• Technical Note: Early harvest of squirreltail seed

      Doescher, P. S. (Society for Range Management, 2001-03-01)
      Squirreltail (Sitanion hystrix (Nutt. J. G. Smith), a native, cool-season perennial bunchgrass of the Intermountain West has been shown to reinvade degraded rangelands invaded by exotic annual weeds. However, one limitation to mechanical seed collection of this species is the disarticulating nature of the rachis at seed maturity. The purpose of this research was to determine if early harvest of the inflorescence before disarticulation would result in viable seed. After anthesis, seeds were collected weekly in 1995 and about every 10 days in 1996 at a research site near Prineville, Oregon. Seeds were germinated for 21 days at a constant temperature of 20 degrees C. Germinable seeds were present at all collection dates from late anthesis to seed shatter in 1995, and all but early anthesis in 1996. Total germination, rate of germination and seed weight increased as seeds were collected later in the summer. Collection of squirreltail seed when a majority of seed awns have moved from a reddish to a divergent, straw colored appearance resulted in germination properties similar to fully mature seed. This occured about 1 week prior to the onset of seed head disarticulation.
    • Technical Note: Physical and chemical comparisons between microphytic and non-microphytic soil seedbeds

      Blank, R. R.; Allen, F. L.; Young, J. A. (Society for Range Management, 2001-03-01)
      In arid and semi-arid climates, the physical and chemical nature of the soil seedbed greatly effects success or failure of plant recruitment. We hypothesized that the presence or absence of microphytic soil crusts may influence the character of soil seedbeds. To test this hypothesis, we compared chemical and physical attributes of the soil seedbed (0-6 cm) between adjacent areas of well-expressed microphytic soil crusts and non-microphytic soil surfaces for 2 sites on granitic alluvial fans in north-western Nevada. As compared with non-microphytic areas, microphytic soil seedbeds were finer-textured and contained more DTPA-extractable Mn, Cu, and Zn. Further research should examine in greater detail the role of microphytic soil surfaces in eolian dust entrapment, its relationship to nutrient deposition, and the interaction with seed recruitment.
    • Technical Note: A simple method for preparing reference slides of seed

      Dacar, M. A.; Giannoni, S. M. (Society for Range Management, 2001-03-01)
      Microhistological analysis has become the most commonly used and successful method for determining micromammal diets. However, this technique has a number of limitations, particularly when used on fecal samples where identification of some items is difficult. This method underestimates those nearly unrecognizable plant parts in the diet, such as seed, and overestimates easily identifiable parts, such as leaf epidermis. In this note we describe a simple technique that uses a macerating solution of 17.5% NaHCO3 for preparing reference slides of seeds. Advantages of the proposed method are discussed and compared with Jeffrey's technique.
    • Cattle treading effects on sediment loss and water infiltration

      Russell, J. R.; Betteridge, K.; Costall, D. A.; Mackay, A. D. (Society for Range Management, 2001-03-01)
      To quantify effects of forage canopy properties, soil surface relief, and hill slope on the hydrologic properties of the soil in a New Zealand hill land pasture, duplicate plots (15 x 3 m) with 3 canopy heights (6, 20, and 47 mm) were trodden with 0, 4, or 8 mature cows for 40 min on a moderate (15-18 degrees) slope (Trial 1), and similar plots with the low and high canopy heights were trodden by 8 mature cows for 40 min on gentle (8-14 degrees), moderate (15-18 degrees), or moderately steep (20-25 degrees) slopes (Trial 2). Pre- and post-treading measurements included canopy heights; bare soil proportions; soil moisture contents; hoof prints and skids; roughness coefficients, surface water detention storage volumes, and soil clump volumes with a 38-pin contometer; and water infiltration and sediment loss by rainfall simulation. In Trial 1, increasing canopy height resulted in lower (P < 0.01) proportions of bare ground, roughness coefficients, soil clump volumes, and sediment losses. Increasing treading damage resulted in higher (P < 0.05) post-treading roughness coefficients, proportions of bare ground, hoof print and skid densities, surface water detention volumes, and soil clump volumes. In Trial 2, soil hydrologic properties did not differ between canopy height or hill slope treatments. In the two trials, water infiltration rate was significantly related to the roughness coefficient (r2 = 0.31) and the number of hoof prints (r2 = 0.26). Results imply that a 20-mm canopy height of the forage species common on a New Zealand hill land pasture is adequate to minimize the effects of a short-term treading event on soil water infiltration rate and sediment loss.
    • Relationship between plant species diversity and grassland condition

