Long-term annual nitrogen fertilisation of Eucalyptus regnans F. Mueller and Pinus radiata D.Don : effects on tree growth, soil chemistry and net nitrogen mineralisation

In the current competitive market for land in Tasmania, Australia, economic forest production may require large nutrient inputs to optimise productivity per unit area of land. Nitrogen (N) and phosphorus (P) fertilisers are often required at planting and in early stages of tree establishment to achieve rapid early growth and high survival rates. In Tasmania, further application ofN and P to plantations ranging in age from 2- to 20 years has also occurred. To effectively manage these plantations, a detailed understanding of nitrogen (N) fertiliser requirements and N retention in forests is required. Two field fertilisation experiments were used in this study, one in a 20-year-old Pinus radiata D.Don plantation growing in the north-east of Tasmania, on a Yellow Kurosol, and one in a 5-year-old Eucalyptus regnans F. Mueller plantation growing in the south, on a Brown Ferrosol. Both of these experiments were established in the early 1980s. Treatments at both sites included various combinations of P and N, applied as single-superphosphate and ammonium sulphate through a period of up to thirteen years. This study examines fertiliser-use, efficiency and impact on forest sustainability, including a detailed examination of the soil profile and litter at both sites. Nitrogen cycling was also examined, concentrating on the effect ofN fertilisation on N mineralisation in the contrasting surface soil horizons. After fifteen years of measurements, nitrogen fertilisation significantly increased volume growth at both sites. Two single applications of Palone (totalling 144 kg P ha\\(^{-1}\\)) doubled P. radiata stem volume from 78 m\\(^3\\) ha\\(^{-1}\\) (Nil) to 192 m\\(^3\\) ha\\(^{-1}\\) (P). Annual N fertilisation for a period of thirteen years (in addition to the P fertiliser) further increased P. radiata stem volume from 192 m\\(^3\\) ha\\(^{-1}\\) (P) to 344 m\\(^3\\) ha\\(^{-1}\\) ((P)Nl Y), at age 34 years. In contrast, applications of P alone (up to a total of 598 P ha1 ) had no effect on E. regnans growth, while annual N (plus P) fertilisation for thirteen years, doubled E. regnans growth from 125 to 281 m\\(^3\\) ha\\(^{-1}\\), at age 19 years. Although fertilisers may be used to increase forest growth there is concern that longterm N application may impact on forest sustainability through changes in the soil chemistry. At both sites in this study, significant changes occurred in the soil profile due to long-term fertilisation. Soil pH decreased due to both N and P fertilisation, at both sites. Significant reductions of0.7 and 0.3 of a pH unit were associated with the highest rates of fertilisation in topsoil (Al, 0-10 cm) and litter (02 horizon), respectively. Substantial reductions in exchangeable Mg concentrations were also measured, particularly in the Kurosol. In association with enhanced growth was a large increase in litter accumulated at both sites. Total litter masses (01 +02) ranged from 34.4 to 91.6 t ha\\(^{-1}\\) under P. radiata and from 21.6 to 102.4 t ha\\(^{-1}\\) under E. regnans. At both sites, the 02 horizon masses were significantly greater with annual fertilisation and were a substantial nutrient pool. Under P. radiata, 02 horizon mass was 40 t ha\\(^{-1}\\) when unfertilised and over 70 t ha\\(^{-1}\\) with fertilisation, while under E. regnans the mass was 14 t ha\\(^{-1}\\) when unfertilised and 77 t ha -l with fertilisation. This indicates that, in the cool temperate climate studied here, litter could be an important pool of nutrients. Long-term, annual applications of N fertiliser had no significant effect on the annual rate ofNNM measured in either the Kurosol or Ferrosol topsoil (0-10 cm). However, average rates in Kurosol topsoil were up to four-fold higher in the annually fertilised treatment. At both sites topsoil in situ net N mineralisation (NNM) rates measured at the end of the experiment were low, ranging between 13 and 52 kg N ha\\(^{-1}\\) yr\\(^{-1}\\). Such low rates of N mineralisation might have been associated with a prolonged period of low rainfall that occurred throughout the 18-month measurement period. To assess mineralisation independent of microclimatic effects that prevailed during the in situ study, rates of NNM were measured during aerobic laboratory incubations. In agreement with in situ studies, NNM in the Ferrosol topsoil was not changed by fertilisation. In contrast, the variation in rates ofNNM between treatments for the Kurosol topsoil was greater than that measured in situ, with fertilised topsoil mineralising ten times more N than that unfertilised. However, results were highly variable across moisture and temperature treatments. Despite the high amounts of N and P that had been applied during annual fertilisation, differences in total N and NNM in soil were small and highly variable. This result contrasted with the large differences in total N content and NNM rates in the 02 horizon from both sites. The influence of fertilisation on N cycling in litter produced clear results, i.e. daily rates of mineralisation (measured by aerobic laboratory incubations) were higher in annually fertilised than unfertilised litters, at both sites if incubations went for 7 days or longer. In contrast, low mineralisation rates in both topsoils often produced similar daily rates of NNM regardless of fertiliser treatments. In both topsoils, 60 days was required to produce a significant cumulative effect, resulting from a divergence in NNM rates between the fertiliser treatments during the later stage of the incubation. Increased replication of samples in laboratory experiments did increase the sensitivity of NNM measurements and significant differences between fertiliser treatments were measured. These results confirm the importance of litter as an N source in cool temperate plantations. The importance of the litter layer in N cycling was particularly evident under E. regnans, where N was most concentrated in the litter layer. In addition, the E. regnans 02 horizon accumulated significantly more P, S and Ca due to annual fertilisation. The effects of air-drying, incubation period, moisture content and temperature on NNM in laboratory studies were examined and found to depend on both the site and fertiliser treatment. This study indicated that higher NNM rates in topsoils would occur if soils were not maintained moist prior to and during incubations. This was particularly important for wetter sites, where canopy closure had occurred, resulting in smaller moisture fluctuations, as observed in the Ferrosol topsoil. Laboratory incubation conditions also influenced correlations between nutrient content (total N, P Sand Ca) and NNM. For example, NNM rates in the Kurosol topsoil were linearly correlated with N, P and Mg concentrations and pH when incubated at 20°C (p <0.05), but only with P at 10°C. This study indicated that minimum disruption of soil processes, particularly by drying was essential if accurate measurements of changes in soil N, in the cool temperate environment, were to be obtained. Hence, I could not identify a reliable indicator ofNNM for these plantations. In agreement with the in situ study, application ofN fertiliser seasonally (June, October, January or April) resulted in a short-term elevation of mineral N (less than six-months), particularly at the wetter Ferrosol site. April fertiliser application provided the longest period of enhancement and October the shortest, indicating that the current operational practice, of applying fertiliser in autumn (March-May), provides an adequate window for fertiliser uptake to occur. These trends also confirm that, independent of the time of fertiliser application, a six-month delay after fertiliser application was adequate to determine long-term fertiliser affects on NNM. This delay allowed mineral-N concentrations to attenuate to a low value that facilitated NNM measurements. In agreement with previous studies, at both sites rainfall appeared to be a strong regulator of mineral N availability after fertilisation. The importance of the litter was also highlighted, because it retained N fertiliser and thereby limited N leaching. In situ rates ofNNM and tree growth in the nil and annual fertilisation regimes in the P. radiata plantation were modelled using the process-based model CABALA. The model was developed with forest managers in mind as part of a silvicultural decision support system and links C, water and N flows through the atmosphere, tree and soil. In this study, CABALA was validated for P. radiata using parameterisation from published and unpublished data derived from an independent study. Predicted growth increases due to annual fertilisation within 15 % at age 34 years. Limitations to the current simulation were often due to the assumption that P was not limited at this site. Other responses to N movement in the forest system were also adequately predicted. However, this study indicated that functions for N mineralisation and canopy development need to be more sensitive to fertiliser inputs to adequately predict N availability in the mineral soil and litter layers. There currently seems to be an overdependence on C: N ratio, because it drives predictions ofN mineralisation in the CERES model, but there is doubt that this is an important controlling variable in forest systems. Further study of in situ NNM rates of litter horizons would be required to clarify the amount and mechanisms of N recycling that occur in these systems. Large growth rate increases from N fertilisation in both P. radiata and E. regnans plantations were often associated with significant changes in soil chemistry and litter accumulation. Substantial reductions in exchangeable Mg concentrations and soil pH indicate that careful site management is required. Significant accumulation of litter under these plantations may act a substantial future source of nutrients, including N availability for further tree growth, particularly when ...