Introduction

Reactive soils undergo significant volume changes following seasonal changes of soil moisture content; shrinking in the dry season due to loss of moisture and swelling in the wet season by absorbing moisture1. This behaviour leads to severe damage to infrastructures, particularly lightweight structures such as residential buildings2,3. Such damages and resulting economic and environmental impacts have been reported in many countries worldwide, including Australia, the United States of America (USA), the United Kingdom, China, South Africa, India, and Spain4,5. In the USA, for instance, the potential harm caused by reactive soils was found to surpass the average yearly damage caused by natural disasters such as floods, hurricanes, earthquakes, and tornadoes2.

Approximately 20% of Australia’s land is covered with reactive soils (Fig. 1a). Many instances of damage to infrastructures, such as residential buildings, roads, and small-scale commercial structures, due to the reactive nature of the underlying soils have been reported in these areas1. The spread of reactive soils covers the major cities in Australia, including Melbourne, Sydney, Brisbane, Adelaide, and Darwin, which are also recognised among Australia’s top ten growth areas6,7 (Fig. 1b). A study by the Housing Industry Association (HIA) in 2011 found that around 1,000 new houses were damaged due to the extensive ground movement in the western suburbs of Melbourne during the millennial drought between 1996 and 2009. The HIA attributed this issue to soil heaving following the heavy rains that broke the drought, even though the homes were built in compliance with Australian Standards and the Building Code8. In the reported incidents, particularly in the western suburbs of Melbourne, the majority of the residential slabs affected were constructed using waffle raft footing systems9.

Fig. 1
figure 1

Distribution of the (a) moderately to highly reactive soils in growth areas, Australia (area in black) (modified after10), and (b) population density in growth areas, Australia (modified after6).

Full size image

Reactive soils cause numerous problems for residential structures, and as a result, many previous studies have been devoted to researching construction practices and footing systems to mitigate these issues. Much of this research has focused on traditional footing types, such as stiffened rafts1,11,12,13. While these investigations have provided valuable insights into the performance of stiffened raft slabs on reactive soils, there is a clear and significant gap in the literature when it comes to waffle raft footing systems. Despite their widespread use in current residential construction, particularly due to their cost-effectiveness and efficiency, waffle raft systems remain underexplored in the context of reactive soil behaviour. A comprehensive understanding of soil-structure interaction in response to seasonal ground movements and temperature changes in reactive soils is essential for optimising waffle raft designs and minimising potential damage. Moreover, the seasonal variations of soil moisture and temperature beneath and around footing systems built on reactive soils are not adequately understood in the literature yet. In particular, detailed field-scale investigations using real foundation systems on this topic are not reported in the literature, according to the authors’ knowledge. Such a study can provide first-hand data on the seasonal changes in soil behaviour under footing systems to optimise their long-term performance. Therefore, this study investigates the seasonal variations in soil moisture content and temperature beneath a large-scale waffle raft foundation constructed and instrumented on reactive soils in Victoria, Australia, aiming to contribute to a better understanding of the performance of waffle raft foundations in reactive soil conditions. The following sections detail the various aspects of waffle rafts, as well as the variations in soil moisture and temperature and their relationship to reactive soils and foundation built on them.

Waffle raft foundation systems

The waffle raft system is a type of footing system that uses a grid of stiffened beams connected to the foundation slab, and the overall system resembles a waffle. The reinforced concrete internal beam network, which transfers the superstructure load evenly to the ground, is in direct contact with the ground, and the remaining void spaces between the slab and the ground are filled with a void former material. Figure 2 schematically illustrates a typical waffle raft section. Compared to other footing systems such as strip footings, pad footings and pier/pile footing systems, waffle rafts are commonly used for residential structures in Australia, mainly due to their simplified construction process14 and low cost15. In the current construction practice, expanded polystyrene (EPS) is the most frequently employed material for creating voids, as compared to other options such as polypropylene, polyethylene, and cardboard formworks15. The demand for EPS waffle pods in Australia is estimated to be approximately 7,000 tonnes annually16. EPS void formers have become integral to residential footing systems due to their advantages, such as cost and time savings during construction15,17. Additionally, EPS exhibits excellent insulation properties18.

