2. LGT Environmental Impacts

2.1 Range Utilization

2.1.1 Direct Indicators

2.1.1.1 Soil Structure

Soil structure can be characterized by the nature of soil aggregate stability and soil surface features. Grazing impacts surface features, bulk density, and aggregate stability. The effect of large herbivores on surface detention of water are mixed and dependent on grazing intensity and physical characteristics of the soil (Thurow 1991). Hoofprints made by livestock increase micro-relief if stocking intensity is moderate. If stocking intensity is heavy, micro-relief is reduced due to disaggregation of soil structure resulting in the soil surface becoming either loose dust if soil moisture is low, or flattened and compacted if soil moisture is high. Heavy grazing can alter micro-relief by reducing litter accumulation and bunchgrass growth forms, thereby increasing runoff velocity. Alteration in soil structure affects infiltration rates.

Livestock can impart substantial pressure loading to the soil surface (Tollner et al. 1990). Chancelor et al. (1962) found that bulk density was not affected below the 25 cm depth when loadings were 1.4 Mpa, indicating that animal compaction is a near-surface phenomena. Surface compaction has been associated with decreased porosity, lower air permeability, high cone index, and higher bulk density (Warren et al. 1986). Both Gill (1971) and Kilmer (1982) found grazing to increase compaction in sandy loam ultisols as grazing density increases, particularly in high moisture, early growth conditions. Tollner et al. (1990) noted that many of the negative compaction impacts by grazing livestock in alfalfa, bermudagrass, and wheat pastures were mediated by freeze-thaw cycles and shrink-swell process in southeastern U.S. soils. In areas prone to salinization, such as the flooding pampa in Argentina, grazing impact was not found to be a significant factor in affecting soil physical features (Alconada et al. 1993)

When contrasting native savanna in Oxisols of Colombia with savanna planted to derived pastures or derived grass-legume pasture, percent of soil aggregates greater .5mm was 9.6, 25.5, and 31.9, respectively, with differences occurring only between native and derived pastures (Rao et al. 1992). Water sorptivity (cm/sec ½) was 0.2, 0.4, and, 0.39 for native savanna, derived grass pasture, and derived grass-legume pasture, respectively. In a Brachiaria humidicola-Desmoium ovalifolium pasture in Peru, manure and leaf litter was the principal carbon source in these pastures. Soil physical and biological properties were affected by stocking rate but soil chemical problems were not. Bulk density increased with increasing stocking rates but also decreased when the animals were removed, suggesting a reversible effect of animal trampling. Increasing stocking rates did not affect root biomass, suggesting that trampling per se is not detrimental for plant production. However, earthworm biomass dropped precipitously when the pasture was overgrazed. Rotational grazing strategies can help root recovery.

Eldridge and Koen (1993) found that surface features on a grazed Eucalyptus intertexta-Callitris glaucophylla woodland were the overriding factors impacting hydrology and erosion of the ecosystem. Macroporosity, aggregate stability, organic matter, and differences in vegetation cover were the principal controlling surface features. Their conclusion was that vegetation cover is a good predictor of hydrological properties of degraded sites while physical properties were more influential on sites exhibiting greater stability of surface processes. Both Ludwig et al. (1994) and Tongway and Ludwig (1990) have stressed the need to understand pattern of surface features as they affect pattern of water, nutrient, and organic-matter flow across landscapes. Nutrients and water patterns flow across and beneath landscapes from a series of sources and sinks with grazing serving as a redistribution agent. As precipitation increases from semiarid to high rainfall humid grazinglands, surface feature conditions shift from primary impact on surface process to proportionally more subsurface, saturated matrix, and macropore lateral flow (Whisenant and Tongway 1995). How grazing impacts these process depends on whether erosion cells are created within the season to accelerate the normal flow process. That is, grazing per se does not alter systems significantly until landscape features are altered to a point where water, nutrients, organic matter, and sediment loss is accelerated. The challenge for resource managers is to identify threshold conditions which lead to formation of erosion cells, rapid deep drainage losses, and excessive lateral-flow leaching.

Many of the measures of grazing impact on soil and water processes must be viewed in the context of scale of measure. Several studies have indicated that runoff and infiltration processes are noticeably impacted by grazing pressure effects at the 1.0 m² level with reduced measurable impact as the size of the instrumented watershed increases to the landscape level (Sallaway and Waters 1994). Caution should be exercised when attempting to scale up small plot grazing responses to the landscape level. Any increase in rainfall intensity and/or storm size will increase soil erosion by water, and the same applies to any increase in wind regimes in more arid zones. As long as the erosion constitutes redistribution within the landscape, it does not lead to irreversible degradation. When it reaches the level where there is a net loss of soil from the landscape (via channels, streams, and rivers), degradation becomes evident (Tongway and Ludwig 1990).

2.1.1.2 Soil Erosion

See Section II-2.1.1.2 and Section V-2.2.11

2.1.1.3 Pesticide/Herbicide Residues

In 1980 agriculture used 72 percent of all pesticides applied in the U.S., and herbicides and insecticides made up 89 percent of the pesticides used by agriculture (National Academy Press 1993). However, unlike many areas of intensive agronomic agriculture, the use of pesticides on rangeland is minimal. Of the pesticides used, herbicides are the most prevalent. The cost of these chemicals and their application and the relatively low productivity potential of rangeland ecosystems limits herbicides treatments, even in developed countries. For example, in the U.S., Texas, a state with recognized woody-plant problems in most of its major land resource areas, only about 1.5 million acres of mechanical and chemical brush and weed control are applied each year. Only about one-half of these acres are treated with herbicides, representing less than 1.0 percent of the total rangeland acres in the State.

Recommended application rates for herbicides vary with the chemical, the target plant species, plant phenology, and, in some cases, soil textural and organic carbon characteristics. However, the common rates recommended for application on rangelands are less than 2 pounds (0.9 kg/ha) of active ingredient per acre and in many cases 1 pound per acre or less (less than 0.5 kg/ha). These low rates are not conducive to major movement of chemicals in runoff following precipitation events on treated areas. However, chemical contamination of water is possible under certain conditions. Major factors contributing to herbicide movement from rangelands into water sources are the amount of chemical applied, kind of vegetation/soil surface intercepting the herbicide, the proximity to surface watercourses, permeability of soil profiles and mobility of the herbicide, and the intensity and timing of precipitation events in relation to herbicide application. For this study, probability for residues in water in the ecozones where herbicides are used will probably be associated with differences in amounts of precipitation typical for the areas.

There are relatively few chemicals that are approved (labeled by EPA) for use on U.S. rangelands as herbicides. The most common of these chemicals include clopyralid, dicamba, 2,4-D, glyphosate, hexazinone, picloram, tebuthiuron, and triclopyr. Some of these herbicides are also commercially available in mixtures or may be tank mixed. It will be most useful to discuss the potential for residues of these materials in water by the individual chemicals.

Pesticides are classified as nonpersistent if they have half-lives of 30 days or less, moderately persistent if they have half-lives longer than 30 days but less than 100 days, and persistent if they have half-lives longer than 100 days. A pesticide is likely to contaminate groundwater (leach) if its sorption coefficient (the likelihood that it will attach to soil particles) is low, its half-life is long, and its water solubility is high. The pesticide residues most commonly found in U.S. groundwater include alachor, aldicarb, atrazine, bromacil, carbofuran, cyanazine, DBCP, DCPA, 1,2-dichloropropane, dinoseb, dyfonate, EDB, metolchlor, metribuzon, oxamyl, simazine, and 1,2,3-trichloropropane (U.S. Environmental Protection Agency 1986). These chemicals are not commonly used as herbicides on rangelands for brush control. Most herbicides used on rangeland are classified as nonpersistent or moderately persistent.

In general, the groundwater table is shallower in humid regions than in more arid regions. A shallow depth to the groundwater offers fewer opportunities for pesticide sorption and degradation. The travel time of the pesticide to the water table may range from days to a week if the depth to the water table is shallow, the soil is permeable, and the amount of rainfall exceeds the water-holding capacity of the soil. In contrast, the travel time may even be decades in arid regions where the water table is tens of meters below the land surface.

Hydrogeologic conditions (underground plumbing) beneath the soil profile may dictate the direction and rate of chemical movement. The presence of impermeable lenses or layers in the soil profile and underlying strata may limit the vertical movement of pesticides. Such impermeable layers may, however, contribute to the lateral flow of shallow groundwater and to the eventual discharge of groundwater and its contaminants into surface waters. However, the presence of high-permeability earth materials, such as sands and gravel, may greatly accelerate the vertical and horizontal flows of contaminants. Of particular concern is the presence of karsts (limestone) and fractured geologic materials that generally transmit water and chemicals rapidly to the groundwater body (National Academy Press 1993).

2.1.1.3.1 Picloram

Bovey (1993) and Bovey and Scifres (1971) summarized the literature on dissapation, movement and environmental impact of picloram and other herbicides on rangelands. Several rangeland herbicides including picloram have been studied to determine movement in runoff water after application. The maximum loss obtained for picloram, dicamba, or 2,4,5-T was 5.5 percent, and the average was approximately 3 percent. The time interval from the first rainfall determined the amount of picloram and other herbicides that moved into the soil profile and/or the amount that moved away from the point of application with surface runoff. Four months after application, losses from all herbicides were less than 1 percent of that lost during the initial 24 hours after application (Trichell et al. 1968). Scifres et al. (1971) indicated that picloram moved in surface runoff when 0.28 kg/ha was applied in the Rolling Plains of Texas for control of honey mesquite. Irrigation the first 10 days after application resulted in a concentration of 17 ppb of picloram in surface runoff. Irrigation at 20, 30 or 45 days resulted in less than 1 ppb of picloram residue in runoff water. No more than 1 or 2 ppb picloram was detected after dilution of runoff water into ponds.

Baur et al. (1972) studied picloram residues from an 8.0-ha watershed treated with the potassium salt of picloram at 1.12 kg/ha in the Post Oak (Quercus stellata) Savannah land resource area of Texas in 1969 and 1970. Samples were collected directly below the treated area and in streams below the plots. Samples taken adjacent to the plots declined to less than 10 ppb by 10 to 12 weeks after application. Eight days after application, water sampled 1.2 km from the plots contained less than 1 ppb of picloram. Less than 1.0 ppb were occasionally detected 1.6 km from the plots 8 months after treatment.

Research shows that herbicide residues can occur in surface water if heavy rainfall occurs soon after treatment. When pelleted picloram was applied at 2.24 kg/ha to a 1.3-ha rangeland watershed, surface runoff of 1.5 cm from a 2.1-cm rain received 2 days after treatment contained an average of 2.8 ppm of picloram (Bovey et al. 1978a). However, picloram content declined rapidly in each successive runoff event, and runoff water contained less than 5 ppb by 2.5 months after application. Loss of the potassium salt of picloram from grassland watersheds in surface runoff water was similar whether the picloram was applied as aqueous sprays or as pellets. Picloram plus 2,4,5-T at 0.56 kg/ha each were applied May 4, 1970, December 4, 1970, May 4, 1971, October 8, 1971 and May 5, 1972 on a Houston black clay soil. (Bovey et al. 1974). No runoff event occurred until July 25, 1971, 72 days after the third herbicide treatment. Concentration of 2,4,5-T and picloram averaged 7 and 12 ppb, respectively, in runoff water after July 29, 1971 herbicide content was usually less than 5 ppb. Highest concentrations of 2,45-T (26 ppb average) for 1971 occurred in runoff water on July 27 and 29 after heavy rainfall. Data indicated that picloram or 2,4,5-T content was typically less than 5 ppb in runoff if major storms occurred 1 month or longer after treatment on the Houston black clay.

