An investigation of runoff from raised beds and other tillage methods in the high rainfall zone of south-western Victoria, Australia.
Raised beds (RB) as a method of tillage for surface drainage of cropping land were introduced in the late 1990s to south-western Victoria, Australia. They are a popular tillage choice for crop production on the texture-contrast soils in this region (Peries et al. 2004). Farmers have introduced RB to reduce the risk of cropping land becoming waterloggcd. Indeed, this region is classified as having very high susceptibility to waterlogging (Robinson et al. 2003), which often causes poor plant growth and in some cases total crop failure (McDonald and Gardner 1987).
Variations in soil properties and field-scale topography partly determine whether water movement is over the land surface or through the soil profile. The combined effect of climate and soil type can result in excess surface water across the landscape of south-western Victoria, and this is due to the poor internal drainage of the texture-contrast soils (Maher and Martin 1987), where winter and spring waterlogging often leads to subsurface lateral flow or surface runoff. Observations on nutrient export and soil loss from cropping land elsewhere have raised concerns about the off-site impacts of runoff, such as through the eutrophication of surface waters (Lovett et al. 2007; Cogle et al. 2011). Despite the potential of these negative side-effects from runoff, it is a component of the hydrological cycle that can assist in reducing the impact of land becoming waterlogged.
Raised beds provide a form of surface drainage that is designed to create improved soil conditions for crop production. The furrows enable a direct pathway for excess water to flow away from the beds. Furrows have been used successfully to provide drainage in irrigated cropping systems, such as in vegetable production (Nguyen et al. 1988). Various forms of RB have been used around the world for hundreds of years to improve surface water movement (Barrow 1999).
The unique surface topography of RB is thought to affect the soil hydrology. Tisdall and Hodgson (1990) reported that RB (also referred to as 'ridge tillage') resulted in greater surface movement of water than alternative tillage treatments in the same production system. In a study over 5 years in Western Australia, Bakker et al. (2005) showed that RB increased runoff compared with control plots at four experimental sites. Several factors play an important role in determining the hydrological partitioning in RB. For instance, Hulugalle et al. (2002) reported that bed width had a significant effect on the runoff volume produced, while soil compaction in the furrows has been shown to cause runoff (Oltenfreiter et al. 2003). The amount of tillage (e.g. zero v. conventional tillage) can significantly affect the volume of runoff produced (Carroll et al. 1997). Many studies have reported on the runoff produced from irrigated cropping systems, but fewer have reported on the runoff generated from the high rainfall zones of southern Australia that are dependent on winter rainfall.
The effect of RB in changing the surface hydrology and, consequently, the runoff and erosion potential may be counterbalanced by improvements in soil structure and soil drainage near saturation. For example, Holland et al. (2007) found that soil in RB had a better connected pore network than a soil managed under conventional cultivation. This was consistent with measurements that showed the RB soil was also able to transmit solute faster. Measurements on intact cores in the laboratory indicated that there was a reduced risk of the soil becoming waterlogged. Field results reported by Holland et al. (2008) confirmed that RB soil was better drained and more aerated than conventionally cultivated soil. Therefore, the available evidence presents a conundrum: does the unique surface topography of RB with the presence of furrows lead to increased runoff, or does the improved soil structure in RB result in similar or decreased runoff compared with other tillage treatments? To resolve this uncertainty, the objectives of this study were: (i) to assess and compare the runoff from RB with other tillage treatments, i.e. namely conventional cultivation and deep cultivation; and (ii) to improve understanding of the factors that influence the amount of runoff from these tillage treatments by analysing rainfall data and other physical variables, including basic soil properties. In addition, crop measurements from each tillage treatment were compared to investigate whether crop biomass affected the runoff volume produced.
Material and methods
The experimental site was near Mt Pollock, 30km west of Geelong (38 [degrees] 10' S, 144[degrees]05' E) in south-western Victoria. The soil and landscape at this site are similar to the surrounding area that is used for crop production or has been targeted for cropping using RB in the future. The soil has a sandy clay loam texture in the A horizon (0-15cm) and a gravely clay in the B horizon (16-100+ cm). It is classified as a vertic, subnatric, grey Sodosol (lsbell 2002) and the soil profile is described according to McDonald et al. (1990) in Table 1.
