Report on MIMA mound processing.
April 2006
Larry Theller
Center for Advanced Applications in GIS
Agricultural and Biological Engineering
Purdue University
Introduction:
This project consists of processing two data sets which were recorded at
separate times for a single location. For clarity the maps use a legend that
relates to when the data was processed. The data sets are referred to as “May
2005” and “March 2006.”
The actual survey names are
Map Legend Survey name
May 2005 60X60mBL_EMI_survey.txt
March 2006 2005_1010emsurveyNS.TXT
March 2006 2005_1010emsurvey6060.TXT
Data:
The first data set (May 2005) was recorded as a single stream starting
from one origin and proceeding across the project area in east-to-west and
west-to-east traverse profiles, then in north-to-south and south-to-north
traverse profiles. (Figure 1.) The end of each traverse involves swinging the
recording instrument in an arc as the direction changes and this makes these
readings unreliable. Hence the ends of every traverse are dropped from the
data set. The points recorded during the warm-up period at the beginning of
each survey were also dropped from the data set.
The second survey (Figure 2) was completed as two data sets, as shown
in the table above. For both data sets the traverses were split into north-
trending or south-trending; and east-trending or west-trending.
In both surveys only north-trending or east-trending traverses were used
in an effort to simplify the presentation of the data by removing high-
frequency noise and other effects. All “turns” data was marked as “end” and
ignored in all calculations. (Figure 3.)
Both surveys were corrected for the time lag between the GPS device
recording position and the instrument recording measured values, by moving
the values in the recorded channels two positions back in time to the earlier
position. This reflects the fact that the GPS responds more slowly than the
instrument. In a series of experiments the data was displayed without any lag
and then with one and two station lag added, to examine the effect of moving
the lagtime on the high-frequency noise (or herringbone effect) visible in the
data. The best removal of “herringbone” came with the two position lag. The
two position lag was suggested by literature and personal communication and is
an industry standard.
1
Survey extent:
May 2005
Total points acquired: 4054
North-trending traverse points 883
East-trending traverse points 886
March 2006
Total points acquired: 3793
North-trending traverse points 795
East-trending traverse points 808
Thermal Drift.
The detection of thermal drift is possible if a non-moving instrument is
monitored for 24 hours and that 24 hour period exhibits normal similar
temperature fluctuations to the survey period. It is not possible to calculate a
drift rate with the data supplied because there is no base station data. The
2005 survey might have benefited from this had it been possible. The AUX 2
channels were not resolved well even with a lag correction. The Figure 5 for
AUX 2 using both north- and east- trending data was still displaying a lot of
“intersection problems” after the lag correction.
One interpretation of Figure 5 could be that values were changing
through the day as the 2005 survey progressed south and west; however this is
not entirely clear. A graph of the entire 2006 survey seems to indicate that
thermal drift was minimal for the second survey, (slope of whole line over time
seemed flat) and so no thermal drift correction is applied to either.
Other data related to project:
Aerial Photo: t2424_234.tif (Used on Figures 1 and 2)
NAD_1983_StatePlane_Washington_North_FIPS_4601_Feet
Geographic coordinate system name: GCS_North_American_1983
DEM:
A Digital elevation Model (DEM) was created from 4361 points recorded
with an RTK GPS unit. These data were provided in feet to three decimal
places with a reported average vertical error of 0.04 feet. These data were
provided as both a shapefile and a TIN projected to
NAD_1983_StatePlane_Washington_North_FIPS_4601_Feet
Geographic coordinate system name: GCS_North_American_1983
This is mapped as “Surface Elevation in feet” on Figure 3 and others.
Methods:
The process used for the creation of the following maps involves fitting a
mathematical surface over the measured values at the sample points. Several
parameters are available to manipulate during this process. The type of
equation used, the density of the resultant values, and the number or distance
2
away from current location that the input samples are drawn from, are all
parameters that are user-controlled.
The Tool used for this is ESRI ArcGIS™ 9.1. If one investigates this tool
through their online help, this description of the process is available:
How IDW works
Inverse distance weighted (IDW) interpolation determines cell values using a linearly
weighted combination of a set of sample points. The weight is a function of inverse
distance. The surface being interpolated should be that of a locationally dependent
variable.
IDW lets the user control the significance of known points on the interpolated values,
based on their distance from the output point...
