Strategic Supplementation to Manage Fine Fuels in a Cheatgrass (Bromus tectorum)–Invaded System

ABSTRACT Management of areas invaded by cheatgrass (Bromus tectorum) continues to be one of the greatest challenges for US Great Basin ecosystems. Targeted cattle grazing in the fall and winter has shown positive results as a management tool to reduce dormant fine fuel biomass within cheatgrass-invaded areas, but management of targeted grazing within large pastures can be challenging. We evaluated the use of strategically placed liquid protein supplement stations over a 4-wk period in the fall to focus cattle grazing along a linear transect stretching away from water to reduce residual cheatgrass biomass on a production-scale, working ranch from 2014 to 2017. Liquid protein supplement stations were moved approximately 1 km farther from water during each week of the study, eventually reaching 4 km from a single water source. Global Positioning System–tracked cattle visited supplement stations 52% ± 4% of the days during the study period and were within 100 m of the supplement station transect line 17.7% ± 2.6% of the time: more than 3 × greater (P= 0.07) than random locations (5.1% ± 2.6%). Week of the study and the subsequent supplement distance from water did not influence the number of visits cattle made to supplement. The duration that cattle remained at supplement was greater in wk 4, when supplement was placed 4 km from water, compared with wk 1 and 2, when the supplement was 1 and 2 km from water, respectively. At the conclusion of grazing, utilization along the supplement station transect averaged 66.0% ± 5.7% and did not differ between supplement stations at 1 km, 3 km, or 4 km from water. Strategic supplementation provides a valuable tool to target cattle grazing at specific locations within cheatgrass-invaded systems to reduce fine fuel buildup during the dormant season.


Introduction
One of the greatest ecological threats to ecosystems of the Intermountain West has been and will continue to be invasive annual grasses. Currently, cheatgrass ( Bromus tectorum L.) is the most problematic invasive annual grass at regional spatial scales, with an estimated 15% ground cover across one-third of the Great Basin (210 0 0 0 km 2 ) ( Bradley et al. 2018 ), and the ecologically dominant ✩ Mention of a proprietary product does not constitute a guarantee or warranty of the product by USDA or the authors and does not imply its approval to the exclusion of other products. * Corresponding author. species on > 20% of the sagebrush steppe ( Knapp 1996 ). Cheatgrass and other invasive annual grass −dominated areas are expanding at a rate of 2 , 300 km 2 yr − 1 in the Great Basin ( Smith et al. 2022 ) and are likely to become an even bigger problem in the future ( Davies et al. 2021a ). Cheatgrass invasion is especially problematic and increases the likelihood of frequent wildfires because it has less fuel moisture for a longer period than native vegetation and increases the amount and continuity of fine fuels compared with sites where native bunchgrasses are the primary herbaceous component ( Balch et al. 2013 ;Davies and Nafus 2013 ). Frequent fire is detrimental to most native species not adapted to a frequent disturbance regime and contributes to a positive feedback cycle between fire and increased abundance of cheatgrass ( D'Antonio and Vitousek 1992 ).
These feedbacks between fire and invasive annual grass fuel loads have caused widespread ecosystem conversions from sagebrush ( Artemisia ssp.; Young et al. 1972 ;D'Antonio and Vitousek 1992 ;Alba et al. 2015 ) and salt desert shrub ( Atriplex ssp.; West 1994 ) to annual grass −dominated systems. Perryman et al. (2018) suggested that these invaded systems should no longer be considered as degraded perennial grass understories but instead viewed as mixed annual-perennial understories.
The accumulation of fine fuels from cheatgrass, as well as the associated deep and continuous plant litter, hinders the establishment of native vegetation and creates an environment that favors increased cheatgrass germination and establishment ( Evans and Young 1970 ;Beckstead and Augspurger 2004 ). Thus, Schmelzer et al. (2014) hypothesized that the amount of standing dead biomass left on site during the fall season regulates the spectrum of dominance that cheatgrass is able to generate. Management must address buildup of cheatgrass fuel loads to both reduce fire risk and decrease the conditions favorable for the germination of cheatgrass seed and subsequent emergence and establishment of seedlings.
