Novel Assay for Attached Filamentous Algae Productivity and Nutrient Removal

Filamentous algae nutrient scrubbers (FANS) have demonstrated potential for cost-effective and sustainable nutrient bioremediation of a wide range of wastewaters. Typically, FANS are seeded with a mixed assemblage of algae species, however, growing a monoculture of one species on FANS could facilitate biomass use by providing a more consistent and high-quality substrate for end-product applications. To date, a standardised bioassay to assess the productivity and nutrient removal of filamentous algae attached to a bottom substrate (that could help identify promising species for FANS monoculture) has not been developed. Therefore, we developed a microscale filamentous algae nutrient scrubber (µFANS) and a protocol to establish monocultures of freshwater filamentous algae to compare performance in terms of attachment capability, nutrient removal and biomass production. Four common filamentous algae species (Cladophora sp., Oedogonium sp., Rhizoclonium sp. and Spirogyra sp.) were seeded by evenly distributing and rubbing the biomass onto µFANS textured liner to “hook” algal filaments, providing initial physical attachment. Within 14 days, a “lawn” of the seeded algae had established and the “hooked” biomass had attached biologically. Depending on species, biological attachment resulted from either holdfast development from filaments that grew from settled zoospores, growth of rhizoids or adhesion of filament fragments to mucilage. Biomass productivity of each species ranged from 2.2 to 5.3 g DW m−2 day−1 while nutrient removal rates ranged from 8.8 to 28.4 mg NO3 g−1 DW day−1 and 2.2 to 8.1 mg PO4 g−1 DW day−1. Oedogonium sp. was the best performing species overall, with the strongest holdfast attachment, high biomass productivity (mean 4.2 g DW m−2 day−1) and high nutrient removal rates (mean 21.8 mg NO3 g−1 DW day−1; 5.6 mg PO4 g−1 DW day−1).

FANS typically grow a mixed-species assemblage of filamentous algae (Mulbry et al. 2008a;Adey et al. 2013;D'Aiuto et al. 2015;Sutherland et al. 2020). The algae are established by either letting them naturally colonise the floway or seeding the floway surface with mixed algal species collected from nearby water bodies. The resulting uncontrolled species composition on the floway can lead to variability in algal productivity and the composition and characteristics of the harvested algal biomass. An alternative approach is to seed a single filamentous algal species onto the floway. While the growth of a single algal species may not be required for effective nutrient removal (Hannon et al. 2010), it can provide a consistent source of high-quality biomass with low variation in biochemical composition for use in a range of end-product applications (Lawton et al. 2013). Such an approach has worked well in suspension (free-floating) algal systems (Cole et al. 2016;Mata et al. 2016) and therefore, may work well in attached algal systems such as FANS.
As FANS are an attached algae treatment system, it is essential to select a target filamentous algae species for seeding onto the floway that naturally grows attached to a substrate (Sutherland and Craggs 2017). Strong algae attachment enables the algae to tolerate higher water turbulence (Dodds 1991), making the algae more resilient to breaking and sloughing off the floways during increased horizontal water velocity resulting from runoff during heavy rainfall events (Craggs et al. 1996;Adey et al. 2011;Sutherland and Craggs 2017). Filamentous algae can initially be attached onto floways during seeding by physically hooking the filaments onto the textured bottom liner or mesh. Biological attachment by a holdfast can occur by two mechanisms, rhizoid formation and zoospore formation and settling.
Algal filament fragments that become physically hooked onto substrates can subsequently develop anchoring rhizoids that become holdfasts (Etherington 1964;Fletcher and Callow 1992). As the filaments grow, attachment is provided by the growth of rhizoids, increasing the surface contact between the algae and the substrate, thereby forming a stronger attachment (Fletcher and Callow 1992). The development of rhizoids is highly dependent on the substrate topography and light source and could also be associated with environmental conditions such as temperature and water pH (Nagata 1973(Nagata , 1977Fletcher and Callow 1992). In general, rhizoid growth is promoted by increased surface roughness, moderate temperature (~ 20 °C), and neutral pH (Nagata 1973(Nagata , 1977Fletcher and Callow 1992). High or low pH can affect rhizoid formation and an extremely low temperature can inhibit rhizoid growth, although this is dependent on the algal species (Nagata 1973(Nagata , 1977Fletcher and Callow 1992).
