (hydrogel) or water.
The remaining post-burn fuel material was collected by means of a random sample of 1.0 m² in the subplots of the controlled burns. After collection, the samples were dried in a forced air oven at a temperature of 65.0 ° C (± 2 ° C), until reaching constant mass, to estimate the percentage of total dry mass of fuel material remaining and consumed during burning.
The normality of the data (residues) was assessed by Shapiro-Wilk test, with subsequent analysis of variance (ANOVA) for the variables of the available fuel material and the factorial (concentrations x spray volumes x times after application of the retardant). Signifi cant diff erences between means were compared by the Scott-Knott test at 5.0% signifi cance. The determination of the ideal concentration of water-retaining polymer was obtained by means of regression analyses (p ≤ 0.05), between the fi re behavior variables (dependent variable) and the retardant concentrations (independent variable) with a spray volume of 2.0 L m -2 at post-application times of 1.0 and 2.0 hours.

Available fuel material and weather elements
The available fuel material was homogeneous between the treatments, since no signifi cant diff erences (p> 0.05) were observed between layer thickness, percentage distribution, total dry mass and moisture content in the plots used in the evaluation of the diff erent concentrations of the water-retaining polymer and its interactions with the spray volumes and post-application times (Table 1).
Regarding variations in air temperature, relative humidity and wind speed (Figure 2), monitored instantaneously during controlled fi res, the fi res occurred under similar weather conditions, when comparing the average values obtained for the diff erent spray volumes applied and presence / absence of the water-retaining polymer. The average values of air temperature, relative humidity and wind speed ranged from 29.1 to 34.8 ° C; 18.5 to 33.0% and 0 to 2.7 m s -1 , respectively.

Fire behavior
The fi re behavior described by the fi re line propagation speed (PS) and the length of the fl ames (L), as well as their reduction rates, showed diff erences between the treatments in the unfolding of the triple factorial (concentrations, spray volumes and post-application times) of the water-retaining polymer (Table 2), indicating together that the preparation and application infl uence the product's eff ectiveness as a retardant. Given the higher effi cacy of the water-retaining polymer in the spray volume of 2.0 L m -2 , the ideal product concentration for this spray (based on the pre-determined concentrations of the polymer) was determined through adjustments of quadratic polynomial models with correlations greater than 80.0% in the two post-application times (Figure 3).

3.3.Survey of remaining and consumed fuel material
In the survey of the post-burn combustible material, the control (without application of retardant and water) showed a higher percentage of combustible material consumed by fi re (above 90.0%) ( Figure 3). As expected, at the most eff ective concentrations of the water-retaining polymer (0.0050 and 0.0075%, in the spray volume of 2.0 L m -2 , after 1.0 h of application), the combustible material consumed (23.0 %) was lower than the remaining fuel material (77.0%) ( Figure 3I). However, after 2.0 h of application of the water-retaining polymer, an inversion of this relationship occurred (fuel material consumed and remaining 80.0 and 20.0%, respectively) ( Figure 3J).

