| |
Final Report, September 1998Vegetative Cover Monitoring, Public Perception Survey, and Public Outreach Programs for
|
Demonstration Cell
Coordinated Resource Management Team members representing the Yavapai Ranch, U.S. Forest Service, Soil Conservation Service, Arizona Department of Environmental Quality, Friends of the Prescott Forest, and Arizona Cooperative Extension Service met on October 17, 1991 at the Demonstration Cell location. The team members visited several locations and selected three tentative sites for intensive monitoring. These sites were a limy upland in fair ecological condition, a loamy upland in fair ecological condition, and a loamy bottomland in poor ecological condition. Team members decided at this time that livestock exclosures would be constructed at each site and vegetation cover data would be collected inside the exclosures protected from grazing and outside to evaluate the time control grazing program. A third fenced area was decided upon for the loamy upland site. This fenced area was selected to demonstrate the effect of periodic high animal impact on vegetation and soil cover characteristics.
Monitoring Site Characteristics
On April 6 and 7, 1992, the final site locations were marked and initial baseline data collected for the seven plots located on the three sites. Table 1 provides information on location and characteristics of the three sampling sites. Methods for the soil sampling and analyses data are in Appendix D.
Vegetation and Soil Cover Sampling
Point sampling was selected as the technique for monitoring vegetation cover change. Since a diversity of personnel were anticipated to participate in the collection of cover data, a 10-point frame technique was utilized to reduce individual decisions regarding cover to a decision at a point. Each plot consists of a permanent 200-foot centerline with perpendicular transects selected at 10-foot intervals to the left and right of the centerline. From this potential of 40 transects, 10 transects (samples) were randomly selected for sampling at each date. Five 10-point frames (subsamples) are systematically placed at one pace (2 steps) intervals along each transect. Data for the five frames per transect are summed as the sample unit for statistical analyses.
The sampling was designed to involve the Coordinated Resource Management Team and other interested individuals in an annual spring data collection and field review of on-site conditions for the demonstration (Fig. 5). Additional information on sampling methods, annual dates of monitoring, participants, and general notes may be found in Appendix E. Ten to 17 participants participated in the annual spring monitoring data collection sessions.
Objective
Vegetative and ground cover monitoring is done for this project to document the changes in prevalence of plant species and ground cover in response to seasonal precipitation and to implemented livestock management practices associated with fencing and water improvements. These vegetative and ground cover data are surrogate measures of the potential for sediment movement by water; thus, affecting water quality.
Vegetation and Ground Cover in the Spring, Photos
A photo was taken of the centerline of each plot for each spring sampling date. These color photos are displayed in Appendix F. The photos of the plots on April 6 and 7, 1992 show an apparent visual similarity of paired plots at each of the three sampling sites. The April 22, 1996 photos (Appendix F-10 and F-11) illustrate the extremely dry conditions for this sample date due to low rainfall in the winter and spring of 1995-96 (Fig. 2). The photos in Figs. 6, 7 and 8 provide the visual appearance of the seven monitoring plots on April 27 and 28, 1998.
Vegetation and Soil Cover Data in the Spring, 10-Point Frame
Data by plot within years, expressed as hits/50 point samples, are provided in Appendix G, including estimates of standard deviation (s) and 95% confidence intervals (CI) for each attribute measured. An awareness of the data variability is provided by a brief discussion of the canopy cover data for Plots Y-1 and Y-2. The coefficients of variation (CV) for the 10-point sample data for canopy cover ranged from 28% to 64% with a mean of 47% for Plots Y-1 and Y-2 sampled from 1992 through 1998. The CV is the standard deviation expressed as a percentage of the mean. For data with 10 samples per mean (n=10), the Least Significant Difference (LSD) between two means at the 95% level of significance is approximately equal to the CV.
A check through the data summary sheets Appendix G shows that variability illustrated for canopy cover of plots Y-1 and Y-2 is representative of other measured attributes, except when zeros are common sample values. Thus, differences between attribute means for grazed and not grazed plots must be on the order of 40 to 50% of their mean value to be statistically different at the 95% confidence level.
