Printer Friendly

Sandblasting damage of narrow-leaf lupin (Lupinus angustifolius L.): a field wind tunnel simulation.


In the last 20 years, narrow-leaf lupin (Lupinus angustifolius L.) has been developed into a major dryland crop species in Australia. Previously, cropping systems in southern Australia involved either continuous cereal or pasture/cereal rotations. Legume species are now an important component of farming strategies, providing an economic benefit in grain production as well as benefits to other crops in terms of weed management, nutrition, and disease break (Nelson et al. 2000). Lupins are a major legume species, especially in Western Australia where there is a predominance of acidic, sandy soils.

Farmers and agronomists in southern Australia have frequently expressed concern about the susceptibility of lupin to sandblasting damage. Unlike cereals where the shoot apex remains underground during vegetative growth, the lupin meristems are above ground after seedling emergence and vulnerable to physical damage. Buds are easily damaged and the plants cannot recover by producing new shoots (Dracup and Kirby 1996). The likelihood is that lupins will suffer greater damage and subsequent yield depression than cereals when subject to an equivalent amount of sandblasting during an erosion event. Also, the likelihood of soil erosion is increased during the establishment phase of lupins, as the accumulation of leaf area is slow during the first half of the crop life, taking 14 weeks to reach a leaf area index of 1 (Greenwood et al. 1975). This slow initial expansion of leaf area by lupin crops compared with cereals grown under the same conditions predisposes the crop to weed competition, insect damage, brown leaf spot (Pleiochaeta setosa), and greater risk of sandblast damage on wind erodable soils (Perry et al. 1986).

The risk of soil erosion varies in space and time, being dependent on the duration, turbulent structure, and strength of winds; soil type; the amount of anchored plant residue; soil moisture; and crop cover (Moore 1998). The interaction between these factors and resulting dust and sand transport has been simulated with computer models. The Wind Erosion Assessment Model (WEAM) has been shown to provide realistic estimates of sediment transport on light soil types in southern Australia (Shao et al. 1996). Although the risk of erosion can be reduced through agronomic practices such as minimising tillage and maintaining vegetation cover, it is not always possible to eliminate this risk due to variable and low rainfall (Marsh and Carter 1983). This is particularly the case for lupins, the preferred legume crop on light-textured soils, which also have a high potential of eroding.

Sandblasting occurs when eroding soil particles strike plant surfaces, thereby combining the processes of abrasion and tissue removal, which results in modification of the physiological performance of the plant (Cleugh et al. 1998). The extent of injury to a plant species will depend upon windspeed, abrasive flux, duration of exposure, size, shape, and density of the abrasive material; growth stage and condition of the plant; growing conditions; and gustiness of the wind (Skidmore 1966).

Several studies have been carried out using a wind tunnel together with pot-grown plants to examine factors influencing the impact of sandblasting. The importance of tissue removal and physiological response was studied in winter wheat (Armbrust 1974) and in grain sorghum (Armbrust 1982). The survival and growth of 4 grass species varied in response to several levels of wind and sandblasting. They showed greater sensitivity at the 2-leaf stage than the 6-leaf stage (Fryrear et al. 1973). Lyles and Woodruff (1960) found that alfalfa and switchgrass showed a linear yield reduction with increased exposure. The general trend in all these studies was a decreased production rate after exposure to wind plus sand.

Armbrust (1984) showed that sensitivity to sandblasting varies with age. Grain sorghum, soybean, and winter wheat seedlings exposed to several levels of sandblast at 5 different stages following emergence. He showed that greatest reductions in growth occurred in plants sandblasted 7-14 days after emergence. However, when millet (Pennisetem glaucum) was subjected to sandblasting, the effects of plant age on dry weight were less pronounced (Michels et al. 1995). The amount of damage as a result of sandblast varies from decreased survival rate, slower development, delayed maturity, and reduced yields to total loss (Greig et al. 1974; Downes etal. 1977; Armbrust 1984). Some crops grow better when subjected to mild sandblast injury than without sandblast (Fryrear et al. 1975). Woodruff (1956) exposed winter wheat (wheat sown in autumn under North American conditions where the crop over-winters under snow) to autumn and spring treatments of wind-blown sediment transport and developed relationships between wheat yield and sediment transport rate. Results indicated that damage was a function of the total amount of soil striking the plant rather than the time interval between exposures and that plants treated in an earlier stage of development showed little final yield response, whereas plants in spring (about 180 days after sowing) had substantially reduced grain yields.

The work summarised above has used several ways to express the level of sandblasting; these include mass transport (mass/width/time), mass flux (mass/area/time), total amount of sediment striking the plant, and total kinetic energy. Mass flux and mass transport are the standard units for presenting erosion rates; however, in this study they are not suitable because it is not only the rate of mass transport or flux, it is the duration and wind speed that affect plant damage. Thus, we have chosen to express sandblasting in units of total transport mass, TTM (kg/m), which is the time-integrated horizontal transport rate for a given wind speed. We chose 13.7 m/s because this wind speed has been recorded in the study area as creating sandblasting. This approach means that we have held the particle momentum constant (i.e. constant particle mass and velocity), but we have changed the total momentum for each treatment by increasing the total number of particles as represented by TTM.

