Characteristics of hummocked and non-hummocked Colorado riparian areas and wetlands

Abstract

Several hummock formation theories exist in the scientific literature, but are rarely mentioned in management publications where only one or two other theories are stressed. Our objectives were to identify and describe the edaphic, climatic, topographic and vegetation characteristics of riparian areas and wetlands in Colorado that do and do not support hummocks, and to classify hummocked sites according to hummock characteristics. 25 natural resource professionals throughout Colorado identified ten sites each that were a mix of hummocked and non-hummocked in proportion to their local abundance, from which 40 hummocked and 40 non-hummocked sites were randomly selected for sampling. Plant functional group canopy cover was determined using Daubenmire cover classes. Soil samples were collected and numerous site characteristics were recorded. Hummock size, shape and density were measured at hummocked sites. Forward model selection and multiple logistic regression were used to determine relationships between site characteristics and odds of hummock occurrence. Mean winter precipitation, mean annual temperature and forb cover were negatively related to odds of hummock occurrence while soil silt content and plant species richness were positively related to odds of hummock occurrence. Cluster analysis resulted in three groups of hummocked sites with differing hummock morphology, vegetation and climatic characteristics. Differential frost heave and plant biomass accumulation were likely involved in hummock formation at sites examined for this study, but conditions expected to occur following cryoexpulsion of clasts were not observed.