Prior to the establishment of this experiment in 1999, the paddock had been a traditional, perennial pasture grazed by sheep. The plot area was 1.8 ha comprising nine 0.2-ha plots (20m width, 100m length). The three tillage treatments compared were raised beds (RB), conventional cultivation (CC), and deep cultivation (DC). Each treatment was replicated three times across the field in an east-west direction using a randomised block design. Each year except for 2001, the tillage plots were cropped with grains or oilseeds. According to best management practice, each year the furrows were sown to minimise the potential for erosion. Prior to beds being formed, the RB plots were deep-cultivated (up to 30 cm depth). The soil was then formed into beds 2m wide and 15-20 cm high with furrows (40 cm wide) separating the beds (Fig. 1). Treatments CC and DC did not have furrows and their surfaces were flat and even. The CC treatment was tilled to a shallow depth (<5 cm), whereas the DC treatment received an initial deep cultivation undertaken with a chisel plough (to 20 cm depth). Subsequent to establishment a minimum tillage policy was adopted, and the only cultivation that the RB treatment received was along the furrows once every 3 years. This was undertaken to ensure satisfactory water flow along the furrows. The slope down whole plot area was <1%. Crop management practices were implemented across the experiment to ensure that the plot management reflected standard local agronomy. For instance, typically there were three herbicide applications, and mono-ammonium phosphate fertiliser (100 kg [ha.sup.-1]) was placed with the seed at sowing.
An automatic weather station was installed at Mt Pollock and equipped to record humidity, temperature, wind speed, and solar radiation at hourly intervals. Rainfall was measured with a tipping-bucket rain gauge that logged every 0.2 mm of rainfall and recorded on an hourly basis. The long-term average rainfall for Mt Pollock was estimated using the SILO Data Drill set (SILO 1998). The data were calculated for the latitude and longitude of Mt Pollock by employing a spatial interpolation technique (Jeffrey et al. 2001). Consequently, the long-term average rainfall is synthetic, but it was derived from actual point data at nearby Bureau of Meteorology stations. Individual rainfall events in this study were defined as periods of rainfall inclusive of breaks <12 h. This is the same criterion as used by Bracken (nee Bull) et al. (2008) and was chosen to allow the ground surface to partially drain between rainfall events. This assisted in identifying independent rainfall and runoff events. Rainfall events were continuously recorded from July 2000 to December 2005. Weather station malfunction resulted in several periods of missing data between January and June 2000. Potential evapotranspiration (E[T.sub.p]) was calculated using the Penman-Monteith equation (Allen et al. 1998) and generally followed a seasonal trend with ~ 1-2 mm [day.sup.-1] during winter and up to 10 [mm.sup.-1] day in summer.
Surface runoff volume per unit area (mm) was measured from the RB, CC, and DC tillage plots. The tillage plots were hydrologically isolated using compacted earth mounds and surface drains along the perimeter of each plot. Flows were measured using modified Replogle-Bos-Clemmens type flumes with a throat width of 100mm (Clemmens et al. 1984) and a hydrostatic water level sensor. These flumes were installed at the lowest point of each plot, and flow heights were recorded on a data logger (Mindata Australia Pty Ltd, Richmond, Australia) every 10 min (the most frequent interval possible), from which the runoff volume was calculated (minimum was 0.2 mm) from each plot. Runoff volumes were recorded for 6 years from 2000 to 2005.
Crop measurements of grain yield and plant dry matter (t[ha.sup.-1]) were taken each year at harvest for each treatment. Regular visual assessments were made on the percentage groundcover and crop condition. Surface (0-10 cm) soil water content [theta] ([m.sup.3][m.sup.3]) was measured 55 times with a soil water sensor (ThetaProbe[R] Delta-T Devices, Cambridge, UK) on the RB and CC treatments in 2 years only (2003 and 2004). Soil bulk density (g[cm.sup.3]) was determined using the core method (McKenzie et al. 2002). Intact cores (height 6.3 cm, internal diameter 7.3 cm) were sampled at the soil surface 0-10 cm depth. A disc permeameter was used to measure unsaturated hydraulic conductivity (McKenzie et al. 2002) of the RB and CC surface soil at four different tensions (-40, -30, -20, and -10 mm) on 29 September 2003 and 08 March 2004.
A natural logarithmic transformation, In(runoff+ 1), was used to stabilise the mean--variance relationship of the runoff data. The log-transfomaed data were modelled using a linear mixed model which included fixed effects for treatment, date, and the interaction of treatment and date. Random effects included replicate and plot effects. The Pearson's product-moment correlation coefficient was used to assess the strength of the relationship between In(runoff+ 1) and rainfall variables. When all 736 rainfall events were considered for regression relationships between runoff and rainfall variables, there were problems with model fit, particularly for rainfall > 10 mm. These problems were due to many rainfall events not producing any runoff (zero-inflated data). The dataset was reduced to include only rainfall events for which there was runoff, thereby overcoming the model fit problems for rainfall >10mm. For this reduced dataset, simple linear regression relationships were then fitted between In(runoff+ 1) and rainfall variables including 'total event rainfall', 'rainfall intensity', or 'maximum rainfall in 1 h'. Following the work of Abrisqueta et al. (2007), a variable formed from the product of rainfall amount and rainfall intensity was also investigated as an explanatory variable for the untransformed runoff data, but the residuals from the model did not meet model assumptions of identical distribution. For each linear regression, the significance of treatment differences in the intercept and slope was assessed. The significance of the fixed effects (P<0.05) in all models was determined using approximate F-tests with the denominator degrees of freedom calculated according to Kenward and Roger (1997). All models were fitted using ASREML 2.0 software (Gilmour et al. 2005). A statistical comparison between the treatments for grain yield and plant dry matter was made by calculating the least significant difference (l.s.d.) using analysis of variance (ANOVA) with treatment and year as factors. Because of the unbalanced structure of bulk density and the hydraulic conductivity measurements, a paired Student's t-test was used to determine significant (P< 0.05) differences between means.