The design details for waffle raft construction are stipulated in the Australian Standards for residential slabs and footings – AS287019. This standard serves as a reference for engineers, architects, and builders involved in the design and construction of residential buildings in Australia. In a standard waffle raft section, the maximum distance between internal beams is 1090 mm, with each internal beam having a width of 110 mm. The edge beams must be at least 300 mm wide for structures that use full masonry or masonry veneer systems. On the other hand, cladding frames and articulated masonry veneer systems require edge beams with a width of 110 mm. AS287019 specifies that the depth of the edge and internal beams can range from 260 mm to 610 mm, depending on the type of construction and site class. A damp-proofing membrane made of polyethylene film with a thickness of ~ 0.2 mm is used under the waffle raft that serves as a moisture barrier (Fig. 2).

Fig. 2
figure 2

Typical cross-section and detailing of a waffle raft foundation (modified after19).

Full size image

The specific design parameters and considerations outlined in AS287019 include site classification, site investigation, and concrete and reinforcement requirements. These design details aim to ensure the safe and reliable performance of waffle raft foundations in various soil conditions. The site classes describe the extent of potential ground movement resulting from moisture variations in reactive soils. The characteristic surface ground movement, \(\:{y}_{s}\), is determined according to Eq. (1), and it is used for site classification as per Table 1.

$$\:{y}_{s}=\:\frac{1}{100}\sum\:_{n=1}^{N}{\left({I}_{pt}\stackrel{-}{\varDelta\:u}h\right)}_{n}$$
(1)
$$\:{I}_{pt}=\alpha\:\times\:{I}_{ps}$$
(2)

Where, \(\:{I}_{pt}\:\left(in\:\%\:pF\right)\) is the instability index, \(\:\stackrel{-}{\varDelta\:u}\) (\(\:in\:pF)\) is the soil suction, \(\:h\) is the layer thickness, \(\:n\) is the number of soil layers within the design depth of the suction change, \(\:\alpha\:\) is the lateral restraint factor, and \(\:{I}_{ps}\) is the shrinkage index. The values of \(\:\alpha\:\) can be obtained as 1.0 for the cracked zone and 2.0 – z/5 for the uncracked zone, where \(\:z\:\left(in\:m\right)\) is the depth from the finished surface level (m) to the centroid of the area defined by the suction change profile and the thickness of the soil layer under consideration in the uncracked zone. The shrinkage index (\(\:{\text{I}}_{\text{p}\text{s}}\)) may be determined using one or more of the methods outlined in AS 2870 − 201119,20.

Table 1 Site classification based on characteristic surface movement (\(\:{y}_{s}\)) (modified after19).
Full size table

Soil parameters relevant to waffle rafts

The soil parameters that constantly vary and affect the foundation stability include soil moisture content and temperature. Understanding spatiotemporal variations of these parameters in the vicinity of foundations is essential to optimise the foundation design.

Soil moisture content

Soil moisture is a crucial component of soil composition with significant implications across engineering and agricultural applications21,22,23,24,25. The moisture content directly affects the behaviour of soils, such as cracking, shrinking, swelling, consistency, and soil density25. In engineering, soil moisture is integral for evaluating the performance of structures such as pavements, foundations, dams, walls, and clay liners24,26,27. The variation of soil moisture around and beneath foundations can occur due to numerous factors, encompassing both natural and man-made influences, such as weather changes, gardening activities, and pipe leaks28.

Extreme weather conditions underscore the importance of monitoring soil moisture, as rainfall can saturate the soil, leading to swelling and potential ground movement. Conversely, prolonged dry conditions can induce soil shrinkage, adversely impacting overall structural integrity. In addition, the characteristic surface movement, \(\:{y}_{s}\) (Eq. 1), which is used for site classification, is a parameter that accounts for the surface movement due to soil moisture content changes. One of the key properties that characterise the expansive behaviour of reactive soils due to moisture variations is soil suction. Suction is defined as the potential energy of soil water undergoing changes29. Total suction, which is influenced by soil moisture, is commonly described using the soil-water characteristic curve (SWCC)30. The SWCC represents the relationship between the soil suction \(\:\left(\psi\:\right)\), and soil water content. Suction can be either matric suction or total suction, depending on the conditions being considered. Soil water content, on the other hand, can be represented as gravimetric water content \(\:\left(w\right)\), volumetric water content \(\:\left(\theta\:\right)\), or degree of saturation \(\:\left({S}_{r}\right)\:\)31. The variation of suction with depth above the water table is known as the suction profile, which is considered in AS287019. The AS2870 standard assumes a linear decrease in suction with depth, from a maximum at the surface to zero at the active depth zone. AS2870 provides design parameters for suction changes, such as a change in suction at the soil surface \(\:\left(\varDelta\:u\right)\) value of 1.2 pF for all specified locations in Australia and a depth of design soil suction change ranging between 1.8 m and 2.3 m for Melbourne. Suction changes have significant implications for ground movement and foundation stability, particularly within the depth of design soil suction change defined by AS2870. Therefore, studying soil moisture variations is essential because they directly affect suction, which drives the shrink-swell behaviour of reactive soils. Since suction is a key parameter used to understand the volume change behaviour of reactive soils, it becomes essential to examine how suction responds to seasonal soil moisture variations. Despite some previous studies that have focused on moisture content variations in reactive soils (e.g11,28,32,33,34. , , spatiotemporal moisture content variations under a waffle raft footing system are not thoroughly understood.

