Methods

Soil Data

First Step

•  Soil data clipped so that only the prairie ecozone remained.
• Erased all of the parks and lakes from the prairie ecozone layer.
• Buffered major Manitoba cities by 1km and then erased the buffered layer from the remaining soil layer.
• From that layer only well-drained soils were selected from the data by using select by attribute and then making a new layer from the selection.

 Second Step- After final vector landcover layer completed

• The soil layer was then clipped to it
• Soil vector layer was then converted to raster using the feature to raster conversion tool.

Landcover Data

• All NTS tiles were merged together into one layer.
• The merged layer was then clipped by the soil vector layer made in the first step so that lakes, parks etc..., and badly drained soils were clipped out
• A layer including only the five most preferable landcover coverage types was then made through selecting by attribute and then making a new layer.

Coverage Types Selected:

• The final vector landuse layer was then converted to raster using the feature to raster conversion tool.

Canadian Digital Elevation Data

• Each CDED NTS mapsheets were converted from elevation USGS format to elevation raster data using the DEM to raster tool.
• All of the converted tiles were then merged together using the mosaic to new raster tool.
• The new mosaic raster was then projected.
• That raster was then clipped through extraction by mask using the final landcover vector layer.
• Slope and aspect were then calculated from the raster using the respective spatial analyst tools.

Weighted sum analysis

Factors to be used in the MCE model

• Slope
• Aspect
• Landcover
• Soil type

 Slope: was normalized using the spatial analyst tool, raster calculator. The expression entered was: 1- ((“slope”-0)/ 56.1)

 Aspect: was normalized using the fuzzy membership tool (Gaussian membership type)

 Landcover: since landcover was nominal uncontinuous data its normalization was subjective. A new column in the final landcover raster layer was added and a value specified for each landcover type. The most preferred landcover type to be converted to switchgrass crops was the perennial cropland and it was assigned a value of 1, next was native grassland (0.9), cultivated agricultural land (0.8), herbaceous land (0.7), shrub land (0.6).

 Soil: soil type was also nominal uncontinuous data so a new column was also added to the soil rastser. The most preferred soil type, black chernozem was assigned a value of 1, the second most preferred soil type, dark grey chernozem was assigned a value of 0.8, and least preferred soil type, regosolic, was assigned a value of 0.6.

2. Weighting each Factor

The weights of each factor were determined using an analytical hierarchical approach online at cci-icc.gc.ca/tools/ahp/index_e.asp. The line-by-line method produced values of 6% for both aspect and landcover, 24% for soil and 64% for slope. This study considered slope to be the most important factor, followed by soil, and then aspect and landcover were considered to be the least important.

3. Weighted Sum Calculation

The weighted sum was calculated through the weighted sum tool using the determined weights for each factor.

The weighted sums of the factors were then calculated once more but instead with all factors being considered equally weighted.