E002 - Long-Term Nitrogen Addition to Disturbed Vegetation

Introduction

Established at the same time in 1982, Experiment 002 is a nitrogen addition experiment whose initial design was almost identical to Experiment 001. Like Experiment 001, and often in connection with it, Experiment 002 has been used to examine the effects of low-level nitrogen addition on nitrogen-limited grassland ecosystems. Unlike Experiment 001, however, the native vegetation in Experiment 002 was disturbed via thorough agricultural disking prior to the start of treatments. Also, the Experiment 002 plots were established in only three locations, all of which were successional grassland fields (54 plots each). In each field, nine treatments are imposed on sets of replicate plots (54 plots per field). Two of the treatments are controls, neither of which receives ammonium nitrate fertilization, but one of which receives other nutrients (P, K, Ca, Mg and trace metals). The seven remaining treatments add ammonium nitrate at varying rates as well as the other nutrients. Plots have been sampled for above ground biomass (sorted by species) annually through 2004, and once subsequently in 2008. Plots are sampled intermittently for soil chemistry (NH4, NO3, Ca, Mg, P and K), belowground biomass, light penetration, and small mammal densities.In 1986, aluminum flashing was buried between the individual plots to prevent plants from spreading by vegetative reproduction. Initially, the plot grids were enclosed by fencing to exclude large mammalian herbivores, however in 2004 this

was removed. Gophers are trapped and removed. In 1992, two modifications were made to the Experiment 002 treatment regime. In two of the three fields, cessation treatments were initiated with all nutrient addition being discont

inued for half of the replicates. In the third field, fertilization treatments are ongoing, but half of the replicates are burned annually.

Key Results

Although nutrient enrichment frequently decreases biodiversity, it remains unclear whether such biodiversity losses are readily reversible, or are critical transitions between alternative low- and high-diversity stable states that could be difficult to reverse. Our 30-year grassland experiment shows that plant diversity decreased well below control levels after 10 years of chronic high rates (95–270 kg N ha-1 year-1) of nitrogen addition, and did not recover to control levels 20 years after nitrogen addition ceased. Thus, the regime shifts between alternative stable states that have been reported for some nutrient-enriched aquatic ecosystems may also occur in grasslands. Isbell et al. 2013 Ecology Letters

Using data from the first 25 years of the N addition experiment, Clark and Tilman (2008) found that chronic low-level N addition (10 kg ha-1 yr-1) reduced plant species numbers by 17% relative to controls receiving ambient N deposition (Fig. 3). Moreover, species numbers were reduced more per unit of added N at lower addition rates, suggesting that chronic, low-level N deposition may have a greater impact on diversity than previously thought. Clark et al. (2009) found that net N mineralization rates remained elevated in plots that had ceased receiving N 12 years earlier. Although these grassland ecosystems had not retained a high portion of the deposited N, the effects of this N retention were surprisingly long-lasting.

Because soils are the largest active terrestrial sink of C, the potential effects of elevated N deposition on soil C stores is of great interest. Earlier CDR work suggested that N deposition did not impact soil C stores (Wedin and Tilman 1996). However, we have recently found that 27-years of chronic N addition to prairie grasslands strongly increased the C sequestration in mineral soils (Fornara and Tilman 2012 Ecology). A key mechanism was an N-induced increase in root mass accumulation with a shift to C3 grasses, which despite their lower N-retention ability still acted as important soil C sinks.

We found that although chronic nutrient enrichment initially increased productivity, it also led to loss of plant species, including initially dominant species, which then caused substantial diminishing returns of productivity from nitrogen fertilization. Our results support the hypothesis that the long-term impacts of global changes on ecosystem functioning can strongly depend on how such drivers gradually decrease biodiversity and restructure communities. Isbell et al. 2013 PNAS

We found that the stability of ecosystem productivity was only changed by anthropogenic drivers that altered biodiversity, with a given rate of plant species loss leading to a quantitatively similar decrease in ecosystem stability regardless of which driver caused the biodiversity loss. These results suggest that changes in biodiversity caused by anthropogenic drivers may be a major factor determining how global changes affect ecosystem stability. Hautier et al. 2015 Science

Within a decade after establishment and after undergoing rapid successional changes, plots that shared the same treatment in Experiment 001 and Experiment 002 had converged in composition, plant abundances (Fig 1; Inouye and Tilman 1995). While Experiment 002 plots were initially dominated by annual species, often not of native origin, these were quickly replaced by native perennials.

Figure 1 Graphs of Percent Similarity (PS) vs. time for convergence within fields. Each set of eight graphs presents data comparing disturbed and undisturbed grids in one field, separated by nitrogen treatment (fertilization rates increase alphabetically). Small circular points represent individual comparisons of two plots of the same treatment, one plot on the undisturbed grid and one plot on the disturbed grid. Each graph contains36 such points for each year (six disturbed plots x six undisturbed plots). Stars indicate average PS for each year. For regression analyses, transormation were done on average PS (arcsine square-root) and (year - 1981).

Thirteen years after nitrogen cessation treatments began, a period comparable to the 10 years of elevated nitrogen addition, relative plant species numbers returned to control levels (Fig 2, Clark and Tilman,2008). Plant species abundances, however, were not seen to recover.

Figure. 2 Recovery of relative species number after cessation of nitrogen addition. Relative species number of all plots that continued to receive nitrogen (+N) and of those plots for which nitrogen addition ceased from 1991 and on (-N) is shown as the average across all nitrogen addition levels each year. There were no significant interactions between the rate of nitrogen addition and either year or the cessation treatment. (Clark and Tilman 2008)

Figure 3. Proportional species loss versus nitrogen input rate for (a) 2002-2004 and (b) 1983-1985. Plot averages for each field over the three year-period fitted to a logarithmic curve excluding controls (95% confidence curves included). P values correspond to the significance of the nitrogen input term (N input = experimental N addition + atmospheric N deposition) in a model of the proportional loss of species regressed on the natural logarithm of the nitrogen input rate, Field, and their interaction. Dashed lines correspond to linear interpolation between the mean effect at the highest nitrogen addition rate and controls. (Clark and Tilman 2008)

Methods for e002