Coefficients of variation ranging from 30 to 60% are common in biological field sampling. There can be individual means that are a poor estimate of a true population mean, but sample means tend to approach true population means as sample numbers increase. Realistic interpretation of data requires good field experience and knowledge of associated interactive factors. A team approach to data collection and interpretation helps separate the biologically meaningful trends from the sampling background variation.
A summary of selected cover attributes expressed as percentage of pin hits for each of the seven monitoring plots from 1992 through 1998 are provided in Table 2. Canopy cover and soil surface cover are shown graphically by site in Figs. 9 (limy upland), 10 (loamy upland), and 11 (loamy bottomland).
Canopy cover on the Y-1 (not grazed) plot increased from 1992 to 1994 (Fig. 9) then decreased from 1994 to 1997. The high canopy cover on the Y-1 plot in the spring of 1994 is not readily explained. The low 15% canopy cover for Plot Y-1 in 1997 may be explained by the very low precipitation of 1995-96 (Fig. 2). The differences between canopy cover of Plot Y-1 and Plot Y-2 (grazed) are the expected result of utilization by cattle (Fig. 3). Utilization of forage by cattle in the area of Plot Y-2 was very light in the 1992-93 grazing period even though Pasture 5 was grazed for 2 periods for a total stocking rate of 10.4 ADAs. Pasture 5 was not grazed in the summer of 1997 (Fig. 3). The canopy cover of 22% for Plot Y-2 in 1998 compared to 25% canopy for plot Y-1 could be either a carry over effect of previous year grazing or could be due to sample variation. The canopy cover on both Plots Y-1 and Y-2 was slightly higher in 1998 than for the base year of 1992.
Total cover on the soil surface of Plots Y-1 and Y-2 tended to follow the same pattern as shown for canopy cover, but with much less variability over years and less differences between Plots. The surface cover averaged about 60% for both plots over the monitoring period.
On the Loamy upland site (Fig. 10) the canopy cover on Plot Y-4 (not grazed) tended to increase from 1992 through 1998. In addition to being protected from livestock grazing, this plot received two overland water flow events that had little influence of either Plots Y-3 or Y-5. The first flow event occurred in August 1993, and was documented by Dave Smith, Natural Resource Conservation Service in Prescott. The second overland flow event occurred in August or early September 1996. The extent of the area affected by the flow was very evident at the time of the fall 1996 monitoring.
Canopy cover decreased severely on Plot Y-5 (grazed) from 1994 to 1997 despite a cattle stocking rate of less than 3 ADAs on Pasture 14 during this period. The extremely poor precipitation year of 1995-96 appears to be a primary cause for the decline of the blue grama canopy cover during this period. Pasture 14 was rested from grazing from August 25, 1996 through the April 27, 1998 sampling. Canopy cover of Plot Y-5 in 1998 was considerably lower than for the non grazed Plot Y-4 but was higher than for the initial sampling in 1992.
The plant canopy cover on Plot Y-3 (Animal impact) tended to remain higher than the Y-5 (grazed) plot in 1996 and l997. The high animal impact treatment of 35 cattle on the plot for 4 hours prior to the April 1996 sampling periods was accomplished in early April 1995, so there was a summer, fall, winter, and spring rest period before the 1996 sampling. The next animal impact on Plot Y-3 was 220 cows for 2 hours on August 20, 1996. The vegetation only had fall, winter, and spring growing periods to recover prior to the spring 1997 sampling. Similar to the Limy upland range site, the cover on the soil surface for treatment Plots Y-3, Y-4, and Y-5 was relatively stable over years, averaging over 60%.
The loamy bottom site baseline sampling in 1992 documented a canopy cover between 70 and 80% and a soil surface cover of 20 to 30% (Fig. 11). The vegetation on the site in 1992 was dominated by an annual mustard weed (Chorispora tenella). Yearlong grazing, with cattle concentrating at this location, is hypothesized to be the cause of the annual weed vegetation and low soil surface cover in 1992. a major increase in cover, primarily litter (Table 2), on the soil surface was attained between spring 1992 and spring 1993 for both the ungrazed (Y-6) and time control grazed (Y-7) plots. The low in canopy cover on Plot Y-6 (not grazed) occurred in the spring of 1996 in response to the dry winter of 1995 and spring of 1996 (Fig. 2). Canopy cover and soil surface cover on Plot Y-7 (grazed) were further reduced at the time of the 1996 spring sampling due to a stocking rate of 4.9 ADAs on Pasture 8, with the second period grazing from October 17 to 29, 1995 (Appendix C). The dry winter of 1995 and spring of 1996 (Fig. 2) produced little regrowth after this grazing period.