On sandy soils Scifres et al. (1977) found only trace amounts of picloram or 2,4,5-T, which had been applied at 0.56 kg/ha each, in surface runoff water following storms about 30 days after application. Mayeux et al. (1984) found that maximum concentrations of picloram were 48 and 250 ppb in initial runoff from an 8-ha area treated with 1.12 kg/ha in 1978 and 1979, respectively. Herbicide concentration decreased with distance from the treated area in proportion to the size of adjacent, untreated watershed subunits that contributed runoff water to streamflow. About 6 percent of the applied picloram was lost from the treated area during active transport.

Bovey et al. (1975) conducted an investigation to determine the concentration of 2,4,5-T and picloram in subsurface water after spray applications to the surface of a seepy area watershed and lysimeter site in the Blacklands of Texas. A 1-to-1 mixture of the triethylamine salts of 2,4,5-T plus picloram was sprayed at 2.24 kg/ha every 6 months on the same area for a total of five applications. Seepage water was collected on 36 different dates, and 1 to 6 wells in the watershed were sampled at 10 different dates during 1971, 1972, and 1973. Concentration of 2,4,5-T and picloram in seepage and well water from the treated area was extremely low (less than 1 ppb) during the 3-year study. No 2,4,5-T was detected from 122 drainage samples from a field lysimeter at another site sampled for 1 year after treatment with 1.12 kg/ha of a 1-to-1 mixture of the triethylamine salt of 2,4,5-T plus picloram. Picloram at levels of 1 to 4 ppb was detected in lysimeter water from 2 to 9 months after treatment. Supplemental irrigation in addition to a total of 85.5 cm natural rainfall was used to leach picloram into the subsoil.

In another study, Bovey and Richardson (1991) found that picloram and clopyralid remained in the uppermost 30 cm of a Houston black clay soil. The herbicides were sprayed at 0.56 kg/ha each on the same area in 1988 and 1989 on a seepy site overlying a shallow, perched water table in the Blacklands of Texas. Approximately 90 days after treatment, greater than 92 percent of the picloram had dissapated. No herbicide was detected in subsurface water from the area in 1988, but concentrations of less than 6 ppb of both herbicides were detected in subsurface water collected 11 days and from 41 to 48 days after treatment in 1989.

Johnsen and Warskow (1968) sprayed a 31-acre watershed in central Arizona with 1.7 pounds of picloram per acre (0.77 kg/ha) in June, 1965, with a helicopter. Less than 0.1 ppm were detected by bioassays in runoff water the first year after treatment. An estimated 0.03 pounds (.0136 kg) of the 53 pounds (24.04 kg) applied left the area during the first 18 months after treatment.

Davis and Ingebo (1971) applied pelleted picloram (10 percent active ingredient) at 9.3 pounds per acre (10.42 kg/ha) to a 2.1-acre (0.85 ha) side slope drainage area in a 46.5-acre (18.83 ha) watershed in central Arizona. Most water samples were collected at a weir directly below the treated area through which the entire watershed drained. The highest concentrations of picloram (370 ppb) were collected at the weir 7 days after the treatment, after a 2.53-inch rain (63 mm). Since the picloram-treated areas was only 4.5 percent of the total area drained, there was a possible 22-fold dilution in picloram content. In 3 months, picloram usually occurred in trace amounts and could not be detected after 14 months and 40 inches (102mm) of accumulated precipitation. The investigators indicated movement of picloram into streams was related to rainfall duration and amount and that water from picloram-treated watersheds may cause damage to crops if used for irrigation.

Research conducted in a South Texas environment indicated that consentrations of 0.055 to 0.184 ppm of picloram were detected in surface runoff water within 2 weeks after its introduction to adjacent treated areas at 1.1 kg/ha (Haas et al. 1971). Within one year, the concentration in runoff water was at or below detection limits (0.001 ppm). In no case did treatment of areas adjacent to domestic water wells (30 to 150 feet deep or 9 to 45 meters) result in detectable residues of picloram in those wells. Once picloram was moved into water catchments in the Rolling Plains of Texas, residues were detected for at least a year following treatment (Scifres et al. 1971). There is general agreement among workers that dilution is important in the dissipation of picloram from impounded water. Some feel photodecomposition is important in reducing picloram concentration in water. In the photolysis of picloram, certain levels of light energy are necessary for degradation of each molecule. Assuming light energy is randomly dispersed, then interception of photons by picloram molecules would be a random occurrence. In such a system, the degradation of picloram would be expected to occur rapidly at first then decrease as fewer molecules were available for light interception.

2.1.1.3.2 Clopyralid

Clopyralid is chemically closely related to picloram but reacts differently to certain weed species and is less persistent in the environment (Herbicide Handbook of WSSA 1989). A 1-to-1 mixture of the monoethanolamine salts of clopyralid and the tri-isopropanolamine salt of picloram was applied at 0.56 kg/ha each in May, 1988, and June, 1989, to the same area (Bovey and Richardson 1991). Neither herbicide was detected in subsurface water from the treated area in 1988, but concentrations of less than 6 ppb of clopyralid or picloram were detected in subsurface water collected 11 days and from 41 to 48 days after treatment in 1989. The study represents a worst-case scenario because the herbicides were applied twice to bare soil and were disked into the soil to prevent loss from photodecomposition. Under normal practices, the herbicides may be applied only once every 5 to 20 years to weeds and brush and are not protected from photodecomposition by disking.

2.1.1.3.3 Phenoxys

Trichell et al. (1968), using gas chromatograph and bioassay detection techniques, investigated the loss of 2,4,5-T, dicamba, and picloram from bermudagrass and fallow plots of 3 and 8 percent slope. When determined 24 hours after application of 2.24 kg/ha, a maximum of about 2, 3, and 5 ppm picloram, 2,4,5-T, and dicamba, respectively, were found in runoff water after 1.3 cm of simulated rainfall. Losses of dicamba and picloram were greater from sod than from fallow plots, whereas 2.4.5-T losses were approximately equal. Four months after application, picloram, 2,4,5-T and dicamba concentration in runoff water from sod plots had diminished to 0.03. 0.04, and 0 ppm, respectively. Maximum loss of any herbicide from the treated area was 5.5 percent and averaged 3 percent.

Bovey et al. (1974) sprayed a 1-to-1 mixture of the triethylamine salts of 2,4,5-T plus picloram at 1.12 kg/ha every 6 months on a native grass watershed for a total of 5 treatments. Plant wash-off was the main source of herbicide detected in runoff water. Concentrations of both herbicides was moderately high (400 to 800 ppb) in runoff water if 3.8 cm of simulated rainfall was applied immediately after herbicide application. If major natural storms occurred 1 month or longer after herbicide treatment, concentration in runoff water was less than 5 ppb.

Norris and Moore (1970) indicated that concentration of 2,4-D, 2,4,5-T, picloram, and amitrole seldom exceeds 0.1 ppm in streams adjacent to carefully controlled forest spray operations in Oregon. Concentrations exceeding 1 ppm have never been observed and are not expected to occur. Chronic entry of these herbicides into streams for long periods after application does not occur.

Bovey and Young (1980) summarized the literature on the fate of 2.4-D and other phenoxys in impounded water. In general, phenoxys decompose rapidly, especially if adapted microorganisms are present. Photodegradation of phenoxys in impounded water is also an important means of breakdown.

Edwards and Glass (1971) applied 11.2 kg/ha 2,4,5-T (an excessively high rate) to a large field lysimeter in Coshocton, Ohio. The total amount of 2,4,5-T found in percolation water intercepted at 2.4 m deep for as long as 1 year after application was insignificant. Bovey and Baur (1972) found little or no 2,4,5-T in soils 12 weeks after treatment in soils at five widely separated locations in Texas with the propylene glycol butyl ether esters of 2,4,5-T at 0.56 and 1.12 kg/ha.

Bovey et al. (1975) conducted an investigation to determine the concentration of 2,4,5-T and picloram in subsurface water after spray applications of the herbicides to the surface of a seepy area watershed and lysimeter in the Blacklands of Texas. A 1-to-1 mixture of the triethylamine salts of 2,4,5-T plus picloram was sprayed a 2.24 kg/ha every 6 months on the same area for a total of five applications. Seepage water was collected on 36 different dates, and 1 to 6 wells in the watershed were sampled at 10 different dates during 1971, 1972, and 1973. Concentration of 2,4,5-T and picloram in seepage and well water from the treated area was extremely low (less than 1 ppb) during the 3-year study. No 2,4,5-T was detected from 122 drainage samples from a field lysimeter at another site sampled for 1 year after treatment with 1.12 kg/ha of a 1:1 mixture of the triethylamine salt of 2,4,5-T plus picloram. Picloram at levels of 1 to 4 ppb was detected in lysimeter water from 2 to 9 months after treatment. Supplemental irrigation in addition to a total of 85.5 cm natural rainfall was used to leach picloram into the subsoil.

Phenoxy herbicides do not persist in water sources, and significant concentrations, if found, occur within a short time after treatment (Bovey and Young 1980). Loss of herbicides from treated areas by movement in runoff water is a very small percentage of the total applied even under intensive natural or simulated rainfall. Phenoxy herbicides rapidly dissipate in streams and are not detected downstream from points of application. In impounded water, phenoxys decompose rapidly, especially if adapted microorganisms are present. Even under large-scale applications to surface water sources, 2,4-D disappeared rapidly after application, and concentrations remained low or undetectable. In surveys of major river systems in the U.S., 2,4-D appeared infrequently and in minute concentrations.

2.1.1.3.4 Dicamba

Trichell et al. (1968) studied dicamba runoff from sloping sod plots in Texas. They found that as much as 5.5 percent of the applied dicamba was recovered in runoff water when 1.3 cm of artificial rain was applied 24 hours after herbicide application. No dicamba was found in runoff water from a similar artificial rain application 4 months later after a 21.6-cm natural rainfall event.

Norris and Montgomery (1975) found maximum dicamba levels of 37 ppb about 5.2 hours after treatment at 1.3 km from the point where the sample stream entered the treatment unit in Oregon, U.S. Dicamba residues detected the first 30 hours after application resulted from drift and direct application to exposed surface water. By 37.5 hours, residue levels had declined to background levels; no dicamba residues were found more than 11 days after application. Dicamba levels found in streams were several orders of magnitude below threshold response levels for fish and mammals.

