Вы находитесь на странице: 1из 18

ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF

FOOTING SYSTEMS ON REACTIVE SOILS FOR THE EFFECTS OF


TREES

D. A. Cameron1 and N. S. Beal2


1
University of South Australia, School of Natural & Built Environments
2
Hughes, Beal & Wright Pty Ltd Consulting Engineers Brisbane Queensland

ABSTRACT
The paper details the method of designing for the effects of trees as set out in a new informative Appendix H in
the revised Residential Slabs and Footings Standard, AS2870, published by Standards Australia in February
2011. Background to the recommendations is given along with an introduction into research efforts that may
improve designs in the future. In essence, the recommendations have been formulated in the light of more than
20 years successful use in South Australia of simple rules promulgated by the Footings Group (South Australia).
Other evidence, generally from case studies of damaged houses, has been gathered in different climates, which
supports the general premises of the recommendations and the extrapolation of the method to areas with wetter
climates.

1 INTRODUCTION
The root systems of trees extract water from the soil as part of their biological activity, most of which is
transpired through their leaves. The presence of trees can increase the degree and depth of variation in soil
moisture compared to that which occurs in their absence. If the soil is reactive, the extra moisture variation will
cause extra volume change and foundation movement. If a tree dies or is removed, the soil within the influence
of its root will commonly be desiccated relative to the surrounding soil. Subsequent wetting up of the desiccated
zone as it equilibrates with the surrounding soil can result in swelling and foundation heave if the soil is reactive.
The Australian Residential Slabs and Footings Standard, AS2870, was first published in 1986. It has provided
guidance for the design of residential footings on reactive clay. The original standard, and the subsequent
editions, including the penultimate edition published in 1996, did not provide any means of allowing for the
exacerbation of reactive clay foundation movement caused by trees. Rather, the standard encouraged the
avoidance of tree influence on footing systems by setting limits on the proximity of trees to the house. The site
classification method in the standard assumed reasonable site maintenance throughout the life of the building
and that trees are kept sufficiently far away, so that they have no significant influence on the ground surface
movement beneath the building.
The separation rules in AS2870 set limits to the ratio of Dt/HT, where Dt is the minimum distance from the
house wall to the tree trunk and HT is the height of the tree. The separation ratios recommended by AS2870
increase with the level of site classification and also if more than one tree is present. There is no attempt to
distinguish between effects of different tree species or the form of the tree: the sole criterion is the height of the
tree.
Due to the continuing trend towards smaller residential allotments and the increasing appreciation of various
benefits of trees, the absence of explicit rules to allow for the effects of trees has become an increasing limitation
on the applicability of the standard. Design engineers have tried to minimise the risk of damage due to associated
soil shrinkage settlements either by excluding trees, or, as in South Australia, by designing footings to cater for
the anticipated extra soil shrinkage settlement. The South Australian design criteria for tree-drying are based
on simplistic empiricism, as very little information is available on the effects of different tree species on soil
moisture regimes in the urban environment. It is likely that this simplistic approach is resulting in some footings
being overdesigned and thus adding unnecessarily to building costs, and in other cases, under-design, resulting in
footing failures, building damage and loss. The experience of design engineers in South Australia since the
implementation of the rules in 1990, as ascertained via a questionnaire to the Footings Group circulated in 2010,
suggested that the approach may be conservative; failures are rare. Others may argue that 20 years is too soon for
some tree species to have had their full effect.
A useful concept in modelling the maximum desiccation caused by any vegetation, including trees, is the wilting
point suction, which is defined as the soil suction beyond which tree roots cannot continue to extract water from
the soil. This concept sets absolute limits on the drying effect of trees on the soil. When combined with
knowledge of the depth extent of the root affected zone, the equilibrium moisture conditions at depth and the

Australian Geomechanics Vol 46 No 3 September 2011 87


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

reactivity of the soil, estimates of the extra ground movement attributable to the tree drying can be made.
Various researchers have used the wilting point concept in their numerical models for predicting the settlement
caused by extraction of water from the soil by vegetation (Fredlund and Hung, 2001; Indraratna et al., 2006;
Hemmati et al., 2008; Navarro et al., 2009).
In arid areas, vegetation will often be unable to extract water from the top metre or so of the seasonally dry
profile because the soil there is already drier than the wilting point. Accordingly, the vegetation must extract
water from deeper within the soil profile to survive. Tree roots also need a reasonable level of oxygen to survive.
It appears that shear surfaces and shrinkage cracks within reactive soil masses can allow ingress of enough
oxygen to sustain roots at significant depths. The literature abounds with reports of deep penetration of tree roots
(Fityus, Cameron and Driscoll, 2007).
The limitation on drying inherent in the wilting point concept has been considered in the current
recommendations in AS2870-2011. However, a pragmatic and simplified approach has been adopted in defining
the tree induced suction profile. The simplifications and departures from the strict wilting point concept include
that:
1. trees can affect design suction changes from the ground surface downwards
2. the depth of drying is curtailed to an extent.
The extension of the South Australian approach to other climates has been made based on equivalent drying
areas (suction change profiles) within the soil profile, with a reduction in the degree and depth of drying for
wetter regions where water is normally more available to vegetation and hence there is less incentive for
development of root systems to the depths seen in semi-arid climates. The general approach has been calibrated
with data from South East Queensland (Beal and Cameron, 2007), which is presented here in support of the
method given in the revised Standard. South Australian and Victorian data reported in Cameron (2001) are also
presented in this paper for completeness.

