Declines in grassland diversity in response to nutrient addition are a general consequence of global change. This decline in species richness may be driven by multiple underlying processes operating at different timescales. Nutrient addition can reduce diversity by enhancing the rate of local extinction via competitive exclusion, or by reducing the rate of colonization by constraining the pool of species able to colonize under new conditions. Partitioning net change into extinction and colonization rates will better delineate the long-term effect of global change in grasslands.

We synthesized changes in richness in response to experimental fertilization with nitrogen, phosphorus, and potassium with micronutrients across 30 grasslands. We quantified changes in local richness, colonization, and extinction over 8-10 years of nutrient addition, and compared these rates against control conditions to isolate the effect of nutrient addition from background dynamics.

Total richness at steady state in the control plots was the sum of equal, relatively high rates of local colonization and extinction. On aggregate, 30-35% of initial species were lost and the same proportion of new species was gained at least once over a decade. Absolute turnover increased with site-level richness but was proportionately greater at lower-richness sites relative to starting richness. Loss of total richness with nutrient addition, especially N in combination with P or K, was driven by enhanced rates of extinction with a smaller contribution from reduced colonization. Enhanced extinction and reduced colonization were disproportionately among native species, perennials, and forbs. Reduced colonization plateaued after the first few (< 5) years after nutrient addition, while enhanced extinction continued throughout the first decade.

Synthesis: Our results indicate a high rate of colonizations and extinctions underlying the richness of ambient communities, and that nutrient enhancement drives overall declines in diversity primarily by the exclusion of previously established species. Moreover, enhanced extinction continues over long time scales, suggesting continuous, long-term community responses and a need for long-term study to fully realize the extinction impact of increased nutrients on grassland composition.

The Nutrient Network (NutNet) is a globally distributed experiment replicated across short-statured, primarily herbaceous ecosystems (hereafter called grasslands) on six continents. Most sites with experimental treatments contain three replicated treatment blocks (21 of 30 sites; range 1 - 6 blocks per site). Within each block, eight different nutrient addition treatments are applied to 5 x 5 m plots. Thus, most sites have a total of 24 experimental units (3 blocks x 8 nutrient treatments; Borer, Harpole, et al., 2014). These nutrient addition treatments represent the factorial combinations of nitrogen (N), phosphorus (P), and potassium and micronutrients (K⁺), plus a control. Nutrient addition rates are: 10 g N m^-2 yr^-1 as timed-release urea, 10 g P m^-2 yr^-1 as triple superphosphate, 10 g K m^-2 yr^-1 as potassium sulfate, and 100 g m^-2 of a micronutrient mix (6% Ca, 3% Mg, 12% S, 0.1% B, 1% Cu, 17% Fe, 2.5% Mn, 0.05% Mo, and 1% Zn). N, P, and K are applied annually, while the micronutrient mix was applied only once at the start of the study to prevent toxicity of largely immobile micronutrients. Treatments are hereafter referred to shorthand by their nutrient components e.g., ‘NK’ for Nitrogen + Potassium and micronutrients, ‘NPK’ for full nutrient additions. All present species are identified in each 5 x 5 m plot every year at the time of peak biomass, or twice annually in highly seasonal sites. For the current study, we only selected sites with at least eight years of treatment data (n = 30 sites) to capture longer-term dynamics in species turnover, as well as control for otherwise uneven sample sizes at disparate time intervals.