In 1984 Muir and Grift (1987) sampled the Ochre and Turtle rivers which flow into Dauphin Lake in western Manitoba, Canada, to determine levels of MCPA, diclofop {(+)-2-[4-(2,4-dichlorophenoxy) phenoxyl] propanoic acid}, dicamba, bromoxynil, 2,4-D, triallate [s-(2,3,3-trichloro-2-propenyl)bis(1-methylethyl)carbamothioate], trifluralin [2,6-dinitro-N,N-dipropyl-4-(trifluoromethyl)benzeneamine], which were used widely in each watershed. Dicamba and 2,4-D were detectable throughout most of the sampling period in both rivers at low levels of less than 1 ppb. Levels of less than 6 ppb of dicamba and 2,4-D were detected in water from the Turtle River before a high-water event, possibly from sprayed ditches or rights-of-way near the river. Even so, discharges of all herbicides monitored in the study were less than 0.1 percent of the amounts used in each watershed.

Dicamba dissipated most rapidly from water under nonsterile, lighted conditions (Scifres et al. 1973). Pond sediment evidently contained microbial populations capable of decomposing the herbicide. Temperature was crucial in dicamba dissipation, especially in the presence of sediment. In some cases, influence of sediment on dissipation rate of dicamba was apparently augmented by light. Under summer conditions, dicamba at 4.4 kg/ha/surface area of ponds dissipated at about 1.3 ppm/day. Dicamba dissipated as a logarithmic function of concentration with time.

2.1.1.3.5 Triclopyr

Schubert et al. (1980) using a helicopter treated the upper part of a watershed in West Virginia, U.S., with 11.2 kg/ha triclopyr. Two streams traversed the treated area. Movement of triclopyr residues in soil and water down slope from the treated areas was insignificant.

Maximum concentration of triclopyr in stream water was 95 ppb the first 20 hours after application, similar to that observed for other herbicides applied to forest streams (Norris et al. 1987). Reduction in concentration the first 20 hours after application was attributed to photodecomposition. In September during the first significant rains after application in May, maximum triclopyr residues of 12 ppb were found in a small pond at the site. A 6-cm rain on November 7, causing a 6,500-L stream discharge, increased triclopyr concentrations to 15 ppb, but after November 11 no more triclopyr was detected.

Triclopyr was applied in both ester and amine formulations on October 24, 1986, to Coastal Plain flatwood watersheds near Gainesville, Florida (Bush et al. 1988). Panicum grasses (Panicum spp. and Dichanthelium spp.) wiregrass (Aristida stricta), gallberry (Ilexglabra), and most herbaceous plant species were controlled by both formulations. Triclopyr application resulted in a shift toward a bluestem-dominated understory. Triclopyr residues were detected at trace levels of 1 to 2 ppb in storm runoff during the first runoff event after application. No triclopyr residues were detected in subsequent runoff events or in any groundwater wells for 6 months after application.

2.1.1.3.6 Tebuthiuron

Pelleted tebuthiuron was applied at 2.24 kg/ha active ingredient to a 1.3-ha rangeland watershed. A 2.8-cm rain, 2 days after application, produced 0.94 cm of runoff, which contained an average of 2.2 ppm of tebuthiuron (Bovey et al. 1978b). Tebuthiuron concentration decreased rapidly with each subsequent runoff event. After 3 months, tebuthiuron concentration was less than 0.05 ppm; none was detected in runoff water 1 year after treatment. Concentration of tebuthiuron applied as a spray at 1.12 kg/ha, decreased to less than 0.01 ppm in runoff within 4 months from a small plot receiving simulated rainfall. On 0.6-ha plots, mean tebuthiuron concentration from sprays and pellets was 0.50 ppm less in water when the first runoff event occurred 2 months after application. Concentrations of tebuthiuron in soil and grass from pellet applications were less than 1 ppm and decreased with time.

Tebuthiuron applied at 1 kg/ha as 20 percent a.i. pellets to dry Hathaway gravelly, sandy loam soil in the spring diminished by 5 percent at the first simulated rainfall event, 37 mm, in runoff water and sediment (Morton et al. 1989). The second and third simulated rainfall events, 22 and 21 mm, respectively, removed an additional 2 percent of tebuthiuron. When tebuthiuron was applied to wet soil in the spring, the initial simulated rainfall events, totaling 42 mm, removed 15 percent of the tebuthiuron. When tebuthiuron was applied to wet soil in the fall, the initial rainfall events, totaling 40 mm, removed 48 percent of the tebuthiuron in runoff water and sediment. No significant differences were found in the total amount of tebuthiuron within the soil profile after application to dry and wet soils. More than half of the tebuthiuron had moved into the upper 7 cm 1 day after application. Tebuthiuron was not detected below 90 cm after 165 mm of simulated rainfall and 270 mm of natural rainfall.

2.1.1.3.7 Hexazinone

Lavy et al. (1989) found relatively small amounts of hexazinone in runoff water from a spot-gun application to a forest floor in Arkansas. The authors concluded that the continued use of hexazinone at recommended lable rates does not appear to be a threat to the environment. In another study, maximum concentration of hexazinone was 14 ppm in the stream that drained an 11.5-ha watershed treated with 2 kg/ha (Bouchard et al. 1985). Hexazinone residues of less than 3 ppm were detected in stream discharge for 1 year after application. The amount of hexazinone transported from the watershed in stream discharge represented only 2 to 3 percent of the amount initially applied.

Neary et al. (1986) found only 0.53 percent loss of hexazinone in streamflow of the applied herbicide in Georgia, U.S. Residues in streamflow peaked at 442 ppb in the first storm but declined rapidly and disappeared within 7 months. Total sediment yield increased by a factor of 2.5 because of increased runoff associated with site preparation using herbicide and salvage logging. However, sediment loading remained below those produced by mechanical techniques, and overall water quality changes were small and short-lived. Leitch and Flinn (1983) applied hexazinone at 2 kg/ha from a helicopter to a 46.4-ha catchment. Only 6 of 69 samples analyzed contained hexazinone, which was well below maximum allowable concentration of 600 ug/l for potable water.

2.1.1.3.8 Glyphosate

Glyphosate has limited use on rangelands. It is rapidly absorbed on soil (Torstensson 1985). Adsorption occurs through the phosphoric acid moiety that competes for binding sites with inorganic phosphates. Glyphospate is virtually immobile in soils. Inactivation of glyphosate through soil adsorption is important to rates required for uptake by target plant species.

Bronstand and Friestad (1985) concluded that regular use of glyphosate in agriculture or forestry allowed only very remote chances of contaminating the aquatic environment. The compound dissipates by microbial degradation, adsorption to sediments, and by photolysis.

2.1.1.4 Water Quality and Pesticide/Herbicide Residues

See Sections II-2.1.1.3 and Section IV-2.1.1.3.

2.1.1.5 Carbon Dioxide Balance in the Atmosphere

The fraction of incoming solar energy which is radiated back out to space from the earth as long-wave radiation is determined by the concentration of several atmospheric gases. The principal long-wave, energy-absorbing trace gases are carbon dioxide, methane, chlorofluorocarbons, and nitrous oxide, all of which are increasing in the troposhpere. CO₂ is the most abundant and is being added in the greatest quantity; it is expected to cause about 50 percent of global warming occurring in the next half century (Johnson et al. 1994).

There are two questions associated with CO₂ and the impacts of livestock grazing on the environment: a) the impact, direct or indirect, of livestock grazing on atmospheric CO₂ and b) the impact of atmospheric CO₂ on livestock grazing and, subsequently, on the environment. Neither of these questions has definitive answers in the literature. However, the postulation that rangeland-grazing animals could increase in numbers from current levels so that their contribution to atmospheric CO₂ would be significant compared to CO₂ generated by industrialized humans and their activities is untenable. The second question, the impact of atmospheric CO₂ on livestock grazing, is dealt with at length in recent literature and is more easily addressed.

Despite media obsessions with tropical moist forests, savannas actually dominate the southern continents, covering 65 percent of Africa, 60 percent of Australia, and 45 percent of South America (Stott 1994). In savanna ecosystems, the majority of the grass species involved belong to the C₄ group of photosynthetic plants. In such plants, the rate of photosynthesis continues to increase with the intensity of photosynthetically active radiation (PAR), instead of the carbon dioxide curve flattening off to a plateau, as is more normal in C₃ species. In C₃ plants the carbon dioxide is fixed initially as a three-carbon compound, phosphoglyceric acid; by contrast, in C₄ plants, the carbon dioxide is fixed as a four-carbon compound, oxaloacetic acid. C₄ plants are the most efficient photosynthetically where light conditions are maximized, so that their optimum temperature for CO₂ fixation lies between 30 and 45^o C, their photosynthetic rate under optimal conditions is between 40 and 80 mg CO₂ dm^-2 h^-1 (3 to 4 times that of C₃ species), and the light saturation is 100 percent. This means that they are ideally adapted to bright, fairly dry environments, precisely the conditions associated with open savanna grasslands of the seasonal tropics. However, there is some evidence that this competitive advantage in hot, dry regions may be significantly diminished as a consequence of predicted greenhouse effects, such as increased atmospheric CO₂, because C₄ species appear to show little photosynthetic response to elevated CO₂ (Woolhouse 1990). Nevertheless, dry-matter production may still be enhanced by an improvement of the plant-water status, through the reduction of stomatal aperture which usually accompanies the effects of high CO₂ concentrations (Squire 1990).

In the savannas, both natural fire, from lightning strike, and human-induced fire are often entirely integral to the maintenance of the ecosystem. In many savannas, it would be impossible to maintain the economic productivity of the grass stratum without the use of fire as a prescribed tool (Stott 1994). It is now recognized that the savanna form is generally governed by the intricate interplay of five key ecological factors, namely plant available moisture, plant available nutrients, fire, herbivory, and major anthropogenic events. Savanna ecosystems have themselves long been a significant contributor to the never-ending story of global environmental change. Even before the impact of humans, fire, through lightning strike, friction, and refraction, was a natural stress in savanna lands. Today, savanna fires contribute the largest percentage of all the CO₂ emitted to the atmosphere through biomass burning in the tropics. In the Brazilian cerrado, for example, the entire humid savanna is fired once every two years, while 75 percent of the humid savannas of Africa burn annually. Thus, 85 percent of the CO₂ emitted from tropical Africa is derived from its savannas. Many ecologists and conservationists thus regard biomass burning in the savannas as a major contributor to global warming. But in reality the issue is much more complex, because the savannas have always burned regardless of human activity. What we need to know is the exact nature of the recent increase in burning over the long-term historical levels. Would this recent increase really be significant at all if it were not being added to the historical contributions over the past 150 years or so by the industrialized countries of the North (Stott 1994)? In many ways our future management of the savanna landscapes of Australia, South East Asia, India, Africa, and Latin America will indicate all too starkly whether we are succeeding in our human response to global environmental change (Stott 1994).

Deforestation and grazing influence species composition, primary productivity, and organic matter decomposition, thereby altering the liberation and sequestering of CO₂ (Archer 1994). Changes in land cover and ecosystem processes may further influence climate by altering surface energy flux and biophysical properties (albedo, temperature, evapotranspiration, air circulation, etc.) and by changing levels of particulate input (e.g., dust) in the atmosphere (Graetz 1991).

