Wild and Free-Roaming Horses and Burros: Current Knowledge and Recommended Research. (1980)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

CHAPTER 3 EFFECTS OF EQUIDS ON OTHER ECOSYSTEM COMPONENTS INFORMATION NEEDS The Public Rangelands Improvement Act of l978 clearly begins with an attitude of concern for the "unsatisfactory condition" of western rangelands, and for the need to improve that condition. In making lengthy provisions for research on, and management of, wild horses and burros, including authorization of the present study, it acknowledges that proper wild equid management must become part of the broader goal of improving and properly managing the rangelands. As mentioned earlier, the repeated reference to "excess" animals in PL 95-5l4 requires some definition of the term, and one among its several connotations is the impact of wild horses and burros on other ecosystem components. Hence some consideration must be given to the ways in which these animals create impacts upon their surroundings, to the criteria by which those impacts are evaluated, and to the standards for the term "excess animals" derived from severity of impact. Each land unit is capable of carrying some potential array of vegetative species if not excessively disturbed. As recognized and defined in PL 95-5l4, the "range condition" on a site at any point in time is "the present state of vegetation...in relation to the potential plant community for that site, and the relative degree to which the kinds, proportions, and amounts of vegetation in a plant community resemble that of the desired community for that site." To establish and maintain some level of range condition would appear to be a desirable management goal for each land unit. What that level should be is a policy decision, but presumably it would at the least ensure perpetuation of the basic vegetation, soil, and water resources. Since vegetation has evolved with herbivory, it is capable of sustaining some level of herbivorous removal while maintaining itself in some equilibrium state. This state may vary between years and be highly dynamic, but it will vary around some long-term, constant average over a period of time. Maintaining the vegetation at some decided-upon state implies that herbivorous removal must be maintained at a level that will perpetuate the equilibrium. One definition of "excess" horses or burros could be that number which, when exceeded, causes the range—including its vegetation, soil, water, and wildlife—to decline from the desired condition. l3l

l32 Hence, the information needed to determine when equids are in excess includes knowledge both of their forage preferences and consumption rates, and of the effects of that consumption on their ecosystems, including vegetation, soil, water, other wild animals, and livestock. Forage preferences and consumption rates were considered in Chapter 2. The ecosystem factors will be reviewed in this chapter. STATE OF KNOWLEDGE Information Sources The search for information on ecosystem effects was concentrated in four areas: l. Impacts on range-plant community created by wild, free-roaming horses and burros and associated livestock. 2. Indications of competitive pressures by wild equids on other wild animals. 3. Impacts on watersheds created by wild equids and domestic grazers. 4. Impacts on range nutrition and feeding ecology created by wild equids and associated domestic and native ungulates. Major attention was given to published, peer-reviewed literature, but considerable unpublished information (internal reports, consultant's reports) was obtained from files of the BLM, USFS, and NFS. Duplicate copies of these reports are on file at New Mexico State University and Utah State University. Range-Plant-Community Impacts In general, wild and free-roaming horse and burro populations in the western U.S. can be discussed in the context of the following vegetation types: pinyon-juniper, sagebrush-bunchgrass, salt-desert shrub, ponderosa pine, and hot-desert shrub, the last being primarily the domain of burros. While this geographic classification is general and considerable physical, climatic, and vegetational variation may exist within a particular type, it offers, at least, a general scheme for discussing and analyzing the problems on an ecological basis. Recent symposia and state-of-the art publications have dealt with the ecology and management of several of these areas (e.g., Clary l975, Currie l975, Gifford and Busby l975, Springfield l976, USU/CNR l979). Such publications provide a broad foundation of information upon which to build an analysis of horse and burro ecology. Considering the scope and alleged severity of grazing impacts by horses and burros on western rangelands, the problem has been studied surprisingly little. General statements by recognized range management experts warn of a severe and ever-growing problem. For example, Box (n.d.) observed that: "In many areas ranges are

l33 seriously overgrazed and the environment degraded." He further cited specific cases of burros in the Death Valley area and wild horse herds in the Rock Springs, Wyoming area posing threats to the range of native ungulates. Similarly, Cook (l975) suggested that under the Wild and Free-Roaming Horse and Burro Act of l97l, horses and burros living in areas where a year-round forage supply was unavailable would cause serious range deterioration that would lead to reduced grazing capacity for wild ruminants and domestic livestock. Subjective observations such as the foregoing are generally based on years of professional experience and may possibly depict the reality of the horse and burro problem. However, the regrettable fact remains that few pertinent empirical data are generally available for scrutiny and evaluation. Until such data can be procured, professional judgments, whether made by scientists or practicing range managers, are certain to remain controversial. Empirical studies on grazing impacts by either horses or burros are few in number. Most of the studies that do exist deal with burros. Hanley and Brady (l977) reported that burros in the lower Colorado River Valley of California and Arizona overgrazed ranges near the river, but that plant utilization, principally on browse species, decreased to light or moderate rates at distances greater than 2.5 km from water. They found that: "Overgrazing decreased the canopy cover of Ambrosia dumosa from about 2.26 to 0.04 percent and decreased total canopy cover for all species from 8.64 to 2.80 percent." They further observed that no particular plants acted as increasers or invaders under heavy burro use. The same general trend of heavy impact adjacent to water sources is evident in a report by Fisher and others (l973), who worked in the Panamint Range. However, at an exclosure of 2 km from the water source, measurement of species density and volume showed higher biomass outside than inside the exclosure. Annual densities and species diversity were higher outside the exclosure, but a bunch grass—Stipa sp.—was more abundant inside the exclosure. Carothers and coworkers (l976), who performed studies in the Grand Canyon area, reported that their results "demonstrate conclusively that the feral ass has a negative effect on the natural ecosystem of the lower reaches of the Grand Canyon." They showed that the total vegetative cover of the Mojave Desert scrub type was reduced from 80 percent on a control plot to 20 percent on an impact plot, and that the total number of vascular species present was reduced from 28 to l9. A later study by the same research group (Phillips and others l977) was much expanded to include the three major vegetation types in and along the Grand Canyon where burros were considered a problem. These included the pinyon-juniper type, the blackbrush (Coleogyne ramosissima)-grass type and the Mojave Desert scrub type of the inner gorge of the canyon. These authors concluded that burros had had great impacts on all three community types. In the pinyon-juniper and blackbrush-grass types, the major trend appeared to be replacement of palatable grasses such as galleta (Hilaria jamesii) and squirreltail (Sitanion hystrix) by unpalatable shrubs such as broom snakeweed (Xanthocephalum sarothrae) and brittle bush (Encelia farinosa). The inner gorge sites sustained the greatest amount of disturbance,

l34 attributed to a concentrated use by burros, the "inherently fragile nature of the arid desert flora," and a relatively low species diversity. Also noted was a "severe" impact on crustose mosses and lichens. The authors observed that the pattern of plant-community response to heavy grazing by burros was essentially similar to that widely documented under heavy grazing by livestock. Koehler (l974) studied burros in Bandelier National Monument and concluded that foraging by a population estimated to number from l07 to l20 individuals had led to degradation of some 4,000 ha of rangeland. He postulated the following order of disappearance of perennial grasses in pinyon-juniper communities overgrazed by burros: (l) Lycurus phleoides and Oryzopsis hymenoides, (2) Stipa neomexicana and Koeleria cristata, (3) Sporobolus airoides, (4) Festuca arizonica, (5) Bouteloua eriopoda and B^ curtipendula, (6) Stipa comata, (7) Sporobolus contractus, (8) Sitanion hystrix, (9) Hilaria jamesii, (l0) Bouteloua gracilis and B. hirsuta, (ll) Aristida s£., and (l2) Muhlenbergia torreyi. In addition to providing a basis for predicting community change, this sequence allows insight into the relative preferences by burros for the grass species listed. Generally, the most preferred species in a community are the first to be reduced by improper grazing practices (Stoddart and others l975) . A later study in the same area by Potter and Berger (l977), conducted after the burro population was reduced in l975, reported forage utilization levels on individual grass species. "Severe" levels of utilization were still noted on several grass species, averaging 6l and 50 percent for winter and summer periods, respectively. Some use of the unpalatable shrub called broom snakeweed was also noted and was interpreted as an indication of overutilization of the preferred grasses. These authors also observed heavy use of mountain mahogany (Cercocarpus montanus) by burros during periods of winter forage stress. Feral burros have also been studied in Death Valley National Monument by several investigators. Moehlman (l974) observed that animals in the Wildrose-Emigrant area were primarily browsers, seemingly due to a low availability of herbaceous species. Working in the Wildrose Canyon area at a later time, Douglas and Norment (l977) observed that browsing pressure by burros on the shrubs Acamptopappus shocklei and Ambrosia dumosa was probably severe enough to threaten removal of these species from the community. They further postulated that the level of use of Artemisia spinescens and Dalea fremontii, while lower than that of the preceding two species, was nonetheless severe enough to cause concern. In contrast to the studies just described, burros in two locations (Granite Wash and Temple Bay) in Lake Mead National Recreation Area apparently did not cause major impacts on vegetation (O'Farrell l978), even though estimated animal densities were considerably greater than those in the Wildrose Canyon area of Death Valley. This lack of impact was attributed to above-average precipitation in the area during recent years, leading to a greater-than-normal level of primary production and grazing capacity for burros, o'Farrell's (l978)

