Improving denitrification in an aquaculture wetland using fish waste - a case study

Abstract

Cost-efficient, end-of-pipe, nitrate removal techniques are called for by the commercial aquaculture industry. This case study examined how simple flow manipulations improved the denitrification performance of a 19,007 m² (13,305 m³) constructed, free water surface (FWS) wetland treating aquaculture effluent. The wetland consisted of two separate streams with a common outlet: one stream treating nitrate-rich but carbon deficient effluent from the production unit at a hydraulic retention time (HRT) of 1.5 days (wetland stream 1); and a second stream treating carbon-rich, fish sludge-based effluent at a HRT of 41.0 days (wetland stream 2). During the course of the study (May–July 2017), three increasing proportions (40, 49 and 56%) of nitrate-rich effluent were re-directed from wetland stream 1 to the sludge-fed wetland stream 2 aiming at improving heterotrophic denitrification conditions in wetland stream 2 and consequently nitrogen removal in the wetland as a whole. Inlet C/N ratio in wetland stream 2 decreased from 1086 ± 57 to an average of 234 ± 56 (p < .05), and the area-based, total nitrogen (TN) removal rate in this wetland section increased significantly from 0.1 ± 0.01 to 8.4 ± 1.4 g/m²/d at the highest manipulated flow. In comparison, the flow manipulations had no effect on TN removal rates in wetland stream 1 averaging 1.4 ± 0.2 g/m²/d throughout the study. For the wetland as a whole, the TN removal rate increased from 1.4 ± 0.2 to 3.9 ± 0.8 g TN/m²/d. The flow manipulations furthermore improved the removal rates of total phosphorous and dissolved organic matter in the wetland as a whole. The study demonstrates that denitrification in a constructed aquaculture wetland may be improved by combining sludge-based and nitrogen-rich effluents in right proportions and leading it through an anoxic section of the wetland.

Introduction

Cost-efficient, end-of-pipe techniques for removing nitrate from recirculating aquaculture system (RAS) effluent are required to reduce nitrogen discharge from aquaculture farms. Constructed wetlands are man-made treatment systems successfully applied in many parts of the world for treating municipal, agricultural, and industrial wastewater (Verhoeven and Meuleman, 1999). They are mechanically simple but biologically complex systems. Nutrient / pollutant removal happens in a rather “uncontrolled” manner based on biological and biochemical processes affected by interactions between plants, microorganisms, and the sediment (Kadlec and Knight, 1996).

Where space allows, constructed wetlands may also be established for treating dilute aquaculture effluent. To improve effectiveness of such engineered systems, underlying factors governing nutrient removal processes need to be understood and ultimately controlled. Several studies have investigated the efficiency of subsurface- and horizontal-flow constructed wetlands treating aquaculture effluent (e.g., Schulz et al., 2003; Schulz et al., 2004; Sindilariu et al., 2007; Sindilariu et al., 2008; Sindilariu et al., 2009; Dalsgaard et al., 2018). Schulz et al. (2004) investigated the effect of hydraulic retention time (HRT) in three parallel, 350 m² free water surface (FWS) wetlands treating effluent from a flow-through rainbow trout (Oncorhynchus mykiss) farm. Mean total nitrogen removal rates of 0.45, 0.71, and 0.82 g/m²/d at HRTs of 3.5, 5.5, and 11 h, respectively, were found, and the authors concluded that a shorter HRT favored aerobic conditions and hindered denitrification. Dalsgaard et al. (2018) investigated longitudinal and seasonal nutrient removal rates in a FWS wetland treating effluent from a Danish Model Trout Farm (MTF) type I (Jokumsen and Svendsen, 2010). The wetland was characterized by a relatively short retention time (7.7 h), and a miniscule net removal of nitrate-N was observed in spring (February–June) with area-based removal rate constants (k_A) averaging 0.1 ± 0.1 m/d. This was succeeded by a small net production of nitrate-N from July to January, and it was argued that denitrification was limited by high oxygen concentrations and low organic carbon availability.

In comparison, a two-year monitoring study in eight constructed wetlands associated with Danish MTFs type III found average removal rates of 1.5–1.9 g NO₃-N/m²/d (Svendsen et al., 2008). All eight wetlands were characterized by relatively long HRTs (20–50 h), and farm effluent concentrations were higher in organic carbon and nitrate-N than in the wetland study by Dalsgaard et al. (2018) (approximately 10.0 versus 3.5 mg total BOD₅/l, and 6.6 versus 3.4 mg NO₃-N/l).

These average removal rates are, however, still relatively low compared to other denitrification technologies applied in aquaculture (van Rijn et al., 2006) and the current, more or less passive, operation of aquaculture wetlands prevents many Danish farmers from increasing their production because of exceeding their discharge allowance (Danish Ministry of the Environment, 2012).

A study by Suhr et al. (2013) demonstrated that RAS sludge may be used as a cost-free carbon source in a single-sludge denitrification process aimed at removing nitrogen from fish farm effluent. By mixing nitrate-rich effluent from a production unit with carbon-rich sludge in a heterotrophic denitrification reactor, nitrogen was removed at a maximum rate of 124.8 ± 15.7 g NO₃-N/m³ reactor/d. The study highlighted that C/N ratios and HRT are essential parameters with respect to maximizing nitrogen removal in this type of setup.

As an alternative (or in addition) to using a separate, rather complex filter set-up as the one by Suhr et al. (2013), we hypothesized that the sludge denitrification process may be exploited within the basins of an existing constructed wetland. To examine this, an 11-week field study was carried out with the aim of improving overall nitrogen removal in a constructed FWS wetland treating effluent from a Danish Model Trout Farm type III. The objective of the study was to mix internally generated, carbon-rich sludge effluent with increasing proportions of nitrate-rich effluent in a wetland side stream and measure the effect on overall nitrogen removal rates.