Landscape Influences on Stream Habitats and Biological Assemblages

Landscape- and Reach-Scale Predictors of Large Wood Abundance in Low-Gradient Streams

Lucinda B. Johnson, George E. Host, Jennifer H. Olker, and Carl Richards

doi: https://doi.org/10.47886/9781888569766.ch8

Abstract.—Wood is an important component of small to medium-sized streams in forested regions, but has been poorly studied in agricultural areas. Our goals were to (1) characterize the abundance, size, and distribution of wood in low-gradient streams in two agricultural regions, (2) quantify the influence of reach- and landscape-scale factors on the abundance and distribution of wood in these streams, and (3) compare trends across two study areas. Wood abundance was quantified in stream reaches in two diverse agricultural regions of the Midwestern United States: central Michigan and southeastern Minnesota. Wood abundance was quantified in 71 stream reaches, and an array of channel, riparian zone, and landscape features were characterized. Multiple regressions were conducted to predict abundance from those explanatory variables. We found that large wood was relatively scarce in these low-gradient streams compared to low-gradient streams in forested regions. Mean log size was greater, but total abundance was lower in Minnesota than Michigan. In Minnesota, greatest wood abundance and greatest extent of accumulations were predicted in wide, shallow stream channels with high substrate heterogeneity and woody riparian vegetation overhanging the channel. Models were dominated by reach-scale variables. In Michigan, largest densities of wood and accumulations were associated with catchments in hilly regions containing urban centers, with low soil water capacity, wide, shallow stream channels, low coarse particular organic matter standing stocks, and woody riparian zones. Models contained both reach- and landscape-scale variables. Difference in the extent of agricultural and forest land use/cover between Michigan and Minnesota may explain the differences in the models predicting wood variables. Patterns in wood abundance and distribution in these Midwestern streams differ from those observed in high gradient regions, and in low-gradient streams within forested regions. This has important implications for ecosystem processes and management of headwater streams in agricultural regions.