The soil names of soil drilling-logs are processed numerically to assess the vertical variation of the soil and thickness of ground.
We analyzed soil drilling-logs down to 12 meters obtained from 149 locations in the eastern region of Muroran, Hokkaido, Japan.
Soil names are classified into 13 sequential categories and a set of numerical scale is given to each of them. The numerized drilling-logs are transformed to the spectra of Haar, Slant, Walsh function system. Using the most substantial function system, spectral zoning maps are drawn to understand the ground condition and to estimate categorized drilling-logs at arbitrary locations.
The results are as follows:
1) Haar function system provides the most abundant information on the ground condition. This function system was adopted in this study since it is most suitable for the spectral analysis.
2) In the average, 99% of numerized drilling-logs can be re-created from the 8 strongest spectra and almost 100% from 16 strongest ones.
3) D. C. spectrum component is most informative on the ground condition.
4) In the zoning map of D. C. spectrum component, zones with small fluctuation of numerized drillinglogs show extremely strong or weak spectrum intensity.
View full abstract