Watershed Erosion Prediction Project Model Results
The spring vegetation sampling with the 10-point frame was designed to provide input parameters to produce simulation runs with the Water Erosion Prediction Project (WEPP) Hill Slope Profile Model, single storm mode. Mary Kidwell and Mark Weltz, USDA Agricultural Research Service, Tucson, Arizona provided the technical support to produce the simulation runs with the WEPP model.
Four input data files are required to run the hillslope portion of the model. These are: (1) a climate file, (2) a slope file, (3) a soil file, and (4) a plant/management file. For the climate file, the predicted 25-year rainfall frequency for 30 minutes for the Yavapai Ranch area was selected as the rainfall event for the single storm simulation. The amount of rainfall for this storm is 1.3 inches/30 minutes (Source: U.S. Dept. Commerce, Weather Bureau. 1961. Rainfall Frequency Atlas of the United States. Tech. Paper No. 40. Washington D.C.)
For the slope file, slope at the sites were as shown in Table 1. The simulation was made for a single, uniform slope segment over the length of the monitoring plots (60 meters).
For the soil file, soil data from Table 2 were input. The formulas, however, in the WEPP User Summary (July 1995) for calculating Ki, Interrill erodibility, Kr, Rill erodibility, t c, critical hydraulic shear, and Ke, effective hydraulic conductivity may not be reliable for many rangeland soils. In fact, calculation of effective hydraulic conductivity based of the WEPP User Summary formulas and data for soils at the Yavapai monitoring sites produced negative and zero values. Values for erodibility, hydraulic shear, and effective hydraulic conductivity developed from a WEPP field trial on a loamy upland range site designated as Location 1 (Prescott, AZ) in the WEPP User Summary manual were used for simulation runs for the Yavapai sites. The vegetation of the Prescott site was dominated by blue grama and the surface texture of the soil (Lonti Soil Series), sand 48% and clay 18%, is similar to the soils for the Yavapai sites.
For the plant/management file, initial conditions section, the standard deviation of pin heights measured in inches in the field and shown in Figs. 12 and 13 were converted to meters as a random roughness index utilized in the model. Canopy cover, soil surface cover beneath canopy (interrill cover), and soil surface cover not beneath canopy (rill cover) as shown in Table 3 also were input parameters.
Simulation runs for Plots Y-1 through Y-7 were made for the 1992 baseline year and for 1998 after seven years of time controlled grazing, a high animal impact treatment, and data from exclosures with no grazing. Runoff volume in millimeter depth leaving the plots and sediment leaving the plots expressed as Kg/m width x 60 m length are shown in Fig. 14.
Predicted runoff from Plots Y-1 through Y-5 for the 33.02 mm storm event over 30 minutes was about 12 mm or 36% of the rainfall amount (Fig. 14). There was very little difference in simulated runoff between years or plot treatments for these five plots. Runoff volume from Plots Y-6 and Y-7 averaged about 9 mm (Fig. 14), 27% of the rainfall amount. Less runoff from the loamy bottom plots (Y-6 and Y-7) compared to the limy upland and loamy upland plots appears to be a function of calculations within the WEPP model based on soil characteristics. The effective matrix potential was calculated as 55.63 mm, 53.12 mm, and 72.96 mm for the limy upland, loamy upland and loamy bottomland sites, respectively.
Simulated runoff in 1998 for the loamy bottomland Y-6 and Y-7 plots was predicted as 8.7 and 8.8 mm, respectively, compared to 9.6 mm for both the Y-6 and Y-7 plots in the baseline year of 1992. Soil surface cover was much higher in 1998 on Plots Y-6 and Y-7 compared to 1992 (Fig. 11), and this additional soil surface cover apparently resulted in the lower simulated runoff in 1998 compared to 1992.
