GIS-based soil bearing capacity zonation maps for the Dhaka metropolitan development plan (DMDP) area, Bangladesh

Abstract

This study develops soil bearing capacity zonation maps for the Dhaka Metropolitan Development Plan (DMDP) area in Bangladesh. A total of 360 borehole records from different locations within the region were compiled and analyzed. The mapping was carried out in ArcGIS 10.5, utilizing the Ordinary Kriging interpolation method. This research has developed 4 (four) shallow foundation bearing capacity zonation maps for the DMDP area at 1.5 m and 3.0 m depth and 2 (two) pile capacity zonation maps at 12.0 m and 20.0 m depth based on SPT-N₆₀ values.Precast pile size of 12″×12″ was used for the calculation of pile capacity at 12.0 m and 20.0 m depth, respectively. The accuracy of the zonation maps was evaluated by comparing estimated pile capacities at 12.0 m with pile load test results at seven locations. The analysis yielded an RMSE (Root Mean Square Error) of 5.76%, MAE (Mean Absolute Error) of 4.61, a coefficient of determination (R²) of 0.6, and a correlation coefficient of 0.77. The allowable bearing capacity in the study area is significantly influenced by soil type and shallow soil conditions. The bearing capacity zonation maps will assist geotechnical designers in the initial selection of the site and depiction of encountered strata. The map conserves time and resources by providing a fundamental overview of the preliminary foundation design, covering the estimated cost.

Introduction

The capacity of the subsoil to support foundation load without causing cracks or excessive settlement is referred to as the bearing capacity. The primary function of a foundation is to distribute the weight of the building uniformly over the ground. A correctly built foundation evenly distributes the load across the earth without overstressing it. The structure could be damaged if the earth experiences significant settlement or shear failure due to overloading. Therefore, when designing foundations, geotechnical and structural engineers need to think about the soil’s load-carrying capacity.

The Standard Penetration Test (SPT) is one of the most frequent approaches for estimating in-situ geotechnical parameters¹. This approach is popular because SPT data is readily obtainable from soil borings. Typically, SPT is carried out either at the stratum transition or every 1.5 or 2.0 m of depth¹.

The geographic information system (GIS) is a highly effective technology for acquiring, displaying, analyzing, and visualizing spatial data. The acquisition of geotechnical information is both laborious and expensive, especially where extensive borings are required at numerous locations and the information is not broadly disseminated among engineers, consultants, contractors, and researchers². Existing geotechnical data can be analyzed to generate additional details through the integration of GIS extension technologies. The outcome of the GIS methodology is a comprehensive understanding of spatial relationships and patterns in the data. In geotechnical applications, GIS facilitates four main functions: integrating data, visualizing and analyzing data, planning and summarizing site tasks, and presenting data³.

There are many related studies that have developed soil bearing capacity zonation maps for urban areas around the world. Ahmad et al.⁴ developed geotechnical zonation maps (GZMs) of USCS soil type and bearing capacity in Peshawar, Pakistan, by profiling and zoning geotechnical subsoil data using GIS. Khadim et al.⁵ created geotechnical maps indicating bearing capacity and groundwater quality for the Al-Imam area (Babil, Iraq). Al-Maliki et al.⁶ created a bearing capacity map for An-Najaf and Kufa employing ArcGIS 10.2.1. Aldefae et al.⁷ analyzed the settlement behavior and strength parameters of soil layers and developed a digital soil database in the Wasit governorate of Baghdad using ArcGIS software. Celik et al.⁸ developed geotechnical and geological maps using ArcGIS software. They evaluated two interpolation methods (minimum curvature and Voronoi polygons methods) to develop SPT-N, soil classification, groundwater table, plasticity index, and bearing capacity micro zonation map in Nidge, Turkey. Khalid et al.⁹ employed the Kriging method to generate contour maps illustrating SPT-N and USCS soil classifications for Islamabad, Pakistan. Ali & Shakir¹⁰ prepared geotechnical maps of Thi Qar governorate in Iraq using GIS. Ijaz et al.¹¹ prepared geotechnical maps (GMs) utilizing SPT-N, shear wave velocity, plasticity index, activity, linear shrinkage, cohesion, frictional angle, and chemical contents. An optimized inverse distance weighting (IDW) algorithm, derived from the modified Shepard method was utilized to create geotechnical spatial models (GSMs) at various depths for Rawalpindi, Pakistan. Daniyal et al.¹² employed the IDW technique to create soil type, groundwater table, SPT-N values, shear wave velocity, and load-carrying capacity maps for Lahore, Pakistan. Civelekler¹³ developed allowable bearing capacity zonation maps for Eskisehir city at multiple shallow depths by employing the IDW interpolation technique. Ijaz et al.¹⁴ assesses multiple geostatistical and spatial interpolation methodologies utilizing extensive geotechnical data from Lahore. They employed the Inverse Distance Weighting (IDW) technique to create Geotechnical Spatial Models (GSMs) at various depths, utilizing SPT-N, shear wave velocity, soil classification, plasticity index, soil activity, and chemical composition. Rahman & Ansary¹⁵ used both IDW and Ordinary Kriging interpolation in ArcGIS 10.5 to prepare the SPT-N₆₀ and USCS soil classification maps for the DMDP region in Bangladesh. Amini et al.¹⁶ produced soil zonation maps (SZMs) of SPT-N and USCS soil type, adopting Kriging interpolation in GIS for New Kabul City, Afghanistan. For each zone, they assessed the bearing capacity and settlement utilizing Kriging interpolation technique, SPT-N values, and soil classifications. Table 1 highlights the features examined in numerous studies in the development of soil-based zonation maps.

