Page 138 - Numerical Analysis and Modelling in Geomechanics
P. 138

STABILITY OF UNSATURATED SOIL SLOPES 119
            rainfall intensity (e.g., 70 mm/hr adopted in Hong Kong) as a landslide warning
            signal could be potentially very misleading. In other words, a slope can be perfectly
            stable if the initial main water table is low, even if subjected to extremely high
            rainfall intensity.

                                Influence of rainfall duration

            The  relevance  of  long  periods  of  rainfall  to  slope  stability  has  attracted
            considerable attention and debate over the years. Based on empirical correlation
            found  between  rainfall  and  landslide  data,  researchers  such  as  Peck  [30]  and
            Lumb  [8]  have  suggested  the  significance  of  10-day  and  15-day  antecedent
            rainfall on slope stability respectively. By making use of automatic rain gauges
            to obtain better and more data of rainfall distribution and intensity, Brand [1] and
            Premchitt  et  al.  [3]  concluded  that  the  majority  of  landslides  are  induced  by
            localised short duration rainfall events of high intensity. Antecedent rainfall is not
            of major significance. However, all these traditional correlations between rainfall
            data and slope failure events ignore local geological and hydrological conditions.
              To investigate the effects of rainfall duration on slope stability, 1 in 10-year
            rainfall  records  collected  by  the  Hong  Kong  Royal  Observatory  between  1980
            and  1990  (Lam  and  Leung  [23])  were  adopted  for  parametric  studies.  The
            rainfall events considered are shown in Figure 4.8. In the parametric analyses, it
            is  assumed  that  the  rate  of  infiltration  is  equal  to  the  rainfall  intensity.  The
            hydraulic heads at HI and FG (see Figure 4.4) were specified at 62 mPD and 6 mPD
            respectively, and other relevant mean values from Table 4.1 were adopted for the
            parametric analyses. At the end of each prolonged rainfall, a 2-hour rainstorm of
            high  intensity  (74  mm/hr)  is  also  included  in  each  analysis  to  investigate  the
            influence  of  antecedent  rainfall  on  the  subsequent  performance  of  the  slope
            subjected to the intensive 2-hour rainstorm.
              Figure 4.9 shows the locations of the main water table between sections JJ and
            KK  of  the  natural  slope  under  various  rainfall  conditions.  It  can  be  seen  from
            Figure 4.9a that the intensive 2-hr rainstorm cause a significant rise of the main
            water table at the toe of the cut slope (near to Section A-A). This can probably
            explain why many slopes with a marginal factor of safety fail after an intensive
            rainstorm. For slopes with a higher factor of safety, a substantial amount of rain-
            water may be needed to cause landslides.
              Figures 4.9b and 4.9c show a typical groundwater profile after 7-day and 15-
            day prolonged rainfall events with an average intensity of 82 mm/day and 46 mm/
            day  respectively.  Because  of  limited  space,  only  some  computed  results  are
            included.  The  main  groundwater  table  rises  globally.  By  comparing  with
            Figure 4.9a, the influence of the 2-hr intensive rainstorm on the level of the main
            water table is more significant after the 7-day and 15-day prolonged rainfall events
            in  terms  of  the  amount  and  extent  of  the  rise  of  the  main  water  table.  This  is
            probably due to the increase of water permeability (hydraulic conductivity) as a
   133   134   135   136   137   138   139   140   141   142   143