Soil temperature

The temperature of the soil depends on the heat flow within the soil and the exchange of heat between the soil and the atmosphere35. Several factors contribute to the soil temperature, including soil moisture, vegetation, evaporation, slope, land surface, and organic matter content36. An increase in soil temperature can impact soil physical properties, altering soil structure through temperature-induced dehydration of 2:1 clay minerals and heat-induced cracks36. Additionally, higher temperatures can affect clay properties by increasing their silt content37,38,39.

Apart from soil moisture, fluctuations in soil temperature also play a crucial role in the performance of waffle raft footing systems. Extreme temperatures, especially in regions with significant climatic variations, can impact the thermal stability and insulating properties of the footing system. Seasonal temperature changes may also influence wall displacement in certain structures, such as U-framed structures, by causing changes in the earth pressure40. A reliable understanding of soil temperature variation beneath foundations is crucial to minimise excessive heating/cooling of the interior of structures and to understand energy losses associated with floor heating. In addition, the air conditioning requirements of buildings are dependent on the soil temperature variations41. Therefore, considering seasonal changes in soil temperature during the design phase is vital to minimise deflections and heat losses and optimise the thermal/energy performances of buildings.

Methodology

We constructed a model waffle raft foundation (referred to as foundation hereafter) on a highly reactive site in Victoria, Australia, to comprehensively understand the spatiotemporal variations of soil moisture content and temperature. The procedure followed four distinct phases (Fig. 3). The initial phase involved a thorough site investigation, during which the subsurface soil profile was examined, and the required soil samples were collected for subsequent laboratory testing. The results of this phase directly contributed to the site classification and design of the model waffle raft foundation according to AS2870 guidelines. In the second phase, specialised monitoring sensors were installed at strategic locations and depths to record the moisture content and temperature under and around the foundation. These sensors played a crucial role in providing real-time data essential to achieving the objectives of this study. In the third phase, the waffle raft was constructed adhering to the guidelines of the AS2870 standard, followed by the last phase of monitoring and data collection. The forthcoming subsections explain the phases mentioned above in detail.

Fig. 3
figure 3

The four main phases of the methodology of this study.

Full size image

Site investigation

The selected site is located in the Sunshine campus of Victoria University, which is in western Melbourne, Victoria. The geological map42, as shown in Fig. 4a, indicates that the site lies in an area of newer volcanic basaltic plains where the subsurface basaltic clay is known to be highly reactive. This was later confirmed through laboratory testing and analyses of the soil profile. In addition, the selection of the site within the premises of Victoria University assisted with the construction logistics and access to subsequent monitoring. Furthermore, as one of the fastest-growing regions in Australia43,44, the western suburbs of Melbourne, including the selected site in Sunshine North, are experiencing significant urban expansion within the existing urban and urban growth boundary (Fig. 4b). Hence, the results of the study will be of direct and immediate practical importance.

Fig. 4
figure 4

(a) Approximate site location in the geological map of Melbourne (modified after42) and (b) the growth corridor plan for the western suburbs of Melbourne (modified after44).

Full size image

Located approximately 14 km west of Melbourne, the Sunshine North region is situated in a temperate climatic zone of Melbourne (zone 3 according to19, and the depth of design soil suction change (Hs) can be adopted as 1.8–2.3 m19. The sub-surface soil profile at the site comprises a fill layer extending to approximately 0.3 m, characterised by silty gravel containing fine to medium brown-grey particles, followed by a residual clay layer between 0.3 m and 3 m, characterised by silty clay with high plasticity. Based on the borehole investigation and following the procedure in AS2870, the site can be classified as H2. The nearest tree (about 6 m high) to the edge of the foundation to be constructed was more than 25 m away, hence, its influence was deemed negligible.

