Judith Bean, NSW Department of Primary Industries, Trangie Agricultural Research Centre, PMB 19, Trangie NSW 2823, PO Box 578, Gunnedah NSW 2380. Email: lund@hwy.com.au
Gavin Melville, NSW Department of Primary Industries, Trangie Agricultural Research Centre
Stephen P. Clipperton, NSW Department of Primary Industries, Trangie Agricultural Research Centre and Mineral Resources, NSW Trade and Investment, Locked Bag 21, Orange, NSW 2800.
The study was carried out in moderately degraded rangelands (shrub invasion was not advanced) in the Cobar district of north-west New South Wales, on sites characterised by reduced numbers of perennial grasses leading to low productivity and damaging erosion. The aim of the study was to determine whether seed of pastorally desirable perennial grasses, already existing in the landscape, was adequate to support regeneration of perennial grasses, and whether regeneration could be enhanced by low-cost minimum-disturbance changes to the landscape.
The strategy was to harness natural processes already operating in the environment (mainly wind and running water) to disperse the available seed over large areas and to implement low-cost strategies to promote germination and survival of seedlings. Study sites were established in two rangeland types characterised by differing soils and vegetation communities. Both study sites included small knolls or ridges with plants of pastorally desirable perennial grasses, surrounded by more extensive slopes virtually bare of such species. In August 1999, at the time of site selection, individual grass butts had been overgrazed and were only visible on close examination of the sites.
Site 1 was characterised by long very gentle slopes of hard-setting red earths with abundant lateritic pebble and a crust of cryptogam, running out from slightly higher areas of non-ferruginous rocky outcrop (Photo 1). The dominant tree species was mulga (Acacia aneura) with scattered bimble box (Eucalyptus populnea). The slopes were characterised by very sparse plants of No. 9 wiregrass (Aristida jerichoensis ) and variable speargrass (Austrostipa variabilis). When selected, this site was very overgrazed and so the sheep were removed from the paddock throughout the study.
Photo 1. Small knoll of non-ferruginous rocky outcrop at Site 1 with a significant population of mulga oats and some mulga Mitchell grass.
Site 2 was characterised by steeper slopes of a medium-textured soil with abundant rock fragments >2 mm in diameter, running off a prominent ridge with abundant in-situ rock and gravel. The dominant tree species on the ridge crest and shoulders was green mallee (Eucalyptus viridis), with red box (Eucalyptus intertexta), wilga (Geijera parviflora) and white cypress pine (Callitris glaucophylla) scattered on the middle slopes; variable speargrass (A. variabilis) was also prominent on the slopes. The site was grazed by cattle prior to and throughout the study at ~0.02 large stock units ha-1.
Photo 2. Slope with numerous plants of variable speargrass, forbs, red box and large wilga trees, leading in the distance to the fenced rocky ridge with green mallee at Site 2 (November 1999 at commencement of the study).
At both sites the tops of the rocky knolls or ridges were fenced using steel pipe as corner posts and stays, star pickets at intervals between the corner posts, hinge-joint netting (90 cm in height with seven horizontal wires plus vertical wires 30 cm apart) and a single barbed wire at the top. Large branched seed heads were able to pass through the netting. The fence was built to prevent grazing by native and domestic large herbivores, with the aim of promoting seed production and protecting plants of desirable perennial grasses from overgrazing and subsequent death. An opening was provided so that if, in the future, it became necessary to graze rank growth within the fenced area, stock could be given short-term access. These fenced areas were termed seed production areas (SPAs). The detail of their shape varied according to the shape of the knoll or ridge and on average measured 50 m in width and from 60 to 200 m in length.
Outside the fences on the slopes leading away from the SPAs, two regeneration techniques and an untreated area were implemented:
- Technique 1: Pits cut into the soil surface at regular intervals by a crocodile (a turning metal drum with projecting metal spades) pulled over the site with a 4WD vehicle; the pits, immediately after implementation, had a mean length of 24.5 cm and mean depth of 3.4 cm (Photo 3). At Site 2, with its medium-textured soil, the pits were less sharply cut and largely obliterated by heavy rainfall within 9 months of implementation.
