Charles D. Bonham

Secondary successional patterns in a sagebrush (Artemisia tridentata) community as they relate to soil disturbance and soil biological activity

The relationship between secondary succession, soil disturbance, and soil biological activity were studied on a sagebrush community (Artemisia tridentata) in the Piceance Basin of northwestern Colorado, U.S.A. Four levels of disturbance were imposed. I: the vegetation was mechanically removed and as much topsoil as possible was left; 2: the vegetation was mechanically removed and the topsoil scarified to a depth of 30 cm; 3: topsoil and subsoil were removed to a depth of 1 m, mixed and replaced; 4: topsoil and subsoil were removed to a depth of 2 m and replaced in a reverse order. Plant species composition, dehydrogenase and phosphatase enzymatic activity, mycorrhizae infection potentials, and percent organic matter were the variables measured. Treatment 4 drastically altered the pattern of vegetation succession. Treatments 2, 3, and 4 started with Salsola iberica as the dominant species but six years later, 3 and to lesser extent 2 changed in the direction of the species composition of 1, dominated by perennial grasses and perennial forbs. Treatment 4 developed a shrub dominated community. The rate of succession was not decreased by the increased levels of disturbance. Both dehydrogenase enzyme activity and mycorrhizae infection potential (MIP) increased with the change from Salsola iberica to a vegetation dominated by either perennial grasses and forbs or shrubs. The intensity of disturbance in 2, 3, and 4 reduced drastically dehydrogenase activity and MIP, but in six years they recovered to levels comparable to 1. Phosphatase enzyme activity and organic matter were unrelated to species composition but related to treatment and time elapsed. In both cases a significant decrease was observed throughout the six-year period.

Vegetation changes induced by prairie dogs on shortgrass range.

Highlight: This study documented some effects of prairie dogs on a shortgrass type of the Central Plains Experimental Range approximately 35 miles northeast of Fort Collins, Colo., and an adjacent area. Prairie dogs changed the plant species composition of the two sites studied, but these changes were not all detrimental. Species diversity was greater and some plant species used by livestock were more abundant inside than outside the prairie dog towns.

Interannual variation in the climate and above-ground biomass of Leymus chinense steppe and Stipa grandis steppe in the Xilin river basin, Inner Mongolia, China

We evaluated the relationship between variability in climate and variability in primary production and efficiencies of water use of Leymus chinense steppe and Stipa grandis steppe during 1980–89. On average, annual precipitation was 313·3 mm, while peak above-ground live biomass (PALB) and peak standing crop (PSC) were 182·68 g.m −2 and 193·48 g.m −2 for L. chinense steppe, 144·43 g.m −2 and 152·12 g.m −2 for S. grandis steppe. The coefficient of variation (CV) in annual precipitation was 22%, while the CV in PALB and PSC were 29% and 26% for L. chinense steppe, 24% and 25% for S. grandis steppe. Rain-use efficiency was 6·3 kgDM.ha −1 mm −1 year −1 for L. chinense steppe and 4·9 for S. grandis steppe, using PSC as the estimate of ANPP. Monthly and seasonal patterns of precipitation were as important as annual precipitation in determining responses of these two steppes.

Temporal variation in aboveground biomass ofLeymus chinense steppe from species to community levels in the Xilin River Basin, Inner Mongolia, China

We analyzed the long-term dynamics of aboveground biomass ofLeymus chinense steppe in relation to interannual variation of precipitation and temperature during 1980–1989 at levels of community, growth form and species in the Xilin river basin, Inner Mongolia Autonomous Region, China. Annual aboveground net primary production (ANPP) varied from 154.00 g m-2 yr-1 in 1980 to 318.59 g m-2 yr-1 in 1988, with a mean of 248.63 g m-2 yr-1 and the coefficient of variation of 25%. ANPP was not significantly correlated to annual precipitation and total precipitation during April–September atp≤0.05 level, but precipitation in May and August accounted for 69% of interannual variation of ANPP. The means of rain use efficiency and water use efficiency ofL. chinense steppe were 8.1 kg DM ha-1 mm-1 yr-1 and 0.89 mg DM g-1 H2O respectively. Aboveground biomass of various growth forms and species had different response patterns to interannual variation of precipitation and temperature. Monthly and seasonal distribution of precipitation and temperature were the key controls of aboveground biomass of species.

Assessing the power of the point-line transect to monitor changes in plant basal cover.