Zoospores are released by the asexual reproduction of cells of the parent filament, these subsequently settle and attach to a substrate by developing a holdfast from which a new filament grows (Rosemarin 1985;De Wreede and Klinger 1988;Rindi 2010). Depending on species, induction of zoospore production is primarily associated with environmental factors such as photoperiod, light irradiance, temperature, and to a lesser degree nutrient concentration or biotic interactions (Maggs and Callow 2003). Zoospore settlement on a surface is dependent on several factors, including the roughness and topography of the substratum, biofilm chemical signals, and attractive surface properties mediated by adsorbed macromolecules, microbial communities and hydrodynamics (Amsler et al. 1999;Maggs and Callow 2003). Roughness and topography of a surface appear to play a critical role in zoospore attachment, leading to the general view that filaments develop more easily on roughened surfaces as these enhance survival of the newly attached spores from dislodgement by water currents Maggs and Callow 2003).
In order to select target macroalgal species for use in FANS, a bioassay is required to provide a rapid and rigorous comparison of species performance in terms of attachment capability, biomass productivity and nutrient uptake rate. Bioassays using filamentous algae typically have been conducted in suspension cultures in a similar way to studies with microalgae. For example, Lawton et al. (2013) compared biomass productivity and biochemical composition of three species of freshwater filamentous algae in 20 L semicontinuous cultures under a range of conditions, including aeration rate and CO 2 addition. Lawton et al. (2014) also used suspension culture assays to compare the growth of isolates of Oedogonium collected from multiple geographic locations under a range of temperature treatments. Similarly, Valero-Rodriguez et al. (2020) compared the productivity and competitive ability of three filamentous macroalgae species in 1 L semi-continuous cultures with continuous aeration under seasonal light and temperature conditions. However, to provide a realistic indication of species performance and suitability for use in FANS, a bioassay must compare the performance of algae growing attached to a bottom substrate under conditions comparable to those on a FANS. To date, few studies have compared the performance of species of filamentous algae when grown attached to a substrate. De Vries et al. (1983) developed a microscope slide bioassay to test for nitrogen and phosphorus limitation in eleven strains of filamentous algae genus Stigeoclonium. However, this bioassay is not suitable for use as a FANS bioassay as the cultivation conditions used are different from those that the algae would experience when grown on a FANS. Although the algae were grown attached, they were maintained in still water with periodic water exchanges (e.g., static batch cultures). In contrast, a FANS bioassay must have continuous flow of nutrient-enriched water over the attached algae on the floway. Additionally, the De Vries et al. (1983) bioassay involved the induction and settlement of spores, followed by a 3-week growth period before the algal filaments had reached sufficient size for testing in the bioassay. Consequently, the use of this assay would require the development of protocols to induce zoospore release for each algal species and a considerable time lag from seeding until the start of the bioassay. Therefore, the overall aim of this study was to develop a reproducible bioassay to rapidly assess the suitability of different species of filamentous algae for FANS based on their ability to attach and grow, remove nutrients and produce harvestable algal biomass. The specific aims were to (i) determine if a single filamentous algae species can be rapidly seeded onto microscale filamentous algae nutrient scrubbers (µFANS) to develop a stable monoculture of uniform biomass composition, and (ii) to compare the biomass productivity and nutrient removal rates of µFANS monocultures of four common filamentous freshwater algae species (Cladophora sp., Oedogonium sp., Rhizoclonium sp. and Spirogyra sp). The bioassay developed here can enable the suitability of locally isolated species for FANS cultivation to be simultaneously assessed with a high number of replicates.

Microscale Filamentous Algae Nutrient Scrubber (µFANS)
We developed µFANS to provide a small-scale approximation of how filamentous algae grow on large-scale FANS (e.g., attached to a bottom substrate floway with nutrientenriched water continuously flowing over and through the attached algae) to enable the performance of different algal species to be tested under relevant cultivation conditions. The µFANS system consisted of a rectangular polypropylene (PP) plastic container (size: 15 × 24 × 5.6 cm) with high density polyethylene (HDPE) textured liner (size: 19.5 × 11 cm or 0.0215 m 2 ). This liner was chosen as studies have shown that textured substratum or surfaces with variable topography can enhance algal biomass attachment (Maggs and Callow 2003;Blersch et al. 2017). The textured liner was permanently attached to the base of the container using a multi-purpose permanent elastic sealant/adhesive glue (Bostik Simson ISR 70-03) (Fig. 1). The lined µFANS rested on the top of a water storage container that provided a sump with 2 L of total working volume. A submersible water pump continuously circulated nutrient-enriched water from the sump to a weir at the top of the µFANS and treated water drained back into the sump through holes at the bottom end of the µFANS. The pump tubing included an adjustable plastic valve to enable the water flow rate to be controlled.