l35 preliminary report was based on a small number of samples, and judgment should await a final report. Although several management-level studies and inventories have led to subsequent population reduction of horse herds over the West, little controlled research has been done on specific grazing impacts by horses. The number of such reports in the literature is certainly less than that for burros. These data for horses seem especially meager in light of the extensive land area and the great number of animals concerned. Apparently much of the management that has been applied to horse ranges over the past 3 to 4 years has proceeded from experience and judgments exercised by the agency range managers. This problem will be discussed later under "Review of Range Survey Methodology" in Chapter 5. It is mentioned here, not for the purpose of criticizing previous horse management work, but rather to reemphasize the need for information on range trends, forage utilization, and associated environmental impacts—as they relate specifically to horses—to serve as a basis for future management actions. Research in this area is needed to assist the manager in distinguishing forage utilization and range impacts by horses from those of other sympatric ungulates, and to determine whether horse impacts differ from those of cows or sheep. There is also a need to monitor range impacts with greater sensitivity than is presently possible with existing techniques, which were developed largely for use with domestic livestock. Since scientific information on specific grazing impacts by wild horses and burros is scarce, additional perspective may be gained through exploration of the relatively extensive literature on grazing impacts by livestock. This literature has further importance: it forms the basis of existing management programs for wild horses and burros. The benchmark review monograph by Ellison (I960) throughly covered the pertinent literature up to l960. Some 260 references are cited in that paper, and discussion is organized on the basis of plant-community types, including those pertinent to wild horses and burros. One of Ellison's major conclusions is that successional trends in plant communities are roughly proportional to grazing intensity: they are pronounced under severe grazing pressure, and in some instances difficult to distinguish at light or moderate levels (Ellison l960:45). This conclusion has generally been sustained by the vast majority of long-term grazing experiments conducted throughout the western United States, and it is the fundamental assumption underlying contemporary grazing-management practices. The following perspective on grazing intensity presented by Ellison in his l960 paper remains germane to the present question and is worthy of consideration when reviewing results of grazing studies or when evaluating effects of animals in a management context: Degrees of grazing are not easily defined. The range that will support a given number of animal months during one period of years may be capable of supporting twice as many during a wetter period or perhaps only a third as many during

l36 a drier period. A depleted range may be heavily utilized by the same number of animals that would make only light use of the same range in good condition. The terms "lightly", "moderately" and "heavily" grazed are not only restricted in significance by being relative instead of absolute, but, unless production of forage and numbers of animals are specified, they are highly subjective. What one author means by "light" grazing may very well be "moderate" or even "heavy" grazing to someone else. (Page 5) As a means of providing more recent information on the relationship of grazing intensity to plant-community change, the data in Table 3.l were assembled. This table summarizes a small sample of long-term grazing experiments that have either been conducted in plant-community types where wild horses now occur, or that have focused on plant species locally important elsewhere on wild horse ranges. No comparable studies were found that dealt with the blackbrush-grass or Mojave desert scrub vegetation types now occupied by feral burros. One of the studies presented in the table (Hutchings and Stewart l953) was discussed in the Ellison (l960) treatise. The rest have been published since then. The table is incomplete in that it focuses mainly on the impacts of "heavy" grazing. This emphasis was prompted by the concern of the land-management agencies over the question of excess numbers of animals. Since major and enduring environmental change is a possible consequence of excess numbers of either wild or domestic ungulates, this emphasis appears necessary to provide perspective for designing research specifically related to horses and burros and to provide an interim management perspective pending the completion of that research. The specific data relating stocking rates to plant-utilization levels and subsequent changes in range-plant communities or particular plant species should prove useful in that regard. An important assumption underlying our analysis of such studies is that properly managed grazing, which takes into account the species and number of animals as well as their season and distribution of grazing, can in most cases be harmonious with most resource needs and values. The general conclusion from studies summarized in Table 3.l is the same as that drawn earlier by Ellison (l960): heavy grazing promotes accelerated plant-communitity change that has generally been interpreted as deteriorating range condition. Apparent deviations from, or variations on, this general statement can be seen in the studies of Hyder and others (l975), Skovlin and others (l976), and Hutchings and Stewart (l953) . These questions are addressed below. Range dominated by the shortgrass species blue grama and buffalograss appear relatively resistant to grazing-induced change. The study by Hyder and others (l975) summarized in Table 3.l led to the conclusion that there was no ecological basis for relating changes in range condition (i.e., plant-community composition) to severity of grazing as separate and distinct from changes promoted by variable weather conditions from year to year. This opinion appears to

l37 ary of Six Long-ten Crazing Studies Measuring Utilization Rates and ity Responses to Heavy Stocking Plant Association Location/ Animal Annual Season Species/Study Precip. of Duration tin. ) Use Stocking Rates tAUD/acrel- Light Moderate Heavy Key Forage Species Plant Utilisation Levels t%) . Light Moderate Heavy Hajor Cossmmity Responses to Heavy Grafting Study Colo./hei fen/ 7 years 2.i Bouteloua graciln i00 Ib/acre of unqrazed herbage remaining and considered "very heavy" Thinning of blue grant standt heavy grazing in April, May, fc June reduced cool-season perennials. Changes in commity botanical composition sore a function of veather than of Months of repeated heavy grazing. Hyder and other* t1975) Festuca ariconica Huhlenbergia son tana 38 74 36 70 Reduction of plant cover and increase in propo,ion of undesirable species such as Antennaria parvlfplia. Decrease in root depth, weight, and branch rootlets Delayed early ssason growth and retarded later growth. Erosion rates Increased 2- to 4-fold over light grazing. smith t1M7) 24 32 Slight increase in pro- duction of A. spicatum and no change in P. sanbergil, both occurring In open grassland situations. Carex sp. occurring under tree stands reduced pro- duction by 50%. Elk use was decreased. Skovlin and others t1976) Ponderosa Colorado/ pine-bunch- hei fers/ grass i6 years oregon/cattle/ pinsr-bunch- ii years grass Salt-desert sagetoruah- pinyon- juniper Utah/sheep/ i2 year* 6.7 Winter Ceratoides lanata 49 oryzopvis hymenoides 65 Total herbage production increased 54, 46, and 34% Cor light, and. and heavy grazing over course of i2-yr. study. Major con- tributors to this increase were C. lanata and A. confertlfolia. Hutching* and Stewa, t1953) Sagebrush- b unchgras Idaho/sheep i3 years/ Aqropyton spicatun 20 to 40 Range condition declined from good to poor in i3 yrs. Sagebrush increased 70%. Total grass produc- tion declined 73%. Blue- bunch wheatgrass decreased Laycock ti967) Created Utah/cattie/ whoatgrass i0 years artificial Hov.- Dec. 12.9 Spring i3.4 Variable, but relatively Bluebunch wheatgrass and heavy t6-i9%) use on Poa nevadensis increased sagebrush. 47%. Sagebrush production decreased 20%. Heavy use on shrubs while grasses were dormant led to nity progression. Agropyron crista^um 53 Reduced basal area of grass plants i8% in comparison to moderate grazing. Reduced grass yields from 433 lb/ acre for moderate gracing to 325 Ib/acre for heavy grazi ng. Accelerated invasion of sagebrush and rabbit brush. Prischknecht and Harris t1961J An animal unit day tAUD) is the grazing pressure exe,ed by a 454-kg tl.ooo-lb) unit slonth tAUM). ^Calculated on the basis of 5.0 sheep days - i.0 AUD. cow, an animal unit, or her equivalent during a 1-day period. Thi,y AUD • an animal

l38 conflict with earlier conclusions from a study on the same site (Kipple and Costello l960) that heavy grazing reduced vigor and yield of the important forage species and led to a deterioration of range condition. However, Hyder and others (l975) attributed the results of the earlier study to improper classification of site differences as differences in range condition. The Oregon study on ponderosa pine-bunchgrass range (Skovlin and others l976) did not show the drastic decline in range condition under heavy grazing that the Colorado study (Smith l967) showed in a similar plant association (Table 3.l). However, stocking rates in the Oregon study were considerably lower than those in the Colorado study, while annual precipitation in Oregon was roughly 30 percent greater than that in Colorado. Given comparable stocking rates, differences in precipitation notwithstanding, declines in range condition in the Oregon study would likely have been larger. Even with the prevailing conditions observed, the important species elk sedge (Carex geyeri) declined 50 percent in the forested habitat, and this was viewed as a serious consequence. This decrease was associated with, but not necessarily the direct cause of, a 20 percent decline in livestock grazing capacity over the ll-year study in the forested habitat. Total herbage production, and presumably grazing capacity, increased over the course of a l2-year winter grazing study at the Desert Experimental Range in southwestern Utah (Hutchings and Stewart l953), where salt-desert shrub, sagebrush-grass, and pinyon-juniper vegetation types were all represented (Table 3.l). Even though the increase was smaller under heavy grazing than under either moderate or light grazing, the general trend is not entirely consistent with that propounded by Ellison (l960). A reason for this response can be proposed. The ranges where the study was conducted were in a generally depleted state at the outset, due to excessive grazing and drought. Therefore, control of animal numbers, even though they were allowed to graze "heavily," apparently allowed some degree of range recovery. Furthermore, subsequent studies on physiological responses of desert plants to season and intensity of defoliation (Cook l97l) have shown that herbage removal during winter is considerably less detrimental than spring or summer defoliation. Thus, control over season of grazing—so that use was confined to winter—probably contributed to range recovery. The general pattern of plant-community change under heavy grazing in the Great Basin has been described by Young and others (l976) as: (a) an increase in native shrubs undesirable for browsing, (b) a reduction in perennial grasses and forbs, and (c) exploitation of these voids by alien annual weeds such as cheatgrass that are highly adapted to intensive grazing. From the standpoint of soil and water conservation, communities dominated by annual species are unpredictable from year to year as far as how much soil cover they will provide, and their soil-binding root masses are small. A later section in this report deals specifically with soil-plant-watershed relationships under grazing pressure. The most extreme change in this regard has been the dessication of upland meadows caused by accelerated erosion and stream-channel cutting (Cottam and Stewart