Available fuel material and weather elements
In all treatments, the percentage distribution of the classes of available fuel material showed a higher leaf composition, followed by medium branches, thin branches, barks and herbaceous material. The average total dry mass of the fuel material was 27.4 t ha -1 , composed of 11.2; 4.5; 8.6; 2.3 and 0.9 t ha -1 for leaves, thin branches, medium branches, bark and herbaceous material, respectively.
These values corroborate other studies of Eucalyptus aged between 5 and 7 years during the dry season (CORRÊA et al., 2013;CARMO et al., 2018). The combustible material available in the experimental area had a high ignition hazard due to the number of leaves, which are of a class with less timelag (time required for the loss of moisture from the fuel to the environment) (ALVES et al., 2018).
In the same area of cultivation, when the plants were 4.5 years old, Alves et al. (2017) obtained an average total dry mass of 14.0 t ha -1 , while Carmo et al. (2018) evaluating the same genetic material in this region, in areas aged 7 years, observed an average total dry mass of 31.0 t ha -1 in the composition of the litter (combustible material), both in the month of August. The increase in total dry mass observed in the experiment is common Table 2 -Fire behavior and rate of reduction in interactions of water-retaining polymer concentrations with spray volumes and postapplication times. Tabela 2 -Comportamento do fogo e taxas de reduções nas interações das concentrações do polímero hidroretentor com os volumes de calda e tempos pós-aplicação.
* Concentration of 0% considers only the application of water, while the other concentrations consider the water-retaining polymer dissolved in water; averages with the same lowercase letter in each row and uppercase in each column do not diff er by Scott-Knott test (p> 0.05). PS: speed of fi re spread; L: length of the fl ames; RRPS: rates of reduction of PS; RRL: rates of reduction in L; Control: portion without application of retardant / water ". * A concentração de 0% considera apenas a aplicação de água, enquanto as demais concentrações consideram o polímero retentor de água dissolvido em água; médias com a mesma letra minúscula em cada linha e maiúscula em cada coluna não diferem pelo teste de Scott-Knott (p> 0,05). PS: velocidade de propagação do fogo; L: comprimento das chamas; RRPS: taxas de redução do PS; RRL: taxas de redução em L; Controle: porção sem aplicação de retardante / água ".  in cultures of hybrid clones of Eucalyptus grandis x Eucalyptus urophylla cl. H13, aged 5 to 7 years, in the dry season (CARMO et al., 2018), in response to the high deposition of combustible material with the growth of the trees.
The moisture content of the fuel material was less than 15.0%, except for herbaceous materials (ranged from 20.7 to 43.1%). These values are considered lower than fi re extinguishing moisture (25.0 to 30.0%), indicating a high risk of fi re occurring in the area during the experimental period (SOARES and BATISTA, 2007). The observed values were higher than those found by Alves et al. (2017) for the same area, at an age of 4.5 years, in the month of August (around 8.0% in the same fuel material partitions).
The increase in moisture content with the age of the forest stand may result from lower rates of evaporation of water from the soil with the deposition of combustible material (MATEUS et al., 2013;SLIJEPCEVIC et al., 2018), since the litter of eucalyptus presents low decomposition rates (CARMO et al., 2018). The evaporation process depends on the transfer of water from the topsoil to the plant fragments in the combustible material. The movement of water from the soil occurs by capillary action, and is interrupted with the increase of the porous space in the fragments of the combustible material (SHARPLES and McRAE, 2011). Therefore, changes in energy balances occur in the fuelatmosphere limit layer, causing surface fragments to dry faster compared to fragments close to the ground (HOFFMANN et al., 2012), altering the behavior of fi re in controlled burns.
Micrometeorological conditions (Figure 2) associated with the characteristics of the combustible material in the study area indicate that the environment presented a high risk of forest fi re (HOFFMANN et al., 2012;ALVES et al., 2017). The environmental scenario reinforces the importance of fi re behavior forecasting models and indirect fi ghting methods, such as the use of retardants.