2 AUSTRALIAN STANDARD AS2870


An underlying philosophy of the Standard since its inception in 1986 and prior to the recent release has been to
provide footing designs for the predictable seasonal and environmental (i.e. normal) changes in the soil-
moisture regime, and to entrust homeowners with the responsibility of managing sites after construction. In this
way extreme (i.e. abnormal) moisture changes can be avoided. The basis for this decision was that Australia
could not afford to waste resources in building houses for all possible extremes of soil moisture in the urban
environment. Trees closer than certain prescribed distances relative to their heights were regarded as a potential
source of extreme (abnormal) moisture change. The previous editions of the standard did not provide guidance
for such cases.
The separation rules that are deemed to exclude tree effects were based on experience gained largely from
investigations of damaged houses in the Melbourne area. Table 1 summarises the separation rules of AS2870
first expounded in the associated CSIRO information sheet (1988) and later by AS2870.2 (1990). It should be
noted that the CSIRO information sheet simply recommended that the distances for single trees be increased by
50% for dense groups of trees.
Table 1: Separation Rules of AS2870 1988 to 1996

Site Classification Minimum Separation, Dt/HT

Single trees Tree groups


M, Moderately reactive 0.75 1.12
H, Highly reactive 1.0 1.5
E, Extremely reactive 1.5 2.25

3 SOUTH AUSTRALIAN FOOTINGS GROUP APPROACH TO THE DESIGN


OF TREES
Since July 1990, the Special Provisions of the South Australian (SA) Footings Group have contained a design
method that accounts for trees in the generally semi-arid climate of the State. The majority of the experience
with this method has been gained in the city of Adelaide and its surrounds. It is common practice in South
Australia to classify sites according to estimates of potential ground movement calculated from design soil total
suction change profile and estimates of soil reactivity. It is a simple and readily accepted extension of this
approach to adopt a modified design total suction change distribution, which includes tree drying effects.

88 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Furthermore, in South Australia, it is common practice to design each and every footing system for the expected
design movement, using beam-on-mound computer programs such as SLOG (Mitchell method) or CORD
(Walsh method).
For normal sites, AS2870 recommends a design suction change regime, which decreases linearly with depth.
This is supposed to represent the range of suction through the profile on normal sites under the influence of
climate. The maximum design suction change is at the soil surface and decreases to zero at the depth of design
suction change Hs, which varies with climatic region. This depth is four metres for Adelaide. The Footings
Group method simply required an increase in the suction change at depth Hs from zero to a prescribed value,
ubase. The additional suction change attributed to tree drying of the soil was assumed to increase linearly from
zero at the surface to ubase at depth Hs. There was no requirement to consider any tree-induced suction change
below the depth, Hs. The design suction change distributions are shown in Figure 1.
In the Footings Group method, the adopted suction change, ubase, at depth, Hs (i.e. 4.0 m for South Australia),
varies depending on the following factors:
1. Separation ratio (Dt/HT)
2. Density of planting
3. Site classification, assuming no tree influence
Generally, the separation ratio referred to in the first point is based on HT, which should be the expected mature
height of the tree. Without expert botanical assistance, this is a parameter which is difficult for engineers or site
classifiers to estimate and so many designers have resorted to applying the most severe condition, a separation
ratio of 0.5. The method attributes no further increase in tree effect for smaller separations.
The Footings Group method increases the calculated mound height in the centre heave mode of foundation
movement as the method is specifically concerned with new construction and the planting of trees during or after
construction. In the 2008 revision of the Special Provisions, commentary was added to warn of the potential
impact of tree removal prior to construction. Tree removal and the consequent equilibration of the desiccated soil
in the root zone can lead to large permanent edge heaves if the tree happened to be outside the building footprint,
or even large mound heave if the tree was within the building footprint.
For design of footings on reactive sites unaffected by trees, AS2870 defines the foundation mound by defining a
notional edge distance parameter. With tree drying, the reach of drying under a slab system can be expected to
be greater than normal, which other things being equal, requires a more robust footing design. The South
Australian practice has been to ignore this aspect, but to compensate by adopting a symmetrical mound in the
design rather than the asymmetrical mound that could be expected with a tree only on one side of the building.

us us
wet dry wet dry

no trees no trees
Hs Hs

ubase ubase

a) Single tree b) Row, group or stand of trees

Figure 1: Design suction change profiles for tree drying, Footings Group SA
Extra requirements for ductility in the reinforced concrete beam sections are prescribed in recognition of the
greater uncertainty of design and the associated need to avoid gross failure of the slab if design movement limits
are exceeded.
Members of the Footings Group (SA) attach a caveat to their designs when designing according to the special
provisions of the Group. This caveat expresses the limitations of the design approach which can be summarized
in point form as follows:
There are no reliable design methods due to the complexity of the problem

Australian Geomechanics Vol 46 No 3 September 2011 89


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

e.g. predicting tree root geometry or future tree growth is uncertain


Precise design is outside current knowledge
- some compensation is provided by increasing footing sizes
Footing designers and civil engineers are not experts on trees
Greater tolerance of possible house distortion is required
The current Standard and its Commentary also recognize the difficulty of dealing with nature. Under the heading
Limitations; Clause H1 states: The risks of underperformance arise from factors that include the inherent
variability and unpredictability of living, growing trees and their interaction with the environment, as well as
imperfections in the method of modelling the effect of trees.

4 CRITICISMS OF THE ENGINEERING APPROACH


As with any simple design method for a complex problem, many criticisms can be levelled at the engineering
approach, and it is hoped that with further research, the method can be improved. Nonetheless the South
Australian approach appears to have been successful in dealing with tree-related settlements over the last 20
years. Some obvious criticisms of the approach are listed in the following discussion.