      Bai, Y.; Abouguendia, Z.; Redmann, R. E. (Society for Range Management, 2001-03-01)
      Although the maintenance of biodiversity has become one of the goals in ecosystem management, the relationships of diversity to ecosystem characteristics such as level of herbivory, productivity, and vegetation structure are still poorly understood. We examined these relationships in 8 native grassland sites differing in grazing histories and range condition in the Mixed Grassland (6), Moist Mixed Grassland (1) and Aspen Parkland (1) ecoregions of southern Saskatchewan. Range condition, assessed using standard methods, ranged from fair to excellent. The Shannon's diversity index followed a curvi-linear relationship with range condition, increasing from fair to good, but decreasing from good to excellent condition, within a range between 0.66 and 2.58. Species evenness was affected by range condition in a similar manner ranging from 0.44 to 0.86. Species richness varied among sites and plots between 4 and 28 plants 0.25 m(-2), but changed little with range condition. Most structural parameters, such as the cover, height, or thickness of standing plants (live or dead) and litter, increased with range condition especially from good to excellent. The Shannon's diversity index was positively correlated with forb biomass, but not with biomass of any other group or their combination. Grazing regimes that maintain good range condition also maintain species and structural diversity of grasslands.
    • Observation: Long-term increases in mesquite canopy cover in a north Texas savanna

      Ansley, R. J.; Wu, X. B.; Kramp, B. A. (Society for Range Management, 2001-03-01)
      It is necessary to quantify rates of woody plant encroachment on southwestern USA rangelands to determine the economic feasibility of treatments designed to manage these plants. This study observed changes in honey mesquite (Prosopis glandulosa Torr.) canopy cover over a 20-year period (1976-1995) in 2 treatments: an untreated area that initially had a moderately dense mesquite stand (14.6% cover), and an area cleared of mesquite with root-plowing in 1974. Canopy cover of mesquite was estimated from scanned color-infrared aerial photograph images by manually delineating mesquite canopies with a computer using ArcView software. During the 20 years, mesquite cover in the untreated area increased (P less than or equal to 0.05) from 14.6 to 58.7%, averaging 2.2 percentage units per year. Cover in the root-plow treatment also significantly increased during the same period from 0 to 21.9% (1.1 percentage units per year), but the rate of increase was significantly lower than in the untreated area because mesquite growth was from new seedlings instead of established plants and/or new seedlings as occurred in the untreated area. Rate of increase was significantly lower from 1976 to 1990 (1.6 and 0.2 percentage units per year) than from 1990 to 1995 (4.1 and 3.7 percentage units per year) in the untreated and root-plow treatment, respectively. These differences were attributed to precipitation which was near normal from 1976 to 1990 but 25% above normal from 1991 to 1995.
    • Demographic characteristics of 3 Artemisia tridentata Nutt. subspecies

      Perryman, B. L.; Maier, A. M.; Hild, A. L.; Olson, R. A. (Society for Range Management, 2001-03-01)
      Previous research suggested that woody plant recruitment may occur in pulses in semi-arid areas. The overall objective of this study was to determine if this pulse phenomena was recorded in the demographic structures of big sagebrush (Artemisia tridentata Nutt.) stands in Wyoming. In 1997, approximately 75 stem cross sections were collected from 9 stands of each of 3 subspecies of big sagebrush in Wyoming along elevation and climatic gradients. Annual growth-rings were used to identify year of establishment and demographic characteristics were analyzed from age-class frequencies. Mean stand ages of the 3 subspecies were different (P = 0.002), and analyses revealed that Wyoming (A. tridentata ssp. wyomingensis) and mountain big sagebrush (A. tridentata ssp. vaseyana) stand ages (32 +/- 9 and 26 +/- 9 years, respectively) were significantly older than basin big sagebrush (A. tridentata ssp. tridentata) (17 +/- 3) stands (P < 0.05). Mean recruitment intervals (years) were shorter for basin (1.6) than for Wyoming (2.3) and mountain (2.2) sagebrush (P = < 0.01). The number of cohorts did not differ among the subspecies (P = 0.11), but the percent of years with recruitment was significantly higher for basin (59%) compared to Wyoming (37%) and mountain (39%) subspecies (P 0.0001). Age-class frequency distributions of each stand and regional stand combination were assessed for dispersion across each associated period of record. Chi-square goodness-of-fit tests were performed for the negative binomial distribution. All stands (with one exception) and all 3 regional stand combinations fit the negative binomial distribution. Age-class frequency patterns indicate that recruitment is clustered or aggregated across each period of record. Recruitment in big sagebrush stands occurs in pulses throughout Wyoming.
    • Supplemental polyethylene glycol influences preferences of goats browsing blackbrush