Table 1.

Data utilized in previous studies for preparing soil zonation maps.

Data used	No. of samples collected	Site	Source
Soil type & bearing Capacity	103	Peshawar, Pakistan	Ahmad et al.4
Bearing capacity, undrained shear strength, liquidity index, and compression index	105	Basrah, Iraq	Khadim et al. (2013)
SPT-N & bearing capacity	464	An-Najaf, Iraq	Al-Maliki et al.6
Bearing capacity, consolidation coefficient, ground water levels, shear strength	464	Wasit governorate, Baghdad, Iraq	Aldefae et al.7
SPT-N, groundwater table, plasticity index, bearing capacity, liquefaction	100	Nidge, Turkey	Celik et al.8
SPT N, groundwater levels, and soil type	50	Islamabad, Pakistan	Khalid et al.9
SPT-N, moisture content, dry density, ultimate bearing capacity	423	Thi Qar governorate	Ali and Shakir10
SPT-N, shear wave velocity, plasticity index, activity, linear shrinkage, cohesion, frictional angle, and chemical contents	406	Rawalpindi, Pakistan	Ijaz et al.11
SPT-N, bearing capacity, shear wave, ground water table	65	Lahore, Pakistan	Daniyal et al.12
SPT-N & bearing capacity	40	Eskisehir, Turkey	Civelekler13
SPT-N, shear wave velocity, soil type, plasticity index, soil activity, chemical composition	810	Lahore, Pakistan	Ijaz et al.11
SPT-N & soil type	360	DMDP area, Bangladesh	Rahman and Ansary15
SPT-N, undrained shear strength, soil type, consolidation parameters	213	New Kabul city, Afghanistan	Amini et al.16

GIS-based geotechnical mapping has been utilized globally; however, Bangladesh has limited published studies on this issue. This research creates the very first GIS-based bearing capacity zonation maps for shallow foundations at 1.5 m and 3.0 m depths and pre-cast pile capacity at 12.0 m and 20.0 m depths with the 12.0 m map validated against pile load tests. This represents a novel approach both locally and internationally, offering a practical framework for integrated geotechnical assessment throughout the DMDP area.

The study area

The DMDP covers an extensive area of approximately 1528 square kilometers, encompassing Dhaka city and its surrounding districts. This includes Dhaka, Narayanganj, Gazipur, and part of Narsingdi District. The DMDP is located within longitudes 90°13′15′′ and 90°36′15′′ east and latitudes 23°33′30′′ and 24°6′15′′ north. The DMDP plan was developed in 1995 with the technical aid of the United Nations Development Program (UNDP), the World Bank (WB), and the Asian Development Bank (ADB). It encompasses a broad geographical area and aims to guide the development of Dhaka and its surrounding regions in an organized manner. Figure 1 displays the study area map.

Fig. 1.

Study area (DMDP map). This map was generated by the 1^st author using ArcGIS 10.5 software (https://www.esri.com/software/arcgis).

Geology characteristics and setting

The DMDP area lies within the Bengal Basin, a vast geological region characterized by thick sediment deposits from the Ganges–Brahmaputra-Meghna river systems. The terrain is generally flat, with elevations varying from 1 to 24 m. The majority of city areas are situated 6–8 m above sea level¹⁷. Regarding seismic vulnerability, Dhaka faces a significant threat. The city lies between the Indian and Eurasian Plates, near a tectonically active plate boundary¹⁸. The city and its environs are covered by two geologic units: alluvial deposits and Madhupur Clay from the Pleistocene¹⁵. The primary landform components are the floodplains, abandoned canals, depressions, and the Dhaka terrace, or highland. Marshlands and wetlands are prevalent geological characteristics throughout the study area. Figure 2 shows the digital elevation model of the DMDP area.