µFANS Seeding
Four locally occurring filamentous algae species (Cladophora sp., Rhizoclonium sp., Oedogonium sp., and Spirogyra sp.; Fig. 2) were compared in this study. These species were selected for cultivation on FANS as they are widely distributed in New Zealand and worldwide and previously have shown promise for nutrient bioremediation (Saunders et al. 2012;Khataee et al. 2013;Lawton et al. 2017;Liu et al. 2020;Valero-Rodriguez et al. 2020). In addition, using common locally occurring algae species in a treatment system is beneficial since these algae are already adapted to local conditions and will be naturally seeded in the system (Grobbelaar 2010; Wilkie et al. 2011). These four filamentous algae species were collected from natural water bodies, ponds, and agricultural drainage systems close to Ruakura, Hamilton, Aotearoa New Zealand. Algae samples were transported back to the National Institute of Water and Atmospheric Research (NIWA) research facility at Ruakura, where they were cleaned and identified to genus level based on morphological characteristics using a compound microscope. Samples were maintained in nutrient-enriched (2 g m −3 KNO 3 -N and 0.2 g m −3 K 2 HPO 4 -P soluble reactive phosphorus; EMSURE®) dechlorinated water in a temperature and light controlled laboratory (20-21 °C, 12:12 light:dark cycle, 250-300 µmol photons m −2 s −1 ). Single species stock cultures were created by isolating individual filaments from samples and scaling these up in sterile Petri dishes, and subsequently transferring cultures into a consecutively larger culture flask.
Five replicate µFANS of each species were seeded by evenly distributing 1.2 g fresh weight (FW) (~ 56 g m −2 ) of dewatered algae by rubbing it down the µFANS textured bottom liner to hook algal filaments, providing initial physical attachment. The seeded biomass rate was enough to cover the whole surface area of the µFANS bottom liner. This seeding method was selected based on pilot trials that compared seeding using algal zoospores and seeding by rubbing biomass onto the liner surface. The rubbing method resulted in a denser and more even biomass coverage over the liner surface in a shorter time than seeding using algal zoospores. Over time, secondary biological attachment occurred through the development of holdfasts following the growth of rhizoids from the hooked basal regions of the filaments and/or the subsequent release and settlement of zoospores on the liner surface.
The seeded µFANS were operated in batch culture using the same nutrient-enriched dechlorinated filtered freshwater used to maintain the algae stock cultures, but with higher nutrient concentrations (10 g m −3 KNO 3 -N and 2 g m −3 K 2 HPO 4 -P soluble reactive phosphorus; EMSURE®). These nutrient concentrations were selected based on preliminary trials, which showed these concentrations provided sufficient nutrient supply for algal growth over a 6-day growth period. The nutrient water was pumped continuously at ~ 0.5 L min −1 to the top of the µFANS floway. This flow rate was used for the seeding stage as preliminary trials indicated that it was high enough to cover the whole surface area of the floway in water, yet low enough to avoid washing the biomass off from the liner. The experiment was conducted under constant laboratory conditions (light intensity of 250-300 µmol photons m −2 s −1 using full-spectrum LED plant growth lights, 12:12 light and dark cycle, 20-21 °C) at the NIWA facility at Ruakura, Hamilton.
This seeding phase lasted for three 6-day batch cycles, and the nutrient-enriched water was replaced at the start of each new batch cycle. Daily reseeding was performed in the first batch cycle by respreading any clumped algal biomass over the liner surface to promote a more rapid uniform establishment of a dense algal turf on each µFANS. The amount of biomass needed for reseeding was determined according to the biomass coverage on the liner of each µFANS.

FANS Operation
At the end of the seeding phase, the algal biomass on each µFANS was harvested by drawing a metal scraper down the length of the liner to cut longer filaments and remove any detached biomass to leave a uniform coverage of attached algal filaments of < 1 cm length on the liner as a standing crop. As much as possible, µFANS were harvested to provide a similar standing crop between replicates and species.
Following harvesting and standing crop biomass measurement, each µFANS was reassembled, and 2 L of fresh nutrient-enriched dechlorinated water was added. The µFANS were then cultured under the same conditions described in Sect. 2.2 above for the next 3-day batch cycle. At the end of each batch cycle, the pump on each µFANS was turned off to allow the water to drain from the liner for 2 min before Photo of three µFANS biomass measurement, biomass harvest and standing crop measurement. This process was repeated five times to give six consecutive 3-day growth cycles.