      Titus, C. H.; Provenza, F. D.; Perevolotsky, A.; Silanikove, N.; Rogosic, J. (Society for Range Management, 2001-03-01)
      Supplemental polyethylene glycol (PEG) increases intake of foods high in tannins, but it is not known if PEG affects preference when herbivores forage on a variety of foods that differ in concentrations of macronutrients and tannins. We investigated how macronutrients, tannins, and PEG affected preferences of goats (Caprus hircus) for current season's and older growth twigs from the shrub blackbrush (Coleogyne ramosissima Torr.). In blackbrush, current season's twigs are higher than older twigs in macronutrients, but goats prefer older twigs because high levels of tannins in current season's twigs decrease preference. We conducted a pen trial and a paddock trial. During the 7-day pen trial, goats were offered current season's twigs and older twigs throughout the day. Eight goats were supplemented with 20 g PEG mixed with 100 g ground alfalfa (Medicago sativa L.) pellets, and 8 goats were supplemented with 100 g ground alfalfa pellets. Goats supplemented with PEG ate more current season's twigs than goats that did not receive PEG (P = 0.04). During the 17-day paddock trial, 10 goats were supplemented with 50 g PEG mixed with ground alfalfa/barley (Hordeum vulgare L.), and 10 goats were supplemented with ground alfalfa/barley. Goats supplemented with PEG preferred current season's to older twigs, whereas PEG-unsupplemented goats preferred older to current season's twigs (P = 0.0001). Goats had equal preference for juniper (Juniperus osteosperma Torr.) trees (P = 0.243). Collectively, our findings show that supplemental PEG can change food preferences.
    • Intermountain plant community classification using Landsat TM and SPOT HRV data

      Clark, P. E.; Seyfried, M. S.; Harris, B. (Society for Range Management, 2001-03-01)
      Rangeland plant communities of the Intermountain West differ in their ecology and management requirements. Successful management of extensive areas at plant community-level resolution first requires an efficient, cost-effective means of plant community classification and mapping. We evaluated the influence of image acquisition date and satellite imaging system on the accuracy of plant community maps created from multispectral satellite imagery of Reynolds Creek Experimental Watershed (RCEW) (234 km2) in southwestern Idaho. Maps delineating 6 native and 2 non-native Intermountain plant communities were created from Landsat 5 TM and SPOT 3 HRV data using a maximum likelihood classification procedure. Map accuracy was assessed using ground reference points. Maps created from satellite data acquired during dry-down (early August) had higher overall accuracy (average = 70.5%) than from data acquired during peak growth (early June) (average = 54.4%). Overall accuracy of maps generated by Landsat (average = 60.1%) and SPOT (average = 65.5%) were statistically similar. Given their broad spatial coverages (3,600 to 31,450 km2 scene(-1), respectively), moderate resolutions (20 to 30 m pixels, respectively), and potential to provide high classification accuracies, the SPOT 3 HRV and Landsat 5 TM satellite systems were well-suited for classifying plant communities in the Reynolds Creek Watershed and similar areas of the Intermountain West. Practical procedures for plant community classification and map accuracy assessment are presented for use by natural resource managers.
    • Revegetating spotted knapweed infested rangeland in a single entry