During most of the study period, annual rainfall was less than the long-term average of 556 mm. The annual rainfall had a major influence on the annual total runoff produced from all treatments (Fig. 2). The wettest year, 2001, had the most runoff and was the only year with annual rainfall greater than average (it was decile 5). The other years were much drier, and as a result much smaller runoff volumes were recorded for all treatments. For instance, 2002 was decile 1; 2000, 2003, and 2005 were decile 2; and 2004 was decile 3. This runoff study was undertaken during a period that was mostly drier than the long-term rainfall records.
There was a wide distribution in the size of individual rainfall events, with a large proportion of small rainfall events. Table 2 shows the number, percentage, and mean of all rainfall events. From July 2000 to December 2005, 736 individual rainfall events were recorded, and of those, 97% did not produce any runoff The smallest rainfall event to produce runoff was 5.6 mm, but there were 51 events with >10 mm rainfall that produced no runoff. Moderate rainfall events (5.6-25 mm) accounted for 17% of the total, and there were only 12 large ([greater than or equal to] 25 mm) rainfall events, which corresponded to <2% of all events. No events with rainfall between 53 and 137 mm were observed. The variability in the relationship between rainfall amount and runoff suggested that other factors such as rainfall intensity and soil properties were also important. There was a large variation in rainfall intensity, but most events (84%) were <1 mm [h.sup.-1]. The many fewer, higher intensity events were most significant in producing runoff.
Treatment RB released the greatest total runoff volume (187 mm) during the study period. In comparison, the total runoff for CC was 146mm (78% of RB) and for DC it was 118mm (63% of RB). Overall, surface runoff was recorded 24 times between 2000 and 2005 at Mt Pollock. However, there was one runoff event from which data could not be analysed (on 28 June 2002) due to flume equipment failure. Also, on 01 February 2005, there was uncertainty about the runoff from four plots because the earth mounds between these plots collapsed. The amount of runoff from the RB, CC, and DC treatments and the accompanying rainfall amount (mm) and rainfall intensity (mm [h.sup.-1]) for each event from 2000 to 2005 are given in Table 3. The largest runoff event (21 April 2001) accounted for >50% of the total runoff produced from all tillage treatments during the study period. There was also large variation in the runoff produced between years as a result of differences in rainfall distribution and intensity, as well as changes in physical factors (such as the soil properties and crop biomass). This variation in runoff between years is not uncommon for this climate and is similar to that recorded by Melland et al. (2008) from 1998-2000 in western Victoria.
Of the 23 runoff events analysed, there was a significant difference (P< 0.05) between the treatments on nine occasions. These differences were mostly found during winter-spring (June-November). Seasonality in the generation of runoff has also been reported elsewhere (Wei et al. 2007). Analysis of the main treatment effect across all runoff events showed that the runoff volume was significantly different between treatments. Treatment RB produced the greatest runoff for seven of the nine events where significant differences were found. On three occasions, runoff from RB was significantly greater than from both of the other tillage treatments (23 July 2002 01 August 2002, 30 August 2005), but on one occasion (13 August 2004), CC yielded significantly more runoff than the other treatments. Overall, the CC and DC treatments were significantly different from each other on only three occasions, while the DC treatment produced the least runoff for each event.
Runoff and rainfall
The relationship between rainfall and runoff in each year (Fig. 2) shows an influence of rainfall on the amount of runoff produced and the effect that this has on differences between treatments. In particular, the large amount of rainfall in 2001 corresponded with RB producing much more runoff than CC and DC. This difference was mostly due to the event on 21 April 2001; Fig. 3 presents the average cumulative runoff for each treatment with the cumulative rainfall. Although the following year, 2002, was the driest of the experiment, again RB produced greater runoff than the other treatments. Subsequently, in 2004 CC produced more runoff than RB, and in 2005 there was very little difference between any of the treatments. While an analysis of the relationship between total annual rainfall and runoff is of some value, it is the interdependent nature of individual rainfall events and runoff that is more informative than the accumulated total in any given year.