Fig. 2.

DMDP area DEM. This map was generated by the 1^st author from Google Earth Pro elevation data using IDW interpolation in ArcGIS 10.5 software (https://www.google.com/earth/versions/#earth-pro & https://www.esri.com/software/arcgis).

According to the Bangladesh National Building Code (BNBC-2020) the DMDP area lies in a region of moderate seismicity, zone-2 with seismic zone coefficient (z = 0.20) among four seismic zones in Bangladesh. It is in close proximity to the active tectonic borders of the Indian and Burmese Plates. The geology, particularly the thick layers of unconsolidated alluvial deposits, makes Dhaka vulnerable to seismic shaking. Intensive extraction has led to a groundwater level reduction of 60 m over the last fifty years¹⁹. The current depletion rates between 0.6 and 2.4 m annually throughout Dhaka²⁰.

Methodology

Geotechnical data from 360 locations and 4680 data points have been obtained across the DMDP area under the project of the Bangladesh Urban Earthquake Resilience Project (BUERP). Figure 3 represents borehole coordinates (longitude and latitude) across the study area. Data was collected at 1.5 m depth intervals for each borehole, extending to a maximum depth of 19.5 m. The SPT (ASTM D 1586) process was used to characterize soil conditions and the collected soil samples were disturbed. All engineering properties were tested in the laboratory. The tests conducted included sieve analysis (ASTM D 422-63), liquid limit, plastic limit (ASTM D 4318-17), and hydrometer analysis (ASTM D 7928-17). Borehole location, SPT-N values, soil classification based on the USCS (ASTM D 2487-17) method, Atterberg limits and proportions of coarse and fine particles at different depths were acquired from geotechnical exploration reports.

Fig. 3.

Data collection points across DMDP area. This map was generated by the 1^st author using ArcGIS 10.5 software (https://www.esri.com/software/arcgis).

Zoning has been carried out based on bearing capacity. One of the most widely adopted approaches for estimating soil load-carrying capacity is based on correlations with SPT values. The SPT is typically regarded as more appropriate for granular soils. However, it continues to be one of the most widely applied in-situ tests for both cohesive and non-cohesive deposits, especially in areas where it is regarded as the benchmark subsurface exploration technique. Corrected SPT N-values have been effectively correlated with undrained shear strength and bearing capacity characteristics in clays^26,27. In this study, all SPT data were rectified and analyzed in accordance with established procedures. Meyerhof²⁸ introduced equations commonly applied for a foundation settlement of 1 inch²⁸. Bowles (1988) revised Meyerhof’s equations, assuming Meyerhof’s equation is conservative²⁶. According to Meyerhof²⁸ and Bowles (1988), suggested expressions for allowable bearing capacity based on SPT-N Values are listed in Eq. (1) and (2).

1

2

where is the allowable bearing pressure for 1- ch settlement (), = 1 + 0.33 1.33 (Meyerhof (1965), is the footing depth , B is footing width , and is the average SPT from 0.5B above to 2B below footing base.

In this research, the bearing capacity of shallow foundations at 1.5 m and 3.0 m was calculated using square foundation of size 6 ft. × 6 ft. (1.83m × 1.83m). The chosen footing size is a practical dimension frequently utilized for medium-load structures in the study area. Field survey data of existing buildings indicate that shallow footings with comparable dimensions and embedment depths are commonly employed for low- to mid-rise structures. Lightly loaded single-storey buildings typically utilize smaller footings (approximately 4 ft. × 4 ft.). The chosen size offers a representative evaluation of bearing capacity across diverse soil conditions while adhering to standard local practices.

The ultimate capacity of a pile comprises of skin friction (), and end bearing ()¹

3

4

where is the skin friction area, is the skin frictional resistance, is the end bearing area and is the ultimate point resistance¹.

The allowable axial capacity is to be considered as:

5

Meyerhof’s (1976) formula, presented in Eq. (6) was used to compute end bearing in cohesionless soil²⁹. Skin friction was determined using Eq. (7). A factor of safety () = 3.0 was considered while determining allowable bearing capacity. The specified value aligns with the minimum safety factor recommended by the BNBC 2020 for precast pile foundations, addressing uncertainties related to soil variability, installation effects, and pile-soil interaction.

Meyerhof (1976) proposed that the ultimate point resistance, in a homogenous granular soil can be derived from the SPT-N value based on field data.

