Significant portions of land east of Great South Road in Takanini have been progressively rezoned for development since approximately 2004. With the introduction of the Auckland Unitary Plan, additional land has been identified for future urban development. This land is typically underlain by soft, compressible Holocene age, Tauranga Group peaty soils to varying depths with a typically high groundwater table. These characteristics present significant settlement risks to structures due to the application of fill and building loads, together with the possibility of groundwater drawdown induced settlement. The Regulatory Authority has required monitoring of groundwater levels and fill induced settlements prior to 224C to confirm attenuation of settlements to tolerable levels. Post 224C monitoring of selected dwellings, kerb points and manholes is also required for a period between 3 and 5 years. The post construction monitoring periods have been completed for a number of developments. This paper reviews and comments on the settlement performance using available data from completed developments constructed on the peaty Takanini soils in the last 15 years. It compares the expected settlement performance against the actual performance. Based on the available data, back analyses have been carried out to estimate primary and secondary settlement parameters. It also reviews the groundwater levels and the potential influence on dwelling settlement. Comments are provided on the implications of the recorded performance with respect to future developments on other land underlain by Takanini peat.
Since the adoption of the Takanini Structure Plan by the Papakura District Council (PDC, now Auckland Council) in November 2000, a number of developments have been completed on the peaty soils in the area. A substantial portion of the land bounded by the NIMT rail line in the west, Airfield Road in the north, Mill Road & Cosgrave Road in the east, and Old Wairoa Road in the south that was pasture 20 years ago, is now used for medium density housing, together with commercial, educational, and recreational type developments. With the adoption of the Auckland Unitary Plan – Operative in Part (AUP-OP) (Auckland Council, 2020), additional areas of land have been zoned as “Future Urban”. The Future Urban zone in Takanini is predominantly located between Mill Road in the east, Porchester Road in the west, extending north from Airfield Road to beyond Ranfurly Road.
The peaty Takanini soils can be described as organic clays and silts with varying quantities of fibrous peat. They are known to extend to depths of up to 20m. They encompass much of the Takanini area (Edbrooke, 2001). The peaty soils typically exhibit variable but low undrained shear strengths and high compressibility with a variable composition across the area. Groundwater levels are usually within 1-1.5m depth below the ground surface during the summer rising to near the surface during winter (Beaumont, et al. 2010). Groundwater levels are generally higher in the east and lower in the west. Horizons of pumiceous materials are also present at varying depths within the peaty soils, with varying degrees of continuity. Their thicknesses have been identified as ranging from 0.5m up to approximately 4m.
Core Geotechnical Constraints
The main geotechnical issues associated with the peaty soils of the Takanini area are the low undrained shear strength and high settlement potential under fill and building induced loads. They are also susceptible to widespread settlement from drawdown of the groundwater table beyond the historic summer low and groundwater recharge is required.
Historic Performance Standards
Developments have been successfully completed in the Takanini area for many years prior to the introduction of the AUP-OP. Acceptance criteria were developed by PDC with respect to maintaining groundwater levels through recharge, fill and dwelling induced settlements, and long term monitoring requirements (e.g. 3-5 years).
PDC required developments to recharge the first 15mm from each rainfall event. This has been adopted in the AUP-OP (rule I4220.127.116.11) as have the long term monitoring requirements.
Settlement criteria developed by PDC required fill induced settlements to attenuate to 3mm/month for 3 consecutive months and for the rate of settlement to have attenuated to less than 20% of the initial settlement rate. Engineers were then required to assess alert and alarm levels for dwelling, manhole and kerb settlements and changes in the groundwater level for long term monitoring.
Due to the high compressibility potential it is important to have a good understanding of the likely fill and dwelling induced settlements in order to provide appropriate recommendations to support successful development. The Addison residential development located immediately to the east of Porchester Road was one of the first developments to commence since adoption of the Takanini Structure Plan by PDC. The geotechnical engineers for the development carried out a preload trial. From the preload trial they estimated fill induced settlements at 10-15mm per 100mm of fill placed, and dwelling induced settlements at 30-50mm for lightweight timber framed dwellings supported on stiffened pod raft type foundations applying a uniformly distributed load of 10kPa to the peaty soils. These values have been widely adopted for developments across the area.
Since the introduction of the AUP-OP the Auckland Council is now requiring deep investigations and site specific settlement assessments to support applications for consent for developments in place of reliance on the settlement rates and values assessed by PDC, and derived from the now historic preload trial. Council is also no longer prescribing that fill settlements must attenuate to less than 3mm/month for 3 consecutive months and be less than 20% of the initial settlement rate. Engineers are required to provide the acceptance criteria.