Biomass Measurement
Attached algal biomass FW growing on each replicate µFANS was measured directly by draining and then spinning each replicate µFANS lined plastic floway (placed vertically with the algae facing inwards) at 2800 rpm for 30 s in a spin dryer (Spindel SPL265, 6.5 kg capacity) to remove excess water before weighing. The weight of each replicate µFANS lined plastic floway (that had been measured before algal seeding) was then subtracted from this weight to calculate the spun fresh weight (g FW) of the attached algae. This method was used to measure the weight of both the standing crop (biomass remaining following harvest) and the final biomass (biomass on µFANS at the end of each 3-day cycle). Biomass growth (g FW) for each µFANS was then calculated as the Final Biomass (g FW) -Initial Standing Crop (g FW). Subsamples of harvested biomass from each replicate µFANS were dried overnight in an oven at 65°C to determine the dry weight and determine the dry weight to fresh weight (DW:FW) ratio of algal biomass for each Fig. 2 Microscopic images of filamentous algae species assessed in this study a. Oedogonium sp., b. Spirogyra sp., c. Rhizoclonium sp., and d. Cladophora sp. Scale bar applies to all pictures replicate µFANS. These specific DW:FW ratios were used to convert the measured FW biomass of the standing crop and the final algal biomass to dry weight (DW) for each replicate µFANS.

Biomass Composition
Following weighing, a small amount of the attached algal biomass was sampled from each replicate µFANS and examined under a compound microscope to check the condition and purity (e.g., species composition) of the biomass.

Nutrient Removal Analysis
Water samples were collected daily from each replicate µFANS to determine the amount of nitrate (N-NO 3 ) and dissolved reactive phosphate (DRP) removed by the algae. Immediately before water sample collection, filtered dechlorinated water was mixed into the sump of each µFANS to replace any water that had evaporated over the 3-day growth cycle (approximately 5-7% of the total water volume, equating to an average of 100-140 mL every 24 h). A 12 mL water sample was collected from each µFANS replicate, filtered using 0.7 µm glass microfibre filter (GF/F) and the concentration of nitrate and DRP in each water sample was determined using the nitrate cadmium reduction method (Hach Method 8039) and the ascorbic acid method (Standard Method Phosphorus 4500-P) respectively. As nutrient adsorption by the µFANS and other processes such as denitrification and evaporation may have contributed to nutrient reduction, one control µFANS for each species was operated throughout the experiment. These were maintained under the same conditions and treated the same way as all other replicate µFANS except that they were not seeded with any algae. Concentrations of nitrate and DRP were determined in water samples taken daily from each control µFANS as described above. The nutrients removed by each replicate µFANS was calculated as the difference in nutrient concentrations between the Day 0 and Day 3 water samples. The difference in nutrient concentrations between the Day 0 and Day 3 water samples from the control µFANS was then subtracted from this value to estimate the total amount of nitrate and DRP removed by the algae over the 3-day experimental period. Nitrate and DRP removal rates were calculated as the amount (mg) of nitrate and DRP respectively reduced per g DW of biomass produced in a day.

Statistical Analyses
Differences in standing crop, DW:FW ratio, biomass productivity, total nitrate removal, total DRP removal, nitrate removal rate and DRP removal rate between species were tested using two-way analyses of variance (ANOVA) with species and batch cycle as fixed factors, or a Kruskal-Wallis test for variables that failed normality and/or homogeneity of variance tests. Linear regression analysis was used to test for a relationship between nitrate removal rate and specific growth rate, and phosphate removal rate and specific growth rate. Data were analysed separately for each species. All statistical analyses were carried out using SigmaPlot software (Systat Software Inc., USA). All data are reported as means ± S.D.

Biomass Seeding and Attachment
All species took about 10-14 days to establish uniformly on the liner of the µFANS (Table 1, Supplementary Information Fig. 9 a-d). Oedogonium sp. displayed the fastest and strongest biomass attachment on the liner. Based on light microscopy observations, the strong attachment of Oedogonium sp. appeared to be mainly a result of asexual zoospore production ( Supplementary Information Fig. 10) with subsequent settlement of the zoospores and holdfast attachment of the new filaments to the µFANS liner (Supplementary Information Fig. 11). There was evidence of filament fragmentation in Spirogyra sp. (Supplementary Information Fig. 12) and some fragments adhered to mucilage, however, microscopic  Fig. 13-14).

