Sie sind auf Seite 1von 24

Assignment D4 (Part-B)

GISC9318- Spatial Statistics

Prepared By:
Md Shakhin Uz Zaman
GIS-GM Program Candidates
Niagara College-NOTL
135 Taylor Road, ON
L0S 1J0

Prepared For:
Mr. Ian D. Smith, M.Sc., OLS, OLIP, EP
Professor
GIS-Geospatial Management
Niagara College-NOTL
135 Tylor Road, ON

L0S 1J0

April 12, 2017


`

April 12, 2017


GISC9318D4

Ian Smith M.Sc., OLS, OLIP, EP


Professor at Niagara College
Niagara College
Niagara-on-The-Lake, ON
Los 1J0

Dear Mr. Smith,

RE: Submission of GISC9308 Deliverable 4b- Geostatistical Analysis of


Student Collected Spatial Data

Please accept this letter as my formal submission Assignment 4b (major


assignment)Geostatistical Analysis of Student Collected Spatial Data. The purpose
of this assignment to derive a working ability to report upon the collection of
geospatial data and to describe the data both geostatistically, creation of a continuous
surface using IDW and kriging techniques for interpolation based upon the sample
points measured (Static Water Level). Overall, the IDW method produced a better
surface result.

Should you have any questions regarding the enclosed documents or if there are
technical issues regarding the files feel free to contact me at (416)-806-8759. I look
forward to your comments and suggestions. Thank you for your time and
consideration.

Sincerely,

Md Shakhin Uz Zaman
GIS-GM Certificate Candidate
Project Manager
M.k/J.X

Enclosure: 1) Deliverable 4b report.

2) Study Area, IDW and kriging Maps

135 Taylor Road Niagara-on-the-Lake Ontario L0S 1J0


Tel- (416)-806-8759 Email- mohamedkazi11@gmail.com
Deliverable D4 (Major Assignment)

EXECUTIVE SUMMARY

The geostatistical analysis of student collected spatial data project, was a major
project that included data collection, preprocessing, and surface creation. The project
study area was defined as the municipality of Wainfleet, a rural township in southern
Niagara Region, Ontario. Data used for the project were obtained online, from
Government of Ontario, and included water well data with static water level in metres
which were used as the z-value. Using these static water level values, two interpolated
surfaces were created; an Inverse Distance Weighted, and Kriging interpolation.

The IDW interpolator was used to create a smooth, gradual surface from the Static
Water Level points. The IDW result came out as desired, which showed detail of the
Static Water level change throughout the study area. The simple Kriging interpolator
was used to create a more exact surface which in turn produced a more a stricter
surface.

Data coverage within the study area posed slight problems with some data gaps and
small clusters of data points. These areas slightly affected the accuracy of both
interpolation techniques, and the surfaces would be better suited if these gaps could
be filled in with supplemental water level measurements.

In all, both interpolation techniques accurately performed the job they were set out to
do. The added geostatistical power of any kriging technique allows for a stricter surface
to be derived than that of an IDW due to the interactive modelling and added
parameters.

i|Page
Deliverable D4 (Major Assignment)

TABLE OF CONTENTS
Executive Summary ........................................................................................................i
List of Figures ............................................................................................................. iii
List of tables .............................................................................................................. iii
1.0 Introduction ....................................................................................................... A
2.0 Project goal and objectives .................................................................................... A
2.1 Project Goal ........................................................................................................ A
2.2 Objective and Deliverables ...................................................................................... A
3.0 Study Area ......................................................................................................... A
4.0 Methodology ....................................................................................................... B
4.1 Data Collection..................................................................................................... C
4.2 Determining Sample Size ......................................................................................... C
4.3 Importing Data to ArcMap ........................................................................................ C
4.4 Data Exploration ................................................................................................... D
4.4.1 Histogram ...................................................................................................... D
4.4.2 Normal QQ Plot ............................................................................................... F
4.4.3 Trend Analysis ................................................................................................H
4.5 Model ................................................................................................................ J
4.5.1 Inverse Distance Weighted (IDW) ............................................................................. J
4.5.2 Kriging ............................................................................................................. L
5.0 Discussion of Results .............................................................................................O
5.1 Statistical Evaluation of Interpolation Techniques ..........................................................O
5.2 Similarties ........................................................................................................... P
5.3 Differences ......................................................................................................... R
6.0 Recommendations ................................................................................................ R
7.0 Conclusion ......................................................................................................... R
Bibliography ................................................................................................................ S
Appendix A- Study Area ............................................................. Error! Bookmark not defined.
Appendix B- IDW Interpolation of Static Level in Wainfleet .................. Error! Bookmark not defined.
Appendix 3- kriging Interpolation of Static Level in Wainfleet

