Preferential Flow of Liquid
Manure to Subsurface Drains

Larry D. Geohring
Sr. Extension Associate, Cornell Cooperative Extension

Tammo S. Steenhuis
Professor, Cornell University

The USGS National Water Quality Assessment data for the Northeast indicate animal manure was the primary source of nitrogen (N) and phosphorus (P) inputs to the watersheds (about 41 lbs/acre). Manure source inputs of N exceeded fertilizer, atmospheric and other point sources of these nutrients. Manure N accounted for 36% of the total N inputs, slightly higher than the N inputs from the atmosphere, the next highest source. The fate and transport mechanisms of nutrients from manures is complex and still not well understood.

Much attention has been focused on water quality contamination from rapid surface runoff of manured areas (Brockamp, 1993). However, information on the water quality from tile beneath manured sites, and especially the portion attributable to preferential flow, is inadequate. A better understanding of these processes would be beneficial to improving manure application management decisions about timing, amount, and application method which maxirnize the nutrient benefits of manure while minimizing the contamination of receiving water bodies.

Preferential transport processes can move significant quantities of contaminants very rapidly, with little alteration, to the shallow groundwater and then to tiles. A significant portion of cropland where manures are applied has also undergone various tile drainage improvements. In these areas, preferential flow contributes to the tile discharge which is outletted to surface waters. Experimental efforts with preferential transport have not been adapted specifically to manure to assess whether manure will also be rapidly transported.

Evans and Owens (1972) and Dean and Foran (1992) reported the application of liquid manures to tile drained fields resulted in elevated levels of nutrients and bacteria compared to normal tile discharges from unmanured sites. Fleming and Bradshaw (1992) reported maximum levels of 88.2 mg/l of NH4-N, and 1020 mg/l total suspended solids; and McLellan et al. (1993) reported peak concentrations of 53,000 organisms/100 nil of E. Coli in tile discharges shortly after application of liquid manures which originally contained 149 mg/l NH4-N and 7,000,000 organisms/100 tnl of E. Coli. For comparison, the accepted maximum effluent standard for most municipal wastewater treatment plants is 400 organisms/100 rnl of E. Coli and 37 mg/l of total suspended solids.

Bacteria typically ranges in size from 0.3 to 12 um and protozoans such as Cryptosporidium parvum and Giardia @lia are approximately 2-5 um and 5-10 um respectively. Thus, these limited observations of the occurrence of bacteria in the drain effluent may also be an indicator that these protozoa are also present. Consequently, it appears that preferential flow processes can have important implications on the fate and rapid transport of nutrients and pathogens from the applications of manures. These observations indicate that transport of liquid manure contaminants may lead to high and undesirable concentrations of nutrients and bacterial pathogens to subsurface drains and subsequently into receiving water bodies.

A field experiment was initiated at Cornell University's Willsboro Farm located in Northern New York and adjacent to Lake Champlain. The experimental plot area consists of 8 tile drained plots where each plot consists of a single drain line discharging into a manhole. The soil in the plots is a somewhat poorly drained Rhinebeck variant fine sandy loam which has similar drainage characteristics to 880,000 acres of somewhat poorly drained soils commonly used for agriculture in NY. No manure was previously applied to these plots for the last 16 years.

After the harvest of corn silage and prior to any manure application four plots were pre-wetted with sprinkler irrigation (Plots # 3,4,5, and 6) to establish a soil moisture differential and until drain flow commenced. A total of 3 inches of water was applied during the pre-wetting process.

Dairy cow manure from a manure storage pond was applied with a bulk tank spreader at a rate -of 5000 gal/acre across all 8 plots. The manure contained 6.7% solids. The nutrient content of this manure was 0.22% Nitrogen of which 0.09% was AmmoniaN. The manure application rate was typical of what farmers in the area would normally spread and was selected to meet part of the nitrogen needs of the next years corn crop.

Irrigation at a rate of 0.5 in/hr was initiated on Plots 1, 2, 3, and 4 within hours of manure application and before the manure had dried on the surface. The remaining four plots (Plots 5, 6, 7, and 8) were irrigated I week later, after the manure had an opportunity to dry on the surface. During and following irrigation, the drain discharge was measured and sampled using a grab sampling technique at 15-minute intervals. Irrigation was continued until all the irrigated plots exhibited drainflow.