The evolution of farming systems in Australia in regions prone to soil erosion where lupin is the preferred legume break crop has stimulated the need for additional information on the response to sandblasting. This is an essential first step in the development of management techniques relevant to these conditions. This may include tillage techniques that retain plant residue cover to increase the windspeed threshold for erosion and/or the integration of perennial vegetation such as windbreaks into farming systems to reduce windspeed and the subsequent severity of erosive flux.

The purpose of this paper is to identify and quantify damage to young plants of narrow-leaf lupin caused by exposure to varying durations of abrasion. The effect of this damage on the grain yield of plants grown in a field situation is quantified. This study contributes to the understanding of sandblasting impact on crop development and yield at one wind speed known to cause crop damage in the study area. While accepting limitations of mobile wind tunnel methodologies in reproducing atmospheric surface layer eddies larger than the wind tunnel characteristic dimensions, the analysis of field-grown plants in situ reflects a more realistic view of plant response within a crop canopy. This practice overcomes some distortions in plant growth and development that may occur when utilising pot-grown experimental plants.


Site description

Narrow-leaf lupin cv. Merrit was sown in furrowed rows at 150-mm spacing, leaving a fiat soil surface, on a gently sloping site (<2.5%) near Strathalbyn, South Australia (313900E, 6095500N; MGA Zone 54) on the 14 May 2000. The site was lightly scarified and harrowed before seeding to remove excess surface trash and the site rolled after sowing. The soil type was by the Australian Classification (Isbell et al. 1997) a Supracalcic, Red Sodosol; thick, non-gravelly, sandy/clayey and moderate (a Hypercalcic Solonetz by the FAO UNESCO classification); and is typical of the light sandy soils in which lupins are cultivated in Australia. The particle size distribution (PSD) of the 0-0.05 m soil layer is shown in Fig. 1. The soil and abrader sediment samples were particle-size analysed by Coulter Multisizer (Lines 1992) within the range 2-600 [micro]m according to the method of McTainsh et al. (1997). Particle-size data are expressed as volume per cent particle-size frequency distributions (Miller and Lines 1988). Volume or weight per cent are the units of convention within the earth sciences (Folk and Ward 1957).


Wind and sediment transport rate measurements

Measurements of wind velocity (u) and soil flux (q) were made within a portable wind tunnel (Raupach and Leys 1990). The wind tunnel was a blower type driven by a 56-kW diesel motor connected to a 1.5-m axial fan. The air passed through 4 m of flow conditioning before passing through the 7.2-m working section of the tunnel. The tunnel is 1.2 m wide and 0.9 m high, giving a 1.1-[m.sup.2] cross-sectional area. Flow conditioning is described in Raupach and Leys (1990). Average crop height at the time of the experiment was 5.3 cm. The tunnel was placed over each lupin plot and run at about 13.7 m/s in the free-stream of the tunnel at 0.6 m height. The wind velocities from within the wind tunnel were converted to the equivalent wind velocity at 10 m above the ground using the logarithmic wind profile (see Eqn 1), in order to make comparisons with the Australian Bureau of Meteorology data for the study area.

The procedure for calculating wind velocity at 10 m height ([u.sub.10]), friction velocity (u*), and surface roughness length ([Z.sub.o]) is the standard as outlined by Leys and Raupach (1991) of fitting a logarithmic profile to the measured profile of u(z) from within the tunnel:

u(z) = (u*/k) ln(z/[Z.sub.o]) (1)

where u(z) is the wind velocity at height z and k is the yon Karman constant (0.4).

Values for u* and [Z.sub.o] were inferred from profiles of mean velocity (u) measured using 5 Pitot-static tubes with a dynamic pressure port diameter of 1 mm. The Pitot-static tubes were placed at heights (z) 25, 50, 100, 150, 200, and 600mm above the mean ground level, which was estimated by eye. The heights (z) 25, 50, 100, 150, and 200 cm were used to calculate [z.sub.o] by calculating the intercept of the linear regression of ln(z) against u. No displacement was used in the calculation of u* or [Z.sub.o]. This implies that both u* and [Z.sub.o] could be overestimates; however, no simple solution for determining displacement height was available. The Pitot tube at 600 mm was used to check any variation between wind tunnel runs and not in the calculation of the velocity profile.

The first 2 m of the 7.2-m-long working section was covered with a board to prevent scouring from the aerodynamic trip (positioned 0.5 m) and the saltation introduction tubes (positioned 1.2 m) downwind of the start. The next 4 m was the treatment area and the remaining 1.2 m was not used due to wind flow complexities associated with the exhaust end of the tunnel.

The treatment area was divided into four 1-[m.sup.2] subplots for the purpose of data collection.

Abrader material was introduced into the wind stream through 6 tubes about 10cm above the surface and 1.2m downwind of the start of the working section at a rate of 20 (s.c. [+ or -] 0.25) g/m.s. The abrader material was sieved siliceous sand with a mode particle size of 0.15 mm. The particle-size distribution (PSD) for the abrader and the site soil is shown in Fig. 1. We did not use the site soil as the abrader due to practical problems of drying and sieving the soil to remove material that would not pass through the saltation introduction system. Six tubes were used, to give an even distribution of sediment across the tunnel. Sediment was released at 10 cm height to enable the particles to gain speed in the 0.8 m before they hit the test section. The soil at the study site would be expected to erode at 100 g/m.s (WEAM estimate) if it was totally dry and bare and the u* was 1 m/s. We used a sediment transport rate of 20% of the maximum likely rate because at sowing we would expect higher soil moisture levels and some residual cover from the previous crop or pasture. With soil moisture of 2% and cover of 18%, WEAM indicated a sediment transport rate of 19 g/m.s, which is close to the 20 g/m.s rate selected for the experiment.