6

Meyerhof (1976) stated that the average skin frictional resistance, for high-displacement driven piles can be derived from SPT values as

7

where is length of pile , is the diameter of pile , is avg. SPT value within 10D above to 4D below the pile tip, ₆₀ is the avg. value of the SPT and is the atmospheric pressure (≈ 2000lb/ft²)²⁷.

In cohesive soil, Hara et al. (1971) correlation between cohesion and SPT-, stated in Eq. (8), was used for the determination of cohesion ( ²⁷. Then skin friction was calculated using the formula given by Shioi & Fukui, stated in Eq. (9). () = 3.5 was considered for calculating allowable bearing capacity. BNBC 2020 recommends a higher factor of safety 3.5 to cohesive soils to address the inherent uncertainties associated with clay behavior, including variations in undrained shear strength, limitations in sampling and testing, and effects related to pile installation.

8

According to Shioi & Fukui³⁰ skin friction () for cohesive soil³⁰,

9

End bearing

End bearing capacity in cohesive soil is given by:

10

where is the bearing capacity factor, (≈9 for deep foundations) (dimensionless), is the undrained shear strength at pile base ), and is the pile tip area .

Two interpolation methods have been evaluated in the ArcMap 10.5 software. Spatial Analyst: IDW and Ordinary Kriging ^21-25. Each method produced a different dataset representation. The deterministic IDW method guesses unknown values by weighting the average of nearby measured points, giving closer samples more weight. Though simple and computationally efficient, IDW ignores spatial autocorrelation among data points.

Ordinary Kriging uses a semivariogram model shown in Fig. 4 to include distance and spatial structure for statistically optimal and unbiased predictions. Kriging produces an estimation variance map that quantifies prediction uncertainty by considering spatial dependence between sampled sites. The Ordinary Kriging prediction is based on³¹ Eq. (11).

11

where estimated value at location , μ the average or the local mean, and is random error term.

Fig. 4.

Semivariogram exhibiting nugget, sill, and range of spatial dependency.

The spatial dependence of the data was analyzed through the empirical semivariogram, defined as

12

where is the semi-variance corresponding to lag distance , number of data pairs separated by distance.

Nugget, sill, and range define micro-scale variability, overall variance, and spatial correlation distance in the semivariogram. Kriging predicts values at unsampled places and estimates prediction uncertainty, improving results interpretability. The ArcGIS Geostatistical Analyst module fitted the experimental semivariogram with spherical, exponential, and Gaussian models. Each model was assessed to determine the most appropriate depiction of the data’s spatial organization. The nugget-to-sill ratio measured spatial dependency.

Cross-validation statistics were used to compare the performance of both IDW and Kriging interpolation methods to construct spatially continuous surfaces of the soil bearing capacity zonation map in the DMDP area. 360 borehole points calculated bearing capacity (Meyerhof, 1956) at 1.5 m depth were compared with cross-validated predicted values at the same locations, where each borehole value was temporarily removed and then forecasted using the remaining data points. Kriging exhibited lower RMSE (1.289) relative to IDW (1.321), signifying enhanced accuracy and impartiality in predictions. The nugget-to-sill ratio further indicated high spatial autocorrelation in the geotechnical parameters, validating the use of a geostatistical technique. Kriging outperforms IDW in terms of RMSE and minimum bias, taking use of spatial correlation to provide more accurate predictions.

In this research, pile capacity at 12.0 m and 20.0 m was calculated. According to the comparison, the Ordinary Kriging interpolation technique was deemed to better represent the bearing capacity zonation maps in the DMDP area. Figure 5 illustrates the methodological framework of the study.

Fig. 5.

General illustration of the flowchart used in this study.

This study primarily concentrates on creating bearing capacity zonation maps for shallow foundations at depths of 1.5 m and 3.0 m, as well as for pile capacity at 12.0 m and 20.0 m. To better understand the soil qualities, the entire region has been split into 10 (ten) distinct zones. Savar, Gazipur Sadar, Dhaka North City Corporation (DNCC), and Dhaka South City Corporation (DSCC), Keraniganj, Narayanganj Sadar, Bandar, Sonargaon, Rupganj, and Kaliganj.

Results and discussions

Bearing capacity zonation maps for shallow foundation

This study assessed the bearing capacity at 1.5 m and 3.0 m depth using different theoretical approaches. The objective was accomplished by utilizing soil investigation data obtained from boring logs and SPT-N across various sites within the DMDP area. The zonation map was prepared using GIS software.

Bearing capacity was evaluated for a footing of size 6 ft. × 6 ft. (1.83m × 1.83m) utilizing the Kriging technique for data interpolation, which is appropriate for vast amount of input data.

