Health and Characteristics of Australian Apple Growing Soils

,


Introduction
Apple production in Australia is valued at nearly $620 million making it one of the highest value fruit industries in Australia [2].Despite the economic importance of the Australian apple industry, little is known about the condition and characteristics of Australia's apple growing soils.Mapping and characterisation of Australia's apple growing soils are scarce.Te limited mapping which exists was mostly conducted at the 1 : 10000 to 1 : 1000000 scale which is not sufcient for farm management or reporting soil condition [3].Intensive soil mapping and characterisation have only been conducted in two apple growing regions, the Huon Valley in Tasmania [4] and the Shepparton Irrigation District in Victoria [5][6][7][8][9][10][11][12].
Knowledge of Australia's apple growing soils and their condition is needed for the Australian apple industry to report on soil health, as a baseline to enable changes in soil condition to be monitored over time, and to provide data required for the development of perennial tree crop-soilclimate models.Te "health" or "quality" of Australian apple growing soils has not previously been reported, nor the soil types or properties of soils in each of Australia's apple growing regions.Assessing soil health or soil quality is not straightforward, as there are no universally agreed parameters, protocols, or values for quantitatively determining soil health [13][14][15], much less agreed values for orchard soils.Difculty reaching agreement is in part due to the complexity of soil systems, abundance of potential soil indicators [16], and lack of data to support linking specifc soil indicators and thresholds to diferent soil types, and the purpose to which the soil is used [15,17,18].For example, Gatica-Saavedra et al. [16] identifed 342 potential indicators of forest soil health, which they coalesced into 62 chemical, 28 biological, and 32 physical indicators.A number of approaches have been developed to quantify the health of cropped soils, for which threshold values for diferent soil indicators have been defned [22][23][24][25][26][27].However, similar approaches have not been developed for orchard soils.Whilst DuPont et al. [28] described orchard soil health as the capacity of soil to support productive trees without negatively afecting the surrounding environment, specifc indicators of orchard soil health or threshold values of what constitutes "good" and "bad" orchard soil health have not been identifed.
Over the last two decades, biophysical models have been developed to support strategic and tactical decision making in most agricultural industries [29][30][31][32], for example, APSIM [33], CERES [34], SWAP [35], and DSSAT [36].Development of the SPASMO (Soil Plant Atmosphere System Model) perennial tree crop model and its user interface SINATA [37] has been hampered by the lack of soil data from Australian apple growing regions.Unlike most soil water models which have relatively simple tipping bucket soil-water modules, the SPASMO model requires knowledge of the soil water retention curve described by the van Genuchten model and saturated hydraulic conductivity for each soil layer [38][39][40], which is not available in any Australian soil databases.
Consequently, this study was commissioned to (i) characterise the physical and chemical properties of key apple growing soils in each of the major apple growing regions of Australia, (ii) provide a "snapshot" of current soil condition or soil health, (iii) establish a baseline for future monitoring of soil health, and (iv) provide chemical and physical data required to further advance development of the perennial orchard model SPASMO and its user interface SINATA.

Site Location and Sampling.
Representative apple growing soils were identifed in each of 10 major apple growing regions in Australia.In Victoria, four sites were located in the Yarra Valley, two in Gippsland, and two at Harcourt.In South Australia, nine sites were located in the Adelaide Hills.In New South Wales, four sites were located in both Orange and Batlow, and one in Bilpin.In Tasmania, four sites were located in the Huon Valley, and one site in the Derwent Valley.In Western Australia, one site was located in each of Manjimup, Pemberton, and Kirup (Table 1).
Given the absence of detailed soil mapping in most regions, orchard selection, and sampling sites were selected on the basis of expert opinion, and on-site recognisance by experiences soil scientists, assisted by local industry advisors.Tis ensured that representative soils, orchards, and sites within orchards, were selected for investigation.At each site, a single soil profle was excavated to around 70 cm depth, within the centre of the tree mound, using a mechanical auger "Dingo" with a 600 mm diameter auger.Te soil profle was described according to NCST [41] and classifed according to Isbell [42].Sampling for chemical and physical properties was conducted on a soil horizon rather than depth basis.For each soil horizon, three 250 cm 3 intact cores were extracted from beneath the tree row for analysis of the soil water retention function at each site.In addition, a 300 g composite sample of disturbed soil was obtained from each horizon from beneath the tree row for analysis of soil chemical properties at each site.

