Manual
ExploStat (updated at 1 december 2001)
Introduction
ExploStat
contains tools for exploring spatial variability, ordinary Point Kriging,
Irregular Block Kriging and other deterministic interpolation methods. Furthermore there
is a method for iterative optimizing sample plans. Its most important power is
visualization of spatial data. For exploration of spatial data it offers
linking the variogram and the histogram with points on the map. Selection tools
are available to guide the user through large data sets. Plotting transects of interpolated surfaces creates the opportunity to evaluate
the effect of interpolation parameters. An important feature is presentation
and processing of drill core information. As a result ExploStat offers tools for mapping sediment layers from drill core data.
ExploStat handles points, lines, polygons and raster data. It offers editing of
spatial data with the mouse. In fact, it is a GIS program. The vector data are
stored as a set of line segments, an arc-node structure is not yet used.
Several ASCII formats are supported for vector and raster data. Attribute data
are read from ASCII-CSV tables and DBASE IV tables. Just as the most well known
GIS-programs ExploStat create project files where information is stored about
data sets and location of data sets on the computer. A example data set and
it’s associated project file is available on the website of ExploStat. The
topics in this manual will be described in an order that is defined by the menu
items and speed buttons of ExploStat.
OPEN
Tables: A new window appears with a menu
bar. Four DBASE tables can be opened. ExploStat creates a database from these
tables. The tables are related to each other by common attributes. This
database represents drill core information. Each table contains a set of
mandatory attributes and a set user defined attributes. The mandatory
attributes are needed for linking the tables and building a database that
represents properties of drill, sediment layer and sediment sample objects. Table 1 contains the attributes of the
drill location. The mandatory attributes are: Drill code (string), x,y and z
coordinate (floating point). Table 3 contains the descriptions of sediment
layers. The mandatory attributes are: Drill code, number of layer, top of layer, bottom of layer and sediment
description (string). Table 4 contains the location of the samples in the drill
core: Drill code, analysecode, top of sample, bottom of sample are the
mandatory attributes. The analysecode links table 4 with table 5. Table 5
contains the properties measured from the sample. Here the attribute
analysecode is only mandatory. Table 2 is not yet implemented, it should
contain the drill method in case of more then one method at a drill location.
All mandatory attributes are listed in the window where the tables are loaded.
Click on the attribute names to select a attribute. Select in the same order as
the list indicates left of the table window. After selecting the mandatory
attributes you can select additional attributes that will be showed when you
query a drill location on the map. When all attributes are selected click on
the OK button at the top of the window. The you can open the next table. Begin
always with table 1. You have to open at least table 1 and table 3 to make the
database active. After loading table 3 a legend have to be defined. Click on
the first name of the list with legend items. After then you can assign a color
to the selected item by clicking on a radio button left of a color. When ready
click on the button cont. legend. To get information about the sediment samples
table 4 and table 5 have to be loaded. All the information about the legend and
the selected attributes is stored in the project file.
Point
layer: Three
ASCII formats of tables with x,y locations of point locations and their
attributes can be loaded. First the well known CSV format where values are
separated bij comma’s or other list separators (the names of the attributes are
in the first row), secondly the GEO-EAS format coming from the first user
friendly Kriging program with the same name, created in the eighties in
California, thirdly the most simple format: only three columns with x,y and z
(without a header). At the moment only one point layer can be loaded.
Line
layer: Five line
layers can be loaded. There
are two ASCII formats supported: first the MOSS format, a format that is
available in most of the GIS programs on the market, secondly a coordinate
format that is created by a ARCVIEW-extension which is available at the
ExploStat website.
Mask
layer: This is a
special case of a line layer. The mask plays a key role in spatial selection.
The area within a mask polygon can be made active and works as a point in
polygon operation. Histogram and variogram are created from the point locations
which are activated by this operation.
Grid
layer: One ASCII
format is supported: the ESRI ASCII grid file format. Up to 30 grids can be
loaded.
SAVE
Point
layer: The active
point layer is saved. The last plotted line layer is active.
Line
layer: The active
line layer is saved. The last plotted line layer is active.
Grid
layer: The active grid layer is saved. The last
plotted grid layer is active.
DISPLAY
Create
Legend: This is
not a pure display function but is controls the presentation of the legend. It
serves to modify an existing legend by choosing other colors for the legend
items.
Size
Drill Plot: The
size of the rectangles that represents the drill core descriptions and the
properties of sediment samples can be changed. The manual choice enable the user
to define separately vertical and horizontal dimensions of the drill core
rectangles.
INTERPOLATION
Three simple interpolation methods are available. With the interpolation results, spatial variability can be visualized. The Kriging interpolation is activated by a speed button, this is described later in the chapter about the speed bar.
Voronoi
spread: This
method assign a value to a gridcel coming from the nearest point location. The
resulting surface is characterized by sudden changes of value. Plotting this
surface on the screen of the map gives a good insight to spatial variability.