With an eightfold increase in population since the eighteenth century to the currently more than 5 billion people, development and industrialization have increased atmospheric level of greenhouse gases. The population now appropriates 40 percent of all organic matter fixed by photosynthesis per year and consumes the equivalent of 2 tons of coal per year. It is no great surprise that CO₂ levels have increased dramatically in the last century and particularly in the last three decades (Byers 1994a). The CO₂ concentration of the atmosphere has increased from as low as 265 ppm only 125 years ago to about 350 ppm at present (Mayeux et al. 1991), an increase of almost 30 percent. However, atmospheric levels of CO₂ are not expected to double before the year 2025 and may not occur within the next century (Trabalka et al. 1985). The biggest contributors to elevated CO₂ are emissions from factories, vehicles, and power plants, primarily within the industrialized countries. CO₂ release from deforestation has contributed as much as half of the atmospheric increase since 1800 and is responsible for at least 20 percent of current emissions (Byers 1994a). The slash-and-burn conversion technique immediately releases both tree and litter CO₂ to the atmosphere. Dead organic matter in the soil holds two times as much total CO₂ as there is in the atmosphere and when released during deforestation is a major source for increasing atmospheric levels of carbon. Deforestation without slash and burn, for example, in the U.S. where the forest heritage was converted into the Midwest breadbasket and logs converted to lumber for construction, does not contribute as much CO₂ to the atmosphere. Using this latter technique leaves the majority of the CO₂ in the lumber. However, past deforestation, especially in Central and South America, has been done to accommodate crops and pastures for livestock grazing (Cross 1994). Thus, there is a relationship, albeit indirect, between livestock grazing and elevated atmospheric CO₂.

Possible effects of the predicted increase can be characterized as a) the direct effects on rangeland vegetation caused by a carbon enriched atmosphere, or b) indirect effects on vegetation caused by global warming as a consequence of increased CO₂ and other greenhouse gases. There is substantial evidence to support the hypothesis of some researchers that there have already been dramatic changes in rangeland vegetation over parts of the world that can be related, at least in part, to elevated atmospheric CO₂ (Byers 1994a).

Mayeux et al. (1991) and Scifres and Hamilton (1993) describe changes in the vegetation of the southern Great Plains and Southwest of the U.S. as essentially an increase in the distribution and density of naturally occurring shrubs from widely separated sites into open grassland stands. Many woody species were a part of the migration from limited sites to become established across the broad range of soils and topography. Archer et al. (1988) describes the mechanisms of establishment of pioneer woody plants in a grassland environment and subsequent changes through juvenile to mature plants and associated components of mottes in south Texas. Major woody species that have increased and/or thickened as a part of this phenomenon include mesquites (Prosopis spp.), creosote bush (Larrea divaricata), and junipers (Juniperus spp.), many shrubby composites including sagebrushes (Artemisia), rabbitbrushes (Chrysothamnus), snakeweeds (Gutierrezia), and a host of less geographically important plants.

The commonly accepted combination of causes for the increase of woody plants on open grasslands (overgrazing, fire suppression, propagule dissemination, climate change) are acknowledged by Mayeux et al. (1991). However, these researchers also believe that the change cannot be explained by these factors only and contend that no compelling evidence exists which substantiates that any single one or combination of them is responsible. Woody plant increases have been observed across extreme heterogeneity of the entire American West as well as in other parts of the world, including Argentina, Mexico, South America, Australia, Africa, and India. Interestingly, in the low veld of Rhodesia, shrubs increased at the expense of perennial grasses more in the absence of domestic livestock than when grazed lightly or moderately (Kelly and Walker 1976). Areas from which grazing has been excluded since before significant woody plant invasion, including an exclosure since 1915 on the Jornada Experimental Range in southern New Mexico, have been encroached by woody vegetation similar to adjacent unprotected sites.

Future changes in plant community structure have been predicted as consequences of continued increases in atmospheric CO₂ in the next century (Bazzaz and Garbut 1988, Overdieck 1986). Mayeux et al. (1991) considered whether the effects of global increases in atmospheric CO₂ levels are already evident as shifts from open grassland to shrubs. There is evidence that higher levels of CO₂ may have already influenced productivity of forests based on studies of annual growth rings of pine species (LaMarche et al. 1984) where tree growth rates were found to exceed those expected from climatic trends, but to be consistent with global trend in CO₂. Historic increases in the yields of other plants, including several crops such as wheat (Gifford 1979, Byers 1994a) and soybeans (Allen et al. 1987), may be at least partly due to increased availability of CO₂.

CO₂ enhancement of the atmosphere will have its greatest effect under hot, dry climates, where leaf temperatures and stomatal closure lead to elevated O₂/CO₂ ratios at the site of CO₂ fixation. In C₄ plants, however, there already exists a CO₂-concentration mechanism, the C₄ cycle, which is not susceptible to O₂ competition, and in which the carboxylating enzyme is protected by a locally elevated concentration of CO₂. It is not surprising, therefore, that C₄ plants show little response to elevated CO₂. This means that their competitive advantage may be reduced under increased atmospheric CO₂, having significant effects on both the composition and the productivity of the grass stratum in the tropical savannas. But, again, there is much uncertainty about this simple prediction because the single variable advantage may be offset or transcended by significant changes in the other environmental factors concerned with dry-matter production, particularly temperature and precipitation (Squire 1990).

Increasing the CO₂ concentration of the atmosphere in which plants grow increases carbon assimilation rates and has favorable effects on other physiological processes of all functional groups of plants (Pearcy and Bjorkman 1983). Idso et al. (1987) found that the literature indicated plant growth would be increased by approximately 30 percent by a 300 ppm increase in atmospheric CO₂. When combined with the 3^o C predicted increase in temperature from the greenhouse effect, the plant growth enhancement factor rises from 1.30 to 1.56. These same researchers found that even higher growth enhancement could be achieved if the non-CO₂ trace gas effect is equally as strong. However, they also found that atmospheric enrichment tends to reduce plant growth at relatively cold temperatures and that predicting the ultimate biospheric consequences of the earth's atmospheric CO₂ concentration may be more complex than originally anticipated.

A study reported by Riechers and Strain (1988) using blue grama (Bouteloua gracilis) bears out that much more specific information will be required to predict changes. Blue grama, a C₄ grass species, may benefit more from CO₂ enrichment than would be predicted based on the response of other C₄ species studied in other experiments, even though the relatively large enhancement in growth is far less than the increases seen for many C₃ species. Therefore, even though some C₄ species, such as blue grama, may be differently enhanced by elevated CO₂, they will still be at a comparative competitive disadvantage compared to C₃ species.

Rochefort and Woodward (1992) modeled the response of global family diversity to global environmental change, including climate and a doubling of atmospheric CO₂. The model assumes that three primary mechanisms define diversity: the capacity to survive the absolute minimum temperature of a site, the ability to complete the life cycle in a given time length and warmth of the growing season, and the capacity to expand leaves in a defined regime of precipitation and vegetation transpiration. Global temperatures are assumed to rise 3^o C and global precipitation increase 10 percent. The direct effects of CO₂ on vegetation transpiration are also included. The addition of CO₂ in the atmosphere along with a 10 percent increase in precipitation appears to counteract the negative effect of increasing temperature on vegetation diversity (global warming was found to be deleterious for global diversity because the increased rate of vegetation transpiration, at the higher temperatures, was not offset by increased rates of precipitation). As a consequence, one-third of the world's floristic regions might increase their diversity or at least maintain a similar value of diversity to the present. The model demonstrated that CO₂ taken as a factor by itself can have a significant effect on global family diversity, and the authors recommend strongly that CO₂ effects should be included when modeling global climate change.

While increasing CO₂ appears to improve water use efficiency of all plants, experimental evidence suggests that effects on other physiological processes are more strongly expressed in plants that possess the C₃ photosynthetic pathway than in those with C₄ photosynthesis (Bhattacharya 1993). Carbon assimilation rates vary widely with species and environmental conditions but are thought to be higher in C₄ than C₃ species at current levels. Mayeux et al. (1991) compared net carbon assimilation rates of little bluestem (Schizachyrium scoparium), a C₄ native perennial bunchgrass widely distributed across the Great Plains of the U.S., and the C₃ woody invader honey mesquite (Prosopis glandulosa) at increasing CO₂ levels. Their studies indicate that photosynthetic rates of the grass are about 20 percent higher at CO₂ levels characteristic of 150 years ago. However, net photosynthetic rate of the shrub equals that of the grass at today's atmospheric CO₂ level, 350 ppm, and will exceed that of the grass as ambient CO₂ level continues to increase. Thus, the C₃ mesquite has realized a greater relative advantage than the C₄ grass as CO₂ increased over the last 150 years, if increasing photosynthetic capacity improves performance at the whole-plant or higher levels.

The work by Mayeux et al. (1991) has profound implications for predicting relative composition changes of rangeland vegetation between C₃ plants (woody and cool-season species) versus C₄ plants (warm-season perennial grasses). However, there may be offsetting conditions not addressed in their experiments. CO₂ can interact in complex ways to alter plant growth, making it difficult to predict the future productivity and composition of plant communities from the results of studies solely examining the effect of CO₂ on plant performance (Coleman and Bazzaz 1992).

Even with the lack of definitive information on interactions between elevated atmospheric CO₂ and other environmental factors, such as temperature and precipitation, the evidence is strong that rangeland vegetation will be affected in both composition and productivity potential by the greenhouse effect, including CO₂. The result could be a decrease in the relative proportions of C₄ to C₃ plants based on the increased competitive advantage to C₃ plants from increasing CO₂. In the studies by Mayeux et al. (1991), there was also a dramatic increase in the response of C₃ herbs to increasing CO₂ at subambient levels representative of the past 150 years. Oats (Avena sativa) and wild mustard (Brassica kaber) were grown in a continuous CO₂ gradient from 150 ppm to current ambient, about 350 ppm. Over the range of increasing CO₂ from 250 to 350 ppm, representative of the change in the last 150 years, net carbon assimilation increased by over 50 percent. Leaf area and oven-dry weight of top growth increased by the same extent or more, indicating that historical changes in CO₂ may have profound effects on the growth of C₃ herbs. C₃ woody plants respond to elevated CO₂ to the same or greater extent as C₃ herbs (Tolley and Strain 1987). Conifers exhibit a pronounced growth increase, suggesting that CO₂ may have already played a role in recent increases in the extent of the piñon pine-Juniper type and the abundance of a number of Juniperus species throughout North America.

Mayeux et al. (1991) hypothesize that favorable effects of increasing CO₂ that apply to all functional groups of plants, especially improved water-use efficiency and amelioration of stress, suggest that overall productivity of rangelands will increase. However, increased productivity will probably continue to be reflected in increased biomass of less desirable C₃ weeds and woody vegetation, as opposed to C₄ warm-season perennial grasses, where the two functional groups occur together. This has very profound ramifications for rangeland management and for the use of grazing animals. The hypothesis that historical increases in atmospheric CO₂ conferred competitive superiority upon C₃ weeds and shrubs and rendered them inherently better adapted to today's rangeland than C₄ grasses implies that, by definition, current climax is characterized by shrub dominance. This seems true not only for the present but increasingly so in the future as CO₂ levels continue to climb. Efforts to define range condition and trend relative to a historical species composition in which shrubs were poorly represented seem unrealistic in light of both the hypothesized CO₂ effect and the well-documented success of woody plants over recent decades. For information on methane emissions see Section IV-2.4.1.1.

