Associated with a project that was based upon the assumption that nitrogen may limit net primary plant production in desert grasslands, this project began measuring available inorganic soil N and potentially mineralizable N of soils at two desert grassland locations. Both available N and potentially mineralizable N were greatest following a drought period in 1989, declined during wetter periods that followed and remained relatively stable until another extended drought period. After drought in 1995-6, both forms of soil N increased, indicating the potential for greater NPP following drought and lower potential NPP during periods of normal precipitation.
This project began as part of a study to identify the interactions between soil inorganic nitrogen supply and El Nino-La Nina oscillations that bring wetter-than-average or drier-than-average (respectively) winter-spring precipitation to the central New Mexico region (El Nino-La Nina Fertilization Experiment). This project was known locally as "The Fertilizer Study" and was initiated by Dr. Sandra Turner. To test whether soil inorganic N was limiting aboveground net primary production, fertilizer was added to designated plots within a block of study plots. At first, the soil sampling was designed to confirm or establish the actual fertilizer contribution to each site. The amount added was the amount of inorganic N in soils from the fertilized plots minus that in the control plots. The amount in the controls plots was interpreted as being the amount of inorganic N that had not been utilized by the plant/microbial community and was "available" at that time for utilization in the future. Potentially mineralizable N measurements were performed to determine how the amount of readily mineralizable N was changed by the fertilizer addition; termed a priming effect when the addition of inorganic N results in even greater increases in mineralizable N. Early results from this study were reported at the 1990 Ecological Society of America annual meetings (White et al. 1990).
Following the first year's work, measurements of soil available N and potentially mineralizable N were continued. Except in 1990 when a computer hard-drive that contained most of that year's soil N data crashed and the data were not able to be recovered, the period from 1989 until the spring- summer of 1999 was covered by at least one collection per year.
Each 30m by 30m plot was given a number at the beginning of the study. Before each year's fertilizer application or collection, plots were chosen by selection of random numbers from a table or by using a phone book and using the last 2 digits of the phone numbers. Plots that were used in prior years were excluded from future use. The outer 2.5 m of each plot were left as a buffer strip, so the inside 25m by 25m plot constituted the actual sampling unit.It is not known if the fertilizer was applied to the 30 x 30 m plot or only to the 25 x 25 m plot excluding the buffer. Thus, values given below are based upon application to the 30 x 30 m plot and should be used cautiously.
In 1989, fertilizer treatments were applied as NH4NO3 in 2 treatment levels with controls (low, high, and no fertilizer). Three plots received 61.8 kg N/ha, three plots received 30.9 kg N/ha, and 3 plots were not fertilized (controls). All of these plots were selected in a random manner. It turned out that plot 7 on the west side was too disturbed to be used, so plot 51 was substituted. On the east side, plot 66 was mistakenly fertilized instead of plot 56, which had been selected, so that plot 66 became the 3rd heavily fertilized plot. Hence the plots used in their respective treatments were as follows:
Treatments East side West side
---------- ---------- ----------
Control 16, 33, 42 16, 33, 42
Low 7, 10, 25 19, 25, 51
High 1, 26, 66 1, 26, 56
The fertilizer was applied in April on the east and west sides. Fertilizer was applied with a cyclone spreader by a person walking 5 equally spaced paths across each plot while cranking the spreader. The light treatment was applied with passes in only one direction, while passes were made in both directions across the plot for the heavy treatments.
Soil samples were collected on April 18, 1989, to verify the nitrogen additions and to measure net N mineralization and nitrification potentials. Soil cores (4.2 mm in diameter) were taken to a depth of 20 cm at 4 m intervals along a diagonal from the NE to the SW corner of the plot (starting at the edge of the 2.5 m buffer zone). These cores (15 in all) were placed in the same bag and comprised one sample. The 3 plots represented the sample units (n=3). Plot 66 was not collected until July 4th, when the mistake was identified because plot 56 had no elevated levels of nitrate or ammonium.
