Microbial community dynamics and operational performance in rice straw anaerobic digestion

Abstract

Waste rice straw (RS) is generated in massive quantities around the world and is often burned or left to rot in the fields creating environmental and health issues. Work has shown anaerobic digestion (AD) of RS is feasible, but is primarily associated with expensive and/or hazardous pretreatment methods. Therefore, this study sought to determine if RS AD was possible without pretreatment and under what conditions. Biomethane potential (BMP) batch tests (~200 days) found methane yields were similar at organic loading rates (OLR) of 1 to 3 g VS/L; 425 μm and 1.0 mm RS particle sizes were superior to 30 mm and 70 mm; and addition of dairy manure reduced yields. Nigerian RS methane yields were higher than Chinese, Indian, and Philippine RS, with BMP yields increasing with lower lignocellulose. Using the BMP data, 2.5 L continuously-stirred reactors were operated (~500 days) at five feeding frequencies (FF) and two different OLRs. Less frequent feeding at low OLR was more effective probably because of the relative recalcitrance of RS; a FF well-suited to acyclic harvesting at RS in rural settings. However, less frequently fed units (1/14 and 1/21) at higher OLRs both failed (soured) due to apparent organic acid accumulation owing to overloading. 16S rDNA amplicon sequencing showed ‘healthy’ reactors were dominated by Methanogens and Bacteroidetes and ‘souring’ of reactors led to dominance by Firmicutes. Overall, increasing the OLR had a much greater impact on AD microbial communities and infrequent feeding did not have a negative impact at low OLR. The substrate, not the feeding regime, dictated the community. In this case the traditional markers of reactor stability e.g. pH, did not suggest impending reactor failure. However, microbial population shifts, such as increases in fermenters and decreases in methanogens, reacted earliest and could be used as an early warning of forthcoming system failure. Further, four RS:DM ratios were tested (to assess different C:N ratios) in the reactors. 100 % RS had the highest methane yield with increasing DM reducing yields. Sequencing data indicated microbial community richness increased with the increasing DM addition. Predominant OTUs in the RS only unit were Bacteroidetes/Firmicutes and the 30:70 RS:DM unit was Proteobacteria. DM reactors had lower abundances of cellulosic hydrolysing bacteria such as, Christensenellaceae and Bacteroidetes. Data suggest DM-amended reactors were underfed. As such, the benefit of co-digestion would be decreased VFA production and VS reduction, which might cope with higher OLR, enabling a greater throughput of RS. Overall, results show that RS AD without major pretreatment is feasible, especially at lower OLRs with less frequent feedings, although co-digestion with manure could allow higher OLR operations, although this needs to be proven.