| Methods |
|
Report
Sustainable Disturbance Levels for Military Training on Gypsic Soils (Phase III)
PROJECT LOCATION: Holloman Air Force Base (HAFB)
TPOC: Jeanne Dye1
Submitted by: Jeffrey E. Herrick 2 and Jayne Belnap 3
149CES/CEV, 550 Tabosa Ave., Holloman AFB, NM 88330-8458,
Jeanne.Dye@holloman.af.mil, 505-572-3931. Original TPOC = M. Hildegard
(Hildy) Reiser.2USDA-ARS Jornada Experimental Range, Box 30003, MSC 3JER, NMSU,
Las Cruces, NM 88003, jherrick@nmsu.edu, 505-646-5194.3USGS, 2290 S. Resource Blvd., Moab, UT 84532, jayne_belnap@usgs.gov,435-719-2333.
Organization
This report is organized into six sections. The Introduction provides a general background and discussion of the problem. Protocols used are outlined under Methods, and the data are summarized in tabular form and discussed under Results and Discussion. The Synthesis section integrates the indicators in order to assess the resistance and resilience (recovery potential) for each of three soil types and three disturbance classes. A preliminary conceptual model is the basis for management and monitoring decisions addressed in the Recommendations sections. A Web-based Decision Tool/Model was developed based on the results of this study.
Table of Contents
1. Introduction
1.1. Soils
1.2. Vegetation
2. Methods
2.1. Study Sites
2.1.1. Dune margin
2.1.2. Transition
2.1.3. Outcrop
2.2. Experimental design
2.3. Treatments
2.3.1. Control
2.3.2. Horse
2.3.3. Infantry
2.3.4. Track
2.4. Chronology
2.5. Measurements
2.5.1. Vegetation indicators
2.5.2. Microbiotic crust indicators
2.5.2.1. Lichen cover
2.5.2.2. Chloropyll content
2.5.2.3. Nitrogen fixation
2.5.3. Hydrology and water erosion indicators
2.5.3.1. Water infiltration
2.5.3.2. Field soil stability
2.5.3.3. Pocket penetrometer
2.5.3.4. Surface roughness
2.5.4. Wind erosion indicators
2.5.4.1. Wind boxes
2.5.4.2. Torvane
2.6. Calculations
2.6.1. Single disturbance
2.6.1.1. Resistance
2.6.1.2. Relative resillience (6 yr)
2.6.1.3. Absolute resilience or percent of control (2003)
2.6.2. Double disturbance
2.6.2.1. Resistance (2X)
2.6.2.2. Resistance (+1)
2.6.2.3. Relative resilience (6 yr)
2.6.2.4. Absolute resilience or percent of control (2003)
2.7. Statistical analysis
2.7.1. Treatment effects
3. Results and Discussion
3.1. Overview
3.2. Specific indicators
3.2.1. Vegetation indicators
3.2.1.1. Resistance and resilience: site comparison.
3.2.1.2. Resistance and resilience: treatment comparison.
3.2.1.3. Double disturbance.
3.2.2. Microbiotic crust indicators.
3.2.2.1. Resistance and resilience: site comparison.
3.2.2.2. Resistance and resilience: treatment comparison.
3.2.2.3. Double disturbance.
3.2.3. Hydrology and water erosion indicators.
3.2.3.1. Resistance and resilience: site comparison.
3.2.3.2. Resistance and resilience: treatment comparison.
3.2.3.3. Double disturbance.
3.2.4. Wind erosion indicators.
3.2.4.1. Resistance and resilience: site comparison.
3.2.4.2. Resistance and resilience: treatment comparison
3.2.4.3. Double disturbance
4. Synthesis
4.1. Soil and site stability
4.1.1. Water erosion
4.1.2. Wind erosion
4.2. Hydrologic function
4.3. Biotic integrity
4.4. Site comparisons
4.5. Treatment comparisons
5. Recommendations
6. Web-based DecisionTool/Model
Literature Cited
1. Introduction
Large areas of military land throughout the western United States have been degraded by military and nonmilitary uses, including livestock grazing, ground defense training and vehicle maneuvers. These activities disturb the soil surface and have direct and indirect effects on vegetation. The net impacts of these activities vary depending on the resistance and resilience of the ecosystem.
