Data Modelling

Scope
Produce a landuse management scheme for the Kanata-West Carleton area of Ottawa-Carleton, Ontario, in a GIS environment. This proposal was produced in accordance to "Basic Modelling" specifications given by the client.

Data Inputs

Region of Interest	Kanata-West Carleton, Ottawa-Carleton, Ontario, Canada
Projection	UTM, zone 18, NAD 27, units decimal degrees
Vector Coverages	31f_ag Area 31F, Agriculture
	31f_for Area 31F, Forest
	31f_lu Area 31F, Land Use
	31f_wa Area 31F, Water
	31g_ag Area 31G, Agriculture
	31g_for Area 31G, Forest
	31g_lu Area 31G, Land Use
	31g_wa Area 31G, Water
	aoi Area of Interest
	roads Roads
	streams Streams
	water Water
	waterways Waterways
Documentation -Canada Land Inventory (CLI)	ag_read.txt Class information for Agricultural Coverages
	for_read.txt Class information for Forest Coverages
	lu_read.txt Class information for Land Use Coverages
	wat_read.txt Water

Pre-Processing

Initial Data Formatting

The initial pre-processing phase consisted of merging and uniting the agriculture, land use, and forest themes together to accurately define the study area, by eliminating border separations.

Various streams, waterways and watersheds were eliminated as they were not applicable to the study area due non-relevance to the study, extending beyond study area borders, etc.

Hardwood Creek and Constance Creek were kept in the study area, containing applicable class data from the forest coverage (see Analysis).

Land use values contained within the water theme were eliminated, as the land was not suitable for productive land use. This was done by making separate views for agriculture, land use and forest, then uniting the water to the land type depending on the view and removing water from the new united table.

This removed the water and the land underneath it. To get water back on the view, the water theme was re-added from the first view. The water theme included land in it and as a result the land was deleted from the table, making it single symbolized. We then had the separate land use themes and water theme on each view. To keep tables simple, data not referring to Class A was removed.

Proposed Classification Scheme

The datasets represented and measured the land value relative to the thematic use (land use, agriculture, forest). The documentation was also examined in order to define the landuse management framework and mapping system.

The need to reclassify was based on simplifying data for the landuse schemes. Key-value pairs were assigned for each landuse feature coverage area. The reclassified data was derived from grid conversions and reclassifying based on numerical code.

As a result, a three-tiered class structure was constructed uniform to the three landuse feature coverages for effective, scalable spatial analysis:

Table 1: Three-Tiered Class structure

Value	Class
1	Good
2	Fair
3	Poor

Reclass Code Rationale

Agriculture: versatility and functionality. Original land use classes 1-4 were classified as value 1 (no limitations/severe limitations). These classes have soils that can be considered capable for sustaining cultivated field crops. Classes 5 and 6 were reclassified as value 2. These classes can only sustain perennial forage crops. Therefore the land has many limitations and needs much improvement in order to have land sustain any type of crop. Class 7 and Organic was reclassified as poor, as there is no capability of growing any crops on this land.

Forestry: volume and functionality. Original land use classes 1-3 were classified as good (1). These classes have productivity greater than 71 cubic feet/acre/year. These classes have only slight limitation affecting growth in a slight to moderate fashion. Classes 4-6 were reclassified as fair. These classes have a productivity of between 11 and 70 cubic feet/acre/year, and have severe growth limitations. Class, 7 was reclassified as poor. This class has limitations that prohibit the growth of any commercial forests.

Land Use: Assorted. The following were classified as good - land use, horticulture, orchards, vineyards, swamps, marsh and bog. These cases were chosen, as our firm was optimistic that the soils can grow crops of value. Urban built-up area, cropland, improved pasture and forage crops, unimproved pasture and rangeland and productive woodland were classified as fair. These cases were chosen, as there was the possibility of growth, although the soil was of fair conditions. In the cropland, tobacco and potatoes can be grown; the soil required can be of high sand content, which is of little of interest for growth of vegetables. Mines quarries, sand and gravel pits, outdoor recreation, non-productive woodland, and unproductive woodland (sand) and unproductive land (rock) were classified as poor. These cases have little value for growth, as there is a high content of sand and gravel in the soil, as well as being non productive for growth.

Alternative Classification Scenario

Since the boundaries for the classification categories are important, it would be essential to include an alternate scenario of our original data re-classification. This alternative provided another possibility for re-classification and placed less limitation on our project. This gives the reader an opportunity to decide on what classification theme (s)he believed to be the more applicable scenario.

The second scenario, (shifting up values assigning fewer land types as good, and one more as "bad", in both the agriculture and forestry and adding two to the "good" in land use), provides an alternative to the proposed classification schema.

Agriculture: good was classified as classes 1-3, where as the proposed is classes 1-4. Classes 4-5 were reclassified as fair. The proposed scenario included classes 5 and 6. The classification of poor entails classes 6, 7 and organic, whereas the proposed only had class 7 and organic. This shift weighs the soil according to the limitations. In the good classification, the limitations are moderate, in the fair classification, the limitations are severe and in poor, the limitations are not feasible.

Forestry: classes 1 and 2 considered good, as opposed to 1-3 in the proposed scenario. The classification fair now entailed classes 4,5. In the proposed scenario, classes 4, 5 and 6 were all rated as fair. The classification poor now entailed classes 6,7. The proposed scenario rated only class 7 as poor. This reclassification now bases limitations are slight, moderate and severe. The good classification has soil with only slight limitations, the fair classification has moderate limitations and the poor soil has very severe to precluding limitations for growth.