Circular, Spherical, Exponential, and Gaussian semivariogram models were used to investigate bearing capacity data. Figure 6 shows these models, and Table 2 lists the statistical markers used to assess model accuracy. Using a leave-one-out procedure for cross-validation criteria, the Exponential semivariogram model produced the lowest RMSE and nugget-to-sill ratio. Spatial dependence is high when the nugget-to-sill ratio is below 0.25, moderate between 0.25 and 0.75, and mild above 0.75³². The Exponential model’s nugget-to-sill ratio is 0.725, indicating moderate spatial dependency.

Fig. 6.

Semivariogram models for bearing capacity (Meyerhof, 1956) at 1.5m depth: (a) Exponential, (b) Circular, (c) Spherical, and (d) Gaussian.

Table 2.

Indices used to cross-validate bearing capacity (Meyerhof, 1956) forecasting models at 1.5m depth.

Model	ME1	RMSE2	MSE3	RMSSE4	ASE5	Nugget/sill
Exponential	0.008	1.289	0.006	1.013	1.273	0.725
Circular	0.010	1.290	0.007	0.996	1.296	0.790
Spherical	0.009	1.289	0.007	0.999	1.291	0.789
Gaussian	0.010	1.291	0.008	0.991	1.303	0.821

¹Mean Error.

²Root Mean Square Error.

³Mean Standardized Error.

⁴Root Mean Square Standardized Error.

⁵Average Standard Error.

Exponential model performed better with the lowest Mean (0.008), Root Mean Square (1.289), Mean Standardized (0.006), Root Mean Square Standardized (1.013), and Average Standard Error (1.273). The results show that the Exponential model predicts most accurately and consistently. Accordingly, Ordinary Kriging with an Exponential semivariogram was selected.

The statistical indices and visual inspection of the interpolated zonation maps of the data indicate that Ordinary Kriging interpolation techniques more accurately represent the bearing capacity zonation maps in the DMDP area. Therefore, the bearing capacity zonation maps for the current study adopted the Ordinary Kriging method.

Bearing capacity at 1.5 m depth

Meyerhof (1956)

Figure 7 shows a variation in capacity at 1.5 m depth within the DMDP region. From Fig. 7, it is seen that the bearing capacity varies between 1.0 ksf and 4.75 ksf. The most prevalent bearing capacity at this depth is 2.0–3.0 ksf. In the south, southeast, east, and some central places, the bearing capacity is between 1.0 and 2.0 ksf. The bearing capacity varies between 3.0 and 4.0 ksf in the northwestern, northern, and a few northeastern regions. A few scattered locations in the north and easterly zones have the bearing capacity of 4.0 to 4.75 ksf.

Fig. 7.

Bearing capacity at 1.5 m (Meyerhof, 1956). This map was generated by the 1^st author using ArcGIS 10.5 software with the Ordinary Kriging interpolation method (https://www.esri.com/software/arcgis).

Following Meyerhof’s (1956) approach, around 20% area exhibits a bearing capacity of 1.0–2.0 ksf, 51% is within 2.0–3.0 ksf, approximately 26% of the area falls within 3.0–4.0 ksf, whereas about 3% of the area showed values in the range of 4.0–4.75 ksf.

Bowles (1988)

Figure 8 shows that the bearing capacity varies between 1.68 ksf and 7.0 ksf. The majority of the area exhibits load-carrying capacity between 3.0 and 4.0 ksf at this depth. In the south, southeast, east, and some central places, the bearing capacity is between 2.0 and 3.0 ksf. The load-carrying capacity ranging from 4.0–5.0 ksf is noted in the northern, northeastern, northwestern, and several central areas. The bearing capacity ranges from 5.0 to 6.0 ksf in the northwestern and northern regions. A few scattered locations in the north and northeasterly zones have the bearing capacity of 6.0–7.0 ksf.

Fig. 8.

Bearing capacity at 1.5 m depth (Bowles, 1988). This map was generated by the 1^st author using ArcGIS 10.5 software with the Ordinary Kriging interpolation method (https://www.esri.com/software/arcgis).

Following Bowles (1988) methodology, less than 1% of the area has a bearing capacity between 1.68 and 2.0 ksf, nearly 20% region falls within 2.0–3.0 ksf range, 35% of the area ranges from 3.0 to 4.0 ksf, 27% territory measures between 4.0 and 5.0 ksf, 5.0 to 6.0 ksf in 14% of the region, and only 3% of the area reaches 6.0–7.0 ksf.

