Winter 2002
Manure sampling, analysis, and applicator calibration
John E. Sawyer and John Lundvall, Department of Agronomy
The goal of a project outlined in the Odor and Nutrient Management newsletter, Fall 2002 issue, pages 4-6, is to demonstrate crop utilization of liquid swine manure nutrients - from understanding the nutrient rate of application to measuring crop response. In this article, we present results for the first step: understanding the rate of nutrient application. For each site in the demonstration project, this step entailed manure preapplication sampling and laboratory analysis, manure sampling during application with laboratory analysis, application equipment rate calibration, and nutrient application rate calculated from the calibration and at-application sample analysis. For the first 3 years of the project (2000–2002), we worked with 16 producer cooperators at 39 production/field sites located in 12 counties (representing 54 manure treatment applications). Specifics on the demonstration procedures are outlined in the Fall 2002 newsletter article.
Preapplication manure analyses compared with at-application analyses. For all sites, the manure source was from swine finishing facilities with storage in under-building pits or outside concrete tanks (two sites). Manure samples were collected 2 to 3 weeks before planned application by either dipping manure off the surface or probing the storage profile. Forty of the 54 applications were based on total-nitrogen (N), with the remaining 14 based on total-phosphorus (P). Multiple samples (up to 11 samples per site) were collected during application to the demonstration sites (98 total manure samples for the 3 years). Manure was agitated during pump-out of the storage structures. Figure 1 shows a comparison between the preapplication sample analyses (total N, P2O5, or K2O per 1,000 gallons) and the average of the samples per site collected during application. Presamples were often analyzed only for total-N if the application was to be based on total-N. Figure 1 represents the ability of the presample to predict the manure nutrient concentration during application. Overall, the presample gave a good prediction of the total-N concentration expected during application. On average, the preapplication sample had 5.7 percent lower total-N than the at-application samples. Across all sites, the average ammonia-N in the liquid swine manure was 83 percent of the total-N. For P, the variation between pre- and at-application sampling was larger, but in some instances the presample was dipped off the manure surface, which is not expected to provide a good representation of P in an agitated pit. Because potassium (K) is contained in the soluble manure solution, the preapplication samples were close to the at-application samples.
Intended manure nutrient rate compared with calculated applied rate. Figure 2 shows the comparison of the intended manure total-N or total-P application rate and the calculated applied nutrient rate. The applied rate was calculated from the average analyses of the manure samples collected during application at each site and the application equipment calibration. For total-N, if one accepts ± 30 pounds N/acre as an acceptable ability to apply manure-N, then 18 percent of the applications (7 of 40 applications) were outside this range (all but one of these was with a vacuum style applicator). In some instances, the calibration process indicated that greater than desired rates were going to be applied because of equipment limitations to reduce the flow rate and/or tractor speed limitations. These sites were kept in Figure 2, and an example is the two very high application rates. The occurrence of applications well above intended rates happened with vacuum-style applicators, and especially when the manure nutrient concentration was high. For total-P, if one accepts ± 15 pounds P2O5/acre as an acceptable ability to apply manure-P, then 29 percent of the applications (4 of 14 applications) were outside this range, mainly due to the presample P analysis being higher or lower than the at-application samples. However, a wider range in P application could be expected as some of the manure samples were dipped from the manure storage surface for total-N measurement rather than probed, which would be expected to not represent P as well.
When based on either total-N or total-P, 19 percent of applications were greater than 25 percent from the intended rate (10 of 54 applications). The majority of applications were within 15 percent of the intended rate. If you take out the two known high application rates from one site, then 13 percent of applications fall outside the ± 30 pound total-N/acre range. Seven of the 10 high application rates were made with vacuum-style equipment. Many of the applicators used in the project were equipped with a flow monitor and rate controller. These applicators calibrated well, and variation between intended and calculated rates generally were due to differences in the pre- and at-application manure analyses. Partly due to the preapplication sample analysis being lower than the at-application sample, the tendency was for the calculated applied rate to be larger than the intended rate.
Variability in nutrient analyses for samples collected during application. Figure 3 shows the comparison of individual manure sample N, P, and K analyses and the site average analyses. Because the project worked with producers from a wide area of Iowa and with different swine production practices, one would expect a wide range in total N, P, and K content, as is seen with the distribution in average site analyses. For total-N, the lowest site had 32 pounds and the highest site had 79 pounds total-N/1,000 gallons. For total-P, the lowest site was 17 and the highest 54 pounds P2O5/1,000 gallons. For total-K, the lowest site was 23 and the highest 48 pounds K2O/1,000 gallons. These differences in site averages highlight the importance of sampling and laboratory analysis rather than using book values. Only if a book value happens to coincide with the actual analysis would the book value be helpful for determining application rates.
Figure 3 also shows the variation within the multiple samples collected during each application. For N and K2O, the ranges are very narrow, with most samples falling within ± 2 pound/1,000 gallons (94 of 98 samples within this range for N and K). For P the variation was wider (22 of 98 samples greater than ± 2 pounds P2O5/1,000 gallons), indicating the tie between P and variation in solids content as a storage structure is emptied.
Summary. The project is documenting the importance of sampling liquid swine manure for determining nutrient concentrations. In conjunction with application equipment calibration, manure preapplication analyses are helpful for achieving desired nutrient application rates. The entire application process requires effort, but can be successful if careful attention is paid to sampling, calibration, and rate monitoring and control. In addition, over time a manure analysis history from the pre- and at-application samples can be developed that will aid future applications and reduce the reliance on preapplication samples.
The ISU Swine Manure Nutrient Utilization Project, part of the Integrated Farm/Livestock Management (IFLM) Demonstration Program, receives funding from the Iowa Department of Agriculture and Land Stewardship, Division of Soil Conservation, USDA Natural Resources Conservation Service, and the Leopold Center for Sustainable Agriculture.
This is the second in a series of newsletter articles highlighting the ISU Swine Manure Nutrient Utilization Project. The final article will appear in the March 2003 ONM newsletter and will highlight crop yield response to manure nutrient application.
Figure 1. Comparison of pre- and at-application
manure nutrient analyses.
Figure
2. Comparison of intended and calculated as-applied manure nutrient application
rates.
Figure
3. Variability in average manure nutrient analyses between demonstration
sites and within multiple samples collected during application.