1.1. Soils. Surface disturbances by humans, livestock and vehicles have multiple effects on soils (Webb and Wilshire, 1983; Thurow, 1991). Few studies, however, have addressed the effects of disturbance on gypsiferous soils covered by microbiotic crusts. These soils cover most of Holloman AFB. Previous studies in arid regions have illustrated the critical importance of soil biological crusts for surface stabilization and erosion control (Belnap, 1995; Belnap and Gillette, 1998), and the importance of crust biological nitrogen fixation for maintaining soil fertility in some arid systems (Belnap, 1994). The effects of soil biological crusts and their disturbance vary in different parts of the world (Webb and Wilshire, 1983; Eldridge and Greene, 1994). A study similar to the one initiated in Phase I of this study was implemented on nongypsiferous soils at a site in the Chihuahuan Desert near Holloman AFB. Observations and preliminary results of this study support the general concluSion that interactions between soils and types of disturbance dramatically affect impacts on ecosystem function in areas dominated by microbiotic crusts. Ancillary studies using a rainfall simulator and a wind tunnel have shown that climate and, especially, the temporal distribution of rainfall plus the frequency and intensity of high-wind events must be considered together with soil and disturbance type. Results of both studies are now being analyzed and prepared for publication (Herrick et al., unpub. data).
Top of page1.2. Vegetation. Direct effects of disturbance on vegetation are widely recognized. Trampling and vehicle traffic tend to have a negative effect on woody vegetation, while herbaceous vegetation can be positively or negatively affected. Indirect effects include changes in soil water and nutrient availability (Webb and Wilshire, 1983; Thurow, 1991).
The objectives of this study were to: (1) identify gypsic soil(s) most suitable for military training exercises; (2) evaluate the impact of different disturbance types associated with military training activities on the resistance and resilience of a suite of vegetation, microbiotic crust, hydrology and water erosion, and wind erosion indicators; and (3) develop a preliminary conceptual model and set of recommendations for where and when military training is most suitable in these semi-arid landscapes.
The information gained from this study is already being applied to assessment (Pellant et al., 2005), monitoring (Herrick and Whitford, 1995; Herrick et al., 2005), and remediation of degraded land (Herrick et al., 1997; Herrick et al., 2006; Rango et al., In Press).
Top of page2.1. Study sites. Three study sites were chosen to represent the dominant soil types at Holloman AFB. A full replication of treatments was performed at these three sites. All sites were flat (<2% slope). All three sites may have been extensively grazed by livestock prior to 1942 when DOD acquired the land on which Holloman AFB is now located. The three sites were:
2.1.1. Dune margin. This site had nearly 100% gypsum soil and was located near the eastern border of the White Sands Missile Range, less than 200 m east (downwind) of active dunes. Highly dispersed low-statured shrubs dominate the site. Little, if any, anthropogenic disturbance has occurred here for at least 35 years.
2.1.2. Transition. This site was located in an area with gypsum intergrading with silty, silica-based material. It is covered by a patchy (1-20 m diameter) mosaic of dense perennial grasses with interspersed shrubs. The site has had very limited anthropogenic disturbance.
2.1.3. Outcrop. This site was located on partially indurated gypsum, exposed at and near the surface. The site is dominated by dispersed sub-shrubs, perennial grasses and forbs. There is some evidence of military training in the area, including several communications wires and foxholes. These would have been generated in the past 30 years; the mission of the base was changed from missile research to supporting tactical fighter aircraft in the early 1970's. Plots were located to avoid clear signs of historic disturbance.
2.2. Experimental design. A randomized, complete block design, with six blocks and four treatments, was applied at each of the three sites. The 24 individual plots at each site measured 8x30 m.
2.3. Treatments. The treatments listed below were first applied in 1997. In 2000, the treatments were re-applied to half of each treatment plot (4x30 m).
2.3.1. Control. These plots were left untreated for the duration of the study.
2.3.2. Horse. For this treatment, horses were guided by their riders back and forth across each plot until it appeared that every point had been disturbed (either having been stepped on or had soil directly kicked onto it) at least once. The number of times the horses passed across each plot was standardized across blocks and sites.
2.3.3. Infantry. For this treatment, booted soldiers crossed across each treated plot a fixed number of times in the same manner as described in the horse treatment above.
2.3.4. Track. For this treatment, a WWII-vintage jeep was driven back and forth across the plots at a speed of 5 to 10 kph. Each point on each treated plot was run across twice by two wheels for a total of four wheel passes. Tire inflation pressure was kept standardized at approximately 15 psi.
Top of page
Oct-Nov, 1997:Baseline measurements completed for selected variables. Treatments applied to each 8x30 m plot. Post-treatment measurements completed.
Oct-Nov, 1998: One-year, post-treatment measurements completed.
Oct-Nov, 2000: Three-year, post-treatment measurements completed. Treatments re-applied to half of each treatment plot (4x30 m). Post-re-treatment measurements completed.
Oct-Nov, 2001: Four-year, post-treatment measurements completed. One-year, post-re-treatment measurements completed.
Oct-Nov, 2003: Six-year, post-treatment measurements completed. Three-year, post-re-treatment measurements completed.
Several additional erosion bridge and wind erosion measurements were also completed (see Results section below) during other data collection periods. Laboratory measurements and erosion bridge photograph analyses were completed during non-field months.
