Source DocumentPrevious PageTable Of ContentsNext Page

Climatic conditions associated with successful perennial grass recruitment events in south-eastern Australia

Roshan Thapa and David Kemp

Charles Sturt University, School of Agricultural and Wine Sciences (SAWS), Leeds Parade, Orange NSW 2800


Recruitment of new perennial grass plants within existing ecosystems is determined by seed availability, suitable microsites, resources and climatic conditions. This paper reports on the modelled soil moisture conditions associated with successful recruitment events in five field experiments at Orange (Phalaris aquatica), Trunkey Creek (Austrodanthonia spp.) and Wellington (Bothriochloa macra) in Central New South Wales (NSW), and the frequency of those conditions during the past 30 years. High seedling numbers were recorded when mature germinable seed was present and when a rainfall event (median 68 mm across the three sites) kept the surface soil moisture (0-50 mm) above the permanent wilting point for at least 15 days. Rainfall events typically occurred in the second half of February, sometimes extending to early March. No recruitment occurred if the surface soil was only moist for 7 days or less. The past data (1975 - 2004) showed that Orange had a median of 20 moist days each year at this time, whereas it was 16 days for Trunkey Creek and 10 days for Wellington. The probabilities of exceeding 7 or 15 days of moist surface soil were 98% and 78% at Orange, 91% and 49% at Trunkey Creek, and 73% and 30% at Wellington. These results show some recruitment is possible in most years, provided seasonal conditions enable adequate seed set over summer.

Key Words

soil moisture, modelling, perennial grasses, late summer, drought, irrigation


Recruitment from seed which is a natural event that occurs within ecosystems is determined by the availability of seed, suitable microsites, other resources and climatic conditions. Seed availability and soil disturbance, either occurring naturally by self-mulching processes in soil or through management interventions such as soil scarifying to create suitable microsites are important to maximise recruitment (Thapa 2010). The key climatic conditions required include suitable temperature and adequate soil moisture. Soil moisture available to the seed is one of the main factors that enables germination and establishment of perennial grass seedlings (Dowling et al. 1971; Fowler 1986; Maze et al. 1993; Lauenroth et al. 1994; O’Connor 1996; Hamilton et al. 1999). Successful recruitment was observed between 2006-2008 at all the field sites at Orange (14907' E, 3314' S), Trunkey Creek (14919' E, 3349' S) and Wellington (14858' E, 3230' S) in central New South Wales (NSW), Australia. Using the data obtained from these three field sites (Thapa 2010), combined with a small plot study using irrigation and seed addition through the year, this paper aims to first identify the soil moisture conditions associated with successful initial recruitment events and then to predict the frequency of attaining those required conditions using climatic data for the 30 years from 1975-2004. Emphasis is placed on the initial seedling recruitment conditions because drought years resulted in a low frequency of survival of young plants. The resulting data proved to be insufficient to extend the climatic analysis to the conditions required to support plant survival through the subsequent summer.


Field experiment data sets used

Successful recruitment events were observed in each of 5 experiments: 2 each at Orange and Trunkey Creek, and a single event in one experiment at Wellington despite prevailing drought conditions (Thapa 2010). The seedling recruitment events observed for Phalaris aquatica L. (at Orange), Austrodanthonia spp. H. P. Linder (at Trunkey Creek) and Bothriochloa macra (Steud.) S.T.Blake (at Wellington) occurred in March (early autumn) after significant rainfall events in the February-March period. Rainstorms in late February at all the sites were the first major event after seed maturation. Seedlings observed in March resulted from these February rainfall events. These February rainfall events and the resulting estimates of moisture in the top 50 mm of the soil surface layer were analysed to determine the soil moisture conditions when the 5 recruitment events occurred. A soil depth of 50 mm was chosen as root growth would not extend much beyond that within the first few days after germination.

Irrigation experiment

The irrigation experiment was done at the same field sites in Orange, Trunkey Creek and Wellington. The experiment started in January 2007 and continued until February 2008 except at Wellington where the experiment was cut short in October 2007 due to the unavailability of the site. Water was applied at approximately 6-week intervals at a rate equivalent to 50 mm rainfall (50 L m-2) over 2 days at 25 mm on each day. Separate plots were used on each occasion. Plots were covered with shade cloth for 2-3 days following irrigation to reduce evaporation. Initially no perennial grass seed was added at each site to test recruitment from the seed bank. However, due to a lack of seedling emergence and accumulated evidence that the perennial grass soil seed bank was presumed to be exhausted (Thapa 2010), the design was modified to add the equivalent of 50 kg seed ha-1 to the watered plots from November 2007 and continued until the end of the experiment in February 2008. Of the total of 21 watering events across the three field sites, 13 coincided with rainfall events.