2.1.1.6. Vegetation Ground Cover

Cover of vegetation has been identified as one of the major mediators of raindrop impact and subsequent infiltration rates, runoff levels, and sediment loading (Blackburn et al. 1982, Thurow 1991, Blackburn et al. 1992, Bari et al. 1993). Vegetation structure also impacts light interception, competitive interactions between plant species, and habitat structure for animals and insects. Vegetation ground cover can be characterized in terms of plant morphology, interception/stemflow characteristic, and contribution to the litter component of the ecosystem. Grazing directly impacts vegetation cover through defoliation and trampling, indirectly affecting competitive interactions and subsequent species composition.

One of the primary impacts of vegetation cover is interception of rainfall. The amount and intensity of precipitation reaching the soil surface may be influenced by grazing to a degree that grazing alters the amount and type of vegetation on a site. Short-term effects of grazing in the form of utilization reduces cover as standing crop declines, exposing more of the soil surface to rainfall impact. Thurow et al. (1987) found that sodgrasses intercept and return 10.8 percent of rainfall back into the atmosphere before impacting the soil. Bunchgrasses intercepted 18.1 percent of the precipitation in this Quercus virginiana savanna and only 54 percent of the annual percipitation reached mineral soil beneath oak mottes as through fall or stemflow. Due to concentrating effects of stemflow, soil near the bases of the live oak trees received about 222 percent of annual precipitation, whereas areas more than 100 mm from the trunk received only 50.6 percent of annual rainfall. Angerer (1991) in a subtropical savanna parkland in South Texas, found that approximately 7 percent of the annual rainfall was intercepted in mottes, with approximately 15 percent being concentrated at the base of woody species in the mottes via stemflow.

Increasing grazing pressure decreases interception loss, reduces litter turnover, increases bareground, and increases raindrop velocity. This results in slower infiltration rates, increased runoff amount and flow rates, and greater sediment loading. If heavy utilization of forage persists for long periods of time and vegetation composition shifts morphology of the plant community to taller, woody species, hydrology of the system can be altered significantly.

Cattle have been documented to be the primary distributor of Prosopis seeds in South Texas landscapes and once established, the Prosopis tree becomes a primary focal point for seed rain via avian defecation while birds are perched in the tree (Archer et al. 1988). Once established, understory woody species begin to concentrate water and nutrients and reducing light for the predominant C₄ grassland species. In this case, vegetation structural change due to grazing is a slow change in vegetation composition induced by successional forces. Woodlands emerging from heavily grazed situations or landscapes with fire impact removal can intercept 20 to 32 percent of annual rainfall (Ahmad-Shah and Rieley 1989).

McGinty et al. (1991) has called for identification of hydrologic thresholds for vegetation cover to reduce the likelihood of accelerated water, soil, and nutrient loss from an ecosystem. Generally, this approach is based on standing crop of herbaceous plants and cover of woody species. The suggested threshold levels of standing crop varies from 300 kg/ha for sodgrass cover in temperate savannas of Texas to 1000 to 1440 kg/ha in temperate grasslands of Pakistan (Bari et al. 1993). For more information see Section III-2.1.1.6 and Section V-2.2.1.3.

2.1.1.7 Botantical Composition

See Section II-2.1.1.7 and 2.2.1.4.

2.1.1.8. Upperground and Root Biomass of Grasslands

2.1.1.8.1. Aboveground Biomass

Grazing-induced modifications of competitive interactions are eventually expressed at the population level through the modification of basal area and tiller demography of individual plants (Briske 1991, Milchunas and Lauenroth 1993). Reduction in basal area of bunchgrasses is the initial and predominate response contributing to their decline (Butler and Briske 1988). Decreases in plant basal area is most likely a consequence of the fragmentation of individual plants into smaller units. Consequently, plant density may remain constant or even increase while basal area declines. Smaller plants have higher tiller density per plant basal area, maintaining a constant density per unit land area. Plants reach a threshold where bunchgrasses cannot compensate if continued excessive grazing occurs, resulting in a decline in tiller density per unit area. A decline in tiller density results in a decline in aboveground biomass and subsequent composition shift in the population. Consequently, the more grazing-resistant species within the community utilize a greater proportion of the available resources (Caldwell et al. 1987). Populations composed entirely of plants with reduced basal area may be jeopardized by the inability to effectively compete with populations of less severely grazed species and increased susceptibility to extreme abiotic conditions (e.g., drought or temperature extremes) (Briske 1991). Persistence of small plants may be deceiving when grazing pressure results in a high proportion of photosynthesizing foliage being removed with each bite and catastrophic drought follows resulting in elimination of a population from a community.

Grazing impact on aboveground biomass is a function of chosen herbivore mix, numbers of animals for each kind of herbivore, and plant growth patterns relative to temporal and spatial weather events. Milchunas and Lauenroth (1993) reviewed vegetation response to grazing in a hierarchical context. Species composition change was viewed as a relatively rapid response while changes in annual net primary production being an intermediate response. Soil nutrient loss was considered a slow, variable response. They concluded that aboveground net primary production does not necessarily change when species composition changes, and can increase or decrease depending on the replacement species, life-history traits, and the manner in which the continuing grazing pressure or stress affects water and light resources and nutrient cycling rates. Increases in short-term nutrient cycling rates may increase primary production over time periods of years to decades while decreasing large, recalcitrant nutrient pools.

Recently, yields of subtropical bunchgrass species were shown to be only slightly reduced (less than 10 percent) with end of season utilization levels up to 60 percent. However, yield was negatively impacted in an exponential manner with end of season use levels from 60 to 90 percent (Stuth, personal communication).

2.1.1.8.2 Root Biomass

Root growth is driven by a combination of soil conditions and allocation of carbohydrate from the shoot and energy provided by photosynthesis (Briske 1991). In most growth situations, shoot growth occurs at the expense of root growth when soil moisture conditions are adequate for plant growth. However, when soil moisture is unable to sustain plant growth, as soil dries out, carbohydrate supply to the roots, and consequently root growth, increases until mechanical impairment of root penetration and water conductivity by the soil became restrictive (Gilmanov 1977). Defoliation by grazing animals commonly results in cessation of root growth, followed by a period of rapid tillering, but the net effect of defoliation on pasture production depends on the relative growth rate of the sward, as well as intensity and frequency of grazing (Schuster 1964, Hilbert et al. 1981) Suppression of root growth is generally proportional to the intensity and frequency of defoliation. Longevity of roots varies between species and generally is reduced when shoots are defoliated (Troughton 1981). However, the significance of root longevity to plant survival and competition has not been clearly ascertained.

A single defoliation which removes 50 percent or more of shoot volume can retard root growth for 6 to 18 days when soil moisture conditions are adequate for plant growth (Crider 1955). Defoliations of a single removal of 80 and 90 percent of shoot volume have been observed to stop root growth for 12 to 17 days, respectively. Multiple defoliations slow root growth even more. For example, 70 percent initial removal followed by three clippings resulted in cessation of root growth for over 33 days. This phenomena occurs within hours of defoliation (Hodgkinson and Baas Becking 1977). In a heavily grazed community of caucasion bluestem (Bothriochloa caucasica), tiller density was higher but root mass was lower than plants in a lightly grazed pasture (Christiansen and Svejcar 1988). Generally, grazing regimes which promote increased tillering density in a sward will benefit root production (Briske 1991).

Root growth cessation due to defoliation affects both lateral and vertical development of root systems (Smoliak et al. 1972) as well as detrimentally influences root initiation, diameter, branching, and total production (Carman and Briske 1982, Richards 1984). These reductive responses in root growth collectively reduce the total absorptive surfaces and soil volume explored for water and nutrients. Nutrient absorption in perennial, temperate grasses parallels the growth responses following defoliation (Briske 1991).

Milchunas and Lauenroth (1993) conducted an extensive review of studies on long-term impact of grazing on root biomass. There were no clear relationship between long-term grazing and root mass in field scale experiments. Soil carbon and soil nitrogen did not exhibit any distinct effects of long-term grazing in this global review of the literature. The primary impact was noted in species composition. These authors felt that "the use of species-based criteria in management may lead to erroneous conclusions about the long-term ability of grazing lands to sustain productivity when changes in species composition are minor and changes in soil nutrient levels are negative and large, or may lead to an overestimate of the impact of grazing when opposite occurs." They stressed that assessment of grazing impact in ecosystems must be multiscaled and that the variable assessed must be scaled to the question being asked.

2.1.1.9 Soil Fertility

See Section IV-2.3.1.1, Section IV-2.8.1.1, IV-2.5.1.1.

2.1.1.10 Overstory/Understory Relationships

See Section IV-2.2.1.3.

2.1.1.11 Habitiat Composition and Biodiversity

See Section II-2.8.1.1 and 2.8.1.2.

2.1.2 Indirect Indicators

2.1.2.1 Cropping Systems

See Section V-2.3.2.3.

2.1.2.2 Animal-Mechanical Power

See Section V-2.3.2.1.

2.1.2.3 Population

See topics on urban and rural development in Section V-2.1.2.5.

2.1.2.4 Land Use and Property Rights

See Section V-2.1.2.4.

2.1.2.5 Feeds and Feeding

See Section IV-2.1.2.2.

2.1.2.6 Micro- and Macroeconomic Indicators

2.1.2.6.1 Budgeting

Budgeting is an internationally recognized method to evaluate current livestock-raising practices. The procedure is divided into whole-farm budgets which include a variety of enterprises, enterprise budgeting for a particular activity such as cow/calf production, and partial budgeting to evaluate the impact on an enterprise from changes in practices. Livestock producers in areas where there are severe climatic conditions have limited options for improving their productivity compared with those in tropical areas. In all cases, much depends on size of operation, management skills, and economics of input/output relationships. In virtually all cases small-scale producers, especially those operating on a very small scale, are severely constrained in technology adoption by lack of facilities.

Following are some examples at the micro (farm level) of some well-known production practices which can have a significant impact on improving productivity and income. The examples are for selected sites, but the concepts hold for all areas of the world.

2.1.2.6.1.1 Parasite Control

Animal health analysis can be carried out in a variety of ways, but two of the most important are modeling and partial budgeting. The latter method is especially important to determine if a practice should be recommended and to determine the extent of benefits. Partial budgeting is a very powerful tool that can provide the relatively quick answers often demanded by veterinarians or planners. An example from Florida illustrates calculations of benefits from parasite control.

