Nutrient Management of Intensively Harvested Oil Mallee Tree Crops

A report for the RIRDC/L&WA/FWPRDC/MDBC Joint Venture Agroforestry Program

by Tim S. Grove, Daniel S. Mendham, Stanley J. Rance, John Bartle and Syd Shea

June 2007

RIRDC Publication No 07/084 RIRDC Project No CSF-63A

Who is the report targeted at?
This report will be of use to landholders and mallee plantation managers in considering how to better manage site nutrition and maintain productivity growth of mallee belts. Use of the prototype toolkit, OMEN, will also be useful for landholders and farm forestry advisors.

Background
Establishment of woody perennial vegetation is an important part of the solution to dryland salinity, which is currently threatening a large proportion of the Australian agricultural landscape. The scale of planting necessary to achieve salinity amelioration requires that adoption of woody perennial vegetation is commercially driven. Woody perennials must be widely dispersed in order to intercept surplus water from adjacent crop or pasture planted areas. This is most effectively done using woody crops in a contour belt configuration. This allows agriculture to continue on interrow lands while adjacent woody perennial belts draw down the saline water table. Potential has been identified in WA for an industry based on oil mallee, and a large investment has been made in extensive planting of oil mallee in belts. Investment has also been made in development of a pilot processing plant to extract eucalyptus oil, energy, and activated charcoal from the oil mallee feedstock. Extraction of these products requires harvest and removal of all above-ground components of the tree. Experience in Australia and overseas has shown that short-rotation woody crops can deplete site nutrient reserves, especially where whole tops are removed. However, sustained productivity of oil mallee crops will be required over many rotations to ensure that the potential environmental and economic benefits are returned to farmers and the community. This project was established to develop techniques to readily assess site nutrient export, and to ensure that appropriate nutrient management can be implemented in oil mallee crops for longer term sustainability of the industry.

Aims and Objectives
The aims of this project were (1) to develop techniques for rapid assessment of the quantity of biomass and nutrients exported from intensively harvested oil mallee crops, (2) to apply these techniques to examine nutrient export scenarios at four representative oil mallee planting sites, and (3) to provide recommendations to ensure longer-term sustainability of the industry.

 Methods Used A series of regression functions were defined using trees from a range of sites within a matrix of 2 focus regions (Kalannie, Narrogin), 2 soil types within each region (high and lower fertility), 2 species within each region (including 1 species common to both regions), and a range of ages and tree sizes.

The 3 species sampled were Eucalyptus kochii ssp. plenissima (Kalanni), E. polybractea (at Narrogin) and E. loxophleba ssp. lissophloia; the latter was in common to both soil types in each region. In total, 139 trees were harvested from 27 sites within this matrix. Each harvested tree was separated into a maximum of 7 fractions. The biomass and nutrient contents were assessed and related to nondestructive measurements through functions. These functions were incorporated into a computational tool, Oil Mallee Estimator of Nutrients (OMEN), for ready application. The OMEN tool was then used to predict belt-level biomass and nutrients at 4 representative focus sites, 2 in the Narrogin district, and 2 in the Kalannie district. Above-ground nutrient contents were compared with the stores in soil to 2 m frequency on site nutrient depletion.

Results and Key Findings
Regressions were developed to predict biomass and nutrients (N, P, K, Ca, Mg, and Na) from both individual stem measurements (heights and base diameters), and from whole-tree measurements (tree height, maximum canopy width and length, height to the base of the crown, number of stems) in uncut trees. Both regression types explained a high proportion of the variation in tree biomass and nutrient content. The individual-stem measurements were more strongly correlated to biomass and nutrients, explaining 94.9-97.1% of the variation in canopy contents of biomass and nutrients, and 95.0-97.6% of the variation in stem (wood, bark, branches) biomass and nutrients. The whole-tree measurements were also highly correlated with both canopy (93.7-97% of the variation explained), and stem (91- 94.8% of the variation explained) components. The factors of region, fertility and species, and their 2nd order interactions were of variable significance, depending on the specific tree fraction and nutrient.

Generally, site factors (fertility, region) showed significant effects on the biomass and nutrient composition of canopy fractions but not on stem and branch fractions. Full models were developed to account for these species, site, and interaction effects, with more than 5000 coefficients being generated within this suite of models. Due to the complexity of this data set, the OMEN computational tool was developed as part of this project. This integrates these models into a useful package and allows the user to enter basic site information (belt length, row width etc.), and tree-based measure information (either individual stem or whole tree). OMEN then integrates this measurement information with the regression model parameters to predict the standing biomass and nutrient content of the belt. The OMEN tool can also be used to predict impacts of variable harvest frequency on site nutrient export.