It is against this backdrop of the reconsidered approach of Auckland Council to assessment of settlement attenuation requirements that this paper reviews the results of settlement and groundwater monitoring results for a number of developments across the Takanini area completed in the last 10 years. This paper is intended to compare the expected settlement performance against the actual performance. Based on available data, back analyses have been carried out to estimate primary and secondary settlement parameters. It also reviews the groundwater levels and the potential influence on dwelling settlement. Comments are provided on the implications of the recorded performance with respect to future developments on other land underlain by Takanini peat.
DATA REVIEW & ANALYSIS
Available Data & Analyses
Data for sites in the Takanini area from the RILEY archives has been reviewed as well as information provided by Auckland Council. In total data from 9 sites spread across the Takanini area and underlain by peaty soils has been reviewed and analysed. A summary of the reviewed data points is presented in Table 1 below.
Table 1: Summary of Analysed Points and Monitoring Data Reviewed
|Sites||Settlement Back Analysis Points||Fill Settlement Monitoring Points||Long Term Dwelling Settlement Points||Piezometer Monitoring Points|
Analyses have been carried out using Settle software by Rocscience to assess corresponding soil compressibility characteristics. This has been done for sites where a full electronic data record of settlements induced by fill placement was available, together with information on fill depths, monitoring plinth construction and sufficient investigation data to assess the thickness of the peaty soils. The corresponding compressibility parameters were used to estimate dwelling settlements at 50 years after construction and compared the calculated values with those estimated from the historic preload trial. The dwelling settlement magnitudes recorded at the end of the Council mandated 3-5 year monitoring period are also compared to the estimate for dwelling settlements for the corresponding timeframe.
A comparison is made between recorded fill induced settlements from monitoring and magnitudes estimated using the rate from the preload trial.
Groundwater levels have been reviewed for visible trends and consideration is given to whether these have had any obvious effect on dwelling settlements.
Back analyses were carried out for a single point within Site 1 beneath a 1m high preload. The remaining 18 points were back analysed for settlement plates installed beneath peaty fill soils placed as part of site development works across Sites 2-4. In all cases, the settlement plates were installed and baseline readings were undertaken prior to fill placement. Survey monitoring then proceeded generally on a regular basis at a +/-2mm accuracy. The following methodology was utilised for the back analysis:
- Plot the data against time and inspect for indications of multiple phases of fill placement, review site observations etc;
- Estimate settlement and time for T90 using the Asaoka method. If T90 isn’t clear review settlement vs time data in a log scale;
- Create the model in Settle using the geotechnical investigation data, fill depths and placement times. A unit weight of 15kN/m3 was used for the insitu peaty soils and for the peaty fill. For Site 1 a unit weight of 20kN/m3 was used for the hardfill preload. Insitu peaty soil thicknesses ranged from 8m to 18m for the back analysed sites. Groundwater was set at 1m depth;
- Add the dwelling load (12m x 12m footprint with a rigid 10kPa load) either at 1 year after first fill placement or at the end of the fill monitoring data, whichever occurs last;
- Adjust the cv parameter through iteration until the calculated timeframe for T90 occurs approximately coincident with the timeframe estimated from the monitoring results. Two-way drainage is assumed in order to account for the presence of preferential drainage paths via sandy/pumiceous horizons and fibrous materials;
- Adjust the mv value through iteration until the calculated settlement magnitude at T90 is similar to the recorded value;
- Where the data record is of sufficient length for secondary settlement to be observed, carryout iterative analyses to assess Cα. Secondary settlement should be visible on the Asoaka plots as well as in the settlement vs time plots;
- Check the R2 value for the degree of fit with the recorded data;
- Check whether the back analysed parameters are realistic.
Using the above methodology, soil compressibility parameters were able to be estimated together with calculated settlements beneath typical dwellings 50yrs after their construction. The results are presented in Table 2 below together with comparisons of recorded fill and dwelling settlements against the historically estimated values.
Table 2: Summary of Back Analysed Soil Compressibility Parameters, Fill & Dwelling Settlements
|Cα||R2||Calculated Dwelling Settlement as % of 50mm||Maximum Recorded Fill Settlement as % of 15mm/100mm Of Fill (mm)||Maximum Recorded Dwelling Settlements as % of 50mm|
* Site 2 has a single mv value of 1m2/MN, Site 3 has three mv values above 1m2/MN. When these values are excluded, the average and standard deviation of mv reduce to 0.38m2/MN and 0.14m2/MN respectively across a range from 0.23m2/MN to 0.65m2/MN.
