Suitability for Fracture Sand Mining in Lower Trempealeau County

Goals and Objectives

This assignment is the final segment to measure the potential areas in the lower portion of Trempealeau County Wisconsin which are highly suitable for fracture sand mining. To compile a meaningful suitability map, negative impact is combined with the suitability of land for mining. Mining of any form is a dusty, messy, and noisy business, so steps must be taken to minimize the impact for the surrounding community. This project will involve making two data flow models and combining them at the end to facilitate an efficient analysis process. The first model will consist of a suitability model. The parameters for this model are geology (rock type), land use / cover, the distance from railroads, the slope of land, and water table access. The second model will be made for potential risk on socio-economic factors and environment. The parameters for this model will be the impact to streams, prime farmland, populated areas, and schools. These two models will be combined in the end to produce a final suitability map which takes into account all the factors listed above to produce a meaningful map for potential placement of a new fracture sand mine.

Methods part 1

Starting from the beginning, the local environment is set in Esri ArcMap, and the spatial analyst extension is turned on. To compose the suitability model, first a geology polygon feature class is added to model, and converted into a raster using ‘Polygon to Raster’ tool. A ‘Reclassify’ tool was used to rank to types of geologic layer for its suitability for mining for sand. For our purposes Jordan and Wonewok rock layers are the ideal mining layers, and were set as High suitability. All other rock layers were set to a low suitability (Figure 1b).

            The next objective was to create a spatial layer for land use and cover. The NLCD landcover landuse layer created in the initial portion of the project is used. The legend for the landcover can be found online, and is used to class the types into suitability ranks. After the land cover types are ranked, another reclassify is used to exclude land types, which are completely unsuitable for mining activities. I excluded any water formations including wetlands, any development or forested area. These land cover types would just be too costly and time consuming to facilitate a mining being built (Figure 3).

            Suitable mines will be close to a rail road, for easy facilitation of sand transport (Figure 4). A rail terminal feature class is added and clipped to the lower portion of Trempealeau County. The extent of environmental setting is changed to the entire county as to not truncate any results. Finally a ‘Euclidean distance’ tool is used on a raster of the lower Trempealeau County to calculate the distance from railroads. Euclidean distance calculates the distance from an object, to each individual cells in the raster. The distance is then reclassified (Figure 1b), into low, average, and high suitability.

            If I mine is located on a hill with too great of a slope, the equipment will not be able to extract the sand from the mine properly. To calculate the slope, first a DEM raster is added from a previous step of the project. A 32-bit raster is used as not to truncate any of the elevation data. A ‘Slope’ tool is used on the DEM, which produced a slope for the map. This calculation resulted in a salt-and-pepper effect, so a ‘Block statistics’ tool was used to average out the slope, which in turn produced a clearer elevation map. The average slope was then reclassified, for the suitability to mine.

            Fracture sand mining requires a great amount of water. Access to a supply of water, without having to drill deep into the ground greatly reduces the overhead costs of operating a mine. To incorporate this into the suitability model, a water elevation table is downloaded from the Internet. A ‘Import from .E00’ tool must be ran to convert the .E00 file into a coverage file. This coverage is then added to the model, and converted into a raster after projecting. This raster possesses the elevation of the water table, not the depth from the surface. The water elevation is then subtracted from the DEM in the last objective to get the true water table depth. This is then reclassified and ranked (Figure 1b).

            The final step in creating a suitability model is to combine all of the spatial layers into one layer displaying the total suitability. One thing to be cognizant about throughout the project is to keep all the layers in the same units (meters), to ensure all of the spatial layer information is accurate. All the layers are added together using ‘Raster Calculator’, and a map is produced. 

Methods part 2

            The risk model will be made in the same way as the suitability, but with different criteria. For the risk model, the parameters consist of spatial layers of the impact by proximity to streams, prime farmland, proximity to schools, residential and populated areas. The first portion of the model will be made to examine closeness to streams and rivers. The base of the spatial layer is the DNR flow line feature class downloaded with the Trempealeau County geodatabase in a previous exercise. Wisconsin is covered in small streams, so running a distance off small streams would render the county with continuous high impact from streams. I choose to eliminate all the smaller subversive streams and rivers and only keep large rivers. The class of rivers I ran the proximity tool off of contained portions of the Black, Rock, and Mississippi rivers respectfully. A Euclidian Distance tool was used, then the spatial layer reclassified.

