Characterisation of the Least Limiting Water Range of a Texture-Contrast Soil
q, È 'c)ç.\
CHARACTERISATION OF THE LEAST LIMITING WATER RANGE OF A TEXTURE- CONTRAST SOIL
Thesis submitted for the degree of
Master of Agricultural Science
ln
The UniversitY of Adelaide
Faculty of Agricultural and Natural Resource Sciences
by
STANLEY RABASHI SEMETSA
Department of Soil and Water January 2000 \Maite Agricuttural Research Institute Glen Osmond, South Australia Thisworkisdedicatedtomylateson,KITOSEANSEMETSA TABLE OF CONTENTS PAGE CHAPTER iv ABSTRACT..... viii STATEMENT... tx ACKNO\ryLEDGEMENTS. x LIST OF' F'IGURES... xlr LIST OF' TABLES... .
I CHAPTER1 : INTRODUCTION 1
3 I.2 Research Questions and Objectives 4 1.3 Structure of the Thesis
CHAPTER2: LITERATURE REVIEIV' """"6 6 2.I Introduction .....'...... " """"""""' temporal variability """"' 6 z.z Definition of soil structure incorporating spatial & 9 2.3 Soil structural quality indices for plant growth""' """" 9 2.3-l Aggregate Characteristics """"".' """"' t2 2.3.2 Bulk density and relative bulk density"""""""' 13 2.3.3 Macroporosity and pore continuity """""" 2.3.4 Plant available water capacity Relevance to Plant Growthl6 2.3.5 Least Limiting water Range (LLWR) and its z.3.sJUpper limit (Wet end)""""" """""'17 18 2.3.5.2lower limit .....'.'. """"" 18 2.3.5.3 Prediction of the LLWR"""' """""' (WRC) 20 2.3.5.4Estimation of the Water Retention Curve (SÃO 24 2.3.5.5 Estimation of the Soil Resistance Curve " the LLWR""""" 25 2.3.5.6 Pedotransfer functions and their use to characterise 26 2.4 Duplex soils and pedotransfer functions 2.4.1 Definition """"')6 soils 2.4.2 Origin, distribution and agricultural use of duplex """"' """"""""27 28 2.5 SummarY """"""""
I CH^PTER3:ESTIMATIONoTLLWRFROMSOILPHYSICAL 30 PROPERTIES ...... """"' 30 3.1 Introduction...... ' "" 31 3.2 Materials and methods """""""' """""""' 31 3.2.1 Location """""' 3.2.2 SamPling """""32 ..33 3.2.2.1Sample identification .. " ' .33 3.2.3 Physical and chemical properties 33 3.2.3.LParticle size distribution"""""" """ 3.2.3.2Organic carbon content....' """""""34 3.2.3.3Carbonate content-...' ""'34 34 3.2.3.4 pH and EC .. 3.2.3.5Water Retention Cvrve(WRQ """"'36 38 3.2.3.6Soi1Resistance Curve (SRC)"""' "" 79 3.2.4 Statistical Methods...... " 4t 3.2.4.|Modelsaccountingforlandscapepositionandsoildepth: 42 3.2.4.2 Models accounting for soil properties: 43 3.2.4.3 Models fromthe literatwe: 44 3.3 Results and Discussion.....""" 44 3.3.1 Soil Properties...... "' 46 3.3.2 Models accounting for landscape position and soil depth...""' 46 3.3.2.1Multiple regression analysis fot WRC 48 3.3.z.2Multiple regression analysis for SRC 50 3.3.3TheinfluenceofsoilpropertiesontheWRCandsRc...'...... 50 3.3.3.1 Influence of soil properties WRCs """""""""" 52 3.3.3.2lnfluence of soil properties SRCs 52 3.3.4 Models proposed in the literature' 3.3.4.1Models for WRCs """""52 53 3.3.4.2Models for SRCs-.. """"' 53 3.3.5 Least Limiting Water Range """""""' 5t
55 3.4.1 Prediction of WRC.... 56 3.4.2 Prediction of SRC..... """"' 3.4.3 Prediction of LLWR """""'56
il LLWR""' 58 CHAPTER 4 SOIL WATER CONTENT VARIATION AND 58 4.1 Introduction...... '...... 60 4.2 Materials and methods ..""""""' 60 4.2.1 Meteorological data.. " 60 4.2.2 Measurement of soil water content 60 4.2.3 Pout.....-..... 61 4.2.4 Statistical analYsis 6r 4.3 Results and Discussion..""""' 6l 4.3.1 DePth eflects 62 4.3.2 Surface soil (0 - 30 cm) effects""" 63 4.3.3 Treatment effects.....' 66 4.3.4 LLWRandPoø.... 66 4.4 Conclusions ......
AND YIELD CHAPTER 5: INFLUENCE OF LLWR ON CROP GROWTII R8SPONS8...... """"" 68 68 5.1 Introduction -....."...'.." """""""' """"""""' 68 ,I 5.2 Materials and Methods""""""' id 68 5.2.1 Treatment Design """"""" il 69 5.2.2 Crop Management """"""' 5.2.3 Agronomic parameters """'70 70 5.2.3.1Establishment counts and dry matter yield 70 5.2.3.3Protein...-. ""'70 5.2.3.4 Pout...... ""'71 5.3 Results and Discussion...... ""' """""""""'71 5.3.1 Agronomic measurements ".'"""" """71 growing season 5.3.2 Crop responses to soil water content during the """"""76 77
80 CHAPTER 6: SUMMARY & GENERAL DISCUSSION .80 6.1 Conclusions .82 for future research 6.2Implications 84 CHAPTER 7: APPENDICE S 93 CHAPTER 8: REFERENCES......
ut k ABSTRACT
This thesis addresses three main questions:
LLWR, of a texture-contrast 1. To what extent does the Least Limiting Water Range,
been conducted on relatively soil change with depth? Most previous studies have with depth in the root uniform soils, where clay content does not vary signiflrcantly water holding zone; the occuffence of clay at depth may provide increased
capacity for use by plants later in the growing season'
the LLIUR during a 2. To what extent does the volumetric water content fall outside between typical mediterranean growing seasoî, Poot? An inverse relationship in moderated climates' but LLWR and Poulhas previously been shown to exist
is strongly seasonal. little is known about this relationship where rainfall performance in a 3. What impact do the LLWR and Pou¡ have on field plant .'I when Itf plant performance is generally thought to improve ìt mediterranean climate? I conducted using LLll/R is large and pou¡ is small, but little work has been in mediterranean cropping patterns designed to maximise water use effîciency
climates
functions were required: the To calculat e the LLWRof the soils used in the study, two volumetric soil water content water retention curves, wRC, (A: ( fl, where d is the
curves, sftc (sà : f(Ø, where and y is the matric potential) and the soil resistance
of these were i) taken Sft is the soil resistance to a cone penetrometer). Examples relevant soil and from published pedotransfer functions and ii) developed from including landscape position' landscape properties collected from undisturbed cores'
¡ density (and of course cþ content, carbonate content, organic carbon content, bulk
lv r soil matric potential)' The volumetric water content and soil resistance as functions of lines: for the published pedotransfer functions included models along the following + dlog 0+e log p' WRC,tog 0=loga+b tog Wand forthesRÇ log SR:logc
e are constants' The functions where p is the soil dry bulk density and a, b, c' d and
models of the following types: developed from soil and landscape properties included log SRi¡*r: C * + P¡+ D¡ for the WRC,\q*r= C + P¡+Di+ Vk+ pband forthe SRÇ
water content and soil + 0* * pt, where hijn and SRy*r are respectively the volumetric
P, the jth depth in the soil resistance corresponding to the ilh position in the landscape' Minor variations to profile, D,Íhekth matric potential, Y andthe fh bulk density, p.
into account the presence the models for the l¿7RCsand SRCs were necessary to take with differing trends' or absence of carbonates in the soil, which seemed to correlate
soil cores were To obtain the data necessary for the wRC and sRÇ undisturbed 'I tlt ,iT cm) at 5 different locations down ! from 6 different depths (between 0 and 80 i collected at Roseworthy Agricultural a topographic transect of a sodic hypercalcic chromosol
ranged from 10 to 37o/o' College, South Australia. Across the landscape, clay contents soil, calcium carbonate organic carbon contents ranged from 0.1 to 10.3 g C kg-t
densities ranged from 1'3 contents ranged from 0 fo 441.29 CaCO3 kg-l soil, and bulk ranging between to I.l g cfrr3. The soil cores were equilibrated at 8 matric potentials
content and cone-penetration resistance --0.001 and -1.5 MPa, and volumetric water
were measured.
of LLzLshowed that no aeration problems were experienced at the wet Í Calculation I of available water at the end, but that high soil strength severely limited the amount ; rather small (all < dry end. Because of high soil strength, all values of LLI{R were r v plant available water in this 0.12 cnr3 cm3; many < 0.07 om' crn3¡ indicating minimal
this was attributed soil. There was a minor trend of increasing LLWRwith depth, and
depth, both of which to an increase in clay content and adecrease in bulk density with of the coincident coincided with an increase in the concentration of CaCO3. Because
established occufïence of these soil properties, no simple relationship could be
between the LLlltR and inherent soil properties'
a typical mediterranean To address the second question (viz. Pou¡ vs. LLÍVR during the growing growing season), the volumetric water content was measured throughout patterns (and thus season as a function of depth for crops having different rooting shallow-rooting, diflererrt patterns of water extraction). Cereal grains, which were ('inter-cropped') with were growïr either alone ('mono-cropped') or else inter-seeded
depths throughout the lucerne, which is deep-rootrng; Pou¡was determined at various .I and [,t to be greater for smaller values of LLÍIR' if profile. values of Poo¡wefe expected I particularly later in this was borne out by the data in a strong curvilinear relationship,
Pou¡ > 0'70 were not the growing season when the soil began to dry out. Values of effect onPout' uncommon in the top 30 cn¡ and both wheat and oats had the same
Poo, on crop performance), To address the third question (viz. impact of LLIIR and grain mass' and dry matter yields at tillering and anthesis were measured, and final
grain protein were measured. Dry matter and grain yields were found to be inter-cropped significantly greater in the mono-cropped cereal treatments than in the water content treatments, and this was due to the higher aveÍage seasonal volumetric related to the in the mono-cfopped treatments. obviously this was strongly
I
! vl with magdtude of P6¡11in the top 30 cm in as much as laxger values of Psú coincided
reduced dry matter and grain yields.
The primary implications from this work afe swnmarised as follows:
o Application of pedotransfer functions to determine the LLÚYR for texture-contrast
(duplex) soils, or indeed any soil whose properties change significantly with
depth, needs to be done with caution.
so that efforts o It is possible in duplex soils for the LLIYR to increase with depth,
to improve subsoil conditions may prove to be highly fruitful interms of water use
efficiency in such soils.
o Further work on calcareous duplex soils is required to determine whether the
to other apparently confounding effect of CaCO¡ on LLIYR is related in any way
measurable soil properties such as clay content or bulk density.
I
r vu STATEMENT
other This work contains no material which has been accepted for the award of any
and, to the best of my degree or diploma in any university or other lrJfüary institution
written by another knowledge and belief, contains no material previously published or
perso& except where due reference has been made in the te>d.
Library, I give consent to this copy of my thesis, when deposited in the universþ
being available for loan and photocopying.
SIGNED: DArE:. t1 ltlz,*
I
vlu ACKNO\MLEDGEMENTS
(Department I would like to thank my supervisors, Dr C. D. Grant, Mr I. T- Grierson of Adelaide), and of Soil and Water, Waite Agricultural Research Institute, University for their helpful Mr H Reimers (Department of Agronomy and Farming Systems) program. discussion and guidance throughout my entire research
with most of the soil I thank Mf c Rivers (Department of soil and water) for helping of pressure chemical analyses. He also provided invaluable help with maintenance would also like to chambers through out the duration of the experimental work' I more especially with thank Mr p Cornelius for the outstanding help in the field, collection of the neutron probe readings'
their help with part of the I am grateful to Ms c Hunt and the Biometrics sA team for laboratory (Department statistical analyses of the data. I would like to thank the ICP analyses of the of Plant Science and Molecular Ecology) for doing the total element 'Water) for their grains. Thanks also to the research group (Department Soil and continuous support through hard times.
Agriculture for financial I am also grateful to the Botswana Government, Ministry of study leave' support to enable me to study in Australia and granting me the
the hard I would like to thank my wife Naledi for her patience and support through times during the studY Period.
lx LIST OF FIGURES
FIGURE PAGE
Z.l Diagrammatic representation of the hierarchical organisation of soil particles and soil structural units ...... '...... I ') ', Influence of air porosity (fui,, crrf atrlc-3 bulk; on relative diffusivity of oxygen in silty clay loam at different bulk densities and
aggregate sizes t9 Relationship between soil strength as measured by a cone-penetrometer
and percentage ofcottontaproots penetrating through cores offour different soils 2t 2.4 Distribution of a) duplex soils and b) sodic duplex soils in Australia ..... 29
3.1 Treatment design of the study area . 35 3.2 Hanging columns used for matric potentials -l to -lOkPa 37 J.J Pressure chambers used for matric potentials a) -33 to -100 kPa and 38 for b) -500 kPa to -1500kPa ...... 3.4 a) Measurement of penetration resistance using a balance & constant-
rate penetrometer, b) Procedure for obtaining penetrometer readings from
one penetratior¡ and c) Protocol for penetrometer measurements ...... 40
3.5 clay content (< 2 prn"%) as a function of depth along a toposequence .. 45
3.6 Organic carbon (g kg-t) as a function of depth along a toposequence '..... 45 3.7 calcium carbonate (g kg-t) as a function of depth along a toposequence .. 46
3.8 Bulk density (g cm3) as a function of depth along a toposequence "'...... 46 55 3.9 LLWR 1cm3crr3; as a function of depth for all positions
4.1 MeanPou¡ as a function of depth for all treatments 62
4.2 Soil water content variability at 20 cm in relation to the LLWR for the
1998 growing season a) for all treatments, and b) for treatments
grouped as either inter-cropped or mono-cropped .. 64
4.3 Soil water content variability at 30 cm in relation to the LLWR lot
the 1998 growing season a) for all treatments, and b) for treatments
x 65 grouped as either intcr-cropped or mono-cropped 67 4.4 Rainfall records for the 1998 growing season 67 4.5 P6y¡ ùîdits variation as a function of LLIIR
stage (p : 0'05)' 72 5.1 Dry matter measurements at tillering : 0'05)" 72 5.2 Dry matter measurements at anthesis stage(p " " 0'05)' 73 5.3 Yield (t/ha) for the different treatments (p: treatments 73 5.4 1000 grain weight (g) for the different (O, W(a)' W(h)' 5.5 Grain protein (%) for mono-cropped (h)+ L 75 inter-cropped treatments: O * L, W(d) + L, W water content' 5.6 Grain yield as a function of average volumetric 75 (cm3 crn3) for the whole growing season " " " and 30 cm' 76 5.7 Grain yield as a function of Po¡ at20 as a functionof Por¡at20 cm " 78 5.8 Dry matter yield at tillering & anthesis a functionof Pou¡ at 30cm 78 5.9 Dry matter yield at tillering & anthesis as
xl LIST OF TABLES PAGE TABLE 23 2.1 Soil water retention models '
WRC 47 3.1 Coefficients for main effects model for position-and.depth 49 3.2 Coeffrcients for /og SR for each level of function for soils 3.3 fuf.rftipt" regressioã analyses of water retention : + logClay + logoC 51 with g without carbonates : log0 C I logty to'p ! curve for soils 3.4 ftfoftipt" regression analyses of soil resistance +logp + + logOC 51 with & without carbonates: log SR : C + log0 logClay function for soils tut rttipt" regression analyses of water retention 3.5 54 with & without carbonates log0: log a + b log V cufve for soils 3.6 n¡"itipr" regression analyses of soil resistance 54 without carbonates: IogSR : log c + d log0 + e logp *ith ¿ 57 J.t LLWR for different position and depth' 30 cm" 63 4.1 Details of Poutfor all treatments at20 cmand " 70 5.1 Seeding rates of croPs
xll CHAPTER 1 : INTRODUCTION 1.1 Introduction
hold and Soil structure has a major influence on the ability of the soil to accept,
development (White' release water and nutrients, to recycle carbon and to support root
managed 1997; Kay, 1998). It is of great importance to devote attention to
These practices can ecosystems, where soil structure is sensitive to land use practices'
Most of these alter structure directly by processes such as tillage and vehicle traffrc'
the sustainable changes are relatively short terrrU and are often reversible. Therefore,
structure of the soil land management systems tlrrt are needed are those that keep the related to crop in an optimal condition over longer periods for a range of processes
production and environmental quality.
our knowledge The structure of the soil varies both temporally and spatially, however,
production has and understanding of this variation from the point of dryland crop
is a need to devise been seriously limited by lack of suitable analytical tools. There properties to soil structt*al indices that could be used widely to relate soil structural the spatial plant growth. These indices should have components that incorporate both
the and temporal variations in the soil structure. They should also characterise (aggregate inherent soil properties, i.e. both the static (mineralogy) and dynamic
properties' stability), as well the influence of management practices on soil structural other, in The influence of each of these indices should be distinguishable from each
have been order to improve the controllable variable (i.e. management). Studies such as bulk carried out to relate some soil structural parameters to plant growth
Håkansson' density and relative bulk density (Busscher et ol., 1987; Carter, 1990;
I (Gardner et a1.,1984), aggtegate 1gg0 and da silva et al.,|gg7),plant available water et al', 1987; Braunack and characteristics (Alexander and Miller, l99l; Logsdon (Dexter' 1988; Carter' Dexter, 1989; Donald et a1.,1987), and pore characteristics as aflected by factors l9s8). However, due to their spatial and temporal variations structural indices is limited (Zhai such as climate and management, their utility as soil Furthermore, in order to be et a1.,1990; van \ùy'esenbeeck and Kachanoski' 1988)' properties that are directly useful, these parameters need incorporate soil physical
related to Plant growth.
called the " Least Limiting Recently the introduction of a measure (described below)
1998), has greatly enhanced water Range,,, LLWR(da Silva et al.,1994;Betz et al.' on soil structure' The our capability to predict the effects of management practices relationship to processes advanlageof this index lies not only in its close functional in its (apparently) highly accurate that are directly related to plant growth, but also soil properties' The prediction based upon routinely measured, readily available "after rapid dtainage has LLzLdescribes the range of soil water contents that exists with water potential' ceased within which limitation to plant growth associated minimal" (da STlva et al', aeration, and mechanical resistance to plant growth is
structurally stable, non-swelling 1997a). The LLllRhas been successfully applied to da Silva et al'' 1997b)' soils in Canada(da Silva et al.,1994; da Silva et al',1997a; et al', 1996b)' and some Australian soils (constantini et al., 1996a; Constantin\
recently some American soils (Bertz et aI',1998)'
in the top 30crn, which of Most of thesc studies have assumed uniform soil conditions growth of deep-rooted species' course is important for shallow-rooted crops and early
2 clegree' but is of less value where soil properties vary with depth to any significant 30cnu which is There is thus a need to characterise the LLWR with depth exceeding
and subsequent important to plant roots in later stages of growth leading to maturity harvest. There is also a further need to characterise the LL\41R in texture-contrast particularly important soils, where soil properties vary in the profile abruptly' This is
the sandy in Duplex soils in Xeric climates where water is the limiting factor in plant survival surface horizons, and where access to subsoil water often determines and maturity as the surface soil dries out'
relationships To determine or predict the LLWR for a given soil, some functional retention curve' between basic (and routinely measured) soil properties and the water wRq phs soil resistance curve, (sRq are necessary. The l{RC describes the matric potential' relationship between the volumetric water content, 0, and the soil
(^SR, using a S,IRC describes the relationship between the soil resistance Vn, 'îdthe important to derive cone penetrometer) and the soil water matric potential (wò' lt is
parameters' models that can predict these functions from routinely measured soil
management These functions need to be derived for different soils, climates and
these all affect how regimes, taking into account their temporal variability, because
period, can be included to the LLlqRvaries. This information, collected over a long
and management' predict the LLWRover a certain period for a specified soil, climate
1.2 Research Questions and Objectives
index of soils To be able to confidently predict the LLI{R and to use it as a structural information is for plant growth in texture-contrast (Duplex) soils, some crucial, basic
required. This study focused on the LLWR and addressed three basic questions:
3 soil change with depth? Ð To what extent does the LLWRof a texture-contrast fall outside the LLIøR ä) To what extent does the volumetric water content
during a typicalmediterranean growing season' Pa"r? plant performance in a äi) what impact do the LLWR and Poo¡ have on field
mediterranean climate?