In 1990, it was decided that the number of treatment levels would be reduced to one fertilizer application rate and that the number of replicate treatment plots be increased to 4. Fertilizer was applied in two applications, instead of the single application in 1989, in attempts to reduce "burning" from the fertilizer. The first application was on March 1, 1990, and the second was April 4, 1990. Over an inch of rain occurred on both the east and west sides between applications. For each application, fertilizer was applied with a cyclone spreader by a person walking 3 evenly spaced passes across each plot in one direction, and then repeating the process in the perpendicular direction. The setting on the cyclone spreader was 3 1/4. The plots used were:
Treatments East side West side
---------- -------------- --------------
Control 24, 34, 43, 48 24, 34, 43, 48
High 3, 22, 30, 60 3, 22, 30, 60
Within each 30m by 30m plot, 3 ea. half-meter-square (1/2 or 0.5 m^2) quadrats were randomly selected for soil sampling within the northwest quarter of the plot (the 12.5 m by 12.5 m portion excluding the 2.5 m buffer). To identify the location of the 0.5 m^2 quadrats, two tape measures were run parallel to each other, one along the top of the plot and the second along the mid-line of the NW quarter. A third tape was run between the two tapes on the edges to locate the long side of the randomly selected 0.5 m^2 quadrats. Along each diagonal, 4 soil cores were taken at equal spacing starting with the corner of the quadrat and ending in the diagonal corner. Soil cores (4.2 mm diameter) were taken to a depth of 20 cm. The 4 cores from the quadrat were placed in the same bag and comprised one sample. Thus, each plot had 3 soil samples, one from each quadrat, with each treatment having 4 plots.
Protocols for 1991 through 1994
Starting in 1991 and continuing through 1994, 4 plots were sampled with 3 quadrats in each plot; however, only three soil cores were collected (2 from the opposing corners and one from the middle of the quadrat) per quadrat. As in the other years (except 1989 where only one sample was taken per plot), soils from each quadrat were analyzed separately and later averaged to get the value for the plot (3 replicate soil analyses, one per quadrat, and these values were averaged to get a plot value for each of 4 plots; n=4).
Protocols for 1995 to present: Soil Bridges and associated soil collections
In 1995, 5 soil bridges were installed about 30 m east of the Fertilizer Plots. The primary function of the soil bridge is to measure small changes in soil microtopography; erosion or deposition. At the time of soil bridge measurements, two soil cores (same dimensions, 4.2 mm diameter to a depth of 20 cm) for initial available N and potentially mineralizable N were collected from vegetated (mostly grass, 2 cores each) and unvegetated (2 cores each) areas that appeared comparable to those areas beneath the soil bridge. The two sample-types (under plants or open soils) were analyzed separately and the "average" for the bridge was determined by weighting the soil- types by their respective coverages determined beneath the bridge. For example, if vegetation cover was 35% (0.35 of area beneath the bridge had vegetation) and bare soil 65%, the N values for the soil beneath the vegetation were multiplied by 0.35 and summed with the N values for the bare soils multiplied by 0.65. This procedure gave 5 average values for soil N at each collection.+
In the summer of 1998, a study was performed to make direct comparison of the two different collection techniques (collection of soils from fertilizer plots or soils associated with the bridges) and the potential differences the collection method could have on soil nitrogen mineralization potentials.
Methods used for comparative study
Collection on fertilizer plots (fp): Four plots were randomly selected as per previous methods; 4 each 30 X 30 m plots within a 300 X 300 m study site. Within each plot, three randomly selected 0.5 m^2 quadrats within the 15 X 15 m NW quarter of the plot were selected. Within each 0.5 m^2 quadrat, a soil core (4 cm diameter by 20 cm long) was taken from the NW and SE corners of the plot. Both cores were composited to make one sample from each quadrat. This sampling method gave 3 soil samples for each of four plots, for a total of 12 samples.
Collection from the bridge sites: The 5 bridges were equally spaced along the 300 m east border of the fp study area. At each bridge, 2 soil sample cores (4 cm diameter by 20 cm long) were taken from under vegetation cover (under) and 2 from non-vegetative soil surfaces (open). The two cores were composited to give a single soil sample from under and open at each bridge.