l39 l940). Young and others (l976) state that a cycle of geologic erosion and deposition would be required to restore these meadows to their past condition. From a standpoint of forage values, communities dominated by stands of relatively unpalatable shrubs (e.g., sagebrush) and by annual weeds (e.g., cheatgrass) offer relatively little to livestock (Cook and Harris l968). Although specific data are scarce, the same is probably true for horses. Table 2.22 strongly suggests that horses make little or no dietary use of shrubs. Exotic annuals may be highly palatable and nutritious during their growing season (Cook and Harris l968), but this is generally restricted to a brief period in the spring and early summer of favorable years. In years when precipitation is scarce, yields of forage from annual-dominated ranges may be negligible. As previously mentioned, responses of plant communities to grazing are greatly complicated by the effects of season and frequency of grazing. For example, Laycock (l967) showed that heavy spring grazing of sagebrush-bunchgrass range by sheep in Idaho led to a deterioration of range conditions, including replacement of bluebunch wheatgrass (Agropyron spicatum) by annual cheatgrass (Bromus tectorum) and sagebrush (Artemisia tripartita). In contrast, heavy grazing of similar range in autumn after the herbaceous understory plants had entered dormancy led to increased grazing pressue on sagebrush and an improvement in range conditon. Such differential responses of individual plant species to season of grazing form the basis for the specialized "rotational" grazing systems now in common practice over much of the western range (reviewed by Herbel l974). Controlled studies concerning the effects of plant defoliation on subsequent vigor and production (Cook l97l) showed that desert shrub and grass species are highly sensitive to season of herbage removal. Species included in the studies were black sagebrush (Artemisia nova), big sagebrush (A. tridentata), shadscale (Atriplex confertifolia), winterfat (Ceratoides lanata), Nuttall saltbrush (Atriplex nuttallii), Indian ricegrass (Oryzopsis hymenoides), and squirreltail grass (Sitanion hystrix). A 50-percent winter utilization of these species maintained vigor and yield, but the same level of use during late spring and early summer was detrimental, particularly for the shrub species. No more than a 25-percent use could be safely tolerated by most species in late spring. The implications of these studies for the management of wild horse ranges are clear. If maintenance of a stable range condition is a desired management objective, control of grazing season may be equally as important as control of animal numbers. This type of control may prove particularly difficult in some areas where options are limited to regulating movement patterns and distribution of animals over the range. Year-to-year variability in precipitation (both in amount and monthly distribution) can radically alter the composition of plant communities and influence the pathway of either secondary succession or retrogression. This has already been emphasized in discussing the findings of Hyder and others (l975). Precipitation seems to be an

l40 overriding factor in most range ecosystems, but especially in the more arid ones. On the basis of a l0-year study of semi-desert range in southern Arizona, Martin and Cable (l974) concluded that year-to-year variation in precipitation influenced short-term changes in vegetation more than any other single factor, including season and intensity of grazing. The periodic stresses imposed by droughts, coupled with stress imposed by heavy grazing, have often interacted to produce profound changes in range conditions. In fact, many of the grazing studies discussed above pointed out that much of the change measured over the duration of a particular study actually occurred during periodic drought periods, rather than as a gradual, continuous process. Range grazing capacity is also subject to the vagaries of annual precipitation. Close positive relationships have been demonstrated between annual precipitation and forage production (hence, grazing capacity) for many areas (Pechanec and Stewart l949, Hutchings and Stewart l953, Dahl l963, Currie and Peterson l966, Martin and Cable l974, Duncan and Woodmansee l975). Variations in precipitation greatly complicate the decision regarding the number of animals an area can sustain on a long-term basis. Even with domestic livestock that are subject to a relatively high degree of managerial control, yearly adjustments of animal numbers in accord with the prevailing forage conditions are not practical. Such control is presently out of the question for wild horses and burros. Stoddart and others (l975) state that 65 to 80 percent of the long-term, average forage production base is usually a realistic starting point for calculating grazing capacity. However, they emphasize that under no plan of constant stocking will it be possible to prevent excessive use in dry years, and underuse in good years. If there is to be a temporary imbalance it should favor the plants rather than the number of animals, because ranges are slow to recover from overuse. For example, in studies by Cook (l97l) cited above, shadscale, squirreltail grass, and black sagebrush had not fully recovered from the effects of heavy (75 percent) utilization after 7 years of rest. Year-to-year fluctuations in forage production are even more serious on ranges in poor condition, where much of the forage supply is based on annual species. For example, Murray (l97l) compared grazing capacities of native bunchgrass in southern Idaho to those of cheatgrass ranges in the same vicinity. During wet years, grazing capacities of the two ranges were similar, but in dry years, they were 60 percent less on the cheatgrass range. Thus, maintenance of plant communities with a major component of perennial vegetation appears desirable not only from the standpoint of long-term site stability, but also from the viewpoint of a dependable forage supply for grazing animals. Competitive Effects on Other Animals Interspecific Competition as a Population-Limiting Influence The question of competitive effects of equids on other animals raises once again the difficulties discussed in Chapter 2 in

l4l demonstrating the existence of competition. Interspecific competition occurs when (a) two different species use the same resources, (b) they reduce them to the point of short supply, and (c) the populations of one or both are constrained in the process. Two species may use exactly the same resource but not be in competition if their combined use does not reduce it to the point of population limitation. A population effect, the acid test of competition, is often difficult to demonstrate. If two species have been coexisting for some period of time and one is declining, it is risky to assume cause and effect. Many environmental variables change over time, and the decline of a species could just as well be prompted by one or more of them as by a suspected competitor. Competition is best demonstrated by an experiment in which one species is increased or reduced, and the response of the other observed. Properly, such treatments should be replicated. In the absence of population tests, at least a preliminary indication of competition can be obtained by measuring the resource in question and calculating the resource need of the two species suspected of competition. If the combined need is greater than the amount of resource available, it is at least presumptive evidence that competition may be occurring. Even where a population effect can be observed, the case for competition is strengthened by measuring the resource and demonstrating its depletion and shortage. It is not stretching the correspondence between concept and reality too far to suggest that interspecific competition can be created by administrative fiat. If the decision is made to limit the number of grazing animals on a given tract of land to a certain number of animal unit months (AUM: see Table 3.l for definition) and a portion of these is allocated to equids, then the numbers of livestock must be limited to the remainder. In a real sense, livestock populations are held at lower levels than they would be in the absence of equids, and the latter therefore exert competitive pressures on the domestics. Feral Asses The most commonly inferred competition of equids with wild species is that of burros with desert bighorn sheep, and it has been reported to involve three resources: water, vegetation, and space. The evidence for water resource competition is conflicting. Numerous authors (Dixon and Sumner l939, Sumner l952, Russo l956, Weaver l959, Thomas l979) have reported for years that in areas where burros are numerous and water scarce, the equids foul or completely use up the water so that sheep cannot drink it. But other investigators studying burro watering behavior have failed to find the fouling problem (Welles and Welles l96la, b; Moehlman l974; Golden and Ohmart l976). Observations on watering behavior of the two species vary, but there may be a tendency for bighorns to water by day while burros are more inclined to water at night (Welles and Welles l96la, b; McMichael l964; Farrell l973; Moehlman l974; Woodward l976). Welles and Welles

l42 (l96lb) reported that "thriving bighorn populations have been watering at the same springs with thriving burro populations since before l937." The question may revolve around the abundance of water: problems may arise where it is quite scarce, and its use is intensive. The problem of competition for forage has been given more attention, but also lacks conclusive resolution. It is quite clear that many forage species are shared by burros and desert bighorns (McMichael l964; St. John l965; Hansen and Martin l973; Seegmiller and Ohmart l975, l976, l980; Seegmiller l977; McQuivey l978; Walters and Hansen l978). Burros appear to have a competitive advantage, since they utilize a wider array of plant species and plant parts (McKnight l958) and can range farther from water sources (Seegmiller and Ohmart l980). However, as discussed above, mutual use of the same resource does not necessarily imply competition. The important point may be whether sheep are sufficiently wide- ranging to be unaffected by the burro vegetation damage close to water holes (which most authors agree takes place), or whether vegetation consumption and alteration extend far enough out from water holes to affect appreciable fractions of sheep home ranges. Some authors (e.g., McMichael l964) have suggested that the vulnerable point in the sheep life cycle may be the food needs of the lambs, especially at weaning time (Thomas l979:l83) during the dry season, when they are confined to a narrow radius around the water holes. The question of the space resource involves social encounters between the two species. McKnight (l958) suggests an implicit dominance of burros over sheep. Thomas (l979:l83) describes observations made during a joint BLM, UPS, and California Department of Agriculture study in which sheep invariably waited to drink at a spring until after burros had vacated the area. Such nonovert submissive responses have been reported for other large herbivores. McCullough and Schneegas (l966) speculate on the behavioral incompatibility between livestock and Sierra Nevada bighorns. Jeffery (l963) and Mackie (l970) reported that elk vacate areas occupied by cattle. Despite all of this indirect evidence, experimental observations on population effects are few and equivocal. Several cases have been reported of thriving sheep populations in areas without burros, and a lack of sheep in areas occupied by the equids. Thomas (l979) describes a number of such cases, including situations on two military bases in San Bernardino and Inyo Counties, California. The Avawatz Mountains on Fort Irwin have abundant perennial grasses, a thrifty sheep population, and no burros. The Argus Mountains on the Naval Weapons Center have a large burro population, show signs of heavy grazing and absence of perennial grasses, and have a declining herd of fewer than l5 sheep. McKnight (l958) reported that in the l950s the Panamint Mountains on the west side of Death Valley were heavily utilized by burros and were practically devoid of sheep. The ranges on the east side of the valley contained no burros and had sizeable sheep populations. Of course the question always arises as to whether these negative correlations involve cause and effect, or whether other variables are involved.