Soil Water Retention.
Te soil water retention function was determined using the KuPF apparatus (UGT, Germany; ICT International, Australia) between saturation and − 80 kPa, supplemented with "dry end" retention data determined in triplicate by either WP4C dewpoint potentiometry (METER Group, Inc. USA) or pressure chamber data at − 1500 kPa.Te soil water retention data were ftted for the van Genuchten equation [43] using Excel Solver software according to in which h is matric potential (cm), n and m are the van Genuchten empirical shape parameters, θ s (cm•cm − 1 ) is the saturated water content, θ r (cm•cm − 1 ) is the residual water content, and α (1/cm) is the van Genuchten soil structure parameter.Saturated water content or total porosity was determined following at least 5 days of saturation for topsoils and 14 days for subsoils.Te plant available water content (PAWC) was calculated as the water flled pore space between feld capacity (FC) at − 10 kPa and the permanent wilting point (PWP) at − 1500 kPa [44,45].Te readily available water content (RAW) was determined as the soil moisture held between feld capacity at − 10 kPa and the refll point at − 50 kPa.PAWC and RAW were calculated to 60 cm depth rather than 100 cm depth, due to the use of dwarf root stock in apple orchards.
Macroporosity is related to the soil's ability to quickly drain excess water and facilitate root proliferation.Macroporosity was measured as the proportion of pores larger than 300 μm calculated as the diference in soil moisture between saturation (0 kPa) and − 1.0 kPa [1,48,52].Macroporosity values over 5% are considered optimal, whilst values less than 4% have been associated with compaction [1].
Te soil retention S value is a measure of soil structure derived from the shape of the infection point in the soil water retention curve, measured as the slope of the 2 Applied and Environmental Soil Science gravimetric water content, θ g (kg kg − 1 ), versus the natural logarithm of the pore water tension head, calculated as where n and m are the van Genuchten empirical shape parameters, θ gs (cm cm − 3 ) is the gravimetric saturated water content, θ gr (cm cm − 3 ) is the gravimetric residual water content, and α (hPa − 1 ) is the van Genuchten soil structure parameter.Reynolds et al. [1] suggest that S values ≥ 0.050 represent very good, 0.035 ≤ S < 0.050 represent good, 0.020 ≤ S < 0.035 represent poor, and S < 0.020 represent very poor soil structure or physical quality.

Infltration and Hydraulic Conductivity.
Infltration and saturated hydraulic conductivity were determined at the soil surface in triplicate using the SATURO Dual Head Infltrometer (Meter Group, Inc. USA) operated for approximately 120 minutes.Te Dual Head Infltrometer applies water at two pressure heads, repeated over three cycles, such that the efect of sorptivity and lateral fow can be excluded from the calculation of saturated hydraulic conductivity.
Failure of the Dual Head Infltrometers at several sites (especially the Victorian sites) resulted in measurement of surface soil infltration and hydraulic conductivity by 200 mm diameter, single-ring constant head infltration in which the calculation of saturated hydraulic conductivity was solved according to Reynolds and Elrick [53] assuming an alpha value of 0.12 cm − 1 .Subsoil infltration and hydraulic conductivity were determined by Guelph permeameter [54] and a purpose built thin tube constant head well permeameter, in which saturated hydraulic conductivity was solved by the single head method assuming an alpha value of 0.12 cm − 1 .Measurements were conducted in triplicate at approximately 600 mm depth (B2 horizon) in which a 100-150 mm head was maintained in a 30 mm radius borehole for at least 20 minutes.Saturated hydraulic conductivity was solved using the Guelph permeameter software based on Elrick et al. [55] and Reynolds and Elrick [56,57].