      Sheley, R. L.; Jacobs, J. S.; Lucas, D. E. (Society for Range Management, 2001-03-01)
      Introducing and establishing competitive plants is essential for the successful management of spotted knapweed (Centaurea maculosa Lam.) infested rangeland where a desirable understory is absent. Our objective was to determine a herbicide-mix that would maximize grass establishment in spotted knapweed-cheatgrass (Bromus tectorum L.) bluegrass (Poa pratensis L.) infested rangeland in a single fall application. On 2 sites in Montana, 8 herbicide treatments [none, glyphosate (N-[phosphonomethyl] glycine) at 0.5 kg a.i. ha(-1), picloram (4-amino-3,5,6-trichloropicolinic acid, potassium salt) at 0.14 kg a.i. ha(-1), picloram at 0.28 kg a.i.ha(-1), clopyralid (3,6-dichloro-2-pyrdinecarboxylic acid, monoethanolamine salt) at 0.21 kg a.i. ha(-1) plus 2,4-D (2,4-dichlorophenoxyacetic acid) at 1.12 kg a.i. ha(-1), picloram at 0.14 kg a.i. ha(-1) plus glyphosate at 0.5 kg a.i. ha(-1), picloram 0.28 kg a.i. ha(-1) plus glyphosate at 0.5 kg a.i. ha(-1), and clopyralid 0.2 kg a.i. ha(-1) plus 2,4-D at 1.12 kg a.i. ha(-1) plus glyphosate 0.5 kg a.i. ha(-1)] and 3 seeded grass species ['Luna' pubescent wheatgrass [Thinopyrum intermedium (Host) Barkworth D.R. Dewey], bluebunch wheatgrass [Pseudorogneria spicata (Pursh.) Love], and 'Bozoyski' Russian wildrye [Psathyrostachys junceae (Fisch.) Nevski] were applied in a split-plot design with 4 replications in the late-fall of 1994 and 1995. Spotted knapweed and grass density were measured in 1995, 1996, and 1997, and biomass was measured in 1997 at peak standing crop. Density data were analyzed as a split-split-plot in time, and biomass data were analyzed as a split-plot using analysis of variance. By the end of the study, picloram applied at either 0.14 or 0.28 kg a.i. ha(-1) in the late-fall consistently yielded lowest spotted knapweed density and biomass. Initially, glyphosate alone lowered spotted knapweed density and increased grass biomass compared to that of the control. However, glyphosate treated plots had more spotted knapweed and less seeded grass established by the end of the study. 'Luna' pubescent wheatgrass consistently yielded the high test density and biomass of the seeded grasses. We believe a single-entry revegetation program applying picloram in late-fall combined with a fall-dormant seeding will maximize seedling establishment in spotted knapweed infested rangeland.
    • Chain diker effects on seeded grass establishment following disk chaining

      Wiedemann, H. T.; Cross, B. T. (Society for Range Management, 2001-03-01)
      Placing basins in rangeland seedbeds to capture rainfall has enhanced seeded grass establishment, but it has not been practical on debris-littered land following rootplowing for brush control. A disk chain tested previously will traverse land littered with debris or shrubs while tilling the soil for seedbed preparation. A chain diker is an implement designed to form 10-cm deep basins with a volume of 3.9 +/- 0.9 liters on tilled land. Combining the 2 implements provided tillage, land smoothing and basin formation in the soil with a single pass on debris littered land. Our 4-year study compared established grass densities on disk-chained seedbeds with and without chain diking at 5 locations in west Texas where land was rootplowed for honey mesquite (Prosopis glandulosa Torr. var. glandulosa) control. At 4 locations seedbeds were aerially seeded with 2 kg ha(-1) pure live seed (PLS) of WW Spar bluestem (Bothriochloa ischaemum (L.) Keng). At 1 location a mixture of 6 grasses, including Spar bluestem, was aerially seeded at 2.8 kg ha(-1) PLS. Evaluations were based on established grass densities at the end of the first growing season. At 3 locations, a seeding-before-diking treatment was included to simulate seed dropping from a seedbox mounted over the disk-chain's roller. In the Rolling Plains region of northwest Texas (4 sites), when growing-season rainfall was less than 400 mm, chain-diked seedbeds increased grass densities 2.6 times (P = 0.05) compared to non-diked seedbeds. When growing-season rainfall was greater than 500 mm, grass densities on diked seedbeds were similar to non-diked seedbeds. In the Trans-Pecos region of far-west Texas (1 site), where average growing-season rainfall is 337 mm, diked seedbeds produced 33% greater grass densities (P = 0.05) than non-diked seedbeds with rainfall of 552 mm. Densities of kleingrass (Panicum coloratum L.) and Spar bluestem were significantly greater in the diked seedbeds compared to non-diked seedbeds, and the other 4 grasses were not significantly influenced by seedbed methods. Grass densities were similar whether seeding after diking or before diking. Since long-term, growing-season rainfall has averaged below 500 mm between 65% to 87% of the time at these locations, our data suggests that it would be advantageous to include chain diking in combination with disk chaining for grass establishment on rootplowed land littered with brush debris.
    • Sequential sampling protocol for monitoring pasture utilization using stubble height criteria

