Rocky Mountain Section - 72nd Annual Meeting - 2020

Paper No. 7-4
Presentation Time: 8:30 AM-4:30 PM


LANDAU, Lydia1, LEVINE, Rebekah1 and DYRESON, Eric2, (1)Environmental Sciences Department, University of Montana Western, 710 S. Atlantic St., Dillon, MT 59725, (2)Mathematics Department, The University of Montana Western, 710 South Atlantic St, Dillon, MT 59725

Incision and earlier seasonal snowmelt reduce plant-available soil moisture in high-elevation headwater tributaries throughout the growing season. Reduced soil moisture results in decreased forb and arthropod diversity, abundance, and persistence. Mesic habitat restoration structures have been built in Little Basin Tributary 7 in southwest Montana to address these issues. Understanding how soil moisture is impacted by this style of restoration is essential to evaluating and improving its efficacy. We used HOBO data loggers and soil moisture probes to gather hourly soil moisture data along 6 reaches of Tributary 7. Three reaches had mesic restoration structures installed in 2018, and three reaches were left as control sites. Each logger was equipped with three soil moisture probes, buried 30cm deep at 0m, 1m, and 3m from the center of the channel at the downstream-most proposed structure location in each reach. In this study, we use principal components analysis (PCA), canonical variates analysis (CVA), and two-factor analysis of variance (two-way ANOVA) to look for structure in the data, determine whether we could discriminate between treatment and control reaches based on the data structure alone, and determine whether or not mesic restoration status (MRS) and/or distance from center of channel are significant factors in determining soil moisture. PCA showed no obvious data structure on the basis of MRS. CVA was not able to discriminate between treatment and non-treatment sites. Mahalanobis-distance-based classification correctly classified treatment and control sites with less than 60% accuracy. Two-way ANOVA showed that distance from channel is a significant factor in determining soil moisture, but that MRS is not a significant factor. However, we used a 0.05 level of significance, and the p-value for MRS was 0.057. With an adjusted experiment design, it may be possible to detect the possible role of MRS in determining soil moisture in future studies.