Inverse
Distance weighting:
This is a classical interpolation method where weighting factor is assigned to
point locations in the neighborhood of
a gridcel. The value of this weighting factor depends on the distance between
point location and the gridcel where interpolation takes place. The resulting
surface is smoother then the Voronoi surface, but still it gets an impression
of spatial variability.
COLORS
Color-palettes:
Four standard
palettes are available. i) Deep blue, ii) From Red to Yellow, iii) From yellow
to red and iv) From orange to green.
Contrast
colors: The value
range of the data in the grid is classified into 8 pieces. Each class is
assigned a different color. Because the colors are quite different the
separation between classes is enhanced visible on the map.
Extremes
active: If active
is set (default) the upper en lower cell values will be plotted, otherwise not.
Stepped/Continuous
256 colors are used over the total range of values (continuous). If
stepped is selected, only eight colors are used, each class in the legend is
assigned a constant color.
SPEED BUTTONS
A series of speed buttons is available for query, zooming and to activate panels with analysis tools.
Open project File, after a project has been
opened the speed button opens a line layer or the latest opened type of file.
Save project file
Activates all functions of ExploStat. These
additional functions are not described in this manual.
Activates a panel with checkboxes that
controls the display and calculations of drill cores and sediment sample
properties.
The panel with drill core properties shows several options to control the
display of drill cores and sample properties. The upper left box shows the
classified concentrations, class 4 means a high degree of pollution. In the
lower right box two drill cores are displayed with mainly sandy sediments. At
the right drill core sediment samples have been analyzed, the results are
plotted at the right side of the sediment description. The attribute panel
contains chemicals that were selected from table 5. The checkbox delete means
removing a drill from the set by flagging it. In table 1 a column is available
that contains the values one or zero. Zero means the drill is out, but the
information about the drill is still in table 1. There are several options to
calculate with attributes of the drill cores and properties of the samples. The
results of each calculation (selection of values) are assigned to the current
point layer. To inspect the results you can display the labels of the point
layer or a histogram. In addition it is possible to create a experimental
variogram and do some Kriging or other interpolation.
Radio button panel
Calc_attributes:
1)
Analysis:
Concentrations of the chemical constituent selected in the attribute radio
button panel are assigned to the point layer.
2)
Sediment:
Thickness, Bottom or Top of a sediment layer is calculated. Select one or
more types of sediment from the panel
Layers with the checkboxes.
3)
Quality:
This function works on classified concentrations. After clicking on the radio
button you have to choose from 5 class values. The radio button in the panel
Calc Layers determines whether Thickness,
Bottom or Top of sample is selected and assigned to the point layer. This
function works only if the numbers 0,1,2,3,4 are available in the selected
column of table 5.
4)
Anal.
Sedi: Only the chemical concentrations of samples in the checked layers (panel
Layers) are selected and assigned to the point layer.
Query mode: Properties of objects as tables
and grid cells are displayed.
Digitize inside mask: This is an important
function, when a vector layer is available (mask layer) with closed polygons
the area within one or more polygons can be activated by digitizing inside a
polygon. This enables selection of points and areas where analysis functions
must be carried out.
Zoom in: First push the mouse button on
the upper right corner or the lower left corner of the area you want to zoom
in. Drag the mouse until the size of the area of your interest is reached and
then release the button.
Zoom to previous zoom
Zoom out 2 times from location of mouse
cursor
Geographical
Window Parameters: ExploStat supports vector and raster structures; therefore a
cell size is always defined. Also when no grid layer is present. When a point
layer or a vector layer is loaded the default number of cell columns is 500.
The corresponding cell size is calculated based on the minimum bounding
rectangle of the point or vector layer.
You can change the Cell Size, the number of columns and rows will be
recalculated. Existing grid layers will be deleted. New values for the extend
of the geographical window can derived from the current zoom. Again existing
grid layers are deleted. Thus before changing window parameters existing grid
layers have to be saved, otherwise they will be lost.
The panel with properties of grid layers will
be activated.
Only the existing grid layers will be shown, thus when ExploStat is
running with by example three layers the panel looks like:
The filename of the grid layer is at the left. The first column contains
the value of the grid cell where the cursor resides. The second column shows
the minimum of the value range that is colored on the map. The third column
contains the maximum of the range. The user can enter different values to alter
the range. After entering new values for a range of a grid layer click on Set
Range to set the current range values for the clicked grid layer. Clicking on
the name of the layer results in drawing the grid layer on the screen. The last
clicked layer will be the active grid layer. This is important because all analysis or calculations concerning grid
layers will be applied on the active grid layer. By example clicking on All
Ranges will set the range of all existing grid layers to the range of the
active grid layer.
Profiles: Transects of the grid layers that are checked
will be drawn along the North-South and West east directions along the sides of
the ExploStat window. The location of the transects in the grid layer is
determined by clicking on a grid cell. If more grid layers are plotted as
transects it is important to assign the same range for all transects, otherwise
comparison between the transects is impossible. If Cross is checked then the
lines are plotted on the map where the transects are located.
