|Schmidt, A -|
|Davis, J -|
|Fan, Z -|
|Kiess, A -|
Submitted to: International Symposium on Air Quality and Waste Management for Agriculture
Publication Type: Proceedings
Publication Acceptance Date: July 7, 2010
Publication Date: September 13, 2010
Citation: Schmidt, A.M., Davis, J.D., Purswell, J.L., Fan, Z., Kiess, A.S. 2010. Analysis of the effect of spatial and temporal sampling densities on accuracy of predicting the heating profile in windrowed broiler litter. International Symposium on Air Quality and Waste Management for Agriculture. ASABE #711P0510. Interpretive Summary: Windrowing of broiler litter has been employed to manage populations of microbes in built-up litter. Temperature of the interior windrow increases due to microbial activity, and reductions in microbial populations have been reported in the literature. However, temperature has only been measured at the core of the windrow and therefore no estimates of temperature distribution or duration are available. Temperature was measured in a dense sampling grid at two minute intervals over a 10 day period in litter windrows. Recommendations for spatial and temporal sampling intervals to ensure accurate characterization of temperature distribution were developed. Temperature should be sampled with a grid spacing of 20 x 20 cm or less at 200 minute intervals or less. Accuracy of temperature distribution prediction is only 45% when using grid spacing used in previous studies.
Technical Abstract: A standard method for monitoring temperature in windrow piles of broiler litter to predict microbial population reductions is described. Temperature data collected every 2 min on a 10 cm x 10 cm spatial sampling grid in five identically-constructed litter windrow piles was utilized in this study. A Weibull distribution was fit to mean temperature response (MTR) curves of each pile. Curves were constructed at sample intervals parsed over a range of two to 1000 minutes. No difference in Weibull shape or scale parameters was observed among the analyzed sample intervals. A difference (P<0.05) in mean standard error of Weibull distribution fit parameters was identified between the 200- and 400-min sample intervals. Further analysis between the 200- and 400-minute sample intervals did not reveal a more appropriate value for optimal temporal sampling frequency. Optimal spatial sampling density was characterized using ordinary kriging analysis. Ordinary kriging was used to predict the cross-sectional areas of piles reaching specified time-temperature goals. Eight spatial sampling grid configurations were analyzed. Mean (n=5) predicted cross-sectional area (CSA) reaching 40°C for 120 h differed significantly (P<0.05) between the 30 cm x 20 cm and 30 cm x 30 cm grid spacing configurations. Accuracy of predicted pile CSA decreased as spatial sampling density decreased. This data will be beneficial when designing future windrow composting temperature monitoring studies.