Exploring temporal and spatial variation in cotton tensile-strength loss to assess the ecosystem health of non-wadeable rivers

Cotton strip assays are becoming a widely utilised functional indicator to assess and interpret ecosystem health of non-wadeable rivers. Cotton decomposition may be sensitive to seasonal variation in key drivers, the aim of this study therefore, was to elucidate the temporal variation in cotton strip decomposition across a land-use gradient to underpin the utility of this technique as an indicator. Twelve sites, spanning a range of water quality, related to upstream land-use intensity were sampled monthly for 16 months. Mean cotton tensile-strength loss per day ranged from 0.8% to 13.2% per day across the study sites. Cotton tensile-strength loss was significantly related to mean water temperature, the concentration of dissolved nutrients, water clarity, and dissolved ions, in linear mixed models (all p < 0.005). Results concur with other research from headwater streams, larger rivers, and mesocosm experiments where the importance of temperature and nutrient concentrations as predictors of cotton decomposition have been highlighted. Significant relationships between cotton tensile-strength loss and both land-use stress and season were detected (F = 49.54, p < 0.01 and F = 3.78, p < 0.01), however there was no interaction between season and land-use stress (F = 1.51, p = 0.19). This suggests that despite the within-site seasonal fluctuations in physicochemical drivers of cotton decomposition that were observed in this study, the pattern across the impact gradient is relatively consistent across time (i.e. sites under greater land-use stress will have higher rates of cotton decomposition regardless of season). Despite intrinsic site-to-site variability, our results suggest that, after temperature, nutrient enrichment is likely to be an overriding factor in determining cotton decomposition. The relationship with land-use stress provides support for the use of cotton strip assays as a standardised method for assessing ecosystem health in non-wadeable rivers at a regional scale, however further reference benchmarks may be required for wider application and this provides an interesting avenue for further research. In addition, further research to determine the sensitivity of the method to changes in catchment management and river restoration is required.