For the 23 runoff events (Table 3), In(runoff + l) was significantly correlated with each of the rainfall variables (rainfall amount, rainfall intensity, rainfall hours, and maximum hourly rainfall) as shown by the correlation coefficients in Table 4. The significance level of the correlation between runoff and the four rainfall variables was generally highest for RB compared with the other two treatments, especially in the cases of runoff v. rainfall intensity and runoff v. maximum hourly rainfall (Table 4). The slopes of the regressions of In(runoff + 1) against rainfall for the three tillage treatments were the same, but the intercepts differed significantly for RB v. DC and RB v. CC. The regression equations fitted to the log-transformed runoff values and their [R.sup.2] values were:
ln(RB runoff + 1) 0.51 + 0.03 x rainfall; [R.sup.2] = 0.46 (1)
ln(CC runoff + 1) 0.28 + 0.03 x rainfall; [R.sup.2] = 0.36 (2)
ln(DC runoff + 1) = 0.12 + 0.03 x rainfall; [R.sup.2] = 0.41 (3)
The back-transformed relationships of the linear model for runoff v. rainfall for each treatment are presented in Fig. 4, which shows that the differences in runoff between the three treatments became progressively larger as the rainfall in an event increased. However, there were insufficient large rainfall events during the study period to confirm this indication.
Comparison of the percentage runoff for each rainfall event shows that, in most cases, runoff was <10% of the rainfall event amount (Table 3). There was some interaction between runoff, rainfall amount, and rainfall intensity as shown in Fig. 5, but again, the shortage of observations for large rainfall events meant that the form of this interaction was inconclusive. Nevertheless, comparison between rainfall events >10mm which resulted in runoff and those which did not showed that the runoff-producing rainfall events were on average larger (rainfall 34 v. 15.9mm), longer (38 v. 28h), and had greater rainfall intensity (1.5 v. 1.39 mm[ h.sup.-1]).
Soil influences on runoff
Overall, the RB soil had significantly (P<0.001) lower bulk density (1.33 v. 1.46 Mg [m.sup.-3]) than the CC soil. However, there were few significant differences in hydraulic conductivity between the RB and CC soils; only once (in September 2003), CC was greater at -40 mm tension than RB (P< 0.05). Although measured surface soil water content showed significant differences between the RB and CC treatments in 2003 and 2004, these measurements were not always made close to (i.e. before or after) runoff events. On four occasions near the time of runoff events, the RB soils were significantly drier than CC, and on three of these occasions, CC tended to produce greater runoff volume. Thus, for these runoff events the soil water content probably played a strong role in influencing the runoff produced; however, this was not observed consistently and the difference in the runoff was not significant for these events. The soil water content was seldom measured near to the estimated value for saturation (derived from the total porosity of the soil in each treatment); thus, saturation excess runoff was unlikely to have taken place. Furthermore, the below-average rainfall during most of the study period meant that there were very few occasions when waterlogging was visible or when the crop showed signs of stress from wet soil (e.g. stunted, yellowing plants). Infiltration excess process (or Hortonian flow type) (Haygarth et al. 2000) was probably the most common mechanism for runoff generation observed over the study period. The occurrence of infiltration excess runoff depends not only on rainfall intensity but also on surface soil conditions. Given that there were few differences in hydraulic conductivity between treatments, some relationship between runoff and rainfall intensity would be expected, as discussed above and evident in Fig. 5.
Effects on crops
Yields of grain and dry matter are presented in Table 5. Although significant differences were found between the treatments, they were not consistent between years; no treatment produced consistently higher yields in the drier years of this experiment. No yields were recorded in 2001 because the sown crop failed due to severe slug damage following the wet conditions in April. In this year, the wettest of the trial, of the eight runoff events recorded, seven showed RB producing more runoff than DC or CC, but the difference was significant in only one of these events. Nevertheless, these results provide some evidence that in the absence of plant cover RB produces more runoff in any one event than the other two treatments. The differences between treatments were most notable for the event of 21 April 2001 (see Fig. 3), the largest event of the whole trial, when runoff amounts were 102 mm (s.d. 27), 71 mm (s.d. 39), and 56 mm (s.d. 19) for RB, CC, and DC respectively.
Runoff from the tillage treatments in the other years, when rainfall was below-average, suggested that differences in crop growth (biomass) between the treatments were not large enough to influence the runoff produced. Although RB produced the greatest runoff on most occasions when runoff occurred, the partitioning of crop biomass was not consistent. For example, in 2003 RB had significantly less plant biomass than the CC treatment, but in the same year, RB produced more grain than the other tillage treatments (Table 5).
Freebairn and Boughton (1981) found that the amount of crop residue reduced the amount of runoff and peak runoff rates. However, there was no clear relationship between crop biomass and the amount of runoff produced in this study, probably because most runoff events took place when ground cover was 100% and there was a full crop canopy. An exception was the large event of 21 April 2001 when the order of runoff amounts was RB > CC > DC. Thus, while the amount of vegetation cover is one factor that could differentiate between contrasting agricultural land uses, such as between cropped land and pasture, the tillage treatments in this study showed no consistent effect of crop biomass on runoff.
Other soil properties measured at the experimental site were reported by Holland et al. (2007) and showed that RB soil had significantly lower bulk density as well as improved solute transport properties. Thus, it was expected that the better structure of the RB soil would allow increased infiltration and result in less runoff. However, field measurements of hydraulic conductivity showed a significant difference (P<0.05) only once at one tension, and this is insufficient to explain the differences in runoff between the treatments. Other soil factors have been identified as having an effect (ranging from dominant to minor) on runoff generation and the amount of runoff produced; for example, on some soils, surface sealing is a significant problem (Mamedov et al. 2001 ; Withers et al. 2007), but it was not observed during this study. Mamedov et al. (2001) were able to show that the wetting rate of soil aggregates was a dominant factor in determining runoff, which in turn was dependent on sodicity and clay content. Certainly, the B horizon of this soil is sodic (see Table 1), which could have had a marked influence on the permeability of the soil exposed in the RB furrows, especially given the depth of the furrows, i.e. the top of the B horizon (see Fig. 1). There may also have been some sodic clay brought to the surface during the initial deep cultivation that preceded the formation of the RB and also was carried out for the DC treatment (see Materials and methods). These are other important soil properties to consider, given the widespread presence of sodic clay subsoils in south-western Victoria, and care must be taken during the formation of RB to ensure that subsoil is not brought to the surface where it could potentially lead to surface sealing.
Despite the paucity of soil measurements taken at times close to the runoff events, those measurements that were taken did not provide a consistent explanation of the differences in runoff between the treatments. Furthermore, the lack of consistent differences in crop biomass meant that the differences in runoff from this factor were difficult to discern and were largely due to other factors such as rainfall characteristics and surface topography. Indeed, we suggest that the predominant factor controlling runoff was the surface micro-topography of the tillage treatments; in particular, the presence of furrows (exposing the sodic subsoil) within the RB was important, as well as the controlled traffic along the furrows, which is associated with the RB tillage treatment (Tullberg et al. 2001). Wheel tracks may have acted as conduits for surface water flow as has been reported elsewhere (Withers et al. 2007). Thus, the furrows of the RB were a key factor that increased the volume of runoff from RB compared with CC and DC over the study period.
This study of runoff from different tillage treatments (RB, CC, and DC) in a rain-fed environment in south-western Victoria showed the sensitivity of runoff to rainfall, such that there was little difference in runoff between treatments in the drier, below-average rainfall years, i.e. for 5 of the 6 years of the trial. In contrast, RB significantly increased the amount of runoff during wetter periods when above-average rainfall was received. Infiltration excess runoff was the predominant process generating runoff observed for all tillage treatments, although this was probably a function of the lower than average rainfall and lower soil water contents that prevailed during most of the study period. Regression analysis established that rainfall event amount was the most important factor influencing the volume of runoff produced, but there was also a differential response to rainfall intensity between tillage treatments. No consistent impact of runoff on crop production was found, and the measured soil properties did not assist in explaining the treatment differences in runoff. The predominant factor that led to RB yielding greater runoff than CC and DC treatments was the surface micro-topography of the RB, primarily the presence of furrows which acted as conduits for the flow of surface water.
Financial support for this work was provided by the Grains Research and Development Corporation under the Sustainable Fanning Systems program. The authors thank Rowan Peel, the farmer of Mt Pollock, for access to the experimental site and assistance with equipment installation in the field. Mr Mark Weedon is thanked for writing a macro to analyse the rainfall data.
Received 16 August 2011, accepted 4 July 2012, published online 15 August 2012
Abrisqueta JM, Plana V, Mounzer OH, Mendez J, Ruiz-Sanchez MC (2007) Effects of soil tillage on runoff generation in a Mediterranean apricot orchard. Agricultural Water Management 93, 11-18. doi:10.1016/ j.agwat.2007.06.002
Allen RG, Pereira LS, Raes D, Smith M (1998) 'Crop evapotranspiration: Guidelines for computing crop water requirements.' (FAO: Rome)
Bakker DM, Hamilton G J, Houlbrooke D, Spann C (2005) The effect of raised beds on soil structure, waterlogging, and productivity on duplex soils in Western Australia. Australian Journal of Soil Research 43, 575 585. doi:10.1071/SR03118
Barrow CJ (1999) 'Alternative irrigation--the promise of runoff agriculture.' (Earthscan: London)
Bracken (nee Bull) LJ, Cox N J, Shannon J (2008) The relationship between rainfall inputs and flood generation in south-east Spain. Hydrological Processes 22, 683 696. doi:10.1002/hyp.6641
Carroll C, Halpin M, Burger P, Bell K, Sallaway MM, Yule DF (1997) The effect of crop type, crop rotation, and tillage practice on runoff and soil loss on a vertisol in central Queensland. Australian Journal of Soil Research 35, 925-940. doi: l 0.1071/S96017
Clemmens AJ, Marinus GB, Replogle JA (1984) Portable RBC flumes for furrows and earthen channels. Transactions of the American Society of Agricultural Engineers, 1016-1024.