4.1 THE GEOMETRY OF THE TREE


Trees transpire water through their leaves and therefore the total leaf area has an important bearing on the
quantity of water a tree will extract. The density and shape of the canopy may be more significant than the tree
height with respect to water use. Leaf area index and leaf transpiration measured over time may be combined to
estimate water use. The leaf area index is the amount of leaf surface area per unit ground area.
It has been reported also that water use of a species in a similar climatic zone increases with the diameter of the
trunk at breast height (McJannet and Finch, 2004). Cameron and Mills (2010) were unable to support this
finding with data from field trials using the heat pulse technique to measure transpiration of black box trees (E.
largiflorens) in north-western Victoria. Instead they found that transpiration varied almost linearly with sapwood
conducting area, which is not necessarily directly related to trunk diameter.

4.2 THE DESIGN SUCTION CHANGE PROFILE


Interpretations of soil suctions in the vicinity of trees on urban sites, which were generally within Adelaide and
its surrounds, have indicated deep drying due to trees, especially for tree groups (Cameron 2001, Jaksa et al.,
2001). Depths of drying were observed to be approximately four metres for a single tree and six metres for a
stand of trees. Therefore the Footings Group approach of increasing the suction change at four metres to account
for deeper drying (Figure 1b) is a crude approximation to the actual influence of tree groups.
Furthermore, the total suction changes induced by tree groups appear to be no different from that for single trees,
which is in accordance with the concept that the wilting point suction defines the physiological limit of the trees
drying influence on the soil, irrespective of whether it is a single tree or one of a group.

4.3 RELEVANCE TO WILTING POINT SUCTION


Vegetation can only extract water from any particular zone in the soil if the total soil suction is below a limiting
value; the wilting point suction. The vegetation must extend its roots far enough horizontally or vertically to
access enough soil that is below the wilting point suction to meet its needs for water, otherwise it will be set back
through the need to drop leaves to reduce water consumption, or it may die.
The wilting point suction varies within fairly narrow limits depending on species and environmental conditions.
Correia, Torres and Pereira (1989) reported that at the soil wilting point, leaf water potentials (expressed as
positive rather than negative values) for non-irrigated E. globulus were approximately 2.0 0.15 MPa (or 4.28-
4.34 pF). OMalley and Cameron (2005) observed leaf water potentials over a two year period on leaves of three
exotic species (Golden Raintree, Ornamental Pear and Chinese Elm) and one native tree (Coral Gum). A
relatively small variation in the maximum leaf water potentials between species was observed, with the maxima
ranging between 1.9 and 2.3 MPa (or 4.29-4.37 pF). The higher value was achieved by an ornamental Pear and a
Coral Gum. The maximum leaf water potential is conventionally assumed to be a proxy measure of the
corresponding wilting point total soil suction.
Total suction at the wilting point rather than either of the components of total suction, namely solute and matric
suction, is the most constant measure of the capacity of vegetation to extract water from soil. However,
knowledge of the wilting point suction on its own is not sufficient to estimate the potential effect of vegetation
on foundation movement. Rather it is the change in suction from equilibrium suction at depth towards the

90 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

wilting point suction which is most the relevant suction parameter, as recommended in Cameron (2001).
Cameron reported values of deep equilibrium suction in Adelaide and its environs ranging between 0.87 and
1.38 MPa (3.95-4.15 pF) for 8 case studies. In these same cases, (uwp-ueq) was estimated to vary between 0.23
and 0.40 pF with an average of 0.31 pF.

5 THE DESIGN APPROACH OF AS2870-2011 FOR A RANGE OF CLIMATES

5.1 REVIEW OF THE RECOMMENDATIONS OF THE FOOTINGS GROUP (SA) IN LIGHT OF


THE WILTING POINT SUCTION
Building on the apparent success and acceptance of the SA Footings Group method for allowing for trees, the
method now given in AS2870 Appendix H is largely based around the Footings Group method. However,
account has also been taken of the wilting point suction concept when selecting the values for the suction change
parameters and depth of tree effect.
For the purpose of design, the suction change, (uwp-ueq), at depths equal to or greater than Hs, was assumed
conservatively to be 0.5 pF close to the tree. For the initial theoretical suction change for tree effects in the
South Australian climate, the depths of tree drying were taken to be four metres for a single tree and five metres
for a group of trees. The area of extra drying in terms of log suction difference (or pF) and the depth over which
it acts was calculated. The area was then transformed into a more tractable equivalent suction change
distribution, similar to that adopted by the Footings Group; the area of the theoretical and equivalent suction
change profiles was made equal.
Figure 2 shows the transformation from the theoretical suction change distribution to the adopted distribution,
for Adelaide and environs. In Figure 2 it can be seen that the normal design suction change in the absence of
any tree effect was assumed to be symmetrical, i.e. equal wetting and drying over the life of the building. This
assumption has been adopted for all climate zones.

