In Australia’s northern subtropical cereal belt, soil fertility loss follows conversion into crop production. Soils with longer crop production durations have become more fertility-depleted and hence more responsive to applied nitrogen and phosphorus — the two most limiting nutrients for crop growth in the region. Sustainable crop production requires understand of both annual and long term responsiveness to fertiliser inputs. Though there have been many experiments in individual years, little is known of the long term impact of nitrogen and phosphorus fertilisers on crop production, resource use efficiency and soil carbon in this region’s farming system.
Two long term fertiliser experiments are reported here: Colonsay in southern Queensland, and Myling in north western New South Wales. Both sites are on black Vertosols, the former having a longer crop production history (≈ 44 years at commencement of the study) and hence lower fertility status than the latter (9 years). They aim to assess the effects of long term application of various rates of nitrogen and phosphorus fertilisers alone, and in combination on summer (grain sorghum, mung beans) and winter (wheat, barley and chickpeas) crops. The individual and cumulative production over 17 crops at Colonsay and 9 crops at Myling were assessed. Resource use efficiency of nutrients and water were explored and the fertiliser contribution to soil carbon was assessed using crop inputs estimated from harvest indices where direct measures were unavailable.
Grain yield responses to nitrogen and phosphorus fertilisers were independent of each other for most crops.
On the fertility-depleted Colonsay site, cumulative grain yield in nil nitrogen treatments was 65% of the nitrogen sufficient treatments; a 25000 kg/ha reduction over the seventeen crops. The largest and most consistent responses to fertiliser nitrogen occurred in grain sorghum following grain sorghum grown the previous summer season. Winter or summer cereal crops sown following long-fallow of ≈14 months either had no response or small increases to 40 kg/ha nitrogen application.
Myling in contrast had few large grain yield responses to nitrogen fertiliser. Nitrogen responsive crops were sown under very high cropping intensities with very short fallow periods, or like Colonsay grain sorghum following grain sorghum.
At Colonsay with nitrogen applied at 80 kg/ha per crop, phosphorus fertiliser at 10 kg/ha phosphorus or greater increased cumulative grain yield by 10% (≈ 7000 kg/ha) compared to nil. Crop yield increases were mainly with winter cereals and long-fallow grain sorghum. Cumulative yield increase with phosphorus at Myling was 4.3% (1417 kg/ha) with responsive crops similar to Colonsay.
Precipitation use efficiency (converting rainfall into grain), agronomic efficiency (converting fertiliser nitrogen into grain yield) and recovery efficiency of fertiliser nitrogen into grain nitrogen (getting applied nutrient into the harvest portion of the crop) were all improved with phosphorus application.
Increasing grain yield from fertilised treatments however also increased nitrogen and phosphorus export. Unfertilised (control) plots had on-going removal without replacement, further depleted soil fertility. Some of the nitrogen depletion in unfertilised plots could be accounted for using soil analysis, but the majority of grain nitrogen exported or lost from the system, presumably by other mechanisms remains unexplained.
At Colonsay, an ‘input-output’ mass balance of fertiliser nitrogen applied and that removed in grain from fertilised plots represented the system reasonably well, but was less effective for the higher-fertility Myling experiment.
Grain phosphorus concentration had been measured since 1998. Where nitrogen application increased grain yield, grain phosphorus concentration appeared to be diluted but overall removal was increased due to the higher yields achieved. Phosphorus application increased grain phosphorus concentration of most crops, but the amount of phosphorus removed was not influenced unless grain yield increased. Grain phosphorus concentrations in the literature were compared with these results, and suggested values for use in nutrient removal calculations proposed. These proposed values were used to generate estimates of cumulative phosphorus removal. The relationship between soil testing for fertiliser responsiveness and sustainability of the soil resource was explored. In plots where a phosphorus deficit existed (export > input) the Colwell-P soil test value either remained unchanged in the nil P plots or increased due to fertiliser application in the surface 0.0-0.1 m.
The lower cumulative grain yields at Colonsay without nitrogen input also reduced soil organic carbon in the 0-0.1 m layer compared to nitrogen fertilised treatments. Individual treatment grain yields were used to estimate carbon contribution from above and below ground biomass using harvest indices as an intermediate variable.
The studies reported here have identified areas for future research that will improve understanding of fertiliser use, crop sequences and agricultural systems in northern Australia’s sub-tropical areas. Principal among these are (i) further study of the underlying mechanisms of nutrient availability and crop-soil relationships for use in parameterisation of models of crop production and system functioning, (ii) and the influences of soil chemistry and soil biology on soil nutrient and carbon balances, especially in relation to nitrogen supply and potential for carbon sequestration in soils in future farming systems.