Chemical Analysis.
Bulk soil samples from each soil horizon from Tasmania, Victoria, and Western Australia were sent to CSBP laboratories for analysis, whilst samples

Selection of Soil Health Indicators.
As soil health indicators have not previously been identifed for orchard soils, the soil health indicators adopted in this study are based on the desirable attributes of apple growing soils.Whilst apples grow in a wide variety of soil types and soil conditions, ideal soils for apple production are genrally considered to be, well drained, well aerated or porous, with good water holding capacity, and slightly acid to neutralpH [28,60].Consequently, for the purpose of this study, soil health indicators focused on measures of soil water availability (PAWC, RAW), macroporosity (Macropore %, S value), drainage status (Ksat, AC, BD), and general chemical soil health indicators (pH, CEC, OC), as well as potential barriers to production (ESP, pH, exchangeable Al +3 ).
Treshold values for what constitutes "good" as opposed to "poor" indicator values (Table 2) were largely based on existing thresholds.

Soil Structure and Macroporosity.
Bulk density in the A1 horizons averaged 1.32 g/cm 3 (SD ± 0.16) in which values for individual sites ranged from 0.88 g/cm 3 for a clay loam Red Ferrosol to 1.67 g/cm 3 in a sandy Red Kandosol (Figure 2(a)).Of the 48 A1 horizons, 29 horizons had moderate to high levels of bulk density [51] (Figure 2(a)).Air capacity in the A1 horizons averaged 9.97% (SD ± 5.76), in which 26 of 48 horizons had values below the 10% threshold required for proper aeration and root function [51,69] (Figure 2(b)).In addition, 43 of the 48 A1 horizons had macroporosity (>300 μm) values which were less than the 5% threshold for good soil structure (Figure 2(d)).In contrast to these data which suggest the A1 horizons were poorly structured, the soil water retention S value indicated that 39 of the 48 A1 horizons had good or very good structure (Figure 2(e)), whilst the average saturated hydraulic conductivity of the A1 horizon was 250 mm/hr (SD ± 328) which is considered as being very high (Figure 2(c)) [51,70].
Te A2 horizons had higher average bulk density, lower hydraulic conductivity, and lower macroporosity than the A1 horizons.Te average bulk density for the A2 horizons was 1.43 g/cm 3 (SD ± 0.21) (Figure 2(f )).Te average air capacity of the A2 horizons was 8.02% (Figure 2(g)), which was similar to the A1 horizon.None of the A2 horizons had more than 5% macroporosity (>300 μm) (Figure 2(i)), whilst the average S value was 0.04 in which 7 of 15 A2 horizons were classed as having poor to very poor structure (Figure 2(j)).Te average hydraulic conductivity of the A2 horizons was classed as moderate at 55 mm hr − 1 (SD ± 160) (Figure 2(h)) [51].