Climate data

Climate data were measured using data loggers (Tain Electronics) at the site for the duration of the experiment. Soil temperatures and moisture (using gypsum blocks) in the soil surface layer (0-50 mm) were recorded as well as relative humidity, air temperature, solar radiation, wind speed, and rainfall. At Wellington, the site for the B. macra experiment, rainfall data were obtained from the weather station located <0.5 km from the experimental site. The long-term climate data were obtained from the National Climate Centre of the Australian Government Bureau of Meteorology (NCC 2009).

Soil moisture model

The percentage volume of moisture in the top 50 mm soil surface for the experiment duration and for the 30 years from 1975-2004 were estimated using the Sustainable Grazing Systems (SGS) pasture model (Johnson et al. 2003), version 4.5.4 (Johnson 2008). The analysis focused on the soil moisture status in the top 50 mm created by the rainfall events before the time of the identified recruitment events. At the start of each model run the soil moisture levels were considered to be zero as an extended dry period was chosen as the starting period.


February - March rainfall

Five successful recruitment events were observed across the three field sites and each occurred after rainfall events in February that in some instances extended into early March. Soil moisture in the top 50 mm was estimated during and after those rainfall events using the SGS model. After initial modelling of soil moisture status and reviewing the associated recruitment events, it was decided to identify the periods where the soil moisture content was between ~40% (i.e. close to field capacity in these soils and where free water was available for a seedling to emerge) and ~20% (i.e. close to the permanent wilting point where there would be little available moisture for a seedling).

Recruitment events in the main field experiments

The five recruitment events recorded across the five experiments occurred at the same time of the year, soon after seed set and with significant moist periods for several days (Figure 1). Differences between species (one C4 and two C3 grasses) were neither apparent in the timing of the recruitment event, nor in the general climatic conditions under which it occurred. On average across the sites the rainfall events were 73 mm falling over 7 days or more which kept the soil moist for at least 14 days. This provided the criteria for ideal number of days (i.e. 14 days) that the soil needed to be moist for a successful recruitment to occur. This rainfall event occurred in the second half of February, sometimes extending to early March, when mature, germinable seed was present. If the amount of rainfall is received outside the late February to early March period, the probability of successful recruitment arguably becomes lower, as no recruitment was recorded at other times in this series of experiments. These conditions set the boundaries to determine the frequency of achieving a successful recruitment event.

Recruitment events in the irrigation experiments

The six-weekly irrigation treatments used across the three sites did not result in any significant seedling recruitment on most occasions even though viable grass seeds were detected in the soil cores. There were 21 watering events across the three field sites, of which 8 events recorded recruitment - 3 at Orange, 5 at Trunkey Creek and 0 at Wellington. Of the 8 successful recruitment events, 6 had both water and seed applied and coincided with the rainfall events. Across the sites there were 6 irrigation events that did not coincide with the rainfall events and had no seed added. As these 6 irrigations did not result in recruitment, the resulting soil moisture conditions in these events were considered inadequate for a successful recruitment event. On average, the model estimated that soil in the top 50 mm was moist for 7 days in these 6 irrigation events which was not enough for a recruitment event to occur. Therefore, this period of moist surface soils (i.e. 7 days) provided the minimum criteria below which recruitment was highly unlikely. The general lack of recruitment in irrigated treatments reinforces the view that native grass recruitment depends predominately on seed availability, microsites and adequate soil moisture conditions at critical times.

Figure 1. Soil water content (% volume of soil) for the top 0-50 mm without irrigation (solid line) and with irrigation (dashed line) generated from the SGS pasture model from December to April at (a) Phalaris aquatica site in Orange, 2005-6, (b) Phalaris aquatica site in Orange, 2006-7, (c) Phalaris aquatica site in Orange, 2007-8, (d) Austrodanthonia spp. site in Trunkey Creek, 2006-7, (e) Austrodanthonia spp. site in Trunkey Creek, 2007-8, and (f) Bothriochloa macra site in Wellington, 2006-7; (triangle = time of identified recruitment event in the main field experiments; circle = the time of irrigation treatment (50 mm over 2 days); square = seed addition in irrigation treatments).