Measurements of sediment transport rate, both introduced and originating from the soil surface, were made 6.2m downwind of the start of the working section and 30 cm from the right side (looking down-wind) of the tunnel with a modified Bagnold type trap (Shao et al. 1993). The horizontal sediment transport rate (or the vertically integrated streamwise flux) is defined as:


where Q is the sediment transport rate (g/m.s), and q is the streamwise sediment flux (g/[m.sup.2].s) at height z(m). The quantity q(z) is defined as:


where m is mass of sediment collected (g) over time ([DELTA]t = 60 s) in a trap with orifice width ([DELTA]y = 0.005 m) and at height ([DELTA]z=l)m. Treatments are described in terms of the total transport mass (TTM), which is the time integrated horizontal sediment transport rate (kg/m):



The treatment application commenced on 6 June 2000, 23 days after sowing (DAS) and was completed on 29 June 2000, 46 DAS. The treatment variable, TTM (kg/m), varied by increasing the tunnel run time while maintaining a constant rate of introduced abrader. All treatment runs were performed with the tunnel aligned parallel to the plant rows to reduce the influence of surface roughness caused by the sowing ridges. Run times were 0 (control), 30, 60, 120, and 180min at a constant windspeed. In addition, 2 treatments were included to determine the impact of wind alone on crop damage and grain yield. These treatments required the securing of the soil surface to prevent any wind erosion. A polymer emulsion soil stabiliser, 225.[Curos.sup.R], was applied in solution containing 60 g of225.[Curos.sup.R] at a rate of 2 L/[m.sup.2]. One treatment was 120 min of wind alone (with no wind erosion) and the second a control of 225.[Curos.sup.R] alone to see if the 225.[Curos.sup.R] affected plant growth (Table 1). These 7 treatments were randomly allocated to plots in 6 replicated blocks.

Background meteorological and soil moisture conditions

The plants were grown under field conditions with the meteorological conditions monitored throughout the growing season by an Automatic Weather Station recording temperature, rainfall, windspeed, wind direction, and solar radiation, and soil moisture was measured with CS615 Water Content Reflectometer probes inserted to average over 0--0.15 m depth. The probes were located with 3 replicates in Treatment 7 plots (the most damaged plots) and 3 in Treatment 1 plots (control). Soil moisture at the time of treatment of the top 0.02m at each plot was also determined gravimetrically. Longer term meteorological data with 15-min observation records were obtained from a Bureau of Meteorology Automatic Weather Station positioned at the Strathalbyn Race Course, about 5 km from the trial site, which commenced recording on 14 May 1996.

Crop response

Crop measurements made at the time of treatment included a rapid assessment of the plant damage made 2-4 days after treatment using a 6-stage damage index. Individual plants were ranked at 8 transverse transects along the longitudinal length of the treated area to obtain an average damage ranking. The assessment was carried out by the same person throughout the trial to minimise observer error. The damage index was based on macro-scale damage symptoms arising from the wind and sediment transport over the plants and ranges from healthy undamaged plants through to dead or likely to die plants as follows:

* Stage 1, healthy plant with minor blemishes and some leaf tip withering.

* Stage 2, some burning of the emerging leaves and a thin sliver of bruised tissue along the windward side of the cotyledon.

* Stage 3, immature emerging leaves are wilted and bruised immediately following the event, becoming shrivelled and dark after 24 h. Mature leaves show less damage, usually withering at the leaflet tips. Bruising on the edge of the cotyledon pronounced.

* Stage 4, leaves at all stages of development are dark and shrivelled but the petioles and lower part of the leaf may still be green and upright. Bruising of the cotyledon is extensive and scoring on the windward side of the stem below the attachment of the cotyledon appears as a brown discoloration.

* Stage 5, all the leaves and the majority of the petioles are gone. Scoring of the stems above and below the cotyledon is apparent. The meristems are still intact and capable of shooting again. Potentially the plant can continue to grow.

* Stage 6, the plant is incapable of growing and will completely brown off within days. The soil may be eroded away from the roots of the plant and the plant fallen on its side.

Plant number and phenological stage was recorded in each subplot before treatment using the method described by Dracup and Kirby (1996). At maturity each 1-[m.sup.2] subplot was harvested and the dry weight (DW), grain yield (GW), pod number (PN), and pod weight (PW) were recorded.



Wind velocities for each treatment measured at 0.6 m height in the tunnel (free stream, [u.sub.0.6]) are provided in Table 2 and show an average windspeed of 13.7 m/s, with no significant difference between treatments. The results for u* and [z.sub.o] show that Treatment 2 had higher values than other treatments applied with the wind tunnel. We believe that 225.[Curos.sup.R] prevents smoothing of the surface during the erosion process and therefore maintains surface roughness.

