The Vs30 obtained from two of the methods discussed above for all the survey sites are presented in Table 1 along with Vs30 from previous study (JICA 2018). The location of sites along with notations used in Fig. 1 are also presented. Kriging interpolation in GIS environment was used to illustrate the Vs30 in a map (Fig. 4.a, 4.b, and 4.c). It is evident that the soil of Madhyapur Thimi and Bhaktapur municipality (near the centre of the basin) are soft with Vs30 of about 209.34 to 251.46 m/s in the northern part of Madhyapur Thimi municipality. Similarly, Suryabinayak and Changunarayan Municipality show a mix result. The Vs30 in the west and central part ranges from as low as 152.50 m/s to 267.28 m/s representing soft soil in these areas. Moreover, in the eastern edge of the study area, Vs30 reaches up to 522.24 m/s from Konno and Kataoka’s method in Suryabinayak Municipality whereas, in Changunarayan Municipality the Vs30 ranges from 165.19-296.34 m/s in the east. As some part of the study area, mainly the south-east and north east, are covered with a dense forest, we could not carry out the field survey. Though these areas are most probably bedrock site, the interpolation gives result of Vs30 as 300 to 350m/s due to lack of survey points.
From the result we can say Vs30 is lowest at western and central part of the study area, which are the areas with thicker basin sediments when compared to the basin-edges in eastern and southern region of the study area. In the core area of Bhaktapur district, the depth of soft sediments reaches 235m to 250m (Dhonju et al., 2023; Kawan et al., 2019). When comparing the Vs30 from JICA (2018) with the map obtained from Konno and Kataoka and from inversion, we can see the similarity with much closer result from the first method. In the northern and eastern part of the study area, despite being marked as a hard soil site (JICA 2018), we found the Vs30 at those areas varying from 183.77 to 296.34 m/s only at the north. It indicates the presence of soft soil at those locations. Also, we found the sandy soil to deeper depth at these locations (Fig. 6.a).
As a basin filled with fluvio-lacustrine deposit it can be observed that there are rock hillocks jutting out of the basin-fill sediments in the Kathmandu Basin (Bijukchhen et al., 2017). These elevated hillocks along with the basin-edges with higher altitude than the lower basin have thinner fluvio-lacustrine deposit and are considered rock-sites with bedrocks found at much shallower depth. If this were true, the areas with higher altitude would have higher Vs30 as well.
In the graph of altitude versus Vs30 (Fig. 6.b), we can see that despite being at high altitude and near basin edges Vs30 was comparatively lower than those at other edges of basin in areas lying in northern and eastern part of the study area. Hideki (2015) mentioned in his study that the elevated site of Changunarayan temple, in the northern part of the study area, is a hillock of softer basin sediments rather than being a rock exposure unlike other elevated places at the basin edge. Our study shows the Vs30 of site around Changunarayan(C13) is 194.07 m/s and 212.89 m/s respectively from two methods corroborating the statement. Similarly, north western region of the study area has a Vs30 ranging from 202.04-268.11 m/s.
Thus, we can observe that at the south eastern territory of the Kathmandu Basin, soil having a high Vs30 is present with the maximum obtained at site Sv 19 (522.24 m/s and 521.98 m/s from both methods). One can clearly see the exposed bedrock at this site (Fig. 7). Similarly, the western and central region of the study area where sediment deposit is thicker, returned a lower Vs30. In this way Vs30 determined in our survey were in accordance to the geology of the study area.
We classified the study area into several classes as per NEHRP site classification system (Fig. 5.b). As the Vs30 at the eastern and south-eastern hills are higher, we categorize them in class C. Similarly, most of the region of study area lies in class D including the core area of Thimi, Bhaktapur, Changunarayan, and Suryabinayak municipalities. The western part of the study area has Vs30 less than 175m/s, and are grouped in site class E. Similarly, junction of Thimi, Bhaktapur and Changunarayan municipality and some smaller areas of these municipalities are categorized in group E (Fig. 5.b). As soft soil with low Vs30 amplify seismic waves, greater damage during earthquake at areas with softer soil can be expected as compared to sites with stiffer soil. The area classified as type E having a low Vs30 were the same areas where 2015 Gorkha earthquake caused serious damage to buildings and infrastructures (Chaulagain et al., 2018; Dixit et al., 2013; GEER, 2015; Shakya & Kawan, 2016). The results obtained during our study will be helpful to concerned authorities for further preparedness to build resilience against possible damages during earthquake.
