Factors affecting effluent quality in on-site wastewater treatment systems in the cold climates of Finland and Sweden

Decentralized sanitation facilities serving single households to small communities can offer a more flexible way of improving global sanitation. In Nordic countries, on-site wastewater treatment systems (OWTS) are mostly used in sparsely populated areas and are regulated for the removal of phosphorus (Tot-P), organic matter (biological oxygen demand, (BOD7)) and total nitrogen (Tot-N). However, available reports, albeit commonly using small data sets, call attention to often low and variable treatment performance. In this work, a statistical analysis of 1301 samples from 395 units reported in 10 studies from Finland and Sweden was conducted. The objective was to increase knowledge regarding the effluent water quality of OWTS and their compliance with regulations. In addition, the goal was to identify possible factors affecting contaminant concentrations. Overall, compliance with base-level BOD7 removal requirements was met by most units while 10% of units exceeded Tot -N and >25% exceeded Tot-P limits. Non-compliance rates were high, especially for Tot-P, considering the studied data set was composed mostly of semi-new units (median 3.8 years). Neither weather (air temperature, precipitation, and snowmelt), age or load factors were found to strongly explain the data variability in either soil -based systems (SBS) or package plants (PP). Tot-N and Tot-P effluent concentrations of the two systems were not significantly different. Effluent BOD7 concentrations were lower in SBS (median 3.0 mg/L) when compared to PP (median 7.7 mg/L). Units with P-sorbing filters (median 0.69 mg/L) and chemical precipitation (median 1.54 mg/L), presented lower Tot-P effluent concentrations in contrast to traditional sand filters (median 4.0 mg/L). The biological process used in PP had a measurable effect on effluent BOD7 concentrations, with attached growth units presenting higher values (median 8.1 mg/L) than suspended growth units (median 7.1 mg/L). Although the data set utilized was large, its analysis revealed intrinsic bias in data collection, especially regarding system types and sampling seasonality with colder winter periods sampling underrepresented. There is a need for data collection to improve to allow conclusive treatment efficiency assessment including possible effects of climate and process-related factors.