Analyses of N mineralization potentials: After determining the water holding capacity (WHC; White and McDonnell 1988), a portion of each sample was adjusted to 50% of determined WHC and 5 subsamples were apportioned into plastic cups. Each cup contained approximately 30 g (+/- 0.05 g) dry-weight mineral soil. One subsample of each sample was immediately extracted with 100 ml 2 N KCl for NH4+-N and NO3--N analyses. The remainder of the cups were covered with plastic wrap, sealed with a rubber band and incubated in the dark at 20 degrees C. Moisture content was monitored by mass loss and replenished as needed. At days 14, 28, 35, and 42, one subsample of each sample was removed and extracted with KCl for 18 - 24 hours. The clarified KCl was filtered through a Kimwipe and analyzed on a Technicon AutoAnalyzer.
Fp: Extractable N at each time interval was averaged for the three soil samples from the 0.5 m^2 quadrats per plot to yield a mean value for each plot. The mean values then were used for statistical analysis (n=4).
Bridges: The relative cover for under and open was obtained from the most recent bridge measurements (made within 2 weeks of soil collections). Extractable N in each soil sample was multiplied by the respective relative cover, and N conc.-times -cover values for soils collected at each bridge were summed to give a cover-weighted value for each bridge (n=5).
The extractable N values for the four fp's were compared to the five bridges over the entire extraction period to determine differences between collection techniques (Factor 1), the effect of extraction time (Factor 2), and their interaction (Factor 1 X 2) at each time interval by analysis of variance. In addition, the variance associated with each collection method was determined and analyzed by analysis of variance to determine if the techniques significantly affected the variance associated with each technique.
The two collection techniques were not significantly different for N mineralization (P = 0.47). There was no significant incubation time x collection technique interaction (P = 0.87). The bridge method did increase variance slightly, but not significantly (P = 0.114).
Net N Mineralization Potentials by different Collection Techniques:Factor d.f. S.S. M.S. F-test P-value------ ---- ---- ---- ------ -------Collection 1 0.316 0.316 0.531 0.47TechniqueExtraction/ 4 330 87.5 147 0.001Incubation timeColl. X time 4 0.719 0.18 0.302 0.87error 35 70.8 0.596 Net N. Mineralization Potentials for Soil Collected by Different Techniques (sum ammonium and nitrate; mg N / kg soil).Incubation Bridge Collections Fertilizer Plots CollectionsTime (days) Mean S.E. Mean S.E.----------- ------------------ ---------------------------- 0 1.96 0.43 2.02 0.24 14 5.55 0.804 5.28 0.45 28 7.33 0.77 7.6 0.17 35 8.29 0.74 8.62 0.63 43 9.59 1.34 10.04 1.12Soil Collection MethodsAll soil samples were taken with a 4.2 mm diameter soil corer to a depth of 20 cm and placed into an ice chest and transported on ice to the University of New Mexico, where they were sieved (2 mm), obvious live roots removed, and stored at 5 degrees C.Soil Analysis MethodsSoil Moisture and Organic MatterFor each collection, soil moisture content of each fresh sample was determined by mass loss upon heating at 105 degrees C for 24 hours. Organic matter was determined by loss-upon-ignition from oven-dried samples placed in a muffle furnace and brought to 500 degrees C for 2 hours. Soil Texture and Total N and P.At select times, texture (percent of sand, silt, and clay) was determined by the hydrometer method (Day, 1965). Total nitrogen and phosphorus were determined by Kjeldahl digestion with copper sulfate as catalyst (Schuman et al. 1973) followed by analysis of ammonium by an automated phenolate method (Technicon AutoAnalyzer Industrial Method #19-69W) and orthophosphate by an ascorbic acid method (Technicon AutoAnalyzer Industrial Method #94-70W).Water-holding capacityWater-holding capacity (WHC) was determine by saturating about a 50 g portion of the sieved soil contained in a funnel with DI water, allowing to saturate for 30 min., and then to drain by gravity for 30 min. The drained soil was transferred to pre-weighed soil tins and dried in an oven at 105 degrees C for 24 hours. The water lost upon drying was the water-holding