Development and application of methods to estimate biofilm activity in recirculating aquaculture systems

Recirculating aquaculture system (RAS) is an innovative and sustainable way to farm fish and other aquatic species with reduced environmental impacts and is posed to play a vital role in improving global food security. Good water quality is crucial for fish health and growth and is achieved in RAS by the actions of several water treatment installations, including biofilters. Biofilters perform nitrification, converting ammonia nitrogen to nitrite and nitrate in a process collectively referred to as nitrification. Nitrification is conducted by biofilm attached to the surface of biofilter elements. Biofilm consists of an aggregate of diverse microorganisms encapsulated in their self-produced matrix, whose activities involving different microbial processes within biofilters are essential for maintaining proper water quality in RAS. Measurements of biofilm activity help monitor microbial processes, assess process efficiency, track biofilm development, evaluate biofilm response to changing environmental conditions, and optimize biofilter performance. However, methods to measure biofilm activity in RAS biofilters are still limited, and the current methods available either require extensive time and effort or a high expertise level, making early detection of biofilter anomalies and on-site application impossible. Consequently, there is an urgent need to develop a simple, fast, and practical method to measure biofilm activity in RAS biofilters.

The main aim of this Ph.D. thesis was to develop new methods to estimate biofilm activity in RAS biofilters. Subsequently, the developed methods were applied to assess biofilm activity changes associated with peracetic acid (PAA) disinfection in RAS. The first two studies focused on developing new methods to estimate biofilm activity based on respirometry via measuring oxygen changes. The other two studies involved using new methods to estimate biofilm response under acute PAA exposure and during chronic PAA disinfection in pilot-scale freshwater RAS (FW-RAS), respectively.

A modified respirometric method to estimate biofilm activity in RAS moving bed biofilters (MBB) by measuring oxygen consumption rates (OCR) was demonstrated (Paper I). The method relied on real-time monitoring of dissolved oxygen (DO) concentrations in a sealed chamber equipped with an oxygen sensor, in which a few biofilter elements were fed with different standard substrates (pure tap water, and with ammonium, nitrite and acetate) at a fixed temperature. The OCR from these measurements can be used as a proxy for biofilm activity, making it possible to distinguish between endogenous respiration, ammonia oxidation, nitrite oxidation, and acetate oxidation processes, respectively. The biofilm activities in three different types of biofilter elements were measured: EPP (extruded polypropylene), IMPP (injection molded polypropylene), and PF (polymeric foam), using respirometric and conventional chemical tracking methods. The biofilter elements tested were collected from MBB treating effluent from the same pilot-scale FW-RAS rearing rainbow trout (Oncorhynchus mykiss). The results showed that specific OCR profiles in the absence of substrates and the presence of nitrite, ammonium, and acetate were observed when using the respirometric method to estimate biofilm activity in EPP, IMPP, and PF, respectively. The highest biofilm activities related to the oxidation of nitrite, ammonium, and acetate processes were found in PF with volumetric oxygen consumption rates of 677±125, 764±156, and 166±36 g O₂·m^-3·d^-1, respectively. The respirometric method matched well with the conventional method to estimate biofilm activity for three kinds of biofilter elements tested, confirming the viability of the respirometric method to estimate biofilm activity in RAS biofilters.

Second study (Paper II) developed and verified a novel method to estimate biofilm activity related to the enzymatic process for hydrogen peroxide (H₂O₂) decomposition. The biofilm activity was quantified by measuring the net oxygen release rate (k_or) via real-time monitoring of DO concentrations in a sealed respirometric chamber developed in Paper I, in which a few biofilter elements were fed with H₂O₂ (C₀=10 mg/L) at a fixed temperature. The k_or serves as a proxy for microbial catalase activity in biofilm-driven H₂O₂ decomposition process. Different numbers (0, 2, 4, 8, and 16) of mature biofilter elements (EPP) from an MBB in a pilot-scale FW-RAS for rearing rainbow trout (Oncorhynchus mykiss) were tested in respirometric chambers with repeated 10 mg/L H₂O₂ exposure for one hour and compared with their autoclaved forms. Results showed a large amount of DO rise (0.92-2.31 mg O₂/L) was observed in the chambers packed with mature elements, and only a small amount of DO increase (≤ 0.27 mg O₂/L) was found in the chambers holding autoclave elements. This indicated that the biofilm-driving 10 mg/L H₂O₂ decomposition process was mainly governed by microbial enzymatic activity in biofilm. The DO release profiles in the chambers with mature bioelements were well fitted by a monomolecular model (R² >0.98), and the deduced k_or (h^-1) was proportional to the number of mature bioelements tested (R²=0.99 for linear fit). Repeated H₂O₂ exposure to the mature elements showed similar k_or, which substantiates that one-hour exposure of 10 mg/L H₂O₂ to biofilm did not inhibit microbial enzymatic activity in biofilm.