Validation with previous works
Table 1 shows Vs30 obtained from both methods employed along with Vs30 from JICA (2018) where one can clearly see a good correlation. There is 79.315% of correlation between Vs30 obtained from Konno and Kataoka’s method with that from JICA (2018) while 73.66% of correlation between Vs30 obtained from inversion and that from JICA (2018). The result of correlation is shown in Fig. 8.a and 8.b.
The Vs30 results obtained by both methods are validated by comparing with results of previous works and bore hole logs for reliability. The result of average shear wave velocity up to 30m obtained by two approaches in this study along with that of JICA (2018) are presented using a line graph in Fig. 9. Here Vs30 are shown in y-axis and survey sites are shown along x-axis presented serially as per Table 1. In the figure, we can see Vs30 results of JICA (2018) at stations like at 25, 26, 27, 32, 61, 67, 70, and 71 are higher compared to our results however they followed same natured trend of data except at station 61 to 71. In both Vs30 results obtained by employing above mentioned two methods at these stations, values of Vs30 are higher than at other stations which is a trend apparent in the result of JICA (2018). These stations where Vs30 is higher are hard rock sites and are located at the south eastern edges like Sv8, Sv9, Sv10, Sv11, Sv15, Sv19, and Sv20 of our study area. It can also be justified by the HVSR result at these sites where a predominant frequency is higher. A typical example of HVSR at one of these sites with high Vs30 i.e. Suryabinayak Tindhara (SV11) and a site with low Vs30 i.e. Lokanthali (TH1) are shown in the Fig. 10 and Fig. 11
The Vs30 from both approaches were also in accordance with previous Vs30 study carried out in the Kathmandu Basin (Table 3). Takai et al. (2015) measured the shear wave velocity at the four permanent seismic stations in Kathmandu Basin. In one of them -THM the shear wave velocity is 150 m/s. In this study, site Th4 which is near THM has Vs30 of 189.541 m/s and 191.446 m/s from both approaches. De Risi et al. (2021) calculated Vs30 at 15 different locations of Kathmandu Valley with direct geophysical method. Among them, four sites were closer to our survey sites-Th15, C11, Th4, and B7. The comparison of Vs30 from these two different studies is shown in Table 1.
Table 3
Vs30 comparison with previous literature
Site | Vs30 from literature (m/s) | Vs30 from Konno and Kataoka (m/s) | Vs30 from inversion (m/s) |
Th15 | 198 (De Risi et al., 2021s ) | 198.67 | 209.22 |
Th4 | 170 (De Risi et al., 2021 ) | 189.54 | 191.45 |
B7 | 216 (De Risi et al., 2021 ) | 267.28 | 220.74 |
C11 | 203 (De Risi et al., 2021) | 216.88 | 222.78 |
Th4 | 150 (Takai et al. 2015) | 189.54 | 191.45 |
Validation with the actual bore hole log data
The result obtained by using the nearby soil profile as initial soil model for inversion along with its dispersion curve returned results which was a good match with actual soil profiles from borehole data. The profile and dispersion curve of site Th7 was used to calculate velocity structure at sites Th-8 and Th-15 (Fig. 5.a) by inversion. At these sites, the inversion results matched the actual soil profile justifying the suitability of the method in predicting the actual soil profiles at the site. Similarly, the soil profile and dispersion curve of site Th8 and Th3 were used as initial soil model for obtaining profiles at sites Th5 and Th4 respectively. The obtained soil profile matches the actual soil profile obtained from soil test report in all the sites tested. The final result of Vs30 obtained by this approach is shown in Fig. 4.b and Table 1. We also determined the good correlation between the results of Vs30 obtained from two approaches with that of Vs30 of JICA (Fig. 8.a and 8.b).