Biomass Productivity
Overall, Oedogonium sp. and Rhizoclonium sp. had the highest standing crop (average of all cycles ~ 0.8 g DW ± 0.2 for both species), except in cycle 4 where Cladophora sp. had the highest standing crop. Standing crop varied significantly among species and cycles (Table 2, Fig. 3). The standing crop of Cladophora sp. (average 0.7 g DW ± 0.1) was more consistent among batch cycles and replicates than Oedogonium sp. and Rhizoclonium sp. Spirogyra sp. had the lowest average standing crop (0.5 g DW ± 0.04) (Table 3), and this was more consistent among batch cycles (0.47-0.57 g DW) than for the other three species (Fig. 3).
The DW:FW ratio was different for each algal species and varied between cycles (Table 2, Fig. 4). Cladophora sp. had the highest DW:FW ratio (0.1 ± 0.01), while Spirogyra sp. had the lowest DW:FW ratio (0.05 ± 0.01, Table 3). The DW: FW ratio of Oedogonium  sp. and Rhizoclonium sp. were similar at 0.07 ± 0.01 and 0.08 ± 0.01, respectively. Biomass productivity varied significantly among species, however, the species with the highest biomass productivity was not consistent among cycles, as evidenced by a significant cycle x species interaction effect (Table 2, Fig. 5). Rhizoclonium sp. had the highest biomass productivity of all species in all cycles except cycle 1 where Cladophora sp. had the highest biomass productivity, while in all cycles Spirogyra sp. had the lowest biomass productivity. Across all cycles, biomass productivity of Rhizoclonium sp., Oedogonium sp., and Cladophora sp. (5.3 g DW biomass m −2 day −1 ± 1.1, 4.2 g DW biomass m −2 day −1 ± 1.1, and 4.2 g DW biomass m −2 day −1 ± 1.0, respectively) was nearly double that of Spirogyra sp. (2.2 g DW biomass m −2 day −1 ± 0.1) ( Table 3). Biomass productivity was reasonably consistent between cycles for Spirogyra sp. (2.1-2.3 g m −2 day −1 ) but more variable between cycles for the other three species.

Nutrient Removal
Some nutrient removal (< 20% nitrate and < 12% DRP) was observed in the control µFANS without algae over the 3-day cycle period. The total amount of nitrate removed over the 3-day batch growth cycle by each species was significantly different (Table 2), but the species with the highest removal rate was not consistent between cycles, as evidenced by a significant cycle x species interaction effect (Table 2, Fig. 6). Across all cycles, Oedogonium sp. had the highest average nitrate removal (5.2 g m −3 ± 1.3), followed by Rhizoclonium sp. (4.9 g m −3 ± 0.2) and Spirogyra sp. (4.2 g m −3 ± 0.7), while Cladophora sp. had the lowest nitrate removal (2.5 g m −3 ± 1.0) ( Table 3). Nitrate removal was reasonably consistent between cycles for Rhizoclonium sp. (4.7-5.2 g m −3 ), but more variable between cycles for the other three species.
Nitrate removal rates based on the biomass produced (e.g. mg N g −1 DW biomass day −1 ) varied significantly between species and were not consistent among cycles for each species, as evidenced by a significant cycle x species interaction effect (Table 2, Fig. 6). Across all cycles, mean nitrate removal rates per biomass produced were highest for Spirogyra sp. (28.4 ± 5.5 mg N g −1 DW biomass day −1 ) followed   (Table 3). Across all cycles, Oedogonium sp., Rhizoclonium sp. and Spirogyra sp. had the highest mean DRP removal (1.3 g P m −3 ± 0.2, 1.3 g P m −3 ± 0.1 and 1.2 g P m −3 ± 0.1, respectively), while that of Cladophora sp. was at least 50% lower (0.6 g P m −3 ± 0.2, Table 3). Phosphate removal over the 3-day growth cycle varied significantly between species and cycles, but was not consistent, as evidenced by a significant cycle x species interaction effect (Table 2, Fig. 7). Across all cycles, Oedogonium sp., Rhizoclonium sp. and Spirogyra sp. had the highest mean DRP removal (1.3 g P m −3 ± 0.2, 1.3 g P m −3 ± 0.1 and 1.2 g P m −3 ± 0.1, respectively), while that of Cladophora sp. was at least 50% lower (0.6 g P m −3 ± 0.2, Table 3). DRP removal rates based on the biomass produced significantly varied between species and cycles, but not in a consistent way as evidenced by a significant cycle x species interaction effect (Table 2, Fig. 7). The mean DRP removal rate across all cycles was highest for Spirogyra sp. (8.1 ± 1.0 mg P g −1 DW biomass day −1 ) followed by Oedogonium sp (5.6 ± 2.1 mg P g −1 DW biomass day −1 ), Rhizoclonium sp. (4.2 ± 0.7 mg P g −1 DW biomass day −1 ) and the lowest was for Cladophora sp. (2.2 ± 1.0 mg P g −1 DW biomass day −1 ). The DRP removal rate significantly varied between species and for each species between cycles, as evidenced by a significant cycle x species interaction effect (Table 2, Fig. 7).

