Multi-scaled drivers of rural prairie stream metabolism along human activity gradients


  • This article is published with the permission of the Controller of HMSO and the Queen’s Printer for Scotland.

Adam. G. Yates, Department of Geography, University of Western Ontario, London, ON, Canada N6A 5C2.


1. Stream metabolism is increasingly used for monitoring and assessment of the biological condition of aquatic ecosystems. However, distal environmental drivers, such as land use, are typically not well connected to the proximate controls, such as stream chemistry, that are usually invoked as driving metabolism. This is particularly true for North American prairie streams and for grassland streams worldwide.

2. Stream metabolism was measured at the outflow of 19 subcatchments of the Red River in southern Manitoba, Canada. Subcatchments represented gradients of nutrient-producing human activities present in the region, that is, wastewater treatment (WWT), livestock production and crop cultivation. Stream metabolism was estimated at all sites using diel changes in dissolved oxygen (DO) concentration over 1 week in the middle of summer. Environmental parameters hypothesised to control stream metabolism were sampled across three spatial scales (stream reach, stream segment and catchment). Model selection using Akaike’s information criterion (AIC) was used to determine linkages between environmental parameters and measures of stream metabolism.

3. Estimated rates of metabolism were within the range of past studies of metabolism in prairie streams, although most streams had negative values of net ecosystem metabolism. However, production-to-respiration ratios were >0.5, at all but three sites suggesting that autochthonous production was an important source of organic matter.

4. The a priori model that best predicted gross primary production (GPP) was the intensity of nutrient-producing human activities (i.e. WWT, livestock and crop cultivation) measured at the catchment scale. Ecosystem respiration (ER) was best predicted by the a priori model comprised of GPP, total nitrogen (TN) and total phosphorus (TP). However, model averaging revealed that prediction of ER could be improved by including riparian cover and removing TP from the model. The positive association between GPP and ER suggested that heterotrophic compartments of the ecosystem were modest contributors to variation in respiration rates.

5. Overall, this study suggests that variation of metabolism in prairie streams of southern Manitoba is controlled by human activities occurring at the catchment scale, a finding consistent with current hierarchically structured riverine paradigms. Moreover, increased understanding of the hierarchical structure of stream metabolism drivers will help to ensure that assessment results can be used more effectively to inform management strategies for prairie ecosystems.