Biodiversity and disease risk in an algal biofuel system: An experimental test in outdoor ponds using a before-after-control-impact (BACI) design

described in McGowen et al. [38], which involved vacuum filtering the 15-mL subsample through a pre-weighed 0.2μm glass microfiber filter, dried at 105˚C, ashed at 500˚C leaving an inorganic compound residue and then re-weighed. The difference in weight before and after ashing gives AFDW in g of biomass L-1. For each sample, three separate measures of AFDW were averaged to produce one estimate per sampling event. A second 1-mL subsample was fixed with a 1% phos- phate-buffered formalin solution to preserve the sample for quantification of algal cell densities and the proportion of cells with fungal infection. This was achieved by performing manual counts of individual species cell density (healthy and infected cells) using a hemocytometer and light microscopy. Following each sampling event, fresh water was added to the raceway ponds to main- tain their original volumes at 400-L and compensate for evaporative losses. In addition to sampling the ponds every 2-days, the feedstock of algae in each raceway pond was harvested weekly to measure areal productivity. Harvesting was performed using a sump pump placed directly into the raceway while the raceway was actively mixed by the mechanical paddle wheel. With exception of the first harvest, for which only 70% of the volume was removed due to the fact that biomass was still increasing, all subsequent harvests were performed by pump- ing out 90% of the volume of the raceways, which is a more standard harvest. Immediately after each harvest, raceways were filled back to their starting volumes of 400-L with fresh water replen-

ished with nutrients to produce the same initial concentrations of BG-11 growth media. Data analysis The aforementioned data collection was used (Section Experimental design) to calculate four response variables for each pond–areal productivity of feedstock, the stability of feedstock bio- mass through time, maximum biomass yield achieved immediately prior to harvest, and the proportion of algal cells infected with fungus. Areal productivity is a measure of the average daily biomass production (g of Biomass m-2 day-1) based on the surface area of each raceway pond. To calculate areal productivity, the average daily volumetric AFDW was converted to the mean daily areal productivity by dividing by 4.2 m2 (the surface area of the 400-L race- ways). Average daily productivity of each raceway was calculated for the time frame between harvests, which resulted in two average daily productivity measures of each raceway for the Before fungal infection period and two measures for the After period. Temporal stability of feedstock biomass was quantified by calculating the inverse of the coefficient of variation (CV) as the mean divided by standard deviation for daily AFDW (g bio- mass L-1 day-1) for each growth phase between harvests. To quantify the overall stability for each treatment (Control-mono vs. Impact-polyculture), individual estimates of the CV-1 for growth phases within each period (Before vs. After infection) were average. This allowed for examination of whether there were any significant differences in stability of daily biomass pro-

duction between Treatments (control vs. impact) and Periods (before vs. after fungal infection). The maximum biomass yield achieved in a raceway pond was taken to be the peak algal bio- mass (g biomass L-1) achieved immediately before harvesting each raceway. Samples taken immediately before each harvest event were used to calculate AFDW, providing an estimate of total amount of biomass by volume produced between each harvest event. Using these mea- surements, any differences between Treatments (control-mono vs. impact-polyculture) and Periods before vs. after fungal infection in the average maximum biomass yield achieved between harvest were able to be determined. Fungal infection and community composition (polycultures only) were measured to evalu- ate any difference in response to infection between treatments. Fungal infections were quanti- fied by calculating the proportion of algal cells infected with the fungal parasite for each species (the total healthy and infected cell density divided by the density of infected cells) for PLOS ONE Biodiversity, biofuel and disease: Ecological engineering in algal feedstocks PLOS ONE | https://doi.org/10.1371/journal.pone.0267674 April 28, 2022 6 / 15