Clicking
on 
will extend the fields with the filenames
over the three columns. This is useful when the name does not fit in the box.
Furthermore when grid layers are created with interpolation, some parameters
will be shown in the same box. Then more space will be necessary to show all
information.
Calculation with grid layers: Some simple point operations on two grid
layers are available. Add, subtract, divide and multiply is possible.
Sample Plans: Sample plans on a grid can be generated. The user can select distances
between sample points or fill in another distance.
In this example grid layer one and grid layer two are added. The result
will stored in the next not used grid layer.
Line layers: The panel with line layers becomes visible. Up to 13 line layers are
possible. Here the same rules as for grid layers. Operations are applied on the
active line layer such as digitizing, editing or selecting.
Digitize Line Layers: The most important line layer mask can be digitized from screen, starting with an empty layer. All other
line layers have to be imported before digitizing is possible A digitized line
segment is finished by double clicking. If a polygon is digitized stop just
before the starting point and choose for close segment. So a mask layer can be created or lines can
be added when a mask layer is already loaded.
The mask is a special line layer, therefore it is yellow highlighted. As already explained by digitizing in mask, the mask layer contains a set of polygons that define sub areas that can be activated. Checking the Fill box means that the activated areas are filled with a red color.
Edit Line Layers: Line segments can be deleted.
Select Line Segment: For generating slopes (taluds) in a DTM line segments have to be
selected as sources and borders of a slope (talud).
Point Layer: There is one point layer. Points locations
can have a set of attributes read from file. For calculations with attribute
values there always only one attribute active. A point layer can also be
created from drill core descriptions (below: drill core descriptions).
Histogram: A histogram is made from point locations inside activated areas
(digitize inside mask). If no mask layer is available or checked all, point
locations will be active. All active point locations are red highlighted. When
clicking on a bar of the histogram the point locations will be deselected that
have attribute values within the class of the histogram bar.
The point locations on the map that are deselected turn to yellow. The open circles are in the not activated area. The statistical parameters above the histogram are calculated from the point locations in selection. You can also click on a point location on the map to deselect or select that specific location. To change the number of bars (classes) fill a number in the edit box and click on N Class.
Variogram: The so called experimental variogram is calculated from the point
locations that are in selection.
In the example here the blue experimental variogram is calculated from
all point locations within the activated area. After deselecting the five
points near the left border of the diagram, a new experimental variogram is
calculated and resulted in the red line. The number of lags and the width of
the lag intervals can be changed. Clicking on New Scale does rescale for the
last plotted variogram and only this one is replotted. Moving the cursor over
the variogram area shows the number of point pairs for the lag where the cursor
is. To fit the variogram click on Fit Vario, a new window appears: see below.
The
red experimental variogram is fitted. Default there is a spherical type. The
nugget, sill and range can be fixed. The fit method assign weights to each lag
based on the number of point pairs. Then the least squares criterium is
calculated. The weights can be disabled but this is not recommended. The values
for nugget, sill and range are used for the Kriging if the user wants. The user
can also enter these values manual.
Lines between point pairs:
To
explore spatial variability you can click on the area below the experimental
variogram. Moving the cursor along the vertical axis for a certain lag show the
number of point pairs that exceed the level of semi variance that is shown in
the edit box semiv>. These point pairs are shown by connecting the points
with lines. In the example below for the first lag and the third lag values
haven been clicked. The colored surface is a result from Voronoi Interpolation.
The light blue areas are relative high values so there the spatial variability
is higher then at other locations.
Kriging and optimizing sampling plans
There
are two types of Kriging possible: Ordinary Point Kriging (OPK) and Irregular
Block Kriging (IBK). Furthermore on the same panel (see below) the sample plan
optimizing can be enabled, because the optimization is based on minimizing of
the IBK variance.
In this panel the default values for the variogram are displayed. A model can be selected. Clicking on Set
Values From Fitted Variogram will activate the parameters from the last fitted
variogram. If the checkbox In Active Layer is checked the results of the OPK will be stored in the active layer
otherwise two new grid layers are created to store OPK values and OPK
variance. Clicking on the radio button
Ir. Block will display the panel below.
In
this case the values of the last fitted variogram are enabled. Three progress
bars are present to monitor the (sometimes it takes minutes) calculations. The
size of the discretization is determined by the current grid cell size. If
there are many cells in a block region the calculation of the within block
variance is the most time consuming. Therefore the user is warned when there
are more then 5000 discretizations, continue will cause long time for
calculation. A solution is to increase the cell size with .
Optimizing
a sampling plan is started by clicking on the button Optimize Sampling Plan.
The panel left appears to enter the parameters of the process of optimizing.
Optimum value set the target IBK value. Start distance sets the distance
between sampling points when the optimizing starts. Step size controls the
amount of distance to decrease during optimizing. After the target IBK value is
passed the step size is increased with a