Effectiveness of Current Farming Systems in the Control of Dryland Salinity

Leakage

Leakage for several case studies

1. Liverpool Plains
2. Wagga Wagga
3. Burkes Flat
4. Mallee

Managing salinity

What are we aiming to achieve?

Leakage for several case studies

1. Liverpool Plains
   
   The Liverpool Plains is a productive agricultural catchment near Tamworth in northern New South Wales. CSIRO and NSW Agriculture have undertaken an extensive study in this area to assess alternative land-use systems. Scientists used APSIM (Agricultural Production Systems Simulator), a computer modelling program developed by CSIRO, to simulate alternative farming systems, including their water balance and crop production. The research included a detailed field experimentation program over four years on a farm called 'Hudson' to check the model predictions. Following the field verification, APSIM was applied to simulate alternative farming systems over a range of soil types and rainfall zones within the Liverpool Plains catchment. Mean annual rainfall for the cropped regions of the catchment ranges from 625—740 mm/year. Eighteen different soil types used to support cropping were differentiated, characterised, and used in the analysis.
   
   Results from the study (Table 1) suggest that some cropping system options, when appropriately matched with soil type and location, were able to reduce leakage to a nominal 2—3 mm/year. However, all the cropping systems had significant rates of leakage if implemented on inappropriate soil types. Some soil types (usually shallow with low water holding capacity) and locations were unsuitable for cropping and needed to be sown to perennial grasses or trees.
   
   
   
   Table 1. Predicted mean annual leakage values over a 41—year period for alternative cropping systems in the Liverpool Plains. Figure 3. Year-to-year variation in predicted leakage under alternative cropping systems at the Hudson trial site in the Liverpool Plains.
   
   The long-fallowing cropping and continuous wheat cropping systems result in more leakage than would have been expected under the native perennial grasses that used to cover much of the area now cropped. Opportunity cropping and continuous sorghum make better use of the summer rainfall—they can reduce leakage to acceptable values providing they are on soil types appropriate for this type of cropping (Table 1, Figure 3).
   
   
   
   Figure 3. Year-to-year variation in predicted leakage under alternative cropping systems at the Hudson trial site in the Liverpool Plains.
   
   For some soils types, appropriate cropping and pasture systems in the Liverpool Plains climate appear capable of maintaining leakage rates approximating those of the native vegetation, although current land-use in the catchment is not achieving that aim.
   
   
2. Wagga Wagga
   
   CSIRO undertook a comparison of different cropping systems for the Wagga Wagga region in southern New South Wales, where the average annual rainfall ranges from < 480 to > 660 mm/year. In collaboration with the Australian National University, detailed measurements were made of wheat and lucerne grown in rotation at an experimental site at Charles Sturt University. The APSIM model was used to simulate each of these phases and the output compared against measurements including crop yield, soil water storage, and evapotranspiration measured using weighing lysimeters. Although still preliminary, agreement between the model and the measurements was good. Runoff is negligible at the measurement site, so that good prediction of evapotranspiration and change in soil water storage implies good prediction of leakage.
   
   Scientists then used APSIM to simulate the behaviour of both a continuous wheat cropping system and a wheat–lucerne rotation (three years of wheat, three years of lucerne) using the 36-year weather record from Forest Hill, just east of Wagga Wagga. There was a 53% reduction in leakage from an average of 87 mm/year for continuous wheat to 41 mm/year when lucerne was included in the rotation. Figure 4 compares leakage from the two cropping systems in each year and shows the variation of the reduction from year to year, depending on rainfall and the phase of the rotation. The reduction resulted partly from less leakage during the lucerne phase, and partly from the creation of a dry buffer, which reduces leakage during the first wheat years. Growing lucerne continuously reduced the average leakage to 4 mm/year, which is comparable to leakage from natural vegetation.
   
   
   
   Figure 4: Comparison of annual leakage from wheat and lucerne.
   
   The effectiveness of lucerne in reducing leakage depends on annual rainfall. For an average rainfall of 480 mm/year (Narrandera) the reduction was 66%, while for an average annual rainfall of 660 mm/year (Cootamundra) the reduction was only 48%. Only at the lower average rainfall did the wheat/lucerne rotation reduce leakage to a value (14 mm/year) approaching that from natural vegetation.
   
   The absolute values of leakage presented here are indicative only. Leakage depends not only on rainfall, but also on crop management factors (such as fertility and weed control) and soil type. While certain combinations of these factors could result in leakage values being up to half of the values shown in Figure 4, the relative difference between leakage under wheat and that under wheat/lucerne will remain unchanged.
   
   
3. Burkes Flat
   
   The Burkes Flat catchment covers 900 hectares. It is a subcatchment of the Avoca River in the 450 mm rainfall belt in the foothills of northern Victoria. The area has had dryland salinity problems for the past 70 years.
   
   The Victorian Department of Natural Resources and Environment (NRE) undertook an extensive project to map the areas of leakage to groundwater as well as areas of groundwater discharge. Between 1983 and 1986, native trees were planted on high leakage areas, perennial pastures were established on low—moderate leakage areas, and groundwater discharge areas were fenced off and planted with salt tolerant grasses.
   
   
   
   Groundwater levels have dropped significantly in comparison to the neighbouring untreated catchment. Watertables in the mid-catchment region at Burkes Flat fell more than 5.5 metres over the six years following the establishment of dryland lucerne in 1984.
   
   Since the start of the project, inspections have shown considerable improvement in the condition of the groundwater discharge areas. The Centre for Land Protection Research (part of NRE) in Bendigo estimates that the salinised area would have increased by up to 25% without the catchment treatment. In this light, it is promising that there is not yet clear evidence of changes in salt affected areas within the treated catchment.
   
