Native grassland productivity is generally nitrogen (N) limited, although the prevalence of phosphorus (P) limitation is increasingly recognized. However, the magnitude of multiple limitation by N, P, and other nutrients remains poorly understood in natural grasslands, despite that they account for approximately one-third of Earth’s terrestrial net primary production. In addition, increasing nutrients loads in grasslands may trigger the loss of plant species through a reduction in light availability, and foster the dominance of invasive exotic species. Here we assessed the co-limitation by nutrients of biomass production and plant species composition through a field experiment in a highly-productive grassland in the Flooding Pampa, Argentina. The design was a complete factorial of nitrogen, phosphorus and potassium +micronutrients addition as main effects, paired with unfertilized control plots (total=8 treatments). Fertilizer (NPK) was applied three times a year to 5 m x 5 m plots established within 3 un-grazed exclosures (total = 24 plots). Biomass production was estimated by clipping 0.1 m2 during peak biomass along 5 consecutive experimental years since 2012. Diversity was estimated by the number of species (richness) and the effective richness (eH) indexes, both calculated from 1 m2 species cover data in spring and summer.
Results/Conclusions
P and N significantly increased biomass production, although P addition showed the biggest increase (25% and 20% for P and N respectively). Indeed, all combinations containing P produced more biomass across the 5 experimental years than controls, but the effects were partially additive. K +micronutrients addition showed no response in plant biomass. The plots that showed the highest biomass increase since the pre-treatment condition were also those containing the highest abundance of Festuca arundinacea, a dominant exotic grass (r2=0.17; p<0.05). The negative effects of nutrients on diversity were stronger with N addition than with any other nutrient. N fertilization reduced plant richness by nearly 5 species/m2 the first year, which represented a 20% loss in richness relative to pre-treatment levels. A structural equation model revealed that plant richness reduction by N and F. arundinacea cover negatively correlates with biomass production. Our results indicate that vegetation production of the Pampas flooding grasslands is more P than N limited, and that co-limitation is not a common phenomenon. The responses to fertilization were led by the promotion of an exotic species which replaced many subordinate and rare native grasses and forbs that might contribute to ecosystem functions in the long term.