ii | P a g e
Deliverable D4 (Major Assignment)

LIST OF FIGURES

Figure 1: Sample Location in Wainfleet area ........................................................................ B


Figure 2: Northing Coordinates histogram from ArcGIS ............................................................. E
Figure 3: Easting Coordinates histogram from ArcGIS .............................................................. E
Figure 4: Static Water Level histogram from ArcGIS ................................................................ F
Figure 5: Northing Coordinate Normal QQ plot ......................................................................G
Figure 6: Easting Coordinate Normal QQ plot ........................................................................G
Figure 7: Static Water Level Normal QQ plot ........................................................................H
Figure 8: Trend Analysis of Static Levels ............................................................................... I
Figure 9: IDW method Summary ........................................................................................ J
Figure 10: Power and search neighborhood .......................................................................... K
Figure 11: Kriging method summary ................................................................................... L
Figure 12: Semivariogram/Covariance Modeling .................................................................... M
Figure 13: Searching Neighborhood window .......................................................................... N
Figure 14: Cross validation Statistics for each interpolation ......................................................O
Figure 15 : Interval for Static Level in meters. ...................................................................... P
Figure 16: Kriging and IDW Surfaces Map .............................................................................Q

LIST OF TABLES
Table 1: Statistical information for static water level, easting, and northing. ................................ C

iii | P a g e
1.0 INTRODUCTION
Groundwater systems are often the dominant water resource available for municipal and agricultural
water supply. Moreover, during periods of dry weather, groundwater is a significant source of water to
sustain of the flow in streams and the water in lakes and wetlands. Ground-water systems are dynamic
and adjust continually to short-term and long-term changes in climate, ground-water withdrawal, and
land use. Water-level measurements from observation wells are the principal source of information about
the hydrologic stresses acting on aquifers and how these stresses affect ground-water recharge, storage,
and discharge (Charles and William, 2001). Water levels provide essential data needed to evaluate
changes in the resource over time, to develop ground-water models and forecast trends, and to design,
implement, and monitor the effectiveness of ground-water management and protection programs.

Therefore, the purpose of this project is to undertake two statistical analysis, kriging and Inverse
Distance Weighted (IDW). This analysis was conducted with the Well Data for the Region of Wainfleet,
Ontario. The Inverse Distance Weighted (IDW) methods is much simpler and takes less factors into
account when interpolating compared to the Simple Kriging method which requires more in-depth
analysis of sample data. A detailed explanation of data collection will be presented, followed by a
detailed outline of how each of the interpolation techniques were completed. Finally, both interpolation
techniques will be investigated on their level of accuracy and validity.

2.0 PROJECT GOAL AND OBJECTIVES

2.1 PROJECT GOAL

The goal of this project is to show the difference in Static Water Levels across Wainfleet, Ontario. This
information will help construction companies make better judgements when designing infrastructures in
Wainfleet, Ontario. Also, this project will give students a better understanding and knowledge in Spatial
Analysis Statistics, which is a course students are currently enrolled in the Geographic Information
Systems-Geospatial Management program at Niagara College Canada.

2.2 OBJECTIVE AND DELIV ERABLES

As mentioned in the Terms of Reference the following task listed below is the main goal of this project
(Smith, 2017):

To derive a working ability to report upon the collection of geospatial data and to describe the
data both geostatistically as well as practically.
To predict geospatial coverage in areas not directly measured/observed by interpolation

The project team at M.J.Z GIS Consulting will produce a report that outlines the data collection method,
methodology, and analysis techniques (kriging and Inverse Distance Weighted) used to accomplish these
goals, followed by a detailed analysis of the two interpolation techniques.