- Technique two: Freshly cut branches of turpentinelaid on the soil surface running parallel to the topographic contours; individual branches, within the lopped section, averaged 1-2 cm in diameter and were pliable, such that the lower branches settled to be continuously in contact with the soil surface to form parallel rows of obstruction to any material being washed down-slope by running water. The network of fine branches was approximately 20 cm high (Photo 4).
Photo 3. Seed production area on right of fence and pits implemented by a crocodile on the left at Site 1 (October 1999).
Photo 4. Piles of branches running parallel to the contours of the slope and located in pairs at set distances from the fence of a seed production area located in the distance at Site 1 (October 1999).
Fixed recording areas, 0.5 m2 in size (quadrats), were established on the slopes outside the fences with two areas at each set distance from the fence of the SPA, measured in the direction of water flow up to 100 m from the fence. At approximately six monthly intervals, individual plants of the dominant grasses were recorded according to species and category – new, continuing or dead.
Seasonal growth conditions for the area were excellent from November 1999 to August 2000 (90–100 percentile range of the previous 40 years of rainfall records) and from August 2000 to February 2001(80–100 percentile range) and poor from March to July 2001 (20–40 percentile range) and from August 2001 to March 2002 (20–30 percentile range).
Germinations and survival varied according to species, site, regeneration technique and seasonal growth conditions.
Mulga oats (Monachather paradoxus)
Mulga oats was the dominant species in the SPAs at Site 1. Germinations outside the SPAs were shown to have come from seed sourced both from within the adjacent SPA and from plants existing within the same 0.5 m2 recording area. Seed from a SPA was demonstrated to have been dispersed by water and/or wind for at least 99 m from the fence of any one SPA (Bean et al. 2015).
During excellent seasonal conditions, the numbers of new plants of this species greatly increased under the piles of fine branches, with up to 70 new plants within the one 0.5 m2 recording area. These new plants were seen to exist in lines of fine silt which had accumulated on the up-slope side of each fine branch in contact with the soil surface. Even during poor seasonal conditions there continued to be germinations of this species under the piles of branches but in smaller numbers. Table 1 gives the new and surviving plants in one 0.5 m2 recording area, as an example of the positive response of this species to the piles of fine branches.
The pits did not have the same positive effect on germinations of mulga oats; for example in August 2000 only 175 new plants of mulga oats were recorded across all the recording areas, compared to 1223 under the piles of branches. At the same time only 43 new plants were recorded in the untreated areas.
Photo 5. Two piles of turpentine branches, one at 0 m from the fence of a SPA and the other at 5.5 m down-slope from the fence at Site 1. The 0.5 m2 area at 5.5 m (in foreground), in August 2000, contained 50 new plants of mulga oats, 2 of mulga Mitchell, 1 of no. 9 wiregrass and 14 of variable speargrass (detail in Table 1).
Time |
Mopa |
Thmi |
Arje |
Auva |
Oct 1999 |
No plants |
No plants |
No plants |
1 plant |
Aug 2000 |
50 N = 50 |
2 N =2 |
1 N = 1 |
13 N = 14 |
Mar 2001 |
4 D 4 N = 50 |
= 2 |
9 N = 10 |
13 D = 1 |
Aug 2001 |
2 D = 48 |
= 2 |
6 D = 4 |
6 N = 7 |
Mar 2002 |
2 D 1 N = 47 |
= 2 |
1 N = 5 |
4 D = 3 |
Table 1 Number and history of plants of the four dominant species at the five recording times in the 0.5 m2 recording area at 5.5 m as depicted in photo 5. Mopa=Mulga oats; Thmi=mulga Mitchell grass; Arje=No.9 wiregrass; Auva=variable speargrass; Time=time of recording; N=new; D=died.
Survival rates of the new plants of mulga oats were also higher under the piles of branches than in the pits; under the branches, rates ranged from 76% in excellent seasonal conditions to 57% in poor seasonal conditions, compared to the mechanical pits where rates ranged from 67% to 41%, respectively.
