What the report is about
This study confirms that the dynamics of salt in the rootzone of trees over shallow, saline watertables is governed by a complex interaction of factors, principally soil hydraulic properties, groundwater depth and salinity, sensitivity of plantation species to salt, soil nutrient status, and plantation management. While some combinations of conditions may lead to salt accumulation in the rootzone, others can maintain salt load and its distribution in the profile in equilibrium, allowing long term sustainable timber production. In certain situations, management interventions can promote salt redistribution back down the profile below the active rootzone.

Who the report is targeted at?
The results of this work are relevant to any stakeholders with an interest in or responsibility for the sustainable management of commercial plantations over shallow saline watertables.

Background
Much of the land threatened by salinity over the next 50 years is classed at risk on the basis of watertables rising to within 2 m of the land surface. There is a need for direct management of areas with shallow watertables to prevent the degradation and loss of production, associated salt accumulation at the surface, and salt discharge into rivers and streams with effects on downstream water quality for extractive uses and riverine ecosystems.

Plantation forestry is widely seen as an option to address the problem of rising groundwater and associated spread of salinity, and provide a commercial return from timber production, in addition to the other agricultural productivity and environmental benefits (shade, shelter, wildlife corridors, etc) that arise from the integration of trees into agricultural production.

However, a number of studies have raised concerns over the sustainability of plantations on saline discharge sites, on the basis of possible salt accumulation in the rootzone, and eventual "salting out" of the trees. As this is a long term process, modelling provides a way to make predictions, and possibly direct corrective management actions.

Prior modelling studies have shown significant relative differences in tree growth rates under varying degrees of salinity and hydrological conditions. These studies have focussed on salt-tolerant, but relatively low-productivity species such as E. occidentalis and E. camaldulensis, or productive but salt sensitive species such as E. grandis and E. globulus. Conclusions from this work indicate the sensitivity of results to tree salt tolerance (Silberstein et al., 1999).

Trials of eucalypt hybrids bred for salt tolerance and commercially viable productivity under moderately saline conditions under a range of site conditions were established across Australia. A subset of four of these trials, spanning a range from non-saline to severely saline, was selected for detailed monitoring to provide data for calibration of the hydrological and growth productivity model, WAVES. The calibrated model was then used to explore a range of long term scenarios on plantation productivity and salt load and distribution in the soil profile.

Methods used
The WAVES model was applied to data collected for hybrids of E. camaldulensis, E. globulus and E. grandis at Mount Scobie in northern Victoria, and at Caldwell, Coleambally and Bethungra, in southern NSW. Modelling focussed on wood production of a selected sub-set of commercial clonal lines within these trials, firstly in relation to the observed average site conditions, and secondly under a theoretical range of possible site situations of groundwater depth, groundwater salinity, irrigation, irrigation salinity and plantation management regime, over long term (30 year) simulations.

It was possible to fit the growth of E. camaldulensis x E. globulus at each site, with varying levels of confidence, using a single set of vegetation parameters, observed soil type, water level and salinity, using nutritional status as a controlling factor. This was done to prevent each site having a completely unique set of parameters for the same hybrids and clonal lines, in order to allow extrapolation outside the range of site conditions included in the study. Fitting of E. camaldulensis x E. grandis followed using Mount Scobie as the reference, and all site specific conditions were held constant as the vegetation type was changed. The data from the trials are still too short-term to be confident of longterm predictions of productivity. However, the model gives a biophysical rationale for making forecasts, and gives some confidence that we can at least identify relative productivity and how this will develop in parallel with soil salt accumulation, and waterlogging.

Unfitted results such as leaf area index and salinity profiles were generally also adequately reproduced by the model, thereby providing more confidence in the model application.

Growth at Mt Scobie was able to be described by altering the nominal parameters in WAVES for a "normal" tree from –200m (20 bars) ?max and 1.0 salt sensitivity, to –300 m and 0.5 salt sensitivity, indicating a higher wilting point and lower effect of salt impact on water uptake by plants. Growth at Caldwell and Bethungra was less than projected by WAVES. In order to reproduce the observed growth at these sites while also constraining the vegetation parameters to be constant between sites, it was necessary to reduce the site "nutrition" parameter below the relative value reflected by actual differences in nutrition between the sites. This means that any uncharacterised site conditions that result in the poor growth at these two sites are reflected in one parameter, which may be compensating for other poorly described input to the model, or the model itself may have an inappropriate or incorrect representation of a process. It is parsimonious and acceptable from a modelling point of view, because we are reluctant to manipulate parameters just to gain a model fit, but it is unsatisfactory that we do not know exactly what it is at these sites that causes the poor growth. The use of a very low nutritional value at the dryland site Bethungra was of particular concern, as was the observed soil salinity profile at Caldwell, which, while high in absolute terms, was lower than predicted by the model given the extreme (approximately 40dS/m) watertable at around 2.5m depth.

