DOI: 10.1111/risa.70330 ISSN: 0272-4332

Temporal and Environmental Influences on the Survival of Escherichia coli in Poultry Litter‐Based Soil Amendments in Georgia Sweet Onion Production Systems

Harsimran Kaur Kapoor, Amelia Payne, Krishna Prabha, Laurel L. Dunn, Govindaraj Dev Kumar, Chris Tyson, Manan Sharma, Keith R. Schneider, Aditya Kumar Mishra, Alda F. A. Pires, Patrick Baur, Abhinav Mishra

ABSTRACT

Previous studies have shown that pathogen survival in the biological soil amendments of animal origin (BSAAO) is mediated by meteorological conditions. Hence, a 2‐year field study was conducted in Georgia with four soil amendment treatments (heat‐treated poultry pellets, HTPPs; composted poultry litter, PL; unamended, UN) inoculated with Escherichia coli TVS 353 (a surrogate for Salmonella enterica ) and negative control plots (unamended, non‐inoculated). Additionally, 14‐day‐old sweet onion bulbs were transplanted into plots managed by HTPP (O‐HTPP), harvested, and left to cure for 14 days in the field. A linear‐mixed‐effect model (LME) was fitted to E. coli population data enumerated from the soil samples collected at specified intervals over 161 days from three replicate plots amended with different soil amendment treatments during both years of the study. Amendment type (UN, PL, HTPP, O‐HTPP) and weather factors (cumulative rainfall (cmrain 2 ), average air temperature (at60 12 ), relative humidity (RH 12 ), soil temperature (ast 12 ) for 2 days before sampling, and wind speed 1 day before sampling (W 1 )) were the predictors for the LME models developed. Significant weather predictors that affected the survival of E. coli across soil amendment treatment plots were RH 12 (+0.35) (Year 1); W 1 (−0.77), at60 12 (−1.66), and ast 12 (+1.57) (Year 2). For plots with and without onions, W 1 (+0.69) (Year 1); at60 12 (−0.92), ast 12 (+1.35), cmrain 2 (+0.37), and W 1 (−0.36) (Year 2) were significant. All coefficient values represent the estimated change in log 10 CFU or Most Probable Number (MPN)/g associated with a one‐unit increase in the respective weather predictor, as estimated by the LME model ( p  ≤ 0.05).