Top of page2.5. Measurements. At each 8x30 m plot, a single, 30 m long transect was set up along one side of the plot. With the exception of the dust collector (BSNE boxes) all measurements were taken along these transects. For each measurement year, transects were moved 50 cm towards the center of each plot to ensure that previously-disturbed areas were not remeasured or resampled. This approach increased the amount of variability in vegetation measurements from year to year, but minimized the impact of previous measurement disturbance. All of the following measurements were taken pre-treatment (1997), immediately post treatment (early November 1997), at the same time of year in 1998, 2000, 2001, and 2003, unless otherwise stated.
2.5.1. Vegetation indicators. Plant cover was determined using the continuous line-intercept method. Percentage total canopy, shrub, and grass cover was calculated from canopy length measurements along each 30 m transect. (While the Jornada has since replaced this method with the more rapid, accurate, and repeatable line-point method, the continuous line-intercept method was used for the duration of this study in order to maintain a consistent dataset.)
2.5.2. Microbiotic crust indicators.
2.5.2.1. Lichen cover.Lichen cover was recorded every 25 cm along each 30 m transect using the line-point intercept method. Percentage lichen cover per plot was calculated from the resulting data. No immediate post-treatment measurements were made for lichen cover since it was impossible to do so accurately until after the dust had been redistributed by rainfall, exposing still-intact crust fragments.
Top of page
2.5.2.2. Chlorophyll content. This indicator of cyanobacterial biomass was estimated using absorbance techniques. Ten dry samples were collected per plot. Chlorophyll was extracted from samples with dimethylsulfoxide (DMSO) in the dark for 40 minutes at 65?C. Samples were then centrifuged. Absorption spectra were measured at 666 nm in a Hewlett-Packard diode array spectrophotometer after calibration with a DMSO blank.
2.5.2.3. Nitrogen fixation potential based on nitrogenase activity. Fifteen dry samples per plot were collected. Samples were placed in clear, gas-tight tubes; the entire crustal surface was wetted equally with distilled water and then injected with enough acetylene to create a 10% acetylene atmosphere. After injection, samples were incubated for 4 hours at 26?C in a chamber lighted with Chromo50 (5000 K) and cool white fluorescent bulbs. Subsamples (0.25 ml) of the head space within the tubes were then analyzed for acetylene and ethylene content on a Carle FID gas chromatograph equipped with an 8 foot, 8% NaCl on alumina column, using helium as the carrier gas (30 ml/min). Results are reported as gas chromatographic units and are not convertible to kg/ha of N without calibration by N.
2.5.3. Hydrology and water erosion indicators.
2.5.3.1. Water infiltration. Water infiltration rates were measured using a 12.5 cm ring infiltrometer (Bouwer, 1986) at five locations along the 30 m transect (at 3, 9, 15, 21, and 27 m) in each plot. The soil surface was pre-wetted to a minimum depth of 4 cm and an aluminum irrigation pipe ring inserted to a depth of 3 cm. The ring was filled to a depth of 3 cm and an inverted, one-liter soda bottle with an air tube was used to maintain water at a constant depth within the ring. At least 2.5 cm of water was allowed to infiltrate into the soil before measurements were begun, ensuring infiltration rates were at or near steady state. Once the water level in the inverted bottle had dropped a minimum of 5 cm, times and water depth were recorded and the rate of water movement calculated in mm/h. Due to time constraints, infiltration was not measured pre-disturbance in 1997.
Top of page
Figure 1. Single ring infiltrometer.
2.5.3.2. Field soil stability. At 2 m intervals along each 30 m transect, soil surface stability was measured using a field stability test (Herrick et al., 2001). A small aluminum sampling scoop was used to gently lift 6-8 mm diameter, 3-4 mm thick soil fragments from the soil surface at each sampling point. Each sample was ranked to generate qualitative stability index values (from 1 to 6: lowest to highest stability) (Herrick et al., 2001).
Figure 2. Field soil stability kit.
2.5.3.3. Pocket penetrometer. A SoilTest pocket penetrometer (Bradford, 1986) was used to measure soil surface resistance to penetration. The flat-tipped end of the penetrometer was 6.5 mm in diameter. Surface resistance was measured to a depth of 6.5 mm. For post-disturbance measurements, a foot extension that was 25 mm wide was used in order to increase sensitivity to treatment differences. Numbers were converted to correct for the differences in foot diameters. Pocket penetrometer measurements were taken in 1998, 2000, 2001 and 2003 only.
2.5.3.4. Surface roughness.Using an erosion bridge method, soil surface roughness was calculated as the standard deviation of the heights of 24 pins placed in a 50 cm line at 2 cm spacing along the surface. Estimates were made at five permanent locations per plot immediately following treatment. These measurements were repeated in December 1997, May 1998, May 1999, October 2000, December 2000, June 2001, October 2001, and March 2004.
Figure 3.Erosion bridge.