The undisturbed core specimens obtained from two different depths – 600 mm and 800 mm – were used to conduct shrink-swell tests following the AS1289.7.1.1standard45. The results showed shrink-swell index values of 4.7% and 4.72% for 800 mm and 600 mm depths, respectively. The characteristic surface movement \(\:{(y}_{s})\) was then determined following Eq. (1) and found to be 68 mm, confirming the site classification as H2 (Table 1). Following the AS1289.6.7.346, the permeability of the soil was determined as \(\:3.58\times\:{10}^{-10}\)ms−1. At 0% moisture content, the soil exhibited a thermal conductivity of 0.51 W/mK. The mineral composition of the soil was also analysed using the X-ray diffraction (XRD) technique, and the results are presented in Table 2. A comprehensive characterisation of the soil condition at the site is crucial for further assessments and understanding of the performance of the model foundation.

Table 2 The mineral composition of the soil at the foundation site.
Full size table

Installation of monitoring instruments

The site was first prepared by removing the topsoil and levelling the ground. The moisture and temperature sensors were then installed beneath and outside the ~ 10 m × ~8 m model foundation to be constructed at different depths. Figure 5 schematically illustrates these strategic locations (i.e., just outside the foundation, under the edge of the foundation, under the middle of the foundation, and at different distances from the edge, along the diagonal and symmetric axes) and depths (0.3 m, 0.9 m, 1.5 m, and 2.3 m) designed to comprehensively understand the soil behaviour beneath and outside the foundation.

Fig. 5
figure 5

The layout of installed sensors (a) plan view, and (b) a three-dimensional view below the foundation (the diagrams were created using SketchUp software: https://app.sketchup.com/).

Full size image

The TEROS-12 sensors, as shown in Fig. 6, were utilised in this study, which is designed to independently measure volumetric soil moisture content, temperature, and electrical conductivity and consist of an epoxy body and stainless-steel needles47. The extended surface area of the needles minimises the substrate disturbance during insertion for higher accuracy and reliability of the measurements. These sensors provide a volumetric water content (VWC) accuracy of ± 1–2% when using medium-specific calibration, and a temperature measurement accuracy of ± 1 °C. In addition, its 70 MHz frequency helps to minimise the salinity and textural effects of the soil substrate, making it accurate for most soil types47. TEROS-12 sensors have also been used in other similar studies in the literature (e.g48,49,50. , .

Fig. 6
figure 6

Details of the sensors used in this study (a) the view of the TEROS-12 sensor, and (b) dimensions of the sensor measurement zone47.

Full size image

For the installation, boreholes were first dug to specified depths, and then the sensors were manually embedded into the ground at the desired locations. The sensor needles were first covered with the soil from their respective depth obtained during the borehole digging. The sensors were tested and confirmed for correct functioning before the boreholes were backfilled. All exposed wires and cables connecting the sensors to the data logger on the ground were secured using PVC casings or UV-resistant flexible conduits. This was necessary for their long-term optimum operation.

Construction of the foundation

After the installation of the sensors and prior to pouring the foundation, a crushed rock layer was placed over the foundation envelope. A 0.2 mm thick polyethylene film was then laid out on the crusher layer, which acted as a vapour barrier to the foundation (Fig. 7a). The ~ 10 m × 8 m foundation needed 48 EPS waffle pods of 1.09 m × 1.09 m × 0.3 m, and the edge beams were 385 mm deep. Some of the EPS waffle pods were cut to accommodate openings for sensor wires beneath the footing system as necessary. After positioning the EPS pods, N16 steel reinforcement bars were placed in the spaces for beams according to the structural design specifications (Fig. 2). These reinforcement bars were securely held in place using bar chairs or spacers to maintain their desired positioning during the concrete pouring. A layer of 8 mm steel mesh was then placed on bar chairs over the EPS pods as per the structural design (Fig. 2). The outer formwork arrangement, waffle pod layout, and reinforcement arrangement according to the design of the foundation are shown in Fig. 7b.

Fig. 7
figure 7

– Various stages of the foundation construction (a) site preparation with a crushed rock layer and a polyethylene film, (b) external formwork, EPS pods and reinforcement arrangement prior to concrete pouring, (c) concrete pouring, and (d) view of the constructed waffle foundation.

Full size image

Once the preparation was complete, grade N20 concrete (compressive strength = 20 MPa) was used to construct the foundation – an experienced external contractor was employed for this task, and the typical procedure of concreting was followed (Fig. 7c). The final view of the constructed foundation is shown in Fig. 7d. Finally, the foundation was cured for about a month, and data recording was started.