Bearing capacity at 3.0m depth

Meyerhof (1956)

Figure 9 shows that the capacity varies from 1.32 ksf to 5.65 ksf at 3.0 m depth. The most prevalent bearing capacity at this depth falls within 3.0–4.0 ksf. In the south, southeast, east, and some central places, the bearing capacity is between 2.0 and 3.0 ksf, while values of 4.0–5.0 ksf are found in the northern, northeastern, northwestern, and a few central zones. A few scattered locations in the north and easterly zones have the bearing capacity of 5.0 to 5.65 ksf. The lowest bearing capacity range, between 1.32 and 2.0 ksf, is identified in certain east and southeast areas.

Fig. 9.

Bearing capacity at 3.0 m depth (Meyerhof, 1956). This map was generated by the 1^st author using ArcGIS 10.5 software with the Ordinary Kriging interpolation method (https://www.esri.com/software/arcgis).

Using the method proposed by Meyerhof (1956), the bearing capacity at this depth vary between 1.32 ksf and 5.65 ksf. Approximately 7% of the area falls within the 1.32–2.0 ksf range, 29% is in the 2.0–3.0 ksf, 36% within 3.0–4.0 ksf, 15% in 4.0–5.0 ksf, and only 3% of the area exhibits the highest capacity of 5.0–5.65 ksf.

Bowles (1988)

Figure 10 shows the bearing capacity variation in the DMDP area at a depth of 3.0m. According to Fig. 10, the bearing capacity varies between 2.0 ksf and 8.5 ksf. In most areas, the bearing capacity is between 4.0 and 6.0 ksf. The minimum range of bearing capacity 2.0 and 3.0 ksf is located in some eastern, southeasterly, and a small part of the central regions of the DMDP area; the lower N-values corroborate this at these places. In the south, southeast, east, and some central places, the bearing capacity is between 3.0 and 4.0 ksf. The bearing capacity of 6.0–7.0 ksf is observed in the northern, northeastern, and northwestern regions. A few scattered locations in the north and eastern places have the bearing capacity of 8.0–8.5 ksf.

Fig. 10.

Bearing capacity at 3.0 m Depth (Bowles, 1988). This map was generated by the 1^st author using ArcGIS 10.5 software with the Ordinary Kriging interpolation method (https://www.esri.com/software/arcgis).

Based on Bowles (1988) method, 7% area with bearing capacities between 2.0 and 3.0 ksf, 19% within 3.0–4.0 ksf, 23% falling between 4.0 and 5.0 ksf, 24% in 5.0–6.0 ksf, 21% lies within 6.0–7.0 ksf, 6% between 7.0 and 8.0 ksf, and under 1% exceeds 8.0 ksf.

Comparison of bearing capacity at 1.5 m depth and 3.0 m depth

Figures 7, 8, 9, 10 represent the differences in bearing capacity changes between 1.5 m and 3.0 m depths while using Meyerhof (1956) and Bowles (1988) formulas for bearing capacity. By analyzing both zonation maps of bearing capacity, we can conclude that bearing capacity increases from a depth of 1.5–3.0 m. This increase in bearing capacity is due to the increase in SPT-N value at 3.0 m depth. Field borehole logs indicated the presence of fill or loose topsoil at several locations up to approximately 1.5 m depth, reflected in lower SPT-N values and consequently reduced bearing capacities. At 3.0 m depth, the soils generally transition to natural, more compact strata with higher N-values, resulting in increased bearing capacity. At the identical position, depth, and footing dimensions (6 × 6 ft) (1.83 m × 1.83 m), the Bowles (1988) technique produced bearing capacity values around 50% more than those derived from the Meyerhof (1956) correlation.

Pile capacity zonation maps

This study assessed the bearing capacity of pile foundations at 12.0 m and 20.0 m utilizing various theoretical methodologies. Two (2) Pile capacity zonation maps were created using the Ordinary Kriging interpolation method. The objective was accomplished by utilizing soil investigation data obtained from boring logs and SPT-N measurements across various sites in the DMDP region. For clayey soil, pile capacity was evaluated using the end bearing equation proposed by Hara et al. (1971) together with the skin friction equation of Shioi and Fukui³⁰. For sandy soil, skin friction and end bearing were calculated using Meyerhof’s formula (1976).

Pile capacity at 12.0 m depth

Figure 11 shows the pile capacity variation in the DMDP area at a depth of 12.0 m. The pile capacity varies from 50 to 200 kip. For these values, a precast pile size of 12″×12″ and a length of 12.0 m is used at a different location depending upon conditions. In most areas, the pile capacity is between 100 and 125 kip. The minimum range of pile capacity 50–75 kip is located in some eastern, northeasterly, and a small part of the central regions of the DMDP area. In the south, southeast, east, and some northeastern places, the pile capacity is between 75 and 100 kip. The pile bearing capacity of 150–175 kip is observed in the northern and western regions. A few scattered places in the north have the pile capacity of 175–200 kip.