Land use: two classes that were previously ranked as fair are now reclassified as good. These classes are cropland and productive woodland. In the proposed scenario, these two classes were not seen as having good soils, although they can still grow productive crops. Therefore in the alternative scenario these two classes were reclassified based on growth of crops and not on soil types.

Exploration

Following the pre-processing phase, the data was run through various routines to examine relationships of the phenomena in detail. Routines used were creating histograms, cross-referencing the landuse feature coverages.

Histograms 1-3 (see Appendix) illustrates landuse patterns within the proposed classification classes of agriculture, forest and landuse zone classes:

Table 1 (see Appendix) shows the zonal statistics cross-referenced for each landuse theme.

What does it Mean?

The following patterns were derived from the histograms and zonal statistics:

Landuse: Evident that areas of fair landuse capabilities contain good to fair agriculture and forest classes.

Agriculture: Good landuse classes exist in most of these zones.

Forest: Most of the agriculture and landuse classes fall within the good-fair zones.

Analysis

Constance Creek and Hardwood Creek have portions that are covered by class 7 land. The following analysis illustrates areas of class 7 land in both sub-basins.

Vector Results

The files were queried for each sub-basin and land7 class and saved to shapefiles for vector analysis. Then both sub-basin themes were clipped against the land7 theme. The calcapl.ave script was run to get the area giving the following results:

Raster Results The files were queried for each sub-basin and land7 class and saved to grids for grid analysis. The following tables illustrate areas and differences for each sub-basin covered by class 7 land, in m2.

As a result, if the cell size is decreased (i.e. increased cell information), the value increases. It is evident that during the grid conversion cell information detail decreases/becomes more generalized with increased cell size. There is a trade-off between cellsize and filesize/CPU processing. This discrepancy is decreasing with the result of cheaper computer hardware/disk costs and faster CPU speeds for cheap.

Vector/Raster Differences

It is evident that as cell size becomes finer, the area of difference decreases between raster and vector GIS environments. However, there is trade-off between clipping and running scripts in vector as opposed to the more flexible, user-friendly raster analysis environment, with these functions built into the raster GUI.

Proposal

The reclassified areas were extracted against a bounding geographic box of left: 416000, right: 435000, bottom: 5010000, top: 5028000. In observing the re-classified grids for this area of interest, poor parcels of land would be the best areas to develop urban growth, as they are of little agriculture or habitat value. Therefore, this scheme will preserve good habitat and agriculture land.

Good agricultural land can be defined as land that is high in versatility and functionality. It is able to sustain crops and produce large yields. Good habitat land can be defined as a high volume forest area that has been left undisturbed by industry and urban growth.

In the alternative scenario, the definition of good is limited in comparison to the proposed scenario. This results in more cases in the fair and poor classifications for the alternative scenario. When the classification scheme changes, there are less good agriculture and habitat lands. Because of the classification change it becomes even more important to make sure that urban growth occurs on the poorer land. This type of development will have minimal impact on lands that can be used for intensive agriculture and thriving habitat areas. If the classification of land goes down, the value of that land will decrease. If land values are less, housing costs will be lower and more urban growth could occur. This could easily put a strain on the surrounding lands that are good for habitat and agriculture.

What is now the urban area was once farmland and forested area. (refer to map). It would have been more environmentally and economically sound if the existing urban area was developed northwest of its current location. This means that urban development would have taken place on the poorer parcels of land. If urban development continues in the direction that it is going, habitat and agricultural lands will be sacrificed and therefore urban development should occur northwest of the good habitat and agricultural areas (refer to map). There are also poor parcels of land in the southwest area of West-Carleton that would be ideal for urban growth, but unfortunately this area is smaller in area, which would push urban growth into agricultural and forested areas.

The formula below illustrates the various combinations after calculating all combinations between the three landuse coverages. These values are valuable when assigning weightings to specific criteria in visualizing potential areas. For example, if a user wanted to find areas to build with a strong emphasis on low agricultural land values, the weighting formula could be formulated as follows:

outGrid = ((agri * 80) + (forest * 10) + (landuse * 10))

The resulting grid would show the best areas for this request, defined by the lowest cell value.

List of Maps / Histograms / Statistics

The following is a list of the layouts used in the report as well as the significance for this report.

Histograms 1,2,3 display values within zones of the three landuse themes.

Matrix 1 displays zonal statistics of the three landuse themes cross-referenced to one another.

Layouts 1,2,3 give an overview of agriculture, forest and land use, for the area of interest of Kanata-West Carleton with the proposed classification scheme.

Layout 4 shows the Land Use Management Scheme. This layout shows the area of interest with the previous three layouts overlapping each other. This gives each section a reference number to refer to the table, Attributes of Map Calculation 1, in order to understand the rating. An example is number 1 signifies that the area is of fair agriculture use, fair forest use and fair land use, as all three have ratings of 2. The client can then decide which land would best suit his/her needs for future purposes based on a weighted analysis.

Layouts 5,6,7 give an overview of the three landuse themes with the alternative reclassification. These layouts show an alternative scenario of how the land use may change if the classification is changed.

Layouts 8,9 show the area of interest (co-ords previously mentioned) for each of agriculture and forest. This allows for the client to view separate cases of agriculture and forest for each area.

Layout 10 is an overlay of the previous two maps to display areas of high agriculture and habitat value and areas of poor interest, relating to the proposal section.