Dormant-season cattle grazing during the fall and winter provides opportunities to decrease standing dead biomass and reduce fuel continuity at minimal risk to native perennial herbaceous plants ( Davies et al. 2016a ). For perennial grass −dominated understories, decreased herbaceous standing dead and litter material alters fire characteristics by reducing flame height, rate of spread, and area burned in Great Basin sagebrush ecosystems ( Davies et al. 2015 ). Fall and winter grazing can also reduce invasive annual grass dominance and favor perennial grass production ( Schmelzer et al. 2014 ;Davies et al. 2021b ). Consecutive years of fall grazing can potentially reduce cheatgrass seed banks; however, the cheatgrass seed bank can rapidly recover with suspension of fall grazing ( Perryman et al. 2020 ). Annual removal of carryover fine fuels could significantly alter fuel loads and annual grass litter characteristics in strategic areas, providing land managers more options in planning for desired future outcomes ( Perryman et al. 2021 ).
Strategic fuels management (e.g., fuel reduction) is a common management strategy worldwide because fuels provide the basic energy source for fire to propagate ( Matsypura et al. 2018 ). Fuel reductions decrease fire intensity by reducing the amount of fuel and disrupting horizontal and vertical fuel continuity within a landscape, potentially reducing rate of spread ( Rothermel et al. 1972 ;Fernandes and Botelho 2003 ). One method to increase livestock grazing pressure at strategic locations to reduce fine fuel biomass within large pastures is the use of protein supplements, or a combination of protein supplements and low-stress herding ( Bailey 2004 ;Bruegger et al. 2016 ;Stephenson et al. 2016 ;Stephenson et al. 2017 ). Use of strategically placed protein supplements and the subsequent increased grazing nearby can reduce the amount of biomass carried over from one fire season to the next on arid and semiarid rangelands ( Bruegger et al. 2016 ). However, more research is needed to understand effective strategies for targeting fall and winter cattle grazing on cheatgrass-dominated systems at extended distances from water and at scales that are operationally and economically productive.
This project evaluated the strategic placement of liquid protein supplements along a single line stretching 4 km from water at a production-scale, working ranch over a 4-yr period. The goal of the project was to evaluate the efficacy of using protein supplements to manipulate cattle grazing behavior and alter dormant season cheatgrass fine fuels at varying distances from water. Specific objectives were to 1) track the efficacy of liquid protein supplements to attract cattle to strategic locations within a cheatgrass-dominated community during late fall, 2) determine the effect of moving liquid protein supplement stations progressively farther from water on cattle behavior, and 3) evaluate the reduc- tion in cheatgrass fine fuels as a result of strategically placed protein supplement stations.

Study site
Research was conducted in October and November from 2014 through 2017 at the Elko Land & Livestock TS Ranch in northern Nevada. The study area was a 2 389-ha pasture located in a relatively flat valley (0 −2% slope) near Dunphy, Nevada (40 °4 4 4 4.2"N, 116 °24 48.3"W). The study pasture is a mixture of rangeland and abandoned farmland heavily invaded by cheatgrass after farming ceased in the late 1970s. Vegetation on the site was dominated by cheatgrass with some areas also having a combination of cheatgrass and greasewood ( Sacrobatus vermiculatus [Hook.] Torr.). Present, but generally uncommon, species include a few native perennial grasses (e.g., Poa secunda Presl) and crested wheatgrass ( Agropyron cristatum L.), a relic from a failed rehabilitation seeding after the abandonment of farming. Annual tumble mustard ( Sisymbrium altissimum L.) and Russian thistle ( Salsola iberica L.) were common forbs.
Weather at the study site was typical of cold Great Basin deserts. Annual precipitation during the study period was near the long-term average (1981 −2010, 232 mm) in 2014 (237 mm) but was 30%, 28%, and 12% greater than the average in 2015, 2016, and 2017 ( Fig. 1 , Prism Climate Group ). In 2015, winter and spring precipitation (January −June) was 17% below the long-term average, but in other years precipitation was near or above average during this critical cheatgrass growth period (see Fig. 1 ). Mean annual temperatures were slightly higher than the long-term average (18.