The detailed objectives of the research were to: c|ay content, organic carbon Ð determine which soil properties (e.gbulk density, influence the wRc and sftc content and carbonate content) most strongly
with dePth.
of the LLWRin a Duplex soil' iÐ characterise the spatial variation (with depth) falls outside the LL|I/R n a iiÐ characterise the extent to which the water content
growing season, Prr¡'
on the variation i[|IPo*' Ð evaluate the influence of some cropping patterns crop growth factors such as grain v) evaluate the effect of Poul and LLl|tR on
content' Yield and grain Protein
1.3 Structure of the Thesis
broad goals were formulated' The In characterising the LLWRof a duplex soil, two of a Duplex soil changed with fnst goal was to find the extent to which fhe LLIItR
depth'Twomainquestionswere'firstly,(CHAPTER3)whatsoilpropertiesmost were considered crucial in the strongly influenced the llRC and SRC? (These and secondly, did the LLWR of a development of reliable pedotransfer functions),
Duplex soil varY with dePth? which the water content fell outside The second goal was to determine the extent to (CHAPTER 4), the efflect cropping pattern the LLWRwith depth in a growing season,
4 fell outside of the LLIryR had on the freque[cy with which the water content plant growth and subsequent (CHAPTER 4), and finally, the effects these had on yield (CHAPTER 5).
5 CHAPTER 2 Z LITERÄTURE RE\rIEW
2.1 Introduction
below), has been proposed as a The Least Limiting Water Range (LLWR, defured growth (da Silva' et al'' 1994)' promising indicator of soil structural quality for plant
and is calculated as a range in soil The LLWRcharacterises the soil structuralform, It integrates the impact water content where crops experience the fewest limitations' over plant growth' and can of various soil physical properties that exert direct control how these aflect crop be used to study the dynamics of soil structure and of soil that change with performance. It is therefore critical that the characteristics growth be accounted for' This time be understood and that their influence on plant relate to plant growth and review highlights the soil structural parameters that plant available water in soils' The evaluates the utility of current models describing properties vary with depth (e'g' applicability (or lack of) for use in soils whose to the project work undertaken for Duplex soils) is also evaluated as an introduction
this thesis.
2.2 Definition of soil structure incorporatíng spatial &
temporal variabitity
that addresses only one Because the structure of all soils is dynamic' any defilrition for the way in which the aspect of structure (e.g. architecture) without accounting
and management is rather limited' architecture changes under the influence of climate
6 (Dexter, 1988; Kay' 1990)' so Soil structure is variable both spatially and temporally (Kay' 1990)' any definition to be adopted should address this variability
the structure has been described To describe both the spatial and temporal variability, "structural is the in terms of forrru stability and resiliency (e.g. Kay, 1990)' form that exists in the soil at a given heterogeneous affangement of solid and void space of primary time,,. The structural form describes the architecture' or the arrangement pore characteristics (ø'g' particles into hierarchical structural units (Figure 2.1) andthe "structurql stability total porosity, pore size distribution and pore continuity)' of solids and void space when describes the ability of the soil to retain its arrangement when characterising the exposed to different stfesses". This property is important change of the structure' structural form as it gives information on the rate of structural form ,'structural resiliency describes the ability of the soü to lecover its
applied to the soil has been through natural processes when stress that has been natural regeneration processes reduced or removed". This describes the influence of
and other natural processes such as wetting and drying cycles, freeztng and thawing
involved inaggregation of soils'
resiliency (Kay, 1990) allows Defining soil structure in terms of its forrr! stability and can be related to crop structure to be characterised quantitatively in ways that
performanc e, aîd is therefore adopted here'
7 lgrc¡-Ètilgî-¡gul ¡ 3or'0 i
I
2000 ¿,í. I îa¡t
t
Foot! o¡d rt9ìo. (¡r¡dh¡¡¡lrrrn arqOniC I Rool Htp¡o 2@ pn aggr$otc3 ol got?rclr3
Plool oîd lungol drbril llìGrutlcd nlñ úìorgonicr (p¡rirl¡¡l orgoaicl Hypho
20 pm Boclcrrum
Poclctr ol clcY porlrclat
Micråbrol ttcDr¡r (huíì¡c molt,io13¡ 2pn
Clot Po?lrclca
Clot plolct
Crmcnl
organisation of soil Figure 2.1 Diagrammatic representation of the hierarchical 1982)' p"ï*i* and Joil structuralinits (from Tisdall and Oades'
8 2.SSoilstructuralquatifyindicesforplantgrowth
proposed as possible indices for A number of soil structtnal parameters have been assessing soil structtnal predicting plant growth. Inadequacy of these indices in The following section quality for plant growtþ has prompted research in this field' their advantages and their will review some of the indices that have been proposed, et al', (1994)'s LLWR is the best limitations and finishes by explaining why da Silva
measure of soil structure relevant to plant growth'
2.3.1 Aggregate Characteristics
stability has been used to Distribution of aggregate sizes, their strength and growth (Alexander et al',1991; Logsdon characterise the structure of the soil for plant properties of aggtegates afe et al., 1987; Braunack and Dexter, 1939). These (especiaþ the mobility of important in controlling air, water and nutrient mobility
and root proliferation' Aggregate elements such as phosphorus' Linquist et al',1997) are important for seedling characteristics control seedbed properties, which and also aeration (oades' 1984; emergence, water and nutrient storage and supply, for example' has been used Braunack and Dexter, 1988). Aggregate size distribution, especially in early days of seedling to characterise soil quality for plant growth, ranges 1-2mm have been growth (Alexander and Miller, l99l). Aggregates of size
1973; Njos, 1979) and soya bean suggested as optimum sizes for cereals (Russell,
(Baligar and Nash, 1978; Tisdall and Adem' 1986)'
plant emergence' and Different aggregaÍe sizes have been related to percentage The difference in growth as overall crop yield (Braunack and Dexter, 1988) Plant
9 to limitations to nutrient and affected by aggregate size is not neccssarily attributed Alexander and Miller 1991)' water supply during early growth (Donald et al.,1987; maize (zea mays L) in a loamy soil' Donald et at.(1987), from a study conducted on weight with increase in aggregate srze' found that there was a decrease in shoot dry
size increased from < 1'6mm to 1'6- This phenomenon was observed when aggregate aggregate sizes. Total root 3.2mm,but no significant change was noticed at larger smaller than in finer aggregates length in the coarsest aggregates was significantly to be a result of larger aggregates (Donald et al.,lgg7). This has been hypothesised nutrient and water absorption' having poor root-soil contact resulting in reduced to reduce plant growtlr' and this has smaller aggregates (< 1 mm) have been observed limitation (Donald et al'' 1987)' been suggested to be attributed to aeration
has also been found to be related to seed The aggreg ate-suelimitation to plant growth Optimum range of aggregate sizes for a characteristics (Braunack and Dexter, 1988). the different crop species' For example' rangeof plants varies with the seed size of (198S) in cereals' it was observed from a study conducted by Braunack and Dexter in earlier emetgence and higher yields that interme diate aggregates (l-2mm) resulted loam soil' Emergence of wheat was also of wheat than large aggregates Q4mm) on a sometimes changed the trend of greatly influenced by climatic conditions, which
optimumaggregalesizes.Indrieryearssmal|aggregates(<1mm)promotedearlier highlights the importance of climatic emergence than aggregates > lmm. This of taking climatic variability into conditions on plant growth, and the importance
account.
10 to affect root and shoot Aggregate strength characteristics have also been observed and the size of aggregates growth (Håkansson and van Polgur, 1984)' The strength by the roots, as high strength will dictates the volume of the soil that will be exploited to the water and nutrients in the restrict root growth and subsequently limit the access polgur that small aggregates with high soil. Håkansson and van (1984) observed
same effect was observed by Braunack strength reduced root and shoot length. The was an optimum soil structure for and Dexter (1989), and they suggested that there and this depended on the maximum root development of a given plant species
a basis for Dexter's (1978) model aggregatestrength. This relationship was used as depth' As soil is a heterogenous on root growth in soil with uniform structure with
cannot be achieved in real field situations' systerry uniformity in structure with depth This model' however' is so use of this model under field conditions is limited'
crop root response is monitored' indispensable in soil strenglh studies where
however, is inadequate' as various characterisation of the aggregate form alone and Wander, 1998)' Other properties factors cause the form to change (Xue-Ming
therefore be included (e.s.stability and that address the temporar variability should be attractive' the influence of factors resiliency of aggregates). For this parameter to changes in the aggtegate s\ze artd such as tillage and cropping practices on the soil properties such that separate stability should be easily separated from intrinsic
effects of each could be quantified'
characteristics is the multitude of methods Further complication of usage of aggregate of properties such as for their determination. Determination and interpretation This is caused by the difficulty of aggregatesize distribution and stability is complex'
11 conditions' Even with these duplicating field conditions under controlled laboratory conducted using standardised limitations, assossment of aggregate characteristics is wet sieving method for methodology for ease of comparison, for example the Rosenau, 1986; Xue-Ming and assessment of aggregale stability (Kemper and
Wander, 1998)
2.3.2 Bulk density and relative bulk density
of the soil or its total Bulk density is mostly used to characterise the compactness porosity(e.g.daSilvas/al',1997),andcanbeusedtomeasuretheefTectof Bulk density is grea|ly management and tillage on soils (Håkansson, 1990). such as texture (Rawls, 1983) and influenced not only by inherent soil properties
et it is also dependent on organic carbon content (Rawls, 1983; da silva al.,lgg4),but (e'g farmimplements and vehicles) and compactive forces arising from management shrinking and swelling)' These other natural factors (e.g precipitation, overburden, bulk density' The bulk density properties alter the total porosity by changing the on pore size distribution (Kay and indirectly influences plant growth by its influence
Grant, 1996).
under different when soils with different inherent characteristics are compared itself can have serious limitations management and climatic regimes bulk density by density, however, has been (Soane et al.,l98l; carter, 1990). A relative bulk
(1990) as a measure of soil structure that proposed by Carter (1990) and Håkansson
carbon content in different locations' can noÍnalise the effects of texture and organic of the observed bulk density to Relative bulk density has been defmed as the ratio (Hakansson' 1990)' For bulk density under some standard compaction treatment
12 example,Håkansson(1990)proposedtheuseofanoedometer,witha200kPa while carter (1990) proposed the use of pfessure to determine the maximum density' bulk density' the Proctor test to establish the maximum
FromthestudyconductedbyCarter(1990)inCanadiansoilstherelativebulkdensity yields' It was observed that optimum yield was assessed for its influence on cereal the relative bulk density was (¿gs%of possible maximum yield) was obtained when
observed, in temperate humid conditions of between 77 a¡d84%. Hå[(ansson (1990) bulk density and yields of spring barley' sweden, a relation between the relative climatic regimes and specific crops and These studies were conducted for specific concept can be adopted in different soils. Therefore, before the relative bulk density in different soils' for different crops regions, further studies should be conducted of climatic conditions on the relative bulk under varying climatic regimes' The effect 1 various bulk densities [{ be the most important variable because r)i density would appearto
havebeenfoundtobeoptimumunderdifferentclimaticandsoilmoistureregimes quantifred the separate (carter, 1990). Recently da silva et at. (1997) successfully properties on bulk density using a multiple eflects of management and inherent soil apparently important inherent soil regression analysis that incorporated all the
properties, management and natural factors'
2.3.3 Macroporosity and pore continuity air- are responsible for rapid dratnage and Macroporosity characterises the pores that which fall in this range are variable filled porosity. Definition of the pore sizes of this study the range proposed by (Dexter, 1988; Kay, 1990), but for the purpose
here. Texture and organic carbon content carter (1990b) viz. >50pm, will be adopted
13 k strongly influcnce macropores by directly controlling aggregation (aggregate size
can distribution and their stability) (Kay and Grant, 1996; Quiroga et al', 1996)' It soil therefore be inferred that macroporosity is not only controlled by inherent soil properties alone, but also by management regimes which have an influence on
properties. The eflect of management on macroporosity is well illustrated in medium
1995)' textured soils where organic carbon contents are often low (Gregotich et al''
is applied Seedbeds of these soils are less stable and easily collapse when stress
resulting in compaction and decreased in macroporosity (Thomas et al',1996)'
(Carter, 1990b)' Macropores influence the different physical characteristics of the soil or Carter (1988, 1990b) observed that soil strength (expressed as shear strength
penetration resistance) could be predicted for a Canadian sandy loam from soil measuïements of macroporosity. Hence, knowledge of the critical lange of 't the soil Itf strength and macroporosity for rooting response could be used to describe
physical quality (Carter, 1990b). The impact of macroporosity on soil physical
properties is best observed when pores are drained from saturation, especially at low
suction (Kay and Grant, 1996). Their effect was demonstrated in studies conducted
to by Vepraskas (1984) and Byrd and Cassel (1930) where soil strength was related
pore stze distribution, especially the macropores which tended to decrease
intergranular or efflective stress when the strength was determined (Carter, 1990)'
Root penetration in soil of high strength takes advantage of macropores (such as
Dexter, biopores) to access water and nutrients in deeper horizons (Jakobsen and
lggS). The probability of a root gaining access to water and nutrients deeper in the
soil is influenced by the extent and dimensions of the macropores leading downward r t4 optimum properties of (Jakobsen and Dextcr, lggg; volkmar and Entz, 1995). The properties (density and macfopores for plant growth depend on their physical
a! characteristics (Jakobsen and dimensions) and other factors such as climate and crop
Dexter, 1988)
as a result of both crop and Loss of macroporosity and decrease in pore continuity of water and solutes in the soil management practices has an impact on the movement
and the redistribution of (white, 1985), as illustrated by the varying infiltration rates the water potentials water in the soil. This effect can be characterised by studying
(Kay and Grant, 1996)' close to zero where macropores are water-filled
density reflect changes ln Changes in both bulk density and relative bulk 7996), hence, the impact of macroporosity (carter, 1990b; Kay and Grant,
'J changes in rçlative yield as a til macroporosity on plant growth can be inferred from '!
function of normalised bulk density'
2.3.4 Plant available water capacity
(PAWq has been used to define The concept of the'þlant available water capacitt'' point over a uniform soil water at potentials between field capacity and wilting water content to the volume rooting depth (Gardner et a1.,19S4)' This concept relates
relevant soil matric potentials (e'S'- fraction of pores (Kay and Grant, 1996) between
however, that water is not equally available 10 kpa to -1500 kpa). It is well known, factors such as inherent climatic t plants across the entire PAWC-range due to other I to Gardner et al'' 1984) and l conditions( e.g. evaporation, infiltration and precipitatiory Letey, 1985)' In some soils' variable soil physical limitations (strength and aeration;
15 3 (-0.01 MPa; Haise et al'' 1955) even when water potentials are between field capacity lg44), water availability is and wilting point (-1.5 MPa; Richards and weaver, poor soil aeration' where air-filled restricted due to physical limitations including: 196S) and high soil porosity (volumetric air content) < 10% (Grable and siemer'
strength,wherepenetrometerresistance'SR>2MPa(Tayloretal''1966)'
as a structural index for Nevertheless, the PAWC concept has been used effectively good correlation between the plant growtlU as shown by Burt (1995), who observed a
PAWC and winter wheat Yield.
and \AWC has also been used extensively to formulate irrigation scheduling
Beemstere, 1989; Madsen et al'' calculating regional soil moisture deficits (Bergt and of crop production 1989; Kern, lgg5), agro-ecological zontng and assessments simulation of global land cover potentials (Reinds et a1.,1992;Luyten, 1995) and for 'I [,f il changes (Prentice et al.,1993;Zuidema et al',1994)'
to 2,3.5 Least Limiting Water Range (LLWR) and its Relevance
Plant Growth
plants include the supply of water, The main processes in soils that impact on growing root system' Porosity and nutrients and oxygen, and the proliferation of an adequate here' The LLWR incorporates strength of the soil invariably exert great influence aeration, soil strength, and limitations of water content on plant growth related to
and wilting point. For example, in water retention between the classical field capacity
(1996), under humid temperate conditions in a preliminary study by da Silva and Kay
I directly related Io LLIYR' and corn Catada,shoot growth of corn (Zea møys L.) was
l6 r the soil water content was grorwth decrcased linearly with increasing frequency that
at the wet end and dry end outside of the LLWR The limits of the range are defined
as follows:
2.3.5.1 Upper limit (Wet end)
the volumetric water content The upper limit of the LLWR is selected as the drier of
volumetric water content that corresponds with at -0.01 MPa (field capaclty) or the bulk volume) (da Silva et al'' 100/o volumetric air content (f,ir:0.10 cm3 att I cri
potential that corresponds to field lgg4). The choice of the -0.01MPa as the matric based on the work of Haise et capacity of an unloaded sample by desorption, is matric potentials have at.(1955), and although arguments supporting other relevant
generally accepted' been raised in the literature, the -0.01 MPa is
oxygen diffusion decreased to In Grable and Siemer's (1968) study they found that and most compact soil' The zero at or near fo¡, : l0 to l2o/o in small aggregates 12% (Figure 2'2)' relative diffusivity increased exponentially as /¿ri exceeded diffr'rsion' Different arguments assuming there is a relationship betweenf't and air when one considers for could be raised as to the applicability of this assumption
sandy soils will differ for the example that diffusion in clayey soils as compared to
of their pore geometry' same volumetric air contents due to the variability
justifu selection of the limiting parameters It is not the intention of this review to the
study area' In summary iffi adopted for characterisation of the LLl|/Rof soils inthe limit of LLInR will be considered to be the > l0o/o at w: -0.01 MPa, the upper i
l1 I at frelrl classical "field capacity'' for the "plant available water" concept. lf .f', .10%
capacity, then the upper limit will be chosen to be that wheref ir: ljYo.