2.5.4. Wind erosion indicators.
2.5.4.1. Wind boxes.Relative differences in soil detachment and transport by wind were estimated by anchoring a Big Spring Number Eight (BSNE) dust trap (Fryrear, 1986) on the soil surface at one end of each study plot, parallel to the long side facing either west (dune margin and transition sites) or south (outcrop site). Wind erosion was calculated from resulting samples as kg of sediment per m2 per month for each plot. Measurements were taken five times: posttreatment in 1997, March 1998, May 1998, May 1999, and October 2000.
Top of page
Figure 4. BSNE wind box.
2.5.4.2. Torvane. Crust strength was measured using a standard Torvane apparatus. Measurements were taken five times every 2 m along the 30 m transect of each plot. For most measurements, a 2.5 cm diameter disk was used. For the post-disturbance measurements, a 4.75 cm diameter disk was used to increase sensitivity to treatment differences. A factory-supplied conversion factor was used to correct for differences in disk diameter.
2.6.1.1. Resistance. Resistance to each treatment was calculated as a percentage of the control value. Some properties change immediately in response to disturbance, while disturbance effects on others are delayed. Consequently, we calculated resistance as the smaller of the post-disturbance values (1997 through 2003).
2.6.1.2. Relative resilience (6 yr). This is the proportion of the function recovered 6 years after treatment as a percentage of function lost following treatment. We adjusted for natural variability using values from control plots at each site. Where this value is negative, the treatment continued to decline relative to the control.
2.6.1.3. Absolute resilience or percent of control (2003). This is simply the 2003 treatment value as a percentage of the 2003 control (the 2004 data was used for the Erosion Bridge). It reflects how far below potential the plot is 6 years post-disturbance.
2.6.2.1. Resistance (2X). This is the treatment value as a percentage of the control value. The smaller of the post-disturbance values was used to calculate resistance.
2.6.2.2. Resistance (+1). This is the ratio of the smallest double-disturbance plot value to the same year single-disturbance, post-disturbance value for the same indicator. This is the resistance of plots to additional disturbance. Plots that have not recovered may show higher resistance than some that had recovered.
2.6.2.3. Relative resilience (6 yr). This is the proportion of the function recovered 6 years after treatment as a percentage of function lost following treatment. We adjusted for natural variability using values from control plots at each site.
2.6.2.4. Absolute resilience or percent of control (2003). This is the 2003 treatment value as a percentage of the 2003 control (the 2004 data was used for the Erosion Bridge).
2.7.1. Treatment effects. An Analysis of Variance (ANOVA) was used to test for treatment effects during each year and univariate comparisons were made with the Control for each site. For those variables that did not meet the assumptions for ANOVA in any given year, treatment effects were tested using a Kolmogorov-Smirnov non-parametric test. In these cases, pairwise comparisons between treatments and controls were performed using Friedman's test. In all cases, the plot was always used as the experimental unit (n = 6 plots per site). Where more than one measurement per plot was made, the mean value of all measurements was used for analyses.
Top of page3.1. Overview. The results for each indicator are presented on a single page in a set of tables and figures (see Table 1 example). There is one table for each of the three sites. Each table includes treatment means and standard errors for each measurement date. Resistance and resilience values (Table 2) for each site are included, as are plots of treatment:control ratios by site and treatment through time. Resistance is generally defined as percent of pre-disturbance. Because some effects are delayed we use the minimum value for up to 6 years following disturbance. In order to account for natural annual variability we use control plots as the reference. Resilience is commonly defined as either percent recovery of what was lost over a particular period of time, or as percent of control at a particular point in time. It can also be defined as the rate of recovery (e.g., percent recovery/yr). This definition was not used here. This set of definitions is commonly referred to as "Engineering Resilience". Note that much of the recent ecological literature combines the concepts of resistance and resilience into a single definition of resilience, as the amount of stress a system can absorb before crossing a threshold. We have not used this definition because it is virtually impossible to calculate and extremely expensive to determine experimentally.
3.2.1. Vegetation indicators. All three treatments significantly reduced vegetative cover relative to the control. The reduction was due primarily to a loss of shrub cover which declined more than grass cover in response to all treatments. It took 4 years for canopy cover to recover at the dune margin site, four times longer than it took at the transition site. The rapid recovery at the transition site was due to grass regrowth: 1998 grass cover in the infantry and track plots actually exceeded cover in the control plots by up to 60%, though the differences were not significant (p>0.15). The transition site differs from the other two sites in that the grass community is dominated by Sporobolus airoides Torrey, while Sporobolus nealleyi Vasey is the dominant grass on most plots at the other two sites. Shrub recovery was slow at all sites, particularly in the track plots which were significantly below the control at all three sites 6 years posttreatment. Grass cover at the dune margin site and shrub cover at the outcrop site were highly variable among plots and treatments in 2003. Overall canopy cover declined at the dune margin and outcrop sites from 2001 to 2003, possibly in response to drought conditions.