l43 Dixon and Sumner (l939) and Brandt (as cited in McKnight l958) reported a sequence of changes at Willow Spring in Death Valley that constituted an inadvertent experiment. In l935, the spring was being used by a local sheep population, and was clean; apparently there were no burros. In l938, the spring was trampled and muddied by burros, and no sheep could be found in the vicinity, according to Dixon and Sumner. By l957, burros had been trapped out of the area and it once again supported a thriving sheep population. Despite the absence of conclusive data, the problem has been alluded to so repeatedly by experienced observers that one must suspect its reality. Thomas (l979) marshalls an impressive array of unpublished and published information on the subject. Morgart (l978) reported year-round coexistence between burros and mule deer on the same range at Bandelier National Monument. In addition, deer numbers were greatly augmented in winter by migrants. Deer were almost entirely browsers while burros were mainly grazers, but in winter a third of the burros' diet was browse. Since the browse was overused, and therefore at or below the amounts needed by both species, there was some possibility of competition in Morgart's opinion. Several authors have investigated possible effects of burros on small mammal populations. Moehlman (l974) counted rodent burrows along transects placed to sample areas with different feral ass densities, and found no difference in burrow numbers in the different areas sampled. Other authors, however, have found rather different results. Carothers and coworkers (l976) and Czaplewski and others (l977) investigated the abundance of small mammals in three different habitat types in Grand Canyon, each with different burro densities, and each with areas subject to burro use and control areas protected from burro use. In the pinyon-juniper vegetation type on the Canyon rim, where burros are scarce, no differences were evident between rodent populations in "impact" and "control" areas. In the desert scrub vegetation of the Tonto Platform and Rampart Cave area, where burro use was intermediate, rodents were more numerous in the impact than in the control area, and were composed in large part of disturbance species in the family Cricetidae. In the riparian-plant community of the canyon's inner gorge, where burro use was heavy and vegetation damage extreme, rodent populations in the impact areas were far smaller than those in the control plots. Douglas and Norment (l977) conducted similar investigations in two areas of Death Valley, one heavily affected by burro use and the other free of use. Efforts were made to select plots with similar elevation, slope aspect, and plant communities. The density of small mammals in the burro-free area was more than 3.5 times as great as that in the burro-inhabited area. Species compositions also differed. The authors acknowledged, however, that since the two areas were not adjacent to each other, the possibility existed that variables other than burro populations were partially or wholly responsible for the difference. Guthrie (l978) compared rodent densities of areas with and without heavy burro use in pinyon-juniper woodlands of Bandelier National Monument. Rodents were less abundant in areas used by burros.

l44 Other authors have marshalled evidence to suggest reductions in bird populations (McKnight l958, Wauer l978, Thomas l979) and desert tortoise where burros range (NFS l979) . Feral Horses The attention given to, as well as the attempts made to evaluate, horse competition with other wildlife is much more limited than are burro investigations. Some studies (Hansen l976, Hubbard and Hansen l976, Hansen and Wolsen l977, Hansen and others l977, Thomas l979) have been conducted to assess the degree of dietary overlap between horses, cattle, and several species of wildlife (elk, mule deer, pronghorn antelope, and bighorn sheep). All overlap to varying degrees at one time or another, and in one area or another. Overlap tends to be greatest between horses, cattle, elk, and bighorn, all of which are primarily grazers in the more northerly portions of the western United States. In the Challis region of Idaho, bluebunch wheatgrass (Agropyron spicatum) is a preferred species for horses, elk, and bighorn (Peek as cited in Thomas l979:l93). This is a climax species that often recedes under any material grazing pressures, and its full utilization by cattle has been reported by Morgan (l97l) to pose genuine competitive pressure on bighorn populations in the nearby Salmon River wilderness area. Peek also reported that horses turn to sagebrush to some degree in winter in the Challis area, and at this point may compete with mule deer for limited winter browse. A similar situation in eastern Oregon was reported to us by Forrest Sneva (USFS, personal communication, l980) . In general, several species of sympatric large herbivores may partition resource use when the resource spectrum is broad and utilization is low. But if their populations and feeding pressures increase to the point where preferred food species are reduced, their food selection may converge and the risk of competition will increase (Wagner l978). Thomas (l979:l94) describes a situation in the White Mountains on the California-Nevada border. Up to 25 years ago, when wild horses were scarce in the region, bighorn sheep occupied all parts of White Mountain Peak, which rises to l4,000 feet, including the foothills on the east side of the mountain. In recent years a growing horse population occupies the eastern slopes of the mountain. Sheep are now found only on the west side and the crest. All of this is again circumstantial and not firm evidence of competition. But it would be a mistake to dismiss the possibility that competition does occur simply because the data are equivocal. The question remains open. Effects of Bquids on Range Hydrology Hydrologic impacts are generated by activities that, in one way or another, affect quantity, quality, or timing of water yields. Impacts

l45 may be measured either on or off site, at upland locations, or directly in a channel. The grazing animal influences watershed behavior through removal of protective cover and through trampling disturbance. Resultant impacts may include altered water-quality parameters (coliform counts, nutrients, water temperatures, sediment, dissolved oxygen, and total dissolved solids), reduced stream-channel stability, increased overland flow, reduced soil moisture, and increased erosion. Currently, no information is available in scientific journals that attempts to quantify the impacts of wild equids on quantity, quality, or timing of water yields. However, numerous anecdotal comments have appeared from time to time, and some limited in-house reports have been issued by various federal agencies that purport to identify the hydrologic impacts of wild equids. These reports suggest that wild equids may adversely impact range hydrology by: (a) compacting the soil surface, (b) developing trails in steep terrain that accelerate erosion, (c) overgrazing their habitat, (d) competing with livestock and big game for forage and water, and (e) polluting water holes (Dixon and Sumner l939; McKnight l959; Weaver l959; Welles and Welles l959, l960, l96lb; Buechner l960; Koehler l96l, l974; Fisher l975; Stoddart and others l975; Woodward and Ohmart l976; Carothers and coworkers l976; Norment and Douglas l977; Zarn and others l977a; O'Farrell l978; Hansen n.d.; Jones n.d.; NFS n.d.). These impacts are somewhat localized, highly dependent on local population and climatic trends, and often confounded by other effects (e.g., present livestock grazing, past grazing history, wildlife, and hikers). The existing knowledge about the impacts of wild equids on range hydrology is scanty at best. However, information is available on livestock grazing impacts, which should be similar to those wild equids. The following review deals with the influence of livestock grazing on vegetation, soil, infiltration, and water-quality parameters. Vegetation and Cover Factors Vegetational Factors, Impacts on vegetation caused by the grazing of wild equids and livestock on range-plant communities were reviewed in an earlier section of this chapter ("Range-Plant-Community Impacts"). As pointed out in that review, grazing animals can drastically influence plant communities, and it is the removal of vegetation that has the largest potential impact on watershed behavior. Overgrazing for extended periods may result in vegetational changes that will: (a) reduce protective cover, thus increasing the impact of raindrops; (b) decrease soil organic matter and soil aggregates; (c) increase surface vesicular crusts; (d) decrease infiltration rates; and/or (e) increase erosion (Blackburn l975; McGinty and others l978; Wood and others l978a, b; Wood l979; Knight and others l980). The end result of overgrazing may be to lower site potential. Excessive removal of vegetation for short periods may:

l46 (a) reduce the protective cover, thus increasing raindrop impact; (b) break down soil aggregates; (c) decrease infiltration rates; and/or (d) increase erosion. Usually these impacts can be reduced by adjusting stocking rates and/or applying proper grazing management (McGinty and others l978, Wood and others l978b, Wood l979). Grazing usually lowers the standing crop when compared at any one point of the grazing cycle (Lodge l954, Johnston l962, Rauzi l963, Rhoades and others l964, Rauzi and Hanson l966, Hazell l967, Hanson and others l970). Knight and others (l980) found midgrass standing crop to be similar on moderately, continuously grazed and short-term grazed pastures (Figure 3.l). Standing crop on a very heavily grazed pasture was significantly smaller than that on pastures grazed moderately and continuously or on a short-term basis. Some authors report a stimulating effect of grazing on vegetation production that is similar to that seen in an exclosure or lightly grazed area (Rauzi and Smith l973, McGinty and others l978). Reardon and Merrill (l976), working in Texas, reported that only heavy grazing showed a decrease in plant production when compared to a livestock exclosure and a four-pasture, deferred-rotation system. Mulch. Although mulch is usually reduced by grazing (Johnston l962, Rauzi and Hanson l966, Hanson and others l970, Wood l979), it was not reduced by moderate, continuous grazing below levels found under light, continuous grazing or in exclosures during two studies (Rauzi and Smith l973, McGinty and others l978). Reardon and Merrill (l976) suggested that exclosure vegetation on the Edwards Plateau, Texas needed the stimulation of grazing to keep producing mulch at levels comparable with that of grazed areas. In a South Dakota study, more mulch cover occurred on a moderately, continuously grazed pasture than on a heavily or a lightly, continuously grazed pasture (Sharp and others l964). Lusby (l970) reported that a heavily grazed pasture at Badger Wash, Colorado displayed a decline in mulch cover when compared to an ungrazed area. Bare Ground and Rock Cover. Lusby (l970) reported that bare soil and rock cover increased with heavy grazing in a paired watershed study (Badger Wash, Colorado) that compared heavy and no grazing. Heavy grazing and short-duration grazing increased the percentage of bare ground cover at Rolling Plains in Texas (Wood l979). Several studies have shown that 60 to 75 percent plant and mulch cover is needed to control surface runoff and erosion adequately (Meeuwig l970a, Orr l970, Packer l95l, Marston l952). Many areas where wild equids are found do not have the potential to produce that much plant and mulch cover. Soil Factors Physical properties of soil change more slowly than does vegetation, although the two may be related (Rauzi and others l968).