The typical response of fecal coliform concentrations being discharged from the subsurface drain is shown in Figure I for Plot 4 which was irrigated within hours after the liquid manure application. Plot 4 had returned to a base flow discharge of 5 n-A/s after the pre-wetting irrigation, and this was the condition during the application of the liquid manure. The preferential movement of fecal conform is demonstrated in the figure where the peak concentration of fecal coliform occurred within 40 minutes after the start of irrigation and 95 minutes prior to the peak in the drain discharge. As the drain discharge began to increase, the fecal coliform concentration began to decline where it hovered around 25,000 to 30,000 colonies/100 ml until the irrigation was stopped. Smaller amounts of fecal coliform continued to be discharged during the next 24 hours as the discharge gradually approached the previous base flow situation. The tile drainage effluent during this event exhibited the color and odor of liquid manure, and the discoloration could be directly associated with the changes in fecal coliform concentration.

Ammonium-Nitrogen from the liquid manure was also measured in the drain effluent. The NH4-N in the drain discharge also exhibits preferential flow phenomena, reaching a peak concentration during the rising limb of the drain discharge hydrograph. Ammonium-N is generally believed to adsorb strongly to the soil, but apparently preferential flow processes temporarily dominate this reaction. The similarities of fecal coliform and NH4 -N preferential transport suggest that either one could be used as an indicator of the other.

There is a similar response of fecal coliform bacteria concentrations in the drain discharge when irrigated 6 days after the liquid manure application. There was a natural precipitation event 2 days after the manure application, which had also delivered a total of 0.9 inches. Figure 2 shows the response of fecal coliform concentrations for Plots 5 & 6 which had been pre-wetted along with concentrations for Plots 7 & 8 which had not received the pre-wetting irrigations. In Plots 7 & 8 there was no drain flow during the application of the manure or prior to the irrigation event. For Plot 8 the peak fecal coliform concentration occurred 3.2 hours after the start of irrigation and 60 minutes before the peak in the drain discharge. The peak concentration for Plot 7 occurred 4 hrs. after the start of irrigation. Figures 1 and 2 show similar preferential flow responses despite the events being 6 days apart and the manure having the opportunity to dry on the surface. The only major difference in Figs. 1 and 2 is in the peak concentration of fecal coliform, which is approximately 1/3 to 1/10 less. It is not clear whether this reduction in fecal coliform peak concentration is attributable to the plot variability, the leaching loss from the interim rainfall event or simply due to the 6-day delay after the manure application. Nevertheless, it is still surprising that this high a concentration was still observed given the time, which had elapsed since the initial manure application.

Liquid manure applied to the soil surface at nominal rates, and followed by a precipitation event, can result in bacterial contamination of a subsurface drain in soils which exhibit preferential flow characteristics. Concentrations of nutrients and bacteria will be at peak sooner than the flow of water. The timing of the precipitation event following the liquid manure application will influence the magnitude of the peak concentrations of bacteria such as fecal coliforms. Liquid manure which had dried on the surface did not eliminate the further risk of fecal coliform transport upon rewetting within a 6 day period. An irrigation event on the same day of liquid manure application resulted in a peak concentration of 110,000 colonies/100ml, and an irrigation 6 days after the manure application still resulted in a peak concentration of 38,000 colonies/1001W.

Given a constant precipitation rate, the initial soil moisture content at the time of a precipitation event will influence the length of time which elapses before preferential flow occurs. Liquid manure application to the soil surface when the drain is flowing will result in rapid movement (within minutes) during a precipitation event. For the case when the soil is drier and there is no drain flow, preferential movement of fecal coliform still occurred during a precipitation event when the event was extended for several hours. The peak concentration of fecal coliform did not appear to be affected by the initial soil moisture content.

Surface applied manure potentially can contaminate tile lines and subsequently the surface waters that the tile outlet to. Preventing this from occurring could include reducing the amount of manure applied or plugging the tile lines for a period after application. On most dairy farms in NY these options are not feasible.

Many farms produce 5,000 gal. of manure per acre of cropland. To reduce the amount of manure spread per acre would require more land or less cows. Plugging tile lines will not be effective on slopes exceeding 2%. The tile will either need a plug every 200 ft. or the water will come out around the plug due to head pressure.

One possible solution would be to incorporate the manure as it is applied. By mixing it with the soil it is more likely that the NH4 will be absorbed to the soil. Incorporation should also reduce the amount of preferential flow occuring by breaking up some of the surface flow paths and also by reducing the amount of manure that surface flows are in contact with.

Currently continued research is underway at the Willsboro site to try to discover if incorporation will reduce the potential for tile line contamination.

Figure 2. Fecal coliform discharge concentrations with respect to time for the pre-wetted plots (5 & 6) as compared to the dry plots (7 & 8) during the irrigation event 6 days after the manure application.

To Top