Time-integrated horizontal sediment transport rate and average sediment transport rate

The TTM (with units kg/m) is the sum of the total mass of introduced abrader (TT[]) and the total mass of the sediment derived from the bed, or floor of the tunnel (TT[M.sub.bed]). TT[M.sub.bed] is a function of run time, soil moisture, and possibly plant size. TTM and TT[] were measured, from which we calculated

TT[M.sub.bed]. The mean TTM separated into TT[] and TT[M.sub.bed] are given in Fig. 2. The TT[M.sub.bed] represented about 10% of TTM. The small amount of TT[M.sub.bed] measured from Treatment 2 after 120 min of wind tunnel run time shows the effectiveness of the 255.[Curos.sup.R] Soil Stabiliser in securing the soil surface. The average sediment transport rate (Q) for Treatments 4-7 was 22.37 g/m.s (s.e. [+ or -] 0.25) and had <5% deviation in Q from the mean. Thus, the aim of having a relatively constant abrasion rate was achieved. Analysis of variance (ANOVA) of Q showed that there was no significant difference (P > 0.05) across the treatments. Consequently, in the following analysis, differences in plant damage and subsequent harvest parameters were attributed to the TTM (kg/m) for each treatment rather than variation in Q.

Phenology and plant density

Wind tunnel assessments were undertaken over 23 days, resulting in a change in phenology (determined by leaf number). The average leaf number was 3.4 leaves in the first replicate and 9.7 leaves in the last. ANOVA of leaf number showed that phenological stage was not significantly different for treatment but was significantly different for each replicate (at P = 0.05). Average leaf number for each replicate increased in the chronological order in which they were treated (Fig. 3).

Variation in plant density between plots occurred across the trial site. Some subplots had very low plant densities due to localised levels of poor seed emergence. Subplots with <15plants/[m.sup.2] were removed from the dataset along with all related harvest and damage data. Removal of these subplots resulted in treatment means for plant density (plants/[m.sup.2]) of 37.8, 34.6, 33.3, 36.0, 33.5, 30.9, and 30.7 for Treatments 1-7 (l.s.d. (P= 0.05)= 6.1). ANOVA for plant density showed no significant difference between treatment means (P > 0.05).


The mean damage ranking scores across all replicates for each treatment is plotted in Fig. 4. ANOVA showed a significant difference in mean damage ranking between the different treatments: the treatments with no abrader (Treatments 1, 2, and 3) were significantly lower than all treatments with abrader; Treatment 7 was significantly higher than all other treatment means; Treatments 4-7 were all significantly different from each other, with the damage at each of these treatments increasing from Treatment 4 to 7.


Figure 4 shows that the mean damage rating for Treatment 7 reached a ranking value of 4, which indicated that the majority of plants treated at this level of TTM experienced extreme macro-damage to the immature and mature leaves and stem scoring. Despite this, they could be expected to have the ability to recover from the meristem and continue growth. There was a range of damage recorded at this level, with individual plants in a sheltered position in the canopy showing a very low level of damage and others in a more exposedposition being killed. There was also a low level of damage evident on the control plants as a consequence of a naturally occurring sandblasting event that occurred immediately before the fieldwork commenced. The TTM occurring during this event was estimated at 8 kg/m with WEAM utilising 15-min AWS windspeed observations from Strathalbyn, soil moisture of 3% (0-0.05 m), and cover of 5% (Shao et al. 1996).

Harvest results

ANOVA of the harvest results for Treatments 1-7 showed a significant difference between treatments (P < 0.05). Figure 5 illustrates the effects of treatments on average GW, DW, and PN. Significant differences between the different treatment means occurred for all of these parameters. Treatment 2 was significantly higher than all other treatment means and Treatment 7 significantly lower than Treatments 1-5.


Treatments 2 and 3 aimed to test the effect of wind alone on the plant growth and grain yield by securing the soil surface to prevent sediment transport from the bed and not introducing abrader. Comparison of these treatments indicated that there was a significant increase in DW, PN, and GW for the 120-min wind treatment on the plot sprayed with 255.[Curos.sup.R] (Treatment 2) over a control plot sprayed with 255.[Curos.sup.R] but not subjected to the wind (Treatment 3). This increase cannot be explained.

The negative impact of wind can be important, with physical injuries to leaves of different species reported in a comprehensive reviews on mechanical damage by wind (Miller et al. 1995; Cleugh etal. 1998) This experiment showed that for the conditions applied, wind alone did not reduce plant productivity. This was also the finding of an unpublished trial by H. Cleugh, M. R. Bennell, and J. F. Leys, in which fababean cv. Fiord (Vicia faba) plants at a mean height of 150mm (pre-flowering) and 400 mm (flowering) were treated within the same wind tunnel for 8 h at a similar windspeed to that used in this experiment. This treatment resulted in minor macro damage symptoms of abrasion and leaf margin fraying with no measurable impact on plant development and productivity (Table 3). Subsequently, analysis of the experimental results was carried out with Treatments 2 and 3 removed from the dataset, and only those treatments without 255.[Curos.sup.R] to bind the soil surface were considered in the analysis.


