Responses of a South African mesic grassland to long-term nutrient enrichment and cessation of nutrient enrichment.

Abstract

Nitrogen and phosphorus are two of the most important limiting nutrients required for plant growth and production within grasslands. South African ecosystems are generally considered to be nutrient poor with acidic soils. Increased soil acidity exacerbated by nutrient enrichment are expected to transform South African grasslands over a long period. This PhD research project investigated the cumulative effect of 70 years of nitrogen (N) – in the form of limestone ammonium nitrate (LAN) and ammonium sulphate (ASU), phosphorus (P) – in the form of superphosphate and dolomitic lime application on the Ukulinga Grassland Nutrient Experiment (UGNE). The UGNE has been in operation since 1951 and is located on the plateau at the Ukulinga Research Farm, KwaZulu-Natal, South Africa. Here, the purpose was to evaluate potential changes in above-ground net primary productivity (ANPP), species composition, soil variables and species diversity. In the first experiment (Chapter 2), nutrient-enriched plots were abundantly dominated by taller grass species like Megathyrsus maximus rather than shorter species like Themeda triandra and Tristachya leucothrix. Limestone ammonium nitrate and P did not affect species richness as strongly as ASU, suggesting nutrient identity to be an important factor to consider. Ammonium sulphate enrichment was also associated with low soil pH, enhancing these plots' elevated aluminium (Al) concentration. Considering this key result, careful monitoring of soil pH and Al concentration is required moving forward as Al toxicity may cause a threat to sensitive plant species. To further emphasize the negative effects of soil acidification induced by N on the UGNE, the greenhouse pot-trial experiment (Chapter 5) revealed that the N-enriched soils are extremely acidic and P- deficient, further stunting nodulation development in Vachellia sieberiana saplings. Vachellia sieberiana is a common nitrogen-fixing species that encroaches mesic environments in South Africa, but its competitive ability may be reduced within grasslands under extremely acidic conditions. Phosphorus enrichment can replenish important soil nutrients and further improve soil fertility. The effects of long-term P enrichment (70 years) and short-term cessation of P enrichment (3 years) has received little to no attention in South African grasslands, therefore this was investigated in the second experiment (Chapter 3). Phosphorus is an important macro-nutrient and essential for biological nitrogen fixation (BNF). Therefore, it is expected that long-term P enrichment would modify soil properties and indirectly influence plant cover percentage and composition. What was unknown was if the cessation of the P enrichment would revert plant cover and composition towards untreated/control conditions. In chapter 3, the results revealed that changes in plant cover among the three treatments of control, P enrichment and cessation of P enrichment were consistent. Furthermore, no difference was observed among the treatments for the ANPP, species diversity indices and richness. However, the Indicator Species Analysis proved that Setaria nigrirostris and Bidens pilosa were strongly associated with the cessation of P enrichment treatment. The third experiment (Chapter 4) provided the opportunity to assess the best models that predict species compositional changes and species richness declines. Here, it was clear that the enrichment of more nutrients shifted both grass and forb species composition. For grass species composition, the shift in cover-abundance was from the shorter Tristachya leucothrix (no nutrients added) species to Aristida junciformis (only LAN added) and Megathyrsus maximus (LAN + P added). For forb species composition, Cephalaria pungens occurred in a high cover-abundance in control plots and Bidens pilosa occurred in LAN + P enriched plots. Interestingly, LAN-only and LAN + P enriched plots showed the greatest shift in both grass and forbs when compared to ASU-only, ASU + P and P-only enriched plots. In terms of overall species declines, the models presented identified the number of nutrients added and ANPP as the best predictors. Light was not identified as a significant predictor. The results provided partial evidence for the nitrogen detriment hypothesis and biomass-driven hypothesis over the niche dimension hypothesis. The key findings from all the experiments highlight that the following factors: 1) nutrient-poor soils, 2) type of nitrogen used for enrichment, 3) soil acidification and 4) high soil aluminium concentrations are important in the observed changes in the soil chemical properties, species composition, species diversity, species richness and species growth dynamics on the UGNE. The results of this research project also emphasize the importance of long-term studies in assessing if the cessation of nutrient enrichment is a strategy for ecosystem rehabilitation. It was identified here that long-term nutrient enrichment heavily modifies a mesic grassland community and alternative rehabilitation methods may need to be implemented.