To assess the power of point data (collected systematically at each meter along a permanently-situated, 100-m line transect) to detect actual changes in plant basal cover, we developed a computational approach whereby a simplified shortgrass steppe community was spatially simulated on a computer screen. Cover was then reduced using a random disturbance pattern. One transect could detect an actual decrease in cover from 12% to 8% with less than 20% probability, while 5 transects increased this power to about 80% (P less than or equal to .05). A reduction in cover from 12 to 6% could be detected with 80% probability with only 2 transects, while a cover reduction to 10% could only be detected with 40% probability using 10 transects (P less than or equal to .05). Artificial populations provide a valuable mechanism for quantitatively evaluating field sampling designs.

Frequency of endomycorrhizal infection in grazed and ungrazed blue grama plants.

Highlight: The frequency of mycorrhizal infection in blue grama roots was determined from two criteria: (1) occurrence of any mycorrhizal element, and (2) occurrence of fungal vesicles. No significant differences were observed with respect to grazing using the first frequency criteria. However, roots of previously grazed plants had significantly higher frequencies of vesicles than those collected from exclosures. Frequency of vesicles was found to increase linearly with increase in rooting depth of blue grama. Significant grazing effects on the frequency of vesicles were observed primarily in the first sample depth, 0-10 cm.

Comparison of Techniques Used for Adjusting Biomass Estimates by Double Sampling

This paper compares the ratio and regression estimator procedures for adjusting ocularly estimated plant species biomass in different sizes and shapes of plots. The study was conducted in northeastern Colorado on shortgrass rangeland dominated by blue grama (Bouteloua gracilis). No significant differences were found in clipped plant biomass in 4 quadrat sizes between 0.18 and 0.50 m* and 2 shapes, circular and angtdar quadrats. For double sampling, the scatter plots of data strongly indicated a linear relationship through the origin for estimation and clipping. There were no significant differences between the adjusted mean weights by use of regression with and without intercept. The intercept was not significantly different from zero. Interpretation of correlation coefficient and variance of regression estimate with no intercept becomes difficult because the regression is forced through zero. Therefore, it is helpful to use regression with intercept. In the present study, estimates of both green and dry weights by ratio and regression estimation were comparable. Regression estimation is a minimum variance estimation comparable to ratio estimation even when the assumption of homoscedasticity is not true. The single factor of greatest importance to range management is an accurate appraisal of the volume of forage available. The task is difficult simply because forage varies in the weight of plant material produced by each species in a highly variable environment. Because all the forage cannot be harvested and weighed, we must obtain a reasonable estimate of the actual weight by sampling. Sampling is often not an easy task because data must be obtained within a span of 1 or 2 weeks so that growth differences are minimized. On summer ranges, the information on residual forage is obtained after the livestock have been removed and before snow makes the sampling impossible. If estimation of biomass for individual species is needed, then available funds and personnel may pose limiting constraints. In order to overcome some of these problems, a weight estimate method was designed by the personnel of the Intermountain Forest and Range Experiment Station during the summer of 1936 (Pechanec and Pickford 1937). The data are collected in two phases. In the first phase, the desired factor (xi) is measured by some indirect method such as ocular estimation. In the second phase, the desired factor is measured both directly and indirectly. The indirectly estimated values (xi) in the first phase are then adjusted by developing a mathematical relationship between the direct estimates (yi) and indirect estimates (xi). The sample size in the first phase is usually large compared with the sample size in the second phase. The sample in the second phase is usually &random subsample from the first phase but it may be drawn independently. The mathematical procedures used for adjusting the indirect estimates are linear regression and ratio estimation. The theory of Authors are with the Range Science Department, Colorado State University, Fort Collins, 80523 and the USDA-ARS Crops Research Laboratory, Colorado State University, Fort Collins 80523. This research was supported by U.S. AID Grant No. 391-80183 to Javed Ahmed and by facilities proved by USDA-ARS at the Central Plains Experimental Range, Nunn, Cola.. and Colorado State University Experiment Station Scientific Series No. 2620. Manuscript received June I, 1981. JOURNAL OF RANGE MANAGEMENT 36(2), March 1983 linear regression requires the assumption that the population regression of y on x is linear, that the residual variance of y about the regression line is constant (homoscedasticity), and that the population is infinite. No assumptions are made about the line passing through the origin. In the ratio estimation procedure, assumption of homoscedasticity need not be made but the estimator works well when up = xi. In many ways, the procedure is analogous to fitting a linear relationship between y and x which passes through the origin. When we are trying to decide what kind of estimate to use, a graph in which yi is plotted against xi is helpful. If the graph shows a straight line, relationship through the origin and variance of points yi about the line seem to increase proportionally to xi, then the ratio estimate is better than the least squares estimate for regression (Cochran 1963). During the summer of 1979, data were collected by double sampling from the Central Plains Experimental Range (CPER) near Nunn, Colo. CPER is administered by the Agricultural Resea Service, USDA. The objective of this study was to compare the ratio and regression estimator procedures for adjusting the ocularly estimated species biomass. The technique of double sampling and its statistical aspects are described in most sampling technique textbooks. The technique is also described by the National Research Council (1962) and by Schumacher and Chapman (1948). Pechanec and Pickford (1937) gave a detailed outline for the training of personnel for double sampling. Burton (1944) reported the ability of different personnel to estimate the yield in plots. Double sampling determination of herbage production in different vegetation types and the results were discussed by Pickford (1940), Wilm et al. (1944), Ragsdale (1956), Hillmon (1959), Hughes (1959), Shoop and Mcllvain (1963), and Tadmor et al. (1975). Double sampling was found desirable for extensive browse inventories by Carhart and Means (1941), Schawan and Swift (1941), Dasman (1948) and Blair (1959). Abstracts of the double sampling technique are given by Morris (1967). Statistical aspects of double sampling were reviewed by Francis et al. (1979). Optimum allocation of resources to direct and indirect methods of estimation is well defined by Cochran (1963) for a single factor under study and for a given sampling procedure. The optimum allocation formulations described by Schumacher and Chapman (1948) and Wilm et al. (1944) are. similar to those described by Cochran (1963). More recently, Ahmed (1980) and Ahmed and Bonham (1980) described a technique for optimum allocation in multivariate double sampling for biomass estimation.