The developed respirometric methods from Paper I &II were applied in a third study (Paper III) to evaluate biofilm response to acute PAA exposure. The biofilm samples tested were collected from MBB at two different FW-RAS facilities (pilot-scale RAS vs. commercial fish farm). They were exposed to increasing PAA concentrations (0, 1, 2, 4, 8, and 16 mg/L) under standard conditions. The two facilities used the same type of biofilter elements (RK element) but adopted different disinfection procedures in daily operation, in which repetitive PAA was used in commercial fish farms for seven years and no PAA was used in pilot-scale RAS. The activity and viability of biofilm samples after exposure were measured by respirometric methods and flow cytometry, respectively. Results showed that PAA led to a clear dose-dependent inhibition of biofilm metabolic activity in bioelements from pilot-scale RAS, in which IC 50 values (the concentration of PAA caused 50% inhibition of biofilm metabolic activity) for nitrite oxidation, ammonia oxidation, endogenous respiration, and enzymatic H₂O₂ decomposition processes in the biofilm were 1.27, 1.59, 2.67 and 4.68 mg PAA/L, respectively. Also, the intact cell counts and the fraction of dead cells in biofilm (viability indicators) declined and increased along with increasing PAA concentrations, respectively. In contrast, the biofilm activity and viability in bioelements from the commercial fish farm were only suppressed at 16 mg PAA/L exposure, resulting in a significant reduction of nitrite oxidation and ammonia oxidation activities by 38.8% and 50.6%, respectively, and a significant increase of the proportion of dead cells by 6.5% in biofilm compared to control without PAA exposure. The results suggested that the inhibitory effects of PAA on biofilm were related to suppressing metabolic activities and the loss of cell viability in biofilm, and PAA-adapted biofilm (i.e., biofilm samples from commercial fish farm) may enhance its resistance to PAA disinfection.

In the final study (Paper IV), the developed respirometric methods from papers I &II were applied to evaluate biofilm activity changes during chronic PAA disinfection. The three disinfection treatments: i) continuous ozone disinfection at a dose of 0.06 mg/L Cl₂ equivalent; ii) continuous PAA disinfection at a dose of 1 mg PAA/L; and iii) control without disinfection, were tested in triplicate and applied in nine identical experimental FW-RAS stocking with Atlantic salmon parr (Salmo salar) during a four-week trial. The activities of biofilm that were attached on the surface of bioelements in MBB from three treatment groups were measured with developed respirometric methods via determining OCR and k_or. Results showed that PAA disinfection increased the endogenous respiration activity and the enzymatic activity for H₂O₂ decomposition in biofilm by 39-133% and 135%, respectively, and partially inhibited the nitrification activity (drop by 36%) in biofilm, causing a significant nitrite accumulation (rise by 45-74%) in system water, compared to control treatments during a four week trial. Ozone disinfection only resulted in an increased endogenous respiration activity of biofilm at the beginning of the trial, and the endogenous respiration activity was reduced to control levels at the end of the trial. These results indicated that chronic exposure to 1 mg PAA/L can alter biofilm metabolic status in biofilter elements, which may cause adverse effects on biofilter performance, while chronic exposure to ozone of 0.06 mg/L Cl₂ equivalent improved water clarity without compromising biofilm metabolic state in biofilter elements.

In conclusion, the respirometric methods can estimate biofilm activities involving metabolic processes of endogenous respiration, ammonia oxidation, nitrite oxidation, heterotrophic oxidation, and enzymatic H₂O₂ decomposition within RAS biofilters by ex-situ monitoring dissolved oxygen changes (i.e., oxygen consumption and production) in respirometric chambers following standard substrate feeding (Studies I-II). Although requiring some specialized equipment, the methods are relatively simple and fast, require a small sample volume, do not destroy biofilm integrity, and can be used as an on-site tool to monitor and diagnose biofilm in RAS biofilter elements. The potential applications of the methods include evaluating the colonization status of biofilter elements during start-up, estimating biofilm activity alterations associated with changes in biotic and abiotic conditions, such as pathogen invasion, salinity shifts, chemical disinfection (Studies III-IV), feed loading, and water flow rate. Regarding PAA disinfection in RAS, PAA can cause biofilm metabolic status changes, which can be beneficial and detrimental depending on the PAA product (PAA/H₂O₂ ratio) applied, PAA dosage, PAA application strategy (chronic vs. acute) and biofilm state (resistance). The targeted dosage of PAA entering biofilter units should be < 1 mg PAA/L to minimize the adverse effects of PAA on biofilter nitrification performance when applying continuous PAA disinfection in RAS.