The effectiveness of the network of fine branches in promoting germination and survival of new plants can, at least in part, be attributed to the microenvironment of the 20 cm high pile which retained moisture at a relatively uniform level, even in the middle of summer. Dew was seen on the fine branches even in mid-morning in the month of February. During the excellent seasonal conditions of November 1999 to August 2000, the initially red-brown crust of cryptogam on the soil surface under the piles of branches, changed to a layer of green moss; this moss remained even through the poor seasonal conditions from March 2001 to March 2002. The network of fine branches also provided protection from grazing.
At Site 1, study of seeds present in the soil seed bank showed that the ratio of number of seeds of mulga oats to number of seeds of No. 9 wiregrass incorporated into the soil was higher under the piles of branches compared to in the pits. Thus laying of branches would over time lead to a relative increase of seeds in the soil seed bank of mulga oats (Bean et al. 2016).
At Site 2 mulga oats was the second most important species in the SPA. The number of new plants outside the SPA did increase under the piles of branches during the excellent seasonal conditions from November 1999 to August 2000 relative to the pits and untreated areas but not to a level that would bring significant regeneration. There was evidence that seeds for these germinations were, at least in part, sourced from the SPA (Bean et al. 2015). On this site there was no evidence that rainwater ran down-slope carrying silt and seed to be deposited on the up-slope side of the branches in contact with the soil surface; it was concluded that on the medium textured soils of this site much of the rain infiltrated into the soil at the point of impact and so this landscape did not function as a runoff run-on system at the scale that Site 1 did.
It is concluded that for significant regeneration of mulga oats to occur available seed needs to be captured by some structure that obstructs running water and/or slows wind speed, then buried in a suitable seed bed such as silt accumulated on the up-slope side of an obstruction. For the seeds to germinate in large numbers and survive there needs to be a microenvironment which maintains moisture, even under severe summer conditions. These conditions were only met at a sufficient level under the piles of branches at Site 1.
Mulga Mitchell grass (T. mitchelliana)
At Site 2 at the commencement of the trials, mulga Mitchell grass was the dominant species on the shoulders of the ridge and slightly more numerous than mulga oats within the fenced SPA. During the excellent seasonal conditions from November 1999 to August 2000 many plants of this species, which had existed initially as overgrazed butts, produced long stems running across the soil surface. During this period 7% of new plants outside the SPA originated from vegetative reproduction at points where these long stems had been pinned to the soil by the piles of branches. From August 2000 to February 2001 the number of new plants produced vegetatively increased to 81% of the total number of new plants; in this period the majority of the new plants produced vegetatively formed where the hooves of cattle had pushed a stem into wet soil after rainfall. During the period of poor seasonal conditions from March 2001 to March 2002, the 57% of new plants formed by vegetative reproduction were seen to be located where stems had been pinned to the soil surface by the piles of branches (Bean et al. 2015).
At Site 2 new plants of mulga Mitchell grass originating from seed during the excellent seasonal conditions did not show a significant preference for either the branches or the areas disturbed by the crocodile. In contrast during the poor seasonal conditions from March to August 2001, there were more new plants in the pits than under the branches; it appeared that seed, buried in the pits when these were filled with soil during the heavy rains of the first period, reached the end of their period of dormancy and germinated.
At Site 1 where plants of this species were very few in the SPAs, new plants produced from seed were small in number, even during excellent seasonal conditions; there were equal numbers of new plants in the pits and under the piles of branches. However the piles of branches were more effective in capturing seed from the SPAs than the pits – new plants under the branches were found up to 24 m from the fence while new plants in the pits were limited to 11 m from the fence. At this site there were no new plants produced by vegetative processes (Bean et al. 2015).
Silky umbrella grass (Digitaria ammophila)
At Site 1, no plants of this species were found at commencement of the study, either within the SPAs or on the slopes around the SPAs. But during the excellent seasonal growth conditions from November 1999 to August 2000, large branches seed heads of this species were transported across the site, presumably by wind, from more distant locations. The piles of fine branches were very effective in capturing these seed heads which were observed imbedded in the network of branches. New plants appeared directly below the seed heads indicating that seed had been released onto the soil surface and germinated in the relatively moist microenvironment. Neither the pits nor the untreated areas were able to capture these seed heads and so no new plants germinated in either of these areas (Bean et al. 2015).
