Results and Discussion

Analysis and Conclusions

Indicators
   Recommendations
   Summary of Concerns
   Suggestions for Further Work

A number of interesting images and datasets were produced through the course of this study. As noted earlier, these images not only facilitated analysis, but also have great educational and inspirational value. For example, the orthographic representations of Vancouver's former drainage system can be most inspiring and informative. Many Vancouverites simply do not know that their street was built overtop of a stream. Only recently did I learn that the street I live on (and have lived on for 22 years) was, in fact, built overtop of a stream. It was partly this realization (which I had several years ago) that inpired me initially to think about urban hydrology and urban stream restoration.

    Simply looking at a map showing lost streams can be a liberating experience, and relates well to the bioregional movement. Streams, after all, know no boundaries, and looking at a map like the one below helps us to think about space and natural processes inspace in new ways.
    In our daily lives, surrounded by built environments, we often forget where all of it came from. Looking at an image like the one below is a bold statement for some; it reminds us that the city was already "green" before we arrived in great numbers! When planners and policy-makers speak of "greening the city", we are reminded by images like this that there exists a "wild" landscape still, hidden under the asphalt and neatly mowed front lawns of modern-day Vancouver.
    Perhaps efforts to "green the city" should instead begin by speaking about "reclaiming" formerly natural spaces, and allowing ecological processes to continue within urban areas.

The MCE images produced in the preceeding section were combined to produce the "suit" image below, which shows the overall areal suitability for urban stream restoration. Note that the numbers are shown in real data here. The results were in byte format, but were converted for the purposes of this display (to create a continuous legend). Note also that these values are in unassigned units (interval data) and therefore should be understood on a relative basis only.

"Suit" was then overlaid with the image below, "loststreams":

... to get the following image of streams and their suitabilities: "canudigit":

The image below, suitmap, shows the product of this analysis. Which streams are more suitable for daylighting? Given all of the constraints listed in previous sections, how should we select which stream to daylight? The map below is a useful tool:

Below is a histogram showing the distribution of suitability indicies along each of these streams.

Note that the histogram shows a fairly normal distribution of suitability. We see some extreme values, but we should be very careful at reading "too far" into this data, as it was derived from other maps at lower resolutions. That would be akin to including seven decimal places when only two "significant figures" are called for. The general trend here is encouraging, we see a great many sections of stream that are moderately suitable.
In looking at the streams as a whole, we see that those streams currently within parks are deemed the most suitable in general. This is not simply because parks were weighted in an MCE, but because it likely also coincides with several othr factors including land use, slope, proxmity to bikeway, and many others.

Indicators of Success:

A good indicator of the sucess of this study is that there are currently daylighting projects in planning for Brewery Creek (Mount Pleasant) and a creek in Kitsilano - both of which rate highly on the above suitability index.
Also it should be noted that BC Hydro is currently sponsoring a daylighting of a stream in Stanley Park - a stream that was rated extremely suitable in the above analysis.

Recommendations:

It is therefore the recommendation of this study that two additional creeks be daylighted - the ones which show the next highest suitability values on the above map. One of these would be a stream in Everett Crowley Park near the southeast corner of the city. This would be ideal, as this is a recently dedicated park. parts of it, interestingly, were previously used the municipal landfill for Vancouver until the early 1960s. The non-landfill areas festure lush hemlock and red alder forests, and the landfilled areas are fields, berry bushes, and low-lying vine maple and red alder.
The second stream that should be daylighted can be selected here, however a more definitive descision should wait for an analysis which includes more relevant data concrning rainfall. The second recommendation would be an additional stream in South Vancouver - however this result may be unduly influenced by the aspect data, which was only included as an imperfect surrogate for rainfall. This second recommendation, therefore is weak and should not be heavily relied upon.

Summary of Concerns:

These concerns have been metioned thusfar in the text, but will be summarized here for convenience:

as with all GIS projects, the fundamental problem if dealing with continuously variable data was encountered. (see glossary). Information is simplified (lost) for analysis; a boolean image is an often-necessary but poor substitute.
again, as with all GIS projects, we encountered issues related to the propagation of error. For example, the streams dataset was derived by digitizing a map that was produced by DFO, and even this was subsequently rasterized, representing at least three "generations" of error. The streams dataset was particularly bad for this, however the problem was encountered with other datasets.
surrogate data was used in some cases where more suitable data was unavilable. For example, aspect was used as a surrogate for rainfall, on the understanding that - ceteris paribus - south-facing slopes receive more rainfall, given the prevailing winds in this area. This was dealt with by weighing aspect more lightly in the relevant MCE.
The selection of a particular stream (the final step) can be done with only a relative degree of certainty given the analysis produced. This, however, reflects the nature of the problem (no pun intended), as there may be multiple streams which are suitable for daylighting. This reflects the general difficulty in quantifying this question.
There was a lot of good data on hand, however much of it was foreign to Idrisi, and crashed upon being imported, through any route. This severely limited the scope of the analysis. (see suggestions for further work)
The basic DEM that was used was limited to a 25 metre resolution
The basic DEM also had a "groove" that could not be adequately compensated for, depite the spplication of pit removal techniques.
Some digitizing was necessary, however in all cases, guides (eg. of higher precision) were applied, (as in the case of the 2m_contour.dxf file)

Suggestions for Further Work

An original goal of this analysis was, after selecting a suitable stream, to use the block outlines data (which are polygons at a very high resolution, showing blocks, alleys, lot lines, and so on) to zero in on the individual stream and come to conclusions regarding the actual site, down to the city block level, and to identify which lots would have to be acquired, analyse the slope along the stream path, work out sufficient riparian and habitat zones, and compensate for adjacent land uses. The block outlines data was so incredibly overjoyed at its importation into Idrist, that it crashed the computer every time I did so. Unfortunate.
The application of socioeconomic data would be a nice addition.
An interesting study would be to take an animal - say a type of waterfowl - and look at its distribution along recently daylighted streams. There are countless databases on riparian habitat, but very few databases on such habitats for daylighted streams.
On an unrelated note, I am still dangerously obsessed with the idea of producing a flexible, queryable DEM for cyclists in Vancouver. For this to be web-published and widely available would be quite desirable for the cycling community. During a stage of this project, I went so far as to digitize my house and the downtown library (a favourite cycling destination of mine) and develop a least-cost path. It was interesting.