2.3.5.2 Lower limit
The lower limit of the LLtl1R is selected as the wetter of the volumetric water the soil contents at a matric potential of -l.5MPa (wilting point) or at which
as the matric resistance is 2 MPa (da Silva et al.,lgg4). The choice of the -1.5 MPa
potential that conesponds to wilting point is based on the study by Richards and
'Weaver (1944), and is adopted here.
The selection of the 2 NlPacut-off value for the soil penetration resistance was based
were inhibited on the study by Taylor et at. (1966), who found that tap roots of cotton
halted when soil penetrometer resistance reached 2 NIPa, and growth was completely
the 2 MPa at2.5 Mpa (Figur e2.3). One can argue the applicability of the selection of
plants have difÊerent limit to cover a wide range of crop species. It is well known that
abilities to penetrate soil with high strength and this depends on plant physiology
(Materechera et a1.,1991). It is not the intention to debate the applicability of this
more strength-limit here, but to advise that it be used as an approximate figure until
suitable values can be incorporated into existing models'
2.3.5.3 Prediction of the LLWR
condition Accepting the LLlItRas the most eflective tool for describing the physical
of the soil for root exploration, it is necessary to measure it, or where this is not
possible, to estimate it from the WRC and SAC.
I
18 o.
05 ¡ o/0 r$ t Ò{05ñn 0 a x o 0.5 - | 0.¡t r l-2 3 a , 2-3 c tt o 3' 6 a c s0lL 1 Voo. 0.68 s I o.2 q s.3 c .c T x l0 I 'o/oo's 0.r I 0
0 ¡o eo 30 40 50 60 70 AIR POROSITY - PERCENT
t relative Figure 2.2.Inf7uence of air porosity (fan c air / cms_bulk) on dilfusivity of oxygen in siltyclay loam at different bulk densities and aggregate sizes (Grable and Siemer, 1968).
t9 2.3.5.4Estimation of the Water Rctention Curve (WRC) unavailable in most Representative databases of soil hydraulic information are
and laboratories countries worldwide. Direct compilation of hydraulic data from field is generally too costly and cumbersome, and therefore, impractical for most Hence, there is applications (William s et al., 1992 and Cresswell and Paydar, 1996)' routinely a need for robust methods to predict the wRC from more easily and content (0) as a measured soil properties. Relationships describing the soil water
forms, and it is function of the matric potential (W) æe available in various (empirical)
predictions can be requisite to find the best model for a set of data from which other
and Corey, made. A wealth of literature has been published in this area (e.g. Brooks and cass, 1987; 1964; King 1965; Campbell, 1974; van Genuchten, 1980; Huston not be derived Ross e/ at.,l99I) and the formulation of each of the many models will provide here. However, some models have been fitted well to Australian soils and water by good fits to the available data across a range of potentials important to
plants.
proposed by The models commonly used (Table 2.1) are the power function models (1930)' The model Brooks and Corey (1964), Campbell (1974), and van Genuchten
points across the proposed by van Genuchten has been found to give a best fit of the
Brooks and entire soil water retention curve' while those of Campbell (1974) and aI., (1991) corey (1964) fit some parts of the data better than other parts. Rawls et
between show three soil water retention models and their assumed relationships
different parameters.
20 }{ v V QTJINLAN 80 x C OLUMBI.A NARON z Y * o o t o MILES l-- x 60 U Y æ. Þ- zlrl ul 40 o- *b+ t- bzo o oT* O T{¡ É. t( k'f Y o Fhto f drvw x 50 o to 20 30 40 soll- STRENGTH ( BARS )
Figure2.3.Relationshipbetweensoilstrengthasmeasuredbyacone penetrating through cores of penetrometer and p.r."ot"ge oj c_oJlon tapîoots iour different soils (Taylor et al',1966)'
2l Sinrple power functions such as that of Brooks and Corey (1964), Campbell (1974), and Hutson and Cass (1937) have been proposed for use where soil water retention models are to be predicted entirely from independent soil physical properties and characteristics. Da Silva e/ al., (1994) proposed the use of Ross et al.'s (1991) model:
0: a,tb l2.ll
where áis the volumetric water content (cm3 cm3),,lris the matric potential (MPa),
and a and b are dimensionless constants. A good fit of data points across a range of
potentials corresponding to plant available water was obtained in the study of da
Silva et al.(1994).
Caution should be exercised, however, when selecting a model for a set of data, so
that all parameters can be easily measured or estimated. For example, the models of
Brooks and Corey (1964) and Campbell (197a) use a parameter called the "air entry"
pressure, which is diffrcult to measure because it occurs as an inflection point in the
water retention function. By the same token the van Genuchten (1980) model has
been criticised by some (Cresswell and Paydar, 1996) due to the relatively large
number of parameters involved. The van Genuchten (1980) model, however, has
been found to fit data well across the entire range of relevant water potentials (Rawls
et a1.,1991 and Cresswell and Paydar,1996), and is therefore the preferred model for
this study, where the various parameters can be estimated with reasonable accuracy,
and where sufflrcient computing power is available.
22 Table 2.1. Soit water retention models (Rawls et a1.,1991)
HTdraulic soil Pa rr mctc r ch ar¡clc rist ic P¡ramctcrs co rrcs po ndc ncc
A. Brooks-Corcy (196a)
Soil ,'¿ater rctcntion 8-e, (Ð' 4 = satura(cd watcr contcnt 7= ) 0r-t, distribution ht=h¿ I : pore sizc indcx 0,':8, å¡ : bubbling pressurc or air 0r= ú ' entry pressure 4 : rcsidualwatcr content / : total porositY B. Campbcll (1974)
Soil watcr reten(ion tlt e Hb 4 : saturatcd watcr content 0r: ú 0, h I/u : bubblíng pressure or air Ht= hu entry Pressure ¿r=l b : constant t-
C. van Gcnuchten (1980)
Soil rvatcr rctention e-e, 4 : saturated water content 8": ú 0,-0n 4 : rcsidual watcr contcnt 0r=0, d : constant a : (år)-t n : constant n=).+l I m : cons(ant m:1-l.+ r
0 = Volumctric wa(cr contcnt. å = Capillary suction (cm). C = PorosítY.
23 2.3.5.5 Estimation of the Soil Rcsistnnce Curve (SnQ
Models to describe the relationships between water content and soil strength have
(1982), and been proposed by Mirreh and Ketcheson (1972), Ayers and Perumpral (1987). Soil Busscher and sojka (1987), and these have been reviewed by Perumpral 1963; strength varies with bulk density, soil water content (Taylor and Gardnet' (Gupta Mirreh and Ketche son, 1972 and Perumprul', 1987), texture, organic matter
differentiate and Larson, 1982 and Spivey et a1.,1986), so it is not a simple matter to
models the sole effect of soil water on soil strength. Busscher (1990) tested various
content, as to describe the relationship between soil strength and a normalised water
follows:
+b pt +d yr+e +Íp,t, 12'21 ^SR=ø p+" ú SR=(upr/@+(0-dh [2'31
Sp=apb 6 12.41
sR = ¿ po ,t" t2'51 (Mg/m3), where SR : soil resistance to a blunt penetrometer (MPa), p: bulk density
and o, b, c' rpr: soil matric potential (MPa), á: volumetric water content (cm3 crn3) ,
d, e, andf are adjustable parameters.
Most of the models fit his data well, but da Silva ef at. (1994) and da Silva and Kay
(1997a) used Busscher's equation [2.5] with greatest success and so it will also be
used a used here. Arguments could be raised on the selection of this model, as it
l3mm diameter cone penetrometer (30' angle), whereas this study used a smaller tip
(2mm) with a 40o cone angle. Nevertheless, without indications of a better model,
which have variations Equation [2.5] will be applied here to determine SRC for soils
in physical properties ranging from being slightly swelling to non-swelling and
increasingly caloareous with depth.
24 2.3.5.6Pedotransfer functions and their use to characterise the LLWR pedotransfer functions have been defined as mathematical expressions relating diflerent soil characteristics and properties to one another, or to land qualities, and can (Bouma be used to calculate more complicated parameters from simple characteristics
(1997a) used and van Lanen, 1987 and Wagenet et al., 1991). Da Silva and Kay pedotransfer functions with great success to predict the WRC and SRÇ for example:
For the WRC:
InT=lna+blny 12.61
where: ln a : -4.1518 + 0.6851 ln Cløy + 0-4025 tn OC + 0.2731 ln p 12'71
b: -0.5456 + 0.1127hn Ctøy + 0.0223 ln OC + 0.1013 ln p [2.81 : o/oby d: volumetric water content (cm3crn3), ry: mattic potential (MPa), Clay
(%), bulk density (g .*'), a and b weight .2 þoV OC:organic carbon content p:
are adjustable parameters.
For the SftC;
lnSR=lnc+dlnQ+elnp 12.el
where: ln c : -3.6233 - 0.1447 Clay + 0.7653 OC [2.10ì
d = -0.4805 - 0.1239 CIay + 0.2080 OC 12.ttl
e:3.8521 + 0.0963 Clay 12.t2l
Sft : soil resistance (MPa), A OC, and pare defined above, c, d, and e are adjustable
parameters.
Fromthe pedotransfer functions above, it is possible to estimatethe LLÍTR from data
onthe clay content, organic carbon content and bulk density of a given soil. Similar
25 frnrctions will be used in the present study to calculate the LLWR of the soils involved
(CHAPTER 3).
2.4 Duplex soils and pedotransfer functions
Duplex soils ofler an excellent opportunity to study the applicability of models to
predict the LLú\1R in the plant root zone. Little account is generally taken of the
enormous variability in physical and chemical properties that occur with depth in
these soils, yet these exert enormous influence on the soil hydraulic properties, and
presumably therefore on plant performance. This section reviews some of the
relevant properties of Duplex soils in preparation for the application of pedotransfer
functions to determine the LLII/R for a Duplex soil in South Australia, where the
properties vary signifrcantly with depth in the rooting zone.
2.4.1 Definition
Duplex soils in Australia have been defined as those with markedly contrasting
texture between surface and sub-surface horizons (Northcote, 1979). Thç texture
contrast is of at least I Vz l.rcrture groups (e.g loam to silty clay loam or greater, cþ
loam to light clay or greater) between A and B horizons (Northcote, 1979;
Chittleborough, 1982). The horizon boundaries are clear to sharp. Even though the
term is not used elsewhere, soils with similar characteristics are contmon in most parts
ofthe world.
26 2.4.2 Origin, tlistribution and agricultural usc of duplex soils
No single explanation accounts for the origin of all duplex soils but three main pïocesses are thought to be involved in their formation, including (Chittleborough,
l9B2; 1992): i) sedimentary layering, ii) formation of clay-sized minerals in situ by differential weathering, and iii) downward translocation of dispersed clay within an initially homo geneous material.
Duplex soils occupy approximately 20Yo of the Australian continent and many of
these soils are sodic andlor saline (Chittleborough, 1992) (Figure 2.4). They have
been used extensively for agricultural production in Australia. Their importance has
prompted extensive research on how to manage them effectively to maximise returns.
Due largely to variation in their inherent properties with depth (e'g' texture,
mineralogical composition), management of these soils is difficult. They experience
waterlogging in late winter and early spring, limited retention and availability of
water and nutrients in the surface soil, and limited root penetration into the cþ
subsoil, which reduces plant-water extraction especially during spring (Robson et al.,
1992; Gardner et al., 1992; Belford et aI., 1992; Dracup et al., 1992; Turner, 1992;
Gregory, 1998). These problems generally reduce crop growth and yields and
certainly make yield highly variable both spatially and temporally (Belfotd et al.,
r9e2).
Physical limitations due to high subsoil bulk density afe coÍrnon (e.g. ¿ 1.8 Mg/m3,
Dracup et al., Igg2) and penetration resistance frequently exceeds 2lli/.Pa (Tentnnt et
at., 1992). Reduced root growth under these conditions invariably reduces yield
(Gregory, l99S). Furthermore, variation in the physical composition of the A-horizon
27 grow into the B- has been fourcl to have an influencc on the ability of the roots to generally horizon(Dr acup et at., t993). Root systems of different crop species are due to shallower and smaller in duplex soils as compafed to other soils, mostly
1996)' physical limitations in the B-horizon (Dracup et al., 1992; Gregory et al',1992'
properties The importance of soil chemical properties in controlling soil physical
These cannot be overlooked, especially in sodic duplex soils (Tennant et al',1992)' A soils are characterised by low infiltration rate and low hydraulic conductivities'
a result review on the soil physical properties limiting crop production in Australia as of sodicity is well documented by Jayawardane and Chan (1994).
2.5 Summary
for plant The LLI(Rhas the potential to be a robust index of soil physical condition root growth because it incorporates soil properties that have a direct impact on
growth. It is, however, difficult to measure (time consuming, costly), and so suitable
pedotransfer functions are needed so that LLWR can be predicted from more-readily to qualiff the measurable I available soil properties. Furtherrnore, research is needed
and applicability of the LLWRto different climatic conditions, management regimes
determinant of soil types. This is important in areas where plant available water is the
mediterranean the success or failure of the crop, such as in the South Australian to climate on duplex soils. The work reported in the following chapters attempts
address these research needs.