Discussion on Analysis & Comparison With Settlement Monitoring Results
In the back analysis results (Table 2), the average and standard deviation of mv is affected by four values of 1m2/MN or greater, which appear atypical. However, they are indicative of the variability of the peaty soils. When the atypical values are excluded, the calculated volume compressibility values are within a range of 0.23-0.65m2/MN. This implies the volume compressibility of the peaty soils here are generally confined to a reasonably narrow range with areas of higher compressibility interspersed.
The cv values range from 40 to 200m2/year. Back analysis returned similar ranges of cv for each of the sites indicating relatively high variability of the peaty soil composition and presence of materials that would provide preferential drainage. This is further evidence of variability of the materials within individual development sites and across the area.
Coefficients of secondary consolidation were found to be consistent with only three values above 0.001 within a single site (Site 3). Secondary consolidation makes up approximately 25-65% of total calculated settlements over the dwelling 50-year design life.
Consistency of the calculated cv ranges on a site by site basis could provide confidence in the timeframes for settlements to occur but the back analysed mv values do not provide the same confidence. With 4 out of the 19 points having mv values of 1m2/MN or greater, there is a risk that settlement magnitudes for some future dwellings may be significantly under predicted if areas of higher compressibility are not considered.
Fill induced settlements were typically less than would be expected using the rate derived from the historic preload trial. Only 5 out of 36 points exceeded 100% of the estimated value and only two exceeded it by more than 20%. These results show that the preload derived rates provide a reasonably reliable estimate of the upper bound of expected fill induced settlement. This could be due the fact that the settlement rates were derived from the results of a hardfill preload whereas the fills considered comprise peaty soils. After adjusting the settlement rate for the difference in material unit weights, two further points exceed the estimate. The success of this method is likely due to the degree of inbuilt conservatism given the material unit weight differences combined with the length of time that the preload trial was in place (approximately 2 years) whereas most fill settlement monitoring results cover a 1-2 year period.
Fill settlement rates sufficiently attenuated at all points to comply with the historic Council requirements. In all cases compliance with the criteria was achieved at some time after achieving T90.
The majority of calculated future dwelling settlements (11 out of 18) were greater than 50mm after 50 years and 3 exceeded 100mm. This is in stark contrast to the maximum recorded dwelling settlement of 37mm (from 24 dwellings) at the end of the long-term monitoring period. Coincidentally, this is generally consistent with the calculated settlements for the remaining 7 back analysed points for the same timeframe. Further interrogation of the back analysis results reveals calculated settlements at 2 years and 5 years after dwelling load application exceed 50mm at 6 of the 18 points. These results indicate that even after considering the settlement timeframe over which the preload derived value was developed, the nominated 50mm settlement value is not entirely reliable and should be viewed with caution. However, despite the significant proportion of back analysed points that exceeded 50mm at 50 years the effect of this additional settlement is considered unlikely to be significant for properly designed dwellings. This is likely due to the use of stiffened pod raft type foundation systems that create an essentially rigid slab in relation to the soft peat (Econcrete=25GPa, Epeat=1-4MPa typically), resulting in negligible differential settlement, combined with the floorslab being set well above minimum floor levels for flooding (with settlement allowances), and the use of flexible service connections that can accommodate a level of relative vertical movement. Problems could arise if settlements resulted in floor slabs settling below the design flood levels, where relative movements between the dwellings and private services exceeds their tolerances or results in a loss of grade. These issues can be readily addressed through setting appropriate floor levels to accommodate calculated settlements, increasing the tolerance of private services to vertical dwelling movement and increasing their gradient.
It would therefore follow that if the 50mm estimate is to be used to estimate future dwelling settlements, it should be used circumspectly, together with provision for the possibility of settlements significantly exceeding this value.
From the groundwater level monitoring data available we were able to generate graphs of the groundwater level with time for five sites. The results are plotted in Figure 1 below. For clarity we have plotted the average recorded groundwater levels for each site.
Figure 1: Average Recorded Groundwater Levels
Discussion on Groundwater Monitoring Results
Regular seasonal fluctuations in response to rainfall events are visible in Figure 1. Most recorded groundwater level readings are between 0.5m and 1.5m depth.
Sites 1, 3 and 4 show a general cyclical trend of increasing groundwater depths. Further review of the individual piezometer data above indicates that Sites 1 and 4 appear to have been affected by abnormally low rainfall over the 2019/2020 summer. Site 3 may also be affected by reduced rainfall, and the presence of the Council stormwater channel located approximately 175m to the west that is currently under construction. Site 2 is located in the vicinity of Site 3 but has instead experienced relatively consistent seasonal groundwater fluctuations. This could be due to the presence of its own internal stormwater channel that is not yet connected to the main Council channel. The water impounded within the internal channel would be acting to reduce seasonal groundwater fluctuations. Google Earth photography indicates that there was still significant water within the Site 2 channel whereas the water levels within the Council channel under construction are much lower and have likely been suppressed to enable construction.