Fig. 11.

Pile capacity at 12.0 m depth. This map was generated by the 1^st author using ArcGIS 10.5 with the Ordinary Kriging interpolation method (https://www.esri.com/software/arcgis).

At 12.0 m depth 3% of the area has a capacity of 50–75 kip, 27% of the region in 75–100 kip, 40% of the area has 100–125 kip, 21% of the territory within 125–150 kip, 8% of the region has 150–175 kip, and 1% of the area with 175–200 kip.

The pile capacity was initially estimated using static analysis equations derived from soil investigation data. A total of 18 static pile load test reports at (10.68 ~ 13.71) m depth of 8 buildings were collected from the Public Works Department (PWD) in order to evaluate the accuracy of the maps and the usefulness of the computed capacity. Figure 12 shows the locations of the pile load test points. Table 3 represents the estimated pile capacities obtained from static analysis equations and the corresponding results from field pile load tests. The models’ accuracy was evaluated using efficiency metrics such as RMSE, Coefficient of Determination (R²), and Coefficient of Correlation. The RMSE was 5.76%, MAE 4.61 and the (R²) was 0.6. The assessment exhibits a coefficient of correlation of 0.77, suggesting a significant positive linear correlation. This indicates that zonation maps developed in this research show a limited error range. This deviation can be ascribed to various factors, such as variability in soil stratification, a limited number of available test sites, and the inherent simplifications of empirical correlations.

Fig. 12.

Pile load test conducted points. This map was generated by the 1^st author using ArcGIS 10.5 software (https://www.esri.com/software/arcgis).

Table 3.

Collected pile load test report information.

SL	Name of the building	Latitude	Longitude	Pile size (mm)	Pile length (m)	Allowable load (kip) from pile load test	Anticipated pile capacity (kip) from zonation map
1	International training complex at BPATC, Savar	23.869596	90.271632	300×300	12.0	122	125–150
2	Construction of multipurpose building (5 storied building with 5 storied foundation) at West Agargaon, Sher-E-Bangla Nagar, Dhaka	23.781342	90.373715	300×300	13.71	133	125–150
		23.781385	90.373747
		23.781442	90.373734
3	Construction of 4 storied 3500 sft sub-station at Paikpara, Mirpur, Dhaka	23.793457	90.352947	300×300	12.19	111	100–125
4	Construction of 14 storied residential building (Grihosuchona, B-1) for low and middle income people at Mirpur-16, Dhaka	23.806375	90.389638	300×300	11.6	111	100–125
		23.806287	90.389741
		23.806254	90.389908
		23.806177	90.390002
5	Construction of 14 storied residential building (Grihosuchona, B-3) for low and middle income people at Mirpur-16 , Dhaka	23.806856	90.390389	300×300	10.68	111	100–125
		23.806857	90.390264
		23.806751	90.390533
		23.806671	90.390591
6	Construction of 1500 sft 13 storied + 1 basement residential building (building No-A) at Zigatola, Dhaka	23.737950	90.373523	300×300	12.19	111	125–150
7	Construction of mosque at 456 flat project at Malibag, Dhaka	23.747458	90.414800	300×300	13.71	111	100–125
		23.747408	90.414770
8	Residential tower building at Motijheel thana compound Dhaka	23.730156	90.417327	300×300	12.19	122	100–125
		23.729928	90.417176

Pile capacity at 20.0 m depth

Figure 13 shows the pile capacity variation in the DMDP area, ranging between 50 and 350 kip at 20.0 m depth. For these values, a precast pile size of 12″×12″ and a length of 20.0 m is adopted at different sites according to local ground conditions. In most areas, the pile capacity was between 150 and 200 kip. The minimum range of pile capacity, 50–100 kip, is located in a tiny northeastern region of the DMDP area. In the southwest, southeast, east, and some northeastern places, the pile capacity is between 100 and 150 kip. Pile capacity ranging from 200 to 250 kip is prevalent in the northern, central, and southern regions. Northern and some southern regions have pile capacities of 250–300 kip. A few scattered places in the north have the pile capacity of 300 to 350 kip.

Fig. 13.

Pile capacity at 20.0 m depth. This map was generated by the 1^st author using ArcGIS 10.5 software with the Ordinary Kriging interpolation method (https://www.esri.com/software/arcgis).