l47 T- o o o CM nr o o U) T o o o T" o s o < 5 e •H N O1 * B n * •a 8 to c E S b^ ^^ C I O O -< C 0) (d .P 4.) tB U rH & W (0 id *o •o » •H '6 z u f>

l48 The soil factors that may be altered by livestock grazing may also change infiltration and erosion rates. Bulk Density. The soil-compacting effects of grazing animals appear to be most pronounced in the first 2.5 to 5 cm of soil surface (Alderfer and Robinson l947). No significant increases in bulk density due to moderate, continuous grazing were reported by Meeuwig (l965), Rauzi and Hanson (l966), and Skovlin and others (l976). However, Klemmedson (l956) reported that grazing-induced changes in range condition also changed the bulk density of the soils. The bulk density tended to decrease as range condition improved, although no statistical tests were reported. Unrestricted grazing of stream bottoms in the Black Hills caused increases in bulk density in the top 5 cm of soil at three out of four sites (Orr l960). Similarly, unrestricted grazing of the Southern High Plains caused increased bulk density of the surface soil (Brown and Schuster l969). In northern Utah, Laycock and Conrad (l967) showed that seasonal changes in bulk density were greater than those that could be attributed to grazing. Rhoades and others (l964) and Linnartz and others (l966) reported increased bulk density due to grazing down to 40 cm below surface or deeper. These findings are contrary to other studies that found grazing impacts to be minimal—restricted to the upper 5 cm of the surface. It is possible that different soil types were compared between the grazed and protected areas. Rhoades also reports that numerous rodent and insect burrows in the protected area may have caused the lower bulk density. Organic Matter/Aggregate Stability. No differences in percentage of organic matter were found between soils of grazed and protected areas by Lodge (l954), Klemmedson (l956), Rhoades and others (l964), and Meeuwig (l965). Wood's (l979) study in Rolling Plains, Texas reported organic matter to be highly correlated with infiltration rates and sediment production, but found little difference due to grazing treatment. McGinty and others (l978) showed the percentage of organic matter in two grazed pastures to be lower than that of a nongrazed area. Wood (l979) reported that the percentage of water-stable aggregates was lower in a heavily grazed pasture than it was in one 20-year-old exclosure, but not significantly lower than that in another 20-year-old exclosure. He also found aggregate stability to be highly correlated with infiltration rates. Infiltration Studies relating livestock grazing practices directly to runoff are rare, in part because most hydrologic concern has been directed toward the recharge of soil moisture for growing forage in arid and semi-arid environments. Most studies have concentrated on infiltration as an indicator of range hydrology, and there exists a modest amount of literature on the subject. Unfortunately, this

l49 information presents several different descriptors, and infiltration is measured with a variety of instruments (Gifford and Hawkins l978). Ideally, infiltration should be measured using rainfall, either natural or simulated, and calculated as the difference between input rainfall and output runoff. As a refinement, initial surface retention and temporary storage may be quantified. Flooding infiltrometers have been used, but represent unnatural conditions: the impact of raindrops is eliminated and exceptionally high infiltration estimates result from the ponded head of water. Data dealing with the impacts of livestock grazing on infiltration rates have been summarized in Table 3.2. Several conclusions can be drawn from these studies. l. The results are often confounded by range-improvement activities, past grazing history, and/or climatic fluctuations. 2. Results may be very site specific. 3. There is usually little difference between light and moderate grazing. On some sites there may not be a difference between no grazing and light or moderate grazing. 4. Heavily grazed areas almost always exhibit a lower infiltration rate than areas grazed lightly, moderately, or not at all. 5. These studies were conducted by sundry methods, mostly on year- or season-long, continuously grazed pastures with varying stocking rates. Infiltration data listed in Table 3.2 that met the following conditions were analyzed statistically: (a) terminal rates, (b) measured with either natural rainfall or a sprinkling infiltrometer, and (c) grazing intensity identified. These data were also used in developing a deterministic model for predicting infiltration rates under various livestock grazing regimes (Gifford and Hawkins l978, Hawkins and Gifford l979). The authors cited were unable to differentiate between the influences of light grazing and moderate grazing, and considered them to be identical. Statistical analyses indicated that there was a difference between infiltration rates associated with moderate/light, heavy, and no grazing. These data were also subjected to regression analyses using the three categories of light/moderate, heavy, and no grazing (Figure 3.2). From these calculations the following conclusions were drawn: (a) "There is an influence of grazing on infiltration;" (b) "There is a distinct impact from heavy grazing which is statistically different from that of light/moderate." The model developed by Gifford and Hawkins (l979) used an average infiltration recovery time from grazing of 4 years. This figure was based on limited data presented by Dortignac and Love (l96l) and Thompson (l968) . They contend that the recovery period is certainly longer than l year, a factor that reduces the potential of improving watershed conditions through improved grazing management systems. The existence of this long recovery period is not supported by McGinty and others (l978) and Wood and others (l978b), who studied several different grazing systems in Texas. The results of these two

l50 i •• J e " C U i|£ S ?TS a * o O.^ ^. 1u "a ai oou at t3 Ov* *r -? "*^ 4fC ^ tu *— QJ 3 C n >l t O U CI O. O ftl T3 ,- - 1- 5 I.CO « *> tt ^ ,— O ^ ^ fl ^ ^ — 0 > X: CD » - 3.i-k. T3 k •a — at C T3 <r w*Xa «|9 9 5 •c i in 0. a .u M H £ v w ^ QJ a **.3 C*. »»- U«-14I k C^ *• ^^^ T3C*'*.»t"5.w»o r^m i o - ^ 1 n t• "° "= c .; .o •« S •ri T3 .0 * * T! .0 « 4 •D .n 1 " O* fai V O £ "° 7 55 ,' ? ?"" ""2 *«'-*«*^g'ww»«>'g«»'- tre-tment tre-tment stoi.ni of F composite ,,r,., S > j 01 u 3 s 3 o ** kl "^ X 14 c cvji1^ ro \o «o'^ t/i ^" D r^ B c i OO O O OO O O ^-- r- rn dd 9 1 . c - 1 CO £ to r- ^ 4i 01 , . .H rH > k 91 m « «o t4 <^ CNJ o» c« oo o a "^ O an 'S ij 3 3 3) 4J 14 A § ^-*u> 3«3 - OJ V *U c 5 0 0 0 "1 ? - •— irt tn eg m ^™ m n •H TJ - 2 o a ^ o o o *— ~ w* -« c u1 o ^ c c 3 *j in — r*~1 vo ^ *r in tn*o otm tntnu)vo -x ^ •H id N CNJ rj rJ t^" — " — ^-" c J ^ -J 00^ •t N w k O O O CM O O r- id a s |s t^ a X "• || i. II 9 o id 1 1 I &'O 14 M O i . •-. 5 j I E '" a. ci = s - C B 'n O" Ul a M 3 ii c CO U | "~ c t/t n O *J L. .r- ^ ^- u i- „ t t*- >.»• I ^ .u 0 1 tn 3~ •3 = I 11 1 n t 3 N Ml M — ,_, — t 1^ 2* 2 * i • B«s *s* .ii ** w — w» aj — - ,X U - i~ ) — «5-: i<s ' 13 wt. ^ 5— : • < — ' t *t §~ C • 41 iJ tVi •— 1>1 ii— i a tt> Z 3 .^ «j v C *" C 3J > ^ ,— .r 1 1 ' a "i « ,- c •»| o o ^"og '« 2^ O •" e "~ Ji 'aj "5 ?D^ ?^ i 'i .X *- «M ,"* ** M* JP §" S-! M 0 fl ? - t ~ ' ' 1 "— - 1 mi -J i

l5l - L. L is .a TJ XX ? S 5s u "* 01 s*s i i *" i1 -*? ! n S 5. -1 I § U^- —5 M ^- e •v 9 C •o * B -0 0 0 »^ u L P 4-1 ^ £ £ H C 4i Ul QJ 01 « || 3 a» " 1 4-J *> less — c l»fc I O 1• 4i n g O .a 0 -3 •9 * > k U 0 •a O o -sr R &0 - Ok U L. o e 1u o» tsfsi •O .f-t .S.3 w1 — T3 C a' i5 X*-1 a * o 3 X tvj QO SO 00 O I/I DO o CM i-•O '-O ^-o o — — ^j "T O ^ f* f.* O O> CVi CO.- CO •— o rg o t%, •1 d>- o oro irt C3 00 0 CO 3 sj eNJ '"^ ^~ CNJ CNJ O 000 k k k I i 3 ™ L I e L. - *> li M 01 li i! .e ? *.• c C X xC JS 4i i >1S 35 i i i *L$ C O I a. M : U o si =.l li ss ;s 1 £~" — N *! o | > O ,T, 30 — " 1— — — * ."g ^^^0* = § 'I ^I« ** 5 > X « E O — O '^ > "^ .8 -i n 0 — — • • • d 41 1V » C«3 — o O *J i O 03 |31 — — rs *j a: o* •0*0^ 0 X '0 "'O tJ ^* BUM '/i J> 5 , £ i •S E ^t e k M I/i • • £ x 5; "o L "•* 5 — « *0 £ J ^ « Oi 0 . M 7 .a — —- T 41 x2 V iT s *"" o o1 £ k n- a x J.O 0. S.c Sf fc Q k J S "5 £ " S E * .° * ..1 •0 , t. C 4) •« Or 0 « O Al 41 ill • 1 k X k w 41 • • • *J O m 2 X It^ 4.* — ^- «— C — * Of U w% 3 iT lT D *n w ^ w .re-tmenl -ge - O fsi .Q jj 41 O O O t .o .1-1 *— «) OJ — LO ,t jQ > > A i O O 4) a vi ""^ "* £ £ . ook ofi X* XO X * 11 "0 - •— W U -CO*— — • O — <QWiQkA*l*f*4l ) ^ t« e E u > B 31 >• o 2 • • ^ O u O ., «J O e « r • t'M D ,- — ' k 2 — O 0 O i-.i 1-i T? o vi o «-• .n ^ £ >,•— X IO X 41 O -~ i J a, > cc ' ^ O £ £..~"i~ &— 5 41 § ji^.^2 X 2 Cj 0 X S ^ XI rt) O 5S.5 ; «a f' •— ,i- •— f C"' s£~ LO 03 rsj v V CO o CM tNJ ^ «vi ra o «*> o d o V m ZZ'22 • w o 2 ^T m co ri OC r* in 00 * 41 S 1 • OJ 1- | I u e ^~ M e e C £3 3 *-* 0* II c e c c £I k w. , 41 41 41 1* 1' i 1 1 H I e o 0 Is. Montana Peterson, •i O •S-S3 fc m IL ||] C lo >«1 k II5 3 f^ OJ k * 25 In'S — k r.|_ 1— s 3 0 j^ — ^25 ,"-* I .S o£ *^w "• X O < O» K ^ i- i— i . •- U ^ 5 » M