Protein supplement stations and cattle grazing management
Cattle grazed the study area in October and November in all years of the study. The grazing herd was made up of dry cows (calves recently weaned before turning out) in the late-first or early-second trimester of pregnancy. In fall of 2014, 2015, 2016, Figure 2. The study pasture (2 389 ha) located near Dunphy, Nevada in northcentral Nevada. Fire break was strategically placed stretching away from water up to 4 km (i.e., wk 4 supplement placement). Supplement stations were moved at weekly intervals beginning at 1 km from water and moving approximately 1 km farther during each week of the study. and 2017, the number of cows turned out was 800, 805, 1250, and 655 cows, respectively.
Protein supplement used to attract cattle to strategic locations was a Loomix 31% crude protein (CP) and 1% phosphorus (P) molasses-based liquid supplement (www.loomix .com). Cattle received the supplement via eight, 160-gallon, self-limiting tanks, with four lick rollers per tank. Supplement was provided based on an estimated consumption of 0.45 kg d −1 animal −1 . Cattle were accustomed to using liquid protein supplement tanks in years before the beginning of the study.
Eight supplement tanks were placed in a rectangular formation spaced 100 m apart along either side of a 4-km long supplement line extending away from water ( Fig. 2 ). The supplement line was established in yr 1 of the study, and supplement tanks were placed at the same locations in each year of the study. Supplement tanks were stationed and filled at the week 1 location before cattle turnout in October in each year of the study. The center point of the supplement tank formation during the first week of the study was 1 km from the only water source in the pasture. The center point of the tank formation was moved approximately 1 km farther from the water source at the end of each week for 4 consecutive wk. This process created four separate supplement stations ranging from 1 km to 4 km from water during a 4-wk (28-d) grazing period during each fall, for 4 yr (2014 −2017).

Global Positioning System tracking analysis
In 2015, 2016, and 2017, Lotek 3300 global positioning system (GPS) collars (Lotek, Ontario, Canada) were fitted on 11 randomly selected cows before turnout. Different cows were selected during each year of the study. Cattle locations were recorded every 10 minutes. To evaluate how GPS-tracked cattle responded to supplement station locations along the established supplement transect, separate random lines were generated at random direction bearings stretching away from water during each year of the study in ArcGIS Pro. Random lines were created at a distance of > 600 m from the transect line based on previous research indicating that supplements were effective at increasing grazing utilization within areas up to 600 m from supplement stations ( Bailey et al. 2001 ;George et al. 2008 ). Direction bearings away from water were created with a random number generator in each year. Comparisons between the amount of time and distance GPS-tracked cattle were within 100 m of the supplement transect line and the randomly generated line in each year were analyzed using repeated measures within Statistical Analysis Software (version 7.0.2, SAS 2008).
Week of the study and treatment (i.e., supplement line or random line) were fixed effects and blocked by year of the study. A compound symmetrical covariance structure was used on the basis of the lowest Akaike Information Criterion value. The subject was the supplement placement station by treatment within year.
We also evaluated cattle location differences over time as supplement stations moved farther from the water source. Variables analyzed for each supplement placement included mean distance from water, distance from supplement station, area covered (minimum convex polygon calculated daily in ArcGIS), mean number of days cattle visited supplement, hours spent within 150 m of supplement per visit, and daily distance traveled. Lastly, GPS point density maps were generated on the basis of supplement station movement using the point density function in the Spatial Analyst toolbox within ArcGIS Pro, with GPS fix data from all years within respective study weeks. Areas near water are expected to be grazed heavily and trampled by loafing cattle, so cattle GPS points near water (i.e., within 150 m) were removed in the point density maps to better highlight use of the pasture away from water where cattle were more likely to be grazing ( Raynor et al. 2021 ). Data from the point density analysis were presented within 10 equal interval classifications for the number of GPS fixes per pixel.

Standing crop estimates
Standing crop data were collected along the supplement transect line at five locations near each supplement station location. Data were collected at the center of each supplement station (0 m) and at 150 m and 300 m in both northeast and southwest directions along the supplement transect away from the center location ( Fig. 2 ). Standing crop data at all supplement stations were collected five times each year, once before grazing and at the end of each supplement placement movement period (wk 1 −4). This provided for cheatgrass utilization to be measured as supplement stations were moved incrementally farther away from the stationary, sole water source.