Past climate data and frequency of recruitment

Across the three sites a successful recruitment event in the main field experiments was achieved when adequate soil moisture in the top 50 mm was maintained for around 14 days. Based on these experimental results 15 days of adequate moisture was set as the safe criterion that should result in a high possibility of recruitment occurring. The irrigation experiments helped identify 7 days as the minimum numbers of days the soil moisture was inadequate and where there was a low probability of recruitment occurring. Midway between the minimum and ideal conditions are arguably the events that will produce noticeable improvements in the recruitment of these desirable perennial pasture grasses. Analyses for the past 30 years showed that Orange had an average of 20 moist days each year following a satisfactory rainfall event in the February-March period, whereas it was only 16 days for Trunkey Creek and 10 days for Wellington (Figure 2-a). On average some recruitment is then possible. The shortest moist period was 2 days at both Trunkey Creek and Wellington, and 9 days at Orange. The longest moist period was 31 days at Orange and 24 days at Trunkey Creek and Wellington. Further analysis (Figure 2-b) showed Orange should get some recruitment every year as the probability of exceeding 7 days of moist soil was 98% and 78% of years exceeded 15 days. Trunkey Creek would only fail in 9% of years to achieve the minimum moisture conditions of 7 days and 41% of years should have a high chance of recruitment. The worst site was at Wellington where there would be little or no recruitment in 27% of years (<7 days moist soils), and that in 73% of years moist soil would last at least 7 days and so some recruitment might occur. This indicates that useful recruitment could occur in most years.

Figure 2. (a) Number of days when the soil was moist in the top 50 mm after a significant rainfall event in the February-March period shown as box plots, and (b) their cumulative probability distribution (quadratic curves fitted to the data to show trends) over the last 30 years across the three field sites at Orange (Phalaris aquatica), Trunkey Creek (Austrodanthonia spp.) and Wellington (Bothriochloa macra); soil moisture (0-50 mm) data generated from SGS pasture model.


Rainfall events in late summer were useful for successful recruitment. This was in contrast to the ‘autumn break’ which normally occurs later, when these analyses show that recruitment does not happen often in central NSW. Over the last 30 years at these study sites reliable rainfall events started in February. No successful recruitment events were observed at other times of the year suggesting that the rainfall events that occurred in February through March are more reliable for recruitment as at other times of the year recruitment is limited by other factors (e.g. low availability of seed) even though reliable rainfall events occur. There is a reasonable probability of adequate rainfall events, usually starting in February each year that would result in a successful recruitment event. Any average or above average years should result in recruitment in paddocks appropriately rested to maximise flowering and seed set and on appropriate soils where a light scarifying was done to increase the microsites for recruitment.


Dowling PM, Clements RJ and McWillliam JR (1971). Establishment and survival of pasture species from seeds sown on the soil surface. Australian Journal of Agricultural Research 22, 61-74.

Fowler NL (1986). Microsite requirements for germination and establishment of three grass species. American Midland Naturalist 115, 131-145.

Hamilton JG, Holzapfel C and Mahall BE (1999). Coexistence and interference between a native perennial grass and non-native annual grasses in California. Oecologia 121, 518-526.

Johnson IR (2008). SGS Pasture Model. Armidale, IMJ Consultants Pty Ltd.

Johnson IR, Lodge GM and White RE (2003). The Sustainable Grazing Systems Pasture Model: description, philosophy and application to the SGS National Experiment. Australian Journal of Experimental Agriculture 43, 711-728.

Lauenroth WK, Sala OE, Coffin DP and Kirchner TB (1994). The importance of soil-water in the recruitment of Bouteloua gracilis in the shortgrass steppe. Ecological Applications 4, 741-749.

Maze KM, Koen TB and Watt LA (1993). Factors influencing the germination of six perennial grasses of central New South Wales. Australian Journal of Botany 41, 79-90.

NCC (2009). Climate data. National Climate Centre, Australian Government Bureau of Meteorology, Melbourne.

O’Connor TG (1996). Hierarchical control over seedling recruitment of the bunchgrass Themeda triandra in a semi-arid savanna. Journal of Applied Ecology 31, 155-171.

Thapa R (2010). Low cost rehabilitation of perennial grass pastures by managing seedling recruitment. PhD, Charles Sturt University, Wagga.

Previous PageTop Of PageNext Page