Main
Index| Methodological and Operational Problems | Main
Index |
Data Collection Problems:
One of the major problems I came across when collecting my data was that most of the data from the s:\\drive had no metadata provided with it to interpret the meaning of the legends. When I visited the VanCORE website where this data was derived from, the website provided minimal additional data, such as reference coordinate type and data units. However, most of the information provided at this website was potentially useless and irrelevant with regards to my project analysis. Consequently, some of the raster and vector layers provided could not be understood so I chose not to use them in my analysis. But who knows, they might have proved to be very useful had they only provided metadata to reveal what information the map was displaying.
My original intent was not to use data from the s:\\drive, however after countless hours searching the internet I could not find any additional raster or vector layers on Vancouver Island to download for my project that were free or relevant for answering my question of "where the safest places to live on Vancouver Island are in the event of an Earthquake". Luckily I was able to find information from BC Stats, on population densities of major cities and earthquake occurrences, to enter into the database workshop by hand. Since I had no data layer showing the locations of the major cities that I included in my analysis, I had to digitize these points in by hand onto a vector layer. The trouble I had was that the scale of my Vancouver Island Map and my raster layer for allroads were not the same. Consequently, some of my digitized points did not fall directly on a road which doesn't really make sense considering they are all fairly large cities. I had to make slight alterations in the locations of some of my points by re-digitizing them onto the roads later on so that my analysis would work out.
Analysis Problems:
One problem, with regards to the spatial analysis of my data, that I came
across several times was related to the number of columns and rows of pixels
in the raster image. Some of the raster layers were provided with
values already, such as 415 columns and 311 rows, yet other layers did
not provide values. The mistake I made was that for these layers
with no column or row values provided I chose an arbritary number of 400
for both columns and rows. This proved to cause me a great deal of
agony and extra work later on in the analysis when I tried to overlay layers.
I was not able to overlay one of the raster layers with 415 columns and
311 rows with a raster layer with 400 columns and rows which I created.
I soon discovered that in order to perform an overlay both layers must
have the same number of columns and rows. Consequently I had to go
back to the initial raster layer I created, change the number of columns
and rows to match those with the provided values, and reclass, re-assign
and in some cases re-buffer the images to return back to the stage I was
at with the overlay. Needless to say, I created a lot of extra work
for myself.
A second problem with my analysis involved buffering the "allroads" image.
Originally I had decided that the safest areas would be within 500 metres
of a major road for quick evacuation during an earthquake. However,
after creating this buffer I realized the distance was too small and the
buffer did not even register on the image. I knew that using this
buffer value would produce an image with no suitable locations to live
if I chose to perform a MCE, so as a result I chose a larger buffering
value of 1500 metres so that it could be seen on the image. I knew
at this point that my final image would not display what I had originally
intended for it to because ideally one would want to live closer than 1500
metres to a road in the event of an earthquake, however this increase in
buffer size was a necessary alteration.
Conclusion:
Overall the project went pretty smoothly with just a few bumps along the way. I realize however, that my final results are very subjective and that any other individual researching the same issue might choose different distance values and different criteria to produce their map. This is the way it is for any type of analysis, results will completely vary from one individual to another, and not necessarily anyone's results may be wrong.
All in all this project created a successful outcome of the safest places
to live on Vancouver Island in the event of an earthquake. Had I
been able to find information on other variables, such as the different
types of substrate underlying Vancouver Island (ie. glacial till - may
collapse in an earthquake, or bedrock-will withstand shaking better), my
results may have been a little more precise. However, as I said,
I plan to move to Vancouver Island in the future and when I do so, I will
take into account the results I obtained from this project so as to find
a safe place to live if the "Big One" should ever strike, which inevitably
it will.
