Raster Modeling

Goal 

The goal of this assignment was for us to use various rater geoprocessing tools to build models for both a sand mining suitability, as well as sand mining impact in the form of environmental and cultural risk in Trempealeau County, WI.

Objectives

There are different objectives listed for each of the models that we needed to build. The first model was the suitability model for mining. The objectives are listed below:

1. Generate a spatial data layer to meet geologic criteria
2. Generate a spatial data layer to meet land use land cover criteria
3. Generate a spatial data layer to meet distance to railroads criteria
4. Generate a spatial data layer to meet the slope criteria
5. Generate a spatial data layer to meet the water - table depth criteria
6. Combine the five criteria into a suitability index model
7. Exclude the non - suitable land cover types

The objectives for creating a model representing the risk for mining are listed below:

1. Generate a spatial data layer to measure impact to streams
2. Generate a spatial data layer to measure impact to prime farmland
3. Generate a spatial data layer to measure impact to residential or populated areas
4. Generate a spatial data layer to measure impact to schools
5. Generate a spatial data layer to measure impact on one variable of your choice 
6. Combine the factors into a risk model
7. Examine the results in proximity to prime recreational areas

Methods

Suitability Model:

To create the suitability model there are five variables that we will be basing the model off of, these variables are listed below:

1. Geology
2. Land Cover
3. Water Table Height
4. DEM (Slope)
5. Rail Terminals

Each of these variables are going to be ranked from values of 1 - 3, 3 being the most suitable and 1 being the least. We used this ranking system because it will best represent what will work the best for frac sand mines. Below is a table that is explaining each of the classifications and rankings of each of the different variables.

(Fig. 1) Suitability Table

1. Geology - Ew/Ej are the most suitable areas for a frac sand mine to find resources which is why those areas are ranked the highest, while the other areas that are ranked the lowest are the least suitable areas for a frac sand mine to find resources. In order to properly rank each of the different types of geology, I had to reclassify the geology feature, where the Ew/Ej formations were in one category which was ranked 3, and all of the other formations were in the category ranked 1. 

2. Land Cover - We needed to go to the NLCD website to find data on the different types of land that are located within the study area. Barren Land, Pastures/Hay are the highest rank because those areas would require the least amount of effort to mine. While the other areas that are ranked lesser than these get more and more difficult to mine, which is why they are lower ranked. To properly distribute the different types of land cover into the correct rank, I used the reclassify tool to differ which types of land were most suitable for frac sand mining, to which ones could be used, and the different land types that could not be used. 

3. Rail Distance - The closer the frac sand mine is to a rail terminal the easier it would be to ship out the materials they mine, and it would also reduce the trucking cost. The farther the distance from the more expensive it is to ship the materials that are mined that is why the longer distances are ranked lower. In order to rank the rail terminals properly I needed to first use the euclidean distance tool so then we could figure out which rail terminals are in Trempealeau county or which rail terminals are closest too Trempealeau county. Once we found which rail terminals were closest or in Trempealeau county we can then rank them according to how far away the terminals are by using the reclassify tool. 

4. Slope - The areas with less elevation is better for mines to be built on, a very steeply sloped area is unsuitable for a mine, therefore the less slope the higher the rank it is given. To find the slope of the study area, we had to use a recent DEM that we had already created in a past assignment. The next step was to convert the DEM so then it would display slope, we utilized the slope tool before reclassifying or ranking anything. Once we had the output from the slope tool we would then use the block statistics tool which is where it groups similar cells with one another so than once we reclassified the output it would be easier to read. 

5. Water -  When looking at the water table we need to assume that the closer the water table is to the surface the higher the rank it gets, meaning the lower the height the higher the rank. In order to find the data needed for the water table we needed to go to a specific website and download the data. We then needed to use the 'topo to raster' tool to convert the contour lines of the water table over to a raster. Once we had all of the data we can now use the reclassify tool so than we can rank the water from how high it is away from the surface. 

After I had classified all of the different features, I used the tool 'raster calculator' to combine all of the data into one map that would provide us with a visual of where the best areas for a sand frac mine to be placed. Below is the Suitability model that was made for this assignment:

(Fig. 2) Suitability Model

Risk Model:

When creating the impact model we need to utilize five different variables than before four of which were chosen for us and the last was our choice.

