Search

A model of the denitrification on a polyhydroxyalkanoate (PHA) based reactor for recirculating aquaculture was developed. PHA is a family of non-water soluble bioplastics produced by bacteria. The PHA formulation used in this work was polyhydroxybutyrate (PHB). The model considered nitrate concentration, dissolved oxygen, organic carbon and biomass concentration as the most significant variables. The developed model represents adequately the nitrate reduction with the medium used, for nitrate under 100 ppm NO 3 − -N. In the conditions tested, an average ratio of 2.92 g PHA to 1 g NO 3 − -N reduced was found. The model results showed a denitrification rate of 2.97 kg m −3 d −1 for ranges from 10 to 50 mg NO 3 − -N L −1 . Using this model as a management tool, the required size of denitrification units and PHA recharging time can be predicted based on the expected nitrate loading and the time between PHA recharges desired. The unit sizing should be done for the maximum load expected. The slow rate and the energy required for PHA hydrolysis, make it unavailable as electron donor after the nitrate is consumed, so it will not promote the formation of sulfides. The model can be modified for other biodegradable non-water soluble medium by changing the hydrolysis constant, which must be determined experimentally.

The effect of salinity, dissolved oxygen and NO3-N concentration on the denitrification of recirculating aquaculture water using polyhydroxybutyrate (PHB) was evaluated. Four PHB media with different molecular weights and configurations were tested. The results show that at higher nitrate concentrations in the influent water, the consumed PHB:NO3-N ratio decreased. An average of 2.9 g of PHB:1 g NO3-N removed at temperatures of 20.8 ± 1.1 °C was measured. Although the molecular weight showed an apparent correlation with the denitrification rates, the correlation was not statistically significant. A moderately biofouled granular media displays a heterogeneity of microenvironments that allow some denitrification to occur in the presence of bulk dissolved oxygen levels approaching 5 mg L−1. As a practical approach, the inhibitory effects of oxygen can be mitigated either by design of the denitrification media bed and/or by control or reduction of the influent dissolved oxygen levels. The high plastic consumption needed for oxygen removal indicates that the second approach is more cost efficient. At a flux of 60 m3 m−2 d−1 the denitrification rate decreases at a constant rate in the first 30 cm of the PHB bed. Below this depth, the denitrification rate decreases very slowly and stays above 1 kg-NO3-N m−3 d−1. In a pragmatic sense, denitrification abilities can be expected to be similar in all salinities. Volumetric nitrate removal rates in the order of 2.5 kg-NO3-N m−3 media d−1 should be broadly obtained in fresh and marine water systems. In the range of up to 250 mg NO3-N L−1, the PHB can be used as a base for a passive denitrification unit that requires little management. The availability of an economic source of PHB such as production waste and the development of the bioplastic industry is determinant for the adoption of this material as a carbon source for denitrification processes.

An estimated 28% of the production cost in HISTAR systems that are artificially illuminated is attributed to the lighting cost. This cost estimated is based on an operational configuration comprised of eight CFSTRs, a system dilution rate ( D s ) of 0.640 d −1 , and 400 W metal halide lamps positioned at an elevation of 38.1 cm over the culture. Deterministic model simulations of the volumetric productivity ( P v ), photosynthetic efficiency ( E o ) and lighting cost (LC) under various management strategies, operational parameters and reactor design configurations were performed and compared to the simulation results obtained for the original configuration. The simulations showed that LC may be reduced by 35.5% by switching from a metal halide (MH) to high-pressure sodium (HPS) light source at an optimum system dilution rate D s = 0.641 d −1 . LC may be reduced by an additional 17.8% through decreasing the lamp elevation to 25.4 cm. Increasing the wattage of the light source from 400 to 1000 W in the last six reactors would reduce the LC by 13% from the original cost. Overall, using HPS lamps at 25.4 cm height, with six 1000 W and two 400 W lamps at a D s = 0.641 d −1 will result in a 54% overall LC reduction compared to the original configuration of HISTAR. This represents a 13% reduction in the overall microalgal production cost for HISTAR.

A deterministic model was developed to predict microalgal productivity within the Hydraulically Integrated Serial Turbidostat Algal Reactor (HISTAR). HISTAR consists of two turbidostats, which concomitantly inoculate the first of a series of CFSTRs. The CFSTRs function as a biomass amplifier of the inoculated culture. The model included an analysis of the internal light dynamics within each CFSTR and the effect of the instantaneous average irradiance ( I a n (PAR)) on the instantaneous growth rate ( μ n ) within CFSTR n . The latter effect was accomplished using Steele's equation. The instantaneous average irradiance was determined by integrating the Lambert–Beer Law over the depth of the culture within the CFSTRs. Fourier series analysis was used to model the biorhythms of the microalgae. The model was calibrated for biomass using four data sets collected at system dilution rates ( D s ) of 0.265 day −1 , 0.385 day −1 , 0.641 day −1 and 1.127 day −1 . The ability of the calibrated model to simulate daily volumetric productivity ( P v ) within HISTAR was determined through calculation of the percent standard error of prediction. The overall mean for the four data sets was 24.8%. The average predicted productivity for the data sets was 24.2 g m −3 day −1 (16.2 g m −2 day −1 ) and the average actual productivity of the data sets were 25.5 g m −3 day −1 (19.9 g m −2 day −1 ), resulting in only a 5.1% error. Simulations produced by the calibrated model were used to estimate the optimum D s (between 0.641 day −1 and 0.884 day −1 ). The model predicted a P v of 70.2 g m −3 day −1 ( P a = 46.8 g m −2 day −1 ) at optimum D s . The maximum P a observed in the model simulations was 39.9 g m −2 day −1 .

Factors such as limitations in water quality and quantity, cost of land, limitations on water discharges, environmental impacts and diseases, are driving the aquaculture industry toward more intensive practices. This will force producers to adopt environmentally friendlier technologies. Recirculating systems, with a biofilter as the most prominent characteristic, treat internally the water contaminated with dissolved organics and ammonia and reduce the amount of water use and discharge from aquaculture operations. This paper reviews the implications of the changing use of recirculating aquaculture systems (RAS) on biofiltration research for freshwater and marine operations. Demand for cost effective biofilters will increase with the expansion of recirculating systems, both as a complement and replacement of traditional ponds. For freshwater aquaculture, emphasis should be placed in cost competitiveness, low head operations, intensification of ponds with RAS biofiltration and the evaluation of suspended growth systems. In the marine systems, an increase in demand of oligotrophic and ultraoligotrophic systems is expected, particularly in the nursery systems. Sizing and cost efficiency of biofilters for nursery operations should be addressed. Problems in marine biofilter acclimation appear to justify the development of new acclimation procedures. Biosecurity concerns, land cost and storm threats will drive nursery systems inland, where saltwater supply and disposal will force an increased water reuse. Denitrification strategies will need to be redefined and optimized for the marine nursery environment.