Te average bulk density for the B2 horizons was 1.44 g/ cm 3 (SD ± 0.18) (Figure 2(k)) which was similar to the A2 horizon (Figure 2(f)).Te average air capacity and hydraulic conductivity of the B2 horizons were 5.28% (SD ± 3.81) and 17.03 mm/hr (SD ± 82.46), respectively (Figures 2(i) and (m)).Notably, 28 of 48 the B2 horizons were considered to have moderate density, and 9 horizons were considered to have high to extreme density.Only 7 of the 48 the B2 horizons had air capacity values above the 10% threshold required for adequate root function (Figure 2(l)).Te average macroporosity in the B2 horizons was only 1.11% (Figure 2(n)), in which none of the 55 horizons exceeded the optimum macroporosity threshold of 5%.Te average retention curve S value in the B2 horizons was 0.03 in which 10 horizons were classed as having good to very good structure (Figure 2  Applied and Environmental Soil Science Reynolds et al. [1] pH h20 : acidity-alkalinity, EC horizons, high for 20 horizons, moderate for 13 horizons, and low in only one horizon [51].By comparison, the average organic carbon content of the A2 horizons was 0.75% (SD ± 0.55) in which 6 of the 16 of the A2 horizons had lower organic carbon levels than the upper B2 horizons (Figure 3(a),   3. 6 Applied and Environmental Soil Science   Applied and Environmental Soil Science the B2 horizons was low at 0.80% (SD ± 0.87%), in which only 5 of 55 horizons were classifed as having high to very high levels of organic carbon.Overall, CEC was substantially lower than expected, given the high organic carbon and clay contents at most sites (Figure 3(d)).Te CEC averaged 10.10 cmol (+) kg − 1 (SD: 7.01) in the A1 horizons, 2.83 cmol (+) kg − 1 (SD: 2.19) in the A2 horizons, and 8.12 cmol (+) kg − 1 (SD: 5.47) in the B2 horizons (Figure 3(d)).Of the 117 soil horizons, 50 were classifed as having low (6 to 12 cmol (+) kg − 1 ) to very low (<6 cmol (+) kg − 1 ) CEC, and this included 16 of 47 A1 horizons, 13 of 15 A2 horizons, and 21 of 55 B2 horizons (Figure 3(d)).Only 3 of the 118 soil horizons were classifed as having high to very high CEC (>25 cmol (+) kg − 1 ), all at the South Australian Black Dermosol site (Figure 3(d), Table 5).
Apples trees are considered to be moderately sensitive or slightly tolerant to salinity [71,72] in which 50% yield reduction is predicted at a soil salinity of 5 EC e dS/m.Analysis indicated that only 7 horizons had EC values greater than 0.3 dS/m, and only 2 horizons, both at the Hydrosol site, had EC values greater than 1.0 dS/m (Figure 3(b), Table 5).Given a texture multiplier of 9 for a clay loam [62], apple yield is predicted to be severely afected at the Hydrosol site in Victoria, with the possibility of a slight yield reduction at the Brown Sodosol and three Brown Chromosol sites in South Australia.No other sites appear to be afected by salinity.