1. Streams
2. Farm Land
3. Residential Areas
4. Schools 
5. Wildlife Areas

Each of these variables are ranked from 1 to 3, 1 having the largest impact and 3 having the smallest environmental impact. This is opposite from the suitability model because when we combine the two maps it will be easier to see where would be an optimal location for a mine within the study area. In order to reach a ranking of 1 (largest impact) the feature must be less than or equal to 2100 ft. To reach a ranking of 2 (moderate impact) the feature must be within 2101 to 4200 ft, and to reach a ranking of 3 (low impact) the feature must be greater than or equal to 4201 ft away from the mine. Below is a table referring to each of the features:

(Fig.3) Risk Table

1. Streams - The further that a mine is away from a stream the less likely it is to get contaminated by frac sand mining process. In order to classify the distance from potential mines in the county we first need to run an 'Euclidean Distance' which will help us locate all of the streams and the distance from potential frac sand mine sites. We will then use the reclassify tool so then we can classify the distances into the three categories.

2. Farm Land - The reason why we are using farm land as a feature is not because of the possibility of mine locations, but rather mining facilities don't want to contaminate their crops or animals with dust and different types of contaminants. We first need to change this feature into a raster, we do this by using the 'feature to raster' tool, then we reclassify that raster so than we can rank the distances once again.

3. Residential Areas - The distance between a mine and residential areas is very important, because a mine doesn't want to cause local communities any health hazards through the dust and contaminates created by frac sand mines. We located the residential areas in a zoning feature by creating a query statement to select all of the different school district locations. Then we ran the 'euclidean distance' tool so than we can find the distance from the potential mine sites. Then we reclassified the output value and ranked the distances.

4. Schools - The impacts of frac sand mining for schools is the same as listed above in the residential areas section. Before we can reclassify the feature we first need to fun an 'euclidean distance' tool so than we can find the distance between local school districts in the study area, and the potential mines. We would then run the 'reclassify' tool and rank the values by distance.

5. Wildlife Areas - The reason why I chose wildlife areas as my final feature is because just like residential areas, these wildlife areas are at risk of having their habitats contaminated by the dust and different contaminants produced by the mines. The process for this area is the same as the residential area category.

After I had ran all of the tools and classified all of the features I can once again combine all rasters by using the 'raster calculator' tool. This output will provide us with the areas in our study area that are most likely to be impacted by frac sand mining.

(Fig. 4) Risk Model

The final step of the assignment was to combine both the final output from the suitability model with the final output of the risk model. This will allow us to analyze the data so than we can discover the areas where it would be best for a sand mine to be located where it would cause the least environmental damage. I used the 'reclassify' tool for both models so than both outputs would have the same amount of ranks. Once the output from both reclassify tools was complete, we use the 'raster calculator' tool once more to combine both outputs into one map.

(Fig. 5) Combined Model 

Results

(Fig. 6) Suitability Model Maps

(Fig. 7)  Risk Model Maps

(Fig. 8) Ranking of best possible frac sand mine locations

Discussion

After looking at all of the data and all of the maps that were created the main area where there is high risk is more near the northwestern portion of the study area. This indicates that in order for the mine to have a low risk it needs to have the mine located in more the southeastern area of the study area. For the best results I would put a frac sand mine in the portions of the map where the value is greater than or equal to 11. The reason for this is because if a mine was to be constructed in an area where the value is less then 11, that mine has that much more of a chance of negatively impacting the environment. If the mine was to be constructed in an area where the value is greater then 11, that mine would be in one of the more optimal locations for frac sand mining both regarding the suitability for the mine and the risk towards the environment. 

Conclusion

Throughout this assignment many different aspects of it were very useful, and will continue to be very useful in the future. Combining multiple rasters is useful in wanting to combine many different features into a single map while keeping the specific weight of each feature. By creating two seperate models it helped us acquire better database management skills. The final result is a very unique map because it is combining multiple variables into one map and keeping the weight from each variable which allows us to find the most suitable area for a frac sand mine while keeping the environmental risks of the mine in mind.