Page 62 - Modern Spatiotemporal Geostatistics
P. 62

Spatiotemporal  Geometry                    43

        Separate    metrical  structures

        These metrics  are convenient  for  many geostatistical applications  because they
        treat  the  concept  of  distance  in  space  and time  separately.  The  main  idea  is
        introduced  by the following definition.

        DEFINITION    2.6:  The  separate  metrical  structure  includes  a  spatial
        distance  \ds\ €  -R+  0  and an  independent  time  interval  dt =  t-z  —1\  e  T,
        en that


            In  Equation  2.11,  the  structures  of  space  and time  are  posited  indepen-
        dently.  Thus  "distance"  has meaning only  at  a fixed  point  in  space  (when  it
        means "time elapsed") or at a fixed time (when it  means  "distance between spa-
        tial locations").  The formulation  (Eq.  2.11)  includes two celebrated space/time
        structures:  Newtonian  and  Galilean.  Newtonian  space/time  is defined  by  the
        following  properties:  between  any two  points  in space/time  (events)  p  and p'
        there  exist  a  Euclidean spatial  distance  \da\  and  a  temporal  interval  dt.  An
        alternative conception  to  Newtonian space/time  is Galilean  space/time,  which
        is defined  so that  between  any two  points  p  and p'  there  always exists  a  tem-
        poral interval  dt,  but only if the two  points  are simultaneous is there any spatial
        distance  \ds\  between  them.
            The  distance  \ds\  can  take  on  different  meanings  depending  upon  the
        particular  topographic  space  under  consideration.  The  following examples ex-
        plore  several  Euclidean  and  non-Euclidean spatial  distances  in  two  and  three
        dimensions.

        EXAMPLE   2.10:  In the  Euclidean  plane  and  in  ordinary  three-dimensional  Eu-
        clidean  space,  \ds\  is  defined  as the  length  of  the  line  segment  between  the
        spatial  locations  Si  and  82  =  Si +ds,  i.e.,  the  Euclidean  distance in  a  rect-
        angular  coordinate  system  is defined by the  Pythagorean  formula





        Frequently,  the  distance concept  most  useful in a particular  application  is non-
        Euclidean.  In  some  physical  situations  it  makes  sense  to  define the  absolute
        distance  between  the  spatial  locations  PI  and  P%  with  coordinates  «i  and
        «2  =  si +ds,  respectively,  as follows  (see, also,  Fig.  2.11)




        Distance  as defined  in  Equation  2.13  could  represent the  length  of the  shortest
        path  traveled  by  a  particle  that  moves from  PI  to  P^,  if  the  particle  is  con-
        strained  by the  physics of the situation to  move along the  sides of the grid; or,
        it  could  represent the  shortest  path  that  a car  must  travel  to  get from  point
   57   58   59   60   61   62   63   64   65   66   67