Fire behavior
The descriptive variables of the fi re behavior showed higher values in the control (without application of retardant or water) in function of the real humidity conditions of the combustible material. Similar values were observed by Alves et al. (2017) in the same planting of hybrid clones of Eucalyptus grandis x Eucalyptus urophylla cl. H13, 4.5 years old in the month of August, with PS of 0.74 m min -1 and L of 100 cm.
The application of water only (concentration of 0%) reduced the behavior of fi re when compared to the control, however, it was less eff ective when compared with the applications of the concentrations of the waterretaining polymer, regardless of the volume of spray solution applied. In this case, the post-application time may have infl uenced the water evaporation process (faster) in the microclimate conditions of the area (BOURNE et al., 2015;PLUCINSKI et al., 2017).
Regarding the post-application time of 1.0 h of the water-retaining polymer on the combustible material, the most eff ective concentrations were 0.0050 and 0.0075%, with maximum reduction of PS and L in the spray volumes of 0.5 and 1.0 L m -2 , with extinction of fi re at 2.0 L m -2 . However, after 2.0 h of the application there was a tendency to increase PS and L in all concentrations and spray volumes. The rates of reduction of PS and L obtained in comparison with the control, were 100% in these concentrations and spray volume.
In this case, the water-retaining polymer penetrated evenly among the plant partitions, leading to increased conservation of moisture in the fuel due to the greater adherence between the water present in the waterretaining polymer molecules and the forest fuel. This characteristic makes the water-retaining polymer eff ective as a fi re retardant, since it hinders the water evaporation process and reduces the behavior of fi re (GIMÉNEZ et al., 2004;RIBEIRO et al., 2006;PLUCINSKI and PASTOR, 2013).
The concentration of 0.0100% in all spray volumes was less eff ective than the concentrations of 0.0050 and 0.0075%; in this case, this higher concentration allowed the formation of a gelatinous layer (lumps of gel), (BALENA, 1998) after 1.0 h of application of the product, which in turn, maintained the moisture only in the surface layer of the combustible material. This behavior diff ered from other studies with retardants, in which increasing product concentrations generated greater effi cacy (RIBEIRO et al., 2006;FIEDLER et al., 2015;CANZIAN et al., 2016).
Among the factors evaluated, the spray volume was determinant for the maximum eff ectiveness of the water-retaining polymer as a retardant in hybrid clones of Eucalyptus urophylla x Eucalyptus grandis cl. H13. Similar results were observed by Batista et al. (2008) in controlled fi res of Pinus taeda L., where application of 0.5 L m -2 retardant spray (Phos-chek ® ) caused a reduction in fi re behavior, whereas increasing the spray to 1.5 L m -2 led to extinction of the fi re in the plots. In this scenario, when applying larger spray volumes, one should consider the availability of water resources for the capture of water in the aff ected areas (FIEDLER et al., 2015;CANZIAN et al., 2018).
The interaction between the largest spray volumes and the most eff ective concentrations of the waterretaining polymer generated greater conservation of moisture in the combustible material. In this case, the applied product inhibited the evaporation of water from the fuel for a prolonged period in response to the slow release of water present in the water-retaining polymer molecules (BALENA, 1998;SOUZA et al., 2012;FIEDLER et al., 2015). By diff erentiating the polynomials, 0.0060% is defi ned as the ideal concentration of water-retaining polymer for application without waste in chemical fi rebreaks for indirect fi refi ghting within 2.0 h after its application on the combustible material of hybrid clones of Eucalyptus urophylla x Eucalyptus grandis cl. H13.
The post-application time of the water-retaining polymer on the combustible material, in general, did not impact the eff ectiveness of the retardant, however fi re extinction was observed after 1.0 h of application of the product. Similar behavior was observed by Souza et al. (2012) applying water-retaining polymer spray of 2.0 L m -2 at a concentration of 0.0010% on Melinis minutifl ora P. Beauv., observing the conservation of moisture up to 24 h after application of the product on the fuel. In this case, the type of combustible material can determine the conservation of moisture for a prolonged period in the forest fuel.

Survey of the remaining and consumed fuel material
In general, after 1.0 h of application of the product at concentrations of 0.0050 and 0.0075%, the fi re was extinguished, thus justifying higher values of the fuel material remaining post-burn. However, in the plots with burns carried out after 2.0 h of application of the product, the fi re continued to spread slowly, leading to an increase in the consumption of fuel material available in the plots. Similar results were obtained by Ribeiro et al. (2006)

in controlled burns of Brachiaria decumbens
Stapf. with application of retardant (Phos-chek ® ) at a concentration of 13.4% in the spray volume of 1.2 L m -2 , where they obtained fuel material consumed below 2.0% due to the extinction of the fi re in the plots. However, in the application of smaller spray volumes, there was a reduction in the behavior of fi re and a higher percentage of combustible material consumed, as also evidenced by Canzian et al. (2016).

5.CONCLUSIONS
Water-retaining polymer was eff ective and can be used as a fi re retardant for indirect use, including in controlled burning in plantations of hybrid clones of Eucalyptus urophylla x Eucalyptus grandis cl. H13, when applied at a spray volume of 2.0 L m -2 and a concentration of 0.0060% (diluted in water) within 2.0 h after application on the combustible material.
The increase in spray volume increases the eff ectiveness of the fi re retardant; however, it is emphasized that the availability of water can be a limiting factor in the fi ght against forest fi res.