l52 tn * g : o c •— c 3T3 «l o ** « S Z •— 4* *J C M C N •£S k c *^ - — .0 t/1 O /i ^a •o u k k W .^ 1 « N f* U G .o ^ — -« CL -o fc fl OJ u > S ISiliSS *J • 01 C Ml o vj c x a X •§1•'1 Pfem |« CV V — ' > X > O ? CM •— ^^ S 3 ^ e •• »* 3 S .0 E 3 iJ r^. O u . - n j iA tsi k 9t CJ k * k k — k Q. • k *J ^ • C ** O — ".' O 3 — e . •fc '— 0 'J s-a "c ? •? §. C *rf ? i 3 i- v, .^ >, Z£g ^ '= 5,- 3- — to w1 a* wt CM r* .j 9 &«*0 V & t* 4 ,c ^ 0 £ k k D«— . J1< "3 ^ fc £ a - — k O O II c"^" a o *•> CT1 O **• § i s i C X M OJ 43 ^* CM CM . ' 1 •— o ^- • — 3 • • . r— J^ E 0 c 13 '- U *S*t **^1 *~ • **> rt * - ": "°. "•? rv ^- 4* ftm •— ^- • o ^f u^ _. •g u o ~ -j JZ f. -. ~ " tN* - u • J b C ~ 8 =i •3 ^. 4 *J iM*l — O IVll r< " Ok | c 3 E.i k i k k k l. k. k V y 4i u ai • - i i C *> 4-* *J 2 >i "3 ~, z •1 c 1l- ^-14- S1 w c c k e c c S £ • * U. 31 V 1 i i * 52 c o 8 £1 la I i Is £ g lo. I I U.- i; ' 3 li c . , X £ C *0 i i ;• f '' • •o £t *^ c 1n « 13 -y-- •A 19 5 ^* o •1 •— »— co o^^ •^ "i - o u S,^ O . -CT1 3 « ^ »Z^— ^ i— i if^. S. "o - ^ — -.* — — .^-j -i .o at ^ o S •— » s_. ft N — DC fc*- ~J - ^ ^~ ~ "a .-) .0 2> - ^ — .^ x • E -^ — .? y ^ 41 . 3. ; W '4 Sg • C * "1 -M. «1 O 1V t 5 — C * > k. k V o * -* C g"5S « — * — 3 O « •- XX >. ** WB • WM .a — J « ft) 2 lA **• ** «l ^- *- .* —1 Ml

l53 lc (in/hr) UNGRAZED I • UNGRAZED SOURCE: Gifford and Hawkins (l978). FIGURE 3.2 Relationships between infiltration rates (f ) on grazed and ungrazed areas.

l54 studies showed no difference between infiltration rates in pastures that were grazed with a Merrill deferred-rotation system and in livestock exclosures that had been ungrazed for 20 to 30 years. However, there was a significant difference between heavily stocked, continuously grazed pastures and ungrazed exclosures or Merrill deferred-rotation pastures. Knight and others (l980) studied the impact of short-term; moderate, continuous; and very heavily stocked grazing systems on infiltration rates (Figure 3.3). These rates were lowest in winter and peaked in late spring and early summer, fluctuations that corresponded very closely with vegetation growth. Generally, there was no difference between the moderate, continuous and short-duration grazing systems. The very heavily stocked system resulted in significantly lower infiltration rates than the moderate, continuous or short-duration systems, except for January l979. The impact of grazing on infiltration rates will vary depending on range condition. Generally, infiltration rates of a site improve as the site's range condition improves (Osborn l952, Rauzi and others l958, Leithead l959, Rauzi l960, Johnston l962, Rhoades and others l964). Several researchers (Packer l95l, Marston l952, Orr l970, Meeuwig l970a) have recommended that somewhere between 60 and 75 percent of plant and litter cover is required to control surface runoff. Recently Dadkhah and Gifford (unpublished manuscript), found that trampling may have a much greater impact on infiltration rates than does plant cover (Figures 3.4 and 3.5). Generally, a plant cover maintained at approximately 50 percent is just as good as an 80 percent litter cover in controlling infiltration rates on slopes of at least l5 percent. As slopes become steeper, increased plant cover is required to control infiltration. Water Quality Grazing is a diffuse activity and thus is a potential source of nonpoint pollution. Available literature indicates that sediment and bacterial water-quality indices are the parameters of greatest importance with wild equids. Sediment from Upland Sites. Lusby (l970) reported that an ungrazed watershed produced 45 percent less sediment than a heavily grazed watershed. These comparisons were based on a paired watershed approach, but the two were not calibrated against one another before the grazing treatment was applied. Lusby assumed the watersheds were similar before treatment. Studies conducted in a Louisiana forest (Duvall and Linnartz l967) and on a Canadian grassland (Johnston l962) did not show an increase in sediment production from heavy grazing. A study on the Manitou Experimental Forest, Colorado found no difference in sediment production between moderately, continuously grazed and nongrazed areas (Dunford l954). Similar results were reported by Rich and Reynolds (l963) in a study conducted on the chaparral lands of Arizona. However, moderate grazing of a subalpine

l55 -5 2 O M i CO i to 8 i to 1 o o tT) ft a T> V to •H 0) fri 10 C0 » id 3 M n) * " CO § o 3 '6 a o .H •U CO r*. 05 N id S 10 CO f» i 8 M

l56 I • • TM -IM -nt SOURCE: Dadkhah and Gifford, unpublished manuscript. FIGURE 3.4 Infiltration rate curves for different soil compaction treatments. Dotted portions of each curve represent time periods after l0 minutes when an interaction between rock cover and trampling existed.

l57 f ' i , -. H\ - - SO\ M Tim* . mlnulas Figure 3.5 Infiltration rate curves for different percentages of grass cover. Dotted portions of each curve represent time periods after 15 minutes when an interaction between grass cover and trampling existed. (From Dadkhah and Gifford, unpublished manuscript.)

l58 range in Utah increased sediment production over that of nongrazed areas (Meeuwig l965). Meeuwig emphasized the need for proper management to reduce grazing impacts. A change in grazing management on part of the upper Rio Grande Basin from heavy, continuous grazing to moderate, summer-deferred grazing resulted in sediment being decreased from l.72 to 0.54 metric tons/ha (Aldon and Garcia l973). Buckhouse and Gifford (l976a) reported that 2 weeks of heavy grazing on a chained pinyon-juniper site in southern Utah did not increase sediment production over that of an ungrazed area. Heavily, continuously grazed pastures on the Edwards Plateau, Texas produced more sediment than a pasture grazed under a four-pasture, deferred-rotation system, but was not different from an exclosure that had not been grazed for 28 years (McGinty and others l978). Hood and others (l978b) measured sediment loss from three vegetation areas in eight different grazing treatments on the Rolling Plains, Texas. In shrub zonal areas, heavy, continuous grazing did not increase sediment production over that of nongrazed zonal areas. Heavily, continuously grazed midgrass-interspace areas produced more sediment than did the four-pasture, deferred-rotation system or a 20-year-old livestock exclosure. Knight and others (l980) studied the influence of short-term; moderate, continuous; and very heavy, continuous grazing on sediment production at Edwards Plateau, Texas. He found no difference in sediment production between moderately, continuously grazed pastures and those grazed for a short amount of time (Figure 3.6 and 3.7). However, there was a large increase in sediment production from the very heavily, continuously grazed pasture as compared to the other two levels of grazing intensity. As previously indicated, soil compaction by livestock has a major impact on infiltration rates. Trampling also has some impact on sediment production, but the overwhelming influence is through plant-cover manipulation. Dadkhah and Gifford (unpublished manuscript) have found that trampling (up to perhaps 30 percent disturbance) has a major impact on sediment production from bare and sparsely vegetated loam soils. However, as plant cover increases beyond perhaps l5 to 20 percent, it dominates control of sediment yield. Once cover extends over approximately 50 percent of the land, sediment production is statistically the same minimal value, regardless of the trampling disturbance. These findings conflict with the suggestions of Meeuwig (l970b) and Meeuwig and Packer (l976), which were based simply on a visual inspection of regression relationships, that 65 to 75 percent cover is required to minimize sediment production. This broad range obviously provides latitude for management recommendations, and the sediment issue becomes especially critical on rangelands that are only capable of supporting 50 percent cover at best. To summarize, most studies show little or no difference in sediment production between lightly and moderately grazed pastures. Likewise, many of these studies show no difference between nongrazed areas and lightly or moderately grazed pastures. However, they almost always show an increase in sediment production from heavily grazed pastures as compared with lightly, moderately, or ungrazed pastures.

