Monitoring the Effects of Pasture Fertilization on Soil Water in the Ferrum Watershed
Background
Agriculture is a major contributor of nonpoint source pollution in southwest Virginia where dairy and beef farms are relegated primarily to low-lying areas in the watershed. Both pastures and hay producing fields are often fertilized with recommended amounts of fertilizer to ensure maximum forage growth and improved quality with less regard to leaching and runoff. Our objectives were to compare nitrate and phosphate levels in soil water from fertilized and unfertilized pastures in the Ferrum watershed and to relate these data to stream samples above and below the pastures and baseline soil water data obtained from wooded plots above the pastures. In addition, surface litter decomposition was monitored in the two fields.
Study Site and Methods
Porous cup soil lysimeters were placed in the Ferrum watershed to compare the fate of nitrate between a fertilized and unfertilized pasture. Ferrum Creek bisected the two pastures, which consisted of primarily Tall fescue (Festuca arundinacea L.). Orchardgrass, redclover, white clover, and various annual weeds were also present in sparse amounts. One pasture (~ 0.9 hectares) was fertilized with 80 kg NPK per hectare on the following dates: May 11, 2004, July 16, 2005, and July 12, 2006. The other pasture (~1.2 hectares) was left unfertilized. Historically, the pasture had been used for grazing horses, but had not been grazed for over six years. Since animal removal, the area had been mowed once or twice per year with no hay removal except June 2004. For background measurements, five lysimeters were installed in an elevated portion of the watershed above the pastures in wooded plots designated Plot 1 and Plot 4. In each pasture, five soil lysimeters were placed along a diagonal transect 10 meters apart ranging approximately 20 to 40 meters from the creeks edge. Because of dry conditions during Spring 2005, no soil water from the pastures could be extracted from the lysimeters. Therefore, during June, 2005, the lysimeters were moved to within approximately 10 meters of the creek at the tree canopy's edge.
Lysimeter Installation
The 61 cm porous cup lysimeters were installed as follows:
For more detailed instructions on installation, see the Soilmoisture Equipment Corp website (1).
Soil Water Collection
Soil water samples were collected as follows:
Chemical Analyses
Aboveground Biomass and Tissue Nitrogen
Aboveground biomass was determined during the summer 2006 by removing stems and culms from six randomly selected areas within each pasture using small circular quadrats (0.056 square meters). Measurements of biomass began immediately after mowing plots and fertilizing plots and continued throughout the remainder of the season. Harvested plant samples were placed in bags and dried in an oven at 90 degrees C to a constant weight. A small subsample (approximately 8 g) of each dried sample was ground in small grinder. A 1.0 g sample was digested and analyzed for total N.
Surface Litter Decomposition
To determine the effects of fertilization on surface litter decomposition in the pasture community, plastic mesh litter bags were assembled. Tall fescue leaves from both pastures were collected and dried in an oven at 90 degrees C to constant weight. The dried leaves were separated into ten gram piles and randomly placed into pre-weighed litter bags, which were sealed using a heating iron. The bags were placed at surface level adjacent to each soil lysimeter in each pasture, secured with a wire staple, and covered with thin layer of soil. Litter bags were periodically removed from the pasture, cleaned of soil and debris, dried and weighed. After weighing, the litter bags were returned to the field for future sampling.
Results and Discussion
Soil water phosphate levels were monitored periodically (Fig. 1). Besides the initial high levels found in the July 2, 2004 samples, soil water phosphate levels ranged form 0.011 to 0.042 mg/L.
Nitrate levels in the soil water extracted from the elevated and wooded portion of the watershed (Plots 1 and 4) were generally low throughout the the study (Fig. 2). The exception was Plot 4 for the July 2, 2004 sampling date in which case levels were as high as those found in the pastures. Nitrate levels in both pastures were consistently higher than background for each sampling date (Fig. 2 and Fig. 3). Higher soil nitrate levels were expected for three possible reasons:
There were significant differences in soil water nitrate levels between fertilized and unfertilized pastures for several samples collected on the same day, particularly those sampled just following pasture fertilization and when significant rainfall had occurred. Nitrate levels between the two pastures, however, returned to near baseline levels with much less difference, if any, between them.
For a better comparison between the fertilized and unfertilized pastures, mean soil water nitrate levels for each pasture and each sampling date were indexed to the baseline mean, which can be defined as the combined mean soil water nitrate levels for Plots 1 and 4 for that sampling date. Each mean soil water nitrate concentration was simply divided by the baseline mean. A clear effect of applying fertilizer to a pasture is shown in Figure 4, which results in a pulse of soil water nitrate when compared to the unfertilized pasture. The added fertilizer, however, appeared to have little effect on stream water nitrate (Fig. 4). Nitrate indices for stream water above and below the pasture differed only slightly with water below the pasture not significantly different from that above.
Surface litter decomposition differed between the two pastures over the sampling period from 50 to 220 days following initial placement (Fig. 5). During those times, a faster decay rate was found in the fertilized pasture compared to the unfertilized one. A greater potential source of nitrate and phosphate pollutants from pastures may result from surface runoff rather than ground water. Although no attempts were made to monitor surface water constituents, the differences in the decay rates shown here indicate that grass leaves may senesce faster in fertilized filed, releasing more nutrients onto the surface of the soil with greater potential loss from surface runoff. Since fertilized leaves are likely to have greater nutrient content, the faster decay rate may have a compounding effect on adding these pollutants to our streams.
We expect fertilizer additions will have a significant impact on plant growth. We will complete our analysis soon to see if our prediction is correct.
WIthout significant rain events that promote surface loss, grasses appear to capture nitrogen from the soil very well, although pulses of nitrogen entering streams and creeks may be difficult to monitor or detect by our sampling methods. In any case, results so far indicate a quick return to baseline of nitrate after fertilization.