During each data collection period, three, 0.25 m 2 quadrats were randomly placed at each distance along the supplement transect line. Beginning with the second sample period, quadrats were placed approximately 2 m distant than the previous week to avoid resampling the same location. All dormant, standing cheatgrass plant biomass was clipped at ground level, placed in paper bags, dried at 60 °C for 48 h, and weighed. Utilization was determined by dividing collected standing crop material at the end of each week by the initial, pregraze sample period and subtracting from 100. Thus, utilization included the percent of cheatgrass standing crop that was either consumed or trampled by cattle.
Standing crop and utilization data were analyzed with a repeated measure analysis of variance. Supplement station location and week of study were treated as fixed effects. Year was treated as a random variable, and the subject for the repeated measure was the supplement station within year. An autoregressive covariance structure, based on the lowest Akaike information criterion, was used. Residual plots were evaluated to ensure assumptions of normality and equal variance were met. We acknowledge that sites were not spatially independent because cattle had free range within the entire area during the study in each year. However, controlled treatment research projects at production management spatial scales are difficult or impossible to replicate across the landscape. With the temporal replication of year at supplement stations and the spatial scale of the assessment (i.e., 1-km distance between the supplement stations), the lack of spatial independence was mitigated.

Grazing behavior
Averaged across all years of the study, individual cattle visited supplement stations an average of 52% ± 6% of days within the study period (14 −15 d; Fig. 3 ). Some individual cows (3 out of the 33 cattle tracked) visited stations frequently or during more than 90% of the study days (i.e., > 25 d out of the 28-d study period. Others (3 of the 33 cattle tracked) visited supplement infrequently, or during < 20% of the study days (i.e., fewer than 6 d, see Fig. 3 ). A majority of cattle (18 of the 33 tracked) visited stations between 36% and 71% of days the supplement was available (i.e., 10 −20 d, see Fig. 3 ).
As the supplement moved farther from water each week, no differences were detected for average number of days GPS-tracked cows visited supplement stations ( Table 1 ). However, the amount of time cattle remained near supplement during each visit was greatest in wk 4 when supplement was placed 4 km away from water and least in either wk 1 or 2 when supplement was only 1 and 2 km from water, respectively. Cattle tended ( P = 0.08) to be farther away from water during wk 4 compared with other weeks of the study (see Table 1 ). No differences were detected among weeks for daily distance traveled, distance from supplement site, or minimum convex polygon area cattle covered (see Table 1 ). Collared cattle were more likely to be found near the supplement transect line compared with the randomly generated transect lines. The amount of time GPS-tracked cattle spent within 100 m of the supplement transect line was > 3 × greater ( P = 0.07) than randomly generated transect lines without supplement ( Fig. 4 ). The average distance of cattle from the center of the supplement transect line was nearly 500 m closer ( P = 0.05) compared with random lines without supplement (see Fig. 4 ).
During each week of the study, the highest mean density of GPS-tracked cattle location fixes occurred at the current supplement station ( Fig. 5 ). The highest point densities typically extended between 100 and 200 m away from supplement station locations, with the highest point densities observed within the supplement station formation in each week of the study (see Fig. 5 ). In wk 1 and 2, regions in the northwest portion of the pasture generally had higher densities of GPS fixes. Cattle were attracted to this location because of the presence of an old watering site cattle were habituated to use from previous years grazing in the pasture. This watering location was nonfunctional and had no water during the study.
Cattle grazing reduced cheatgrass standing crop at all supplement stations along the supplement transect line ( Fig. 6 ). Before grazing initiation each year (i.e., wk 0), standing crop at supplement station #1 was 59% and 31% greater ( P < 0.03) than stations #2 and #3, respectively. However, by wk 4 of the study, after supplement had been cycled through all stations, standing crop was similar ( P > 0.21) among each of the supplement station locations. Averaged across years, the mean final standing crop at all stations was between 175 and 276 ± 132 kg ha −1 . However, standing crop was variable by year with the lowest final standing crop at a supplement station being only 36 kg ha −1 (supplement station #1 in 2014) and the greatest being 606 kg ha −1 (station #3 in 2016).