Monitoring of the foundation performance

A CR300 series data logger was used to retrieve the data from the moisture and temperature sensors. The data was then downloaded via PC400 software for subsequent analyses. The data recording frequency was set to 15 min to enable rigorous and granular analyses of data for understanding the reaction of soil to moisture and temperature changes in a range of time scales. The following section presents the results after analysing recorded data for over one year (June 2022 – August 2023), covering an entire dry and wet cycle over the four seasons.

Results and analysis

The spatiotemporal variations of volumetric soil moisture content and temperature were analysed alongside the average monthly rainfall and air temperature variations. To explore the seasonal behaviours, the Australian climate is considered across four main seasons51, categorised by grouping calendar months – Spring: September, October, November; Summer: December, January, February; Autumn: March, April, May; Winter: June, July, August.

A summary of average monthly rainfall and air temperature for the monitoring period from June 2022 to August 2023 is shown in Table 3. These data were obtained from the Bureau of Meteorology, Australia, at station 086038, located at Essendon Airport, approximately 6.1 km away from the site, which is the nearest bureau station for the project site. Monthly temperature data were derived from the average of the monthly mean maximum and minimum temperatures. The mean maximum temperature is the average daily maximum air temperature for each month, and the mean minimum temperature is the average daily minimum air temperature observed over all years of the record51. Monthly rainfall data was obtained by aggregating all available daily rainfall for the respective month.

Table 3 Summary of the average monthly rainfall and air temperature for June 2022 to August 2023.
Full size table

Soil temperature and soil moisture variations at 0.3 m depth

To analyse the soil temperature and soil moisture variations at 0.3 m depth, the sensors A & M – outside, B – top edge, G – middle, H – along the symmetric axis, I – bottom edge were considered (sensor C encountered a technical issue leading to no data from that). Figure 8 shows the soil temperature and soil moisture variations throughout the considered period, along with average monthly rainfall and air temperature.

Fig. 8
figure 8

Monthly variations of parameters at 0.3 m depth (a) soil temperature, air temperature and rainfall, (b) soil moisture and rainfall, and (c) soil moisture and air temperature.

Full size image

According to Fig. 8a, the soil temperature shows negligible spatial differences, where all the locations show nearly the same monthly temperatures. This could be attributed to the efficient heat distribution in the soil within the monthly time scale considered here. It can also be observed that the soil temperatures closely follow the trend of air temperature variation, which is a reflection of the prevailing seasonal weather conditions. For example, the soil temperature peaks in the driest months (in the summer) around January and February 2023 (Fig. 8a) and is lowest during winter months (i.e. June-August). The open nature of the foundation without the superstructure may have helped the soil temperature below the foundation follow the air temperature trend. It is also apparent that the heaviest rainfall month in October 2022 has not impacted the trend of soil temperature variation, suggesting a more profound effect of seasonal weather conditions on soil temperature than inconsistent rain events.

Soil moisture contents in Fig. 8b and c demonstrate spatial variations with distinct differences for different locations considered. In general, the moisture contents at all locations show sensitivity to rainfall, with higher rainfall resulting in higher soil moisture and vice versa. Figure 8b demonstrates that during the dry season (November 2022 – May 2023), the moisture content outside the foundation is lower compared to that under the foundation, which reverses during the wet season (July 2022 – October 2022). This is a reflection of a more efficient reaction of soil moisture content in the exposed ground to weather changes compared to the ground below the foundation. In addition, Fig. 8b and c depict that the fluctuations become less dramatic towards the middle of the foundation, as shown in Table 4. It should be noted that, within the first 1–2 months after construction, the moisture content shows random variations potentially influenced by the construction process and initial adjustments of the foundation. After the completion of the construction, the moisture variations become typical. In general, while the soil outside the foundation undergoes significant seasonal changes in soil moisture content, the soil moisture content below the foundation is relatively stable. However, the seasonal moisture content changes under the foundation are not negligible, as assumed in the standard design practice (this aspect is detailed in the discussion section later).

Table 4 Maximum and minimum moisture contents at 0.3 m depth.
Full size table

Soil temperature and soil moisture variations at different depths

As illustrated in Fig. 5, two locations were selected to analyse the soil temperature and soil moisture variations at different depths, one under the middle of the foundation and the other just outside the foundation. The depths considered for the analysis were 0.3 m, 0.9 m, 1.5 m, and 2.3 m.

Under the middle of the foundation

Figure 9 shows the soil temperature and soil moisture variations at different depths below the middle of the foundation throughout the considered duration, along with the average monthly rainfall and air temperature given in Table 3.