The characteristics of the interpolated surface can also be controlled by limiting the
input points for calculating each interpolated point. The input can be limited by the
number of sample points to be used or by a radius within which there are all points to be
used in the calculation of the interpolated points.
References
Philip, G.M., and D.F. Watson. "A Precise Method for Determining Contoured
Surfaces". Australian Petroleum Exploration Association Journal 22: 205-212. 1982.
Watson, D.F., and G.M. Philip. "A Refinement of Inverse Distance Weighted
Interpolation". Geoprocessing, 2:315-327. 1985.
ESRI Desktop Help, ArcGIS 9.1 May 2006.
Interpolation Procedures:
For all data used in this project the point measurements were
interpolated using Inverse Distance Weighting, with a search radius of 10
meters and a sample number limit of 20 surrounding points. This indicates that
the first 20 points within 10 meters were used for a distance-weighted average,
where closer points are more important than more distant ones.
The grid interval, after some experimentation, was chosen as 0.5
meters.
Figure 4, for example, displays the separation of the input points over a
representation of one of the gridded results. A low-pass filter is applied in all
instances and a “hillshade” or shaded-relief version of this is calculated and
made 65% transparent.
This “hillshade” layer is superimposed on the data layer in the display to
provide the “shading” effect that accentuates the surface.
3
Results:
Apparent Conductivity (AUX2):
The AUX2 and AUX 4 channels are Apparent Conductivity at two different
depth resolutions. For AUX 2 and then again for AUX 4 data, the surface is
fitted to all north and east-trending profiles, and for May 2005 survey the WE
and SN profiles are separated into discrete data sets and a raster created for
each as discussed below.
Figures 4 through 7 portray the Surface Elevation contours over Apparent
Conductivity for the two surveys. Mapped versions of AUX 2 are each presented
after low-pass filter was applied. These are displayed with hillshade set to 65%
transparent.
It is obvious in Figures 4 and Figure 5 that the earlier survey, May 2005
had significant problems with differences in the measured values between the
east-west and the north-south profiles. Lag corrections which helped with the
other layers were also applied to these data, of course, but that was not
sufficient to remove these discrepancies. When the north-trending profiles are
separated from the east-trending profiles, and each is used to create a new
surface, the mismatch between profiles is removed (Figures 6 and 7). The basic
character of the two layers created in this fashion is very similar when both are
displayed in a 3 D view (Figure 8).
Differential in Apparent Conductivity (AUX2):
Figures 9 and 10 display the AUX 2 Apparent conductivity for the two
surveys. For this process the low-pass filtered west-east profiles were
subtracted from the low-pass filtered south-north profiles. This is the
differential between North-trending and East-trending profiles for AUX 2. It is
clear in comparing these two figures that the May 2005 survey (Figure 9.) had
an increasing differential, as the values clearly ramp from south to north.
Figure 10 displays a generally flat differential between the north-and east-
trending profiles.
Figure 11. displays the Apparent Conductivity over the DEM in a 3 D view
that helps with the visualization of the relationship between the surface
topography and the conductivity data.
Figure 12. is a 3D view of the difference in the two differentials for AUX
2 data for the two different surveys – with 2006 over 2005. This illustrates the
different character between the differentials in the two surveys.
Apparent Conductivity (AUX4):
The deeper apparent conductivity maps from AUX 4 were created from
the total set of north- and east-trending points. This map layer is differential
between North-trending and East-trending profiles for AUX 4. This was done for
each survey. A low-pass filter is applied; hillshade of this layer is calculated
and made 65% transparent.
4
The differences between the two surveys were slight and are illustrated
in Figure 13. 3D View: AUX 4 Apparent Conductivity from 2005 over AUX 4
Apparent Conductivity from 2006. The two surveys are displayed individually in
Figures 14 and 15. Figure 16 is an image portraying the AUX 4 layer of May 2005
over a depiction of topography.
Figure 17 is a rendering of the characteristic differences between AUX 2
and AUX 4 values. In this case it is created from March 2006 data.
Differential in Apparent Conductivity (AUX4):
Figures 18 and 19 display the differential in AUX 4 Apparent Conductivity
for the two surveys. Because of the large spikes along the north and east, the
March 2006 data has twice the magnitude of the 2005 data set. There does not
seem to be a strong correlation either with topography (represented by
elevation contours) or between the two surveys.