These difficulties in modelling indicate that the level of site characterisation is currently inadequate to determine good definition of a growth–site condition matrix, and hence extrapolation from the measurements at these sites to sites with different conditions is really only qualitative. The modelling indicates the likely productivity at these sites and enables some basis for making these qualitative extrapolations. Clearly, measurements at new sites should be made to allow a determination of relative site qualities, and from these assessments the results presented here can be interpolated to give an indication of growth rates.

Results/Key findings
Modelling to calibrate actual growth performance against observed site conditions, as well as long term scenario modelling, provide support for sustainable plantation growth of salt tolerant tree species on sites over shallow saline watertables. More importantly, long term simulations indicate that the load and distribution of salt in the profile can be positively influenced by management interventions of thinning and irrigation.

Modelling of the distribution and load of salt in the soil, commencing from an assumed non-saline profile over a shallow saline watertable, indicates that salinity increases, but reaches a distribution and load similar to that observed in the pre-planting condition at the sites modelled. It then reaches a dynamic equilibrium, fluctuating seasonally, but maintaining a consistent lead and profile distribution.

Importantly, the projections for thinning from 350 spha to 150 spha at year 12 at Mt Scobie (a thinning level required to produce sawlogs) produced flushing of salt down the profile, and reduction of salt load in the upper profile. This occurred despite the simulated thinning corresponding to a dry year. The effect of this flushing was maintained in the simulations to years 17 and 18, when seasonally heavy rain produced an even greater flushing effect. Beyond this, salt gradually accumulated again in the upper profile by age 24 to 25, but was again flushed by a period of simulated heavy irrigation with saline water.

Implications for relevant stakeholders
The results of this work inform the potential for sustainable management of commercial plantations over shallow saline watertables, with soil salt loads and profile distribution reaching an equilibrium level that can be positively influenced by management interventions such as thinning. This result establishes an important role for salt tolerant trees as an environmentally sustainable land use option that can make commercial use of the existing and expanding areas of shallow saline watertables, and form an important component of salinity management moving forward.

Afforestation of saline discharge areas would reduce salt movement into drainage systems and limit impacts on downstream habitats and ecosystems. The potential for commercially viable plantation forestry on saline discharge areas and for management of saline drainage and wastewaters has enormous environmental, social and economic benefits. The establishment and management of such plantations on, say, 500,000ha of land at risk from rising watertables would require a capital investment of around $1.5 billion, but could potentially generate revenues of $7.6 billion. Timber production has a multiplier effect of around 2 for employment, value adding, income and output, providing a potential benefit to rural communities of over $760 million/yr. The capacity of commercially driven forestry to achieve such critical scale, coupled with savings to the public in expenditure on environmental remediation, and the potential for re-direction of limited public funds to other environmental projects, provides an otherwise unachievable benefit to environmental spending.

Recommendations
The findings of this report are positive for management of land with shallow saline watertables using salt tolerant timber trees. However, the projections have necessarily been limited to, at most, an observed growth period of seven years. Given model projections are a complex interaction of factors, principally soil hydraulic properties, groundwater depth and salinity, plantation species salt sensitivity, soil nutrient status, and plantation management, and given the recognised limitations of long term modelling, it will be important to continue monitoring the trial sites to enable verification of model projections. This, in turn, will also enable refinement of modelling tools.

In addition, while the work described in this report has provided a solid benchmark of site conditions over the early years of the rotation, the attempt to use this data for modelling has revealed a number of inadequacies. To address this, future site monitoring needs to better stratify sites to limit variability within modelled strata. Also, more work is required to characterise soil hydraulic properties.

While the model was able to describe observed growth at Mt Scobie, the ranking of E. grandis and E. globulus hybrids differed at other sites. This was most confusing at Bethungra where the commercial E. globulus hybrid cohort performed better on the high salinity stratum than the low salinity stratum, while the reverse was true for the E. grandis hybrids. WAVES was unable to adequately describe these variations in hybrid performance. While this is always likely, it indicates a potential area to improve functions in WAVES.

The work indicates the need for short term verification across other sites outside those studied, particularly across more dryland sites; verification of soil salinity profile and load changes following thinning; separation of confounding effects of a variable climate file against a background of management interventions; and improvement of the allometric functions of WAVES to accommodate changes in growth patterns in plantation stands over time. Also, the current work has necessarily utilised single tree plot and five tree line plot trials, and extrapolated performance of commercial subsets of these trials using site average soil conditions. The precision of modelling would potentially be improved using trials comprising only the target set of commercial genotypes. Finally, while WAVES output provides growth curves for a given set of conditions, interpretation of these for different combinations of salinity and groundwater would be facilitated by comparison to a benchmark growth curve for the soil type under non-limiting conditions. This would avoid confounding effects that limit the capacity to draw conclusions, such as uncertainty over whether the decline in growth frequently observed toward the later part of the 30-year simulations is due to the age constant programmed into WAVES or due to salinity, waterlogging, or other factors.
 
 

Site map Contact us Feedback Links Disclaimer Privacy Statement