l59 N « s- ID s K I en «0 1 S- l CD Sc i- •l- 4-> to OL Vt OJ « g. o N C .i- n> O 4-> i- O «J CO L. O) 3 4) 4-1 Q (O •O Z O tf) I/I £ cr> s- O (T) 0i •M T3 X O CC h< oo CM O tN I to I CM g M CM (VH/SN01) NOiionaoyd iN3wia3s aaivinwnoov

l60 o> Nouonaoyd (VH/SNOi)

l6l Sediment from Riparian Zones. Sediment production within the riparian zone is often a serious matter, as problems with bank stability are often encountered. The bank-stability problem is often a combination of such things as destruction of vegetation, mass wasting, and bank cutting. In general, natural, stable, well-vegetated stream banks help maintain stream-channel integrity (Ames l977, Davis l977, Glinsk l977, Kennedy l977). Knight (l978) studied the influence of deferred rotation; rest rotation; moderate, continuous grazing; and no grazing on riparian habitat in the Blue Mountains, Oregon. He found that overwinter effects were more extreme than could be accounted for by livestock grazing. However, several recent papers (e.g., Cooper l978) suggest that there is very little scientific information available on the relationships between grazing and riparian habitat management. Much of the available information is based on observations and a few case studies. Bacteriological Quality. Stephenson and Street (l978) investigated both limited and intensive grazing systems at Reynolds Creek Watershed in Idaho. Results indicated that fecal coliform counts in the waterway increased after the cattle were introduced and remained at elevated levels for a period after the cattle were removed. No correlation was found between fecal coliform and stream discharge, but there was a correlation between total coliform and stream discharge (r = 0.85). Sites subjected to summer grazing followed the same general pattern, with dilution and other inhibiting factors reducing concentrations as sampling progressed downstream. Inconsistencies were attributed to the tendency of cattle to congregate around trees for shade and the creation of a holding area for the fall roundup near the channel. This latter situation possibly resembled a feedlot. On winter range sites, total and fecal coliform counts were dominated by runoff events. Again, counts increased when cattle were returned in the fall, and a partial snow cover and frozen ground allowed runoff and flushing during the winter. Irrigation return flow washed the bacteria into the stream from April to August, when the cattle were no longer on the pasture. An attempt was made in the Reynolds Creek study to use a proposed BLM allotment plan to study fecal contamination, but incomplete construction of fences allowed cattle to move back into previously grazed allotments, thereby confounding the results. Counts increased as cattle were moved onto the allotments and fell after they were removed. An intense storm 3 weeks after the cattle were removed from both allotments caused considerable increases in counts (between 200 and 2,000 percent). Johnson and others (l978) studied the effects of flood-plain grazing by l50 cattle in the Colorado Front Range. An 85-ha pasture bisected by a perennial stream was grazed from early April to mid-June, l977. Six samples were taken prior to, and after, the removal of cattle. Sampling did not begin until early June. Results indicated significant increases in both fecal coliform and fecal streptococci counts over an upstream control. Within 9 days after

l62 removal of the cattle, neither type of bacteria was statistically different from its counterpart in the control. Darling (l973) investigated cattle and sheep grazing on allotments in the Logan River Basin. Fecal coliform were monitored at low levels (<10 counts/l00 ml) before the cattle and sheep were moved onto the allotments; after they moved in the fecal coliform count rose, but so did that in the control watershed. Fecal streptococci counts showed much variation, but differences were significant for all indicator bacteria when compared with an ungrazed watershed. The influence of various land-use practices on bacterial water quality in a humid region was investigated by Kunkle (l970). The land-use categories included forest, pasture, barns, village, and a composite, and the quality of water stemming from the forested area was considered the control. Fecal coliform counts were closely related to stream discharge levels in the grazed area. Also, counts exhibited a hysteresis loop with stream flow much like suspended sediment on the same watershed. The author also found that bacteria were stored at the bottom and bank areas of the streams, and if agitated, became part of the stream load. By wading a reach of the channel, Kunkle found an increase in fecal-coliform concentration 30 m downstream and concluded that a floodwave could disturb these storage areas and release the bacteria. The influence of channel storage and lack of overland flow on the watershed led to the conclusion that very few bacteria were flushed during the events originated on upland areas. The channel and zones adjacent to the channel where overland flow was observed were cited as source areas for the bacteria. A grazing study conducted outside the channel to determine bacterial pollution was reported by Buckhouse and Gifford (l976b). The study site was a chained pinyon-juniper community in southeastern Utah. The slash was windrowed and seeded to crested wheatgrass (Agropyron cristatum), and part of it was grazed at a rate of 2 ha/AUM. Data were generated with a simulated 28-minute rainfall. Results indicated no statistically significant difference between grazed and ungrazed aras. Analysis of source material revealed that sufficient numbers of bacteria existed for several weeks after deposition. Therefore, low counts were attributed to the low density of fecal material, which at this stocking rate covered approximately 0.2 percent of the area. Doran and Linn (l979) monitored runoff from a Nebraska pasture grazed by cows and calves for total coliform, fecal coliform, and fecal streptococci during l976, l977, and l978. Bacteriological counts in runoff from both grazed and ungrazed pastures generally exceeded recommended water-quality standards. Runoff from the grazed pasture contained 5 to l0 times more fecal coliform than runoff from the ungrazed pasture. The authors reported little difference in total coliform counts between the grazed and ungrazed areas, but fecal streptococci counts were higher in runoff from the ungrazed pasture and reflected the contributions from wildlife. They found the fecal coliform/fecal streptococci ratio in pasture runoff useful in identifying the relative contribution of cattle and wildlife. Ratios below 0.05 were indicative of wildlife sources and ratios above 0.l were characteristic of cattle.

l63 Buckhouse and others (l979) studied the influence of livestock grazing on the bacteriological quality of Meadow Creek in the central Blue Mountains of eastern Oregon. They found no detectable increase in coliform counts due to livestock grazing. Observed increases in bacteriological counts were due mainly to natural causes. NEEDED RESEARCH Overview The Committee recommends that studies on habitat selection, food preferences and consumption rates, plant responses to grazing, animal population responses, and hydrologic effects be paralleled with empirical observations on ecosystem responses to grazing by different numbers of equids, with and without other herbivores. We accord first priority to studies of horses with and without cattle for several reasons. l. Horses are more numerous and widespread than burros. 2. Dietary data indicate that horses are more likely to compete with cattle than with sheep. 3. Burro-livestock overlap and possible competition are potentially less extensive, if for no other reasons than that burros are more limited in distribution and the desert areas they occupy have a limited carrying capacity for livestock. 4. The following research designs will make it clear that the facilities and numbers of replications needed to study horse-cattle interactions alone will be demanding of personnel, space, and finances. Adding the study of sheep would double these demands. Horse-sheep interactions need to be investigated, but in our view not until horse-cattle studies as well as some of the other research advocated in this report have been provided for adequately. Impacts on wildlife need attention, but we are not recommending formal studies at this time for two reasons: l. The possible combinations of interactions are too numerous to be studied at present, given the amount of resources that are likely to be provided: horse-elk, horse-bighorn, horse-pronghorn, horse-deer, burro-desert bighorn. Such a large number of controlled studies seems out of the question for now. 2. Experimental controls of the sort advocated in the horse-cattle studies are difficult to establish for wild ungulates, particularly when replication is demanded. What we do urge is that the federal and state agencies be alert to the possibility of doing before-and-after censuses on wildlife populations in areas where horse or burro herds are to be reduced. If several desert bighorn populations, for example, could be censused before planned burro reductions and for several years afterward, these

l64 counts could serve as experiments. If, in addition, a number of populations in nearby areas where burro reductions were not undertaken could be similarly counted, the overall result would be a roughly controlled experiment that hopefully would give more accurate answers than we now have on burro-sheep competition. Similar opportunities involving horses and wild ungulates should also be exploited. Accordingly, we propose the following three projects, which concern horse-cattle impacts on vegetation, range hydrology, and riparian zones. Project 8. Grazing Impacts of Equids and Cattle on Range-Plant Communities Rationale PL 95-5l4 requires that a program of research be developed that defines what constitutes excess numbers of animals. "Excess" can be interpreted from both the standpoint of animal population welfare and welfare of the range. On the latter point, there appears to be little quantitative documentation in the published literature of grazing use by either horses or burros. Several studies have documented plant utilization levels by both burros (Koehler l974, Carothers and coworkers l976, Hanley and Brady l977) and horses (Salter and Hudson l979), but grazing by other ungulates using the same range has often confounded the results, and the attribution of specific levels of defoliation to specific numbers of animal days of grazing use has not generally been done. Protracted studies (>5 years) on range trend are quantitatively needed to document adequately the effects of horses on plant communities. However, in view of the immediacy of the question and the brief amount of time available to conduct the required studies, quantitative data on plant utilization levels can be used to arrive indirectly at an approximate definition of "excess" from the standpoint of range-plant community use and stability. It is quite apparent that herbivores affect the physical structure and botanical composition of the plant communities upon which they graze. However, for a given stocking rate or level of herbage removal, the grazing effect is not uniform across animal species because various animals possess different forage preferences, dietary habits, and grazing behaviors. To plan effective grazing management programs, the range manager needs to know more than merely which plant species are consumed or preferred by his grazing animals and the extent to which their diets overlap. He must also understand the temporal and spatial patterns of such grazing use, the relative degree of foliage removal on the major forage plants in the community, and the parts of various plants that are consumed. All of these factors contribute to the persistence or the demise of individual plants in a community and, ultimately, to shifts in competitive relationships and community succession. These relationships have been studied relatively well under cattle and sheep grazing regimes, but have received little attention where equids graze, either alone or in combination with other large herbivores.