At 20 m depth, the pile capacity ranges between 50 and 350 kip. The distribution reveals that 1.5% of the region has 50–100 kip, 23% falls in the range of 100–150 kip, 41% has (150–200) kip, 8% of the area within 250–300 kip range, while 3.5% reaches 300–350 kip.

Conclusions

A variety of geotechnical features within the DMDP region are depicted in the geotechnical maps produced by this study. GIS-based bearing capacity zonation maps were obtained by adopting the spatial analyst Ordinary Kriging method. 4680 SPT- and different soil type data were used to develop bearing capacity zonation maps of shallow foundation at 1.5 m and 3.0 m depth, as well as pile foundations at depths of 12.0 m and 20.0 m.

At a depth of 1.5 m, Meyerhof’s (1956) method represented that the highest percentage (51%) fell under the 2.0–3.0 ksf bearing capacity zone, while 26% corresponded to the 3.0–4.0 ksf zone. Similarly, the Bowles (1988) method illustrated the most frequent zones as 3.0–4.0 ksf (35%) and 4.0–5.0 ksf (27%), which together constituted more than 60% of the area.

From the values obtained by Meyerhof’s (1956) method, at 3.0 m the bearing capacity is between 2.0 and 4.0 ksf for about 65% of the area, and 25% between 4.0 and 5.0 ksf. Bowles (1988) method, however, gave much larger values, where the dominant ranges occurred between 4.0 and 7.0 ksf, which made up almost 70% of the area, and the capacity more than 7.0 ksf occurs only in nearly 7% of the region.

The bearing and pile capacities estimated through the Meyerhof (1956) and Bowles (1988) formulations were adjusted with standard correction factors to account for local conditions; nonetheless, the original calibration showed no regional specificity. Thus, the findings must be interpreted carefully, acknowledging the uncertainty linked to empirical assumptions. It is advisable for future research to integrate CPT data or utilize locally calibrated models to improve the precision and regional relevance of these correlations.

The pile capacity zonation map at 12.0 m shows that 40% of the region is in the 100–125 kip range, 27% is in the 75–100 kip range, and 21% is in the 125–150 kip range. Pile capacity over 150 kip is comparatively limited at this depth. The pile capacity increases from 50 to 350 kip at a depth of 20.0 m. The peak zones, which made up about two-thirds of the area, are between 150 and 200 kip (41%) and 200 and 250 kip (23%). About 12% of the region has a capacity of more than 250 kip.

The created geotechnical maps will decrease time and costs needed for the initial assessment of major projects and act as a database for smaller residential buildings. The zonation map will offer a clear understanding, even for the anticipated future settlement. The geotechnical zonation maps created in this study provide a realistic basis for preliminary foundation design. Besides assessing soil strength, these maps serve as a decision-support tool for identifying suitable foundation types. Areas with appropriate bearing capacity at shallow depths are ideal for shallow foundations, whereas zones with lower capabilities require pile foundations to reach competent strata. Thus, the maps enable designers in making informed, site-specific, and economically efficient foundation decisions.

Limitations and practical use

The resulting maps’ accuracy is strongly dependent on the amount and quality of data available from soil investigations; however, their dependability could be compromised by data shortages. The conclusions of this research are site-specific and may not be immediately applicable to places with diverse geological conditions or environments. Bearing capacity estimates at 1.5 m depth are subject to higher uncertainty, primarily due to the presence of potential fill or disturbed soil layers near the surface, which may not be adequately represented by the existing subsurface data. Additionally, this study did not take into account a variety of geotechnical criteria that may have an impact on the final evaluation, including consolidation parameters, moisture level, fiber content, wet/dry unit weight, specific gravity, and seismic parameters. Further subsurface explorations and the gathering of further geotechnical data inside the research region could enhance the bearing capacity zonation maps’ validity and dependability. Furthermore, plate load tests on full size footing can be conducted at sites to validate the results of bearing capacity derived from zonation maps of shallow foundation, and more pile load tests can be undertaken to validate the pile capacity derived from zonation maps.

It should be highlighted that the maps generated in this study are intended to be useful for early-stage design considerations, risk detection, and viability assessments. These maps do not replace extensive, site-specific geotechnical investigation. Comprehensive analysis and field testing are still required to ensure safety, dependability, and conformity with technical requirements, particularly when designing and constructing critical infrastructure. Nonetheless, the maps of bearing capacity shown here are aimed to offer a broad insight into the load-carrying capacity of the DMDP area.

Author contributions

M.S.R. wrote the draft, developed the graphs and analyzed the data Z.S.A edited the draft M.A.A. supervised the work and edited the paper.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.