3.0 STUDY AREA


Deliverable D4 (Major Assignment)

The study area for this project is a small township South of Ontario named Wainfleet. The Township of
Wainfleet occupies an area of 53,143 acres (21 507 ha) in the southwestern part of the Regional
Municipality of Niagara (Ontario Geological Survey, 1985). The Static Water Level maps is prepared based
on the Static Water Levels observed in water wells completed at depth less than 25m and assumed all
well are under unconfined condition. Figure 1 below shows the study area that will be investigated for
this project. (Refer to Appendix A)

Figure 1: Sample Location in Wainfleet area

4.0 METHODOLOGY

B|Page
Deliverable D4 (Major Assignment)

4.1 DATA COLLECTION

The data was collected from the Government of Ontario website. Simple random sampling method was
used. In simple random sampling, each member of a population has an equal chance of being included in
the sample. Also, each combination of members of the population has an equal chance of composing the
sample.

Simple random sampling is the easiest method of sampling and it is the most commonly used. Advantages
of this technique are that it does not require any additional information. To draw a simple random sample
from a wells dataset, each entry has been numbered sequentially. There were 755 entries in the wells
dataset with Static Water Level in meters. 255 entries have been sampling randomly generated without
replacement by a software R using the following parameters:

The margin of error is 5%, the confidence level is 95%, the population size is 755 and the response
distribution is 50%.

4.2 DETERMINING SAMPLE SIZE

The project team collected 349 well data points on February 26 th, 2017. On the government of Ontario
website the region of Wainfleet, Ontario had 755 surveyed points on well data (which showed the Static
Water Level depth). Using the simple random sampling method well data sample size was reduced to
255 observations.. Form the random observation points the median, maximum, minimum and standard
deviation were calculated. Table 1 below shows calculations of basic statistical information for Northing,
Easting and Static Water Level.

Table 1: Statistical information for static water level, easting, and northing.

Static Water Easting (Meters) Northing (Meters)


Level (Meters)
Mean 4.00 633999.16 4750616.32

Median 3.70 635145.31 4748733.18

Standard 3.17 4283.33 4294.71


Deviation
Maximum 18.63 639810.41 4761685.47

Minimum 0 616624.96 4746132.46

4.3 IMPORTING DATA TO ARCMAP

C|Page
Deliverable D4 (Major Assignment)

The first step the before importing the excel data file into ArcMap was creating a file geodatabase in
ArcCatolog. This will help the technicians at M.J.Z GIS Consulting keep the data organized and
presentable to the client. The excel file was added into the new file geodatabase and exported as a new
shape file. The new shape file had a Longitude, Latitude and Static Water Level information. Using the
add XY coordinates the Longitude and latitude points were converted into Northing (X value) and Easting
(Y value) coordinates. Using the Mr.SID file from the X drive was also imported to the new file
geodatabase. Note all the data was projected into NAD 1983, UTM Zone 17N. Finally, M.J.Z GIS Consulting
requires all employees to create metadata to insure accountability. Metadata shows who created the
data, what files it contains, why was the data generated and how the data was generated. (Stanford
Libraries,2017)

4.4 DATA EXPLORATION

Once all the data has been added into the new file geodatabase the data was explored. Using the
Geostatistical Analyst tool the data was examined. This allows the Technicians at M.J.Z GIS Consulting
to understand and make better decisions when analysing that data. When exploring the data special
attention was payed to distribution, global trends, direction influence and so forth (Esri Exercise 2, 2017).
The following data was examined:

Histogram
Normal QQ Plot
Trend analysis

4.4.1 HISTOGRAM
Histogram for the Wainfleet, Ontario Static Water Levels, Northing and Easting Coordinate systems were
created using the Geostatistical Analyst Tool in ArcGIS. The number of bins (16) was used, which is the
square root of the sample size data (255). Histogram show a good representation of how frequent values
show up in a dataset. Figure 2, 3 and 4 below shows the histograms for the Northing Coordinates, Easting
Coordinates and Static Level. The Histogram for the Northing Coordinates has a negative skew, while the
Histogram for the Easting Coordinates has a positive skew and the Static Water Level Histogram has a
positive skew.