Patterns of tree species richness in Jalisco, Mexico: relation to topography, climate and forest structure.

AbstactThe objective of this study was to identify the major environmental variables and components of forest structure associated with variability in tree species richness on a network of 806 permanent plots in the State of Jalisco, Mexico. Tree data recorded on the sample plots were used to characterize tree species richness by forest type and climatic conditions (temperature and precipitation) in the State. Species composition and other diversity indices were also calculated. Explanatory variables identified in a Poisson regression identified forest cover type, elevation, tree basal area, canopy closure, and winter precipitation as being important to changes in tree species richness. An “extreme quantile curve estimation” approach was then used to approximate the boundary that represented the maximum potential species richness response to the various levels of important variables. Maximum tree species richness decreased with increasing elevation. The relationships between maximum species richness and tree basal area, canopy closure, and winter precipitation followed a hump-back unimodal model, with intermediate values supporting the largest species richness. We believe that results of the current study will contribute to further development of a conservation plan for tree species in the State of Jalisco, Mexico.

Least squares regression vs. geometric mean regression for ecotoxicology studies

Least squares linear regression and geometric mean regression analysis procedures are compared for use in analyzing ecotoxicology data. These data are characterized by both response data and environmental factor data being measured with error. According to cited literature, when factor data are measured with error, regression analyses for response as a function of environmental factors should be performed using geometric mean regression. A case is made in the current paper that it is more appropriate to use least squares regression in analyses of such data rather than geometric mean regression.

Influence of spatial autocorrelation on a fixed-effect model used to evaluate treatment of oil spills

A study was conducted during 1989 by the EPA to evaluate the effectiveness of two methods of nitrogen fertilizer application to treat the Exxon oil spill on beaches of Alaska. The current study, using the data, was to determine the spatial relationship of oil residue on these beaches and to evaluate the influence of spatial autocorrelation on the power of a fixed-effect model used by EPA investigators. The distribution of residual dry weight (mg oil residue/kg of beach material), was found to be spatially independent on individual beaches and for individual sampling times. The presence of spatial autocorrelation resulted in an overestimation of treatment means and an underestimation of their respective standard errors. In turn, an increase in the significance of differences in individual treatments and in the power of the test resulted. A spatial autoregressive model was used to correct for the presence of spatially autocorrelated residuals and the model was used to obtain the best linear unbiased estimates of parameters. Analysis of variance showed that oil residue concentration was significantly higher on the control beach (Raven Beach) than on either of the beaches fertilized with two different nitrogen treatment applications (Kittiwake Beach or Tern Beach). These results also showed that the two fertilizer treatments reduced the concentration of oil residue significantly faster than would have occurred without treatment. The power of the experimental design, a randomized block design used by the EPA, adequately detected significant differences among treatments. A fixed-effect model was used to analyze the data. Furthermore, the current study also found that increasing the sampling intensity or replicating the experimental design could have decreased the power of the test because of the presence of spatially autocorrelated data. In general, the power of the fixed-effect model for the design to detect a significant difference among treatments increased over time. In contrast, the power of the fixed-effect model to detect a specified difference across all times decreased over time because of decreases in means and their associated standard errors.