28 Ð
,
rß" (ö) ,' o
a 'a¡ t a ) 1 1! / Þ
w Figure 2.4. a) Distribution of duplex soils in Australia, b) Sodic duplex soils (after Chittleborough, 1992). 29 CHAPTER 3:ESTIMATION OF LLWR FROM SOIL PHYSICAL PROPERTIES 3.1 Introduction The influence of soil properties on the LLWR, especially in soils where physical and chemical properties vary with depth, as in Duplex soils, is poorly understood- The LLWR is obtained from the IYRC and SÃÇ so its dependence on more readily measurable soil properties should be obtained from these properties and these functions. An understanding of the influence of soil properties on the LLll/R n texture contrast soils will be helpful in the management of these soils. Computation of the LLWR requires that the relationships between water content, á, and water potential, % soil resistance, ,Srl, and aeration .fo¡r, be obtained' The relationship between 0 and Y Q.e the WRQ can be modelled using various functions (e.g. van Genuchten, 1980; campbell, 1974; Brooks-corey, 1964). The Brooks- q1.,1991) Corey function and its variants (ø.g. Hutson and Cass, 1987; Ross et have been found to provide a reasonable fit to water retention curves in the region important to water available to plants (da Silva and Kay, 1997a). Da Silva and Kay (I997a) used the model by Ross et al., (1991) in non-swelling calcareous soils with great success. The same model will be applied to the soils in this study because they share common properties (e.g. nominally non-swelling and calcareous at depth) and thus satisfu some of the criteria for the effective application of pedo-transfer functions. 30 Soil resístance (Sft, as ilìeasured using a penetromctcr) is strongly influenced by both 0 andbulk density, p (Busscher, 1990; Becher, 1994). The relationship between Sft, 0 and p cffiLbe described in functions, which will be termed here, Soil Resistance Curves (.SRCÐ. The model by Busscher (1990) which has been used successfully by da Silva et at. (1994) and Bertz et al., (1993) to describe the relation between the soil resistance to penetration by a metal probe and the soil water content will be used here. The aeration status of the soil can be quantitatively described from the air-frlled porosity (Xu et al., 1992),f'¡', which can be calculated from the total porosity, f, and the volumetric water content, 0. Aîinverse relation exists betweenfi¡ and 0. Derivation of the LLWR from the above-mentioned functions is both technically diffrcult and time consuming. This makes the LLIIR less attractive as a routine analytical tool unless pedotransfer functions can be developed. This was achieved by da Silva et al., (lgg4) for a silt loam and a loamy sand soil. To do the same for a Duplex soil in the present study, two objectives needed to be met: (i) identify which soil properties most strongly influence the l{RC and SftC and (ii) develop pedo- transfer functions that relate WRC and SAC functions to soil properties with depth. 3.2 Materials and methods 3.2.1 Location The study was conducted at the Roseworthy campus of the University of Adelaide (lat. 35. 30'S, long. 138' 40'E), 50km north of Adelaide, South Australia. The site has a mediterïanean climate, characterised by cool, wet winters (June to August) and 31 warm, dry summers (December to Februaly). Tlre average annual rainfall of the area is about 440 mm most of which is received during the winter and early spring months. The farming systems in the aÍea are dominated by cercaVlegume-based pasture rotations. The site was chosen because two crops with different root morphology were growïr and two cropping patterns (viz. mono-cropping and inter-cropping) were being practised, such that soil water variation between treatments was expected. Furthermore, soil properties were thought to vary across the experimental site down a slope, so replicated field plots were established across the topo-sequence from the bottom (Rl) to the top (R5XFigtue 3.1). The soil is classified as a red-brown earth (Dr 2.43, Northcote, 1979) or as a Sodic, Hypercalcic, Red Chromosol ( Isbell, 1996). The soil was a sandy surface sharply overlying a red poorly structured sandy clay, highly calcareous with depth. 3.2.2 Sampling A hydraulically-driven coring device (mounted on a four-wheel drive vehicle) was used to push a 5 cm-diameter steel tube to a depth of 80 cm. The soil was too hard to obtain samples below this depth with the equipment available. Sampling was conducted between August and November, 1998 on five replicate-plots of lucerne stands, rather than on all the different treatments in the field experiment. The paddock had previously been an established lucerne stand prior to the introduction of cereals in the field experiment. (No major soil structural changes were expected from the relatively recent changes to the crop and soil management in the paddock)' Six cores were collected in a circle of I m radius to minimise within-treatment sample- variation. This was done at two locations in each plot for a total of 12 cores per plot. 32 Samples were collected close to the edges of the plots (15 m x 5 m) to avoid plant disturbance. The soil cores were sealed in plastic and placed in cold storage (4" C) until they could be analysed in the laboratory. 3.2.2.1 Sam ple identification Due to the large number of samples collected, notations were used to identifu samples collected from different positions and depths (ø.g. Rl I denoted samples collected at position I (similar to R1 in Figure 3.1) at depth I (0 - 10 cm)). Positions were from I (R1, bottom of the landscape) to 5 (R5, top of landscape)(Figure 3.1), and the depths were: 1 (0 - 10 cm),2(10 - 20 cm), 3 (20 -30 cm), 4 (30 - 40 cm), 5 (40 - 60 cm) and 6 (60 - 80 cm). 3.2.3 Physical and chemical properties 3.2.3.1 Particle size distribution Particle size distribution was carried out using the dispersion, sedimentation and sampling method (modified method of Day, 1965). Air-dried soil samples were passed through a 2 mm sieve before analysis. Samples of the < 2 mm air-dried soil fractions, weighing between 10-20 g, were dispersed with 10cm3 of 10% calgon (sodium hexametaphosphate) solution, 0.5 cm3 of 0.6 M NaOH and 150 cm3 distilled water. The samples were agitated in an end-over-end shaker for up to 24 hrs. Additional distilled water was added to ensure mixture contained < lyo suspended material, and the mixture was stirred and allowed to settle according to Stoke's Law. At an appropriate time the < 2 pm fraction was sampled, then the entire sand fraction 33 (20 - 2000 pm) was isolated and collected. The silt fraction (2 - 20 ¡rm) was calculated by difference. 3.2.3.2 Organic carbon content Organic carbon content of the <2 mmfraction was determined using the Walkley and Black titration procedure (Nelson and Sommers, 1982). Soil samples (1.0 g for samples in the top 30 cm;2.0 g for samples in 30-80 cm depth where the organic content of soil was low) were digested with 10 cm3 I N potassium dichromate and20 cm' concentrated sulphuric acid. The mixture was agitated for one minute and allowed to stand for one hour. Approximately 200 cnt' distilled water, 10 cm3 concentrated orthophosphoric acid and 0.5 cm' of o-phenanthroline solution (indicator) were then added. The mixture was titrated with 0.5 N ferrous sulphate. Calculation of the organic carbon was based on the assumption that each crn3 of I N .I ìrl ,,i potassium dichromate was equivalent to 3 mg of carbon. ,J 3.2.3.3 Carbonate content Carbonate content was determined by the volumetric-calcimeter method (Allison and Moodie, 1965). Soil samples (weighing between 0.16 and 0.68 g) were treated with 25 cri hydrochloric-ferrous chloride reagent, and the displaced CO2 measured using the calcimeter (volume displaced). The calcimeter was calibrated using 50 mg CaCO¡ assuming 5 mg CaCO3 produces 1 cm3 COz. 3.2.3.4 pH and EC The pH was measured in two different solutions aI a ratio of 1:5 (soil to solution ratio), where I part of soil (< 2 mm fraction) was mixed in 5 parts of either water or r 34 N ¡ 1 x x X x x R1 w(h) w(d) o L & w(d) w(h) (bottom) &L x x x x R2 w(h) o w(d) o& (d) L \ry(h) &L &L L x x x x x R3 w(d) o& \il(h) o w(h) w(d) &L L &L x x x ï 1 R4 L \il(d) o L w(h) \il(d) w(h) il &L &L I x x X x x R 5 (top) L o w(d) o& w(d) w(h) w(h) L &L &L x x x x x Figure 3.1. Treatment design of the study area. i t I I I Ì 35 solution of 0.01 M CaClz, agitated 15 min, left to settle 30 min then pH determined in the supernatant. Electrical conductivity was determined on the l:5 mixture. 3.2.3.5 Water Retention Curve WRq The 80 cm soil-cores were identified according to their positions in the experimental site (Figure 3.1), i.e. position 1 to 5 (Rl-R5). Preliminary soil investigations showed a distinctive variation in texture with depth, so the 80 cm cores were segmented into six depths: D : 0-10, 10-20, 20-30, 30-40, 40-60 and 60-80 cm, from which subsamples were taken (5 cm long). The sub-samples were carefully transferred into stainless steel rings (5 cm long x 5 cm diameter), and the bottom of each soil core was covered with a mesh attached by rubber bands. The samples were then saturated in deionised water. ll (1986). i The llRC was determined following the methodology described by Klute Eight matric potentials were chosen, -1, -3, -5, -10, -33, -100, -500' and -1500 kPa' These pressures were chosen because they corresponded to pores relevant to plant available water (mesopores, pore diameter between 0.2-30 pm), and macropores (> 30 pm in diameter) that are important for aeration (Thomasson, 1978) and soil strength (Carter, 1990). For equilibration of samples at matric potentials between -1 to -10 kPa hanging water columns were used (Figure 3.2). For matric potentials between -33 to -1500 kPa, pressure chambers were used (Figure 3.3(a) and 3.3(b). Equilibration of samples at Ì the various matric potentials was done systematically by desorption, starting with the highest potential (-l kPa, left for several days) and progressing to the lowest (-1500 r 36 kPa, left for 30 clays). Afler equilibration at each applied matric potcntial the mass of each sample was determined, after which it was re-saturated with deionised water and equilibrated at the next potential. Volumetric water contents were calculated from the measured gravimetric water contents and the measured bulk densities. I 'I Ìi ',ìJ Figure 3.2. Hanging water columns used for matric potentials -1 to -10 kPa. i ? 37 Figure 3.3(a). Pressure chambers used for matric potentials -33 to -100 kPa. t Ì , Figure 3.3(b). Pressure chambers used for matric potentials -500 to -1500 kPa. I 38 3.2.3.6 Soil Resistance Curve (SttC) The same samples used for the WRC were also used to independently derive the SftC at similar matric potentials. Penetration force was measured using a purpose-made constant-rate cone penetromet er (2 mmbasal diameter, 40" angle) on the core samples placed on a top-loading electronic balance (Figure 3.4a). Penetration force at a specified depth was measured for each sample using a constant rate of penetration (2 mn/min). Penetration resistance (pressure) was calculated from the force, and the basal surfac e area of the cone. Due to the destructive nature of the technique and the limited sample number, a way of obtaining numerous readings from a single sample was devised. Two readings were obtained from a single penetration, but at different depths (12 mm and 24 mm)(Figure 3.4b). A total of seven penetrations was taken on each sample, and the highest and the lowest readings were discarded; thus the average of 5 readings was obtained. The seven penetrations were conducted as follows (Figure 3.4c); after the I't penetration, 5 others were taken equi-distant to each other in a clockwise direction. The last penetration was made in the centre of the sample. 3.2.4 Statistical Methods Variability in the data for the IYRC and SftC was expected from three main sources: i) position in the sloping landscape, ii) depth in the soil profile, and iii) differences in soil physical and chemical properties such as clay content, organic carbon content, and bulk density. For this reason, the data were examined using statistical models to separate the various effects on the WRC and the ,SftC. One model focussed on landscape position and soil depth as the primary variables. Another model ignored 39 Figure 3.4 a). Measurement of penetration resistance using a balance and constant-rate penetrometer. b) cJ t_ 6 7 2mm \ 1_ 7 1 24 mm 2 5 I 4 Figure 3.4. b) Procedure for obtaining two readings from one penetration' c) Protocol for soil penetrometer measurements. 40 landscape position and focussed only on tlre soil properties as thc primary variables. The above two models were compared with some less comprehensive models proposed in the literature. 3.2.4.1Models accounting for landscape position and soil depth: To predict the efflect of position and depth on the lllRc, the measured variables (excluding clay content and organic carbon content (o/roq, which were not replicated) were fitted to the linear model: &j*t= C + P¡+ D¡ + Yh + pt [3'11 where hq*t is the volumetric water content 1cm3 crn3) at the lh position, P, in the landscape (i : L.. 5) the jth depth, D, in the soil profile (i : 1... 6), and the /rth matric potential, { expressed as levels: k:1...8, representing 0.001, 0.003, 0.005, 0.01, 0.033, 0.1, 0.5, and 1.5 MPa, C is a constant , and pris the bulk density (g t*') of the /th soil core. All terms were tested for signifrcance (P:0.01) and a final model was derived by dropping all non-significant terms. It should be pointed out that the linear addition of the terms: depth, D, landscape position, P, and bulk densitY, p, to the matric potential term has no physical or mathematical basis. The rnatric potential, of course, is the independent variable, while D, P, and p all represent influences on the &Y relation. The rationale for starting with Eqn 3.1 was simply to minimize the complexity of the statistical analyses involved. measured Similarly, to predict the effect of position and depth on the ^SRC, the variables were fitted to the nqn-linear model: log SRq*r: C * P¡+ ù + Q¡* pt 13'21 4t 1... where s&i¡ais the soil resistance (MPa) at tlre fh position, P, hthc landscape (i = water 5), the/ deptlU D, in the soil profile (i : 1... 6) and at the measured volumetric content (0*, cri crn3) and bulk density (pt, gcm3). Due to the exponential relation of between ,SrR and 0, the SlR data were log-transformed to normalise the distribution results For both the WRC and the SftC, an analysis of variance was conducted to determine the signifrcance of the main effects plus interactions. The square of the correlation coefficient, R, from the ANOVA showed the amount of variance accounted for by the to the model. Where nÍ > O.gO the model was not extended beyond the main effects interactions. 3.2.4.2 Models accounting for soil properties: To generate a model to predict the eflect of soil properties on ll[RC and SÌC, all data were pooled regardless of landscape position and soil depth, because clay and organic carbon contents were not replicated. Furthermore, because there appeared to be considerable differences between soil samples that contained significant quantities of calcium carbonate, Ihe l{RC and SftC models were developed separately for soils with and without carbonates, as follows. Non-carbonate soils To predict the effect of soil properties on the lryRc,the measured variables were fitted to the non-linear model: 1) Iog 0= C + log Y¡(MPa) + loS p(g cm3¡ + tog clay(%) + log OC(g kg t3.31 42 where 8 C, Y*, p anJ OC are as defmed above. To predict thc cfFect of soil properties on the SfiC the measured variables were fitted to the non-linear model: tog SR= C * log0 + tog p (S c^t) + tog clay(%) [3.41 Soils with carbonates To predict the effect of soil properties on the WRC, the measured variables were fitted to the non-linear model: log 0: C + log W@Pa) + togp(g cma¡ + to7 ctaJ,(%) + log OC(g kC-') + log CaCOj(g kE') Í3'51 Similarly, for the SRÇ the measured variables were also fitted to the non-linear model: tog SR= C t togT + tog p(S cm3¡ + log clay(%o) + tog CøCOj(S kgt) 13.61 3.2.4.3 Models from the literature: The WRC data were fitted to the model proposed by Ross et al. (1991): 0: a yb [3.71 which, when log-transformed, gives: log O:logø+blogY [3.81 where a and b arc adjlstable parameters. The ,SftC data (MPa) were fitted to the model proposed by Busscher (1990): SrR : c 0o p" [3.9] which, when log-transformed, gives: tog SR=logc+dlog Q+etog p [3.101 and were where c, d, ande are adjustable parameters. Models for both the lüRC ^SrRC also generated to take into account the presence or absence of significant amounts of soil carbonates. 43 3.3 Results and Discussion 3.3.1 Soil ProPerties for soils near the Figure 3.5 shows that cþ content increased significantly with depth was more variable bottom of the toposequence (i.e. positions Rl & R2) whereas this (p<0'001) in the uppef part of the slope (i.e. positions R3 - R5). A significantly 70 cm at positions R1 higher mean clay content of 36.5Yo was observe d at a depth of slope in R3 and R5' &,F(2,as compare dto l4.lYoclay at the same depth higher in the part of the toposequence The B-horizon occurre d ar a shallower depth at the bottom the toposequence compared to the top. Organic carbon content was low throughout depths (Figure 3'6)' and was not significantly different for various positions and between the A- The presenc e of CaCOs marked the boundary of the texture-contrast in clay content' At the and the B-horizons, and this also coincided with an increase at a depth of 25 crc' bottom of the toposequence (position Rl), CaCOj was present (p<0.05) progressively up the whereas this occurred at increasingly greater depths CøCOs coincided with an toposequence to position R5 (Figur e 3.7). The increase tn was established in this increase in clay content, although no relation between the two at a depth of 70 cm in study. The highes t CaCOs content of 388.2 g kg-t was found 5'6 g kg-l was the lower part of the toposequence (Rl), and the lowest value of (R5)(Figure 3.7). encountere d at adepth of 35 cm at the top of the slope of the toposequence Bulk density generally decreased with depth in all the positions gcrnt¡ at a (Figure 3.8). The top position (R5) had the highest bulk density (p:1.65 3) the bottom position at a depth of 15 cm, and the lowest QD: 1.31g cm was found at 44 Glay content (%) 0 5 10 1520 % 30 35 0 O . R1 (bottom) 10 n ^30E .!) =4O CL upper slope profiles o50o 60 70 lower 80 F'igure 3.5. Clay content (<2 pm) as a function of depth along a toposequence' oG (g kg'î) 2 4 6 I 10 1 0 u r slope profiles 10 20 ^30 profiles IE lower slope .c, 40 fL âsoo O R1 (bottom) 60 70 80 Figure 3.6. Organic carbon content (g/l¡g) as a function of depth along a toposequence. 