8097 - Laserfiche WebLink

Home Browse Search

2) aspen and conifer; 3) cottonwood, buffaloberry and hawthorn communities tended to have similar understory species. Generally aspen and conifer com- munities occurred at higher elevations, while buffaloberry and hawthorn communities occurred at lower elevations. Analysis of variance and pairwise comparisons were used to determine which cluster analysis groups, based on importance values (i.e., riparian plant community types), had significantly different mean values for selected parameters. While significantly different mean values were found for nearly every parameter measured between at least two cluster groups, some parameters tended to produce more cluster groups with significantly different means. Physical parameters related to plant moisture stress had the greatest differ- ence between cluster groups. Fewer significant differences were shown be- tween cluster groups for soil moisture measurements, determined for three levels in the upper 30 cm of soil, than for plant water potential as determined by the pressure chamber technique. Water potential values reflect how stressed the plant really is. It takes into consideration stresses of the plant root system that may not be reflected by gravimetrically determined soil moisture measurements. For example, soil moisture measurements may indicate dry surface soil conditions. At the same time plant water potential values indicate that the plant is experiencing little if any moisture stress, possibly due to having its roots deep into the water table. Examples of mean water potential of the dominant species woods rose (a ubiquitous species found at each site) , and the average mean water potential for all woody species found at a site are shown in Figure 1. Note that differences between cluster groups are less apparent for mean soil moisture (Figure 1). Significant differences in soil texture were also found between cluster groups (Figure 1). Cluster groups dominated by willlows tended to have low soil pH (6.0-6.5) and high clay content. This resulted from alluviation associated with seep areas or reduced stream gradients caused by terminal glacial morraines or beaver dams. Cluster ~roups dominated by conifer species tended to have rather low soil pH (6.0-7.0) and sandy soils where stream grad- ients were rather steep. Alder stands also had relatively low soil pH (6.0- 7.0) and low electrical conductivity (less than .05 millimhos/cm at 250C- saturated soil paste extract), while hawthorn and buffaloberry had rather high soil pH and electrical conductivity (pH 7.8-8.0 and 1.0-3.5 millimhos/cm at 250C-saturated soil paste extract, respectively). No biologically significant differences in vegetative parameters within cluster groups were shown for solar irradiation exposure produced by shading from adjacent canyon walls and hills. From these results it ~an be concluded that reduction in photosynthesis due to shading produced no significant dif- ferences in the vegetative parameters measured. Forward, stepwise discriminant analysis was used to select combina- tions of environmental variables according to their relative importance in discriminating riparian classification groupings (determined by cluster analysis). Six of 55 variables were selected (a = .05). In order of importance they are: (1) percentage of the soil sample volume filled with air in the B sampling depth (13-17 cm); (2) mean water potential of the dom- inant species; (3) electrical conductivity of soils in the A sampling depth (0-5 cm); (4) soil pH in the B sampling depth (13-17 cm); (5) index of soil stoniness and/or compaction; and (6) percentage of sand in the soil in the B sampling depth (13-17 cm). The model correct+y classified about 64% of 62