Discussion
We developed a novel attached filamentous algae bioassay (µFANS) to assess the performance and suitability of filamentous freshwater algal species for cultivation on large-scale FANS. Our µFANS bioassay provides a comparative assessment of algal productivity among species under controlled growth conditions and nutrient supply. The µFANS bioassay was effective as it enabled rapid uniform establishment of attached filamentous algae species, with simple operation and maintenance for laboratory comparison of performance under growth conditions representative of those that would be experienced on a large-scale FANS. Moreover, the biomass productivity rates we recorded on our µFANS (~ 4 to 5 g m −2 day −1 ) were comparable to those reported for pilot scale (30 m 2 ) outdoor FANS with continuous flow (2.5 g m −2 day −1 , Mulbry et al. 2008b), suggesting that our µFANS bioassay provides a good indication of how species will perform on FANS operated in an outdoor setting with continuous flow. The small-scale of µFANS makes the experimental set-up portable, space-efficient (allowing multiple replicates to be tested simultaneously), and manageable to operate for the cultivation of several macroalgae species in the laboratory. In addition, the µFANS were economical to construct (~ NZ$40) and were assembled easily from readily available materials. Moreover, the µFANS can be operated in batch culture with recirculating nutrient-enriched water, enabling this approach to be implemented where a constant supply of flow through nutrient-rich water is not available.
The attachment capability of a species is the most important criterion when selecting species for use on FANS. Stronger algae attachment could significantly reduce biomass sloughing off the floway (Gross et al. 2016), especially during heavy rainfall events (Sandefur et al. 2011;Sutherland and Craggs 2017). We found apparent differences in the relative abilities of species to attach to the liner surface, most likely resulting from differences in the mechanisms of attachment. Oedogonium sp. formed biological attachment primarily through the release of numerous zoospores which subsequently settled and developed new filaments with holdfasts anchoring them to the liner surface. These findings agree with previous studies which show that asexual reproduction in Oedogonium sp. takes place by the formation of young filaments developed from zoospores that attach to substratum with the help of a basal cell holdfast (Hoffman 1965(Hoffman , 1967Arora and Sahoo 2015). In contrast, both Rhizoclonium sp. and Cladophora sp. became firmly anchored onto the liner surface probably through the development of rhizoids. This type of attachment has been observed in previous life history studies of these species (Nienhuis 1974;Parodi and Cáceres 1993). Both Rhizoclonium sp. and Cladophora sp. can also produce zoospores through asexual reproduction (Bellis and McLarty 1967;Parodi and Cáceres 1993;Aroca et al. 2020), however spore production was not observed in these two species in the present study. Spirogyra sp. filaments did not readily hook onto the textured liner and easily fragmented, with fragment attachment probably by adhering to mucilage or by rhizoid formation. Mucilage secretion and adherence has been observed in fragmented filaments of Spirogyra sp. (Nagata 1977) and rhizoid formation has been induced in cut Spirogyra sp. fragments under laboratory conditions (Nagata 1973). However, not all developed rhizoids will anchor to a substratum, and some may remain floating in still water (Ikegaya et al. 2008). There is also evidence that anchored Spirogyra sp. filaments can detach themselves due to changes in surface tension (Nagata 1973). Spirogyra sp. spore production was observed in this study, however they did not settle on the liner surface. This highlights the importance of considering not just the mechanisms of physical attachment (hooking) and biological attachment (zoospore, production, settling and holdfast development; mucilage secretion, and; rhizoid formation), but also how easily filaments fragment in the absence of zoospore formation when assessing the suitability of species for FANS cultivation. Overall, Oedogonium sp. had the fastest biomass establishment with the strongest biological attachment of all four species. It had the most attachment points (holdfasts), and filaments were not easily fragmented or detached.