Technical problems with data logging resulted in the climate and soil records for the crop-growing season being incomplete. The onsite AWS broke down early in the season and the records from the start of the trial to Julian Day 250 were obtained from the AWS station located at the Strathalbyn Race Course. Onsite recording resumed on Day 263. Soil moisture readings were also intermittent, with those obtained being similar for the control (Treatment 1) and most damaged plot (Treatment 7) throughout the growing season. Some divergence was noted late in the crop season after leaf shedding, with Treatment 7 showing lower soil moisture than the control. The reason is unknown but may be related to greater evaporation from the soil surface in these plots where crop dry weight and, consequently, cover was lower. Figure 6 describes the period in which the treatments were applied in more detail and includes the values for soil moisture in the 0-25 mm zone. Regular rainfall occurred during the experiment and soil moisture limitations during the growing period of the crop were minimal.




Analysis of the sub-diurnal AWS data supports the use of the wind-speed treatment selected for this trial. Windspeeds exceeding 13 m/s (at 2 m) were recorded on 18 observations over the 5-year period or on 1.5% of total observations, demonstrating that severe wind events are not an unusual event in this area. The magnitudes of TTM treatments used in this trial are supported by an estimation of TTM occurring during a natural sandblasting event on the 18 July 1999. This event caused severe plant damage to canola, lupin, and cereal crops and resulted in yield losses of 5% for a wheat crop and 10% for canola. Using WEAM, an estimate of TTM on this occasion was about 260kg/m, which correlates with the level causing severe crop damage in this experiment.

Sediment transport rate

Although the sediment transport rate, Q (g/m.s), was not significantly different between treatments, a marked variation was measured between replicates. Sediment transport rate (Q) from Replicate 4 was significantly higher than that from most of the other 5 replicates. Minor fluctuations in the [] occurred during the trial (Table 4); however, these differences were not statistically significant (P > 0.05) between replicates or treatments. The difference between replicates arose from variation in the rate of sediment transport from the bed (Fig. 7). Several factors including increasing windspeed, soil moisture, cover (plant size), or soil properties may account for this variation in rate. Replicate 4 treatments (Julian Days 166 and 167) were conducted at a time of low rainfall, and soil moisture content in the surface layer (0-25 mm) was <5% (Fig. 6) and these conditions would lead to a larger bed transport mass, [TTM.sub.bed].


It is presumed that the kinetic energy of the abrader would most likely decrease down the plot; however, the position of the 1-[m.sup.2] subplot has no effect on the degree of damage, indicating a relatively even abrasion of the treatment area. It is therefore suggested that factors of plant orientation, densities, and early growth stages used in this experiment did not appear to influence the damage levels over the 4-m plot length. There is a strong linear relationship (Fig. 4) between damage index and increasing tunnel run time (0-180 min), with a maximum average damage rating of about 4 at which there was a measurable yield reduction. This rating reflected severe plant damage with most of the leaf structure destroyed and clearly visible stem scoring; it does, however, leave the growing points relatively undamaged and able to resume vegetative growth. Treatment run times of 30 and 60 min (TTM of 42 and 78kg/m, respectively), while causing damage with an average ranking of 2.6 for the 60-min treatment, do not cause a reduction in plant productivity. The plants successfully recovered from this degree of damage and reached a yield defined by other growth-limiting factors. Damage symptoms at this level include burning of the emerging, immature leaves, showing a degree of wilting and bruising immediately following the event, becoming shrivelled and dark after 24 h, and a thin sliver of bruised tissue along the windward side of the cotyledon. Mature leaves show little macro-damage, perhaps withering at the leaflet tips.

Phenology and plant density

Phenological stage varied over time throughout the experiment with a significant increase in leaf number with each successive replicate. ANOVA of damage with leaf number as a covariate showed that leaf number had a significant effect on the relationship between the treatment and the mean damage rating (P < 0.05). Statistically, this effect is mostly accounted for at the replicate level with reduced damage for later replicates.

Crop development will have 2 general effects on the sandblasting process. Firstly, the increase in LAI (cover) will reduce the sediment transport produced from the soil surface, reducing the risk of crop damage. This effect was not noted to occur at a significant level during this trial but will be important in a field situation when the risk of sandblasting damage is to be considered throughout the crop cycle. Secondly, as the crop develops the susceptibility of the plant to sandblasting damage will decrease. Sandblasting damage to soybean, winter wheat, and sorghum was shown to decrease with plant age (Armbrust 1984). Exposure 7-14 days after emergence reduced the dry weight in all 3 crops more than when plants were younger or older. Armbrust (1984) ascribed this result to the depletion of energy reserves in the seed and where the plant had become dependent on energy from photosynthesis. Loss of leaf material at this stage places an additional burden on the plant's energy supply as energy is needed to repair the damage.

Agronomic trials show that plant density can have a significant effect on grain yield. French et al. (1994) examined the impact of plant density on yield for several cultivars of narrow-leaf lupin and determined a relationship, which described response across low, moderate, and high yield potential sites, with the yield potential at this site fitting into the high yield potential relationship. Herbert (1978), working in New Zealand with the cultivar 'Unicrop' developed a relationship (s.e. = 35) between plant density (D, plants/[m.sup.2]) and crop yield (g/[m.sup.2]) for a high-yielding site described by:

Crop yield = 201.45 + 5.85D-0.02[D.sup.2] (5)

An ANOVA of the average plant number recorded after each sandblasting treatment does not show a significant difference between treatments. There is a trend of decreasing plant number with increasing tunnel run time (i.e. sandblasting). This is probably because as sandblasting "increased the proportion of plants dying and being blown off, the plot increased.