Curly windmill grass (Enteropogon acicularis)
At Site 2 at commencement of the study, small numbers of plants of curly windmill grass existed both within the SPA and on the slopes outside the fence. The large branched seed heads of this species were transported, presumably by wind, through the netting of the fence and across the slopes. New plants were small in number but most numerous under the piles of branches; as with silky umbrella grass new plants were located directly below seed heads which had been caught in the network of branches (Bean et al. 2015).
At Site 1 only four plants of this species were found; all germinated under the piles of branches during poor seasonal conditions from March 2001 to August 2000.
No. 9 wiregrass (Aristida jerichoensis)
Numbers of plants of No. 9 wiregrass in the SPAs were small compared to numbers of the dominant pastorally desirable mulga oats and mulga Mitchell grass. At Site 1, during the excellent seasonal conditions from November 1999 to August 2000, seed production by this species was excessive, both from clumps located at irregular points across the slopes and from isolated plants. Germinations during this same period were very high under the piles of branches with up to 183 new plants within one individual 0.5 m2 recording area. During subsequent periods numbers of new plants were much lower with the highest numbers either in the untreated areas or under the branches (Bean et al. 2015).
Study of seed in the soil seedbank showed that the ratio of seeds of No. 9 wiregrass compared to seeds of mulga oats was higher in the pits than under the piles of branches. Thus if pitting was implemented as a regeneration technique, with time, there would be increasingly greater numbers of seed of this pastorally undesirable grass in the soil seed bank; but the high numbers of seed did not translate into high germinations (Bean et al. 2016).
Variable speargrass (A. variabilis)
At both sites plants of this pastorally-undesired species were few within the fenced SPAs. In contrast, at both sites at commencement of the study it was the dominant species on the slopes. It is probable that the high density of plants of this species on the slopes at Site 2 (Photo 2) inhibited the dispersal of seed of pastorally-desirable species from the SPA to potential germination sites. During the poor seasonal conditions from March 2001 to March 2002 large numbers of plants of this short-lived perennial died (Bean et al. 2015).
Additional trial to assess criteria for effective SPA
On mulga country close to Site 1 described above, seven variations in the landscape, visible at the scale of 20 m x 20 m, were assessed to identify criteria necessary for successful function of a SPA. Within each of these seven variations, two areas 10 m x 10 m were fenced. At intervals over a 29 month period a large number of surface types were identified down to a scale of 2 mm and plants recorded according to the surface type on which they germinated. Survival was also recorded according to surface type.
In the period from December 1999 to mid 2001, during excellent seasonal conditions deteriorating to poor, numbers of new plants of mulga oats were highest in the 1-2 cm wide area adjacent to insitu rock outcrop – this surface type was restricted to rocky knolls similar to those fenced as SPAs at Site 1. During this same period, germinations of mulga Mitchell grass were very low, but the highest number were similarly located in the 1-2 cm catchment area at the margins of insitu rock on the rocky knolls (Bean et al. 2017).
In contrast to all other species, during the poor seasonal conditions from mid 2001 to April 2002, new plants of mulga Mitchell grass were numerous – again located in the 1-2 cm catchment at the margins of insitu rock on the rocky knolls. It is concluded that large numbers of seed of this desirable perennial grass were produced in the excellent to poor seasonal conditions, were then held in a dormant state in the cracks between the embedded rocks and the soil for about one year. They then accessed moisture held at depth at the rock-soil interface and germinated in large numbers. No other surface type was able to retain enough moisture to germinate seeds of this species which remain dormant for significant periods of time (Bean et al. 2017).
Optimum proportion of landscape allocated to piles of branches at the paddock scale
Based on observations at Site 1, if the 0.5 m2 piles of branches (dimensions 71 cm x 71 cm) were less than 5.5 m apart in a down-slope direction, no new plants appeared under the down-slope pile. It was concluded that the ratio of area covered by the branches and the intervening area with no branches is critical to the effectiveness of this regeneration technique. Also the surface of the area with no branches needs to experience minimum disturbance such that it remains relatively impervious allowing water to run off in a down-slope direction, transporting silt and seed with it.
