Epsilon Archive for Student Projects

Abstract

Large areas in the Baltic Sea suffer from eutrophication, due to excessive input of nutrients, mostly nitrogen and phosphorus, from surrounding coasts. The transfer of phosphorus is mostly riverine. Where the concentration of phosphorus bound to particles is highly variable over time, with short time events of high flow possibly having significant impact on the total flux. Monitoring of riverine phosphorus is regularly done by collecting monthly grab samples, with the risk of potentially missing high concentration events caused by momentarily high flow. Due to these risks there is a common use of proxies. Where transfer functions are used to calculate the concentration of phosphorous, to help fill in the gaps between catchments. The aim of this study was to evaluate if high frequency measurements of turbidity could be used as a proxy, in order to improve the flux estimations of phosphorous in rivers. The study was conducted in Kilaån, a river located in Nyköping, Sweden. In Kilaån grab samples are taken weekly, and turbidity measurements are done continually, every 15 minutes, using a sensor installed in the river. By analyzing the river data, a local relationship between turbidity and tot- P could be evaluated. The data from Kilaån was also used in different comparisons to evaluate how frequency of the grab samples affect the transfer estimations, and if these estimations changed when high frequency data was used, as a complement to grab samples. Linear regression showed a positive significant relationship between the two parameters, turbidity and tot-P, p < 0,0001, R2 = 0,661 n=52. The equation from this relationship was then used in order to estimate daily concentrations of phosphorus. When comparing the modelled results, to grab samples linked by linear interpolation, it showed big differences in concentration depending on frequency of the grab samples. Where almost all high concentration peaks during the first period of the study, where missed by monthly grab samples. These peaks where shown to be compensated for, if sensor data was used to fill in the gaps between the grab samples. The total flux estimated with the use of high frequency data to complement the grab samples, was then compared to conventional methods of flux calculations. Showing a, non-significant, tendency of higher flux values when turbidity measurements was used in the calculations. Further on it also showed that monthly grab samples underestimated the annual flux, compared to weekly samples. In conclusion, the results showed that high frequency measurements with sensors in combination with monthly grab samples can generate representative flux estimations. Furthermore, the sensor data can contribute to a better understanding of short time fluctuation and variation of riverine phosphorus concentration.