Research Report 1.5
Impact of Manure Application Methods on Water Quality,
Wall, B.A. Grant, D.J. King
Land Resource Division, Agriculture and Agri-Food Canada
70 Fountain St., Guelph, ONT N1H 3N6
Dr. N. McLaughlin
Eastern Cereal and Oilseed Research Centre, Ottawa, ON K1A 0C6
COESA Report No.: RES/MAN-005/97
In a joint effort between Agriculture and Agri-Food Canada, the Upper Thames River Conservation Authority and the Ontario Ministry of Agriculture Food, and Rural Affairs, a study was conducted with the following objectives:
Three (two ha) field sites were selected in the Mixedwood Plains ecozone with contrasting soil textures (Site 1(medium)-silt loam, Site 2 (light)- sandy loam and Site 3 (heavy)- silty clay loam), systematic tile drainage and a history of no-till corn crop management. A no-till corn crop was planted in May at each field site by the farm cooperator using commercial no-till planters outfitted with various coulter and trash whipper arrangements to manage the corn residue. Starter fertilizer was applied with the planting units according to the farm cooperators preference with rates ranging from 5 - 36kg N/ha, 6 - 14kg P/ha, and 0 - 26kg K/ha. Control plots were fertilized at the time of planting according to soil test recommendations.
Soil and manure nutrient analysis conducted in about the 3rd week of June was used to compute the manure application rates that would be required to meet the crop N requirements according to soil test recommendations. The required liquid manure application rates ranged from about 56,165 to 71,890 L/ha. Crop performance (plant populations, time to silking, leaf tissue analysis, weed pressure, lodging, grain moisture and yield) was monitored from the time of emergence to harvest. The cost effectiveness of the manure application equipment in terms of draft and fuel consumption was measured with an instrumented research tractor from Agriculture and Agri-Food Canada.
The liquid hog manure was side dressed with a 6,800 L tanker around the fourth leaf stage by surface application and two injection techniques (conventional injection and injection modified by slight tillage in front of the injectors). Before application of the manure, bacteria tracer and strontium chloride (StCl) and potassium bromide (KBr) were added to the manure. The area being drained by each of 12 tiles (7m by 100m) was treated as an individual plot. Three treatments (surface applied, conventional injection, modified injection) and a control (inorganic fertilizer) were replicated three times, and crop and water quality response to the liquid manure application was monitored. On the day following the manure application, rainfall was simulated using a travelling gun irrigation system to apply about two to three cm of water to the plots over a 20 to 30 min period.
Water quality samples were taken weekly throughout the growing season while the tiles were flowing. On the days of manure application and the simulated rainfall event, tile water quality samples were taken at 15 min to three hour intervals. An automated meterological station at each study site provided temperature, relative humidity, wind speed/direction, and 15 minute precipitation data. Soil and hydrologic data from the study sites were used to evaluate a tile flow prediction model (DRAINMOD 4.0). Two years of data were obtained for Sites 1 and 2 while a single year of data was collected at Site 3.
Agronomic results showed that the use of liquid manure at side dress did not impair performance of crop growth and development. The combination of side dressed manure N and the starter fertilizer N when applied at soil test recommended rates provided no-till corn yields that were not statistically different from the control treatments where inorganic fertilizers were used. There was also no significant difference in grain yield between the injection and surface manure application methods, although both plant populations and grain yields were marginally lower than the control plots in three of the five crop years. The tractor energy consumption data clearly showed the extra power required for the modified injection configuration over the conventional injection, however, the difference was relatively small and the modified injection configuration was considered practical from an energy consumption perspective.
Liquid manure application for all treatments resulted in increased rates of tile flow volumes within 30 minutes of application and returned to base flow conditions within three hours. Flow rate increases were greatest when the tiles were flowing prior to the manure application. The simulated rainfall event increased tile flows significantly by approximately 20L/min at Sites 1 and 2 and by 7 L/min at Site 3 and did not return to base flows for several days. Flow increases in the tile drains after liquid manure application represent less than 3% of the applied manure, while tile drain flow increases after the simulated rainfall represent about 10% of rainfall volumes at Sites 1 and 2 and less than 5% at Site 3.
Tile water quality impairment was observed through increased turbidity and measured through the presence of bacteria and ammonium, for all manure application treatments within 7 to 30 minutes of application and continued for two to three hours. While the total volume of applied manure reaching the tile was small (<2%), water quality guidelines for bacteria, ammonium and phosphorus were exceeded for several hours. The presence of the bacteria tracer and chemicals in the tile water samples after manure application provided verification that manure was the source of contamination. The simulated rainfall event resulted in increased levels (but at lower concentrations then following manure application) of ammonia, bacteria tracer and phosphorus within 30 minutes and peaked within 60 minutes. Since the bacteria and chemical tracers were not detected in the tile water a few days after the rainfall event it appears that the impact of the manure application on the tile water quality is relatively short-lived.
The movement of the tracers strontium, bromide and chloride mirrored the bacteria movement to the tile drains both in time and concentration. The percentage of the applied non-reactive tracers bromide and chloride reaching the tile drains was found to approximate the percentage of liquid manure reaching the tile drains (<2% of applied). While <1% of the applied reactive tracer (strontium) was recovered in the tile water, it provides evidence that the macropore pathways are contributing to tile flows even under the unsaturated soil moisture conditions of the experiment.
Regardless of the method of liquid manure application, tile water contamination occurred both immediately following the manure application and the simulated rainfall event. In this no-till system, it may only be possible to stop tile water contamination by applying liquid manure during the growing season periods when soil moisture content is low and tile drains are not flowing.
The tile drainage model (DRAINMOD 4.0) provided statistically good predictions of tile flow for both years at Sites 1 and 2 compared to measured flow values. Further study of the water quality components of the model that are currently under development may be warranted.
Study results have led to the following recommendations for the application of liquid manure in no-till cropping systems:
Last Updated: May 16, 2011 02:59:56 PM