capacity of the soil sample.Field available N and N mineralization/nitrification potentialsAfter determining water-holding capacity (WHC), a portion of each sample was adjusted to 500f determined WHC and subsamples were apportioned into plastic cups. Each cup contained approximately 30 g dry-weight mineral soil. One subsample of each sample was immediately extracted with 100 ml 2 N KCl (with 5 ppm phenyl mercuric acetate as a preservative) for NH4+-N and NO3--N analyses. The remainder of the cups were covered with plastic wrap, sealed with a rubber band, and incubated in the dark at 20 degrees C. The plastic wrap minimized water loss during incubation, yet exchange of CO2 and O2 was sufficient to keep the subsamples aerobic during incubation. Moisture content was monitored by mass loss and replenished as needed. At approximately weekly intervals, one subsample of each sample was removed, 100 ml KCl was added and shaken with the plastic wrap coveringthe top, and allowed to settle for 18-24 h. The clarified KCl was filtered through a Kimwipe and analyzed for NH4+-N and NO3--N+NO2--N on a Technicon AutoAnalyzer (Technicon, Terrytown, NY) as described in White (1986). This procedure gave up to three different measurements for inorganic ammonium and nitrate: The initial extractions (time 0, wetted but not incubated) indicated field available concentrations (their sum gave total field available N); The amount of ammonium and nitrate extracted after 35-day incubation; and The maximum sum of ammonium and nitrate in any extraction up to the 35-day incubation (some samples peaked prior to day 35).
File assembled 1/19/2001 JAC. File updated with data through 2005, 10/25/2005 Carl White.
File updated with data through 2008, 5/20/2009 Carl White.
Additional Information on the Data Collection Period
Data collected four times per year (usually January, April, July and October). Sampled one week per season, four seasons per year.
Site DescriptionThe fertilizer study was initiated at two locations on the Sevilleta NWR; one on the east side and one on the west side of the refuge. The east side study site was located within a 300m by 300m block on MacKenzie Flats. The vegetation was dominated by black grama and was thought to be a "typical" black grama desert-grassland. Within that area were 100 ea. 30m by 30m plots (10 plots along the side). The outer 2.5 m of the plot were left as a buffer strip, so the inside 25m by 25m plot constituted the actual sampling unit. When the fertilizer study was no longer active, the plots were sampled to continue the long-term soil record. In 1995, 5 locations along the east 300 m border of the fertilizer study area were established for future soil monitoring. Each of these new sites had soil erosion bridges installed to measure changes in soil microtopography (erosion/deposition). In the methods section are the results of a study that verified that sampling at the 5 new locations gave results comparable to those from the 4 random plots within the fertilizer study area. In 2003, the 5 new locations of 1995 were burned by prescribed fire on 19 June 2003. Additional bridge and soil sites were established within the original fertilizer study area (established in control or untreated plots) and sampled before the prescribed fire to establish similarities to continue the control or untreated soil collections at that area. These new bridge sites are numbered 7196, 7197, 7198, 7199, and 7200.The original study also had a grassland site on the west side of the Sevilleta NWR. This site was west of the Field Station, accessed by entrance at the north gate near Bernardo and following the roadbed of the old highway alignment. This site was dominated by C-3 grasses with C-4 grasses a minor component. Shrubs were primarily 4-wing saltbush. The west side only had room for 61 each 30m by 30m plots. When the fertilizer study was ceased, no further soil collections were made on the west side for long-term study.
Method ReferencesDay, Paul R. 1965. Particle Fractionation and Particle-size Analysis. IN: C.A. Black (ed). Methods of Soil Analysis. Part 1. American Society of Agronomy, Inc., Madison, USA.
Schuman, G.E., Stanley, M.A., and Knudsen, D. 1973. Automated total nitrogen analysis of soil and plant samples. Soil Sci. Soc. Am. Proc. 37:480-481.