3.2.1.1. Resistance and resilience: site comparison. Vegetation at the dune margin site had both low resistance and low resilience compared to the other two sites. Resistance was high at both the transition and outcrop sites. These site differences reflect differences in species composition. The transition and outcrop sites support much higher grass cover, which is both more resistant and resilient than shrub cover. Also, shrub cover at the dune margin site is dominated by fourwing saltbush (Atriplex canescens (Pursh) Nutt.) which is very brittle and, therefore, susceptible to trampling. It is also possible that there is increased competition from the rapidly recovering grasses, though additional work would be needed to test this hypothesis.
3.2.1.2. Resistance and resilience: treatment comparison. While there wasn't a highly significant difference between the different treatments, the track treatment had the most negative and persistent impacts on vegetation at all sites. Shrub cover was particularly affected by the track treatment, with significant reductions persisting for 6 years at the dune margin and transition sites. Horse and infantry disturbance had a significant impact on vegetation for 3 years at the dune margin and transition sites.
3.2.1.3. Double disturbance. The pattern for the second disturbance was similar to that of the first applied 3 years earlier. The dune margin site was most severely affected and the track treatment had the most negative effects.
Top of page
3.2.2. Microbiotic crust indicators. The viability of the microbiotic crust community is reflected in three indicators: lichen cover, chlorophyll content, and nitrogen fixation potential. Chlorophyll content is an indicator of the total biomass of photosynthetically active organisms at the soil surface, including both free-living and lichen-associated algae and cyanobacteria. Some of the lichen species fix nitrogen. The potential to fix nitrogen is correlated with nitrogenase activity measured in the laboratory. Results are reported as gas chromatographic units and are not convertible to kg/ha of N without calibration by N15, which was not done. No post-disturbance lichen results are reported because it was impossible to do so accurately until after the dust had been redistributed by rainfall, exposing still-intact crust fragments.
3.2.2.1. Resistance and resilience: site comparison. Vegetation at the dune margin site had both low resistance and low resilience compared to the other two sites. Resistance was high at both the transition and outcrop sites. These site differences reflect differences in species composition. The transition and outcrop sites support much higher grass cover, which is both more resistant and resilient than shrub cover. Also, shrub cover at the dune margin site is dominated by fourwing saltbush (Atriplex canescens (Pursh) Nutt.) which is very brittle and, therefore, susceptible to trampling. It is also possible that there is increased competition from the rapidly recovering grasses, though additional work would be needed to test this hypothesis.
Top of page3.2.2.2. Resistance and resilience: treatment comparison. While there wasn't a highly significant difference between the different treatments, the track treatment had the most negative and persistent impacts on vegetation at all sites. Shrub cover was particularly affected by the track treatment, with significant reductions persisting for 6 years at the dune margin and transition sites. Horse and infantry disturbance had a significant impact on vegetation for 3 years at the dune margin and transition sites.
3.2.2.3. Double disturbance. The pattern for the second disturbance was similar to that of the first applied 3 years earlier. The dune margin site was most severely affected and the track treatment had the most negative effects.
Top of page
3.2.3. Hydrology and water erosion indicators. The effects of the three disturbance types on site susceptibility to water runoff and erosion were quantified by four measurements: infiltration, soil stability, penetrometer and erosion bridge. Single ring infiltration measurements of water infiltration capacity reflect relative changes in the infiltration rate of water when the soil is saturated. Despite the fact this measurement can overestimate infiltration rates during rainstorms by a factor of 10 or more, it is a very useful indicator of changes in near-surface soil structure (e.g., destruction of soil macropores) that are closely related to infiltration under natural conditions. Field soil stability and pocket penetrometer resistance are also indicators of near-surface soil structure. Increases in pocket penetrometer values are correlated with compaction in the top 1 cm or less, reflecting a loss of potentially water-conducting pores. Reductions in soil stability values reflect a losss or weakening of bonds (usually organic matter) between soil particles. As these bonds are lost, the soil becomes more susceptible to both water erosion and physical crusting during rainfall events. Physical crusts are very dense and tend to have a platy structure that conducts water laterally instead of vertically. The fourth indicator, soil surface roughness, is calculated from erosion bridge data (see photos associated with the data). The erosion bridge is not sufficiently sensitive to detect soil loss rates as low as those occurring at these three sites during such a short study. It does, however, accurately reflect changes in surface roughness. Water moves more slowly across a rougher surface, so it has more time to infiltrate and less energy available for erosion. Rougher surfaces also slow near-surface wind speeds, reducing wind erosion (see next section).