1.2 pF 1.2 pF

no trees no trees
4m 4m

0.43 pF
0.5 pF

Single tree (wilting point approach) Single tree (approximation)

1.2 pF 1.2 pF

no trees no trees
4m 4m
4.5 m
5m

0.55 pF
0.5 pF

Tree group (wilting point approach) Tree group (approximation)

Figure 2: Design suction change profiles for tree drying for Adelaide

Australian Geomechanics Vol 46 No 3 September 2011 91


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

The adoption of 4 m as the depth of influence for single trees reflects a reluctance to impose a requirement to
investigate to greater depth than is already required by the 4m deep normal site suction distribution.
Nevertheless, an extra half metre was added for the tree group distribution in acknowledgement of the empirical
evidence that trees will resort to extracting water from deeper than 4 m, when tree root systems have to compete
with neighbouring trees.
A comparison of tree modified mound heights (ym tree) in centre heave for the Footings Group approach and
mound heights calculated according to the guidance provided by AS2870-2011, is given in Figure 3. Mound
height has been calculated according to Equation 1:
y m tree = y m + yt = 0.7 y s + yt (1)
The mound height in the absence of trees is ym, and, for this comparison, yt is the maximum tree-induced
settlement. The comparison has been made based on a uniform soil profile, cracked through it full depth, and for
shrinkage indices ranging between 1.7 and 4.2 %/pF. With these values of site parameters and soil properties, the
normal design site surface characteristic movement, ys, ranged between 40 and 100 mm.
It is evident that for the single tree case there is little difference between the outcomes of the two methods,
except at the Class E-D end (ys = 100 mm) where the AS2870-2011 method provides an 8% lower estimate of
design mound height. However, for the tree group comparison, it is seen that the AS2870 approach is more
conservative across all classes of site, producing on average 7.3 mm extra mound height.

140
Mound Height Estimate (mm)

120
ys = 70

100
ys = 60

ys = 100
80
ys = 40

60 AS2870
40 Ftgs G p
20
0
0 1 2 3 4 5
Shrinkage Index (% strain/pF)

a) Single tree
140
Mound Height Estimate (mm)

ys = 70

120
ys = 60

ys = 100

100
ys = 40

80
60 AS2870
40 Ftgs G p
20
0
0 1 2 3 4 5
Shrinkage Index (% strain/pF)

b) Group of trees

Figure 3: Comparison of modified centre heave mound heights between methods for Adelaide

92 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

5.2 MODIFICATION OF DESIGN SUCTION CHANGE FOR OTHER CLIMATES


In more moist climates around Australia it was assumed that the area of extra drying should reduce as Hs
decreases. The assumed suction change values, (uwp ueq), are presented in Table 2 and are similar to that
proposed by Cameron and Beal (2007). It should be noted that there is no reason why the wilting point suction
should be reduced in more temperate climates.
However the reduced suction change values reflect the possibility that, owing to the greater availability of
moisture in the soil in such climates, the wilting point is less likely to be reached. The depth of drying was also
reduced with decrease of Hs as indicated in the same Table. A rationalisation for the reduced depth effect is that
in wetter climates the tree should not need to extend its root system as far as in arid climates in order to satisfy
its usual requirements for water and thus in drought conditions the tree is unable to influence the soil to as great
a depth as in areas where drought is more common.
The proposed design suction change due to extreme tree drying (Dt/HT 0.5) is illustrated in Figure 4 and is a
compromise between the Footings Group approach and the wilting point concept, shown to the left of the same
Figure (Ht-theory is the theoretical depth of drying). The areas of suction change for each climatic region, as
indicated by Hs, are equivalent between the wilting point concept and the design compromise. Parameters for the
design suction changes are provided in Table 3 for a design surface suction change of 1.2 pF.
With the exception of the case of a single tree on sites with Hs of 4 m, the design suction change profiles,
including tree effects, generally extend to greater depths than the normal limit of suction change, Hs, but are less
than the depths which would apply if the wilting point concept were strictly adopted. These guidelines have
practical implications for site investigations because it requires exploration to a greater depth than at present, if
the movement due to trees is to be estimated from actual rather than assumed soil profile data.

us us

no trees no trees
Hs
Hs
Ht-theory Ht

(uwp ueq)
ubase
Figure 4: The design suction change proposed in AS2870-2011
Table 2: Assumed theoretical variation of suction change and depth of drying with
Hs, : Wilting Point Suction approach

Hs (uwp-ueq) Depth of drying, Depth of drying,


(m) (pF) Single tree Tree group or row
1.5 0.34 3 4
1.8 0.38 3 4
2.3 0.42 3.15 4.15
3 0.46 3.5 4.5
4 0.5 4 5

Australian Geomechanics Vol 46 No 3 September 2011 93


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Table 3: Design suction change as a function of Hs (AS2870-2011)

Single tree Trees


Hs
(m) Ht ubase Ht ubase
(m) (pF) (m) (pF)
1.5 2.5 0.30 3.0 0.38
1.8 2.7 0.33 3.3 0.40
2.3 3 0.35 3.6 0.43
3 3.4 0.38 4.1 0.46
4 4 0.43 4.5 0.55
Based on the extra suction change due to the tree as defined by Figure 4 and Table 3, the ground movement
attributable to the tree and near the tree, yt max, can be calculated as usual by integration over the suction change
profile, taking account of the variation in soil reactivity through the profile (refer section 2.3 of AS2870-2011). It
may be assumed for the purposes of this calculation, but not for the calculation of ys, that the soil profile is fully
cracked (shrinkage cracking) and so the lateral confinement factor is equal to unity throughout for calculation
of the tree effect movement.

6 THE RECOMMENDED DESIGN PROCESS


In the absence of reliable botanical data, some practitioners may choose to adopt the full tree effect without
regard to moderation due to separation between the building and the tree. Alternatively, the maximum potential
movement due to the tree, ytmax, may be modified to take account of the separation of the building from the tree.
For ratios of separation to tree height Dt/HT greater than 0.5, a linear reduction in tree induced movement is
adopted, which reaches zero when the separation is equal to the influence distance, Di. Interpretation of ground
movement near a row of trees reported by Bozozuk (1962) suggested that movement varied almost linearly with
distance away from the trees.
The maximum influence distance is dependent on competition between the tree root systems and this will tend to
vary depending whether the tree is isolated or is in a group or row. Unlike the tree separation rules in earlier
versions of the Standard (refer the separation rules given in Table 1), the site Class is assumed not to impact on
root spread. Predicting root spread is a complex problem, particularly in an urban environment, however it is
well known that if water is plentiful near a tree, then roots will not reach as far as in the case of low soil water
availability (Yeagher, 1935). Therefore it seems more reasonable to consider climate rather than site Class in
relation to root spread. Although the effect of climate has not been considered directly on the maximum
influence distance, suction change profiles have been reduced from the South Australian profiles for less harsh
climates, and so the calculated movement will be reduced.
The equation for the extra movement, yt, at the building due to the tree for Dt/HT in the range 0.5 < Dt/HT <
Di/HT is:
yt D / HT 0.5
= 1 t (2)
yt max Di / HT 0.5
where Di/HT = 1.0 for a single tree
Di/HT = 1.5 for a group of trees
Di/HT = 2.0 for a row of trees
If the design separation ratio, Dt/HT, is 0.5 or less, then yt equals yt max. If the distance to the tree, Dt, exceeds the
influence distance Di then yt is zero.
Design may then proceed for a tree-affected differential, centre heave, with a modified mound height, ym tree,
equal to (0.7ys + yt) as given in equation 1. The contribution of the vegetation to the mound height is not
modified by the factor 0.7, as it is likely that the extra tree effect drying and resulting extra foundation
movement under the building will occur to the full extent calculated.
The following sections consider the movement estimation method in the light of nine case studies of damaged
houses in SE Queensland and a further eight from South Australia (Adelaide area) and one from Melbourne.

6.1 THE CASE STUDIES FROM SE QUEENSLAND


Pertinent details of the sites and the conclusions from site investigations are summarized in Tables 4 and 5. In
most cases the trees existed prior to the construction. For example in case 6B, a large eucalypt that was located

94 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

within the house footprint was felled shortly before construction of the house. Generally the initial failure
investigations were conducted within four to six years after construction. The deformations observed were not
always entirely attributable to vegetation; there were instances of poor drainage and leaking drains likely leading
to lower than normal suctions in bore holes remote from the tree. As would be expected the total soil suctions on
these SE Qld sites were generally lower than is commonly seen in the more arid climate prevailing in South
Australia.
In Table 5, the floor deformations are given as the maximum out-of-level across the slab (max). The floor levels
were measured using the manometric technique described in Beal (1993). Typically, for an averaged sized
house, about 100 spot levels were taken on the floor in each set of measurements. Soil suction samples were
taken down to depths of only 2 to 2.5 metres in most cases. The suction profiles nearest to the tree were
generally not taken right at the tree but rather near to the building at its closest approach to the tree. The
reactivity of representative samples of the soil profile was determined by shrink/swell test but, as always,
engineering judgement was needed in assigning shrinkage indices to untested parts of the soil profile.
The observed rebounds of the slab deformations attributable to tree removal are shown in Table 5. The rebound
generally occurred within a relatively short period of a year or so after tree felling or root exclusion by a barrier.
Measured total suction profiles across the site at the time of the investigation are shown in Figure 5 for two case
studies concerning isolated trees. As in all the plots of the suction profiles for the case studies, relative separation
(Dt/HT) has been indicated for the borehole closest to the tree (plotted with open circular markers). The borehole
away from the tree has been plotted with square markers).
Plausible design wet and dry suction profiles have been imposed on each Figure, together with the additional
design suction profile for the tree effect. For best comparison with the measured tree-affected suction profile,
the design normal suction profile should be centred, wherever possible, on a deep equilibrium suction value,
ueq. However, a value for the deep equilibrium suction could not always be inferred from the available data, due
to the sometimes limited depth of investigation.
In presenting the data, an allowance is required for the fact that the boreholes nearest to the tree and the
associated suction measurements are not taken right at the tree but rather at the house. For case studies where the
tree was located beyond a Dt/HT value of 0.5, the design tree effect extra suction profile was scaled back
according to the reduction factor allowed for tree separation given in Equation 2 so as to be more directly
comparable with measured suctions. The reduced design suction profile is indicated by a dotted line. This
scaling is not part of the design method but rather, for the purposes of this paper, it allows a fairer comparison
between the design profile and the available suction data.
Data from case study 1B, which involved a number of trees, is similarly presented in Figure 6. This case study
was of particular interest as total suctions were determined on samples taken from almost identical borehole
locations at the tree affected end of the building between 1996 and 2007. Over that period, Dt/HT decreased
from 1.25 to about 1.0, as two of the trees grew rapidly before they were ultimately removed. As expected, the
data suggest that at their initial height of 4.5 m and hence initial separation ratio of greater than 1, the trees did
not have much influence on the soil suction profile. However, as the trees grew further, the slab deformation
patterns that developed and the measured suction data confirmed the growing influence of the trees. After the
trees were removed in 2004 the suction profile at the hole near the tree location reverted to slightly wetter than it
had been when the trees were small.
This case (1B) also illustrated the difficulty of permanently and reliably excluding roots from a foundation by
means of a physical root barrier. A 1.8 m deep root barrier was installed in late 2000 and this induced recovery
of up to 36 mm from November 2000 to March 2002, but by June 2003 all of that recovery and more had been
lost. At that stage it was observed that Tipuana roots of up to 40 mm diameter had grown out of the ground and
over the polythene root barrier. It is unknown if the roots had also circumvented the bottom or the ends of the
barrier. The trees were removed in May 2004 and recovery of up to 56 mm occurred in the period to Oct 2005.
Movement between Oct 2005 and Jan 2007 was negligible, which is the first period since 1996 that the building
had been substantially stable between successive level measurements.
Further suction profile data from SE Queensland are presented in Figure 7, which have been derived from a
further four case studies of groups or rows of trees.

Australian Geomechanics Vol 46 No 3 September 2011 95


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Table 4: Design and Construction Details, SE Qld Cases


1
Case ys Date Dates of
Class Location Construction2
(mm) constructed investigations
Completed Late 1996,
1B 151 E Silkstone MV, duplex, stiffened raft
1993 ongoing
Early and late
Redbank MV, stiffened raft, but Completed
2B 93 E 1988,
Plains poor beam layout 1984
Aug 1989
MV, non-stiffened slab
Completed
3B 49 H Morayfield with perimeter piers, Late 1991
Sep 1985
2 story
MV duplex on a raft slab
Late 1993 and
4B 61 H Thornside with shallow internal late 1991
mid 1998
beams (300 mm)
MV/non-stiffened slab Completed mid 1994, mid
5B 60 H Loganholme
with 900 deep edge beam mid 1990 1996
CF, part slab on ground,
6B 41 H Toowoomba early 1992 mid 1994
rest timber floor stumps
MV, non-stiffened slab
Late 1992,
7B 41 H Morayfield with perimeter strip mid 1989
early 1994
footing
Full masonry units
Completed
8B 30 M Hamilton construction, Jan 1993
early 1988
2 story
MV, timber floor on Completed Early 2003 to
9B 34 M Logan Village
stumps Apr 1995 mid 2005
Notes to Table 4:
1. For calculation of ys, the depth of soil cracking, Hc, was assumed to be 0.5Hs
2. MV = masonry veneer construction, CF = clad frame construction
Table 5: Estimates of Tree Effects and Measured Movements
Depth yt max2 Rebound3
Case 1
Vegetation Dt/HT Hs estimate (mm) after tree removal
(m) (mm) (mm)
Tipuana Tipu (2) &
1B 14 2.3 48 585 42
eucalypt
E. maculata or
2B 0.55 2.3 39 53 47
camaldulensis?
E. maculata or
3B 0.3 2.3 55 63 40
camaldulensis?
Paper bark & 57 30 to 606
4B 0.65 1.5 58
eucalypts (uneven)
Native & exotic
5B 0.37 1.5 57 57 26
trees
58
6B Eucalypt < 0.3 2.3 32 58?7
E. maculata,
33
7B camaldulensis & < 0.3 1.5 30 30
crebria
Native shrubs and
8B <0.3 1.5 48 50 NA
trees
9B Row of trees 0.6 2.3 25 51 23
Notes to Table 5:

1.Trees:
Eucalyptus camaldulensis - commonly known as river red gum
Eucalyptus maculata - commonly known as spotted gum

96 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Eucalyptus crebria - commonly known as narrow-leaved red iron bark


Tipuana Tipu is an exotic from South America. It was planted extensively in suburban areas of Brisbane
in the 1970s and is now considered a weed species.

2. As of early 2000
3. Between July 1996 and November 2000 successive level measurements showed settlement at the tree
end of the building of up to 58 mm as the Tipuana trees grew from a height of 4.5 m to in excess of 8 m
all at a distance of 8.5 m from the building.
4. Except where noted otherwise, max is the maximum level difference on the slab, which will only
exactly reflect actual differential movement if the slab was constructed level to start with.
5. The rebound level change was derived from successive sets of level measurements and thus is
unaffected by any initial out-of-level of the slab from any cause.
6. The recovery was achieved following installation of a vertical root barrier, which may not have
excluded all tree effects.
7. The timber framed floor section of the house was built over the stump of the subject tree. After two
years, the floor was up to 58 mm out of level in a pattern of apparent heave approximately concentric on
the tree location.

Total Suction (pF) Total Suction (pF)

3.0 3.5 4.0 4.5 5.0


2.5 3.0 3.5 4.0 4.5
0
Case 2B: single 0
tree
Dt/HT = 0.55

1
1
Depth (m)

Depth (m)

2
2

Case 3B:
Eucalypt
Dt/HT = 0.3
3
3

Figure 5: Measured and design total suctions - single tree cases, SE Queensland

Total Suction (pF)

2.5 3.0 3.5 4.0 4.5


0

1
Depth (m)

Case 1B: Trees


Dt/HT = 1.05

4 Series1 Series3 BH4 - 2007

Australian Geomechanics Vol 46 No 3 September 2011 97


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Figure 6: Measured and design total suctions - Case Study 1, SE Queensland

Total Suction (pF) Total Suction (pF)

2.5 3.0 3.5 4.0 4.5


3.0 3.5 4.0 4.5 5.0
0
0

1
1

Depth (m)
Depth (m)

3 Case 4B: trees


Case 9B: row Dt/HT = 0.65
of trees
Dt/HT = 0.6 3
4

Total Suction (pF)


Total Suction (pF)

2.5 3.0 3.5 4.0 4.5 2.5 3.0 3.5 4.0 4.5
0 0

1
1
Depth (m)
Depth (m)

Case 5B: trees Case 7B: trees


Dt/HT = 0.5 Dt:HT = 0.2
3
4

Figure 7: Measured and design total suctions tree groups or rows, SE Queensland

6.2 THE CASE STUDIES FROM SOUTH AUSTRALIA AND VICTORIA


The following case studies were reported by Cameron (2001) and are presented here in light of the new
recommendations of AS2870-2011 (refer Figures 8 to 10). As with the South-East Queensland case studies, total
suction profiles near and away from trees are presented with superimposed design suction profiles. The design
suction profiles have been based, wherever possible, on a deep equilibrium suction value, ueq, apparent from the
data from the borehole taken away from the tree influence. Wilting point suctions, uwp, are indicated by dashed
vertical lines. Deduced values of both ueq and uwp are provided in Table 6.

98 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Table 6: Case studies from South Australia and Victoria


utree
ueq uwp
Site Trees (H) Dt/HT (log10MPa or
(MPa) (MPa)
pF)
1A Broadview Gum (9 m) < 0.5 0.98 2.2 0.35
e. Torquata (8.3
2A Greenacres1 0.4 - 0.7 1.15 1.95? 0.23
m)
e. Mannifera
3A Greenacres1 (19.3 m) and 2 0.6 1.15 1.95 0.23
smaller gums
Row of
4A Ingle Farm eucalypts 1.5 1.1 1.95 0.25
(8 m)
Row of
5A Ingle Farm eucalypts 0.4 1.1 2.6 0.37
(10 m)
Row of mature
6A Klemzig < 0.5 1.38 3.4 0.39
eucalypts
Native
7A Hallett Cove 0.1 0.87 1.95 0.35
plantation
Native
8A The Levels < 0.5 0.98 2.45? 0.4
plantation.
1V Williamstown,
Row of gums < 0.5 0.42 0.93 0.35
Victoria

7 DISCUSSION
The concept of wilting point suction is supported by the majority of these case studies from the southern states.
It can be readily seen that values of wilting point suction can be identified from the suction data. Furthermore,
having suction data to sufficient depth near and away from vegetation, a value for the equilibrium suction below
the depth of seasonal influence can be estimated in most cases. Therefore estimates of the relative drying
influence of vegetation on a site can be established and used as a basis for design for tree effects on sites with
similar climates.
The extension of the method for estimating tree induced foundation movement to wetter climates is predicated
on sparse data. The present calibration of the method which is based on the case studies from South East
Queensland, moderates the estimates of tree induced movement in wetter climates. Given that the equilibrium
suction will be lower in areas of wetter climate, and hence further from the wilting point suction, this moderation
of the estimated movement is, on the face of it, at odds with the wilting point concept. Whilst justifications for
the moderation of movement in wetter climates can be proposed, the need for further research into this and other
aspects of tree driven foundation movement is readily apparent.

8 SUMMARY
The extra reactive soil foundation movement caused by tree roots extracting moisture from deep in the soil
profile can be estimated, and footing systems designed accordingly, but with less certainty than we can classify
and design for non-tree affected sites. Accordingly, it is advisable for practitioners to qualify their designs for
tree-affected sites with an appropriate caveat stating the uncertainties and limitations of the resulting designs.
The guidance given in Appendix H of AS2870-2011 have been based in part on the well-established approach of
the Footings Group of South Australia and the field data presented in this paper. Modifications have been
proposed for climates less severe than that of Adelaide and environs. Limited data have been presented from
case studies from South Australia, Victoria and South-East Queensland, which indicate that the new approach
has merit. More data are needed to further substantiate and refine the design approach.

1
data from Jaksa 1998

Australian Geomechanics Vol 46 No 3 September 2011 99


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Total Suction (pF)


Total Suction (pF)

3.0 3.5 4.0 4.5 5.0 3.0 3.5 4.0 4.5 5.0
0 0

1 1

Depth (m)
Depth (m)

2 2

3 3
Case 2A: Tree
Dt:HT = 0.4
4
Case 1A: Tree
Dt/HT < 0.5 4

5
5

Figure 8: Measured and design total suctions - single tree cases, South Australia

Total Suction (pF)

3.0 3.5 4.0 4.5 5.0


0

1
Depth (m)

3
Case 3A:
Tree
Dt:HT = 0.6
4

Figure 8 continued: Measured and design total suctions - single tree cases, South Australia

100 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Total Suction (pF) Total Suction (pF)

3.0 3.5 4.0 4.5 5.0 3.0 3.5 4.0 4.5 5.0
0 0

1
1

2
2
Depth (m)

Depth (m)
3
Case 5A:
3 Row of Trees
Dt/HT = 0.5
4
Case 4A: Row
of Trees
4 Dt/HT = 1.5
5

5
6

Total Suction (pF) Total Suction (pF)

3.0 3.5 4.0 4.5 5.0


3.0 3.5 4.0 4.5 5.0
0
0

1 1

2
2
Depth (m)

Depth (m)

Case 6A: Row of


Trees Dt/HT = 0.4
3
Case 7A:
3 Row of
trees
4

4
5

6 5

Figure 9: Measured and design total suctions cases involving groups or rows of trees, South Australia

Australian Geomechanics Vol 46 No 3 September 2011 101


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

Total Suction (pF)

3.0 3.5 4.0 4.5 5.0


0
Case 8A:
Group of trees
Dt:HT < 0.5
1

2
Depth (m)

Figure 9 continued: Measured and design total suctions cases involving groups or rows of trees,
South Australia

Total Suction (pF)

3.0 3.5 4.0 4.5 5.0


0

Case 1M:
Row of trees
Depth (m)

Dt:HT < 0.5


2

Figure 10: Measured and design total suctions case of a row of trees, Victoria
It is recognised that the approach set out in Appendix H of AS2870-2011 does have limitations. However the
method does take into account the wilting point suction concept, which is a major determinant of the maximum
effect that vegetation can exert on the soil moisture regime. For Adelaide and similar semi-arid climates, a
conservative value for (uwp-ueq) of 0.5pF has been adopted for all species. The values of (uwp-ueq) for wetter
climates are reduced. In wetter climates, if the tree or trees actually do dry the soil to the wilting point value,
then the value of (uwp-ueq) may exceed the values adopted in the proposed method, in which case the method will
likely underestimate the tree induced movement.

102 Australian Geomechanics Vol 46 No 3 September 2011


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

9 FURTHER RESEARCH
Much more research needs to be done to investigate the tree induced suction change and the depth over which it
occurs in a wide range of climates. Fundamental research is being undertaken to monitor water demand of trees
and the consequent effect on ground movement. However to improve the guidance for practitioners in the short
term, there is an urgent need to extend the database of case studies, particularly case studies including suction
data, as they can provide usable information, relatively quickly.

10 REFERENCES
Beal, N. and Cameron, D. A. (2007). A method for evaluating the influence of trees on expansive soil movement
in light of case studies from SE Queensland. Proc. 10th ANZ Conference on Geomechanics, Common
Ground, Brisbane, Oct 2007, V2, pp 200-205.
Beal, N. S (1993). Techniques for using a water level apparatus as applicable to the investigation of building
foundation movements and structural distress. Australian Civil Engineering Transactions, V CE35 (2),
pp 107-111.
Bozozuk, M. (1962). Soil shrinkage damages shallow foundation at Ottawa. Engineering J, V45 (7), July, pp 33-
37.
Cameron, D. A. (2001). The extent of soil desiccation near trees in a semi-arid environment. Int. J. Geotechnical
and Geological Engineering, Kluwer Academic Publishers, V19 (3 & 4), pp 357-370.
Cameron, D. A. and Mills, K. G. (2010). Tree water use, soil suctions and reference evapotranspiration in a
semi-arid climate. Proc. 4th Asia-Pacific Conf. Unsaturated Soils, eds Buzzi, O., Fityus, S. and Sheng,
D., Taylor & Francis, London, Newcastle, Australia, Nov, V1, pp 295-300.
Correia, M. J., Torres, F. and Pereira, J. S. (1989). Water and nutrient supply regimes and the water relations of
juvenile leaves of Eucalyptus globulus. Tree Physiology, Heron Publishing, Victoria, Canada, V5, pp
459-411.
Fityus, S., Cameron, D. and Driscoll, C. (2007). Observations of root architecture and their implications for
modelling water movements in partially saturated soils. Proc Third Asian Conference on Unsaturated
Soils (eds: Z. Yin, J. Yuan and A. F. Chiu), Science Press, China, April, pp 207-212.
Footings Group. (1990). Special Provisions for the Design of Residential Slabs and Footings. IEAust SA,
August.
Fredlund, D. G. and Hung, V. Q. (2001). Prediction of volume change in an expansive soil as a result of
vegetation and environmental changes. ASCE, Geotech. Sp Publication 115, Expansive Clay Soils and
Vegetative Influence on Shallow Foundations, eds. Vipulanandan, C., Addison, M B. and Hasen, M.,
24-43.
Hemmati, S. and Gatmiri, B. (2008). Numerical modelling of tree root water-uptake in a multiphase medium.
Proc, First European Conf on Unsaturated Soils: Advances in Geo-Engineering, eds, D. G. Toll, C. E.
Augarde, D. Gallipoli and S. J. Wheeler, Taylor and Francis, pp 785-790.
Indraratna, B., Fatahi, B., Khabbaz, H. (2006). Numerical analysis of matric suction effects of tree roots. Proc.
ICE, Geotechnical Engineering, 159 (2), April, pp 77 90.
Jaksa, M. B., Kaggwa, W. S. and Woodburn, J. A. and Sinclair, R. (2002). Influence of Large Gum Trees on the
Soil Suction Profile in Expansive Soils. Australian Geomechanics, V37 (1), March, pp 23-33.
McJannet and Finch (2004). A flexible and easy constructed heat pulse system for monitoring sap flow in trees.
CSIRO Land & Water, Tech Report 39/40, Dec.
McKeen, R. G. (1992). A model for predicting expansive soil behaviour. Proc., 7th Int. Conference on
Expansive Soils, Dallas, V1, pp 1-6.
Navarro, V.,Candel, M., Yustres, A., Snchez, J. and Alonso, J. (2009). Trees, soil moisture and foundation
movements. Computers and Geotechnics, 36, pp 810818.
OMalley, A. P. K and Cameron, D. A. (2005). The Influence of Trees on Soil Moisture, Dwellings and
Pavements in an Urban Environment. Report to Local Government Association and City of Salisbury,
November, 90 pages,
http://www.lga.sa.gov.au/webdata/resources/project/The_Influence_of_Trees_on_Soil_Moisture_Dwell
ings_and_Pavements_in_an_Urban_Environment_-_2005_Report_-_Section_1.pdf.
Standards Association of Australia (1986). Residential Slabs and Footings. AS2870-1986, Standards House,
North Sydney, NSW.
Standards Association of Australia (1988). Residential Slabs and Footings. Part 1: Construction. AS2870.1-
1988, Standards House, North Sydney, NSW.
Standards Australia (2011). Residential Slabs and Footings. AS2870-2011, Sydney, NSW.
Ward, W. H. (1953). Soil movements and weather. Proc. 3rd Int. Conf. Soil Mech., Zurich, V2, 477-481.
Yeagher, A.F. (1935). Root systems of certain trees and shrubs grown on prairie soils. J Agricultural Research
51(12):1085-1092.

Australian Geomechanics Vol 46 No 3 September 2011 103


ESTIMATION OF FOUNDATION MOVEMENT AND DESIGN OF FOOTING SYSTEMS ON REACTIVE SOILS
FOR THE EFFECTS OF TREES CAMERON & BEAL

104 Australian Geomechanics Vol 46 No 3 September 2011

Вам также может понравиться