D|Page
Deliverable D4 (Major Assignment)

Histogram
Transformation: None
-1
Frequency 10 Count : 252 Skewness : 1.136
5.7 Min : 4746100 Kurtosis : 2.9166
Max : 4761700 1-st Quartile : 4747500
4.56 Mean : 4750600 Median : 4748800
Std. Dev. : 4313.5 3-rd Quartile : 4753100

3.42

2.28

1.14

0
4.75 4.75 4.75 4.75 4.75 4.75 4.76 4.76 4.76 4.76 4.76
-6
Dataset 10

Dataset : WainfleetSample_projected Attribute: POINT_Y

Figure 2: Northing Coordinates histogram from ArcGIS

Histogram
Transformation: None
-1
Frequency 10 Count : 252 Skewness : -1.3308
6.6 Min : 616620 Kurtosis : 4.5301
Max : 639810 1-st Quartile : 632530
5.28
Mean : 633980 Median : 635150
Std. Dev. : 4303.4 3-rd Quartile : 636800

3.96

2.64

1.32

0
6.17 6.19 6.21 6.24 6.26 6.28 6.31 6.33 6.35 6.37 6.4
-5
Dataset 10

Dataset : WainfleetSample_projected Attribute: POINT_X

Figure 3: Easting Coordinates histogram from ArcGIS

E|Page
Deliverable D4 (Major Assignment)

Histogram
Transformation: None
-1
Frequency 10 Count : 252 Skewness : 1.4771
5.4 Min :0 Kurtosis : 6.265
Max : 18.3 1-st Quartile : 2.1
4.32
Mean : 4.0355 Median : 3.7
Std. Dev. : 3.189 3-rd Quartile : 5.2

3.24

2.16

1.08

0 0.18 0.37 0.55 0.73 0.92 1.1 1.28 1.46 1.65 1.83
-1
Dataset 10
Dataset : WainfleetSample_projected Attribute: STATIC_LEV

Figure 4: Static Water Level histogram from ArcGIS

4.4.2 NORMAL QQ PLOT


Normal QQ Plots for the Wainfleet, Ontario Static Water Levels, Northing and Easting Coordinate systems
were created using the Geostatistical Analyst Tool in ArcGIS. Figure 5, 6 and 7 below shows the Normal
QQ Plot for the Northing Coordinates, Easting Coordinates and Static Level. The QQ plot for the Northing
Coordinates is not normally distributed, while the Histogram for the Easting Coordinates and Static Water
Level are moderate distributed. This means that Northing Coordinates are not evenly distributed along
our study area but Easting Coordinates and Static Water Levels have a moderate distribution along the
entire study area.

F|Page
Deliverable D4 (Major Assignment)

Normal QQPlot
Transformation: None
-6
Dataset 10
4.76

4.76

4.76

4.75

4.75

4.75
-2.88 -2.3 -1.73 -1.15 -0.58 0 0.58 1.15 1.73 2.3 2.88
Standard Normal Value

Dataset : WainfleetSample_projected Attribute: POINT_Y


Figure 5: Northing Coordinate Normal QQ plot

Normal QQPlot
Transformation: None
-5
Dataset 10
6.4

6.35

6.31

6.26

6.21

6.17
-2.88 -2.3 -1.73 -1.15 -0.58 0 0.58 1.15 1.73 2.3 2.88
Standard Normal Value

Dataset : WainfleetSample_projected Attribute: POINT_X


Figure 6: Easting Coordinate Normal QQ plot

G|Page
Deliverable D4 (Major Assignment)

Normal QQPlot
Transformation: None
-1
Dataset 10
1.83

1.46

1.1

0.73

0.37

0
-2.88 -2.3 -1.73 -1.15 -0.58 0 0.58 1.15 1.73 2.3 2.88
Standard Normal Value

Dataset : WainfleetSample_projected Attribute: STATIC_LEV


Figure 7: Static Water Level Normal QQ plot

4.4.3 TREND ANALYSIS


Trend analysis rend analysis often refers to techniques for extracting an underlying pattern of behavior
in a time series. (Wikipedia, 2017) Trends in the data can cause an interpolated surface to be unrealistic
(Esri Exercise 2, 2017). Due to the unique geographical topology of the Niagara Region, the study area
does not show any trends in particular relating to Static Water Levels (Refer to Figure 8). The area in the
North-South is gradually increasing. Also, large Static Water Level changes occur over a short distance.