45 GaGo. (g kgr) 0 100 200 300 400 0 O R1 (bottom) fo lower slope profiles 20 ^30E .!) .c 40 cl o50o tr 60 70 upper slope prof¡les 80 Figure 3.7. CaCO, (g kg-t) as afunction of depth along a toposequence. Bulk density (g cm3) 1.8 1.1 1.2 1.3 1.4 1.5 1.6 't.7 0 - O R1 (bottom) 10 -- E- R2 -å-R3 20 -€-R4 - -å(- R5 (toP) ^30E (J E40 CL ôsoo 60 70 lower slope profiles upper slope profiles 80 Figure 3.8. Bulk density (g ..-t) as a function of depth along a toposequence' 46 depth of 50cm. 'l'he decrease in p with depth at the bottom of tlre toposequence coincided with an increase in coarse-grained, porous CaCO1 3.3.2 Models accounting for landscape position and soil depth Results of the multiple regression anaþsis for the IVRCs and SRCs are given in the following sections. 3.3.2.1Multiple regression analysis Tor WRC The IVRC was fitted to the model shown in Eqn [3.1] and coeffrcients for the factors glVo are listed in Table 3.1. The model accounted for of the variability' Table 3.1. Coefficients for main effects model for WRC' Factors Values assigned to factors Coeffrcients C 0.3638 p (bulk density) -0.0691 P þosition in landscape) 1 (position Rl , bottom) 0.041 2 0.0196 J -0.036 4 -0.0109 5 (position R5, top) -0.0137 D (depth in soil profile) I (surface depth) -0.0132 ,, -0.008 I -0.0118 4l -0.00s6 5l 0.0057 6 (sub-surface depth) 0.0329 ty(matric suction) 0.001 0.01653 0.003 0.1077 0.005 0.0723 0.01 0.0019 0.033 -0.0529 0.1 -0.0772 0.5 -0.1026 1.5 -0.1143 47 From the above analysis, increase in p resulted in a decrcasc in d. This was shown by the negative relationship between á and p. This is consistent with what has been reported in the literature (Hilt and Sumner, 1967; Rajkai et a1.,1996). Clearly there was a decrease in áwith increase tn yas expected, and as well as a decrease in d from Position I to Position 5. 0 increased with an increase in depth for similar 1ø. The coefficients listed in Table 3.1 confirm this. The observed variations (0 : f (ù) at diflerent positions and depths were due to the variability of the bulk density. The influence of other soil properties could not be assessed in the model due to the fact that sub-models were derived for individual positions and depths, where only single measurements of clay content, OC content andCøCOt content were made. 3.3.2.2 Multiple regression analysis for SftC The soil resistance data was fitted to the model shown in Eqn : [3.2] and the result of the multiple linear regression is given nTable 3.2. Fitting the main effects gave an nf : 0.74. The interactions accounted for a change in the variatiorl and therefore were considered in the model to increase the R2 to 0.86. A four-way interaction was present indicating that SrR was dependent on all four parameters: landscape position, soil depth, water content. and bulk density. There was a strong negative relationship between SR and 4 which generally increased from positions Rl to R5 and also increased with depth (Table 3.2). This implies that slight changes in á cause larger changes in SR at the top of the slope than at the bottonr, and that the limiting SrR at the top of the slope would be encountered at gteater water contents than at the bottom of the slope. 48 Table 3.2. Coefficients for log SR for each lcvcl of Position and Depth Position and depth C p e pxe Rll (Rl, depthl) -3.88 3.80 2.62 -6.66 Rl2 (Rl, depth 2) -16.38 12.38 37.49 -30.87 Rl3 (R1, depth 3) 1.53 1.14 -r7.52 4.47 R14 (Rl, depth 4) 1t6.37 -87.07 -291.56 215.57 Rl5 (R1, depth 5) 10.12 -5.82 -27.r3 13.93 R16 ß1. depth 6) 7.75 -3.68 -22.25 10.36 R21 -7.50 5.92 16.36 -14.97 R22 9.72 -4.83 -30.46 14.20 R23 -s.65 4.98 13.53 -14.23 F.24 -7.54 6.55 23.33 -22.02 R25 9.08 -5.60 -31.24 18.50 R26 19.81 -t2.58 -57.26 36.30 R3l -0.90 1.42- 2.23 -2.73 R32 -0.35 r.40 -13.54 2.39 R33 -0.30 1.38 -47.26 22.24 R34 -17.13 12.02 10.94 -13.44 R35 -13.53 9.88 15.34 -16.55 R36 19.75 -13.01 -79.78 49.73 R41 -12.54 8.60 52.77 -37.26 F..42 -2.51 2.50 8.82 -10.73 R43 -30.09 20.34 87.70 -62.56 R44 -4.14 3.33 -9.42 1.74 R45 -0.01 0.64 t6.66 -15.74 R46 3.s0 -1.90 -9.18 2.88 R51 (R5, depthl) -31.60 20.44 103.s6 -69.00 R52 (R5, depth 2) -54.31 32.99 208.t4 -127.53 R53 (R5, depth 3) -23.27 15.3s 7t.33 -50.00 R54 (R5, depth 4) -8.26 6.42 16.95 -17.84 R55 (R5, depth 5) -11.54 8.99 2.79 -9.24 R56 (R5, depth 6) -77.45 13.52 27.75 -2s.32 49 3.3.3 The influence of soil properties on the ll/RC and sfic The influence of soil properties on the IryRC and SÃC was determined using multiple linear regression analysis as shown in Tables 3.3 &3-4. 3.3.3.1 Influence of soil properties on IVRCs Soils with carbonates A similar response as for the non-carbonate soils was found except that 0 increased with a decrease trl OC, contrary to the results for the non-carbonate soils. This could be due to the fact that carbonates were encountered at deeper þers whete OC contents were very low and hence had less influence on water retention. Water content increased wlth CaCO1 but there does not appear to be any reason for this except that the presence of CøCOs coincided with an increase in clay content, so this effect could be a secondary influence due to the clay. Soils withCaCosgenerally had lower bulk densities than soils without any carbonates, and the greater the carbonate content, the lower the bulk density. The model explained 79%o of the variability. Soils without carbonates V/ater retention curves (0: f (y)) followed a common pattern (Table 3.3) with 0 positively correlated to both cþ content and OC content (Hall et al., 1977; Rawls and Brakensiek, 1989), and negatively correlated with p (Hill and Sumner, 1967; Bertz et a1.,1998). The model explained 8l% of the variability. 50 Table 3.3. Summary of multiple regression analyses of the water retention + function for soils with and without carbonates: log0= c * logty+ losp+ logclay logOC. Variable Coeffrcient Standard Error T-value P-value( 2Tail) of Mean 6.8 C -1.309 0.032 -40.765 < 0.0001 log -0.152 0.002 -83.794 < 0.0001 V < log p -0.288 0.079 -3'653 0.0001 < tog ctay 0.350 0.020 17.340 0.0001 < log OC -0.034 0.006 -5.260 0.0001 < log CaCOs 0.034 0.007 4.593 0.0001 (F": 1639, P < 0.0001; Adjusted RÍ : 0.81, N : 1972) Non-carbonate soils c -1.276 0.029 -44.242 < 0.0001 los V -0.168 0.002 -88.303 < 0.0001 log p -0.237 0.077 -3.066 0.002 < Iog clay 0.266 0.019 13.987 0.0001 < log OC 0.101 0.010 10.598 0.0001 (F :2014, P < 0.0001; Adjusted RÍ:0.82, N: 1798) Table 3.4. Summary of multiple regression analyses of the soil resistance curve for soils with and without carbonates: log sR : c + log0+ logp+ logclay + logOC. Variable Coefficient Standard Error T-value P-value (2Tarl) of Mean C -l c -3.322 0.150 -22.110 < 0.0001 log 0 -2.069 0.072 -28.589 < 0.0001 log clay 1.442 0.102 14.r l8 < 0.0001 tog OC -0.082 0.035 -2.374 0.0r8 : : çf : ztl,P < 0.0001; Adjusted RP 0.67, N 424) Non-carbonate soils c -2.707 0.165 -t6.377 < 0.0001 tos 0 -2.045 0-375 -33.234 < 0.0001 tog p 2.045 0.375 5.4s2 < 0.0001 log clay 0.483 0102 4.745 < 0.0001 log OC 0.513 0.052 9.774 < 0.0001 (F:289, P < 0.0001; Adjusted Ñ :0.73,N:420) 51 3.3.3.2lnfluence of soil properties on SACs Soils with carbonates Soil resistance curves (^sR : f (0)) also followed a common pattern (Table 3.4), but neither CaCOs nor pcontributed signifrcantly to the model. S,tR correlated positively with clay content (Canarache , 1990; da Silva et al., 1994), but negatively wlth OC, which was contrary to most studies (Spivey et al., 1986; da Silva and Kay, 1997). The relatio*hip, however, was comparatively weak (cf. Table 3.4), and may have been caused by the small variation and total content of OC. The model accounted for 670/o of the variability. Soils without carbonates In this case, soil resistance correlated positively wrth p, clay content and OC content (Table 3.4), which was also found by Canarache (1990) and Grant (1989)' The greater contents of clay and OC presumably increased cohesion (and thus SR), despite their effects on water content. The model explained 73%o of the variation. 3.3.4 Models proposed in the literature. 3.3.4.1 Models for IYRCs Results for the regression of 0 v. r¿ according to the model shown in Equation [3.8] for soils with and without carbonates are given in Table 3.5 and are all highly significant. The magnitudes of the coeffrcients were similar for non-carbonates and carbonate soils, indicating the water retention curves were similar according to Equation [3.8], which described >70% of the variation. 52 3.3.4.2 Models for SrRCs' The influence of á on ,Sft was similar for soils with or without carbonates, but for non-carbonate soils ,Sft increased with p (TabIe 3.6), which may be expected in typical granular soils (e.g. da Silva et a1.,1994). The opposite effect was observed for soils with carbonates, where,Sft increased with a decrease n p. Carbonates occurred mostly as hard granules in a weak skeletal matrix and were characterised by low p. 3.3.5 Least Limiting Water Range The LLIYRs for the different positions in the landscape and depth in the soil profiles are shown in Table 3.7. The values were calculated using models derived from Table 3.1 &,3.2 asthey had the highest R2 values. Perhaps the most significant observation to consider in these results is that the magnitude of LL\í(R as a function of depth was rather small for all positions in the landscape. No LLÍYR- value exceeded 0.116 cm3 crn3, and many were < 0.070 cm3 cÍL3. Because the range n LLWR - values was so small, interpretation of the variation n LLWR is somewhat difficult. No simple relationship between LLtryRand soil properties (e.g. clay content, P, OC, and CaCOj) could thus be found. Furthermore, the complexity of any interactions between the variables influencing LLWR was certainly not possible to isolate from the IVRCs and ,SftCs. Hence, the apparent increase nLLWRwith depth in the upper - slope profiles (Figure 3.9) should not be regarded with undue alarm. It could safely be stated that the LLWRs were uniformly low everywhere in the landscape. The upper timit of the LLII/R was invariably characterised by the water content at field capacity for all the positions and depths (Table 3.7). The relatively sandy 53 textures ensured that aeration was not a limiting factor in these soils. The lower limit was mostly characterised by the water content at 2NlPaSR. In 83% of the cases, soil strength was too high well before the classical wilting point. It could therefore be suggested that SR is the main limitation in these soils, and that management should aim to reduce soil strength by improving soil structure (e.g. retention of organic residues and minimised traffic). Table 3.5 Summary of multiple regression analyses of the soil water retention function for soil with and without carbonatesz log0: loga + b logry. Coeffrcient Magnitude Standard Error T-value P-value(2 Tail) Mean of Mean log a -0.827 0.004 -193.783 < 0.0001 b -0.153 0.002 -67.626 < 0.0001 (F :4573,P < 0.0001; Adjusted Rf :0.70, n:1972) Non-carbonates soils log a -0.924 0.004 -238.404 < 0.0001 b -0.167 0.002 -81.449 < 0.0001 (F :6634,P < 0.0001; Adjusted RP : 0.79, n: 1798) Table 3.6. Summary of multiple regression analysis of the soil resistance curve for soil with and without carbonatesz log SIR : log c+ d log 0 + e log p- Coeffrcient Magnitude Standard Error T-value P-value (2 Tail) Mean of Mean Soils with carbonates (16.8-4419 C kg-' soil) log c -0.856 0.073 -tt.69l < 0.0001 d -1.889 0.082 -23.029 < 0.0001 e -3.070 0.432 -7.lll < 0.0001 (F:268, P < 0.0001; Adjusted RÍ :0.56, N:424) Non-carbonate soils log c -1.641 0.095 -r7.293 < 0.0001 d -1.751 0.061 -28.768 < 0.0001 e 1.724 0.418 4.124 < 0.0001 (F:418, P < 0.0001; Adjusted * :0.67,N : 420) 54 LLWR (cm3cmi) 0.1 o.12 o.14 tl 0 0.o2 0.04 0.06 0.08 0 10 20 E (botto o - 9- R1 30 IL -r+- R2 o o 40 -*- R3 50 -€- R4 (top) 60 ìlê R5 70 80 ['igure 3.9. LLWR (cm3cm-3) as a function of depth for all the positions. ,T iil "!j 3.4 Conclusions 3.4.1 Prediction of WRC pedotransfer functions to describe the IYRC were successfully derived from soil properties. Most of the variability in d was accounted for by [, whereas the contributions of inherent soil properties, while statistically significant, were found to be small in magnitude. This result is similar to that of Williams et al., (1992) who found that the influence of OC was minimal in prediction of the IURC. In my study area, OC contents were relatively low (< l%) and may therefore not have occurred with suffrcient variation to impact significantly on water retention' t 55 3.4.2 Prediction of SRC By comparison to 4 inherent soil properties made minor (but statistically significant) contributions to the models predicting the SftC. The ef[ect of 0 and soil properties was consistent with that of most studies in the literature (e.g. Byrd and Cassel, 1980; Callebaut et al.,1985). The only deviation was the effect of CaCOs on the models. Soil with CøCOs resulted in a negative relationship between SR and ¿ although this was a secondary effect because CaCOs alone did not contribute significantly to the SftC. In soils with greater quantities of CaC03, however, this may need to be taken into account in developing pedotransfer functions for predicting the SftC' 3.4.3 Prediction of LLWR Derivation of pedotransfer functions to model WRC and SÀC from soil properties 1 Ìl despite 'tlj made it possible to calculate the LLWR in this texture - contrast soil complications arising from the occurrence of carbonates in some of the soil profiles. The presence of carbonates coincided with an increase in clay content marking the boundary of the textwe - contrast, and this made it diffrcult to attribute efflects unambiguously. Furthermore, while the overall magnitude of the LLWR - values was relatively small (S 0.116 cm3 cm-3), LLWR appeared to increase somewhat with depth in many of the profiles. A reduction n p as a function of depth may have been responsible for greater water contents at field capacity, which were never limited by I i restricted aeration. I The main limiting parameter in the study, however, was the SA, which exceeded 2 I MPa well before the wilting point and thus kept the LLIAR quite small. r 56 Table 3.7. LLWR for different position and depth' Code 0vo¡, 0pc 0sn 0wp LLWR (cm3 cm-') (cm'cm-') (cnt' crrr') (cnt' cm-') (cmr crrr') Rll 0.3r9 0.286 0.219 0.171 0.067 Rl2 0.330 0.294 0.221 0.178 0.073 Rl3 0.383 0.300 0.245 0.184 0.055 Rl4 0.400 0.311 0.254 0.195 0.057 Rl5 0.410 0.323 0.252 0.207 0.071 Rl6 0.391 0.346 0.300 0.230 0.046 R21 0.296 0.261 0.220 0.145 0.041 F.22 0.296 0.267 0.2r9 0.151 0.048 F.23 0.34s 0.272 0.187 0.156 0.08s R'24 0.368 0.282 0.181 0.166 0.101 R25 0.402 0.300 0.204 0.184 0.096 P.26 0.402 0.327 0.31 I 0.2t1 0.016 R3l 0.315 0.209 0. r55 0.093 0.054 R32 0.277 0.208 0.173 0.091 0.035 R33 0.300 0.208 0.134 0.092 0.074 R34 0.319 0.278 0.111 0.101 0.107 R35 0.341 0.233 0.087 0.1 17 0.116 R36 0.368 0.265 0.114 0.149 0.116 .'Ì R41 0.323 0.236 0.07s 0.120 0.1 16 tì,{ P!42 0.304 0.238 0.140 0.121 0.097 "f R43 0323 o.237 0.091 o.121 0.116 R44 0.338 o.246 0.076 0.130 0.116 R45 0.349 0.260 0.099 0.143 0.116 R46 0.360 0.289 0.09s 0.173 0.116 R51 0.2s8 0.221 0.207 0.105 0.014 R52 0.255 0.226 0.181 0.110 0.044 R53 0.285 0.228 o t43 0.111 0.085 R54 0.296 0.236 0.148 0.120 0.088 R55 0.315 0.251 0.182 0.134 0.069 R56 0.357 0.285 0.197 0.169 0.088 (NB. Deflrnition of codes is given in Section 3.2.2.1) t Ì ; 3 57 CHAPTER 4: SOIL WATER CONTENT VARIATION AND LLWR 4.1 Introduction The impact of changes in soil structure as influenced by cropping systems on plant growth is strongly influenced by water content. It is therefore important to study the temporal variation in soil water content because it afÊects other soil physical properties used in the characterisation of the LLWR. Changes in soil water content are dependent on prevailing climatic conditions (i.e. precipitation and temperature). For most uniform soils the probability that variations in water content are within or outside the LLWR depends on its magnitude. It is therefore logical to suggest that crops growing in soils with a naffow LLWR would be more wlnerable to high strength, drought and waterlogging than crops growing in soils with a wider LLWR :i rf (Kay, 1989). This may not necessarily be so, however, for soils whose properties lli 't vary with depth. For example, if the magnitude of LLI{R is greater in one part of the soil profile than another, opportunities for compensatory root growth in the region with greater LLWR may occur in periods of water stress. Being able to characterise the probability that the water content falls outside the LLll/R, P¿a1, ãs a function of the LLll/R in a given climate over long periods can be valuable as a management tool to assess risk of crop failure. This is imperative in Australian conditions, where plant available water in many soils is restricted by inherently hostile physical (e.g. low infiltration rates and hydraulic conductivities) and t I chemical (e.g. sodicity and salinity) properties. In addition, management practices ; r 58 leading to the destruction of the soil structure by, for examplc conventional tillage, further reduce the LLllIR and thus inctease Pon. Management practices modiff soil water content and its temporal variability and may have an eflect orr Por¡ irrespective of the magnitude of LLIAR (da Silva and Kay, lggT). For example, tillage systems that ret¿in residue (e.g. stubble retention) have is been reported to conserve soil water over longer periods compared to fallow. There also consistent spatial variation in soil water content between row and inter row positions (Van Wesenbeeck and Kachanoski, l98S)' Crop species due to their variation in water use effrciency (variation in rooting morphology) would systematically influence the soil water content throughout the profile. Shallow - rooted crops predominantly extract water from the top layers, while deep-rooted crops generally extract moisture throughout the soil profile. Farming systems increasingly employ both monocrop and intercrop treatments in order to maximise water use efficiency and minimise leaching of N-fertilisers. Both the choice of cropping pattern data are as well as the climate would be expectedto have alarge influence ortPou6 yet not available for such crops in relation to soil structure (as measured by LLIYR)- The objectives of this study were to: (i) determine whether da Silva's et al., (1997b) observation that the Pou¡ is inversely related to the magnitude of LLI{R in uniform soils, applies to a texture-contrast soil in a Mediteffanean climate (ii) identiff the primary factors contributing to Po¡ and (iii) evaluate the influence of some inter- cropping and mono-cropping treatments oî Poo¿ ï I J I 59 4.2 Materials and methods The study area was described in Chapter 3' 4.2.1 Meteorological data Campus Climatic data were obtained from the University of Adelaide, Roseworthy gauge' Weather Station. Rainfall was measured with a tipping-bucket rainfall 4.2.2 Measurement of soil water content the Soil water content was measured using the neutron scatter method, following at either end technique of Greacen (1931). Two neutron access tubes were installed tubes were of each of the 15 plots shown in Figure 3.1. All the neutron access 503 Hydroprobe was installed to a depth of 1.2nr, and a Cambell-Pacific Nuclear DR (16 s count time) at used to measure water contents. Neutron counts were measured growing 20,30,40, 50, 60, 80, 100, and 120 cm every week for the duration of the the season and season. As both the temporal changes in the soil water throughout values, the changes with the LLWR are based on relative, rather than absolute These calibrations supplied by the manufacturer were considered suffrcient' (1979)' calibrations (Appendix Table At.5) were obtained using the method of cull, 4.2.3 Po"t parameter' i'e The Poutwas characterised to distinguish the influence of the limiting as the the upper limit (P'u1upp"r)), or lower lkrit (Pout (tower))' Pou wãs calculated the LLIIR' number of measurements, ns , taken when the water content was outside divided by the total number of measurements (N:24): lle Pou, [4.11 N 60 More e¡rphasis was placed on characterisation of Pro¡ for the top 30 cm because this is where most of the roots were expected to be growing (Greacen and Russell, 1977; Schultz, 1977). 