l65 Ideally, such studies should endure for more than 5 years, so that successional changes in plant communities could be directly evaluated. However, it is possible to draw inferences from short-term studies in which utilization patterns on individual plants are carefully monitored and related to published information on the physiological responses of these plants to various levels of defoliation. Such studies would also form the beginnings of longer term investigations where successional change is monitored. Emphasis is given here to horse-cattle impacts because they appear to constitute the most prevalent problem all over the West. However, we are not implying that other domestic animals (e.g., sheep) should not be considered if they are important locally. Nor do we imply that burro-livestock or burro-wildlife interactions should not be considered. Objectives l. Determine grazing distribution patterns (habitat selection) of horses and cattle when grazed as single species or in combination. This objective is to be addressed in Project l (see Chapter 2). 2. Within important habitat segments, determine for the major forage species: a. Temporal level of utilization (percentage of production removed during current year) b. Use of standing dead or previous year's (in the case of shrubs) plant material c. Plant parts utilized d. Frequency and timing of defoliation e. Production of forage regrowth and its utilization. Methodology l. In contrast to Project l, this grazing-impact study should be conducted in large paddocks, say from about l00 to 320 (40 to l30) acres in size. These can be set up in small temporarily or permanently fenced pastures so that local environmental variation can be adequately controlled. The actual size of these pastures will depend upon the present grazing capacity of the areas selected for study and the seasons and durations of the grazing periods. 2. Domestic horses generally representative of the locale's wild types might serve as experimental models in studies of plant utilization patterns and rates, daily forage-consumption rates, dietary nutritional value, and—to a partial extent—forage preferences. Representation of social units and sex-age structures typical of wild herds will obviously not be possible in such small-scale, highly controlled studies. However, major attention should be focused on the reproductive female. Domestic cattle used in these studies should represent local conditions in terms of breed,

l66 sex, age, and reproductive status. For example, steers should not be used to represent lactating cows simply for the sake of technical expediency. 3. The season(s) of grazing should also represent local conditions. For example, if existing conditions include year-long use by horses of sagebrush-bunchgrass range, with typical spring use by cows and calves, the small-scale pasture studies should be designed to adequately represent and sample these conditions. 4. The treatments in this project should be the same as those in Project l: a. Grazing by eguids at carrying capacity b. Excessive grazing by equids c. Grazing by cattle (or sheep) at carrying capacity d. Excessive grazing by livestock e. Grazing by equids and livestock at carrying capacity f. Excessive grazing by equids and livestock. In addition, there should be control pastures with no grazing by either class of animals. The stocking rates should be determined through consultation with professional range managers and scientists (BLM, USFS, Soil Conservation Service [SCS], and university scientists and extension specialists) who are familiar with problems and conditions of the particular study area. The combination treatment should be stocked so that the year-long level of grazing use is on an equivalent AUM basis for the two animal species concerned. 5. Where possible, grazing use by native ungulates should be incorporated into the design. This factor might be treated as a separate main effect with appropriate blocks for combinations of horses (burros), domestic livestock, and native ungulates, or it might be analyzed through an alternate experimental design (e.g., split-plot) in which separate blocks are not considered. The use of tame, hand-reared animals in either of these design contexts would provide an avenue for highly controlled studies of equid-wild ungulate interactions. With the exception of bighorn sheep, considerable research experience has been amassed from tame-animal studies of various habitat features (e.g., see Currie and others l977, McMahon l964, Neff l974, Reichert l972, Smith and others l979, Wallmo and others l972 for studies on mule deer; Collins l977 for elk; and Schwartz and others l977 for pronghorn antelope). As a minimum consideration, range and forage use by native ungulates should be monitored in any study location where their presence is important. Attention should be focused both on social interactions with horses (burros) and on possible interactions mediated through the food or water resource. 6. Utilization patterns and rates should be carefully measured as a basis from which plant-community change can be inferred. Utilization rates and levels should be closely tied to amount of grazing usage (in AUDs or AUMs; see Table 3.l for definitions) by the animal species concerned. This approach necessitates a careful quantification of growth (and regrowth) curves for ungrazed representatives of plant species studied.

l67 7. Findings on seasonal levels and rates of species utilization should be related both to appropriate published data and to inferences drawn regarding the likely directions and rates of community change under various levels of grazing pressure and kind of animal use. Examples of literature pertinent to such information might include West (l968), Willard and McKell (l973), Pechanec and Stewart (l949), West and others (l972) , Hyder and others (l975) , Smith (l967) , Cook (l97l), Ellison (l960), and Laycock (l967). The recent bibliographic source by Vallentine (l978) should also be consulted in this pursuit. 8. The approach advocated here will yield rapid, although relatively imprecise, answers to questions about grazing impacts. An essential part of such a study is early establishment of permanent range-trend transects in both grazed and ungrazed pasture. If the experimental grazing treatments could be carried on past the 2-year limits imposed upon this study, such transects would provide the long-term data necessary for ultimately determining the effects of wild, free-roaming horses and burros on plant-community change. Such long-term research could be maintained beyond the initial 2 years by applying only low levels of labor and money. The potential payoff is high. Project 9. Hydrologic Impacts Rationale Virtually no information exists on the impacts of wild equids on range hydrology. Likewise, few data have been recorded on the influences grazing livestock have on range hydrology for the western rangeland now supporting wild equids. To plan effective management programs, the rangeland manager needs to know the impact of wild equids, livestock, and/or wildlife on hydrologic parameters. Hydrologic responses are often much slower and sometimes less obvious than vegetation changes. In order to evaluate hydrologic change adequately, such studies should endure for more than 5 years). Emphasis is given here to the impact of wild equids and wild equid-cattle interactions. This does not mean that other animals should not be considered if they are locally important. Objectives l. For important habitat segments, determine the influence of wild equids and wild equid-livestock interactions on selected hydrologic parameters. Use the data collected in objective l to: 2. Construct new models or develop variables for existing hydrologic models such as the Hawkins and Gifford (l979) model or the universal soil-loss equation.

l68 Methodology The following research should be carried out in conjunction with Projects 2 and 8. l. Because of the relatively low and highly variable precipitation on the wild equid range, and the relatively short time frame of this study, data obtained with simulated rainfall should be the most reliable. This recommendation does not preclude the use of supplemental natural runoff plots or small watershed data if local conditions are favorable and funds are available for such studies. Ideally, this study should be both intensive and extensive in design. a. Intensive Study. This study should be conducted using small runoff plots (l.83 x 22.l3 m or 6 x 72.6 ft) located on 9 percent slopes. This plot size should be large enough to include the variability present in an area, particularly that associated with vegetation patterns (i.e., brush-canopy areas and interspace areas). Data obtained from a plot of this size could be used to refine the universal soil-loss equation for use on western rangelands. The runoff plots should be instrumented with a l5.2-cm (6-in.) H-flume and a 20.5-cm (l-ft) Coshocton wheel sampler or similar water measuring and sampling devices. These runoff plots would measure and sample natural and simulated rainfall events. A rainfall simulator (something similar to Meyer's but placed in line) should be used at selected time intervals during the year to apply rainfall to the runoff plots. The rainfall application rate should exceed infiltration rates, and raindrop size and terminal velocities should approximate a natural storm of the intensity simulated. A minimum of three runoff plots should be used per grazing treatment and vegetation-soil unit sampled. From these runoff plots the following information should be collected at predetermined intervals throughout the year: • Infiltration rate • Sediment production • Additional water-quality parameters • Vegetational cover by species • Bare ground cover • Shrub-canopy area and interspace area • Microtopography. From shrub-canopy and interspace areas adjacent to the runoff plots, the following information should be collected at the same predetermined intervals:

l69 • Soil bulk density Soil organic matter Soil aggregate stability Grass standing crop. b. Extensive Study. This study should use a small portable rainfall simulator (scaled-down model of the one described above) and variable runoff plots that will allow for separation of shrub-canopy areas and interspace areas. Runoff plots should be 0.5 to l nr in size. Simulated rainfall rates should exceed infiltration rates, and raindrop size and terminal velocities should be similar to that of a natural storm. These plots should be relocated at selected intervals on a similar site before the next simulated rainfall is applied. A minimum of six runoff plots should be used per grazing treatment and vegetation-soil unit sampled. The following data should be collected from each runoff plot: Infiltration rate Sediment production Vegetational cover by species Bare ground cover Microtopography Soil bulk density Soil organic matter Soil aggregate stability Grass standing crop. 2. The data collected in the study described above should be used in the construction of new models or in developing variables for existing models, such as the Hawkins and Gifford (l979) model or the universal soil-loss equation. Project l0. Riparian-Zone Impacts Rationale The riparian zone is extremely important, not only in terms of total watershed behavior, but also in terms of its individual components—fish, wildlife, insects, vegetation, birds, and soil. This study would concentrate on defining the impacts of wild equids and cattle on hydrologic behavior and water quality, both of which affect all other components of the riparian zone. Objectives l. Determine the amount of suspended sediment, coliform, total dissolved solids, and dissolved oxygen; measure the water temperature

l70 and general channel stability of selected portions of the riparian habitat grazed by wild equids. 2. Relate measured water-quality and channel-stability characteristics to the characteristics of the adjacent riparian habitat. Important characteristics would include percentage of vegetative cover, grazing intensity, and rest periods. Other important considerations include general behavioral characteristics of the study animals with respect to use of the riparian habitat. Methodology The design of this study will depend to some extent on whether the grazing animals are using springs, established water holes, or live streams as a source of water. It would be desirable, if possible, to include riparian zones with the headwater originating in the pastures and in relatively good biological condition. The stream characteristics should be as similar in all treatments as possible. If these criteria cannot be met and a reach of a river or stream has to be used, water-quality sampling will have to be carried out on a "flow-through" basis. Samples will have to be taken as water enters a specific reach representing a given grazing treatment, and again as water leaves the reach. This scheme dictates that intensity of use must increase in a downstream direction rather than in a random pattern. When assessing the impact of animals on ponds, seeps, or springs, similar water sources will have to be replicated in each pasture. Data should be collected at predetermined intervals throughout the year.