Regressions were developed based on uncut sapling trees. Their application to coppice regrowth was examined at one site. The coppice material showed broad agreement with model predictions, but foliage concentrations of cations were much lower in the coppice material at that site. Hence the model predictions overestimated the biomass and amounts of nutrients in coppice shoots, and more work is required to allow accurate prediction of coppice biomass and nutrient contents.

Using the OMEN tool, a 5-year old E. polybractea belt at Narrogin was predicted to have a total of 25 tonnes ha-1 biomass, with 189 kg ha-1 N, 12 kg ha-1 P, 87 kg ha-1 K, 241 kg ha-1 Ca, and 25 kg ha-1 Mg.

This contrasted with a 4-year old E. loxophleba ssp. lissophloia tree belt at Kalannie, which had a predicted 4.0 t ha-1 biomass, and 38 kg ha-1 N, 2.5 kg ha-1 P, 14 kg ha-1 K, 49 kg ha-1 Ca, and 4.3 kg ha-1 Mg. The sites were assessed after two very dry years in the Kalannie region, which partly explained the large difference in biomass between the sites.

The quantity of nutrients in the tree belts at the focus sites were also related to the soil stores to 2 m depth. Above-ground nutrient stores were generally around 1% of the soil stores in the higher fertility soils, and up to 5% of the soil stores at the lower fertility sites, suggesting that the lower fertility sites may require nutrient amendment at an earlier stage to ensure optimal productivity over multiple rotations.

Implications

• The allometric regressions derived during this study can be used to estimate the amounts and distribution of biomass and major nutrients, from either whole tree or single stem measures.
• The prototype toolkit OMEN simplifies the calculation of amounts and allocation of biomass and nutrients in tree-belts using tree dimension measurements. OMEN could be used to generate lookup tables for tree size classes for application across the industry. There is potential to further refine and expand the toolkit through incorporation of information from other studies or by linkage to growth model output.
• Relationships developed for sapling trees were found to over-predict the biomass and nutrient contents of coppice regrowth at one site – further refinement of relationships will be required for use with coppice.
• Biomass removal over several harvesting cycles will significantly deplete total site stores of some nutrients unless replacement strategies (return of ash, fertilizer application) are implemented. Site depletion will be moderated by the potentially large depth of root penetration of mallee and the possible large amount of leached nutrients in deep subsoil storage. Site N fertilizing will be required to return N that is lost through burning of the biomass.
• The impacts of tree-belt management on labile (plant available or potentially available) forms of nutrients in soil is more important than the effects on total nutrient balances. Comparisons of nutrient properties of tree-belt and alley soils indicate that concentrations of labile forms of N and P are lower in tree-belt soil. If soil nutrient supply in tree belts is not managed, trees may increasingly rely on nutrients acquired from the alley soils. One form of management may be to provide greater inputs through modification to the annual crop and pasture rotations in the alley where greater proportions of legumes, or greater fertiliser inputs might be used.

Recommendations

• Procedures to assess oil mallee tree-belts and estimate amounts and relative proportions of different product categories will become increasingly important once the integrated processing plant is operational. The models and toolkit developed in this study are appropriate for providing this predictive capability.
• Choice of use of whole tree or single stem model should be based on accuracy required. The whole tree model using tree dimension measurements requires fewer measurements than the single stem model in providing estimates of the biomass and nutrient content of leaf, twig and whole canopy. The single stem model using stem dimension measurements provides more accurate estimates of the biomass and nutrient contents of woody components.
• The models and toolkit which have been developed for sapling trees, will need to be recalibrated for use in predicting the biomass and nutrient composition of coppice regrowth.
• The OMEN toolkit can be used to produce simple lookup tables for ready estimation of the biomass and nutrient composition of different tree size classes. It should be further refined through incorporation of biomass and tree measurement data from other studies.
• Ash return to sites is one way of maintaining site nutrient capital. The ash could be recycled selectively to sites, for example targeting sandy sites with relatively low reserves and acidic soils which may benefit from the liming effect of ash.
• Supplementary fertilizer application, particularly N and P, will be required to sustain high growth rates over several harvesting cycles.
• Knowledge of whether growth of oil mallee will respond to nutrient application on some sites is not available. Establishment of small trials is warranted to develop guidelines to help growers in targeting sites for ash recycling and fertilizer application.
• There is a need to define the size of the deep soil nutrient store and the scope for modification of the adjacent annual plant agriculture to better complement the tree crop by greater input of N from legumes or fertiliser.