Page 74 - Modern Spatiotemporal Geostatistics
P. 74

Spatiotemporal  Geometry                   55

        and  the  particular  experimental  procedure that  we decide  to  use in  order  to
        obtain  measurements of the  natural  field.
        EXAMPLE   2.21:  One  cannot  decide  by  means of  purely  geometrical  notions
        how  the  distance  di 2  between  the  points p l  and p 2  at  a  3-km  spatial  lag
        and  a  2-day  time  lag  in  a  porous  medium  compares with  the  distance  d\y
        between  the  points pj  and p 2  at  a  2-km  spatial  lag  and a  3-day  time  lag.
        Which  spatiotemporal  distance  is  larger  can  be  decided  only  by  means  of  a
        physical  process as follows:  An  experiment  is performed during which  a tracer
        is  released at p 1.  If the tracer  is detected  at p 2 but not at p 2, then di 2 <  A\y
        with  respect  to the  particular  medium  and experimental  setup.  Also, measuring
        distance  by  means of  fluid-tracer  dispersion  can  lead to  very  different  results
        than  measuring it  by means of electromagnetic  propagation.
        EXAMPLE   2.22:  Consider  two  geostatisticians  A  and  B  who  both  use the
        same  Euclidean  system,  say,  rectangular  coordinates.  As  a  result  of  physical
        considerations,  the  two  geostatisticians  use different  metrical  systems.  Geo-
        statistician  A  employs the  usual  linear  Euclidean  metric  system s itA  =  r)i,AU
        (with a unit distance u), and geostatistician  B  uses the nonlinear metric system
        s itB  =  exp^s] w  (with  a unit  distance  v);  rjj^  and 77^5  are the  number  of
        distance  units  for  A  and B,  respectively.  Since both  systems  are Euclidean,
        the  spatial  coordinates  of  a specified point  P  in  R 3  provided  by the  two  geo-
        statisticians  are related  as follows


        where  EJ  are  real-valued coefficients.  By  adjusting  the  unit  distances  of  the
        two  geostatisticians  along  the  si  axis  so that  SI,B  =  Si tA,  the  remaining
        coordinates  are altered  in the  same ratio; then,  Equation  2.30 yields


         In  other  words,  since  the  two  geostatisticians  use different  metric  systems,
        there  is no way to  adjust  their  measurements for  all three coordinates.  For this
        reason, the two  geostatisticians  obtain  very different  representations of  objects
                                                    =
                                                s
        and  processes in space/time.  The sphere J^ILi f A  P*  °f the first  geostatis-
        tician,  e.g., is for the second geostatitician  an ellipsoid
        the  angles  measured by the two  geostatisticians  are different;  etc.
            The  situation  presented  in the  following example has interesting applica-
        tions in the  domain of relativistic  geostatistics  and has already been  mentioned
        in  Example  2.20.
         EXAMPLE  2.23:  In  space/time,  distances  are  physically  measured  most  effi-
        ciently  using  light  pulses.  Let a flash of light  be emitted from  point PI at time
        t  and reach point P 2 at time t + dt.  The spatial  distance  \ds\  between the two
         points  is given  by

        where c is the speed of light            are the orthogonal  projections
   69   70   71   72   73   74   75   76   77   78   79