Suspended solids and nutrient retention in two constructed wetlands as determined from continuous data recorded with sensor

Constructed wetlands (CWs) are recognized as a useful method to reduce sediment and nutrient loading from arable land to surface waters. Automatic monitoring systems with continuously measuring sensors were utilized for several years to obtain accurate, up-to-date information of the retention performance of two differently dimensioned and located CWs in agricultural catchments in southern Finland. Here, results derived from the data recorded by the sensors measuring height, turbidity and nitrate-nitrogen (NO3-N) concentration of the water entering and exiting the Hovi (established in 1998) and Rantamo-Seitteli (est. in 2009) CWs are presented. Study periods ended in autumn 2014 and they lasted 7 years at Hovi and 4.5 years at Rantamo-Seitteli. Annual total suspended solids (TSS), total phosphorus (TP) and total nitrogen (TN) per cent retentions were much higher in the older, more generously dimensioned (CW-to-catchment area ratio 5%) Hovi CW (TSS 74%, TP 58% and TN 54% on average during 7 years) than in the newer, more scantly dimensioned (CW-to-catchment area ratio 1.3%) Rantamo-Seitteli CW (TSS 7%, TP 12% and TN 9% on average during 4.5 years). Moreover, the annual retentions of the Rantamo-Seitteli CW appeared to decrease with increasing runoff, while the retentions at Hovi were not affected by runoff fluctuations. In terms of TSS and TP retentions in the Hovi CW, they were in this study at similar levels to those observed shortly after the establishment of the CW. Meanwhile, the retention of dissolved nutrient fractions, i.e. NO3-N and dissolved reactive P (DRP, based on water sampling) at Hovi were clearly higher in this than in the previous study. The results confirm the importance of CW dimensioning and locating near the sources of loading, and refer to improved performance as CWs get older with increasing vegetation and biological activity. Our experiences of CW measurements with sensors have been positive. The information obtained with new technology has provided retention estimates that are more accurate than they would have been if they were derived from sparse grab sampling.