Magnetic Susceptibility (AUX 3)
AUX 3 is mapped as magnetic susceptibility (Figures 20-23). This is generated
from lag-corrected data and filtered. These maps used both east- and north-
trending profiles, and have hillshade of this calculated and made 65%
transparent. Figures 22 and 23 portray the magnetic susceptibility over the DEM
in a 3D view to show the striking negative correlation with topography.
5
List of Illustrations
Aerial photo and shaded elevation with traverse. Figure 1.
May 2005.
Aerial photo and shaded elevation with traverse. Figure 2.
March 2006.
East-trending traverse over contour lines. March Figure 3.
2006.
AUX 2 from all north and east trending points. May Figure 4.
2005.
AUX 2 from all north and east trending points. March Figure 5.
2006.
3 D Comparison of AUX 2 both surveys. Figure 6.
AUX 2 from east trending points. May 2005. Figure 7.
AUX 2 from north trending points. May 2005. Figure 8.
AUX 2 differential between north and east trending Figure 9.
points. May 2005.
3D View: AUX 2 east-trending versus north-trending Figure 10.
over elevation. May 2005.
3D View: AUX 2 differential between north and east Figure 11.
trending points. May 2005.
3D View: AUX 2 differential both surveys. 2006 over Figure 12.
2005.
3D View: 3D View: AUX 4 Apparent Conductivity Figure 13.
from 2005 over
AUX 4 Apparent Conductivity from 2006.
AUX 4 from all north and east trending points. May Figure 14.
2005
AUX 4 from all north and east trending points. March Figure 15.
2006.
3D View: AUX 4 from all north and east trending Figure 16.
points over elevation. March 2006.
3D View: AUX2 and AUX 4 over elevation. March Figure 17.
2006.
AUX 4 differential between north and east trending Figure 18.
points. May 2005
AUX 4 differential between north and east trending Figure 19.
points. March 2006
Magnetic Susceptibility. (AUX3) May 2005 Figure 20.
Magnetic Susceptibility. (AUX3) March 2006 Figure 21.
3 D View, Magnetic Susceptibility. (AUX3.) May Figure 22.
2005.
3 D View, Magnetic Susceptibility. (AUX3.) May 2005 Figure 23.
6
Illustrations:
Figure 1. Aerial photo and shaded elevation with traverse. May 2005.
7
Figure 2. Aerial photo and shaded elevation with traverse. March 2006.
8
Figure 3 East-trending traverse over contour lines. March 2006.
9
Figure 4. AUX 2 from all north- and east-trending points. March 2006.
10
Figure 5. AUX 2 from all north- and east-trending points. May 2005.
11
Figure 6. AUX 2 from all east trending points. May 2005.
12
Figure 7. AUX 2 from all north trending points. May 2005.
13
Figure 8. 3 D View: AUX 2 both profiles. Separated by profile direction
to remove noise.
North-trending over East-trending. May 2005
14
Figure 9. AUX 2 differential between north and east trending points. May 2005.
15
Figure 10. AUX 2 differential between north and east trending points. March
2006
16
Figure 11. 3D View: AUX 2 differential between north and east trending points,
displayed over topography. May 2005.
17
Figure 12. 3D View: AUX 2 differential both surveys. 2006 over 2005.
18
Figure 13. 3D View: AUX 4 Apparent Conductivity from May 2005 over
AUX 4 Apparent Conductivity from March 2006 over topography.
19
Figure 14. AUX 4 Apparent Conductivity from all north and east trending points.
May 2005
20
Figure 15. AUX 4 Apparent Conductivity from all north and east trending points.
March 2006.
21
Figure 16. 3D View: AUX 4 Apparent Conductivity from all north- and east-
trending points over elevation. May 2005
22
Figure 17. 3D View: AUX2 and AUX 4 over elevation. March 2006.
23
Figure 18. AUX 4 differential between north- and east-trending points. May 2005
24
Figure 19. AUX 4 differential between north- and east-trending points. March
2006
25
Figure 20. Magnetic Susceptibility. (AUX3) May 2005
26
Figure 21. Magnetic Susceptibility. (AUX3) March 2006
27
Figure 22. 3 D View, Magnetic Susceptibility. (AUX3.) March 2006.
28
Figure 23. 3 D View, Magnetic Susceptibility. (AUX3.) May 2005
29