Fig. 9
figure 9

Monthly variations of parameters at different depths below the middle of the foundation (a) soil temperature, air temperature and rainfall, (b) soil moisture and rainfall, (c) soil moisture and air temperature.

Full size image

As Fig. 9a shows, the temperature fluctuation is the greatest closer to the surface and becomes less significant with the depth; according to Table 5, which shows the temperature range at each depth, at 0.3 m, the temperature varies in the range 12 –22 °C, while it is at 2.3 m is 15 °C − 19 °C. This is a reflection of the diminishing effect of weather changes above the surface with increasing depth.

Table 5 Soil temperature ranges for different depths under the middle of the foundation.
Full size table

It is also apparent from Fig. 9a that soil temperature variations at different depths are dictated by the seasonal weather conditions, where drier months in the summer/autumn (December 2022 – March 2023) recorded higher temperatures, while winter months (June 2022/2023 – August 2022/2023) generated lower soil temperatures. Figure 9a also infers that the deeper soils maintain a higher temperature during the wet season and a lower temperature during the dry season than the soil closer to the surface. The soil and air temperatures in Fig. 9a indicate that the air temperature variation closely resembles the soil temperature variation at 0.3 m depth, and the increasing depth leads to an increasing contrast.

There are two locations of interest in Fig. 9a – the maximum rainfall event around October 2022 and the second highest rainfall event around April 2023; the soil temperature variation curves cross each other in these locations, meaning that the soil temperature is nearly the same throughout the depth. Although the expectation after a heavy rainfall event can be a decrease in soil temperature, particularly at shallower depths, Fig. 9a displays that it continues to rise after the heaviest rainfall event in October 2022. As described before, this can be understood considering the air temperature variations in Fig. 9a, where, despite the rainfall, air temperature kept increasing from July 2022 beyond October 2022 (spring) until the peak in January 2023 (summer), which negated the effect of rainfall on soil temperature. Similarly, around April 2023, the air temperature is at a steep drop leading to the winter season, and the soil temperature starts to decrease dramatically around that time. The higher rainfall around April 2023 has possibly expedited the cooling, which resulted in the temperature decrease occurring at a higher rate than it increased in October 2022 (Fig. 9a). Therefore, it is clear that seasonal weather conditions (i.e. air temperature) play a superior role in soil temperature variations compared to sporadic rainfall events.

According to Fig. 9a, the seasonal soil temperature variation closer to the surface closely follows that of the air temperature, including the timing of the peak temperature. However, as the depth increases, a time lag can be observed for corresponding peak temperatures at different depths compared to the timing of peak air temperature and the peak temperature at 0.3 m. This indicates that the deeper soil is more immune to surface temperature changes, and the response to surface temperature changes is not immediate.

In contrast to the soil temperature, the soil moisture variations (Fig. 9b and c) against depth reveal a general declining trend, with surface soil moisture levels notably higher than other depths. However, interestingly, soil moisture variations at depths of 1.5 m and 2.3 m exhibit no noteworthy differences and are higher than the moisture levels at 0.9 m. This can be attributed to the soil below the middle of the foundation maintaining the difference in initial moisture content at different depths throughout the monitoring period. Overall, the trends in Fig. 9b and c lead to two important observations – (1) the soil moisture content changes are non-trivial below the middle of the foundation even at 2.3 m depth, and (2) the surface weather changes uniformly impact the soil moisture content changes at different depths below the middle of the foundation.

Outside the foundation

Figure 10 shows the soil temperature and soil moisture variations at different depths outside the foundation throughout the monitoring duration, along with average monthly rainfall and air temperatures obtained from Table 3.

Fig. 10
figure 10

Monthly variations of parameters at different depths outside the foundation (a) soil temperature, air temperature and rainfall (b) soil moisture and rainfall (c) soil moisture and air temperature.

Full size image

The soil temperature outside the foundation at different depths (Fig. 10a) exhibits a similar trend to the soil temperature variations observed under the middle of the foundation at different depths (Fig. 9a). Although the patterns are comparable, the actual corresponding temperatures are different for outside and under the middle of the foundation; Table 6 shows the range of temperatures observed at each depth considered for both cases. The comparable patterns of soil temperature are possibly influenced by the foundation being open, and the slightly higher temperatures observed outside the foundation are due to the direct exposure of the outside areas to the atmosphere. Table 6 also shows that the temperature range closer to the surface is plausibly higher for the exposed ground outside the foundation than that under the middle. As the depth increases, the effect of the foundation on the temperature variation becomes trivial, such that the range of temperature variation becomes comparable.