Discussion
Liquid protein supplements, moved 1 km farther from water each week for 4 wk, effectively attracted cattle grazing on the study pasture during the October −November grazing period and reduced standing cheatgrass fuels. Daily visits to supplement were highly variable (11 −96% of study days) among individual GPStracked cattle. However, moving supplement stations farther from water did not influence the mean percent of days cattle visited supplement stations, and the amount of time cattle spent at stations tended to increase at greater distances from water in wk 3 and 4 of the study. This suggests that even though attraction to the supplement stations varied by individual, moving supplement stations in consecutive weeks at distances of 1 km farther from water provided an efficient management practice to influence cattle grazing behavior in order to reduce dormant season cheatgrass biomass along a designated line.
Past research has indicated a high degree of variability among individual members of a cattle herd for use of self-fed supplements ( Bowman and Sowell 1997 ). For liquid protein supplements, up to 30% of cattle may not consume any supplement, with a 107% coefficient of variation for individual animal intake ( Bowman et al. 1995 ). Variation in cattle age, body condition, and learned behaviors can influence supplement intake and individual pasture resource utilization patterns ( Wyffels et al. 2020 ). The large herd sizes used in this study (655 −1 250 cows), which are common on extensive rangelands in the Intermountain West, can likely compensate for a small number of individuals within the herd that do not consistently visit supplement locations. Strategic supplementation can still modify the grazing behavior of hundreds of animals within a larger herd, facilitating more intense grazing near supplement stations, leading to greater removal of dormant cheatgrass biomass at desired locations.
Liquid protein supplements are commonly used on western ranges to improve cattle nutrition while they consume dormant, low-quality forage in the fall and winter, especially if product delivery trucks can easily access grazed areas. Little research has evaluated the efficacy of using liquid protein supplements to manipulate grazing animals to increase biomass removal from specific locations as a fuels management tool. The amount of time GPStracked cattle spent near liquid supplement stations (2.2 −4.3 h visit −1 ) and along the grazed supplement transect line ( ≈4 h d −1 , 17.7% ± 2.6% of the study period) was similar to other research that used different protein delivery methods (e.g., low-moisture block protein supplements) to attract cattle to specific locations. In northern Montana, GPS-tracked cattle spent 3.2 ± 0.3 h d −1 to 4.6 ± 0.4 h d −1 within 100 m of low-moisture block supplement locations . Stephenson et al. (2016) reported cattle spending greater amounts of time (i.e., 7.8 ± 1.1 h d −1 ) within 150 m of supplement using a combination of low-moisture block and low-stress herding every other day for 7-to 10-d periods during the dormant season in southern New Mexico. Our study demonstrated that liquid protein supplements may equal the effectiveness of other protein supplement delivery methods for attracting cattle to specific areas away from a water source during the dormant season on landscapes invaded with cheatgrass.
The increase in length of time cattle spent at the supplement stations placed 4 km from water highlights the effectiveness of the supplements at attracting cattle to substantial distances from water. Distances > 3.2 km from water are typically excluded from grazing capacity estimates because travel distance limits expected use, but some research has indicated that cattle will use areas farther than this in certain situations ( Holechek 1988 ;Millward et al. 2020 ). Protein supplements have been used to attract cattle to specific areas > 1.3 km from water on California foothill rangelands during the dry summer season ( George et al. 2008 ) and up to 3.1 km from water with a combination of herding and lowmoisture block during the dormant winter season in the southwestern United States ( Stephenson et al. 2017 ). Topography may influence the ability of protein supplements to lure cattle longer distances from water. Cattle will typically spend greater time at supplements and graze more forage away from water when protein supplements are placed in moderate terrain compared with rugged terrain with steeper slopes ( Bailey and Welling 1999 ). While more research is needed to fully understand the extent of how far protein supplements can successfully attract cattle away from water in large pastures, our research indicates this distance can be up to 4 km on relatively flat cheatgrass-invaded areas during the fall to early winter period in northern Nevada.
Strategic placement of protein supplements has utility for multiple vegetation management objectives, including the reduction of fine fuels carried over from one fire season to another and to improve cattle distribution and focus grazing at strategic locations when available forage is dormant and has low nutrient quality ( George et al. 2008 ;Stephenson et al. 2017 ). Cattle are attracted to supplement locations because protein supplement helps meet nutritional deficiencies during periods with minimal green or growing vegetation. Cooler temperatures during the fall and winter and the resulting lower water demand from cattle ( Rasby and Walz 2011 ) may enhance opportunities for cattle to travel greater distances from water, similar to what was observed in the current study. The ability to easily move supplement and attract cattle grazing pressure to new locations based on observed spatial or temporal variability in cheatgrass biomass without large, and often permanent, fencing infrastructure investments allows for more flexibility in management. Identifying variability in annual grass biomass production and areas with greater probability for wildfire with remote sensing technologies may provide opportunities to better select appropriate supplement placement locations and fuel breaks ( Jones et al. 2021 ).