Statistical analyses by ANOVA of grain weight, where plant number was included as a covariate, found that plant number does not account for a significant proportion of the variation between treatments. Consequently, plant density, although likely to be causing minor differences between treatments, does not significantly affect dry matter and grain yield results. The reduction in final yield is a consequence of physical damage of leaf loss and stem damage and the physiological effects of this damage on the surviving plants rather than a reduction in plant density. This follows the findings of Armbrust (1974, 1982) where reduced growth of winter wheat and sorghum after sandblasting treatment was related to loss of viable leaf tissue and physiological changes that reduced photosynthesis and increased respiration.

Sandblasting effects on yield

The effect of sandblasting on the harvest results, average GW and DW, are significant, with an inverse relationship between production and increasing sandblasting (Fig. 8). Plotting these parameters against the average TTM for each sandblasting treatment shows a decline in DW and GW with increasing TTM and related plant damage. This represents an average decrease in crop yield of 18% as a consequence of sandblasting the crop with a TTM of 248 kg/m. Since soil water status after the treatments was adequate, the measured yield reductions were likely to be near the minimum expected.


As there is no earlier work on field grown legume crops, a direct comparison with prior work cannot be made. However, the impact of sandblasting treatments on crop yield of several other field crops has been undertaken (Woodruff 1956; Skidmore 1966). Woodruff(1956) undertook a laboratory-based experiment with potted plants of winter wheat treated in spring and autumn using a TTM ranging from 8 to 220 kg/m, but applied over much shorter duration (maximum of 20 min), giving a sediment transport rate (Q) of 183 g/m.s v. about 22.37 g/m.s for this experiment. The yield of spring-treated plants was 46.4% less than those treated in autumn, whereas the autumn-treated plants show little final response at harvest to the treatments. The paper did not provide the growth stage of the plants at the time of treatment but the spring treatment occurred at around 180 DAS, so the plants must have been advanced. The usefulness of this information is questionable in its application to a field situation, as the degree of crop cover usually present at this stage of crop development would reduce the risk of erosion and saltation occurring to a negligible level. In addition, the methodology uses a single row of plants positioned normal to the flow of sediment transport, which eliminates self sheltering of plants downwind of the saltation source, as would occur in a normal canopy situation in a field crop. Finally, the sediment transport rates (Q) are extremely high and this could be expected to result in an exaggeration in response to the applied treatments. The author reported greater yield reduction in spring than autumn, apparently due to burying and subsequent sheltering of the leaves from ongoing damage in the autumn treatments. This is a likely result of the very high rate of sediment transport. Consequently, a greater response occurred in autumn at the lower rates of sediment transport where there was less sand deposition and burying. As a result, the work does not provide much insight into plant response to damage during the establishment phase of crop development.

Skidmore (1966) treated container-grown plants of green bean (Phaseolus vulgaris L.) at combinations of windspeed (5.6, 8.3, and 11.1 m/s at 300mm above the tunnel floor) and TTM of 0, 3.4, 6.8, and 10.4 kg/m (at a Q of 112g/m.s); 6.7, 13.5, and 20.2kg/m (at a rate of 224g/m.s) and 10.1, 20.2, and 30.3 kg/m (at a rate of 336 g/m.s). At a low windspeed of 5.6 m/s and 5 min of exposure, some damage occurred but yield reduction was minimal. At higher wind velocities and transport mass, damage increased in a linear trend resulting in a maximum yield reduction of approximately 50%. Our study used similar windspeeds ([u.sub.0.6] = 13.7 m/s) but a lower sediment transport rate (22.37 g/m.s) with a longer duration. Clearly, the green bean is much more susceptible to sandblasting damage than narrow-leaf lupin, with severe damage and yield reduction occurring at lower level of total transport mass than in our experiment. This is not unexpected as beans have large fleshy leaves and are taller than lupin at a similar developmental stage. Also, the growth cycle is shorter, possibly providing less time for recovery from the damage event. In addition, the methodology of the Skidmore's trial is similar to Woodruff (1956) where a single row of plants were placed in the saltation stream, which may have exaggerated the result when compared with a field population where self sheltering of plants would occur.


Narrow-leaf lupin (Lupinus angustifolius) is a major dryland crop species in Australia and is susceptible to sandblasting damage. Observations of physical damage to the crop indicated that at the highest levels of sandblasting the majority of plants experienced extreme macro-damage to the immature and mature leaves and stem scoring. Despite this, they were able to recover from the meristem and continue growth but with a yield penalty.

The yield response of the lupin to total transport mass (TTM) is curve-linear (Fig. 8) under strong wind conditions (21.1 m/s at 10 m height). Lupins can suffer moderate levels of sandblasting (TTM of 71 kg/m) at this wind speed with only negligible yield reductions (1%). Above TTM levels of 138kg/m, yield losses >5% were measured. A yield loss of 18% would be expected with a TTM of 248 kg/m when the plants were still immature with average leaf number to 9.7 (<46 days after sowing). For the soil type on which this crop was grown, application of WEAM shows that sediment transport rate of this magnitude requires windspeeds of at least 13m/s, which occur 1.5% of the time.

The mechanism was conclusively shown to be damage caused by impaction of soil particles, not wind alone or the effect of reduced plant density arising from plant death following the sandblasting event. Also, it was found that younger plants were more susceptible to damage than older plants over the range of average leaf number per plant of 3.4-9.7. Erosion control methods, such as windbreaks, strip cropping with faster growing species, and stubble retention farming, which reduce sediment transport rate onto and within lupin crops obviously have yield benefits in wind erosion prone areas during the crop establishment phase.


