Data Issues
A large part of the analysis is based on Environment Canada and NAPS data. The main problem that lies with using their data is that the collection stations are too sporadic over the region, especially for aerosol data. With only eight aerosol stations collecting usable data over such a large region, it is simply impossible to have any accurate analysis. The second problem in association with data is that their archives are not available for many stations even when they do collect the data. For example, the GVRD has close to 30 aerosol stations that collect data, but archives are only available for 16 of them of which only 8 contain suitable data for this analysis. Therefore, the standard of data collection and archiving varies within the stations for aerosol data, have made it hard to have a good coverage for the entire GVRD region.
GIS Issues
GIS is a common tool for modern day geographic and environmental
science modeling of spatial phenomenon. It is a very powerful tool as
it allows the user to visualize and analyze real life spatial patterns
of various topics. Despite its potential robustness, GIS has its limitations,
some of which are recognized on the course of this aerosol and precipitation
study.
The digitized points in Map 1 represent data collection points. Interpolations in Maps 3 and 4 spreads out as equal-intervals from one point to another. The program treats each data point as a maximum value in its proximity and it spreads out as decreasing values from the point. In reality, the data points are the stations. Since many of these stations are located in open areas it might not be a maximum value in its proximity. When there is a single large point source emitting huge amounts of aerosols that are not close to the data collection stations, the decreasing trend from the collection point cannot represent the large point source in the real world. Similarly, if collection points are very far apart and interpolation is made between them, the overall regional concentration will be biased towards the data collection points. An example is demonstrated on the 2001 aerosol distribution map. In Richmond, the concentration of aerosol is represented as one of the highest in the GVRD. However, many Richmond residents find their living area has generally clean air. The contrasting opinion between the GIS interpolation and the Richmond residents is caused by the location of the aerosol collecting station, which is very close to the heavy traffic flow Hwy-99. Actually the data point in Richmond is increased drastically by its close proximity to the highway. Coupled with the fact that there are only 8 aerosol data collection points in all of GVRD, the interpolation is therefore biased due to the high aerosol concentration of Hwy-99, making the whole of Richmond seem to have a dirty atmosphere. In this case, the opinion of the residences of Richmond is right and the GIS interpolation is incorrect.
One way to overcome this issue by GIS users is to do individual location observations for the digitized points and parameterize correction values for them to either increase or decrease their values according to the situation. However, this cannot always be done due to time constraints and costs associated with traveling to all the sites for observations. Another method of overcoming interpolation errors is to incorporate spatial modeling into the analysis. This can be done when a knowledgeable modeller incorporates special algorithms to mimic the real environment.
Another example of interpolation inaccuracy can be found in the precipitation distribution map of 2001. The map shows that the northern mountains across from Indian Arm and north of Coquitlam don’t have extremely high precipitation in comparison to the mountains in the northern part of North Vancouver. This is untrue as many mountain climbers and hikers realize that these mountains north of central Coquitlam usually have the highest precipitation in the entire GVRD region. The reason for this inaccurate representation on the map is that there is no data station up in these northern mountains, causing the interpolation to naturally decrease the precipitation values over those areas.
By looking at the previously mentioned limitations of GIS, a map reader should be alert of the believability of interpolations done on the map he or she is reading. GIS limitations inherently ties with data availability issues. With more data points, a more accurate model of the world can be represented on the map. An important point to keep in mind when using any GIS map is that no matter how many data points there are, the map is not going to be a representation of the real world because the real world is a continuous space and it would require infinite data points to be accurate. Hence, striving for more data points in doing an analysis is very important, but at the same time, the project planner has to take time and cost constraints into mind as a tradeoff.