Based on the 0.71 cm and 5.5 m noted above, it is calculated that on country similar to Site 1 the bands of branches to be effective need to constitute approximately 12-15% of the total area; for example, if the bands of branches each measured one metre in the down-slope direction (= run-on area), they would need to be separated by bands without branches measuring at least 6.4 m (= runoff area) in the down-slope direction. A higher proportion of branches would lead to insufficient nutrients being concentrated on the run-on areas and thus a lower potential for new plants. The more porous the surface in the runoff areas, the smaller the percentage of the landscape that should be allocated to branches. Every landholder would need to experiment on their own land to determine the optimum ratio.
Conclusions
In planning to increase the number of pastorally desirable perennial grass plants in a designated area, the following steps are recommended:
- Locate a number of small (<1 ha in area) knolls or ridges within the landscape characterised by rocky outcrop (rock is embedded in the soil in original position), presence of plants of at least one of the desirable perennial grasses and no or few plants of undesirable grasses.
- Fence this area to exclude domestic and native large herbivores, but allowing passage of large branched seed heads of desirable grasses. Provide an opening in the fence so that if, in the future, rank growth develops, grazing stock could be given short-term access.
- Identify the dominant desirable species within the fenced areas and the nature of the soil on the slopes leading away from the fenced areas:
- Slopes virtually bare of trees and large plants are ideal as they provide freedom for dispersal of seed over large distances.
- Soils which are relatively impervious to water, such that much of the rainwater runs off in the down-slope direction, are ideal for the success of branches placed along the contour lines (fine branches of freshly-lopped turpentine shrubs were very effective). Note that ‘lopping’ should not include removal or killing of the shrubs.
- In contrast, if the soil is medium textured such that much of the rainwater infiltrates at the point of impact, placing of branches on the surface will be less effective in capturing seed of grasses such as mulga Mitchell and mulga oats.
- If the desirable grass species produce large branched seed heads, such as Curly windmill grass or Silky umbrella grass, regardless of soil characteristics, placing of branches will be effective in forming a network to capture these seed heads and enhance germination and survival of new plants.
- Ideally the erection of fences around the rocky knolls or small ridges and the laying of branches in bands along the topographic contours can be carried out during a dry period while waiting for rain.
- The first rain will produce seed on the already-existing plants in the fenced area. Follow-up rain, if sufficient to produce runoff, together with wind, will distribute these seeds to areas outside the fences. The piles of branches will be effective in capturing individual seeds of mulga oats and large seed heads of curly windmill grass and silky umbrella grass. Then the shaded environment within the bands of branches will retain moisture and enhance establishment and survival of new plants.
To increase plants of mulga Mitchell grass further good rain is required after the breaking of a dormancy period of approximately a year. If existing plants of mulga Mitchell grass are characterised by long stems lying on the soil surface, placing cattle in the paddock at normal stocking rates before or immediately after rain will be effective in multiplying the number of plants – hooves of the cattle will push the stems into the wet soil and roots will develop at each of these points.
Acknowledgements
Input by landholders and many Departmental staff, in both implementation and assessment of the trials, is gratefully acknowledged.
References
Bean, J.M., Melville, G.J., Hacker, R.B. and Clipperton, S.P. (2015). Seed availability, landscape suitability and the regeneration of perennial grasses in moderately degraded rangelands in semiarid Australia. The Rangeland Journal 37, 249-259.
Bean, J.M., Melville, G.J., Hacker, R.B, Anderson, S., Whittington, A. and Clipperton, S.P. (2016). Effects of fenced seed production areas and restoration treatments on the size and composition of the native grass seedbanks in moderately degraded rangelands in semiarid Australia. The Rangeland Journal 38, 47 –56.
Bean, J.M., Melville, G.J. and Hacker, R.B (2017). Assessment of the potential of a range of microhabitats for use as seed production areas in moderately degraded rangelands in semiarid A