In general, the treatments reduced water infiltration and slightly reduced soil stability. There was one notable exception to this pattern. The horse and infantry treatments appear to have increased infiltration capacity at the transition site immediately post-disturbance, although the differences were not statistically significant. By 1998, however, infiltration in both treatments had returned to control levels or dropped below them. There is much less lichen cover at the transition site than at the other two sites. Instead, it appears to have a surface crust stabilized by cyanobacteria. Destruction of this type of crust by infantry and horses could temporarily increase infiltration. After the first severe storm, the physical crust re-forms. While the density of living cyanobacteria may take several years to return to control levels (see "Chlorophyll"), polysaccharides and other organic matter previously generated by the cyanobacteria, or even the dead cyanobacteria themselves could easilly stabilize the new crust (see "Field Soil Stability"). Litter generated by the relatively high total plant cover could also contribute to the relatively rapid restabilization. The track treatment tended to compact rather than disturb the crust.
The infiltrometer and penetrometer data together support the hypothesis that, even in those few cases when the initial effect of the treatments may possibly facilitate water entry and seedling emergence, the formation of a physical crust rapidly negates any potential benefits. There was no evidence of any benefits of disturbance for the track treatment.
Top of page3.2.3.1. Resistance and resilience: site comparison. For infiltration rates, resistance to disturbance was relatively unaffected and resilience was low at the dune margin and outcrop sites. The transition site was relatively unaffected and may have even increased in response to horse and infantry treatments. Soil stability showed lower resistance to disturbance at the outcrop site. There was recovery in soil stability over 6 years at all sites, although a decline occurred between 2001 and 2003, matching that observed in vegetation cover. Soil compaction as measured by the pocket penetrometer was significantly higher than the control at the dune margin and outcrop sites by the time the first measurements were made in 2000.
3.2.3.2. Resistance and resilience: treatment comparison. There was no consistent difference in resistance and resilience of infiltration rates and surface soil stability between treatments. Both the penetrometer and erosion bridge measurements differed in their response to different treatments, however. The track treatment had a much more persistent effect on the penetrometer readings, particularly at the dune margin site. The horse treatment had the biggest impact upon erosion bridge measurements. The track treatment reduced surface roughness, whereas the horse treatment increased it, with the infantry treatment having little effect. Belnap and co-workers have reported similar effects for the Colorado Plateau where freeze-thaw processes interact with microbiotic crusts to create greater microrelief. There was a steady decline in surface roughness for the horse treatment plots through time, but erosion bridge levels did not reach control levels after 6 years at the dune margin or outcrop site. At the ddune margin site, infantry and track surface roughness remained below control levels and continued to decline from 2001 to 2003. These treatments were not significantly different from the controls at the transition site. None of the track treatments reached control levels for penetrometer readings.
Top of page3.2.3.3. Double disturbance. The double-disturbance data suggest that at the dune margin site, at least, penetrometer resistance may be initially reduced by horses due to the destruction of the biological crust; but, within a year, the reformation of the physical crust had already pushed the resistance past control levels. The pattern for infiltration was nearly identical to the first disturbance with a further reduction occurring across all sites and treatments, except for the transition site, infantry and track treatments, where there were non-significant increases relative to the single-disturbed plots. Field soil stability was again minimally affected. The horse treatment significantly reduced stability at all three sites. Sampling problems at the outcrop site probably contributed to variability in the data.
3.2.4. Wind erosion indicators. BSNE eolian sediment collection boxes ("wind boxes") placed parallel to the long side facing either west (dune margin and transition sites) or south (outcrop site) provided a direct indicator of sediment movement. The method was limited by the low level of replication and especially by plot size (8x30 m). The torvane test generated a less direct but more sensitive indicator of the shear strength of the soil surface, which is positively related to its resistance to wind erosion.
At both the dune margin and outcrop sites, all three treatments had increased sediment trapped by a factor of two or more following disturbance. Mean values in horse and track treated plots also exceeded control values at the transition site, but results were more variable and non-significant.
The torvane test showed a short-term reduction by all treatments at all sites followed by a rapid recovery. There was some indication that subsequent crust re-formation was actually increasing treatment values to exceed control values, especially in the track treatment at the dune margin and outcrop sites. While this is good for wind erosion resistance, it has potentially negative implications for seedling emergence and water infiltration.
Top of page3.2.4.1. Resistance and resilience: site comparison. Treatment impacts were clearly more severe at the dune margin site. Resistance was high and resilience was low at this site. Treatment effects were non-significant for wind boxes at the transition site where resistance was high. The combination of higher vegetative cover and finer textures at the transition site probably limited detachment and transport of soil particles. For the torvane, resistance was low at all three sites, but there was high resilience throughout. Recovery to control levels occurred at all sites for torvane.
3.2.4.2. Resistance and resilience: treatment comparison. Resistance to horse and track treatments was greater for wind box measurements at the dune margin and outcrop sites. The horse treatment also had the greatest effect at the dune margin site, and was statistically significant. The horse treatment tended to pulverize the surface, significantly increasing the wind erodibility of the soil surface. The track treatment also broke the surface but not to the same degree. Torvane demonstrated little in the way of treatment differences.