Cattle parasites are divided into two classifications, internal and external. Examples of internal parasites now presented are liver flukes, intestinal worms, and grubs. The external ones are horn flies and lice. Myriad others that plague developing countries, such as tick-born diseases, could be evaluated. However, sufficient examples can now explain the importance of this type analysis when considering improvements in range and pasture management and also show how evaluations for projects can be carried out relatively quickly and inexpensively when severe data restrictions exist.

A summary of the way in which treatments have been determined for the five parasites is provided in Table III.3. A table of this type is the first step, as it provides a complete description of the procedure and the cost.

Multiplication of the cost per treatment times the recommended number of treatments for each type of parasite yields the annual cost for each of the three classes of cattle analyzed: mature cows and bulls, replacement heifers, and calves. Recommended treatments and associated costs for intestinal worms vary with animals on a high, medium, and low plane of nutrition. Three control methods for horn flies are evaluated. Costs include a charge for additional labor (when applicable) as well as materials and equipment.

Benefits from parasite control are very difficult to quantify due to a multitude of physiological interactions and various geographical/climatic conditions. Consequently, well-documented data and the benefits from control of the various parasites are seldom available. But this problem can be overcome by using estimates from production and veterinary specialists. The estimates can then be divided into two levels, low and high, with the expectation that the actual benefits will fall within those limits. In all cases, efforts should be conservative in the estimation procedure.

Potential benefits in the example provided are derived from a reduction in death loss, a reduction in weight loss, increased calf crop, or a combination of them. It is estimated, for example, that mature cow death loss in areas infested with liver flukes can be reduced by 2 to 4 percent with the recommended treatments (Table III.4). That is, liver flukes account for about 2 to 4 percent death loss in mature cows, and treatment is assumed to stop these losses. Flukes are also estimated to cause cows to lose about 18 to 45 kg, a loss that is realized when they are sold as culls. Additionally, mature cows with liver flukes probably have a 6 to 12 percent reduction in calf crop if they are not treated.

An economic benefit, in dollars per cow unit (one cow and one calf, that is, 1 AU), is calculated for each of the three benefit categories by multiplying the various production loss estimates by the appropriate values. For example, multiplying the high level in death loss (4 percent) from flukes for mature cows times the projected value of $333.00 per cow (a 432 kg cow at $0.77 per kg) gives $13.32 as the annual loss (Table III.5). The economic impact of weight loss is calculated by multiplying the weight loss times value per kilo times culling rate. The latter is assumed to be 10 percent for cows and 3 percent for heifers. The high estimation of weight loss per cow of 45 kg times $0.77 per kg equals $35.00 per cow, which, times 10 percent, is $3.50.

Reproductive loss is calculated by multiplying the average value per calf of $235.00 (188 kg calf at $1.25 per kg) times the reproductive loss in percent, adjusted for a herd composition of 90 percent for mature cows and 10 percent replacements. As an example, the high estimate of losses from flukes in mature cows is 12 percent times $235.00, which equals $28.20. That coefficient times 0.90 equals $25.38.

Analyses of net benefits from production practices typically have the annual cost subtracted from the annual benefits to calculate the net benefit per unit. However, because the benefits, although reasonable and defensible, are almost always going to be speculative, this approach should not be taken. Another reason not to take this approach is the wide variability among cattle operations. Thus, benefits are better considered as potential losses in income. In other words, the calculated benefits can be thought of as opportunity losses or potential income that is foregone from not carrying out the practice.

The various dollar benefits for the three categories--mature animals, heifers, and calves--for each of the three areas of potential loss are summarized in Table III.6 to arrive at an annual potential income loss. This is done for both the high and low levels. Then the sum of the potential income loss is divided by the cost to arrive at a benefit risk/cost ratio. High ratios indicate more potential benefit or, alternatively, the potential risk if the practice is not followed.

The term benefit risk/cost ratio is used because the more common term, benefit/cost ratio, cannot be properly applied here because it is used in project analysis and carries with it the connotation that discounted (long-term) net benefits are divided by discounted costs. In addition, standard use in benefit/cost analyses assumes some initial major capital investment is made.

The high, or largest, probable ratio for controlling flukes in mature cows and heifers in fluke areas is 16.4 to 1, while the low side is 8.0 to 1. This means that even the most conservative estimate in this analysis places risk at 8 times more than costs. Thus, considering the high incidence of fluke infestation, this practice will usually return much more than the costs. Alternatively, the benefits are high relative to cost.

The results of treating cows and heifers for intestinal worms definitively show that the poorer the nutrition level, the more benefit from treating for worms. Thus, in years when feed is short, analysts would recommend that special care be taken to carry out a good worming program. Likewise, cattle on poor-quality pasture should be systematically wormed to obtain the highest possible benefits.

2.1.2.6.1.2 Economics of Anabolics

Anabolics, also called implants, metabolic modifiers, or growth promotants, are primarily used in cattle and to a much lesser extent in sheep; only minimal use of anabolics is made in goats. This example uses beef cattle for all examples, but the analytical method can be used for other types of livestock. Anabolics, such as IMC's Ralgro, are employed in a variety of situations, including suckling calves not intended for breeding purposes (most anabolics are not cleared for use in breeding stock), growing cattle, (after weaning but prior to the final fattening phase), and during that latter finishing period. Finishing cattle, (those being fattened for slaughter) can be further divided into two types, those on pasture and those in a confinement feedlot.

There are a number of potential physical benefits that depend on the type of cattle and system. Table III.7 contains a summary of typical benefits found around the world. There are, of course, considerable response differences depending on cattle condition, feed quality and quantity, and climatic variations, just to mention the major factors. However, the range provided, such as 0.088 to 0.135 kg of additional gain in suckling calves, is expected under usual conditions. The averages provided, such as the 0.111 kg of additional gain in suckling calves, typical of results from hundreds of on-farm trials, are used to calculate the economic benefits.

One principal benefit attributable to anabolics in all classes of cattle is additional gain, which may be measured in at least three ways: additional average daily gain as calculated in kilos, additional average daily gain as calculated in percentage terms, and total additional daily gain.

Partial analysis is the proper approach to measure benefits from anabolics. The first examples include six methods in suckling calves; the choice of method depends on its relevance to an individual producer (Table III.8). However, in all cases, the analysis centers on additional gain obtained from using an implant.

The first method is a calculation of return per implant using the typical values of 10 kg of additional gain per implanted suckling calf. The analysis indicates that although a typical cost per implant is about $1.40, return would be about $12.50 if calves were priced at $1.25 per kg, thus providing a net return to management of $11.10. That translates, as shown in method two, to $8.93 for every dollar invested--and only during a 90-day period. If, as analyzed in method three, the evaluation is placed on a 1-year basis, the return is $9.47. An appropriate question for the cow/calf producer then is: Do I have alternative investments on my farm or ranch that will yield me more than $9.47 annually for each $1.00 invested. If so, then those investments should have priority over implants.

Another method that can be used to evaluate the financial soundness of implants is to calculate the additional gain required to pay back the cost. The results in method four indicate that if calves were priced at $1.25 per kg, and the cost were $1.40 per implant, then only 1.12 kilos of additional gain would be required to break even.

Return per hectare is the fifth method provided in Table III.8. If carrying capacity were 1 cow on 2 hectares (0.50 cows per ha), a 70 percent calf crop obtained, and a net return per implant of $11.10 (as calculated earlier), then the net return per hectare would be $3.89 per implant. Calves can be implanted the first month after birth, so suckling calves could also be reimplanted prior to weaning. The net result would be a return of $7.78 per hectare.

Another approach is to calculate the return per hectare for each $1.00 invested. Results, shown as method six, indicate that $8.93 would be gained, which is the same as in method twp because the analysis is on a constant-unit basis. Similarly, the return per dollar invested would also be the same if placed on a brood cow basis.

The type of analysis for growing and pasture-finishing cattle is the same as for suckling calves. Computations are not made, but results are provided in Table III.9.

Table III.3. Cattle parasite treatment determination, recommended treatments, cost per treatment, and annual cost per cow unit, Florida, U.S.

Type of Parasite			Treatment Determination	Recommended Treatment
				Mature	Replacement	Calves
Internal
	Liver flukes		Flukes-only area	1/yr	1/yr	-
	Intestinal wormsa		High nutrition	0/yr	1/yr	0
			Medium nutrition	1/yr	2/yr	0
			Low nutrition	2/yr	2/yr	1
	Grubs		All Florida	1/yr	1/yr	-
External
	Horn fliesb
		Sprayc	50 or more flies/animal	6/yr	6/yr	-
		Dust bagsd	50 or more flies/animal	8mo./yr	8mo./yr	-
		Ear tags	50 or more flies/animal	2/an/yr	2/an/yr	-
	Licee		Presence of lice	2/yr	2/yr	-
Internal
	Liver flukes		Fluke-only areas	2.50	1.80	-
	Intestinal wormsa		High nutrition	0.00	1.80	-
			Medium nutrition	2.50	1.80	-
			Low nutrition	2.50	1.80	1.20
	Grubs		All Florida	0.60	0.50	-
External
	Horn fliesb
		Sprayc	50 or more flies/animal	0.29	0.29	-
		Dust bagsd	50 or more flies/animal	1.24	1.24	-
		Ear tags	50 or more flies/animal	1.50	1.50	-
	Licee		Presence of lice	0.32	0.32	-
Internal
	Liver flukes		Fluke-only areas	2.50	1.80	-
	Intestinal wormsa		High nutrition	0.00	1.80	-
			Medium nutrition	2.50	3.60	-
			Low nutrition	5.00	3.60	1.20
	Grubs		All Florida	0.60	0.50	-
External
	Horn fliesb
		Sprayc	50 or more flies/animal	1.74	1.74
		Dust bagsd	50 or more flies/animal	1.24	1.24
		Ear tags	50 or more flies/animal	3.00	3.00
	Licee		Presence of lice	0.64	0.64

Source: Simpson, 1988.

^a Fluke treatments also control most intestinal worms. No additional labor charge assumed.

^b Other methods such as pour-ons or back rubbers are also used.

^c Sprayer costs $500 and lasts 5 years = $100/year and with 1,000 cows = 0.10 cow/year = 0.02 per treatment. Labor cost is 250 for 1,000 cows per treatment and 4 treatments in addition to 2 regular workings = $1,000 for 1,000 cows = $1.00 per cow per year = $0.17 per treatment. Materials cost is $0.10 per cow per treatment.

^d Dust bags. One half kg of dust per head per season. The kits cost $26.50 and have 5.7 kg of dust. Additional dust costs $17.00 for 23 kg. The cost of the kit and dust is $1.04 per head. Also, cost of $0.20 per cow per year is for fence and all associated labor.

^e Has no labor charge as cattle can be sprayed along with other activities.

Type of parasite			Matures	Replacements		Calves
			High	Low	High	Low	High	Low
			Net Reduction Death Loss, Percent
Internal
	Liver flukes		4	2	4	2	0	0
	Intestinal Worms
		High nutrition	0	0	0	0	0	0
		Medium nutrition	2	0	2	0	0	0
		Low nutrition	5	2	5	2	2	0
	Grubs		0	0	0	0	0	0
External
	Horn flies		0	0	0	0	0	0
	Lice		0	0	0	0	0	0
			Weight Loss, Kilos
Internal
	Liver flukes		45	18	45	18	5	0
	Intestinal Worms
		High nutrition	9	0	9	0	5	0
		Medium nutrition	18	7	18	7	9	5
		Low nutrition	27	14	27	14	23	11
	Grubs		9	5	9	5	0	0
External
	Horn flies		23	9	23	9	14	5
	Lice		14	5	14	5	9	5
			Reproductive Loss, Calf Crop, Percent
Internal
	Liver flukes		12	6	12	6	-	-
	Intestinal Worms
		High nutrition	0	0	0	0	-	-
		Medium nutrition	6	3	6	3	-	-
		Low nutrition	12	6	12	6	-	-
	Grubs		2	0	2	0	-	-
External
	Horn flies		8	4	8	4	-	-
	Lice		5	2	5	2	-	-

Source: Simpson, 1988.

Type of parasite			Maturesa		Replacementsb		Calvesc
			High	Low	High	Low	High	Low
			Dollars per cow unit
			Death Lossd
Internal
	Liver flukes		13.32	6.66	3.31	1.66	0	0
	Intestinal Worms
		High nutrition	0	0	0	0	0	0
		Medium nutrition	6.66	0	1.66	0	0	0
		Low nutrition	16.65	6.66	4.14	1.66	3.53	0
	Grubs		0	0	0	0	0	0
External
	Horn flies		0	0	0	0	0	0
	Lice		0	0	0	0	0	0
			Weight Losse
Internal
	Liver flukes		3.50	1.40	1.35	0.54	3.42	0
	Intestinal Worms
		High nutrition	0.70	0	0.27	0	3.42	0
		Medium nutrition	1.40	0.53	0.54	0.20	6.84	3.42
		Low nutrition	2.10	1.05	0.81	0.41	17.10	8.55
	Grubs		0.70	0.35	0.27	0.13	0	0
External
	Horn flies		1.75	0.70	0.68	0.27	10.26	3.42
	Lice		1.05	0.35	0.41	0.13	6.84	3.42
			Reproductive Lossf
Internal
	Liver flukes		25.38	12.69	2.82	1.41	3.42
	Intestinal Worms
		High nutrition	0	0	0	0	-	-
		Medium nutrition	12.35	6.35	1.41	0.71	-	-
		Low nutrition	25.38	12.69	2.82	1.41	-	-
	Grubs		4.23	0	0.47	0	-	-
External
	Horn flies		16.92	8.46	1.88	.94	-	-
	Lice		10.58	4.23	1.18	.47	-	-

Source: Simpson, 1988.

Type of Parasite			Annual costa	Annual potential income lossb		Benefit risk/cost ratio
				High	Low	High	Low
			Dollars per cow unit
Internal
	Liver flukes		3.04	49.68	24.36	16.4:1	8.0:1
	Intestinal worms
		High nutrition	0.41	4.39	0.00	10.7:1	--
		Medium nutrition	4.36	30.86	11.21	7.1:1	2.6:1
		Low nutrition	6.98	72.53	32.43	10.4:1	4.6:1
Flukes and wormsc
	High nutrition		6.08	54.07	24.36	8.9:1	4.0:1
	Medium nutrition		6.08	80.54	35.57	13.2:1	5.9:1
	Low nutrition		6.08	112.21	56.79	20.1:1	9.3:1
	Grubs		0.75	5.67	0.48	7.6:1	0.6:1
External
	Horn flies
		Spray	2.23	31.49	13.79	14.1:1	6.2:1
		Dust bags	1.59	31.49	13.79	19.8:1	8.7:1
		Ear tags	3.84	31.49	13.79	8.2:1	3.6:1
	Lice		0.82	20.06	8.60	24.5:1	10.5:1

Source: Simpson, 1988.

^a All costs are in $/cow unit. A cow unit is 1.05 mature animals (to account for bulls), 0.23 replacement heifers (1- and 2-year heifers) and 0.75 calves. For example, taking data from Table III.1, liver flukes only are $2.50 times 1.05 = $2.63 plus $1.80 times 0.23 = $0.41 totaling $3.04.

^b Benefits are from Table III.5 by summing death, weight, and reproductive loss for matures, replacements, and calves. For example, $13.32 + $3.31 + $3.50 + $1.35 + $25.38 + $2.82 = $49.68 as the high annual potential income loss from liver flukes.

^c Control of worms as well as flukes is obtained by using some commercial flukacides.

Benefit		Suckling Calves	Growing Cattle	Pasture-Finishing Cattle
Typical weight
	Range	Birth - 200	200 - 360	360 - 550
	Average	100	280	455
Additional average daily gain (kg)
	Range	.088 - .135	.111 - .178	.133 - .222
	Average	.111	.155	.178
Additional average daily gain (pct)
	Range	10 - 14	12 - 15	15 - 20
	Average	12	14	16
Additional total gain (kg)
	Range	8 - 12	10 - 16	15 - 20
	Average	10	12	15
Additional total gain (pct)
	Range	4.0 - 6.0	2.8 - 4.4	2.7 - 3.6
	Average	5.0	3.3	3.1

Source: Simpson and Conrad 1992.

Item	Parametera
Method 1 Return per Implant
Cost
1 implant ($) 1/500 implant gun @ $20 each ($) 1 minute @ $4.00/hr ($) Misc. expenses ($)	1.25 0.04 0.07 0.04
Total	1.40
Return
10 kg @ $1.25 per kg ($)	12.50
Net return to management ($)	13.00
Method 2 Return per $1.00 Invested
Return per implant period (10 kg @ $1.25 per kg) ($)	12.50
Cost per implanted animal (See method 1) ($)	1.40
Return per $1.00 invested (Return divided by cost) ($)	8.93
Method 3 Annual Return per $1.00 Invested
Implant period (days)	90 275
365 days minus implant period (days)	.75
Remaining period as fraction of year	8
Opportunity value on money (pct)	6
Equivalent opportunity value, remaining period of year	8.93
Return per $1.00 invested, end of 90 days ($)	9.47
Return per $1.00 invested, end of year ($8.93 times 1.06) ($)
Method 4 Additional Gain Required to Pay Implant Cost
Total cost ($)	1.40
Value per Kilo ($)	1.20
Kilos required per 90 day period (cost divided by value per kg) (kg)	1.12
Method 5 Return per Hectare
Cows per hectare (no)	0.50
Calf crop (pct)	70
Net Return per implant per calf (See Method 1) ($)	11.10
Return per hectare per implant (Return per implant per calf times calves per hectare) ($)	3.89
Implants per calf per year (no)	2
Annual net additional return per hectare from implant ($)	7.78
Method 6 Annual Return per Hectare per $1.00 Invested
Calves per hectare (no)	.35
Gross return per calf per implant ($)	12.50
Number of implants (no)	2
Total gross return per hectare ($)	8.75
Additional cost (cost per implant times number of implants times calves per hectare) ($) ($1.40 × 2) (0.35)	.98
Return per $1.00 invested ($) (Gross return divided by additional cost)	8.93

Source: Simpson and Conrad 1992.

^a U.S. $ in all cases.

Class of Cattle	One Implant	$1.00 Invested	Return per Annum per $1.00 Invested	Hectare	Additional Gain to Breakeven
	U.S. $				kg
Suckling calves	13.00	8.93	9.47	7.78	1.12
Growing cattle	13.00	10.29	10.91	52.00	1.17
Pasture finishing	13.60	10.71	11.35	54.40	1.40

Source: Simpson and Conrad 1992.

Most countries derive the majority of cow milk from dairy cattle that obtain a substantial portion of their nutrients from pastures. In some countries, such as much of China, and particularly among their seminomads, virtually all nutrients come from rangeland grazing. In other areas, such as tropical and semitropical Latin America, dairy cattle derive most of their nutrients from improved forages, with some mixed feed (concentrate) supplemented at milking. A wide range of technologies is employed, and size of operation varies considerably (Simpson and Wilcox 1982). In almost all cases, substantial increase in productivity and net income is possible from changes in the operation.

Following is an example from the humid areas of Mexico, which describes how economics can be used to evaluate the impacts on productivity, cost, and income. The results are shown in U.S. dollars to facilitate international comparisons.

The farm has 22 head of cattle, of which 11 are actually in lactation (Table III.10). The current system, also called the "basic model," reveals cows in lactation produce 5.8 kg per head per day averaged over a 3- to 5-month lactation period. There are 1.4 cull cows and, 0.2 heifers aged 1 to 2 years and 5.0 calves sold annually.

The economic analysis shows that their direct production cost, which is essentially cash costs, is a negative $0.05 per kg produced. The reason is that this particular operation, like many small operations in developing countries, has almost no purchased inputs. Thus, sales of calves and breeding animals, which are subtracted from costs to calculate cost of milk produced, are greater than cash outlay. In the example shown, income from culls and calves equals $2,359 ($736 plus $1,623), while basic production costs are only $1,360 (Table III.11).

A dairy-industry improvement project might focus on management and expanded productivity. A model titled "improved system" is in the column next to the basic model. In this way, changes in parameters can easily be seen. Focusing back on Table III.10, a strategy calculated is expanding stocking rate, resulting in 33 head of cattle rather than 22. Improved management and nutrition result in an increase in calf crop, the number of calves per year expands, calving internal decreases, etc. Days in lactation decrease, but production per cow per year increases from 5.8 kg to 8.0 kg.

The expanded production of milk and animals sold is the result of feeding concentrate to lactating cows, increasing minerals and molasses fed, using more fertilizer on pasture, and paying more for contracted labor. As a result of expanded purchased inputs, production cost becomes a positive $0.03 per kg. However, net income per year above direct production costs increases from $5,926 to $8,456. The important factor is that substantial improvement can be made in productivity and net income if producers of this type adopt some technologies and expand input use. The problem is that changes like this require much more of a business orientation, a change which many smaller, more subsistence-oriented producers are not willing to make.

Table III.10. Analysis in Second Country Currency Dollars

ITEM			UNITS	BASIC MODEL	IMPROVED SYSTEM
CURRENT (ACTUAL) COW INVENTORY
	COWS IN LACTATION		HEAD	11	16
	DRY COWS		HEAD	2	3
	HEIFERS (>2 YEARS)		HEAD	4	6
	HEIFERS (1-2 YEARS)		HEAD	4	6
	BULLS		HEAD	1	2
		TOTAL	HEAD	22	33
INVENTORY BASED ON CALCULATIONS
	COWS IN LACTATION		HEAD	11	16
	DRY COWS		HEAD	2	3
	HEIFERS (>2 YEARS)		HEAD	1	2
	HEIFERS (1-2 YEARS)		HEAD	1	2
	BULLS		HEAD	1	2
		TOTAL	HEAD	16	25
PERCENT ON PASTURE			PCT	80	80
PASTURE			HA	9	9
STOCKING RATE			HEAD/HA	1.4	2.2
OTHER LAND			HA	15	15
TOTAL LAND			HA	24	24
DEATHS AND LOSSES
	WEANED CALVES		PCT	9	5
	HEIFERS FOR REPRODUCTION		PCT	0	0
CALF CROP (BIRTHS)			PCT	64	70
NUMBER OF CALVES PER YEAR			HEAD	7	11.2
CALVING INTERVAL			MONTHS	16.3	15.6
FIRST CALVING			MONTHS	30	24
REPLACEMENT AGE			YEARS	8	7
REPLACEMENT RATE			PCT	12.5	14.3
NUMBER OF REPLACEMENTS			HEAD	1	2
COES ARTIFICIALLY INSEMENATED			PCT	0	0
FORAGE/ANIMAL/DAY (APART FROM PASTURE)
	LACTATING COWS AND BULLS		KG	0	0
	DRY COWS		KG	0	0
	HEIFERS (>2 YEARS)		KG	0	0
	HEIFERS (1-2 YEARS)		KG	0	0
	MALE CALVES		KG	0	0
	FEMALE CALVES		KG	0	0
		TOTAL PER YEAR	KG	0	0
PURCHASED FORAGE OF TOTAL			PCT	0	0
TOTAL FORAGE PURCHASED			KG	0	0
COST/KG OF PURCHASED FORAGE			$	0	0
CONCENTRATE
	PER ANIMAL PER DAY
		LACTATING COWS AND BULLS	KG	0	2
		DRY COWS	KG	0	0
		HEIFERS (>2 YEARS)	KG	0	0
		BULLS	KG	0	0
		MALE AND FEMALE CALVES	KG	0	0
	FEEDING PERIOD
		LACTATING COWS AND BULLS	DAYS	315	305
		DRY COWS	DAYS	50	60
		HEIFERS (>2 YEARS)	DAYS	0	0
		HEIFERS (1-2 YEARS)	DAYS	0	0
		BULLS	DAYS	0	0
		MALE AND FEMALE CALVES	DAYS	0	0
	TOTAL PURCHASED/YEAR		KG	0	9760
	COST/KG		$	0.323	0.323
MILK
	PRODUCTION/COW/YEAR		KG	5.8	8
	PROPORTION SOLD		PCT	95	95
	LACTATION PERIOD		DAYS	315	305
	AMOUNT PRODUCED		KG	20,097	39,040
	PRICE/KG		$	0.258	0.0258
	KG MILK/KG CONC/LACT COW		KG	0	4
SALT
	PURCHASED PER MONTH		KG	40	62
	FED/ADULT/DAY		GRAMS	82	82
	COST/KG		$	0.097	0.097
MINERALS
	PURCHASED PER MONTH		KG	8	23
	FED/ADULT/DAY		GRAMOS	16	30
	COST/KG		$	0.258	0.29
MOLASSES
	PURCHASE PER MONTH		KG	1500	2000
	COST/KG		$	0.016	0.016
OTHER FEEDSTUFFS PER MONTH
	#1 PURCHASED		KG	500	800
	#1 COST/KG		$	0.019	0.016
	#2 PURCHASED		KG	0	0
	#2 COST/KG		$	0	0
MANURE
	MANURE/ANIMAL/YEAR		TONS	2.76	2.76
	TOTAL/YEAR		TONS	45	70
	SALE OR USE VALUE/KG		$	0	0
FERTILIZER
	AMOUNT PURCHASED		KG/YEAR	2000	4000
PRICE/KG			$	0.065	0.065
LABOR (NUMBER OF PERSONS)
	CONTRACTED
		FULL TIME	PERSONS	0	0
		DAY LABOR	DAYS/YEAR	20	30
	FOREMAN OR SUPERVISOR		PERSONS	0	0
	FAMILY LABOR		PERSONS	2	2
LABOR (INCLUDING BENEFITS)
	CONTRACTED
		FULL TIME	$	0	193.55
		DAY LABOR	$	6.45	6.45
	FOREMAN OR SUPERVISOR		$	0	0
	FAMILY LABOR		$	193.55	193.55
SALE WEIGHT OF ANIMALS
	CULL COWS		KG/HEAD	400	510
	BULLS		KG/HEAD	450	600
	HEIFERS (>2 YEARS)		KG/HEAD	350	450
	HEIFERS (1-2 YEARS)		KG/HEAD	200	300
	MALE AND FEMALE CALVES		KG/HEAD	200	250
SALE PRICE/KG
	CULL COWS		$	1.13	1.13
	BULLS		$	1.16	1.16
	HEIFERS (>2 YEARS)		$	1.29	1.29
	HEIFES (1-2 YEARS)		$	1.29	1.29
	MALE AND FEMALE CALVES		$	1.61	1.61
SALE PRICE/HEAD
	CULL COWS		$	451.61	575.81
	BULLS		$	522.58	696.77
	HEIFERS (>2 YEARS)		$	451.61	580.65
	HEIFERS (1-2 YEARS)		$	258.06	387.10
	MALE AND FEMALE CALVES		$	322.58	403.23
ANNUAL SALES
	CULL COWS (CALCULATED)
		NUMBER	HEAD	1.4	2.3
		NUMBER	KG	550	1666
	CULL COWS (ACTUAL)
		NUMBER	HEAD	1.4	2.3
		TOTAL WEIGHT	KG	560	1173
	BULLS
		NUMBER	HEAD	0.1	0.2
		TOTAL WEIGHT	KG	45	120
	HEIFERS (>2 YEARS)
		NUMBER	HEAD	0	0
		PESO TOTAL	KG	0	0
	HEIFERS (1-2 YEARS)
		NUMBER	HEAD	0.2	0.4
		TOTAL WEIGHT	KG	40	120
	MALE AND FEMALE CALVES
		NUMBER	HEAD	5	8.4
		TOTAL WEIGHT	KG	1006	2089
INVESTMENTS
	LAND		$	61935	61935
	CONSTRUC. & BUILDINGS		$	4194	4935
	FENCES		$	6452	6452
	EQUIPMENT & TOOLS		$	1613	2194
	HORSES		$	645	645
DEPRECIATION
	CONSTRUC. & BUILDINGS		YEARS	20	20
	FENCES		YEARS	8	8
	EQUIPMENT & TOOLS		YEARS	5	5
	HORSES		YEARS	10	10
OPPORTUNITY COST
	LAND		$	483.87	903.23
	ALL OTHERS		$	387.10	1290.32
VALUE OF ANIMALS (PER HEAD)
	COWS		$	483.87	903.23
	BULLS		$	387.10	1290.32
REPAIRS AND MAINTENANCE (PER YEAR)
	CONSTRUC. & BUILDINGS		$	38.71	45.16
	EQUIPMENT		$	38.71	41.94
	FENCES		$	193.55	225.81
ANIMAL HEALTH (PER MONTH)
VETERINARIAN SERVICES PRODUCTS ARTIFICIAL INSEMINATION			$	0	32.26
			$	25.81	29.03
			$	0	0
OTHER COSTS (PER MONTH)
	ELECTRICITY		$	3.23	4.84
	GASOLINE & OIL		$	16.13	33.87
	OTHERS AND MISC.		$	6.45	12.9
	MARKETING COSTS		$	0	0
OTHER EXPENSES (PER YEAR)
	TAXES		$	83.87	83.87
	INSURANCE		$	0	0

Source: Simpson, 1995

ITEM				BASE COST OR INCOME DOLLARS	PCT %	IMPROVED COST OR INCOME DOLLARS	PCT %
INVESTMENT
	LAND			61935	73	61935	61
	CONSTRUCTIONS & BUILDINGS			4194	5	4935	5
	FENCES			6452	8	6452	6
	EQUIPMENT AND TOOLS			1613	2	2194	2
	HORSES			645	1	645	1
	BREEDING ANIMALS			9516	11	25548	25
		TOTAL		84355	100	101710	100
DIRECT COSTS PER YEAR
	PURCHASED FORAGE			0	0	0	0
	FERTILIZER			129	9.49	258	4.23
	CONCENTRATE			0	0	3148	51.59
	SALT			46	3.42	72	1.18
	MINERALS			25	1.82	80	1.31
	MOLASSES			24	1.78	32	0.53
	OTHER FEEDSTUFFS			116	8.54	186	3.04
	REPAIRS AND MAINTENANCE			271	19.93	313	5.13
	VETERINARIAN SERVICES			0	0	387	6.34
	VETERINARY PRODUCTS			310	22.77	348	5.71
	ARTIFICIAL INSEMINATION			0	0	0	0
	ELECTRICITY			39	2.85	58	0.95
	GASOLINE AND OIL			194	14.23	406	6.66
	OTHERS, MISCELLANEOUS			77	5.69	155	2.54
	MARKETING COSTS			0	0	0	0
	LABOR
		DAY & PERMANENT		129	9.49	658	10.78
		FOREMAN/ADMINISTRATION		0	0	0	0
			TOTAL DIRECT COSTS	1360	100	6102	100
OTHER COSTS PER YEAR
	CAPITAL COSTS
		LAND		12387	53.66	12387	45.89
		CONSTRUCTIONS, EQUIPMENT		2581	11.18	2845	10.54
		BREEDING STOCK		1903	8.24	5110	18.93
		OPERATING CAPITAL		82	0.35	366	1.36
			SUBTOTAL	16953	73.44	20708	76.71
	OWNERSHIP COSTS
		DEPRECIATION		1403	6.08	1556	5.77
		TAXES		84	0.36	84	0.31
		INSURANCE		0	0	0	0
			SUBTOTAL	1487	6.44	1640	6.08
	FAMILY LABOR			4645	20.12	4645	17.21
			TOTAL OTHER COSTS	23085	100	26994	100
TOTAL, ALL COSTS				24445	33096

Table III.11. Analysis in Second Country Currency Dollars

ITEM			BASE COST OR INCOME DOLLARS	PCT %	IMPROVED COST OR INCOME DOLLARS	PCT %
ANNUAL INCOME
	MILK		4,927	67.62	9571	65.74
	CULL ANIMAL		736	10.10	1619	11.12
	CALVES		1,623	22.28	3369	23.14
	MANURE		0	0.00	0	0.00
		TOTAL	7,286	100.00	1,4558	100.00
INCOME PER YEAR ABOVE:
	DIRECT PRODUCTION COSTS		5,926		8,456
DIRECT PRODUCTION
	COSTS AND TAXES		5,842		8,372
DIRECT PRODUCTION
	COSTS, TAXES AND DEPRECIATION		4,439		6,816
DIRECT PRODUCTION
	COSTS, TAXES, DEPRECIATION AND FAMILY LABOR		(206)		2170
ANNUAL NET INCOME PER LACTATING COW ABOVE:
	DIRECT PRODUCTION
		COSTS	539		528
	DIRECT PRODUCTION
		COSTS, TAXES, DEPRECIATION AND FAMILY LABOR	(19)		523
COST PER KG OF MILK PRODUCED
	DIRECT PRODUCTION
		COSTS	-0.05		0.03
	DIRECT PRODUCTION
		COSTS, TAXES, DEPRECIATION AND FAMILY LABOR	0.24		0.18
ALL COSTS			1.16		0.76
DIRECT COSTS AS A PERCENT OF ALL COSTS			5.6		18.4

Source: Simpson, 1995