Applied and Environmental Soil Science
Exchangeable aluminium is considered to be toxic for apple production at levels above 0.4% [67].Exchangeable aluminium was absent (<0.1%) from all but one of the A1 horizons (Figure 3(f )).Only 12 of the 71 tested soil horizons had exchangeable aluminium levels above 0.4%, 6 of which were B21 horizons, and 5 of which were B22 horizons.Tis included 2 Red Ferrosols, 1 Brown Ferrosol, 1 Red Dermosol, and 1 Grey Kurosol, all of which had a pH H20 between 4.5 and 5.4 (Figure 3(f ), Table 5).

Discussion
Plant available water content is a key soil attribute for apple production [28].PAWC and RAW varied greatly within the same soil order, which has implications for irrigation management.Comparing the two South Australian Brown Chromosols, site SA6 had 40 mm RAW, whilst site SA8 had 109 mm RAW to the same depth.Assuming a crop factor of 1.2 [73] and an average peak summer reference evapotranspiration (ET 0 ) of 45 mm per week [74], site SA6 would need to be irrigated every 5 days in summer, whereas site SA8 would need to be irrigated every 14 days in summer.Furthermore, variance between sites within the same soil order means that irrigation scheduling should be guided by soil moisture probes/sensors rather than relying on generic district-based irrigation guidelines or soil type-based irrigation guidelines.
Drainage appeared to be a key soil limitation to production in many apple growing regions including Tasmania, Victoria, and parts of South Australia.Apple production requires good soil drainage [28], yet at least half to twothirds of all sites showed some evidence of poor subsurface drainage, and almost one-third of sites showed some evidence of impaired drainage in the A1 horizons.Evidence for poor drainage was supported by pedological observations (reported elsewhere) which revealed that 23 of the 34 sites had colour mottling due to temporary waterlogging in at least one soil horizon.Notably, six sites had mottling in the A1 or A2 horizons which indicated drainage and aeration in both topsoil and subsoil horizons.
As most apple orchards utilise dwarfng rootstock, the A1 horizon is the most important soil layer for production and management, in which good drainage, including having low soil density and high hydraulic conductivity, is required [28].Being perennial, and thus infrequently cultivated or disturbed, it was expected that the A1 horizons would be very well structured with high carbon contents, high water retention, and high levels of macroporosity.Tis was not the case; whilst the A1 horizons generally had high to very high levels of organic carbon (average: 2.46%, SD ± 1.12), many of the A1 horizons were poorly structured.Values for bulk density were higher than expected (average: 1.32 g/cm 3 , SD ± 0.16); in which 19 of 34 sites had insufcient air capacity to facilitate proper root function, whilst only 4 of the 34 sites had high levels of macroporosity.However, the high Ksat values indicate the macropores that were present must have been highly connected and relatively efcient at transporting water and air within the A1 horizons as evidenced by the generally high values for saturated hydraulic conductivity and S value.
Te "health" and function of the B2 horizons appeared to be largely related to inherent soil properties rather than management practices.Te majority of B2 horizons appeared to be somewhat hostile to root growth and function, due to their low air capacity (average: 5.27%, SD ± 3.81), high bulk density (average: 1.44 g/cm 3 , SD ± 0.19), high proportion of unavailable soil water to 60 cm depth (average: 37.73 mm/100 mm, SD ± 33.22), suboptimal macroporosity of 1.19%, and at some sites (notably the Ferrosols), low pH and high exchangeable aluminium.Notably, subsoil pH was less than 5.5 at 13 of the 34 sites in which 7 sites also had exchangeable aluminium levels above the 0.4% threshold for toxicity.Many subsoil horizons also had surprisingly low CEC (average: 8.12 cmol (+) kg − 1 , SD ± 5.32).In fact, 10 of the 34 sites had low to very low CEC (<6 cmol (+) kg − 1 ) throughout the whole soil profle.Tese sites were not associated with any particular soil order; they included a Hydrosol, a Ferrosol, three Kurosols, a Chromosol, and a Dermosol.Soils with low CEC are more likely to develop defciencies in potassium (K + ) and magnesium (Mg 2 +) and are at risk of nutrient leaching beneath the root zone [75], which is both inefcient for production, and potential cause of environmental harm [76].Sites with low CEC (<6 cmol (+) kg − 1 ) require frequent, small amounts of fertigation to prevent nutrient leaching below the tree root zone.
Recommendations for soil management are limited by the highly variable nature of the data for the diferent soil health indicators; however, despite the high to very high levels of organic carbon in the A1 horizons, soil structure was poorer and more dense, with lower air capacity, RAW, PAWC, and macroporosity than would otherwise be expected given the lack of cultivation and soil disturbance in perennial orchards.Of particular note was the relatively poor status of many of the Ferrosols and Dermosols which were expected to have better soil structure and water retention than the Chromosols and Kurosols.
For growers seeking improved soil structure or soil carbon levels, normal orchard foor management practices including application of composts, mulches, "living mulches," or throwing cuttings from the inter-row onto the tree row [77][78][79] have been shown to increase soil carbon and improve under tree soil characteristics, especially on degraded or sandy soils.However, in soils with A1 horizons that already have moderate to high levels of soil carbon, further increasing soil carbon may prove frustratingly slow as soil carbon levels are likely to approach equilibrium or saturation over time [80].Consequently, application of further organic material is unlikely to greatly increase soil carbon, although small changes in soil structure may still be achievable.Importantly, living mulches and organic residues may confer improvements in soil structure and macroporosity associated with root growth, reduced raindrop impact, and increased soil fauna burrowing [81,82].