3.2.4.3. Double disturbance. The second disturbance had similar effects to the first with the horse treatment at the dune margin site and the infantry treatment at the outcrop site having the most noticeable effects. Again, the small sample size and even smaller plots (now 4x30 m) significantly limited our ability to interpret these data.
Top of page
4. Synthesis. The sustainability or health of rangeland ecosystems can be described in terms of three attributes: soil and site stability, hydrologic function and biotic integrity (Pellant et al., 2000). These three attributes are key to maintaining the capacity of ecosystems to support DOD land management objectives.
4.1.1. Water erosion. Both the dune margin and outcrop sites have naturally high resistance to water erosion. Slopes are low at both sites and infiltration capacity is high at the dune site due to coarse soil textures. Field soil stability values are inherently high due to high microbiotic crust cover. The lichens and cyanobacteria suffered minimal reductions following disturbance and recovered relatively quickly. Transition was the only site with significant evidence of overland flow. The control plots at this site also had the lowest and most variable (highest coefficient of variation) infiltration rates and the lowest soil stability, suggesting it is inherently more susceptible to water erosion. Fortunately, however, it was also relatively resistant and resilient to all three types of single disturbances.
4.1.2.Wind erosion. The dune margin site is inherently susceptible to wind erosion, and wind erosion was significantly increased at this site by all three treatments during all measurement periods and, at the outcrop site, by all treatments during at least one measurement period. Loss of vegetative canopy cover ensured that sediment movement remained high even after torvane and pocket penetrometer measurements showed restabilization of the soil surface. The transition site appears to be more resistant to wind erosion due to higher vegetative cover and finer soil texture.
Top of page4.2. Hydrologic function. The greatest threat to hydrologic function at all three sites is clearly vehicle traffic. Just two passes of a small jeep with extremely low tire inflation (15 psi) on dry soil reduced equilibrium infiltration rates by 40-50%. Recovery was relatively rapid at the outcrop and transition sites, particularly when compared with the dune margin site where infiltration rates were still 40% below control levels 4 years after the first disturbance. Under more typical training conditions with more passes, higher tire inflation pressures and occasionally moist soils, the effects would be expected to be even greater and more persistent. These data, together with the penetrometer resistance values, suggest vehicle traffic has a long-term, significant effect on soil structure at the dune margin site.
4.3. Biotic integrity. The significant and persistent reduction in shrub cover at all three sites, particularly in response to the track treatment, has significant implications for biotic integrity. It is correlated with reduction in foliage height diversity (not measured) which, in turn, is correlated with a number of wildlife species. The relatively slow recovery of nitrogenase activity ("Nitrogen Fixation") suggests a loss of the integrity of the microbiotic crust community despite relatively rapid recovery of crust cover.
Top of page4.4. Site comparisons. The dune margin site was clearly the most sensitive to disturbance and the slowest to recover. Several factors combine to make this site particularly sensitive to all types of disturbance. One is the low vegetative cover and dominance by low-stature saltbush which appears to be particularly susceptible to breakage and recovers slowly. The second is that gypsic soils have very low strength and are highly susceptible to compaction. Both the transition and outcrop sites have characteristics making them more resistant to degradation. The near-surface petrogypsic horizon at the outcrop site appears to provide greater resistance to both compactive and trampling-type disturbances, while the relatively high cover of resistant plant species and possibly higher water-holding capacity give the transition site a comparative advantage. However, the extremely slow rate of lichen and nitrogenase recovery at this site suggests that disturbance may have some effects on plant production not apparent in this relatively short-term (6 years) dataset. One factor which may have affected resistance and possibly recovery results is the disturbance history at each site. We believe the dune margin site was in a relatively pristine state when we initiated the experiments, while it was clear that the outcrop site had been previously used for training activities. It is possible, therefore, one of the reasons the outcrop site appeared to be so resistant to degradation is that it had already been degraded. Based on our analysis and an examination of variability within the site, however, we do not believe this factor is sufficiently important to change our conclusions.
4.5. Treatment comparisons. Off-road vehicle traffic is clearly the greatest threat to all three sites. Intensive horse trampling and even trampling by humans can also negatively affect soil stability, hydrologic function, and biotic integrity. However, the magnitude and persistence of the impacts on most indicators is generally much less. This conclusion is reinforced by the fact that the horse and infantry treatments imposed were relatively intense compared to what would normally occur during training exercises (infantry), wildlife and livestock management, or recreational activities (horse). The vehicle treatment imposed (jeep) was relatively mild with tires set at a low inflation pressure and a slow driving speed with no turns.
Top of page5. Recommendations. The following recommendations assume that the primary objective is to sustain the capacity of the land to support military training activities and other land use values. This is achieved by planning training so that recovery time is minimized. Recovery time is minimized by selecting site/training combinations that cause relatively little degradation (high resistance) or which result in rapid recovery (high resilience). The key variable to consider in each case is the number of years required for recovery.