We thank the Joint Venture Agroforestry Program of the Rural Industries Research and Development Corporation for their financial support of this work. We are also grateful to Debra Partington for statistical analysis, and Rob Murphy and Kim Tomkinson for technical and field support during this project.


Armbrust DV (1974) Physiological response to wind and sandblast damaged winter wheat plants. Agronomy Journal 66, 421-423.

Armbrust DV (1982) Physiological response to wind and sandblast damage by grain sorghum plants. Agronomy Journal 74, 133-135.

Armbrust DV (1984) Wind and sandblast injury to field crops: effect of plant age. Agronomy Journal 76, 991-993.

Cleugh HA, Miller JM, Bohm M (1998) Direct mechanical effects of wind on crops. Agroforestry Systems 41, 85-112. doi: 10.1023/ A: 1006067721039

Dracup M, Kirby EJM (1996) 'Lupin development guide.' (University of Western Australia Press: Nedlands, W. Aust.)

Downes JD, Fryrear DW, Wilson RL, Sabota CM (1977) Influence of wind erosion on growing plants. Transactions of the American Society of Agricultural Engineers 20, 885-889.

Folk RL, Ward WC (1957) Brazos River Bar: A study in the significance of grain-size parameters. Journal of Sedimentary Petrology 27, 3-26.

French RJ, McCarthy K, Smart WL (1994) Optimum plant population densities for lupin (Lupinus angustifolius L.) in the wheatbelt of Western Australia. Australian Journal of Experimental Agriculture 34, 491-497. doi: 10.1071/EA9940491

Fryrear DW, Armbrust DV, Downes JD 0975) Plant response to winderosion damage. In 'Proceedings of the 30th Annual Meeting of Soil Conservation Society of America (Land Use: Food and Living)'. 10-13 August 1975, San Antonio, Texas, pp. 144-146.

Fryrear DW, Stubbendieck J, McCully WG (1973) Grass seedling response to wind and windblown sand. Crop Science 13, 622-25.

Greenwood EAN, Farrington P, Beresford JD (1975) Characteristics of the canopy, root system and grain yield of a crop of Lupinus angustifolius cv. Unicrop. Australian Journal of Agricultural Research 26, 497-510. doi: 10.1071/AR9750497

Greig JK, Bokhari N, Armbrust DV, Anderson LC (1974) Residual effects of wind and sandblasting damage on tomato plants at different stages of development. Journal of the American Society for Horticultural Science 99, 530-534.

Herbert SJ (1978) Plant density and irrigation studies on lupins. 3. Seed-yield relationships of Lupinus angustifolius cv. 'Unicrop'. New Zealand Journal of Agricultural Research 21, 483-489.

Isbell RF, McDonald WS, Ashton LJ (1997) 'Concepts and rationale of the Australian soil classification.' (CSIRO Publishing: Melbourne, Vic.)

Leys JF, Raupach MR (1991) Soil flux measurements using a portable wind erosion tunnel Australian Journal of Soil Research 29, 533-552. doi: 10.1071/SR9910533

Lines RW (1992) The electrical sensIng zone method (The Coulter Principle). In 'Particle size analysis'. (Eds NG Stanley-Wood, RW Lines) pp. 350-373. (The Royal Society of Chemistry: Wiltshire, UK)

Lyles L, Woodruff NP (1960) Abrasive action of wind-blown soil on plant seedlings. Agronomy Journal 63, 533-536.

Marsh BB, Carter D (1983) Wind erosion. Western Australian Journal of Agriculture 2, 54-57.

McTainsh GH, Lynch AW, Hales R (1997) Particle-size analysis of aeolian dusts soils and sediments in very small quantities using a Coulter Multisizer. Earth Surface Processes and Landform 22, 1207-1216.doi: 10.1002/(SICI) 1096-9837(199724)22:13 < 1207::AIDESP820>3.0.CO;2-K

Michels K, Armbrust DV, Allison BE, Sivakumar MV (1995) Wind and windblown sand damage to pearl millet. Agronomy Journal 87, 620-626.

Miller BV, Lines RW (1988) Recent advances in particle size measurement: a critical review. CRC Critical Reviews in Analytical Chemistry 20, 75-116.

Miller JM, Bohm M, Cleugh HA (1995) Direct mechanical effect of wind on selected crops: a review. Technical Report No. 67, CSIRO Centre for Environmental Mechanics.

Moore G (1998) Soil guide: A handbook for understanding and managing agricultural soils. Agriculture Western Australia, Bulletin No. 4343.

Nelson P, Hawthorne WA, Knight R (2000) Development of lupins as a crop in Australia. In 'Linking research and marketing opportunities for pulses in the 21st Century. Proceedings of the 3rd International Food Legumes Research Conference'. Adelaide, SA, 22-26 September 1997. pp. 549-559. (Kluwer Academic Publishers: The Netherlands)

Perry MW, Delane RJ, Tennant D, Hamblin AP (1986) The growth and grain yield of lupins in relation to the soil water balance. In 'Proceedings of the Fourth International Lupin Conference'. Geraldton, Western Australia. pp. 112-118. (Western Australian Department of Agriculture: South Perth, W. Aust.)