Cogle AL, Keating MA, Langford PA, Gunton J, Webb IS (2011) Runoff, soil loss, and nutrient transport from cropping systems on Red Ferrosols in tropical northern Australia. Soil Research 49, 87-97. doi: 10.1071 / SR10069
Freebaim DM, Boughton WC (1981) Surface runoff experiments on the eastern Darling Downs. Australian Journal of Soil Research 19, 133-146. doi:10.1071/SR9810133
Gilmour AR, Gogel B J, Cullis BR, Thompson R (2005) 'ASReml User Guide Release 2.0.' (VSN International Ltd: Hemel Hempstead, UK)
Haygarth PM, Heathwaite AL, Jarvis SC, Harrod TR (2000) Hydrological factors for phosphorus transfer from agricultural soils. Advances in Agronomy 69, 154-178.
Holland JE, White RE, Edis R (2007) The relation between soil structure and solute transport under raised bed cropping in south-western Victoria. Australian Journal of Soil Research 45, 577-585. doi:l 0.1071/SR07068
Holland JE, White RE, Edis R (2008) Improved drainage and greater air-filled porosity of raised beds in south-western Victoria. Australian Journal of Soil Research 46, 397-402. doi: 10.1071/SR08003
Hulugalle NR, Rohde KW, Yule DF (2002) Cropping systems and bed width effects on runoff, erosion and soil properties in a rainfed Vertisol. Land Degradation & Development 13, 363 374. doi: 10.1002/ Idr.510
Isbell RF (2002) 'The Australian Soil Classification.' (CSIRO Publishing: Melbourne)
Jeffrey S J, Carter JO, Moodie KM, Beswick AR (2001) Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environmental Modelling & Software 16, 309-330. doi:10.1016/S 1364-8152(01)00008-1
Kenward MG, Roger JH (1997) The precision of fixed effects estimates from restricted maximum likelihood. Biometrics 53, 983-997. doi:10.2307/2533558
Lover S, Price P, Edgar B (2007) 'Salt, nutrient, sediment and interactions: Findings from the National River Contaminants Program.' (Land & Water Australia: Canberra)
Maher JM, Martin JJ (1987) Soils and landforms of south-western Victoria. Part 1. Inventory of soils and their associated landscapes. Research Report No. 40. Department of Agriculture and Rural Affairs, East Melbourne.
Mamedov AI, Levy G J, Shainberg I, Letey J (2001) Wetting rate, sodicity, and soil texture effects on infiltration rate and runoff. Australian Journal of Soil Research 39, 1293 1305. doi: 10.1071/SR01029
McDonald GK, Gardner WK (1987) Effect of waterlogging on the grain yield response of wheat to sowing data in south-western Victoria. Australian Journal of Experimental Agriculture 27, 661 670. doi: 10.1071/EA9870661
McDonald RC, Isbell RF, Speight JG, Walker J, Hopkins MS (1990) 'Australian soil and land survey--Field handbook.' (Inkata Press: Melbourne)
McKenzie N J, Coughlan KJ, Cresswell HP (2002) 'Soil physical measurement and interpretation for land evaluation.' (CSIRO Publishing: Melbourne)
Melland AR, McCaskill MR, White RE, Chapman DF (2008) Loss of phosphorus and nitrogen in runoff and subsurface drainage from high and low input pastures grazed by sheep in southern Australia. Australian Journal of Soil Research 46, 161 172. doi: 10.1071/SR07084
Nguyen VQ, Hermus R, McGlasson WB, Meldrum SK, Trethowan ME (1988) Fresh market tomatoes for production on raised beds in New South Wales. Australian Journal of Experimental Agriculture 28, 127-135. doi:10.1071/EA9880127
Oltenfreiter G, Gabriels D, Conrnelis WM, Van Meirvenne M, Schiettecatte W, Verbist K, Hartmann R, Hofman G (2003) Effect of a compacted ploughsole on the occurence of a temporary saturated zone. In 'International Soil Tillage Research Organisation Conference'. Brisbane, Australia. pp. 875-880. (ISTRO)
Peries R, Wightman B, Bluett C, Rab A (2004) Raised bed cropping in southern Victoria--A snapshot of a productive and sustainable option for waterlogging prone soils. In '4th International Crop Science Congress'. Brisbane, Australia. (The Regional Institute Ltd: Gosford, NSW)
Robinson N, Rees D, Reynard K, MacEwan R J, Dahlhaus PG, Imhof M, Boyle G, Baxter N (2003) A land resource assessment of the Corangamite region. Department of Primary Industries, Victoria.