Of the four species we tested, Rhizoclonium sp. had the highest biomass productivity, while Spirogyra sp. had the lowest. Spirogyra sp. also had a consistently lower standing crop and took at least 50% longer to reach the same standing crop biomass compared to the other three species. It was challenging to get a similar standing crop for all species to start each batch cycle as 1) biomass moisture content of each species was different, which affected the DW:FW ratio and the standing crop estimation of all species based on the standardized DW; 2) mechanisms of attachment and texture of biomass were different for all species and therefore each Fig. 8 Relationship between NO 3 and PO 4 removal rate (g m −3 day −1 ) and specific growth rate (% day −1 ) for four species of algae over six growth cycles. Values for each replicate µFANS for each species and cycle are plotted species required a different approach to be taken to harvest the attached biomass; and 3) they grew at a different rate which affected the biomass establishment on the liner. In general, the standing crop in attached algae cultivation systems like FANS is equivalent to the stocking densities of macroalgae in suspension culture, where a higher stocking density can result in higher biomass productivity (Pereira et al. 2006). Therefore, a higher standing crop may have contributed to the higher biomass productivity of Rhizoclonium sp., Oedogonium sp. and Cladophora sp. compared to Spirogyra sp. in the current study. However, standing crop may not be the dominant factor affecting algae biomass productivity in FANS. The tolerance of algal species to the particular growth conditions (including light, temperature, pH, and type of cultivation system) can also impact biomass productivity (Singh and Singh 2015). The low biomass productivity of Spirogyra sp. in the current study compared to the other three species may have been caused by elevated pH of the recirculated nutrient-enriched water, as high pH can limit algal growth (Grobbelaar 2010;Park and Craggs 2011). Spirogyra sp. appears to be less tolerant to high pH than the other algae species tested as it has only been found in streams with a pH below 8.6 (Khanum 1982;Pikosz and Messyasz 2015). Furthermore, Spirogyra sp. had higher growth rates in suspension culture when CO 2 was added, resulting in a reduction in pH (Lawton et al. 2013). In the current study, pH was consistently recorded between 8.5-9.5, which could be too high for optimum growth of Spirogyra sp.
The high biomass productivity of Rhizoclonium sp. in the current study relative to the other three species demonstrates its suitability as a target species for FANS cultivation. This finding is supported by previous studies, which have found that Rhizoclonium sp. is common in mixed-species assemblages in attached algal cultivation systems and often outperforms other species of filamentous algae. For example, Rhizoclonium sp. was consistently the most abundant species on self-seeded FANS floways operated year round to treat dairy manure effluent (Mulbry et al. 2008b). Similarly, Rhizoclonium sp. was one of the most abundant species on self-seeded FANS treating drainage water from a citrus farm (D'Aiuto et al. 2015). Rhizoclonium was also one of only two species of filamentous macroalgae that were successfully isolated from industrial wastewater and maintained indoors under a controlled conditions for more than a month (Nwoba et al. 2017). Further supporting the selection of this species as a target for cultivation on FANS, Rhizoclonium has a wide tolerance to factors such as temperature, salinity, photoperiod and nutrient concentration (Aroca et al. 2020). Both Oedogonium sp. and Cladophora sp. also had relatively high biomass productivity in the current study and as such would be suitable targets for FANS cultivation. Supporting this conclusion, these species are often dominant components of the algal community on self-seeded FANS (Mulbry et al. 2008a(Mulbry et al. , 2008bD'Aiuto et al. 2015) and have comparatively high biomass productivities to other species in suspension monocultures (Lawton et al. 2013(Lawton et al. , 2021Liu and Vyverman 2015).
Nutrient removal rates, as determined by milligrams of nitrate and phosphate removed per gram DW of biomass produced, were highest for Spirogyra sp. and lowest for Cladophora sp. The variation in nutrient removal rates among algae species tested may have been caused by variability in nutrient assimilation efficiency and nutrient storage capacity (Fujita 1985;Ohtake et al. 2021). Nutrient assimilation efficiency of algal species is affected by cell wall permeability and mechanisms of active transport across the cell wall (Kuffner and Paul 2001;Vermeij et al. 2010). For example, Spirogyra sp. has been found to have higher cell wall permeability than Cladophora sp. (Osterhout 1913;Bergman 1949). This may explain the higher nutrient removal rates of Spirogyra sp. in the current study compared to the other three species. In addition, the nutrient removal rate of algae could also be associated with the surface area to volume ratio (SA:V) of algal cells due to the effect of algae cell thickness on nutrient transport distances (Rosenberg and Ramus 1984;Hein et al. 1995). Algae with greater SA:V have been found to have higher nutrient uptake rates (den Haan et al. 2016). In the current study, higher nutrient removal rates were observed in species with thinner cells, such Spirogyra sp., Oedogonium sp. and Rhizoclonium sp. compared to Cladophora sp.
Although Spirogyra sp. was the best performing species in terms of nutrient removal per gram of biomass production; Oedogonium sp. had the highest overall nutrient removal when considering the total concentration of nutrients in the water before and after FANS treatment. This was despite the fact that Oedogonium sp. had 20-30% lower nutrient removal rates per gram of biomass production (mean 21.8 mg NO 3 g −1 DW day −1 ; 5.6 mg PO 4 g −1 DW day −1 ) compared to Spirogyra sp. These differences were due to the much higher biomass productivity of Oedogonium sp. (mean 4.2 g DW biomass m −2 day −1 ) compared to Spirogyra sp. (mean 2.2 ± 0.1 g DW biomass m −2 day −1 ). The higher biomass productivity recorded for Oedogonium sp. in addition to it's high overall nutrient removal make it a better target for combined bioremediation and biomass production than Spirogyra sp. More generally, these results highlight the importance of considering multiple performance metrics when selecting target species for cultivation on FANS. Overall, our results suggest that macroalgae nutrient uptake rate is not directly correlated with growth rate as linear regression analyses showed a weak relationship between these two variables for all species. This could be because the assimilated nutrients are not always associated with biomass yield since different macroalgae species may have different mechanisms of nutrient uptake and storage and, to varying degrees, for example, luxury consumption (Gerloff and Krombholz 1966;Reef et al. 2012).
We have several recommendations and insights for future studies using a µFANS bioassay. The optimum duration of a three-day growth period for each cycle was selected to minimise microalgal contamination of the culture water that could affect macroalgae growth and nutrient removal results. Thus, the recirculating nutrient-enriched media needs to be replaced every three days and we advise cleaning the sump at the same time. We found four to six cycles of a three-day growth period was sufficient to compare the performance and suitability of macroalgae species on FANS as this provides more than a month (including the seeding stage duration) for the algae to adapt to and thrive in the µFANS set up. The time for algae establishment could be reduced from the 14 days observed in the current study, and algae attachment could be sped up by employing daily biomass reseeding and redistribution on the liner surface. This approach could reduce the seeding duration to one week based on preliminary tests. However, seeding too much biomass (more than 56 g m −2 ) onto the liner could limit light and therefore growth. Additionally, redistributing the biomass too frequently on the liner (< 24-h interval) can disturb any attachment that has partially formed. Estimating the amount of FW algae biomass that needs to be harvested from each µFANS to provide for a standardised standing crop in DW across all replicates was challenging as each species' DW:FW ratio varied. For this step, we therefore recommend that estimation of how much FW biomass needs to be harvested for each µFANS should be based on the DW:FW ratio of a biomass subsample collected a day before the actual harvesting day.

Conclusion
We developed a novel, standardised, reproducible bioassay to assess the suitability of different filamentous algae species for use on FANS based on their ability to attach and grow, remove nutrients and produce harvestable algal biomass. The developed microscale FANS bioassay is practicable and effective and provides a comparative assessment of algal productivity and nutrient removal among species under controlled growth conditions. It can be used as a first step to rapidly and easily identify potential target species for FANS cultivation from a large number of candidate species. As the results from a laboratory-scale bioassay may not fully reflect the complex interaction between abiotic and biotic factors in large-scale outdoor FANS, potential target species identified through the bioassay should then be cultivated on outdoor FANS over an extended period under ambient conditions to provide a realistic test of likely performance in the field, including how long target species are able to maintain dominance on the floway. Importantly, this study has proved that a single target algae species could achieve nutrient mitigation and biomass production on a small-scale FANS. Generally, attached algal systems are self-seeded with mixed-species assemblages. However, our results demonstrate that single target species can be maintained on FANS and achieve nutrient mitigation and biomass productivity at least over month timescales. Oedogonium sp. was the best performing algal species in terms of most uniform biomass distribution and attachment, high productivity and high nutrient removal rates. These criteria make Oedogonium sp. an ideal species for cultivation on large-scale FANS, however further largerscale testing under ambient outdoor conditions is required to confirm this.