      Turner, D. L.; Clary, W. P. (Society for Range Management, 2001-03-01)
      Stubble height, a measure of the amount of vegetation remaining after grazing, is thought to be a useful variable in the management of riparian areas. A number of riparian and grazing processes appear to be directly or indirectly affected by the residual stubble height. Thus, average stubble height is often used to evaluate the livestock impact a pasture has received, particularly in riparian zones. Stubble height sampling methodology has received little previous attention. A sequential sampling procedure for stubble height was investigated. The procedure provides statistically defensible answers in the shortest possible amount of time. The procedure does not require a rigid sample size and involves simple yes/no answers at each observation. A small initial sample of readings is selected and evaluated. If there is sufficient information to make a clear decision, then grazing is either continued or stopped. If the initial evidence does not clearly support either decision, then sampling proceeds. This may continue for several iterations before a decision is reached. Statistically supportable decisions can typically be made within a short time frame using this method. This method may also be applied to evaluate trampling and other yes/no responses.
    • Cover for wildlife after summer grazing on Sandhills rangeland

      Reece, P. E.; Volesky, J. D.; Schacht, W. H. (Society for Range Management, 2001-03-01)
      Livestock production and wildlife habitat objectives become antagonistic on grasslands when the architecture of standing herbage needed for key wildlife species limits the amount of forage that can be used by livestock. However, quantitative information needed to achieve cover objectives for wildlife is not available for summer-grazed grasslands. Three replicates of 7 grazing treatments were applied to the same 1.0-ha pastures for 3 years. Treatments included ungrazed control, and grazing at 16, 32, or 48 animal unit days (AUD) ha(-1) for 5 to 7 days during mid-June or mid-July. Cover was estimated after killing frost in September by measuring the average height below which complete visual obstruction occurred. Cumulative grazing pressure (AUD Mg(-1)) was used to describe grazing effects because of measurable differences in herbage among pastures and dates. Grazing in June reduced the average height of autumn cover at a constant rate from 11.0 to 7.0 cm (R2 = 0.34) as cumulative grazing pressure increased from 16 to 90 AUD Mg(-1). In contrast, declines in cover after grazing in July were about 2.6 times greater for cumulative grazing pressures up to 40 AUD Mg(-1) (R2 = 0.62), indicating a measurable decline in plant growth and an increasing dependence of autumn cover on the remaining herbage when grazing ended. Relatively low predictability of autumn cover after June compared to July grazing was offset by more plant growth during the balance of the growing season. Frequency of low-cover patches (less than or equal to 5.0 cm) within pastures was highly correlated (R2 = 0.94) with mean estimates of autumn cover. Consequently, the quality of cover near potential nesting sites also declined as the average height of cover declined, regardless of grazing date. The interdependence of low-cover patches and mean visual obstruction indicates that either variable could be the primary criterion for nest site selection up to 12 cm in visual obstruction.
    • Paddock shape effects on grazing behavior and efficiency in sheep

      Sevi, A.; Muscio, A.; Dantone, D.; Iascone, V.; D'Emilio, F. (Society for Range Management, 2001-03-01)
      Two grazing trials were conducted during early winter (December 1996-January 1997) and spring (April-May 1997) to evaluate the effect of shape of pasture on forage use and behavior of grazing sheep (Ovis aries). Two treatments were tested, square and rectangular paddock, with 2 replicates for each treatment of 9 ewes each. Groups were homogeneous for age and weight. Paddock size furnished 10 m2 per sheep per day. Each paddock was divided into 8 equal plots to determine herbage intake and grazing efficiency along the boundary and in the middle of paddocks. The shape of paddock affected sheep grazing efficiency and herbage intake both in the winter and in the spring. Because of a greater amount of herbage destroyed within boundary plots, the ewes in rectangular paddocks grazed less time, had lower herbage intake and used forage less efficiently than ewes in square paddocks. These results suggest that the shape of pasture can affect the behavior and herbage intake of sheep grazing in small paddocks and indicate that square paddocks should be used for research studies on sheep grazing behavior.
    • Range research: The second generation