SILO (1998) 'The SILO Data Drill.' (NR&M, Queensland Government: Brisbane) Available at: www.longpaddock.qld.gov.au/silo/
Tisdall JM, Hodgson AS (1990) Ridge tillage in Australia: a review. Soil & Tillage Research 18, 127 144. doi:10.1016/0167-1987(90)90055-I
Tullberg JN, Ziebarth P J, Li Y (2001) Tillage and traffic effects on runoff. Australian Journal of Soil Research 39, 249 257. doi: 10.1071/SR00019
Wei L, Zhang B, Wang M (2007) Effects of antecedent soil moisture on runoff and soil erosion in alley cropping systems. Agricultural Water Management 94, 54 62. doi: 10. 1016/j.agwat.2007.08.007
Withers PJA, Hodgkinson RA, Bates A, Withers CL (2007) Soil cultivation effects on sediment and phosphorus mobilization in surface runoff from three contrasting soil types in England. Soil & Tillage Research 93, 438 451. doi: 10.1016/j.still.2006.06.004
J. E. Holland (A,D,E), T. H. Johnston (B), R. E. White (A), and B. A. Orchard (C)
(A) Melbourne School of Land and Environment, University of Melbourne, Vic. 3010, Australia.
(B) Department of Primary Industries, Geelong, Vic. 3220, Australia.
(C) Department of Primary Industries NSW, Wagga Wagga Agricultural Institute, Pine Gully Road, Wagga Wagga, NSW 2650, Australia.
(D) Current address: 14 Dalton St, Wagga Wagga, NSW 2650, Australia.
(E) Corresponding author. Email: firstname.lastname@example.org
Table 1. Soil profile description for the experimental site at Mt Pollock ESP, Exchangeable sodium percentage Horizon Depth (cm) Colour, texture, morphology, pH, ESP A1 0-15 Dark brown (1OYR 3/3); fine sandy clay loam; weakly pedal; 20-30% gravel; pH 6.2; ESP 5.0 B21 16-75 Dark greyish brown (1OYR 4/2); fine gravely clay; fine polyhedral; 50% gravel; pH 7.4; ESP 12.4 B22 76-100 Brownish yellow (1OYR 6/6); sandy clay; quartz; little gravel; pH 9.1; ESP 23.6 B23 >100 Weak red (2.5YR 5/2); gravely clay; slickensides; 50% gravel; pH 9.0; ESP 25.3 Table 2. Number, percentage, and mean of all rainfall events according to rainfall event amount and intensity from July 2000 to December 2005 Event amount (mm) [greater than or <5.6 5.6-<15 15-<25 equal to]25 No. of events 600 95 29 12 of events 82 13 4 2 Mean 1.2 8.9 18.7 45 s.d. 1.4 2.6 2.7 33.4 Event intensity (mm [h.sup.-1]) <1 1-2 >2 No. of events 620 88 28 of events 84 12 4 Mean 0.4 1.3 2.9 s.d. 0.2 0.2 2.5 Table 3. Runoff amount for the tillage treatments (raised beds, RB; conventional cultivation, CC; deep cultivation, DC), rainfall amount, and rainfall intensity at Mt Pollock from 2000-2005 Rainfall intensity was calculated on whole hours of rainfall only and excluded hours during rainfall events when no rainfall was recorded. Values are means (s.d. in parentheses); within dates, means followed by the same letter are not significantly different at P=0.05; for dates where no letters are given, means are not significantly different Date RB Runoff amount (mm) CC 23 Oct. 2000 2.2 (1.6) 1.6 (1.4) 0l Nov. 2000 0.3a (0.5) 0.3ab (0.4) 21 Apr. 2001 101.8 (27) 70.6 (39) 13 June 2001 5.9 (0.6) 3.6 (3.1) 20 June 2001 2.6 (0.4) 1.2 (1.0) 18 Aug. 2001 2.4 (0.4) 2.3 (1.1) 20 Aug. 2001 13.3 (2.1) 13.4 (1.5) 