Unfortunately post 224C (issue of lot titles) settlement monitoring data is not available for Site 3 to comment on the possible effects of the lower groundwater levels.
At the commencement of work on Site 4, there was no stormwater management infrastructure and the site was very wet. Site earthworks, installation of stormwater service lines, road under channel drains, and individual lot recharge pits are likely to be acting to limit the upper bound groundwater level, while also providing recharge when there is rainfall. It is considered that the deeper 2019/2020 summer groundwater levels are likely to be the result of the combined effect of the works and very low rainfall. This would normally be of concern but in this case there has been no obvious response in the settlement data to these lower groundwater levels. This implies that these levels have occurred previously. Regular, stabilised seasonal fluctuations would be expected to resume after completion of construction and installation of all the individual lot recharge pits.
Graphical data provided to us for Sites 6 to 9 shows relatively consistent seasonal cycles in the groundwater levels with the highs and lows from year to year being approximately the same, indicating a flat groundwater level trend.
The back analysis results show that while the peat volume compressibility characteristics are generally confined within a relatively defined range, they also show that there are areas within individual sites and across the Takanini area that exhibit significantly higher compressibility characteristics. For this reason, the calculated volume compressibility values should be used with caution and the engineer should consider the potential presence of higher compressibility zones within individual developments.
The rate of consolidation range derived from back analysis could be used to estimate the likely upper and lower bound timeframes for settlement attenuation noting that achievement of T90 does not necessarily mean that settlement rates have attenuated to less than 3mm/month.
Secondary consolidation is a significant contributor (25-65%) to overall settlements although the quantum is less than anticipated with almost all back analysed Cα values being between 0.001 and 0.002. Allowance should be made for variations in Cα.
The historic preload derived settlement values have been shown to be a relatively reliable method for estimating fill induced settlements up to the end of site development works. However, they have also been shown to potentially under predict settlements for a significant proportion of the dwellings over a 50-year design life. If it is desired to use an arbitrary value to estimate long term dwelling settlement, then based on the back analysis, a value of 100mm would significantly increase the reliability of the approach. This would still need to be combined with measures such as adjustment of constructed floor levels, and increasing the tolerance in the private service connections to dwelling settlement and the gradient of the services to ensure that grades do not fall back towards the dwellings. Nevertheless, the geotechnical engineer should still carry out site specific assessments to estimate settlements over the dwelling design life to confirm the applicability of the arbitrary approach for their specific site and ensure that appropriate recommendations are made.
The requirement for fill settlements to attenuate to less than 3mm/month and be less than 20% of the initial settlement rate prior to dwelling construction has been found to ensure that T90 is reached and that differential settlements across site fills are suitably low. The continued use of this criteria needs to be weighed against the acceptable calculated future dwelling settlement. If lower long term dwelling settlements are required then fill settlements would necessarily need to attenuate to a lower value or other alternative settlement inducing measures adopted (e.g. preload).
Groundwater levels are observed to follow a seasonal cyclical trend across the Takanini area. The results presented for some sites appear to be affected by reduced rainfall, site development works and the the Council stormwater channel under construction. Monitoring is continuing on a number of sites and equilibrium levels are expected to be re-established for these sites once construction is complete. From the available data there is no obvious increase in settlement related to groundwater level effects, indicating that the recorded groundwater levels and the resulting induced overburden pressures have occurred previously. Groundwater recharge is an important measure to ensure that widespread settlement doesn’t occur and it appears that it has resulted in groundwater levels being maintained across the Takanini Area with a few aberrations as mentioned above.
Auckland Council. 2020. Auckland Unitary Plan Operative In Part. May 2020. https://unitaryplan.aucklandcouncil.govt.nz/pages/plan/Book.aspx?exhibit=AucklandUnitaryPlan_Print
Beaumont, J.L., Giles, E., Russell, D., McCarrison, G. 2010. Safe Urbanisation Across The Takanini Peat Flats – Medium To High Density Housing In A Geologically Challenged Area. Proceedings of the 11th IAEG Congress, Auckland, New Zealand. CRC Press.
Edbrooke, S.W. (compiler) 2001. Geology of the Auckland Area. Institute of Geological & Nuclear Sciences Limited. 1:25000 geology map 3. 1 sheet + 74p. Lower Hutt, New Zealand.