H|Page
Deliverable D4 (Major Assignment)

Trend Analysis

Dataset : WainfleetSample_projected Attribute: STATIC_LEV

Figure 8: Trend Analysis of Static Levels

I|Page
Deliverable D4 (Major Assignment)

4.5 MODEL

4.5.1 INVERSE DISTANCE WEIGHTED (IDW)

With Geostatistical Extension loaded, Inverse Distance Weighted (IDW) was selected to create the first
prediction surface in ArcMap. Uses the measured values surrounding the prediction location to predict
a value, where the location are not sampled, based on the assumption that things that are close to one
another are more alike than those that are farther apart (ArcGIS Help, ESRI). In Geostatistical Wizard,
IDW was launched while the Wainfleet well sample being selected as the source dataset and STATIC_LEV
being selected as data field. A summary of the IDW model used in this projected was listed in Figure 9
below.

Figure 9: IDW method Summary

J|Page
Deliverable D4 (Major Assignment)

As Figure 10 shows, the power was set to be the default value 2 because the static level should be way
more depend on closer points rather than distant points. The maximum search neighbourhood was 15,
roughly 5% of the sample size and the minimum was 10, roughly 4%. This means that at most 15 points
around a certain location will be used to predict its value while at least it was 10. It was assumed that
there was no directional influence in well static level so the sector type was set to be 1 sector which will
make points in different direction count equally.

The map of IDW interpolation can be found in Appendix B.

Figure 10: Power and search neighborhood

K|Page
Deliverable D4 (Major Assignment)

4.5.2 KRIGING

Simple kriging was also conducted in the Geostatistical tool as it would predict a smoother surface and
work efficiently. Kriging assumes that at least some of the spatial variation observed in natural
phenomena can be modeled by random processes with spatial autocorrelation, and require that the
spatial autocorrelation be explicitly modeled (ArcGIS Help, ESRI).

With well sample as data source and STATIC_LEV as the data field, simple kriging was conducted through
Geostatistical Wizard and a summary of the kriging method was listed in figure 11 below.

Figure 11: Kriging method summary

As discussed above, the dataset was considered to have no trend so the order of trend removal was set
to be none so as the transformation type.

Due to the fact that the static level is a continuous surface and the semivariogram model intercepts the
y-axis at around 0.5, chances are there was nugget effect inside this dataset. In order to get rid of the

L|Page
Deliverable D4 (Major Assignment)

nugget effect, in Semivariogram/Covariance Modeling, the Model Nugget was enabled and the Calculate
Nugget was set to true. All other parameter were left as default (figure 12).

Figure 12: Semivariogram/Covariance Modeling

In Searching Neighborhood window, the sector type, the maximum neighbors as well as minimum
neighbor was altered to be the same value as IDW method in order to make contrast to the IDW. All
other parameters was left as default as ArcMap automatically calculated these (Figure 13).

M|Page
Deliverable D4 (Major Assignment)

Figure 13: Searching Neighborhood window

The map of Kriging interpolation can be found in Appendix C.

N|Page
Deliverable D4 (Major Assignment)

5.0 DISCUSSION OF RESULTS

5.1 STATISTICAL EVALUATI ON OF INTERPOLATION TECHNIQUES

To evaluate the results of both surfaces, the cross validation comparison dialog box was launched, which
can be seen in Figure 14 below. The objective of cross-validation is to help you make an informed
decision about which model provides the most accurate predictions. It gives you an idea of how well the
model predicts the unknown values. (Esri Exercise 3, 2017)

Figure 14: Cross validation Statistics for each interpolation

The results shows both method produced a mean prediction error close to 0, which mean the results are
unbiased. The root mean square error for both surfaces are close to 1 which indicated the standard errors
are accurate. This tests and validates the accuracy of both interpolation technique to predict a
reasonable Static Water Level in Wainfleet, Ontario (Esri Exercise 3, 2017). Finally, based on the
interpolation of both methods, the results are very similar to one another.

