Epsilon Archive for Student Projects

Abstract

The current state of the world’s freshwater is threatened because of human perturbations. It is important to protect our streams and waterways and solutions are required. One solution is the concept of forested riparian buffer zones which, if designed carefully, can buffer harmful substances and reduce impacts from agricultural land use, give shade and detritus, as well as significantly increasing the overall biodiversity among macroinvertebrates and diatoms in the water body.
In this study, a data analysis of riparian data from the streams in lake Ekoln basin was conducted in order to examine which attributes of a riparian buffer zone are most important for improving ecological condition in agricultural streams. Linear modelling with backwards model selection was used to assess the effects of six predictor variables on 15 response variables and their relationships were presented through tables and graphs.
I found that the simple presence of a forest buffer, regardless of any particular property, reduces the amounts of fine sediment and is associated with increased biodiversity of macroinvertebrates, and increased scores for the macroinvertebrate monitoring index EPT and diatom IPS. The results further revealed that some buffer properties were more likely to be associated with stream ecosystem improvements than others, both in terms of number of responses affected and the strength of those relationships (based on R2).
Buffer width and buffer length had the highest average R2 values out of all the six predictor variables, and was significantly correlated with six and three response variables respectively. Tree species richness was associated with five relationships to the response variables, and had most effect on diatom IPS, but had a low average R2 value of 0,027. Tree cover density had only two relationships, but with very low R2 (<0,0001).
Riparian condition index had the third highest average R2 value and was strongly correlated with sediment and macroinvertebrates. But since riparian condition index is a multimetric variable, further investigation is needed for determining which attributes of the RCI that were the driving factors for sediment and macroinvertebrates.
Since all predictor variables showed relationships with at least two response variables, and reasonably no predictor variables act alone in contributing to an efficient buffer, it is important to highlight the combined importance of several variables. However, overall, these results indicate that focusing on aspects of buffer size (length and width) when designing forest buffers is important for maximizing stream ecosystem biodiversity.