Responses of grazed New Zealand hill pastures to rates of superphosphate application

Effects of superphosphate application rates on pasture production were measured between 1975 and 1988 within 12 grazed farmlets on steep hill country at the Ballantrae Hill Country Research Station in southern Hawke's Bay. The low fertiliser (LF) treatment received 125 kg ha−1 superphosphate pa, and the high fertiliser (HF) an average 625 kg ha−1 pa from 1975 to 1979, then 375 kg ha−1 pa subsequently. No fertiliser was applied to one LF (LFNF) and one HF (HFNF) farmlet after 1980. Soil Olsen phosphorus (P) level declined slightly for the LF treatment as maintenance requirements exceeded P inputs; rose in response to increased fertiliser inputs in the HF treatment but more slowly than anticipated maybe because of the poor fertiliser application history and low starting Olsen P; and declined in the HFNF treatment after cessation of fertiliser inputs. Pasture production was 19% greater in HF than LF in year 1, and on average 39% greater in years 2 to 13. Size of earthworm populations measured up to 30 years after treatments commenced was highly related to level of pasture production, and there was no evidence of deleterious effects of superphosphate application on earthworm populations. Pasture production declined by 3% pa in HFNF after fertiliser application ceased. Variability of annual pasture production was large, year-to-year variation about the mean being 35% for HF and 47% for LF (CV = 8.5% and 12.7%, respectively). This suggests comprehensive farm system redesign will not be required in the near or medium term to cope with the gradual onset of climate change, as farmers already manage large year-to-year variability with existing systems and their careful implementation should suffice. Estimated maximum annual production with non-limiting P supply for low-slope (1–12°), medium-slope (13–25°) and steep (>25°) sites was estimated to be 14.8, 11.5 and 9.4 t DM ha−1 from curves derived using Bayesian statistics. The Olsen P level for 97% maximum pasture production was 18 µg g−1; converted to a volumetric basis this equates to about 15 µg mL−1 which is lower than the commonly accepted industry guideline for near maximum pasture production of 20 µg mL−1.