Conclusions
Tis study has provided valuable insight into the types and properties of soils used for apple production in Australia.Apple production was reported from a diverse range of soil types including Ferrosols, Chromosols, Kandosols, Hydrosols, Sodosols, Kurosols, and Dermosols.Te data presented in this paper serve as a baseline for future soil condition monitoring for the Australian apple industry, as well as soil data required for use and development of the SPASMO and SINATA perennial tree crop models.
Chemical and physical soil properties were noted to vary greatly both within and between soil orders.Tere is no one soil type which is ideal for apple growing, in which most sites and all soil orders were prone to some form of either physical or chemical soil limitation to production.Whilst topsoils generally had high to very high levels of organic carbon, evidence suggested that the A1 horizons at most sites were moderately to highly compact, poorly aerated, and lacking large macropores.Yet the pores that were present were highly connected and efcient at transporting water and air within the topsoil.Almost all sites had restricted drainage and potential for poor aeration in the subsoil.A small number of sites also demonstrated potential issues with aeration in the A1 horizon as evidenced by poor air capacity and mottling.
Only one site was found to be saline; however, almost one-third of all soil horizons were found to be sodic (ESP > 5), and 10 of the 34 sites had CEC values less than 6 cmol (+) kg − 1 throughout the entire soil profle, indicating potential for leaching of nutrients from the root zone and potential for aluminium toxicity because low pH was inferred at 6 of the 34 sites.
Soil water availability (RAW and PAWC) varied enormously between sites, even within the same soil order or region such that irrigation scheduling needs to be guided by infeld soil moisture probes/sensors, rather than relying on generic district or soil type-based irrigation guidelines.
Management recommendations include use of soil moisture sensors/probes for scheduling irrigation, improved subsoil drainage and mounding of the tree row at planting, and use of living mulches and organic residues to improve soil structure and maintain or improve soil carbon.Recommendations for future study include extending the analysis to all major apple growing regions in Australia, and that analysis be repeated 10 years after the initial study to determine trends in soil condition over time, and further studies to identify improved indicators and threshold values for perennial orchard soils.

CEC:
Cation exchange capacity ESP: Exchangeable sodium percent pH: Acidity-alkalinity EC: Electrical conductivity measured as a 1 : 5 solution ECe: Equivalent electrical conductivity RAW: Readily available soil water PAWC: Plant available water content KuPF: Hydraulic conductivity-matric potential device WP4C: Soil water potential dew point hygrometer AC: Air capacity FC: Field capacity PWP: Permanent wilting point Ɵ gs : van Genuchten gravimetric saturated water content Ɵ gr : van Genuchten gravimetric residual water content α: van Genuchten soil structure parameter S value: Reynolds et al. [1] soil structure parameter.

Table 1 :
[58] location and soil type.South Australia were sent to DPI lab in Lismore, QLD.Electrical conductivity (EC) and pH were measured using a soil to solution ratio of 1 : 5. Acidity was measured in both water and CaCl 2 solution (Rayment and Lyons[58]methods 4A1, 4B3, 3A1).Exchangeable cations were determined following leaching with both an alcohol and glycerol solution (prewash) to remove soluble salts from the soil prior to extraction using 1M ammonium chloride.Exchangeable cation concentrations were determined using inductively couple plasma (ICP) spectroscopy (Rayment and Lyons[58]method 15A2).Exchangeable aluminium was measured for three sites in Tasmania and all sites in South Australia and New South Wales following 1M potassium chloride extraction (Rayment and Lyons[58]method 15G1).Soil organic carbon (SOC) was determined by wet oxidation[59](Rayment and Lyons[58]method 6A1).CEC was calculated as the sum of cations (Ca 2+ , Mg 2+ , Na + , K + ) excluding Al 3+ due to missing data from Victoria.

Table 5 )
. Te average organic carbon content of

Table 5 :
Key soil chemical properties.