For the Limy upland site, the predicted sediment leaving Plots Y-1 and Y-2 was greater in 1998 than for the 1992 baseline year even though canopy cover was greater in 1998 than in 1992 and the surface cover was similar between the two years. The high estimate of sediment leaving Plot Y-1 for the 1998 simulation appeared inconsistent with the other estimates. The Y-1 data for 1998 were rerun with only the random roughness changed (Appendix H-42). The change increased the roughness index of Plot Y-1 in 1998 from 0.3 inches (.008 m) as measured (Fig. 12) to 0.4 inches (.010 m) which were the measured values for this plot in 1995 and 1997 (Fig. 12). The 30.8 kg/m x 60 m for the initial simulation as shown in Fig. 14 was decreased to 20 Kg/m x 60 m for the sensitivity test simulation. The amount of sediment leaving the sample plots for the simulation runs was very sensitive to the magnitude of the random roughness index. The 1993 roughness indices were utilized for the l992 simulations, as this measurement was not taken in 1992.
The cover on the soil surface was not greatly different among treatments or between 1992 and 1998 for Plots Y-3, Y-4, and Y-5. The sediment yield differences for plots Y-3, Y-4, and Y-5 appear to be, in large part, explained by the random roughness differences. For plots Y-6 and Y-7 the random roughness magnitudes are the same for 1993 as for 1998. The decreased predicted sediment yield in 1998 compared to 1992, like the runoff volume, for Plots Y-6 and Y-7 appears to be due to more litter on the surface in 1998 compared to 1992.
Vegetation and Ground Cover in the Fall, Point and 40- x 40-cm Quadrat Data
The Yavapai Ranch Coordinated Resource Management Monitoring Team met at the Demonstration Cell on the ranch each fall beginning in 1992 as well as each spring. In 1992 and 1993, each of the seven monitoring plots within the Demonstration Cell were visited and team members discussed project progress and observations at each plot. Notes and comments are summarized in Appendix I. Photos of each plot also were taken and these are found in Appendix J.
Discussion among team members identified the need to collect data at these fall visits to the monitoring sites to supplement the observations and the spring 10-point frame data. The 10-point frame data are not sensitive to identify changes in individual plant species on the monitoring plots and are too time consuming to be a method to extend monitoring to many sites over the ranch. A quadrat sampling method that incorporated a point within the quadrat was proposed.
Quadrat sampling was begun in the fall of 1994 (Fig. 15). Detailed sampling procedures are provided in Appendix I. Quadrat size is 40- x 40-cm, and 20 of these quadrats are placed at regular (1- or 2-pace) intervals along each of 5 transects for a total of 100 quadrats. Hits at the point in the quadrat are recorded to document soil surface cover, presence of a plant species in the plot is recorded to determine plant frequency, and the three most prevalent species in each plot are ranked based on their dry weight for use in determining species composition of the plant community. A photo of the monitoring plot also is taken, as is done for the spring monitoring. The photos are in Appendix J. The annual data are summarized by plot in Appendix K.
Data for presence or absence of plant species in a plot fit a binomial population distribution. Confidence intervals for percentage frequency of occurrence in plots is a function of number of plots in which a species occurs and number of sample plots. Confidence intervals for percentage frequency for 100 sample quadrats are displayed in Table 4. Frequencies within the range of 36 to 64% have a 95% confidence interval of plus or minus 10 with a sample of 100 quadrats. The binomial confidence intervals are not applicable to the point data or dry-weight rank data, as these data are not yes or no decisions.
In general, the total soil cover recorded in the fall sampling from 1994 to 1997 (Figs. 16, 17, and 18) showed relative high soil cover and stability over the study period, similar to that shown by the spring sampling (Figs. 9, 10, and 11). The differences reflect a combination of summer growth of warm season plants and a tendency to view the point in the quadrat as a broader area than for the sharp points of the 10-pint frame used in the spring data collection. Soil surface cover data tended to be about 10% higher in the fall than in the spring for both the limy upland and loamy upland range sites. The relative magnitude of cover on the loamy bottom site was similar at spring and fall sampling except, cover on the Y-7 (grazed) plot was 50% in the spring 1996 and was recorded as near 30% in the fall. The dry year (Fig. 2), an annual plant community, and heavy cattle use of the plot area with a stocking rate of 2.4 ADA in the summer of 1996 (Fig. 4) account for this low soil surface cover in the fall of 1996.
Plant frequency data for Plots Y-1 and Y-2 are displayed in Figures 19, 20, and 21. The grass frequency data (Fig. 19) show that frequency of sideoats grama, needle and thread, and threeawns all are higher on Plot Y-1 compared to Plot Y-2, and blue grama has a higher frequency on Plot Y-2 than on Y-1. The 10-point frame data for 1992 and 1993 for Plots Y-1 and Y-2 (Appendix G), although not sensitive to provide a quantitative estimate of species abundance, showed more blue grama on Plot 2 than on Plot 1, and more needle and thread and threeawns on Plot Y-1 than on Y-2. Sideoats grama was recorded about equally on both plots in 1992 and l993. Sideoats grama decreased in frequency on both Plots Y-1 and Y-2 in 1996, apparently associated with the low precipitation year (Fig. 2). Locoweed (Fig. 20), snakeweed, and pingue (Fig. 21), all poisonous plants to livestock, also decreased in frequency on both plots in response to the drought year of 1995-96 (Fig. 2).
The species composition data collected during the fall sampling is shown in the plot annual data summaries provided in Appendix G. These composition data help to show the relative dominance of a species in a community. For instance, frequency of Evolvolus on 40- x 40-cm quadrats of Plots Y-1 and Y-2 are about 50%. The percentage composition of the community based on weight for this species (Appendix G) is less than 10%. Composition data, however, are relative data, a change in one species creates a change in other species. As an example, sideoats grama on Plot 1 in 1992 made up 49% of the plant community and annual forbs only 4%. In 1993 annual forbs were abundant and made up 28% of the community composition. Sideoats share of the composition was decreased to 26%. The plant frequency data are utilized in this report rather than composition data to discuss trends over time.
On the loamy upland range site (Plots Y-3, Y-4, and Y-5) blue grama remained near a frequency of 100% for the 40- x 40-cm quadrats (Fig. 22). This quadrat size is too large to document changes in the dense blue grama sod present of this site.
Western wheatgrass frequency was near 50% in 1994 on Plot Y-4 (ungrazed) and remained at near 40% for the 1995 and 1997 sample dates (Fig. 22). The 40% frequency of western wheatgrass on Plot 4 in 1997 is a response to protection from livestock grazing and two overland water flow events (August 1993 and August or early September 1996) that had little effect of Plots Y-3 or Y-5.
Perennial forbs are sparse on the loamy upland site. Gara had near a 20% frequency in 1994 and a 30% frequency on Plot Y-3 (animal impact)in 1995, but this is the only treatment or years that perennial forb species showed frequencies over 10% (Fig. 23).
The loamy bottom range site (Plots Y-6 and Y-7) is dominated by annual forbs, but seedlings of sand dropseed and spread of vine mesquite by stolons were beginning to establish on both the grazed and not grazed plots. In the fall of 1996 frequency of sand dropseed plants on both plots was near 70% (Fig 24). Most of these plants were small and poor in vigor and did not survive to the fall of 1997. Perennial ground cherry is the most prevalent perennial forb on the site and appears to vary greatly from year to year (Fig. 25).
Commercial Fuelwood Sale
Baseline monitoring data for a commercial fuelwood cut in the Demonstration Cell area were completed on April 28, 1995 and September 23, 1996. Information and data for Plots Y-8 (not to be cut) and Y-9 (to be cut) are provided in Appendix L. Photos of the study area taken in 1996 are in Figure 26. Figure 25 displays cover and frequency data. The juniper canopy cover on the site is near 30% and the soil surface is covered with 20 to 25% gravel and 35% litter (mostly beneath tree canopy) (Fig. 27). Understory vegetation is very limited. Blue grama, 4% frequency, is the major understory species (Fig. 27). The photo of Plot Y-10 (Fig. 26) shows gully erosion on the site.
The wood sale project was temporarily stopped by a July 1996, injunction by the 9th U.S. Circuit Court of Appeals. The injunction was obtained to halt timber harvest until the Forest Service adopted plans to protect rare and endangered species on Arizona and New Mexico Forests. The injunction has been lifted and cutting within the fuelwood area has begun, but Plot Y-8 area had not yet been cut by April 28, 1998.
Historical Exclosure
The monitoring team first visited an historical exclosure within the Demonstration Cell on April 30, 1993. Plot Y-8 (later changed to Y-11) was established as a photo point located 5 paces west and in line with the south fence of the exclosure. The photo was taken to the north, perpendicular to the south fence line of the exclosure. Plot Y-9 (later changed to Y-12) was located at the center of the south fenceline of the exclosure and the photo was taken to the north. Plot Y-10 (later changed to Y-13) was established as a photo point located 5 paces east of the southeast corner of the exclosure. This photo also was taken to the north. The notes for this initial visit to the exclosure may be found at Appendix E-8.
On April 15, 1994, 50, 40- x 40-cm quadrats were systematically placed along paced transect inside the exclosure (Y-12) and to the east of the exclosure (Y-13).
Plant species frequency was measured and species composition determined by the dry-weight-rank method.
Historical information and data for the exclosure were located in 1998. A copy of the historical information, the 1994 data, and photos of the exclosure and adjacent plots are in Appendix M. The exclosure was fenced in 1939, and plant intercept data inside the exclosure and adjacent to the exclosure to the east (our Y-13) were measured in 1940 and 1953. Species frequency and dry-weight rank data were collected at the exclosure site on April 15, 1994. Although not comparative data without some bias, percentage plant composition based on line intercept data and percentage composition based on dry-weight-rank data are the data available to compare the historical data and recent data inside and outside the exclosure. These composition data are displayed in Fig. 28. From 1940 to 1953 there appeared to be little change in plant community composition inside and outside the exclosure (Fig. 28). By 1994, the plant composition inside the exclosure had changed to a community dominated by squirrel tail, and the plant community outside the exclosure, available for yearlong cattle grazing, was a plant community dominated by blue grama, as it was in 1940. No data were collected to document the changes in the plant communities between 1953 and 1994.
The photos taken on September 29, 1997 confirm that the appearance of Plot Y-13 (grazed) is much the same as it was in 1940 (Fig. 29 and Appendix M). Plot Y-12 (inside the exclosure and protected from livestock grazing) had abundant bare areas in 1997, as much of the squirreltail present at the time of the 1994 sampling died as a result of the low seasonal precipitation of winter 1995 and spring 1996 (Fig. 2).
The soil within the exclosure was extremely soft and powdery at the surface in 1997 (Appendix I-14). Footsteps across the plot were readily visible. The soil outside the plot was firm and footprints were not obvious. A complete evaluation of the effect of 50 years of livestock exclusion on the watershed characteristics of this plot will require not only soil surface cover data, but also data to document soil characteristics that determine effective hydraulic conductivity, critical hydraulic shear, rill erodibility and interrill erodibility. Random roughness of the soil surface also needs to be evaluated.
Deep Well and Powerline Pastures
Monitoring was expanded to Deep Well and Powerline Pastures on the west side of the ranch beginning October 11, 1995. Plot 14 is a loamy upland site located in Deep Well Pasture, and Plot 15 is located on a loamy bottomland site in Powerline Pasture. Photos of these plots as they appeared September 29, 1997 are displayed in Fig. 30. The annual information, photos, and data for these plots are contained in Appendix N.
The data displayed in Fig. 31 show total surface ground cover increased on Plot Y-14 from 62% in 1995 to 76% in 1997. Total cover on Plot Y-5 decreased from 78% in 1995 to 71% in 1997. The soil surface cover on both plots average about 70% and likely are not at risk of serious soil erosion.
Squirreltail frequency on Plot Y-14 and western wheatgrass on Plot Y-15, both classed as cool-season grasses, were greatly reduced in 1996 compared to 1995 (Fig. 32) in response to the dry winter of 1995 and spring of 1996 (Fig. 2). Squirreltail also decreased in the historical exclosure (Y-12) in 1996. Rabbit brush on Plots Y-14 and Y-15, snakeweed on Plot Y -14 (Fig. 33) and globe mallow on Plot Y-14 decreased (Fig. 34) in response to the dry winter to spring, 1995-96 (Fig. 2). The effect of the low precipitation in 1995-96 is dramatic and is the dominant factor affecting vegetation changes on Plots Y-14 and Y-15 during the 1995-1997 monitoring period.
Next Section: Public Participation and Education