O|Page
Deliverable D4 (Major Assignment)

5.2 SIMILARTIES

Both the Kriging and Inverse Distance Weighted produced surfaces which separated Static Water Level
into categorize in Wainfleet, Ontario. Figure 9 below shows how Static Level are classified.

Figure 15 : Interval for Static Level in meters.

Both the Kriging and the IDW surfaces show moderate similarities in Static Level in the North-East section
of the graph. Figure 11 on the follow page shows the generated make (Please refer to Appendix B for full
extent layout). Also, in Burnaby which is located in the south next to Lake Erie show an increase in Static
Levels. Both surfaces have data holes in the North and North-West and East due to the lack of evenly
distributed point in the sample data. Both IDW and Kriging Map can be found in Appendix B.

P|Page
Deliverable D4 (Major Assignment)

Figure 16: Kriging and IDW Surfaces Map

Q|Page
Deliverable D4 (Major Assignment)

5.3 DIFFERENCES

When analyzing both surfaces the IDW interpolation produced a more detailed Static Level. The IDW
results show a more graduated transition between different Static Level. Form Figure 11 the difference
between the two surfaces can clearly be seen in the center of Wainfleet, Ontario. In the IDW
interpolation Static Level between the intervals 0 to 2 meters, while the Kriging interpolation just shows
majority of Static Level at 2 to 4 meters.

6.0 RECOMMENDATIONS
After analysing both interpolation methods, visually the IDW interpolation methods showed a more
accurate and detailed Static Level in Wainfleet, Ontario. However checking the RMS Error, IDW showed
larger value comparing with Kriging.

The issue encountered when completing the project, was the lack of adequate data coverage in certain
areas. Major data gaps create areas where it is very difficult for software to interpolate a surface due to
a lack of reference points. Even smaller data gaps can pose large problems if they occur in areas of high
variability.

Therefore, use of all the water well data with more static water level measurements would provide a far
more accurate result for the study area.

7.0 CONCLUSION
The Geostatistical Analysis of Student Collected Spatial Data project successfully was completed with
the completion of two interpolated surfaces. Water well data, with static water level measurements,
was obtained for the Town of Wainleet, and was used as the basis for interpolating each surface.

Two surface interpolation techniques were used, the Inverse Distance Weighted, and the Empirical
Bayesian Kriging. The IDW technique was used to create a much smoother surface with graduated
transition, which gave a good overall estimate of static water level in the area, whereas the Kriging
technique was used to create a much stricter surface.

Both interpolation techniques produced surfaces that achieved the original goals and desires of the client
and project. Limited errors occurred throughout the project; however, overall data coverage in the area
could be improved which in turn would improve the interpolated surfaces.

R|Page
Deliverable D4 (Major Assignment)

BIBLIOGRAPHY

ArcGIS Desktop. (2016). Tutorial, Introduction to the ArcGIS Geostatistical


Analyst Tutorial. Exercise 1 to 5, Esri ArcMap 10.4.
ArcGIS Help,2017) How Inverse and kringing woeks. ESRI Help desktop.
Charles J. Taylor, William M. Alley, 2001- Ground-Water-Level Monitoring and
the Importance of Long-Term Water-Level Data, U.S. Geological Survey Circular
1217.
Hering,J (2014)- A virtual flood of information: open data for sustainable water
management
Smith, I. (2017) GISC9318 D4 Introduction to Statistics Geostatistical Analysis
Smith, I. (2017) GISC9318, Lecture 4: Spatial Analyst. Niagara College. PP
Smith, I. (2017) GISC9318, Lecture 7: Regression and Interpolation. Niagara
College. PPT
Smith, I. (2017) GISC9318, Lecture 8: Geostatistical Analyst. Niagara College.
PPT
Stanford Libraried. (2017). Internet Retrieved from
http://libary.stanfors.edu\reserach\data-management-services/data-best-
practices/creating-metadata
William D. Banks, 2005-NPCA Groundwater Study, Niagra Penisula OCndevation
Authority, Waterloo Hydrogeologic, Inc.

S|Page

Das könnte Ihnen auch gefallen