Search

Dissolved organic matter (DOM) can be an important source of energy and nutrients in aquatic ecosystems and play important roles in carbon and nutrient dynamics of natural and impacted environments. In order to assess the effects of early stage fish farming on dissolved organic carbon (DOC) concentrations as well as DOM fluorescence indices, we took water samples along transects from six fish farms towards reference sites in a large tropical reservoir (Furnas Reservoir, Southeast Brazil) between April 2013 and December 2016. While DOC concentrations did no change along transects, small but significant changes in fluorescence indices were detectable in up to 100 m from fish farms. Higher fluorescence index (FI) values near fish farms pointed to small increases in microbial production due to fish farming. Only in the more pristine of the two main reservoir branches, small differences in the freshness index (β:α) and the peak T/peak C ratio along transects indicated that the relative contribution of recently produced DOM and the biochemical oxygen demand of DOM increased due to aquaculture. The humification index did not respond to fish farming. In summary, current early-stage fish farming did not appear to cause major impacts on DOM quantity and quality, but future increases in net cage and fish farm densities, as well as longer operation times may change this assessment. The fluorescence indices FI, β:α and peak T/peak C ratio may be useful early warning signals for monitoring fish farm impacts in tropical reservoirs.

Stream nutrient uptake is an important ecosystem service, because it supplies human needs such as water purification. Uncertainty in nutrient uptake measurement determines, in combination with effect magnitude, whether effects of land use, stream restoration or climate change can be detected by this method. However, measurement uncertainty for nutrient uptake metrics is rarely reported. Here, we compared whole-stream phosphate and nitrate uptake among three pristine and three agricultural tropical Cerrado streams. We quantified uncertainty in uptake metrics estimated by kinetic nutrient-addition experiments and evaluated their potential to detect land-use effects. Ambient phosphate and nitrate uptake lengths (SW) ranged from 11 to 106 m and from 29 to 357 m, respectively, in pristine streams and from 41 to 586 m and from 76 to 1372 m in agricultural streams. Moderate measurement uncertainty in SW allowed for the detection of land-use impacts with this metric. However, derived nutrient uptake metrics, such as ambient uptake rates and uptake velocities, as well as parameters of whole-stream saturation kinetics, were associated with high measurement uncertainties that prevented the detection of potential land-use effects. Sensitivity analyses suggested that avoiding high enrichment levels in kinetic addition experiments and increasing sampling effort for plateau nutrient and tracer concentration at the first and last sampling stations of investigated stream reaches are the most promising strategies to reduce measurement uncertainty. Analyzing nutrient uptake without considering measurement uncertainty might lead to poor interpretation in case studies and meta-analyses, such as incorrect evaluations of human impacts or comparisons among systems.

Organic matter decomposition and ecosystem metabolism have been suggested as promising functional indicators of stream ecosystem integrity, but tests of their applicability to tropical systems are rare. Here, we tested the suitability of the ecosystem processes leaf decomposition (LD), community respiration (CR) and gross primary production (GPP) as functional indicators of catchment and riparian corridor land use impacts on 16 rural, tropical headwater streams. Agricultural and urban land cover contributed 0.0–78.3% and 0.0–8.3%, respectively, to total catchment land cover, and 0.0–61.2% and 0.0–7.8%, respectively, to land cover in a riparian corridor of 20 m width of the investigated streams. Median rates of GPP and CR in the studied streams were 0.4 g O2 m−2 d−1 (0.0–1.6, min–max) and 3.3 g O2 m−2 d−1 (1.8–15), respectively. According to the double exponential model, median leaf litter decay coefficients of the recalcitrant and the labile litter fraction were 0.006 d−1 (−0.017 to 0.016) and 0.15 d−1 (0.06–0.86), respectively. While land use in the riparian corridor was correlated to the studied ecosystem processes, whole catchment land use did not exhibit such relations. Nitrogen enrichment and increased CR, both related to urban land use, were the main impacts, indicating that already low levels of urban land cover can have considerable effects on stream integrity in the investigated rural tropical landscape. Although spanning a much wider gradient, agricultural land cover was a less important stressor, and stimulated GPP was related to stream morphology, but not to nutrient enrichment. Leaf decomposition was negatively correlated with both the agricultural and the urban impacts, suggesting that high rates of leaf decomposition are an indicator of pristine conditions in the studied streams. In conclusion, measurements of LD, CR and GPP appear to be suitable, and complementary functional indicators of stream integrity and provided insights into mechanisms of land use impacts on streams in tropical rural catchments.