4.2.4 Statistical analYsis Statistical analysis of the data was carried out using Genstat 5, Release 4.1. Analysis of variance tests were used to establish the influence of cropping treatments, depth, and their interactions oîPou¡. Initial analyses were applied to inter-cropped versus mono-cropped treatments, before comparing individual treatments. A student t-test was used to analyse the main contributing factor to Poo1. 4.3 Results and Discussion 4.3.1 Depth effects The mean value of Pool for all treatments was invariably larger in the soil surface layers than it was at depth (Figure 4.1); Pou¡diminished rapidly with depth to about 40 cm, and then remained relatively constant. That is, Poo42q cd t Poa4¿0,60, 80 cn¡ but Pou(30 (p < 0.01). .Ãlso Pou4jï c@ wãs Pou(20 cm) wãs not signific antly greater than "d not significantly different from PoøØ0,60,80 cm). This could be expected in situations where most of the plant roots are restricted to the top 30 cn¡ and for this reason' further analysis will be focussed on the top 30 cm region' 6t o . out 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 10 20 30 840 ; otso 60 70 80 Figure 4.1. Mean Pout as a function of depth for all treatments (horizontal bars represent +/- std deviations). 4.3.2 Surface soil (0 - 30 cm) effects The soil water content variations throughout the whole growing season in relation to the LLltr1R at 20 cm and 30 cm are shown in Figures 4.2a and 4.3a. Detailed characterisation of Po,rfor both depths is given in Table 4.1. The upper limit of the LLWR at both depths generally corresponded to water contents at "field capacity" (tf^: -10 kPa), and the lower limit corresponded with water contents where penetration resistance exceeded 2 MPa. This meant that while there were no aeration problems when the soil was wet, there were considerable restrictions to water availability at the dry-end simply due to high soil strength. This is illustrated by the fact that at20 cmPout(ower)was always significantly greater (p:0.001) thanPou¡¡ooo",¡ at20 cm(same at 30 cm but to a lesser degree (p < 0.1)). The LLWR at 20 cm (0'05 62 cm3 cm'3) was slightly smaller than at 30 srn (0.07 cÑ cm3). The water contcnt pcaks in Figure 4.2a,b,4.34b, coincided with major rainfall events (Figure 4.4). Table 4.1. Details of Poul for all treatments at 20 cm and 30 cm. 2O cm 30 cm Treatment Pour Pout Por, Pout Pout Pout L 0.75 0.25 0.50 0.71 0.25 0.46 o 0.63 0.30 0.33 0.58 0.33 0.25 w 0.66 0.24 0.42 0.54 0.29 0.25 w/L 0.71 0.17 0.54 0.67 0.17 0.50 olL 0.75 0.25 0.50 0.75 0.29 0.46 4.3.3 Treatment effects Whilst not reflected in the physical data, visual diflerences were observed between the various treatments. An overall analysis of variance for all the depths and treatments showed there were no significant þ : 0.05) treatment eflects on Prrr (this was due to the high variability of Poø between treatment replicates, and the low replication (3) that was possible for this study). For the 20 and 30 cm depths, however, Pou¡ for inter-cropped treatments tended to be greater than that for the rnono- cropped treatments. While it is not strictly valid to proceed with treatment comparisons, the LSD (p: 0.05) indicated that the O and OlLtteatments were significantly different (O/L had a higher Pout than the O treatment). Generally the inter-cropped treatments used more water than the other treatments, as expected, resulting in quicker depletion, even after rains (Figure 4.2b and 4.3b). Differences in residual water content for surface soils at the beginning of the season suggested that planting date could be crucial in minimising the impact of Po6 on growth and yield (Figure 4.2aand4.3a). 63 035 ---()_- L ----''l-----w ----E-----'o _+_wL +?^o.30 ts o/L (J limit LLWR E r llo.zs - Lower limit LLWR oc, -Upper o o.zo $ III I¡ aú 3 o ør 0.'15 Planting Harvesting + + 0.10 (¡) (¡) (r) N N l\) N) o o (o @ @ { L L a, z ot tc o o() o ol a= T (o (o "l (o (o (o ?(o (o (o o (o @ (I) o (Þ @ o @ Date Figure 4.2a. Soit water content variability at20 cm in relation to the LL\YRfor thé 199S growing season for all treatments (individual treatments shown). 0.35 ---è-L -----B- ---Mono-croPPed ___l- lnter_cropped 0.30 limit LLWR E o r Lovuer limit LLWR E o o.25 otr c o o o.20 o Ir IT ITT rII - "r aú ='o |t, 0 1 5 Planting Harvesting + 0.'t0 (¡) (/) (/) t\) N N NJ o o (o @ @ --¡ L L @ o z o, c, c c o () o tf 0) (o ! (o *l= = (o (o f(o o (f) (o ct (o (o dt o @ @ Date G' @ Figure 4.2b. Soil water content variabitity at 20 cm in relation to the LLWR for the 1998 growing season (treatments grouped as either inter-cropped or mono- cropped) 64 rn ¡r \o ¡r 27-Nov-98 27-Nov-98 €Ö ct € v.( Et v( 1t c, ËE o s JJ r\Ê =á,JJ o EtocL rì ¡- .€ ii È { E EE o \ì rä {o L! ó? o é)éq) oo 2&Oct-98 é) 28-Oct-98 -E'ij J I CL> bs ìo Êi t --¡Etr J -¡lolõ51 I l;l s3 I fql ¡ ã? lrl Ëq) I 28-Sep-98 l,'1il: I 28-Sep-98 crt ê)r 0) trF cã != rt Éo cj|(Ð eö 29-Aug-98 I 29-Aug-98 (f) br I ó*¡ o +.õ tE cËo t o ;r C) ô *È B5 ;'i L I 30-Jul-98 ¡ào 30-Jul-98 =8-oL .Fe L6 €Ë= I Ë9 Þg I ¡Ë +¡ 'F C)- I 30-Jun-98 .Ðs 30-Jun-98 e= t T 9-=ij O-ã'.> ¿¡á V rt) 31-M >ä 31-May-98 u: tt ED +àO =è¡ '=É Ê oÊ + a! Ø';. õ'a o- 6l ¡i Ë¡¡P à¡ (1 à¡ 1 -May-98 !læ;' 1 -May-98 sæ |r) o t¡) o rl) o |r) o .. o\ (Ð (Ð c{ c! g. E8ÑR Ëo\ ci o o o o o Ee ci ooo cj ci -!ãA ¡ale/$ è0Ð- o ¡a¡B¡^ à0(Ð Çtuc¡uc) lua¡uoc llog Çnrcrutc) ¡ualuoc llos E5 b The luue¡re i¡ the inter-cropped trcatment, for example, came out of summer dormancy in early May and began to deplete the soil water before any mono-crops had been planted. The mono-crops could thus not use the water available to them while the temperatures were generally higher. On the other hand, the inter-cropped treatments quickly experienced high strength as ádropped outside LLWR. 4.3.4 LLIryR rîd Pout Considering data for the entire profile, Poot deqeased as LLll/R increased' and '-Poorincreased variations withLLIYR (Figure 4.5). It should be noted that the range ¡¡^LLryRvalues was rather small (0.05 - 0.09 cm3crn3), so variation within this range might be expected. Nevertheless, the data would appear to support the hypothesis of Kay (1989) that crops growing in soils with a naffow LLWR have greatet Poøtlnart crops growïr on soils with large LLWR. Whether this relationship is linear, or curvi- linear as in Figure 4.5,cannot be confidently determined fromthe data shown' 4.4 Conclusions Da Silva et al's., (1989) observation that Ponincreases with decreasng LLÚltR is supported by the results, despite the fact that the small range n LLll/R may account for some of the variation in the Pon data. There was a significant decrease ill'Pootwith depth and this coincided with an increase n LLWR. P¡u¡was large at the surface due to abundance of roots and evaporatiorl and the high strength that resulted. There was no treatment effect oî Poulin the surface soils (20 cm and 30 cm), except that O/L had a significantly greater Poottlnnthe O treatment. This is somewhat surprising in view of the contrasting root habits of mono- and inter-cropping, but greater replication may provide a clearer picture in future. Visual differences were certainly observed between the different treatments, even though these were not reflected in the physical data. 66 35 30 25 E 20 E G .s .U 15 É, 10 5 0 (, l\) N) N) N) o (t (, (o @ \¡ @ I o oI I I L C- U) z c c o ôo o l¡) o) l (o ! I I (o I (o (0I (o f(o (o o (o (o o o o @ @ o @ Date Figure 4.4. Rainfall records for the 1998 growing season 080 0.70 40 cm 0.60 6O cm 20 cm 0.50 30 cm I o 040 o. Po,¡ = 476LLUÆ.2 + æLLVttR -1 R2 0.87 0.30 = o.20 0.10 80 cm 0.00 0.04 0.045 0.05 0.0s6 0.06 0.065 0.07 0.075 0.08 0.085 0.09 LLWR Figure 4.5. Pon and its variation as a function oI LLWR. 67 CHAPTER 5: INFLUENCE OF LLWR ON CROP GRO\ilTH AND YIELD RESPONSE 5.1 Introduction under This study was initiated from the work of da Silva and Kay (1996), who found temperate climatic conditions that the rate of shoot growth of maize was strongly under correlated withLLWR and Pou, The extent to which this relationship was valid different soil and climatic conditions was considered worthy of further investigation. in Studies on the LLWR carried out by da Silva and Kay (1996) were conducted some climates where soil moisture conditions were not especially extreme (Poot:0 in soils). In semi-arid to arid environments where P¿ur commonly exceeds 0.8 and where inherent soil properties can exacerbate moisture limitations, it is critical to examine the impact of PourandLL\VR on plant response. An existing trial at the Roseworthy Campus of the University of Adelaide was chosen for this study (same as for chapters 3 e, Ð because it included araîge of crops with diflerent rooting patterns and thus different patterns of soil water depletion. The growth objective of the study was to evaluate the eflect of Pou¡ and LLWR on crop (dry-matter) and quality (grain yield, grain weight, and grain protein content)' 5.2 Materials and Methods 5.2.1 Treatment Design The crops (see below) were established in l5m x 5m plots in a randomised block design with seven treatments replicated five times (Figure 3.1, p. 35)' The plots were 68 laid out across and down thc slopc to include any variability in soil properties along the toposequence (Figure 3.1) All data were subjected to an analysis of variance test using Genstat 5, Release 4-1. Initial analyses were applied to all inter-cropped versus all mono-cropped treatments, before comparing individual treatments. 5.2.2 Crop Management Crops that were growïr included two varieties of wheat (durum(W(d)) and hard (V(h), Marloo oats (O) and Genesis lucerne (L), each established as either a mono- crop or inter-crop with lucerne. Lucerne was planted in autumn (April) 1996, and cereals were sown into the existing lucerne stand during 1998, using a Connor Shea- sod seeder with narrow points to establish the seeds at a depth of 10-30mm. Cereal planting occurred on2ll05l98 and harvesting took place on 17l7ll98. Management of the plots was similar to that practiced on most South Australian grain - legume farms, with initial application of triple superphosphate + 5%o Zn fertilisers at arate of 100 kgltta at the time of seeding. Seeding was conducted using zero tillage. 4Llha Roundup+ MCPA herbicides were used to kill lucerne on the 23112197, followed by 600m1 Wþout and 400m1 Lontrel on the 3013198 in plots where cereals were planted as mono-crops. 2Llha Sprayseed was used to retard lucerne on the l2l5lg8, in inter-cropped plots. Seeding rates for the different crops are shown in Table 5.1. 69 Tablc 5.1 Sccding rntes of croPs. rop Varietv Seedrn ù Wheat (Hard) Triticum aestivum 88 Wheat (Durum) Triticum durum 96 Oats (Marloo) Avenafatua t0 Lucerne (Genesis) Medica sativa 8 5.2.3 Agronomic parameters 5.2.3.1 Establishment counts and dry matter yield Establishment counts and lucerne density were measured approximately 3 weeks after seeding. Average dry matter yield of lucerne and crops was determined at tillering (2118198) and anthesis (l2ll0l9S) by randomly placing a quadrat (50 cm x 50 cm) at two different locations on a plot and cutting the plants within the area to lcm above the ground surface. The dry matter weights were recorded after 48 hours of drying in an oven at 80'C. 5.2.3.2 Crop yields Harvesting of both wheat varieties and oats was done using a KEW plot-harvester and the total mass of grain from each plot determined. The yields from mono- and inter- cropped treatments l¡/ere measured separately. Grain size was determined by measuring the weight of 2 samples of 1000 grains for each treatment. 5.2.3.3 Protein Protein content was calculated from the total nitrogen content of grain samples ground to a fine powder in a CYCLOTEC 1093 Sample Mill. A 0.25 g sub-sample was digested using 4 cm3 sulphuric acid and 2 crÊ ZOy" w/v hydrogen peroxide. The sample was then placed in a digestion block using the following three settings i) 150' C, l0 min. ramp time plus 20 rrlln., ii) 250" C, l0 min. ramp time plus 20 min., iii) 70 300" C, 10 min. ramp time plus 2 hrs. The digest was allowed to cool to room content temperature, and the volume made up to 50 cm3, mixed, and the total nitrogen was determined using an autoanalyser. The following multipliers for calculation of protein contents were used: 5.7 x total N for wheat and 6'25 x total N for oats (Tkachuk, 1969) 5.2.3.4 Po"t 20 and The influenc e of Pou¡on cfop parameters was examined for soil samples from The 30cm depths where most of the root growth was expected (Holloway, 1996). following periods were used in the calculation of Poui i) for dry matter at tillering: for May - August, 1998; ü) for dry matter at anthesis: August - October, 1998 and iii) grain yield: May - October, 1998. 5.3 Results and l)iscussion 5.3.1 Agronomic measurements Lucerne density was unaffected by the intercropping, with mean density of 30'4 plants/m2 with crops, and 29 plantsh* in monoculture. Dry matter yield was than significantly greater (p < 0.05) for all the mono-cropped treatments (O and V/(h) for the inter-cropped treatments (o+L and w(h)+L) at both tillering and anthesis matter stages of growth. At tillering there was SlYo and 69Vo rcduction in mean dry yield for the O+L and W(h)+f treatments, respectively' compafed to the mono- in cropped treatments (Figure 5.1). At anthesis there was anï5o/o and 81% reduction mono- mean dry matter for O+L and W(h)+L treatments respectively relative to their cropped equivalents (Figure 5.2). Dry matter measurements of lucerne at crop tillering and crop anthesis are given in Appendix Table A7.7. 7l 3.0 a o b Ð z.o I I .9 t 690/o reduction I 1o/o reduction c o= 1.0 + + c I I 0.0 o O+L w(h) W(h)+L Figure 5.1. Dry matter measurements at tillering stage (different letters are significantatp=0.05). The dry matter yield-reductions associated with lucerne inter-cropping were almost certainly due to competition for water, nutrients and light (Turner, 1966). The extent to which these acted may be determined from the impact of treatments on grain yield, size and protein content. 12.O Iq 10.0 a .E Ê 8.0 85% reduction I I 6.0 .9 I 1.o/o red 4.O =o I b + b 2.O I 0.0 o O+L w(h) W(h)+L Figure 5.2. Dry matter measurements at anthesis tage (different letters are significant at p = 0.05). Figure 5.3, for example, shows that the inter-cropped treatments (O+L, W(d)+L, and ï/(h)+L) had consistently lower grain yields than the mono-cropped stands (p < 0.05). 72 Inter-cropping caused yield reductions in the order of 79%, - S$o/n relative to the mono-cropped plots. To some extent, the differences in grain yield occurred because germination and emergence of grain seedlings was poorer on the inter-cropped treatments (Fussi, 1998). This was likely caused by the drier conditions on these plots, and thus somewhat lower plant densities. 4.s0 e a I 3.50 ! a! b 2.50 so 88% a4% t 7s% reduction .E 1.50 reduct¡on reduction c c o c 0.50 .L ra, ,t ì¿ I w¿hì t^, lhl+l -0.50 ô+l W ldì Figure 5.3. Grain yietd for inter- and monocropped treatments (different letters significant at p = 0.05). Furthermore, grain size was not significantly affected by the different treatments (Figure 5.4), contrary to the work of Fussi (1998), who found significant reductions in 50 L o40 ! I ! 't ¡o I ! c 'E 20 ctt o 310c, 0 o o+L w(d) w(d)+L w(h) w(h)+L Figure 5.4. 1000 grain weight (g) for the different treatments. 13 grain size for thc intcr-cropped treatments as compared to the mono-cropped stancls in an adjacent trial during the previous season. Grain size depends largely on the effrciency of photosynthetic processes that convert dry matter into grain yield. Even with signiflrcantly higher dry matter yields observed in the mono-cropped stands, however, this was not reflected in the grain size, and the reason for this was not obvious from the study. Finally, no significant differences were observed in the protein content of the irrter- cropped and mono-cropped treatments (Figure 5.5). This response is contrary to what has been reported in Fussi's (1998) earlier study, and in the literature where significant increases in nitrogen þrotein) were observed when cereals were inter- cropped with legumes (Dalal, 1974). This is particularly true for situations where severe water stress is experienced. In situations where water was non-limiting and, with adequate nitrogen, grain yield increased without any change in protein content (Terman et a1.,1969 and Turner, 1966). Fussi's (1998) greater protein contents may have resulted from water stresses experienced by the inter-cropped treatments, because !997 was a drier - than - average year at anthesis and afterward to harvest time at Roseworthy (University of Adelaide, Roseworthy V/eather Station, 1998). By comparison, during 1998 the growing season was relatively wet (c/ Figure 4-4), and even thoughPoo, exceeded 0.6 in most of the treatments, this may not have generated suffrcient water stress to affect grain protein. 