   This evidence confirms the salinity control and productivity benefits of projects such as the one at Burkes Flat. This study indicates that well managed perennial pastures and trees on ridges in a local groundwater system of this type can provide some control of dryland salinity.
   
   
4. Mallee
   
   A modelling and field study was carried out as part of a joint project between NSW Agriculture, Agriculture Victoria and CSIRO to study leakage rates under fallow and non-fallow systems in the Mallee at both Walpeup (Victoria), and Hillston (New South Wales).
   
   Model results at Walpeup indicated that by removing the fallow and including mustard into a wheat-pea crop rotation, long-term average leakage was reduced by between 25% and 40%. The variation in the recharge figures is due largely to differences in rooting depth.
   
   
   
   The study concluded that removing the fallow period reduced the number of small rainfall events that led to leakage. It also found that none of the cropping systems were able to remove the effects of the larger rainfall events, although eliminating the fallow period could reduce the amount of leakage for large rainfall events. However, much of the Mallee overlies large regional groundwater systems for which this reduction in leakage is insufficient.

Back to top

Managing salinity

Dryland salinity generally occurs when a groundwater system cannot carry all the water put into it through leakage. It has various causes. In the Australian landscape there are often many physical restrictions to groundwater systems. They include geological formations that reduce the size of the aquifer and prevent sufficient groundwater from leaving the catchment. Also, as the groundwater moves from the upper hills of a catchment to its lower plains, a combination of flatter slopes and impermeable soils further restricts the capacity of the groundwater systems to carry water.

The amount of water that a groundwater system can carry is called the discharge capacity. If the total leakage from an entire catchment to the groundwater system is less than the discharge capacity, salinity should not occur because the groundwater system can cope with the supply. This concept can provide a useful method for estimating a leakage target—the allowable leakage for a catchment to avoid dryland salinity. This can be done by estimating the discharge capacity and then using the result to estimate the average leakage over an entire catchment. Previous CSIRO studies have used this method for groundwater systems in the Murray-Darling Basin and estimated the maximum amount of leakage for which there would be no dryland salinity. For example, the leakage target for the Upper South-East region (South Australia) was only 2 mm/year (0.4% of rainfall), while for the Liverpool Plains catchment (New South Wales) it was 1 mm/year (0.1% of rainfall). These values are comparable to those found under native vegetation.



It is clear from the results of the case studies that generally, the leakage rates under our current agricultural systems far exceed these targets. The Burkes Flat example is an exception.

As we do not know the precise discharge capacities of our groundwater systems, a conservative approach to controlling dryland salinity is to aim for leakage values comparable to those under native vegetation.

In any case, changes are not likely to produce quick results. The excess water that has been leaking into groundwater systems combined with the time scale of groundwater processes means that it is unlikely that the effects of an instantaneous reduction of leakage will be discernible immediately. Even if we reduce leakage to below the discharge capacity, it still will take some time to influence the current salinisation rate.

For example, the results of groundwater modelling scenarios for catchments in the Liverpool Plains suggest that even if the leakage to groundwater were reduced to zero, the levels of groundwater in the lower parts of these catchments would begin to drop only after 20—60 years. The slow response means that when we reduce recharge rates, even if they are greater than the discharge capacity, we would still expect dryland salinisation to persist and in many cases, to expand.

The discharge capacity and time responses will vary according to the type of groundwater system and the individual aquifer characteristics. The large groundwater systems that characterise the Riverine Plains and Mallee regions generally have low discharge capacity (<5 mm/year) and response times ranging from hundreds to thousands of years. The local groundwater systems that predominate in many of the upland areas may have higher discharge capacities (<100 mm/year) and quicker response times (5—50 years).

Using biological control of leakage is one of three broad options for controlling dryland salinity. The other two options include engineering options (such as groundwater pumping or surface drainage), and adapting to the more saline conditions. A combination of all three of these types of options is likely to be needed.

Back to top

What are we aiming to achieve?

Option A — continue to expand. This is a 'do nothing' scenario, where salinity continues to increase in area and magnitude, until it reaches a new 'more saline' equilibrium.

Option B — buying time. This is where the management strategies we implement slow down the onset and expansion of dryland salinity, but eventually reach the same salinity equilibrium as Option A, the 'do nothing' approach.

Option C — continue to expand, but not as much. This is where our management strategies make sufficient change to reach a stable salinity equilibrium, with fewer saline areas than for Option A, the 'do nothing' approach.

Option D — improve the situation. This option is to reverse the trend of expanding salinity and recover saline land.

These scenarios can be considered at scales ranging from local groundwater systems to the whole of the Murray-Darling Basin. When considering individual groundwater systems, Option B would correspond to a situation in which we have reduced recharge, but it is still much greater than the discharge capacity. This would be equivalent to half-turning off a tap when filling a bath. The bath still fills, but takes longer to do so. Option C corresponds to the situation where recharge is reduced enough to decrease the discharge significantly. Option D corresponds to the situation where the system begins to drain.

For example, suppose we are growing wheat over a regional groundwater system with a discharge capacity equivalent to 1 mm/year. If the deep drainage is 80 mm/year and we reduce this to 40 mm/year by incorporating lucerne in rotation, the groundwater system will still fill up, but it will take twice as long. This is equivalent to Option B.

On the other hand, if we replace annual pasture with a combination of perennial pastures and trees over a local groundwater system with a discharge capacity equivalent to 60 mm/year, we may achieve a situation similar to Option D. When we aggregate the effects over the whole Murray-Darling Basin, the most likely outcome is Option C, but how much of an improvement this is in comparison to Options A or B needs to be determined.

Back to top

Back to Contents