Across years, cheatgrass utilization at the conclusion of the study averaged 48 −81% along the grazed supplement transect line, with no differences detected between the closest and farthest supplement stations from water. Frequency of visits by GPS-tracked cattle to all supplement stations was similar, but the length of visits and utilization of cheatgrass tended to be lower near supplement station #2 compared with other stations. This station had the lowest cheatgrass standing biomass compared with the other supplement stations before grazing initiation on the study pasture. The amount of forage near supplement stations can influence the efficacy of cattle remaining near supplement stations and the percent reduction of standing dormant vegetation nearby ( Stephenson et al. 2017 ). Cattle tend to preferentially select foraging areas with higher biomass production and forage quality relative to availability on rangelands ( Wyffels et al. 2020 ). We speculate the observed lower utilization at station #2 was caused by less available forage in this area compared with other regions of the pasture. Following supplement consumption, cattle likely moved away from the supplement station sooner to graze areas with greater forage availability.
This study supports other research that has highlighted the benefits of managing cattle grazing as an effective tool to reduce fine herbaceous fuels on arid and semiarid regions in the western United States ( Strand et al. 2014 ;Bruegger et al. 2016 ). Our results add to the research base by establishing that strategically placed supplements along a desired corridor can establish a grazed fuel break to reduce buildup of fine fuels during the dormant season for up to 4 km from a single watering point on a relatively flat, cheatgrass-invaded area during fall and early winter. The reduction in cheatgrass along the supplement transect line was within ranges reported by others using cattle grazing on cheatgrass-invaded rangelands. In a study where cattle grazed a 285-ha pasture in the fall to remove standing cheatgrass biomass, utilization rates of 58 −80% were effective at reducing the cheatgrass seedbank while increasing perennial grass standing crop ( Schmelzer et al. 2014 ). In a highly controlled, small-scale targeted spring grazing study, 80 −90% utilization of aboveground biomass reduced flame length and rate of fire spread during the following October ( Diamond et al. 2009 ). On sagebrush and native perennial grass plant communities, Davies et al. (2016b) reported that 40 −60% reductions in biomass through winter grazing reduced flame height, rate of spread, and area burned compared with an ungrazed control. Utilization of cheatgrass and the reduction of standing cheatgrass biomass at levels similar to those in our study should decrease ignition potential and fire intensity and behavior by reducing annual carryover of cheatgrass fuels.

Management Implications
Under a scenario of near monocultures of cheatgrass, fall cattle grazing is a logistically viable tool to reduce the amount of carryover fine fuels at large pasture scales, and strategic placement of supplement can direct this grazing to effectively create a linear fuel break. Cheatgrass can provide an important forage resource for cattle in much of the Great Basin and Intermountain West during the dormant season. With proper protein supplementation, cattle can maintain or increase body condition and weight while providing targeted grazing on strategic areas of dormant cheatgrass-dominated pastures during the fall and early winter season ( Schmelzer et al. 2014 ). This study highlights the efficacy of strategically moving protein supplement stations to create a grazed fuel break up to 4-km long from a single water source. Managing supplement station placements and cattle grazing distribution near or bordering areas with high ecological value or social importance provides options for land managers to reduce fine fuels at targeted, manager-defined locations. Strategically placed supplements can reduce the cost of developing fence infrastructure, decrease fence and wildlife conflicts, and provide greater flexibility to change management locations depending on fluctuations in precipitation, cheatgrass biomass availability, and management goals. Flexible grazing management options will facilitate the use of targeted grazing fuel reduction projects at strategic times (e.g., fall or winter) on rangelands of the Intermountain West ( Perryman et al. 2018 ) and provide more opportunities to better match livestock production and vegetation management objectives in a "win-win" situation within annual grass −invaded systems.