74 18 16 J I I 14 ¿ s I I c 12 o o 10 È I c aú 6 o 4 2 0 o o+L W(d) W(d)+L w(h) w(H)+L Figure 5.5. Grain protein content ("/fi for mono-cropped (O' W(d)'W(h)) and inter-cropped (O+L, W(d)+L, W(h)+L) treatments. .1 4 H i,,: 3.5 wo ! Yield = 2E+11a1604 (úó R2 = 0.76 Ð z.s oo I(l,t .E 1.5 (ú o1 o+L W+L 0.5 o 0 0.180 0.185 0.190 0.195 0.200 0.205 0.210 0.215 e cm3cm3 Figure 5.6. Grain yield as a function of average volumetric water contentrO (cm3 for the whole growing season. ".-'), ! 75 4.00 w(h)r 3.50 3o cm w0)o 20 cm Grain Yield 0.02181655 3.00 = 21 (ú Grain Yield = O.OBP out'7 R2 0.51 2.50 = Oo r, Or R2 = 0.59 o 2.00 (ú r.50 o r.00 O+L O+L 0.50 W(h)+¡ W(h)+¡. 0.00 0.40 0.50 0.60 0.70 0.80 E' , ottl Figure 5.7. Grain yield as a function of Po',¡ at 20 & 30 cm. 5.3.2 Crop responses to soil lYater content during the growing season 't tl,t Despite the implication from the previous section (5.3.1) that soil water was not ,rj I suffrciently restricted to influence grain protein content, various indicators suggest the cereals nevertheless responded to water stress. For example, the lower grain yields for the inter-cropped treatments correlated inversely with the overall average volumetric water content for the growing season (Figure 5.6). More specifically, the impact of water stress can be seen to be strongly related to Pou, in the top 30 cm (Figure 5.7). Furthermore, dry matter yields during the growing season responded strongly to Poon For example, at both tillering and anthesis stages of growth, dry matter yield of I cereals decreased as Pou¡ at 20 and 30 cm increased (Figure 5.8 & 5.9). Negative power functions appeü to describe these relations well, with the exponents at anthesis r 76 bcing co¡siderably larger than at tillering. This, of courss, simply reflectecl the differences in water demand at the different growth stages. There are certain crucial stages of growth that have more bearing on the overall yield than others. In a study conducted by Turner (1966) it was observed that water stress experienced by wheat plants beyond ear emergence stage increased grain yield, whereas water stress earlier than that reduced ear emergence and fertility of any ears that emerged. 5.4 Conclusions Although it was observed in Chapter 4that Poolbetween the inter-cropped treatments and mono-cropped treatments was not significantly different, dry matter yield and grain yield were significantly greater in mono-cropped treatments (p < 0.05). Yield differences could be attributed to competition for at least water by inter-cropped ,) il treatments ¿rs compared to mono-cropped stands, resulting in quicker drying of the 'lü i soil and hence greater Psolandtherefore more limiting conditions as a result of high soil strength. Increase tn Port always resulted in a decrease in both dry matter yield and grain yield. This was expected because crops experiencing more limitations generally do not perform as well as those with fewer limitations. High soil strength in inter-cropped treatments restricted roots from exploring large volume of soil for the necessary nutrients and water and resulted in a reduction in both dry matter yield and grain yield. No significant differences were observed for grain size and protein content between the inter-cropped and mono-cropped treatments. This could be I r 77 14 Anthesis 12 oo DM = o.26Pour-1318 l0 R2 = 0.85 (ft ow p I Tillering o DM 0.12Por-tu' F 6 = =o R2 = 0.78 4 o w(h) o+L 2 o I W(h)+L o+L W(h)+¡ 0 0.90 0.40 0.50 0.60 0.70 0.80 P ot¡ Figure 5.8. Dry matter yield at tillering and anthesis as a function of Poutat20 cm. r.{ rlt '! 14 12 Anthesis 10 -u t' G DM = 0.1 1P or, I It illering R2 = 0.87 o 6 = O-14P out R2 = 0.48 o= 4 w(h) O+L 2 o O+L W(h)+¡ 0 0.40 0.50 0.60 0.70 0.80 0.90 P ou, yield at tillering and anthesis as a function of Pou 30 Figure 5.9. Dry mâtter ^t cm. I t 78 explained by the fact the small variation of the P¿z¡ values between the treatments. Significant differences may have been observed in a much drier year, (e.g. 1997) where competition for moisture would have been high in inter-cropped treatments as compared to mono-cropped treatments (cl Fussi, 1998). An implication of this work is that Poot might better explain short-term crop growth responses than it does final grain yield. This is well supported by the good correlations between P¿y¡ aîd dry matter at tillering and anthesis as compared to that of the overall grain yield. l I I 79 CHAPTER 6: ST]MMARY & GENERAL DISCUSSION 6.1 Conclusions 6.1.1 Pedotransfer functions were developed using data collected from surface and subsurface soils firstly to predict water retention curves, WRC, soil resistance curves, SRC, and then to calculate the Least Limiting Water Range, LLI|aR. Models were either taken from the literature or else developed using soil and landscape properties (e.g. position in the landscape, depth in the soil profile, and inherent soil properties such as bulk density,Yo clay,Yo organtc carbon, ando/o carbonates). The variability accounted for by the models in the literature was considered satisfactory, but never as effective as that from custom-built models. Models that take into account relevant soil conditions are thus superior to he more general models found in the literature. Using a model that is effectively 'taken off the literature shell is therefore to be done with caution, particularly where soil properties exhibit significant spatial variability laterally and with depth. 6.1.2 The LLWR increased to some extent with depth in many parts of the landscape, and while this increase was relatively small it was attributed to increasing clay contents and decreasing bulk densities with depth. This suggests that attention to subsoil root- zone constraints (which limit water uptake by plants) might signihcantly increase crop yields on texture-contrast soils, which sufler water stress in the surface horizons during Spring when thermal conditions are considered ideal for crop growth. 80 6.1.3 One of thc confounding factors in this study was the coincident increase n LLll/R with increase tnCøCO¡ This meant tlnt aconclusive analysis of the relation between clay content, bulk density and LLIAR could not be attempted. Despite the lack of a mechanistic connection between the soil properties examined here, however, the models developed to predict LLWRmay nevertheless be useful in soils where these properties exhibit a more pronounced expression(e.g.highly calcareous, texture- contrast soils). There are plenty of examples of these soils in the mediterranean regions of southern Australia. 6.1.4 Soil resistance to penetration was the main factor limiting the amount of plant available water in the soils of this study. There were no significant problems with soil aerution, so that the upper limit of the LLWR was always characterised by the water content at Field Capacity. One might speculate that, in drier climates, the success of native plants in high-strength soils may lie in their ability to use water quickly, before it drains to Field Capacity. Selecting characteristics from native species to breed crops capable of growing well in hard soils might thus involve evaluating their performance in wet soils. Desirable plants might draw the soil water content down from saturation more rapidly. 6.1.5 This study confirmed that Poøfor a duplex soil increased with decreasing LLWR, n a strong curvilinear relationship. In terms of dry matter yield, this was more significant in the later stages of crop growth (anthesis) than earlier in the growing season (tillering), primarily because the more mature plants had greater demand for water, and there was less frequent rainfall. 81 6.L6 Dry matter and grain yields were significantly higher in mono-cropped treatments compared to inter-cropped treatments, and this was directly related to the magnitude of Poo¡ for the different management practices. Increases in Poø in the top 20 and 30 cm resulted in lower dry matter and grain yields, and this correlated strongly with the average seasonal soil water content for each treatment. 6.2 Implications for future research 6.2.1 The fact that the custom-built models developed here invariably were found to be superior to any taken from the literature, calls into question the actual transferability of pedotransfer functions. To be fair, however, the application of models from the literature produced reasonqble predictions in this study. Obviously the utility of a given pedotransfer function must be evaluated in relation to the accuracy required of a given prediction and also in relation to the resources available for collecting data for custom-built models. However, considerably more resources would have been required in this study to collect the necessary datato make any claims about the degree of generality of the models produced here. Work thus needs to be done on other calcareous duplex soils to determine how transferable the data from this study are, and thus how useful this work has been. 6.2.2 This study was conducted during ayear where climatic conditions were considered to be relatively 'aveÍage' for the region. The previous year, 1997, was substantially drier at crucial stages of plant growth, and Fussi's (1998) thesis indicates plant responses differed considerably. It would therefore be important to observe the effects of the various treatments (mono- y. inter-cropping) during a wetter year. This 82 could be achieved in a shorter time period pcrhaps by using irrigation as a variable, which would have the advantage of not limiting water supply by the inter-cropping treatments throughout the entire soil profile, and testing the efÊects of different Poois at different depths on grain yield and quality. 6.2.3 In designing this experiment, I did not anticipatethatthe presence of CaCOs would coincide with the increase in clay content at the boundary of the texture-contrast in the soil profile. This seriously confounded an analysis of the effects of increasipg cþ content on the LLWR. Further work on texture-contrast soils should exclude CøCOs as a variable, or at least ensure that it does not coincide with other variables of consequenceto LLÍYR 6.2.4 Pedotransfer functions are developed most importantly to guide land-users to assess which soil properties to improve to better manage their land. Therefore, variables that are greatly affected by management such as organic carbon content and relative bulk density should be included in their formulation. In the present study organic carbon levels were relatively low throughout the landscape and therefore had minimal impact on the derived functions. Relative bulk densities were not determined due to time constraints and also because a single tillage system was used. Further studies should be directed to areas where there is a range of organic carbon contents to establish their eflects on pedotransfer functions and also the effects of different tillage systems by determination of relative bulk densities. 83 CHAPTER 7: APPENDICES 84 Appendix Table A7 ,l; Chemical properties of soils in the study area (H. Reimers, unpublished). Depth pH pH EC l:5 Available Available SO¿ Boron cm HzO CaCh ds/m P K mdkg m/kg m/kg múke 0-8 7.2 6.7 o.12 33 327 l9 1.2 8-30 8.2 7.5 0.13 6 333 4.8 2.3 30-60 8.9 I 0.15 2 252 5.2 5.4 60-120 9,7 8.4 0.43 I 423 l4 20 120-145 9.8 8.5 0.59 2 49't 30 23 t45-200 9.7 8.4 o-'t4 I 5r4 49 23 Depth Trace:.Elements rng/kg cEc Exchangeable Cations cm (EDTÀ) cmol cmol(+)/kg Cu Fe Zà Mn (+)/ke Ca Mg Na K ESP 0-8 1.2 58 2.3 t7 6.9 4.3 t.4 0.1I 0.78 1.6 8-30 I 33 1.2 22 19.6 10.2 6.3 0.41 l.l 2.1 30-60 0.69 4 1.9 l.l t2.6 6;1 6.3 0.s8 0.69 4.6 60-t20 0.34 3.7 2.t 0.5 13.1 2.7 6.8 4.2 l.l 32.2 t20-t45 0.3 5.2 2.1 1.6 r4.2 1.8 6.1 6.3 1.2 44.2 145-200 o.32 4.5 1 1.7 15.7 2.6 6.4 7.6 1.2 48.3 85 Appendix Table Ã7.2i|Aain effects analysis of variance table for the water retention curves (WRC). Factors Df SumSq Mean Sq F-value Pr(F) P 4 3.463 0.866 8t7.487 <0.001 D 5 1.535 0.307 289.893 <0.001 v 7 35.660 5.094 4810.513 <0.001 p 1 0.110 0.110 103.683 <0.001 Residuals 374 r 3.962 0.001 R'= 0.91 Appendix Table Ai .32 Full analysis of variance table for the water retention curves (wRC) Factors Df SumSq Mean Sq F-value Pr(F) P 4 3.463 0.866 1376.2r9 <0.001 D 5 1.535 0.307 488.O27 <0.001 V 7 35.660 5.094 8098.385 <0.001 p I 0.110 0.110 r74.547 <0.001 KD 20 0.5r2 0.026 40.693 <0.001 Pxty 28 0.41t 0.015 23.673 <0.001 Dxv 35 0.210 0.006 9.536 <0.001 Pxp 4 0.110 0.027 43.626 <0.001 Dxp 5 0.028 0.006 8.951 <0.001 vxp 7 0.138 0.020 3r.326 <0.001 hrDxty r40 0.203 0.001 2.306 <0.001 PxDxp 20 0.157 0.008 12.500 <0.001 PxVxp 28 0.038 0.001 2.r52 <0.001 Dxtyxp 35 0.028 0.001 r.283 0.t24 h 86 Appendix Table Full analysis of variance table for the soil resistance curves ^7.4. (sRC). Pr(F) Factors Df Mean F-value 38.001 <0.001 P 4 24.667 6.167 31.77r <0.001 D 5 25.7',18 5.156 3423.630 <0.001 0 1 5)5.)ô0 55s.s66 19.429 <0.001 p 1 3.153 3.t53 10.070 <0.001 PxD 20 32.68r r.634 30.762 <0.001 Pxe 4 19.968 4.992 2.479 0.031 Dx0 5 2.0rr 0.402 4.r20 0.003 Pxp 4 2.674 0.669 8.943 <0.001 Dxp 5 7.257 t.45r z.5JZ 14.372 <0.001 vxp 1 2332 3.743 <0.001 PxDx0 20 9.607 0.607 12.500 <0.001 PxDxp 20 0.157 0.008 5.054 0.001 Px0xp 4 3.281 0.820 3.462 0.004 Dx0xP 5 2.809 0.562 2.086 0.004 hrDx0xP 20 6.769 0.339 Residuals 721 117.000 0.162 = 0.86 87 û æ ra tt) q) a a €É) Probe Unit 5132 Probe Unit 6835 ¡¡ È Depth cms Slope lntercept Slope Intercept L 20 0.001559 -1.007 0.002962 -1.0726 (¡) q) 30 0.001559 -1.007 0.002962 -1.0726 1.357 L 40 0.001604 1.425 0.003049 50 0.001507 3.33 1 0.002863 3.2675 E 60 0. 00 1 458 3.331 0.002771 3.3132 80 0.001321 4.18 0.00251 4.1236 á ¡r 100 0.001363 3.386 0.002591 3.3284 120 0.001448 -0.201 0.002752 -0.2618 U ìn Þ- I 3 x E q) Èe ø\ æ o aÀ G É d Allvalues are in m Mg Na K P S AI cd F Fe Mn B Cu Mo Co Ni Zn Ca < q) w(h) 33 24 < 1.8 11.2 < 1.1 1.6 < 1.6 24 330 1120 39 5000 3200 1490 36 0.46 çh Notes: 1. Samples were digested with nitric acid and analysed by lnductively Coupled è0 Plasma Atomic Emission Spectrometry (ICPAES). t 2. For nitric acid digests: a a. The Fe values are indicative only as the nitric acid digest does râ not always give total recovery. The recovery % varies between 60 - 100% depending on samPle tYPe. b, The Alvalues are indicative values as a nitric acid digest does not give q) a total recovery. q) 3. The limit of determination is calculated as 10 X the standard deviation of the blank, I 4. Symbols < indicates the result is less than the limit of detection of determination of the method Fr W(h) indicates hard wheat; O indicates Oats \o F' q) ¡ F-( X E q) a eding: Mean crop establishment counts 3 weeks after se w(h) o w(h)+L O+L 108.3 78.5 99'5 70 Lucerne dry matter production during growlng season: Tillering Anthesis lntercropped W(h) 205.5 105'9 99'7 GI lntercroPPed O 241 .1 131 174'7 I Lucerne onlY .7 o ¡r U ì- t- c) Er ¡< Ð a Appendix X'igure A1: Volumetric water content, 0 (cm3/cm3¡, as a function of applied levels of matric potential, Y (see Section 3.2.4.1) for all positions and depths in the landscape. 0 (cm3 cm'3) oooooo oooooo oooooo -NO¡60 :ir¿ri¿Ào tÀOO (ñ) æ(D cEl @ (IIÐ tffi ffi olE) CEÐ (D(!O {IEI m CE ú crc (ID CTE o@o G' CE cn!¡ (dD crñ @ (D clxün (iED @lErt(¡ C) dÐo CDûD o @ GD oæ ooo (rm øñ N ED (re æo ED (Ð ÜID CD d)o rmf, cEl) cræ CIE) -D m CIIID cEn o {D oo GID m o (D crt (DO G' GTDO o (@ rE @ .rTI CIID @ T' Ú lm CED e a - @ (ID G o dt (ID IED o @ (D æ (r (u¡ @o (¡Ð ¡ o CED ô o o(I@D .lilì (E) o@ G u, drrÐ oo @ cú @ o¡![D(D ilo (I arrm @D @o (Erc (t GID (ÍD ffio ID - c¡D cE! rEn O (r ú ù cfD 6 mo GID (D (D (ID (EIIIID G@ (E) @ drrrÐ rñtD cro (-! mÛno ÍEID oo (@ CED rùFTlo @D oo EIÐ qETD @aDc) cfrt ocD lM CED O l,l.mo cñt -o irm r¡E)oo @D orD oÐ (Tllrc (D (Ð@ (rD (D@ IID @CD o dDo o OIDO o @ (EIICDO ocuÐ arT¡ú @qD m c¡E @ GD @IEIO cilD Ga¡ f@o @ cffr@ m m IIIÐ lm CEIÞCD crÐ (IIID GE (IfD (I¡ @)o@ m olD EI, cD@ rEì3 Cf'ñ m -) (ID CD mtft r'ì ûmdÐ oooogg ooooQo -NOÀOO ¡NOàOø 91 (rl o\ 05 0 5 0.1 0-2 03 04 o. t o.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 . :..¡ir¡' .'.. @ ''a(c :, ' : ts,fìI1 ¡Ë i,ri: ÞE? : ::-it:: ø ir lilr; . r:l : : .Þqitr-.T,r:¡,:ti¡ii;rirlEì:: L:!:::r ¡r; $ '-t 2 (:,' ,''í ,Fj .r.-r) (L i l-¡ u*Lro ',Bn o(E q) 1 :i'i I ru, t:l i,l¡ i;i r1i{,r lr", lllt tr 0 tl . [l:iL.,i¡ 8'go., o o [:þ¡, 'ri.pr.r .: 8.'' rJ- (¿ ($,1 D,9 .t * (.ì ¡r ;l:l ,r,, (t{,(.): '%, q) tttl ïl¡, [ì '-tJrlt .2 1l.i 63 I /s .3 ir 2 O ¡r ((t(_:¡ i q) r.ilì ór F.lr tl r.l É'+gB 0 4,1i tl r:, rili (i ,ti sü {tt .1 ir.q {:l () í:ì g no' ,¡¡,ìjr;, ii rili i, o c) -2 o P8,$ B iJ Él:)tcl fÏ (.ì B o .3 uo) EÈ 2 .e8 I t,. o (j 9, l;:1. I l¿,h fJ Ë,3 t'f' tii tr"l r Ë=ú 0 l{i fui ì ,ii il niï t:.' ..rÌ ,.-- $ l:tt ,E i,]t l,i I ü 5? -1 EEoc¡ ffiuu *, diÐa¡ ü ,:i G¡(l)O d.) r.¡.¡rù ô l:,, tt r,irt at) I -2 o Ir GtË rì('ì -3 ;'t 'r:-"' ll pr i: ': ¡l'.'rtrtt ' t! 2 E.=ÅU) dP t i,, a I:J *t{ s.rQìoo .=ËË. I. u o é) \ry "Ëþh, B W r.1. r'.¡ ¡ít l¡ ¡; tt',Èrr[j eB .1 äE fio ,{,, o t$ 'fir ¡rõ uç .8%fÞ ..1i, i"n g 'i5 .2 Oct ar) .3 ã0É ar) j.E 2 () fit ii:l FìÐ ñ3 t iì' ,-, &E u.(l:1., rl't0¡ ö '', i,ßP o )rì (.r) t" d\1. ,, 0, qB c¡= lrl L¿3 -t !,,iâ "tla,6g' ,'rr iii fl,å' ,u, illjoæ ötf:J frì â0= -2 -3 .= 0.3 0.4 0.5 0.1 0.2 0.3 0,4 0.5 EE "1 o.r o.2 0.3 0.4 0.5 o.t o.2 trC¿ 0 (cm3 cm-3) lJr9-- P CHAPTER 8: REF'ERENCES Alexander, K.G. and M.H. Miller (1991). The effect of soil aggregation size on early growth and shoot-root ratio of maize. Plant and soil 138: 189-194. Allison, L.E. and C.D. Moodie (1965). Carbonates. In C.A. Black (ed.) Methods of Soil Analysis. Part l. ASA Madison, Wisconsin.Pp.1379-1400. Ayers, P.D. and J.V. Perumpral (1982). Moisture and density effect on cone index. Amer. Soc. Agr. Eng.25: 1169-1112. Baligar, V.C. and V.E. Nash (1973). Sorghum root growth as influenced by soil physical properties. Comm. Soil Sci. Plant Anal.9: 583-594. Becher, H.H. (1994). Soil compaction around a small penetrating cylindrical body and its consequences. Soil Technol, 7:83-91. Belford, R.K., M. Dracup, and D. Tennant (1992). Limitations to growth and yield of cereal and þin crops on duplex soils. Austr. J. Exp. Agr.32:929-45. Betz, C.L., R.R. Allmaras, S.M Copeland, and G.V/. Randall. (1998). Least Limiting Water Range: Traffic and Long-term Tillage Influences in a Webster Soil. Soil Sci. Soc. Am. J. 62:1384-1393. Bouma, J. and H.A.J. van Lanen (19S7). Transfer functions and threshold values: from soil characteristics to land qualities. Quantified land evaluation, Proc. Int. Worksh. ISSS/SSSA, Washington, DC. Enschede, the Netherlands, ITC Publ. p. 106-1 10. Braunack, M.V. and A.R. Dexter (19SS). The effect of aggregate size in the seedbed on surface crusting and growth and yield of wheat (Triticum aestivum, L. cv. Halberd) under dryland conditions. Soil Till. Res. 1l: 133-145. Braunack, M.V. and A.R. Dexter (1989). Soil aggregation in seedbed: a review. II. Effect of aggregate sizes on plant growth. Soil and Till. Res. 14 281-298. 