Table 6 Soil temperature ranges for different depths outside the foundation.
Full size table

Interestingly, Fig. 10a shows that, even for variations outside the foundation, during the period of the highest recorded rainfall around October 2022 and the second-highest rainfall around April 2023, soil temperatures at all these depths were comparable. In addition, the soil temperatures are observed to closely follow the trend of air temperature variation despite the influence of rainfall events. As described before, this observation reinforces the notion that air temperature has a more decisive impact on soil temperature dynamics, both within and outside the foundation, illustrating the close connection between meteorological conditions and soil thermal behaviour. Furthermore, when considering the soil moisture variations at different depths outside the foundation, it was found that during periods of heavy rainfall, the soil moisture closer to the surface is higher compared to other depths (Fig. 10b). Conversely, the lowest soil moisture content is recorded at a depth of 2.3 m. As rainfall decreases and air temperature increases, the soil moisture content at 2.3 m depth is higher compared to other depths (Fig. 10b and c). It is also worth noting that there is a time lag for soil moisture content response at deeper depths with respect to the timing of the rainfall events. For example, at 2.3 m depth, soil moisture content peaked in November 2022, after the highest rainfall month in October 2022.

Overall, although the seasonal temperature variations below and outside the foundation at different depths are comparable, the moisture content changes are notably inconsistent. The more efficient and fluctuating nature of temporal soil moisture changes outside the foundation is understandable, considering the direct interaction of the soil with the above-surface weather changes. On the other hand, soil moisture changes below the foundation are comparatively less affected by the above-surface weather changes but are not entirely irresponsive.

Discussion

This study revealed important insights into the seasonal variations of the behaviour of reactive soils below and around a foundation based on first-hand data. These weather-driven dynamics are critical considerations to optimise the design of lightweight structures on reactive soils. While the following discussion on the application of the results of the present study for the design of residential foundations mainly focuses on the relevant Australian Standard19, the key findings are also applicable to other design procedures.

Influence on the mound shapes and edge distance

Both soil moisture content and temperature showed spatial fluctuations. Despite the fluctuations becoming less significant towards the middle of the foundation, both temperature and moisture content displayed non-trivial seasonal fluctuations even below the middle of the foundation. In other words, the soil below the centre of the foundation still actively contributes to ground/foundation movements as dictated by the seasonal soil moisture content variations. This observation has direct implications for the design procedure of residential footing systems. For soil-structure interaction analysis, both AS287019 and Walsh method52 idealised mound shapes of foundations in response to ‘centre heave’ and ‘edge heave’ scenarios for the dry season and wet season, respectively, as shown in Fig. 11.

Fig. 11
figure 11

Idealised mound shapes (a) centre heaving profile in the dry season, and (b) edge heaving profile in the wet season (modified after19).

Full size image

As Fig. 11 illustrates, the edge distance, \(\:e\), defines the length in which movement occurs and beyond that, the foundation stays flat. The equations for edge distances for centre and edge heaves are expressed in Eqs. (3) and (4), respectively.

$$\:e=\frac{{H}_{s}}{8}+\frac{{y}_{m}}{36}$$
(3)
$$\:e=0.2L\le\:0.6+\frac{{y}_{m}}{25}$$
(4)

Where, \(\:L\) is the foundation length (in m), \(\:{H}_{s}\) is the depth of the design soil suction change (in m), \(\:{y}_{m}\) is the differential mound movement (in mm), and \(\:e\) is the edge distance (in mm). The differential ground movement, \(\:{y}_{m}\), is dependent on the moisture conditions, the footing system, and the edge paths. In cases where determining \(\:{y}_{m}\) is challenging, AS287019 recommends using 0.7\(\:{y}_{s}\), and 0.5\(\:{y}_{s}\) for centre heaving and edge heaving, respectively, based on the Walsh method52,53,54. These estimates are of an empirical nature, and the authors are not aware of reliable data that substantiate them.

For the foundation considered in this study, with calculated characteristic surface movement \(\:{(y}_{s})\) of 68 mm, active depth zone \(\:{(H}_{s})\) of 2.3 m and a foundation length of \(\:\left(L\right)\) 9.7 m, using the Walsh method (Eqs. 3 and 4), the edge distances (\(\:e\)) for centre and edge heaving are 1.61 m and 1.96 m, respectively. However, the results of this study suggest that the seasonal moisture content fluctuations extend up to the middle of the foundation (i.e., 3.65 m), which may lead to considerable ground movement even below the middle of the foundation. Therefore, revisiting the standard design practice of residential footing systems is strongly recommended, and further research is needed to improve the overall design practice therein.