Raupach MR, Leys JF (1990) Aerodynamics of a portable wind erosion tunnel for measuring soil erodibility by wind. Australian Journal of Soil Research 28, 177-191. doi: 10.1071/SR9900177

Shao Y, McTainsh GH, Leys JF, Raupach MR (1993) Efficiencies of sediment samplers for wind erosion measurement. Australian Journal of Soil Research 31,519-532. doi: 10.1071/SR9930519

Shao Y, Raupach MR, Leys JF (1996) A model for predicting aeolian sand drift and dust entrainment on scales from paddock to region. Australian Journal of Soil Research 34, 309-342. doi: 10.1071/SR9960309

Skidmore EL (1966) Wind and sandblast injury to seedling green beans. Agronomy Journal 58, 311-315.

Woodruff NP (1956) Wind blown soil abrasive Injuries to winter wheat plants. Agronomy Journal 48, 499-504.

M.R. Bennell (A,D,), J.F. Leys (B), and H. A. Cleugh (C)

(A) Corresponding author: Department of Water, Land and Biodiversity Conservation, GPO Box 2834, Adelaide, SA 5001, Australia. Email:

(B) Department of Natural Resources, PO Box 462, Gunnedah, NSW 2380, Australia.

(C) CSIRO Atmospheric Research, Pye Laboratory, PO Box 1666, Canberra, ACT 2601, Australia.

(D) CRC Plant Based Management of Dryland Salinity, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia.

Manuscript received 16 May 2006, accepted 9 January 2007
Table 1. Sandblasting treatments showing the mean total transport
mass (TTM) for each treatment

Treatment Description TTM

1 Control, no sediment transport or wind, 0
 no soil adhesive
2 Plot sprayed with soil adhesive; wind for 0
 120 min, no sand sediment transport
3 Plot spayed with soil adhesive; no wind or 0
 sediment transport
4 Tunnel run time of 30 min 42
5 Tunnel run time of 60 min 78
6 Tunnel run time of 120 min 153
7 Tunnel run time of 180 min 248

Table 2. Mean treatment: tunnel free stream wind velocity at 0.6 m
([U.sub.0.6]) and calculated outdoor wind speed at 10 m ([U.sub.10])
based on tunnel friction velocity (u*) and surface roughness length

Treatment [U.sub.0.6] [U.sub.10] u* [z.sub.o]
 (m/s) (m/s) (m/s) (mm)

2 13.8 22.6 1.05 2.11
4 13.7 21.1 0.90 1.21
5 13.9 22.0 0.98 1.57
6 13.7 21.2 0.92 1.06
7 13.7 21.1 0.89 0.94
l.s.d. (P = 0.05) 0.29 0.112 0.94

Table 3. Dry weights (DW) and grain weights (GW) (g/plant) of
fababean cv. Fiord plants in the direct damage experiment contrasting
response after an 8 h, 13.7 m/s (free stream, [U.sub.0.6]) windspeed
treatment with a control

Treatment (A) Days after treatment:

 18 33 46 83 (maturity)

13.7 m/s, 8 h 5.3 25.1 50.9 62.6 32.8
Control 5.4 25.1 43.1 73.5 33.1
Critical value 2.6 7.8 18.5 28.9 9.3
 for comparison

(A) Average plant height 150 mm.

Table 4. Introduced total transport mass of abrader
(TT[]) for each treatment showing mean and mean
standard error across the 6 replicates

Treatment Mean (kg/m) s.e.m.

1 0.00 0.00
2 0.00 0.00
3 0.00 0.00
4 36.41 0.71
5 74.76 0.51
6 144.6 1.85
7 214.9 2.99

Fig. 2. Mean total transport mass (TTM, kg/m) and its components of
total transport mass of abrader (TT[]) (solid bar) and the
calculated total transport mass from the bed (TT[M.sub.bed]) (hatched
bar) for each treatment. The very low TTM for the Curos[R] treatment
(2) after 120min of wind treatment shows its effectiveness in reducing
wind erosion.


1 0.0 0.0
2 1.9 0.0
3 0.0 0.0
4 5.6 36.4
5 3.7 74.8
6 8.8 144.6
7 34.0 214.9

Note: Table made from bar graph.

Fig. 3. Number of leaves for each replicate, showing the slow growth in
the plants over the duration (23 days) of the experiment.

l.s.d. (P = 0.05) = 0.6


1 3.4
2 4.5
3 5.5
4 7.3
5 8.2
6 9.7

Note: Table made from bar graph.

Fig. 7. Variation in sediment transport rate from the bed for each

1 1.3
2 2.0
3 3.0
4 6.3
5 0.0
6 0.8

Note: Table made from bar graph.
COPYRIGHT 2007 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bennell, M.R.; Leys, J.F.; Cleugh, H.A.
Publication:Australian Journal of Soil Research
Geographic Code:8AUST
Date:Mar 1, 2007
Previous Article:A specialised soil corer for sampling tree roots.
Next Article:Indexing soil quality: a new paradigm in soil science research.

Related Articles
Turnover of nitrogen from components of lupin stubble to wheat in sandy soil.
Finding non-fish diets for prawns: Wendy Pyper finds that meat and lupin meals are showing promise as replacements for fish-based feeds in the...

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters