assessment of the impact of climate change on ...

ASSESSMENT OF THE IMPACT OF CLIMATE CHANGE ON EXPANSIVE SOIL MOVEMENTS AND SITE CLASSIFICATION Xi Sun, Jie Li and Annan Zhou School of Engineering, RMIT University, Melbourne, Victoria, 3001

ABSTRACT There is overwhelming evidence that climate change leads to a wide range of climatic and weather changes that can affect the performance of built infrastructures. Climate change is likely to have significant impacts on the performance of residential buildings constructed on expansive soils. The Thornthwaite Moisture Index (TMI) as a useful climate parameter has been widely employed to estimate the depth of design soil suction change (Hs) which is needed for the determination of characteristic ground movement (ys). Precipitation and temperature are the primary weather parameters required for the TMI computation. By applying the projected rainfall reduction and temperature increase in 2030, 2050 and 2070 in the TMI calculations, the effects of climate changes on expansive soil movements and site classification can be quantified by the use of the predicted TMI. In this study, TMI values of various areas of Victoria were calculated under A1B and A1FI emission scenario using climate projections generated from 23 climate models. These predicted TMI indices were then used to delineate TMI isopleth lines on the map of Victoria to visualise and compare climate conditions in 2030, 2050 and 2070. A case study was also carried out to assess the effect of climate changes on the magnitude of ground surface movements in the top five most densely populated cities (i.e. Sydney, Melbourne, Brisbane, Perth and Adelaide) in Australia for three specific years (i.e. 1990, 2030 and 2070). The results show that both Hs and ys values are expected to increase significantly with climate change.

1

INTRODUCTION

Climate change is defined as “a change of climate which is attributed directly or indirectly to human activity that alters the composition of the global atmosphere and which is in addition to natural climate variability observed over comparable time periods” (UNFCCC, 1992). There is a lot of evidence revealing that the earth’s climate is changing and a number of climate scientists believe that some of these changes are the result of human-induced effects, which is attributable to greenhouse gas emissions (GHGs). Carbon dioxide (CO2) is the most crucial anthropogenic GHGs. The global mean CO2 level for 2016 is almost certain to be higher than 400 parts per million (ppm), a 44 percent increase from pre-industrial (1750) concentrations (278ppm). Current climate models predict that the steady increase of global average annual carbon dioxide (CO2) will, directly and indirectly, impact the performance of infrastructures, including deterioration of paints, sealants and other protective surface finishes as a result of extreme temperatures; a significant increase in the heating and cooling energy requirement of residential houses and a higher rate of the road pavement deterioration. In Australia, climate change over the coming decade is anticipated to result in increasing temperature, drier conditions and increased frequency of extreme weather events such as bushfires and droughts (CSIRO, 2007). Climate change-induced drought has a significant impact on the performance of residential buildings. From late 1997 to early 2010, an extended and serious drought affected much of Australia, including the Melbourne area. In December 2011, the Housing Industry Association (HIA) estimated that more than 1000 new houses in the western suburbs of Melbourne were damaged due to soil heave. These new houses have all been built in drier weather between 2003 and 2010 and have been subjected to larger ground differential movements induced by abnormal moisture changes after the construction of gardens/lawns and watering system around buildings and the breaking of the long drought in 2011 (Li and Sun, 2015). Climate affects the extent of soil moisture variation likely to occur through time and hence the amount of foundation movement likely to occur on moisture reactive foundation soils. This study examines the effects of climate changes (the projected changes in temperature and precipitation for 2030, 2050 and 2070) on the depth of design soil suction changes (Hs) that can be estimated by using the Thornthwaite Moisture Index. TMI is a measure of the water supply (rainfall) at a site relative to the climatic demand for water (potential evapotranspiration). As a useful climate tool, it has been employed by a number of researchers for the delineation of climate zoning maps. In 1948, TMI values of different locations across the United States were calculated by Thornthwaite, and these values were then used for the delineation of the first TMI isopleth map for the United States (Thornthwaite, 1948). In 1965, the first TMI contour map of Australia was produced by Aitchison and Richards (1965). Based on this map, Smith (1993) developed a TMI contour map for Victoria by employing the same TMI values that were derived by Aitchison and

AUSTRALIAN GEOMECHANICS | VOLUME 52: NO.3 SEPTEMBER 2017


39

ASSESSMENT OF THE IMPACT OF CLIMATE CHANGE ON EXPANSIVE SOIL MOVEMENTS AND SITE CLASSIFICATION

SUN ET AL

Richards (1965). This map has been endorsed by the Australian Residential Slabs and Footings Standard AS2870 (1996 and 2011) and used with minor modification as guidance for design engineers to infer the depth of design soil suction change, Hs, which is required for estimation of the characteristic surface movement, ys. However, this map is outdated since it was produced using climate data for the period of 1940-1960 (McManus et al. 2004) and may not reflect current climate conditions in Victoria. A few different versions of the TMI based contour map of Victoria have been published over the last two decades. McManus et al. (2003) created TMI maps for Victoria using the climatic data for the period of 1961 to 1990; Lopes and Osman (2010) calculated TMI using the 1948-2007 data and Leao and Osman (2013) used weather data from 1913 to 2012. More recently, Li and Sun (2015) produced three climate zoning maps of Victoria over the period of 19542013. In this study, the projected temperature increase and rainfall decrease are used to develop the TMI isopleth maps of Victoria for 2030, 2050 and 2070.

2

THORNTHWAITE MOISTURE INDEX

The Thornthwaite Moisture Index (TMI) was derived from P-E Index (i.e. precipitation minus evaporation) by Thornthwaite (1948) to supersede the means of climate classification, which was originally categorised in terms of general descriptive observations. This climate parameter has been widely used in geotechnical engineering and other disciplines to evaluate the soil moisture changes in the unsaturated zone and predict the depth of seasonal moisture changes. In Australia, TMI index is mainly used for classifying climate zones and estimating the depth of design soil suction changes. TMI is a number, indicating the degree of aridity for an area and is particularly useful in quantifying variations in climate. TMI is an annual index and is computed for a statistically significant number of years to allow a true climatic condition to be obtained for a particular locality. A positive TMI presents a humid climate with a surplus of precipitation and generally high soil moisture, while a negative TMI presents an arid climate with a deficit of precipitation relative to potential evapotranspiration (PET) and generally low moisture in the soil. Zero TMI index means that, over the long term, under average conditions, the inflow from precipitation is just equivalent to the loss of soil moisture through evapotranspiration. TMI is a function of aridity and humidity index, which requires both precipitation and temperature data to be available for consecutive periods of time and is calculated from the moisture deficiency and surplus by conducting a water balance approach. The TMI has traditionally been calculated for the statistically significant number of years using historical records of climate data and assuming that past events are statistically representative of what could happen in the future (Thornthwaite, 1948). However, it seems this assumption must be revised by taking into account the effect of climate changes (e.g. the expected changes in precipitation and temperature). The original TMI calculation equation (Thornthwaite, 1948) is expressed as:

TMI= I h − 0.6 I a

(1)

where, Ih and Ia are respectively humidity and aridity indices, and are determined as follows:

= Ih

100R 100D = Ia PET PET

(2)

where, R represents the moisture surplus or runoff (mm), D represents moisture deficit (mm), and PET is the adjusted potential evapotranspiration (mm) and represents the water need. As shown in Equation 1 and 2, TMI is computed by three parameters; (a) monthly moisture surplus (R), (b) monthly moisture deficit (D) and (c) monthly adjusted potential evapotranspiration (PET). The first two parameters can be derived using the water balance approach and the calculation procedure is described by McKeen and Johnson (1990). Initial (S0) and maximum (Smax) water storage values are needed in order to carry out the water balance analysis. However those values are usually unknown and have to be assumed. The adjusted potential evapotranspiration for the month, i (PETi) is often determined by the Thornthwaite (1948) evapotranspiration equation (Equation 3) since it requires only temperature data which are readily available from most of the weather stations in Australia. DN  PETi = ei  i i   30 

(3)

where Di is the day length correction factor for the month i; Ni is the number of days in the month i; and ei is the nonadjusted potential evapotranspiration (cm) for the month i given as:


40



 10t ei = 1.6  i  Hy 

  

SUN ET AL

a

(4)

where ti is the mean monthly temperature in °C and calculated as the average of tmax and tmin. The heat index for month i is determined as follows: hi = ( 0.2 ti )

1.514

(5)

The annual heat index, Hy is simply calculated by summing the 12 monthly heat index values. The power term a in Equation 4, is given as: a= 6.75 × 10 − 7 H y3 − 7.71 × 10 −5 H y2 + 0.017921H y + 0.49239

and

(6)

0 < a < 4.25

PET plays a critical role in TMI calculation since TMI indices can vary greatly if different PET estimation methods are employed. A relatively simple temperature-based PET model (Thornthwaite, 1948), which was developed based upon global climate pattern distribution and the concept of plant physiology relating to moisture availability (Jewell and Mitchell 2009), has been widely used by many researchers including Fityus et al. (1998); Fox (2000, 2002); Chan and Mostyn (2008, 2009); Jewell and Mitchell (2009); Mitchell (2012, 2013); Er and Rifat (2014) ; Li and Sun (2015) and Karunarathne et al. (2016). A review and comparison of different assumptions and methods for calculating TMI can be found in Sun et al. (2017).

3

THE CORRELATION BETWEEN TMI AND DEPTH OF DESIGN SOIL SUCTION CHANGE

The depth of design soil suction change Hs, is the depth below which has no significant seasonal suction changes and hence no soil volume changes occur in response to climate. Hs data obtained from field measurement are very rare and regionally specific, which means its application range is very limited. Ideally, the value of Hs is determined by long-term site monitoring of soil profile behaviour which involves extremely dry and wet conditions. However collecting data of ground movement, soil suction and moisture changes may take years even decades, which is not only time-consuming but also a tough and challenging work to perform. Fortunately it has been found that Hs correlates well with the climate index, TMI (Aithison and Richards 1965). In AS2870, the approach which was originally proposed by Smith (1993) is adopted to estimate Hs (Fityus and Olivier 2008). AS2870 (1996) recommended that at least 20 continuous years of climate data are needed for the estimation of Hs based on Table 1 while the current version (AS2870, 2011) suggests using 25 continuous years of climate data records as the datum for TMI calculation. It is worth noting that the theoretical basis for TMI-Hs relationship is limited and the correlations outlined in AS2870 are based more on anecdotal evidence and empirical experience rather than a large quantity of scientific research. Therefore further work needs to done to investigate a better correlation between TMI and Hs. Table 1: Correlation between TMI, Hs and climatic zone (AS2870, 1996 and 2011) Climatic Zone

Description

1 2 3 4 5 6

Alpine / Wet coastal Wet temperate Temperate Dry temperate Semi-arid Arid

AS2870(1996) > +40 +10 to +40 -5 to +10 -25 to -5 < -25 -



TMI

Hs (m)

AS2870(2011) > +10 -5 to +10 -15 to -5 -25 to -15 -40 to -25 < -40

1.5 1.8 2.3 3.0 4.0 > 4.0

3

41


SUN ET AL

Surface soil suction change ∆us =1.2pF for all climate zones

Hs =1.5m (Zone 1) Hs =1.8m (Zone 2) Hs =2.3m (Zone 3) Hs =3.0m (Zone 4)

Hs =4.0m (Zone 5) Hs >4.0m (Zone 6)

Figure 1: Typical soil suction change profiles for climate zones (AS2870, 2011). An additional climate Zone 6 was added to Australian Standard AS2870 (2011) edition, indicating the arid climate zone where it’s corresponding Hs could exceed 4 m due to the severe aridity of the climate. The threshold of climate Zone 1 has been changed from TMI >+40 (AS2870, 1996) to TMI >+10 (AS2870, 2011). This amendment is based on the field observation that ground movement is very unlikely to arise in a wet and humid climate (Lopes and Osman 2010). Climate Zone 3 with TMI ranging from -5 to +10 in the 1996 edition has been re-classified as Zone 2 (refer to Table 1) in the AS2870 (2011). In addition, TMI of -15 has been introduced as a new threshold to reduce the TMI range from -25 ≤TMI ≤-5 (Zone 4, 1996 edition) to -15 ≤TMI ≤ -5 for climate Zone 3 (2011 edition) and -25 ≤TMI ≤ -15 for climate Zone 4 (2011 edition) respectively. Table 1 reveals that a range of TMI index corresponds to a specific value of Hs rather than a scoped manner. For example, any TMI index that falls in -15 ≤TMI ≤ -5 corresponds to a Hs depth of 2.3m rather than a range (e.g. 1.8 ≤Hs ≤ 2.3). This would result in a discontinuous Hs isopleth on a map of TMI contours. For instance, a site with a TMI of -15 corresponds to a Hs value of 2.3 m while an adjacent site with a TMI value of -16 has a Hs value of 3 m in 2011 edition. This indicates Hs may increase abruptly across a TMI isopleth. This shortcoming has been noted by a number of researchers. Fityus et al. (1998) proposed a revised correlation to ensure Hs is continuous across an isopleth, which allows Hs values to be interpolated with a known TMI value. Chan and Mostyn (2008) also proposed a curvilinear relationship between TMI and Hs.

4

CHARACTERISTIC SURFACE MOVEMENT

The Australian Residential Slabs and Footings Standard AS2870 was first published in 1986, then revised and published in two parts in 1988 and 1990, followed by two more complete editions in 1996 and 2011. Common to all five versions of the Standard, sites are classified according to soil profile and regional climate influence on soil moisture state (Li et al. 2014). The site classification for reactivity is based on ys, the estimated design site surface movement, which is based on design soil suction change profiles for different climatic regions of Australia (Li and Cameron, 2002). Australian Standard AS2870 introduced a soil suction based method for estimating ys, which has been widely employed by geotechnical engineers with success over the last two decades. For sites where ground movement is predominately due to soil reactivity, and where abnormal moisture conditions are not expected, the standard provides for sites to be categorized into one of five classes in accordance with Table2. M, H1, H2 and E site can be further classified which incorporate with deep-seated moisture changes (denoted by ‘D’) if the corresponding Hs is equal or greater than 3 m. ys is the expected free surface movement (e.g. without the presence of building and excluding load effect) and it is the moisture variation induced vertical movement of the ground surface in a


42



SUN ET AL

reactive soil site. It has less than 5% chance of exceeding ys in the life of the structure (e.g. 50 years). The proposed equation for ys determination is as follows: = ys

1 N ∑ I pt uh 100 n =1

(

)

(7)

n

Where, Ipt is the instability index and can be determined from Ips in the laboratory tests, namely shrink-swell, core shrinkage or loaded shrinkage test. A shrinkage index of 4%/pF would be regarded as a highly expansive soil, 6%/pF very highly expansive and 8%/pF, an extremely expansive soil (Li et al. 2016). ∆u is the soil suction change averaged over the thickness of the layer under consideration, the maximum surface suction change ∆us is 1.2pF for Victoria. h is the thickness of layer and N is the number of layers within the depth of Hs. Values of Hs for various locations in Australia are summarized in AS2870 and for other locations, the site classifier has to have recourse to a Thornthwaite Moisture Index (TMI) map and use Table 1 to obtain the depth of design suction change Hs. Table 2: Site classification by ys (AS2870, 2011)

Characteristic surface movement (ys) mm 0 < ys ≤ 20 20 < ys ≤ 40 40 < ys ≤ 60 60 < ys ≤ 75 ys > 75

5

Site classification

Description

S M H1 H2 E

Slightly reactive Moderately reactive Highly reactive Very Highly reactive Extremely reactive

TMI CALCULATION FOR VICTORIA IN 2030, 2050 AND 2070

Future climate changes cannot be simply extrapolated from past climate. A common tool used for projecting climate change is a climate model which is a mathematical representation of the Earth’s climate system. To project spatially dependent climate under different emission scenarios, 23 climate models from the Coupled Model Inter-comparison Project 3 (CMIP3) database have been used to generate projections of future climate for Australia (CSIRO, 2007). A total of 40 emission scenarios have been developed under Special Report on Emissions Scenarios (SRES) (IPCC SRES, 2000). A1FI (fossil fuel intensive) may be considered as the extreme high greenhouse gas emissions (GHGs) scenario, A1B (balanced) is the medium, and A1T (predominantly non-fossil fuel) is the low emission scenario. Probability distributions have been employed to present the uncertainties caused by the differences among results of climate models and 50th percentile or the median gives the best estimate. The primary meteorological parameters governing TMI are precipitation and potential evapotranspiration which can be derived from local temperature data. The projected climate trends from 23 climate models provide by CSIRO, BOM and DCCEE (2014) are applied to estimate precipitation and temperature needed for the calculation of TMI for 49 stations for Victoria in 2030, 2050 and 2070. Mid-range emission scenario (A1B scenario) in 2030 is adopted since there are marked variations among results of climate models due to near-term changes in climate, while climate changes centred on 2050 and 2070 are more dependent on the greenhouse gas emissions scenario and thus the extreme climate change scenario (A1FI scenario) is employed. This paper introduced a method for predicting and quantifying climate condition in 2030 (A1B scenario), 2050 (A1FI scenario) and 2070 (A1FI scenario) for Victoria using TMI. 5.1

METHODOLOGY

Step 1: TMI calculation in 1990 TMI values of various weather stations across Victoria in 1990 were calculated using the purchased climate data (i.e. monthly precipitation and monthly temperature) from the Bureau of Meteorology (BOM). It should be pointed out that although there are 204 accessible weather stations on the BOM website, only 49 stations across Victoria have the fully recorded climatological data required for this study. The Thornthwaite PET model (i.e. Equation 3) is adopted and the original Thornthwaite equation (i.e. Equation 1) is employed for TMI calculation. The initial (S0) and maximum water storage (Smax) were taken to be 0 mm and 100 mm respectively.



5

43


SUN ET AL

Step 2: Climate trend determination The best estimate (i.e. 50th percentile) of the projected seasonal precipitation and temperature trend maps for Victoria in 2030 (A1B scenario), 2050 (A1FI scenario) and 2070 (A1FI scenario) was obtained from ‘Climate Change In Australia’ website (CSIRO, BOM and DCCEE, 2014). 49 weather stations of Victoria were plotted on the projected seasonal trend maps based upon their coordinates (i.e. latitude and longitude) so that the range of seasonal variation trend for each weather station can be obtained in accordance with the given colour scale bar (Figure 2). Seasonal average variation trend (i.e. the average of the maximum and minimum threshold of the variation trend) was used in the calculation.

Figure 2: The projected summer precipitation map for Victoria in 2030 under A1B emission scenario. (CSIRO, BOM and DCCEE, 2014). Step 3: TMI calculation in 2030, 2050 and 2070 Apply seasonal average precipitation and temperature trend to weather condition in 1990 to allow precipitation and temperature in 2030, 2050 and 2070 to be obtained. These predicted weather parameters were then used for TMI computation.

6

THE DELINEATION OF TMI ISOPLETH MAP FOR VICTORIA

The calculated TMI indices were employed for the delineation of isolines on the map of Victoria for 2030, 2050 and 2070 by the use of Surfer® 11 (2012). Climate zoning for each map is also established for comparison of changes. TMI isopleth maps presented in Figure 3-5 have confirmed the belief of many local geotechnical engineers that the climate of Victoria will become warmer and drier, and a significant increase in clay soil movement is expected for the State. It seems the most noticeable climate change will occur in 2070 due to the significant TMI variation compared to TMI map for 2030 and 2050. This extremely arid climate may lead to an acceleration of soil desiccation. Generally speaking, the driest zone (i.e. Zone 6) spread out over the north-west region of the State while the wettest zone (i.e. Zone 1) is mainly distributed along the southern and eastern coastal areas. The substantially negative TMI values present in the north-west part are due primarily to the prolonged and high-frequency drought whereas large positive TMI values occuring in southern regions can be attributed to a humid coastal climate. It is worth noting that the significant growth of aridity can also affect Melbourne Metropolitan areas as the prevailing climate Zone 3 and Zone 4 in 2030 will suffer an increased growth of drying and thus superseded by Zone 5 in 2070. The soil in south-western suburbs of Melbourne such as Laverton and Avalon is expected to experience the growth of desiccation compared to those in the north, north-east and south-east. This implies a greater depth of design soil suction change and a higher incidence of excessive ground surface movements.


44



SUN ET AL

Figure 3: TMI contour map for Victoria in 2030.

Figure 4: TMI contour map for Victoria in 2050.



7

45


SUN ET AL

Figure 5: TMI contour map for Victoria in 2070. The variations in TMI isolines in Figures 3-5 reveal: • • • • • •

Climate Zone 6 (Arid climate) at the far north-west is expected to suffer severe aridity and it expands further towards the south and east to include Kerang, Nhill, Warracknabeal, Horsham and Longerenong. Climate Zone 5 (Semi-arid climate) expands to its southern direction to include Bendigo, Maryborough, Redesdale, Castlemaine and Geelong. Melbourne CBD and its inner south-west suburbs (i.e. Avalon, Laverton) which are predominated by climate Zone 4 are superseded by Zone 5. This leads to a marked reduction of Zone 4. Climate Zone 4 (Dry temperate climate) has a significant contraction and move towards the east and south-west to include Casterton, Hamilton, Ararat and Benalla. The prevailing temperate climate (i.e. Zone 3) in East Sale areas is replaced by Zone 4. Climate Zone 3 (Temperate climate) moves to south and south-west coastal areas to include Portland and Cerberus. Climate Zone 2 (Wet temperate climate) extends further towards the south and east to include Port Fairy, Beechworth, Coldstream and Scoresby. There is a marked expansion in south-east inland and east coastal areas to include Omeo and Point Hicks. Climate Zone 1 (Wet coastal climate) retract to the coast in Warrnambool areas and a notable contraction takes place in the east part of the State.

7

ASSESSMENT OF IMPACTS OF CLIMATE CHANGES ON YS

To assess and compare the effect of climate changes on the amount of characteristic ground surface movements ys, a case study has been conducted for the top five largest cities in Australia (i.e. Sydney, Melbourne, Brisbane, Perth and Adelaide) in three particular years (i.e. 1990, 2030 and 2070). The Thornthwaite model (i.e. Equation 3) was adopted for PET calculation and the original Thornthwaite equation (i.e. Equation 1) was employed for TMI computation. The initial water storage (S0) of 0 mm and maximum water storage (Smax) of 100 mm were adopted. Meteorological data used to calculate TMI were extracted from the following weather observation stations: Sydney Observatory Hill (066062), Melbourne Regional Office (086071), Brisbane Aero (040223), Perth Regional Office (009034) and Adelaide Kent Town (023090). The summary of climate projections for Australia’s major cities based on the best estimate (i.e. the 50 percentile) and a given emission scenario in 2030 and 2070 is shown in Table 3. It can be seen that the projected temperature will rise across Australia with a increase of approximately 1°C and 3°C in 2030 and 2070 respectively and this pattern varies little seasonally. Different from the temperature pattern which is always increasing, the trend of precipitation change exhibits both increase and decrease. As presented in Table 3, most cities are expected to experience fewer rainfall events in all


46



SUN ET AL

seasons except for the summer period in Sydney, which has a precipitation growth of 1% in 2030 and 2% in 2070. It is also noted the projected precipitation changes vary seasonally and there is a significant decrease in winter and spring precipitation. Table 3: Temperature and precipitation projections for Australia’s major cities Major City

Climatic Variables Temperature (°C)

Sydney Precipitation (%) Temperature (°C) Melbourne Precipitation (%) Temperature (°C) Brisbane

Precipitation (%) Temperature (°C)

Adelaide Precipitation (%) Temperature (°C) Perth Precipitation (%) Note:

Season

2030 A1B 50p

2070 A1FI 50p

Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter

+1 +1 +0.9 +0.8 -6 +1 -2 -5 +0.9 +1 +0.8 +0.7 -7 -1 -2 -4 +1 +0.9 +0.9 +1 -6 -1 -3 -6 +0.9 +0.9 +0.9 +0.8 -8 -2 -1 -6 +0.9 +1 +0.8 +0.7 -7 -1 -2 -4

+3.3 +3.1 +3 +2.6 -17 +2 -6 -16 +2.9 +3.1 +2.7 +2.2 -21 -4 -5 -12 +3.2 +3 +3 +3.1 -18 -3 -9 -18 +3 +3 +2.8 +2.4 -23 -5 -4 -19 +2.9 +3.1 +2.7 +2.2 -21 -4 -5 -12

Climate projections above are relative to average rainfall and temperature in 1990, except for days over 35°C



9

47


SUN ET AL

Based on climate condition in 1990, precipitation and temperature values were estimated using the projected climate trend provided in Table 3 to allow TMI for all cities in 2030 and 2070 to be computed. The depth of design soil suction change Hs was estimated according to TMI-Hs correlations outlined in Table 1 (AS2870, 2011). Surface soil suction change (Δus) of 1.2 was adopted for all cities in accordance with AS2870 (2011). The characteristic ground surface movements estimation method (i.e. Equation 7) was employed for ys determination. For simplicity, the soil profile is assumed to be uniform clay of Ips = 4.0%/pF. As shown in Table 4, all cities except for Sydney will experience an increased ground surface movements as the result of climate change. Of the five sites studied in the ys determination, a highly reactive (H1) site is assigned for Brisbane and Sydney where ys of 49mm is expected for all three studied years. The same magnitude is estimated for Brisbane in 1990 and 2030 but will increase to 51mm in 2070. The substantially negative TMI value of -35 for Adelaide is estimated in 2070, reflecting a marked aridity in its climate, and the resulting excessive foundation movement of 100mm is predicted which rank top among all the cities. Site class for Melbourne (H2) and Perth (H1) remains unchanged in 1990 and 2030, but both cities are expecting greater potential soil movement in 2070 and are thus superseded by E and H2 site respectively. Table 4: The comparison of ys in 1990, 2030 and 2070 Variable City

TMI

Interpolated Hs(m)

ys (mm)

Site Class

1990

2030

2070

1990

2030

2070

1990

2030

2070

1990

2030

2070

Sydney

126

110

76

1.50

1.50

1.50

49

49

49

H1

H1

H1

Melbourne

-11

-16

-24

2.10

2.37

2.93

65

72

85

H2

H2

E

Brisbane

42

31

7

1.50

1.50

1.56

49

49

51

H1

H1

H1

Adelaide

-22

-26

-35

2.79

3.07

3.67

82

88

100

E

E-D

E-D

Perth

1

-5

-17

1.68

1.80

2.44

54

57

74

H1

H1

H2

Figure 6 shows the predicted TMI, Hs and ys in 2030, 2050 and 2070. Generally speaking, climate change is expected to lead to a remarkable decrease in TMI and a significant increase in both Hs and ys. From Figure 6, it can be seen that the calculated TMI for Sydney in 1990 is extremely high compared to other cities. This can be attributed to frequent and intensive precipitation events. 150

110

4

120

100

3.5

90

90

TMI

60 30

2010 Melbourne

2030

2050

2070

Brisbane

Adelaide

Perth

Year

1 1990 Sydney

80 70 60

1.5

-30

Sydney

2.5

2

0

-60 1990

ys (mm)

Hs (m)

3

50

2010 Melbourne

2030

Year

Brisbane

2050 Adelaide

2070 Perth

40 1990 Sydney

2010 Melbourne

2030

2050

2070

Brisbane

Adelaide

Perth

Year

Figure 6: TMI, Hs and ys for all the cities.


48



8

SUN ET AL

CONCLUSIONS

The use of the climate index, TMI, as a predictor of the depth of design soil suction changes (Hs) is now widely accepted by geotechnical engineers and practitioners. TMI is mainly a function of precipitation and temperature. Therefore the rainfall reduction and temperature increase obtained by the use of the projected climate trend derived from 23 climate models (CSIRO, BOM and DCCEE, 2014) can be used to assess the effects of changes in climate (temperature and precipitation) on TMI and Hs. This paper introduced a method for predicting TMI and Hs in 2030, 2050 and 2070. Calculated TMI indices were employed for the delineation of three TMI isopleth maps for Victoria in 2030, 2050 and 2070. These maps show an overall significant growth of drying for Victoria, where the most noticeable increase of aridity is expected to occur in 2070. The general decrease in TMI value across various areas of Victoria indicates a marked reduction in average soil moisture availability. This implies that residential slabs are likely to experience greater ground movements due to the greater depth of Hs, which in turn may result in a higher incidence of slab edge heave and an increase in the occurrence of distortion of residential buildings built on expansive soils. A case study has also been performed to assess and compare the effect of climate changes on the characteristic ground surface movements (ys) for Sydney, Melbourne, Brisbane, Perth and Adelaide. The results show that a significant increase in the predicted ys value is expected in 2070. This implies that further modification of the Australian Residential Slabs and Footings Standard AS2870 would be required to accommodate anticipated future changes in climate.

9

ACKNOWLEDGEMENT

This research was funded partially by the Australian Research Council via the ARC Linkage Grant No. LP16160100649.

10

REFERENCES

Aitchison, G.D. and Richards, B.G. (1965) A broad-scale study of moisture conditions in pavement subgrades throughout Australia: Parts 2 & 3. Moisture Equilibria & Moisture Changes in Soils Beneath Covered Areas. Butterworths, Sydney. AS2870, (1996) Australian Standard: Residential Slabs and Footings. Standards Australia. AS2870, (2011) Australian Standard: Residential Slabs and Footings. Standards Australia. Chan, I. and Mostyn, G. (2008) Climatic factors for AS2870 for the Metropolitan Sydney area. Australian Geomechanics. Vol. 43, No. 1, p 17-28. Chan, I. and Mostyn, G. (2009) Climatic factors for AS2870 for New South Wales. Australian Geomechanics. Vol. 44, No. 2, p 41- 46. CSIRO and BOM. (2007) Climate change in Australia. Technical Report, 140 p. CSIRO, BOM and DCCEE. (2014) Climate change in Australia. [ONLINE] Available at: http://www.climatechangeinaustralia.gov.au/[Accessed 01 October 2014]. Er, Y. and Rifat, B. (2014) Evaluation of climatic factors for the classification of oklahoma pavement regions. GeoCongress 2014 Technical Papers. p 4037-4046. Fityus, S.G., Walsh, P.F. and Kleeman, P W. (1998) The influence of climate as expressed by the Thornthwaite index on the design depth of moisture change of clay soils in the Hunter Valley. Conference on Geotechnical Engineering and Engineering Geology in the Hunter Valley, Conference Publications, Springwood, Australia. p 51-265. Fityus, S. and Buzzi, O. (2008) On the use of the Thornthwaite Moisture Index to infer depths of seasonal moisture change. Australian Geomechanics, Vol. 43 (4), p 69-76. Fox, E. (2000) A climate-based design depth of moisture change map of Queensland and the use of such maps to classify sites under AS 2870-1996. Australian Geomechanics. Vol. 35, No. 4, p 53-60. Fox, E. (2002) Development of a map of Thornthwaite Moisture Index Isopleths for Queensland. Australian Geomechanics. Vol. 37, No. 3, p 51-59. Holland, J.E. (1981) The Design, performance and repair of housing foundations, Swinburne Institute of Technology, Victoria, Australia. IPCC SRES (Special Report on Emissions Scenarios), (2000) Special Report on Emissions Scenarios: A Special Report of Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. 599 p. Jewell, S.A. and Mitchell, P.W. (2009) The Thornthwaite Moisture Index and seasonal soil movement in Adelaide, Australian Geomechanics. Vol. 44 (1), p 59-67.



11

49


SUN ET AL

Karunarathne, A.M.A.N., Gad, E.F., Disfani, M.M., Sivanerupan, S. and Wilson, J.L. (2016) Review of calculation procedures of Thornthwaite Moisture Index and its impact on footing design. Australian Geomechanics, Vol. 51(1), p 85-95. Leao, S. and Osman-Schlegel, N.Y. (2013) TMI for urban resilience: measuring and mapping long-term climate change on soil moisture”, 7th Australasian Housing Researcher’s Conference, 6-8 February 2013, Fremantle, WA, Australia. Li, J. and Cameron, D.A. (2002) A case study of a courtyard house damaged by expansive soils. Journal of Performance of Constructed Facilities. ASCE, Vol. 16 (6), p 169-175. Li, J. Cameron, D.A. and Ren, G. (2014) Case study and back analysis of a residential building damaged by expansive soils. Computers and Geotechnics, Vol. 56, p 89-99. Li, J. and Sun, X. (2015) Evaluation of changes of Thornthwaite Moisture Index in Victoria. Australian Geomechanics, Vol.50 (3), p 39-49. Li, J., Zou, J., Bayetto, P., and Barker, N. (2016) Shrink-swell index database for Melbourne, Australian Geomechanics, Vol. 51(3), p 61-76 Lopes, D. and Osman, N.Y. (2010) Changes of Thornthwaite’s total moisture indices in Victoria from 1948-2007 and the effect on seasonal foundation movement, Australian Geomechanics, Vol. 45 (1), p 37-48. McKeen, R.G. and Johnson, L.D. (1990) Climate-controlled soil design parameters for mat foundations. Journal of Geotechnical Engineering, Vol. 116 (7), p 1073 -1094. McManus, K.J., Lopes, D. and Osman, N.Y. (2003) The effect of Thronthwaite Moisture Index changes on ground movements in expansive soils in Victoria, Australian. An International Conference on Problematic Soil, 29-30 July, 2003 - Nottingham, United Kingdom, CI-Premier, Singapore. McManus, K.J., Lopes, D. and Osman, N.Y. (2004) The effect of Thronthwaite Moisture Index changes on ground movement predictions in Australian soils. 9th Australia-New Zealand Conference on Geomechanics. Auckland. Mitchell, P.W. (2012) Footing design for tree effects considering climate change. Proceedings 2012 ANZ Geomechanics Conference, Melbourne. No. 1.3.11, p 290-295. Mitchell, P.W. (2013) Climate change effects on expansive soil movements. Proceedings of the 18th ICSMGE, Paris. Richards, B.G., Peter, P. and Emerson, W.W. (1983) The effects of vegetation on the swelling and shrinking of soils in Australia. Geotechnique, Vol. 33, p 127-139. Smith R. (1993) Estimating Soil Movements in New Areas. Seminar-Extending the code beyond residential slabs and footings, The Institution of Engineers, Australia. Sun, X, Li, J. and Zhou, A.N. (2017) Evaluation and comparison of methods for calculating Thornthwaite Moisture Index. Australian Geomechanics, Vol. 52. (2), p 61-75. Surfer® 11, (2012) Golden Software, LLC, Colorado: Golden. Available at: http://www.goldensoftware.com/. Thornthwaite, C.W. (1948) An approach toward a rational classification of climate. The Geographical Review, Vol. 38, No. 1, p 55-94. United Nations Framework Convention on Climate Change (UNFCCC). (1992) The United Nations Framework Convention on Climate Change. New York: UNFCCC


50