Influence on the depth of design soil suction change

Understanding the depth over which moisture content is expected to vary is crucial for foundation design. This is because, to estimate ground heave beneath residential buildings, the designer must calculate the heave by integrating the soil strain across the depth range where moisture changes are anticipated55,56. The depth of the design soil suction change, \(\:{H}_{s}\) represents the depth to which seasonal effects will cause the wetting and drying of the reactive clay layers57. The soil moisture content variations with depth show considerable fluctuations even at 2.3 m depth, which is the highest of the range of \(\:{H}_{s}\) for Melbourne, according to AS2870. This was more pronounced in the natural ground outside the slab. The \(\:{H}_{s}\) depth is directly linked to the estimation of characteristic surface movement \(\:{(y}_{s})\), as it defines the vertical extent over which seasonal suction changes are considered in residential footing design. While AS2870 provides standard values for \(\:{H}_{s}\), the site-specific conditions such as moisture variation due to vegetation, drainage, the presence of bedrock or groundwater, may result in suction profiles that extend beyond these empirical limits. If the actual depth of moisture variation exceeds the standard \(\:{H}_{s}\) range (typically 1.8–2.3 m for Melbourne), there is a risk of underestimating \(\:{y}_{s}\), which in turn may lead to incorrect site classification. Inaccurate estimation of \(\:{y}_{s}\) can compromise the selection of appropriate footing systems and design parameters, particularly for on-ground slab systems such as waffle raft slabs, which are more susceptible to differential movement. Therefore, it is essential to assess the actual depth of soil suction variation at a given site to verify whether the standard \(\:{H}_{s}\) values from AS2870 adequately capture the depth of design soil suction change. Over- or underestimating this depth may lead to significant errors in site classification and, ultimately, an inadequate foundation design. Therefore, the non-trivial moisture content changes even at 2.3 m depth observed in this study suggest revisiting the specified \(\:{H}_{s}\) values in AS287019, given that these values are based on historical climate data and may no longer reflect current conditions due to the consequences of ongoing climatic changes. Fardipour, Gad58 also found that \(\:{H}_{s}\) values can vary at the same site when changing the average period used to calculate the Thornthwaite Moisture Index (TMI), which is an index used to classify the climate conditions of different areas.

It should be noted that although a large-scale foundation was used in this study, only the foundation was constructed without the superstructure. The exposed foundation offered better interaction of the ground underneath with the above-surface weather changes, which is absent in the case of real residential structures. However, this can be viewed as the worst-case scenario or a short-term situation representing the duration after the foundation is constructed and before the superstructure is completed.

Conclusions

The following conclusions are made after monitoring and analysing the soil moisture content and temperature data below and outside of a waffle raft foundation.

  1. 1.

    (1) Soil temperature under the foundation at 0.3 m depth at different locations across the foundation varied following seasonal weather conditions but did not show noteworthy spatial differences. Soil temperature closer to the surface, both outside and under the middle of the foundation, showed similar fluctuations to air temperature, and they became progressively less sensitive to surface weather changes with increasing depth. The seasonal soil temperature variation trends at different depths were identical for outside and under the middle of the foundation. Rainfall events show an inverse relationship with soil temperature, where values tend to decrease during periods of high rainfall and increase during periods of low rainfall.

  2. 2.

    (2) At shallow depths, the soil moisture content outside the foundation showed the greatest fluctuations, mainly driven by the rainfall patterns. The soil moisture content fluctuations become less dramatic towards the middle of the foundation, but a considerable fluctuation is still observed under the middle of the foundation. At different depths below the middle of the foundation, the soil moisture content showed seasonal fluctuations up to 2.3 m depth, but the initial moisture content differences at different depths were maintained. More dramatic fluctuations of soil moisture content with depth were observed outside the foundation up to 2.3 m. A revisit to the stipulated \(\:{H}_{s}\) values in AS2870 is recommended based on these results.

  3. 3.

    (3) The empirical estimates of the edge distance as per AS2870 and Walsh method resulted in less than 2 m, while notable seasonal soil moisture content variations were still observed under the centre of the foundation (i.e., 3.65 m from the edge). Therefore, active shrink-swell movements can still be expected beyond the edge distance, recommending a revisit to the standard design procedure of residential foundation systems in AS2870.

Overall, the results of this study set the evidence base for improvements in the standard residential foundation design procedures for improved performance and stability throughout their design life.