Military planners can do this by controlling what types of activities occur, where, they occur, and when they occur. One of the primary conclusions of this study is that the effect of each of these variables (what, where, and when) depends on the others.
The outcrop site clearly represents the most suitable soil for both single and repeated disturbance. However, vehicle disturbances are the most destructive and generally require the longest recovery at all three sites. The dune site is the most sensitive to all three types of disturbance. Consequently, we recommend focusing training activities on soils similar to those found at the outcrop site and avoiding the dune margin. Based on the results, we also strongly recommend limiting traffic to existing roadways: our data show that just two passes with a light vehicle and low tire pressure can cause damage requiring 5 years or more for recovery.
6. Web-based Decision Tool/Model. A simple decision tool based on the results of this study is included on the attached CD and is posted at http://usda-ars.nmsu.edu/JER/Monit_Assess/monitoring.php.
Top of page
Belnap, J. 1994. Cryptobiotic soil crusts: basis for arid land restoration (Utah). Restoration and Management Notes. 12:85-86.
Belnap, J. 1995. Surface disturbances: their role in accelerating desertification. Environmental Monitoring and Assessment. 37:39-57.
Belnap, J. and D.A. Gillette. 1998. Vulnerability of desert biological crusts to wind erosion: the influences of crust development, soil texture and disturbance. Journal of Arid Environments. 39:133-142.
Belnap, J. and D.J. Eldridge. 2001. Disturbance and recovery of biological soil crusts. In Biological Soil Crusts: Structure, Function and Management. Belnap, J. and O.L. Lange (eds.). Berlin: Springer-Verlag.
Bradford, J.M. 1986. Penetrability. In Methods of Soil Analysis. Part I. Second ed. Klute, A. (ed.). Madison, WI. ASA and SSSA. Agronomy Monograph. 9:463-478.
Bouwer, H. 1986. Intake rate: cylinder infiltrometer. In Methods of Soil Analysis. Part I. Second ed. Klute, A. (ed.). Madison, WI. ASA and SSSA. Agronomy Monograph. 9:825-844.
Eldridge, D.J. and R.S.B. Greene. 1994. Microbiotic soil crusts: a review of their roles in soil and ecological processes in the rangelands of Australia. Australian Journal of Soil Research. 32:389-415.
Fryrear, D.W. 1986. A field dust sampler. Journal of Soil and Water Conservation Society. 41:117-120.
Herrick, J. and W. Whitford. 1995. Assessing the quality of rangeland soils: challenges and opportunities. Journal of Soil and Water Conservation. 50:237-242.
Herrick, J.E., K.M. Havstad, and D.P. Coffin. 1997. Rethinking remediation technologies for desertified landscapes. Journal of Soil and Water Conservation. 52:220-225.
Herrick, J.E., W.G. Whitford, A.G. de Soyza, J. Van Zee, K.M. Havstad, C.A. Seybold, and M. Walton. 2001. Soil aggregate stability kit for field-based soil quality and rangeland health evaluations. CATENA. 44:27-35.
Herrick, J.E., J.W. Van Zee, K.M. Havstad, L.M. Burkett, and W.G. Whitford. 2005a. Monitoring Manual for Grassland, Shrubland and Savanna Ecosystems. Volume I: Quick Start. USDA-ARS Jornada Experimental Range, Las Cruces, NM. Distributed by University of Arizona Press.
Herrick, J.E., J.W. Van Zee, K.M. Havstad, L.M. Burkett, and W.G. Whitford. 2005b. Monitoring Manual for Grassland, Shrubland and Savanna Ecosystems. Volume II: Design, Supplementary Methods and Interpretation. USDA-ARS Jornada Experimental Range, Las Cruces, NM. Distributed by University of Arizona Press.
Herrick, J.E., G.E. Schuman, and A. Rango. 2006. Monitoring ecological processes for restoration projects. Journal of Nature Conservation. 14:161-171.
Pellant, M., P. Shaver, D. Pyke, and J. Herrick. 2005. Interpreting Indicators of Rangeland Health, Version 4.0. Denver, Colorado: Bureau of Land Management.
Rango, A., A. Laliberte, C. Steele, J. Herrick, B. Bestelmeyer, T. Schmugge, A. Roanhorse, and V. Jenkins. In Press. UAV utilization for rangelands: current applications and future potentials. Environmental Practice.
Thurow, T.L. 1991. Hydrology and erosion. In Grazing Management: an Ecological Perspective. Heitschmidt, R. and J. Stuth (eds.). Portland, Oregon: Timberline Press.
Webb, H. and H.G. Wilshire. 1983. Environmental Effects of Off-Road Vehicles: Impacts and Management in Arid Regions. New York: Springer-Verlag.