      Young, J. A.; Clements, C. D. (Society for Range Management, 2001-03-01)
      The decade of the 1920s was somewhat of a paradox for range science. A. W. Sampson published 3 books that were widely used as text for higher education classes in range management. The United States Department of Agriculture, Forest Service expanded their mandate to manage grazing on National Forest and began to apply the principles of plant ecology and physiology that were being enumerated by range scientists. At the same time millions of acres of public domain outside the National Forest remained as free range and continued to decline in productivity. Progress was made in applying animal behavior technology to improve the uniformity of range forage utilization. This was especially apparent in regard to sheep and goats which were herded on rangelands. The management tools utilized were herding techniques, salt distribution and water developments. Restoration of range productivity and the place of wildfires in range ecosystems remained very controversial subjects.
    • Economics and demographics constrain investment in Utah private grazing lands

      Peterson, R.; Coppock, D. L. (Society for Range Management, 2001-03-01)
      In Utah during the early 1990s it was theorized that substantive change was under way in the management of private grazing land. Change was thought to be spearheaded by grazing permittees who feared losing access to public forage and thus wanted to increase carrying capacity of private grazing land as a hedging tactic. We synthesized results from socioeconomic surveys conducted among a target population of 5,067 grazing livestock producers during 1993, 1996, and 1997. This population was evenly divided between permittees and operators wholly dependent on private grazing (e.g., private operators). Our primary objectives were to: (1) test the hypothesis that a sustained upswing in management change was occurring; (2) identify factors associated with operations that "actively" invested in their properties versus those that were "passive"; and (3) identify producer priorities for applied research. Mail and phone surveys were used. Data analysis included descriptive statistics and logistic regression. Compared to private operators, permittees controlled far more private land and livestock and were more profit-oriented and dependent on live-stock-derived income. Managers of both groups were aged—37% of the population was >65 years old. Eighty percent of 393 managers surveyed in 1996-7 classified their operations as passive and ranked factors related to aging and economics as main reasons for passivity. Logistic regression and ranking exercises revealed that the active minority was most associated with higher gross annual incomes, more stewardship values, greater willingness to incur debt, and being a permittee. Permittees were more inclined to be active managers because of a greater entrepreneurial orientation compared to private operators, who tended to be hobby ranchers. Our work supported an alternative hypothesis that passivity in land management has been maintained in Utah during the 1990s, largely because incentives were lacking for most of the population to do otherwise. A wealthier minority, however, could still make large investments in their operations because of a superior risk tolerance. We concluded that demographic and economic factors exert the most control over producer behavior today, not access to information or new technology. One consequence is that demand for information and technology can be episodic due to coincident economic, demographic, and policy factors, which also implies that applied research, extension, and policy formulation need to be more opportunistic in response to change. Producers felt that forage improvements, policy, and economics were top research priorities. A looming crest of retirements among traditional landowners bodes for substantive and rapid change in the use of Utah private grazing land. Nearly one-third of those planning retirement hope to sell property to land developers.
    • Vegetation and water yield dynamics in an Edwards Plateau watershed

      Wu, X. B.; Redeker, E. J.; Thurow, T. L. (Society for Range Management, 2001-03-01)
      Woody cover, when expressed at the scale of the 207 km2 Cusenbary Draw basin, remained unchanged (approximately 23%) from 1955 to 1990. When expressed at the scale of range sites, woody cover declined on sites with relatively high production potential and increased on sites with relatively low production potential. Change in woody cover distribution at sub-range site scales, increased low and high woody covers and decreased intermediate woody cover, would be expected to lead to increased water yield at the basin scale because there was an apparent threshold woody cover (approximately 20%) above which simulated evapotranspiration (ET) changed little with increasing woody cover. This potential increase, however, was more than offset by the decreased water yield due to increased ET loss associated with compositional changes of woody vegetation from oak to juniper. A set of woody cover-ET regression curves was developed for different range sites based on simulation studies using the SPUR-91 hydrologic model. Based on these woody cover-ET regression curves and GIS analysis, no brush management would result in a 35% decrease in water yield, while a hypothetical brush management cost-share program would increase water yield by 43% over the 1990 level. Benefits in water yield and forage production from brush management differ in different range sites. A brush management cost-share program that preferentially allocated brush management to sites with deep soil and the highest forage production potential increased water yield by 50%, compared to a 100% increase if brush management were preferentially allocated on sites with shallow soil and highest water yield potential. These model results illustrate that the spatial scale of assessment and spatial distribution of brush management among range sites should be important concerns associated with developing and evaluating brush management policies.