93 Bregt, A.I( and J.G.H. Beemster (1939). Aocuracy in predicting moisture deficits ancl changes in yield from soil maps. Geoderma 43: 301-310- Brooks, R.H. and A.T. Corey (196$. Hydraulic properties of porous media. Fort Collins, CO., Colorado State University. Burt, O. R. (1995). Pooling experiments over time to qualifu the dynamics of dryland winter wheat yield response. Agron. Journal 87:156-164. Busscher, W.J. (1990). Adjustment of flat-tipped penetrometer resistance data to a common water content. Amer. Soc. Agr. Eng 33: 519-524. Busscher, W. J. and R.E. Sojka (1987). Enhancement of sub-soiling efflect on soil strength by conservation tillage. Amer. Soc. Agr. 30: 888-892. Busscher, W.J., L.D. Spivey, and R.B. Campbell. (1987). Estimation of soil strength properties for critical rooting conditions. Soil and Till. Res. 9:377-386. Byrd, C.W. and D.K. Cassel (19S0). The efflect of sand content upon cone index and selected physical properties. Soil Sci. 129: 197-204. Callebaut, F., D. Gabriels, W. Minjauw, and M. De Boodt (1985). Determination of soil surface strength with a needle-type penetrometer. Soil Till. Res. 5,227-45. Campbell, G.S. (1974). A simple method for determining unsaturated conductivity from moisttue retention data. Soil Sci. 177:311-313. Canarache, A. (1990). PENETR - a generalised semi-empirical model estimating soil resistance to penetration. Soil and Till. Res. 16: 5l-70. Carter, M.R. (1983). Temporal variability of soil macroporosity in a flure sandy loam under mouldboard ploughing and direct drilling. Soil Till. Res. 12: 37-51. Carter, M.R. (1990). Changes in soil structure under short-term tillage systems on fme sandy loam in Prince Edward Island. In J.A. Stone, B.D. Kay, D.A. Angers (ed.) 94 Soil Structure Research in Canada: proceeding of the Eastern Canada Soil Structure workshop, September 10-11, 1990. Guelph, Ontario, Herald Press Ltd. Carte¡ M.R. (1990a). Relative measures of soil bulk density to clnractettze compaction in tillage studies on fìne sandy loams. Can. J. Soil Sci. 70 425-433- Carte4 M.R. (1990b). Relationship of strength properties to bulk density and macro- porosity in cultivated loamy sand to loam soils. Soil and Till. Res. 15:257-268- Chittleborough, D.J. (1982). Textural Differentiation in Australian Soils. Adelaide, SA, PhD thesis. unpublished. chittleborough, D.J. (1992). Formation and pedology of duplex soils. Aust. J. Exp. Agr.32:815-25. Constantini, 4., D.Doley, H.B. So. (1996a). Early Pinus caribaea vat. hondurensis root development l. Influence of matric suction. Austr. J. Exp. Agr.36:839-46' Constatini, 4., D.Doley, H.B So. (1996b). Early Pinus caribeavar. hondurensis root development 2. Influence of soil strength. Austr. J. Exp. Agr. 36: 847-59. 'Water Cresswell, H.P. and Z. Paydar (1996). retention in Australian soils I. Description and prediction using parametric functions. Austr. J. Soil Res. 34: 195-212. Cull, P. (lg7g).Irrigation scheduling of cotton by computer. PhD Thesis, University ofNew England. da Silva, A.P. and B.D. Kay (1996). The sensitivity of shoot $owth of corn to the least limiting water fange of soils. Plant and Soil 184: 323-329. da Silva" A.P. and B.D. Kay (1997a). Estimating the least limiting water range of soils from properties and management. Soil Sci. Soc. Am.J 6l:877-883. da Silva, A.P. and B.D. Kay (L997b). Effect of soil water content variation on the least limiting water range. Soil Sci. Soc. Am. J. 61: 884-888. 95 water range of soils. Soil Sci. Soc. Am. J. 58: 1775-1781' da Silva, 4.p., B.D. Kay, and E. Perfect (1997). Management versus inherent soil properties effects on bulk density and relative compaction. Soil and Till Res. 44: 8l-93. Dalal, R.C. (1974). Eftects of intercropping maize with pigeon peas on grain yield and nutrient uptake. Exp. Ag. l0:219-224- Day, p.R. (1965). Particle fractionation and particle-size analysis. In C.A. Black (ed) Methods of Soil Analysis. Part 1. P.545-566. Dexter, AR. (1978). A stochastic model for the growth of roots intilled soil. J. Soil Sci.29:102-116. Dexter, A.R. (19SS). Advances in cha¡actertzationof soil structure. Soil and Till. Res. lI:199-238. '¡ r,{ Donald, R.G., B.D. Kay, and M.H. Miller (1937). The eflect of soil aggregate size on ,i; I early shoot and root growth of maize (Zea mays L.). Plant & Soil 103:251-259. Dracup, M., R.K. Belford, and P.J. Gregory 0992). Constraints to root growth of wheat and þin crops in duplex soils. Austr. J. Exp. Agt.32:947-61. Fussi, C. (199S). Intercropping in lucerne (Medica sativa). Masters Thesis' University of Adelaide. Gardner, E.4., R.J. Shaw, G.D. Smith and K.J. coughland. (1984). Plant available water capacity: concept, measurement and prediction. In J.W. McGarity, E.H. Hoult and H.B. So (ed.) The Properties and Utilization of Cracking Clay Soils.. N.S.W, Australia, University of New England. Reviews in Rural Science 5: 164- 175. I 96 Ga¡dner, W.K., R.G. Fawcctt, G.R. Steed, J.E. Pratley, D.M. Whitfield ancl H' van Rees. (1992). Crop production on duplex soils in south-eastern Australia. Austr. J. Exp. Agr. 32: 915-927. Grable, A.R. and E.G. Siemer (1968). EfÊects of bulk density, aggregate size, and soil water suction on oxygen diffusion, redox potentials and elongation of corn roots. Soil Sci. Soc. Am. Prcq32:180-186. Grant, C.D. (1989). Soil structure and tensile strength in relation to the microtopography of soil ûacture surfaces. PhD Thesis. University of Adelaide' Greacen, E.L. (1977). Mechanisms and models of water transfer. Í¡ J.S. Russell and E.L. Greacen (eds). Soil Factors in Crop Production in a Semi-arid Environment. St Lucia, Queensland. Univ. Queensland Press.p 163-96- Gregorich, 8.G., D.A. Angers, c.A. campbell, M.R carter, c.F Drury, B.H. Ellert, P.H. Groenevelt, D.A. Holmstron¡ C.M. Monreal, H.W' Rees, R.P Voroney, and il T.J. Vyn. (1995). InD.F.Acton and L.J. Gregorich (ed.) Changes in Soil Organic ,\ rl Matter. The Health of Our Soils - Towards Sustainable Agriculture in Canada.. ottawa. centre for Land and Biological Resources Research. p. 4l-50. Gregory, P.J.(199S).Alternative crops for duplex soils: growth and water use of some ceteal, legume, and oilseed crops, and pastures. Aust. J.Agric. Res. 49: 2l-32' Gregory, P.J. and J. Eastham (1996). Growth of shoots and roots, and interception of radiation by wheat and þin crops on a shallow, duplex soil in response to time of sowing. Aust. J. Agric. Res. 47: 427-447. Gregory, P.J., D. Tennant, and R.K Belford (1992). Root and shoot growth, and water and light use efficiency of barley and wheat crops grown on a shallow duplex soil Ì in a Mediterranean-type environment. Aust. J. Agric. Res. 43: 555-3. t 97 Gupta, S.C. and W.E. Larson (1979). Estimating soil water characteristics from particle size distribution, organic carbon and bulk density. Water Resour. Res. 15:1633-1635. Haise, H.R., H.J. Haas, and L.R Jesen. (1955). Soil moisture studies of some Great plain soils: II. Field capacity as related to l/3-atmosphere percentage and "minimum point" as related to 15- and26-atmosphere percentages. Soil Sci. Soc. Am. Proc. 34:20-25. Hfüansson, I. (1990). A method for characterurngthe state of compactness of the plough layer. Soil and Till. Res. 16: 105-120. Håkansson, I. and J.von Polgar (1934). Experiments on the effects of seedbed characteristics on seedling emergence in a dry weather situation. Soil Till. Res. 4: I 15-135. Hill, J.N.S., and M.E. Sumner (1967). Effect of bulk density on moisture ,I iù{ ,i characteristics of soils. Soil Sci., 103:234-8. I Holloway, R.E. (1991). Factors aflecting the growth of wheat roots in the subsoils of upper Eyre Peninsula, South Australia. MAgSc Thesis. University of Adelaide. Huston, J.L. and A. Cass (1987). A retentivity function for use in soil-water simulation models. J. Soil Sci. 38: 105-113. Isbell,. R.F. 1996. The Australian Soil Classification. CSIRO. Collingwood, VIC. Jakobsen, BF. and A.R. Dexter (1988). Influence of biopores on root growth, water uptake and grain yield of wheat (Triticum aestivum) based on predictions from a computer model. Biol. Fert. Soils 6: 315-32I. Jayawardane, N.S. and K.Y. CI:F¿1 (99$. The Management of Soil Physical Properties Limiting Crop Production in Australian Sodic Soils- A Review. Aust. i J. Soil Res. 32: 13-44. ! 98 Kay, B.D. (1989). Asscssing the suitability of different soils for new cropping systems in terms of rates of change in soil structure. In W.E. Larson, G.R. Blake, R.R. Allmaras, W.B. Voorhees and S.C. Gupta (eds). Mechanics and Related Processes in Structwal Agricultural Soils. DordrechtNetherlands, Kluwer Academic Pub. P. 223-33. Kay, B.D. (1990). Rates of change of soil structure under different cropping systems' Adv. Soil Sci.12: l-52. Kay, B.D. (199S). Soil Structure and Organic Carbon: A Review. In R. Lal, J.M. Kimble and B.A. Stewart (ed.) Soil Processes and the Carbon Cycle. Boca Raton, CRC Press. p.169-197 . Kay, B.D. and C.D. Grant (1996). Structural aspects of soil quality. 1z M.R. Cartet and R.J. MacEwan (ed.) Soil Quality is in the Hands of the Land Manager: advances in soil quality for land management: science, practice and policy: .I tì,t proceedings of an international symposiurr¡ 17-19 April 1996, University of tii Ballarut,Victoria. Australia, Centre for Environmental Management. p.37-41. Kemper, W.D. and R.C. Rosenau (1986). Aggregate stability and size distribution. Methods of Soil Analysis. Part 1. A. Klute. Madison, WI, Amer. Soc. Agron. p.425-42. Kern, J.S. (1995). Geographic patterns of soil-water holding capacity in the contiguous United States. Soil Sci. Soc. Am. 59: 1126-1133. King, L.G. (1965). Description of soil characteristics for partially saturated flow. Soil Sci. Soc. Amer. Proc.29;359-62. Klute, A. (1936). Water retention: Laboratory methods. In A. Klute (ed) Methods of t I Soil Analysis. Part L ASA. Madison. P.635-62. ; l 99 Letey, J. (19S5). Rclationship between soil physical properties ancl crop production' Adv. Soil Sci. I :277-294. Linquist, 8.4., P.V/. Singleton, R.S Yost, and K.G. Cassman' (1997)' Aggregate size effects on the sorption and release of phosphorus in an Ultisol. Soil Sci. Soc. Am. J.61: 160-166. Logsdon, S.D., J.C. Parker, and R.B. Reneau (1987). Root growth as influenced by aggregate size. Plant and Soil 99 267-275. rWageningen, Luyten, J.C.(1995).Sustainable world food production and environment. Research Institute for Agribiology and soil Fertility (AB-DLO). Madsen, H.8., K.A. Holst, and S.A. Mikkelsen (1989). The use of the EC soil map in modelling and mapping the root zotte capacity and irrigation need: a case study from Denmark. Agriculture. InF.. J. A. Jones and B. Biagi (ed.) Computerization of Land Use Data. Luxembourg, Commission of the European Communities. p.74-84. Materechera, S.A, A.R. Dexter, and A.M. Alston (1991). Penetration of very strong soils by seedling roots of different plant species. Plant and Soil 135: 31-41- Mirreh, H.F. and J.V/. Ketcheson (1972). Influence of soil bulk density and matric pressure on soil resistance to penetration. Can. J. Soil. Sci. 52: 477-483' Nelsoq D.W. and L.E. Sommers (1986). Total carbon, organic carbon, and organic matter. In A.L. Page (ed) Methods of Soil Analysis. Part2. No.9. ASA Madison. P.539-77. Njos, A.(1979).Aggregate size distribution in the seedbed: effects on soil temperature, matric suctior¡ and emergence of barley (Hordeum volgare, L.) a review of some i( r I research on clayey soils in south-eastern Norway. Proc. 8th Conference I I 100 International Soil Tillage Research Organisation, Llniversity of Hohenheim. Stuttgard, BRD. Northcote, K.H. (1979). A Factual Key for the Recognition of Australian soils. Glenside, S. Aust', Reltim Technical Publ. 76: Oades, J.M. (1984). Soil organic matter and structural stability. Plant and Soil 319-37. Perumpral, J.V. (19S7). Cone penetrometer application- A review' Amer' Soc' Agr' Eng.30: 939-944. Prentice, I.C., M.T. Sykes, W. Cramer (1993). A simulation model for the transient efFects of climate change on forest landscapes. Ecol. Model' 65: 512-70' (1996). Soil Organic Matter Particle Quiroga, 4.R., D.E. Buschiazzo,N. Peinemann Size Fractions in Soils of the Semiarid Argentinian Pampas. Soil Sci. 161: 104-8. Rajkai, K., S. Kabos, M.Th. Van Genuchter¡ and P-E. Jansson (1996). Estimation of water retention characteristics from bulk density and particle-size distribution of Swedish soils. Soil Sci. 161:832-45. Rawls, V/.J. (1933). Estimating soil bulk density from particle size analysis and organic matter content. Soil Sci. t35:123-124. Rawls, W.J. and D.L. Brakensiek (1985). Prediction of soil water properties for hydrological modelling. In c.B. Jones and T.J. Ward (ed) watershed Management in the Eighties. ASCE Convention. Denver. P. 293-99. Rawls, W.J., T.J. Gish and D.L. Brakensiek (1991). Estimating soil watcr retention from soil physical properties and characteristics. Adv. Soil Sci. 16: 213-231. Reinds, G.J., G.H.J. De Koning, and J.D. Bulens (1992). Crop production potential of rural area within the European Communities: PartIII. Soils, climate and 101 administration regions. The Hague, Netherlancls Scentific Council for Government PolicY. Richards, L.A, and L.R. V/eaver (1944). Fifteen atmosphere percentage as related to the permanent wilting point. Soil Sci. 56: 331-339. Robson, 4.D., N.E. Longnecker, L.D. Osbourne (1992). Eftects of the heterogeneous nutrient supply on root growth and nutrient uptake in relation to nutrient supply on duplex soils. Aust. J. Exp. Agric. 32:879-886. Ross, P.J., J. Williams, and K.L. Bristow (1991). Equations for extending water retention curyes to dryness. soil sci. soc. Am. J. 55: 923-927. Russell, E.W. (1973). Soil Conditions and Plant Growth' Longman, London, p'849' Schultz, J.E. (1974). Root development of wheat at the flowering stage under cultural practices. Agric. Record, I: 12-17. Soane, 8.D., P.S. Blackwell, J.W. Dickson, and D'J. Painter (1981). Compaction by agricultural vehicles: a review. l. Soil and wheel characteristics. Soil and Till Res. 15: 1-14. Spivey, L.D., W.J. Busscher, and R.B. Campbell. (19S6). The eflect of texture on strength of southeastern Coastal Plain soils. Soil Till. Res. 6:351-363' Taylor, H.M. and H.R. Gardner (1963). Penetration of cotton seedling taproots as influenced by bulk density, moisture content, and strength of the soil' Soil Sci. 96:153-156. Taylor, H.M., G.M. Roberson, and J.J. Parker (1966). Soil strength-root penetration relations for medium to coarse-textured soil materials. Soil Sci. 102: 18-22- Tennant, D. and G. Scholz (lgg2). Physical and chemical characteristics of duplex soils and their distribution in the south-west of Western Australia' Aust. J. Exp. Agric. 32:827-843. r02 Terman, G.L., R.E. Raming, A.F. Dreier, R.A.Olson (1969)- Yield-protein relationships in wheat grain as affected by nitrogen and water. Agron.J.6l:755-9. Thomas, G.W., G.R. Haszler, and R.L. Blevins (1996). The eflects of organic matter and tillage on maximum compactability of soils using the Proctor test. Soil Sci. 16l:502-8. Thomasson, A.J. (1978). Towards an objective classification of soil structure- J. Soil Sci.29: 38-46. Tisdall, J.M. and J.M. Oades (1932). Organic matter and water stable aggregation in soil. J. Soil Sci. 33:141-163. Tisdall, J.M. and H.H. Adem (1986). Effect of water content of soil at tillage on size- distribution of aggregates and infiltration' Aust. J. Exp. Agric. 26: 193-195' Tkachuk, R. (1969). Nitrogen to protein conversion factor for cereals and oil feed meal. Cereal Chem. 46:419-23. Turner, N.C. (1966). Grain production and water use of wheat as affected by plant density, defoliation and water status. PhD Thesis, University of Adelaide. Turner, N.C. (1992). Crop production in duplex soils: an introduction. Aust.J' Exp. Agr.32:797-800. van Genuchten, M.T. (1980). A closed form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci. Soc. Am. J. 44:892-898. van'Wesenbeeck, I.J. and R.G. Kachanoski (19S8). Spatial and temporal distribution of soil water in the tilled layer under corn crop. Soil Sci. Soc. Am. J. 52: 363-8. Vepraskas, M.J. (1984). Cone index of loamy sands as influenced by pore size distribution and effective stress. Soil Sci. Soc. Am. J. 48: 1220-1225. 103 Volkmar, K.M. and T. Entz (1995). A mcthod for characterung the effect of root- formed pores on growth of roots in a Chernozemic clay loam. Can. J. Soil Sci' 75:293-298. 'Wagenet, RJ., J. Bouma, and R.B. Grossman. (1991). Minimum data sets for use of soil survey information in soil interpretative models. In M.J. Maubach and L.P. V/ilding (ed.) Spatial variabilities of soil and landforms. SSSA Spec. Publ- 28. SSA, Madison, WI. White, R.E. (1985). Transport of dissolved and suspended matter through soil. Adv. Soil Sci. 3:95-113. White, R.E. (1997). Principles and Practice of Soil Science. Oxford, Blackwell Science. V/illiams, J., P. Ross, and K. Bristow. (1992). Prediction of the Campbell water retention function from texture, structure, and organic matter. p.427-441. In M. T. van Genuchten and L.J. Lund (ed.) Indirect methods for estimating the hydraulic properties of unsaturated soils. Univ. of California, Riverside. Xu, X., J.L. Nieber, and S.C. Gupta (1992). Compaction effects on the gas diffirsion coeffrcient in soils. Soil Sci. Soc. Am. J. 56: 1743-50. 'Wander Xue-Ming, Y. and M.M. (199S). Temporal changes in dry aggregate size and stability: tillage and crop effects on a silty loam Mollisol in Illinois Soil Till. Res 49:173-183. Zhai,R., R.G. Kachanoski, and R.P. Voroney. (1990). Tillage effects on the spatial and temporal of soil water. Soil Sci. Soc. Am' J. 54:186-192. Zuidema, G., G.J. Van den Born, and G.J.J. Kreileman (1994). Simulation of global 'Water land cover changes as affected by economic factors and climate. Air Soil Pollut. 76:163-198. 104