22 Aug. 2001 13.1 (1.3) 11.9 (0.4) 25 Sept. 2001 1.3a (0.1) 0.2ab (<O. I) 11 Nov. 2001 0.9 (0.3) 0.3 (<0.1) 07 Feb. 2002 11.6 (5.4) 6.6 (4.7) 23 July 2002 4.2b (0.9) 2.36 (2.0) 01 Aug. 2002 l .Ob (0.1) 0.07ab (<O. I) 23 July 2003 0.24b (0.3) 0.04ab (0.1) 23 Aug. 2002 0.8b (0.4) 0.13ab (0.1) 01 Oct. 2003 0.2 (0.2) 0.05 (0.1) 30 Oct. 2003 0.13 (0.2) 0.01 (<0.1) 25 July 2004 6.4b (1.9) 5.5b (3.5) 13 Aug. 2004 0.43a (0.5) 1.9b (0.4) 10 Sept. 2004 2.2 (1.3) 6.3 (3.3) 01 Feb. 2005 (A) 12.9 (1.1) 13.9 04 Feb. 2005 2.8 (0.6) 3.3 (1.2) 30 Aug. 2005 1.2b (0.7) 0.1ab (0.1) Total 187.4 145.6 I.s.d. (P=0.05) 1.95 Rainfall Rainfall intensity Date DC amount (mm (mm) [h.sup.-1]) 23 Oct. 2000 1.4 (0.3) 52 1.3 Ol Nov. 2000 0.9b (0.8) 16 1.1 21 Apr. 2001 56.5 (19) 138 2.5 13 June 2001 2.8 (1.3) 11 0.9 20 June 2001 1.4 (0.5) 8 0.9 18 Aug. 2001 2.0 (0.4) 23 0.8 20 Aug. 2001 12.3 (3.6) 32 1.5 22 Aug. 2001 11.8 (1.9) 18.2 1.2 25 Sept. 2001 0.1b (0.1) 18 1.2 11 Nov. 2001 0.3 (0.1) 22 1.1 07 Feb. 2002 7.9 (7.1) 53 2.7 23 July 2002 0.1 a (0.1) 21 2.6 01 Aug. 2002 0.03a (<0.1) 12 0.8 23 July 2003 0.02a (<0.1) 27 1.5 23 Aug. 2002 0.07a (0.1) 19 1.4 01 Oct. 2003 0.06 (0.1) 29 1.1 30 Oct. 2003 0.02 (<0.1) 23 1.4 25 July 2004 2.8a (4.6) 21 1.9 13 Aug. 2004 0.94a (1.6) 15 0.8 10 Sept. 2004 3.5 (5.4) 19 0.8 01 Feb. 2005 (A) 9.2 (0.7) 90 1.8 04 Feb. 2005 3.1 (1.8) 6 1.1 30 Aug. 2005 0.2a (0.1) 28 1.7 Total 118.2 I.s.d. (P=0.05) (A) This runoff event had missing values for two CC plots (6 and 8), one RB plot (7), and one DC plot (9). Table 4. Pearson's correlation coefficients between ln(runoff + 1) (RO) and rainfall variables (rainfall amount, R; rainfall intensity, [R.sub.i]; rainfall hours, [R.sub.h]; and maximum hourly rainfall, [R.sub.m]) for tillage treatments combined and for each tillage treatment (raised beds, RB; conventional cultivation, CC; deep cultivation, DC) * P<0.05; ** P<0.01; *** P<0.001 Correlations ln(runoff + 1) Overall, RB, CC, DC, n=189 n=63 n=63 n=63 RO: R 0.64 *** 0.68 *** 0.61 *** 0.65 *** RO: [R.sub.i] 0.44 *** 0.56 *** 0.41 *** 0.35 ** RO: [R.sub.h] 0.47 *** 0.45 *** 0.47 *** 0.51 *** RO: [R.sub.m] 0.35 *** 0.41 *** 0.30 * 0.34 ** Table 5. Crop grain yield and plant dry matter (t [ha.sup.-1]) at harvest for raised bed (RB), conventional cultivation (CC), and deep cultivation (DC) treatments from 2000 to 2005 No grain yield or plant dry matter recorded for 2001 as the crop failed. Means followed by the same letter are not significantly different at P=0.05 Grain yield Dry matter RB CC DC RB CC DC 2000 5.06c 4.27a 4.65b 14.4a 13.8a 14.3a 2002 2.03a 1.99a 2.00a 8.9a 10.7c 9.8b 2003 6.93c 6.32a 6.67b 18.6a 20.7b 18.7a 2004 3.92a 4.64c 4.23b 12.5a 15.9b 13.3a 2005 1.45a 1.5a 1.62b 8.4a 9.2b 9.3b I.s.d. (P=0.05) 0.07 0.80
|Printer friendly Cite/link Email Feedback|
|Author:||Holland, J.E.; Johnston, T.H.; White, R.E.; Orchard, B.A.|
|Date:||Aug 1, 2012|
|Previous Article:||Evaluation and modelling of furrow infiltration for uncropped ridge-furrow tillage in Loess Plateau soils.|
|Next Article:||Shift in C and N humification during legume litter decomposition in an acid tropical Ferralsol.|