Next came impact to prime farmland. This layer was relatively easy to make, the value field was used to convert it to a raster and a reclassify organized the types of farmland into impact levels.

The next impact layer to be produced is the proximity and impact to populated areas. This task contained more data preparation, as there is no feature class, which we have that solely contains population data. I used the NLCD feature layer along with the zoning layer to make a population area layer. An important issue involved with mining operations is mining and dust, and to comply with state regulations, a mine can be located no closer then 640 meters from a residential area. A noise shed is created using Euclidian distance, and eliminating the values lass then 640. Extra steps were taken to ensure the map contained the same units as the others (meters). Similarly to the populated areas layer, a distance from schools layer is also created. For this is used a schools feature class in the Trempealeau County database to look at the language used to name the school district buildings, and query them in the zoning feature class. I combined the population and school district data into one layer and ran a final distance tool from this layer, and reclassified it into 3 categories.

Finally, all the raster’s are added together to form a cohesive risk model using the “Raster calculator’ tool. Once the model is completed, overlay analysis is conducted on the suitability model and risk models. The risk is subtracted from the suitability, producing an overall suitability layer. The layer is reclassified, with the outcome resulting in areas of high, average, and low suitability.

Results:

Figure 1: Table showing the rankings given to the suitability and impact spatial layers. A 3 signifies a high suitability/impact, 2 an Average suitability/impact, and 1 represents a low suitability/impact.

Figure 2: The completed model from suitability of Trempealeau County  for fracture sand mining. The upper portion of the model is the suitability portion, and the lower portion of the model is the risk/impact model.

Results

Suitability Maps

A.B.

C.D.

F.

Figure 3: Suitability of lower Tempealeau County for fracture sand mining. Map A represents the land use, land cover suitability. Map B represents the geologic suitability. map C represents the proximity to train terminals. Map D represents the water table depth. Map D represents the Slope suitability. For all the maps above, Blue represents a High suitability, Tan is average suitability, and Yellow is low suitability. 

Risk Maps

A.

 B.

C.

Figure 4: Risk layers for fracture sand mining in lower Trempealeau County. Map A displays risk to streams and rivers. Map B shows impact to prime farm land. Map C shows impact to populated areas, and schools. 

For maps A and B, Red represents high risk, tan represents average risk, and gray represents low risk.

For map C, white/ light gray is low risk, while black is high risk.

Figure 5: Displaying the total suitability for fracture sand mining in lower Trempealeau County. Areas of high suitability are shown in blue, average suitability shown in tan, and low suitability shown in purple. Continuing, the suitability map shown beneath. Accompanying the suitability map is a risk map containing all the criteria for impact.

Discussion: 

The conclusion of this semester long project resulted in a comprehensive analysis of environmental and socio-economic factors which would lead to a productive and suitable mine location. While most of the previous projects focused on the entire state of Wisconsin, the last portion only involved the lower portion of Trempealeau County, in order to cut down on the run time of the raster tools. Focusing on raster analysis to produce the final product maps and model, students were exposed to a great deal of tools and practices. During the process of this final project, students were purposely given slightly vague guidelines to produce the final project. The aim of the faculty for this was to introduce the students to a less structured project in order to familiarize students with real data, and not simply ‘hand-hold’ them through the process. An outcome of this is students really learn how to solve problems themselves by looking at available resources and actively trying to problem solve. As students shift towards a more professional setting, problem-solving abilities are high lighted.

           

         The spatial layer analysis for the suitability of lower Trempealeau County resulted in discrete areas, which inhabit the best suitability for a future mining operation. The final output displayed approximately 4,500 miles² of land with prime suitability for fracture sand mining following the environmental and socio-economic criteria set for this project (figure 10). Most prominently the suitable land was located in the west-central portion of lower Trempealeau County. There are a few notable rational for this. Firstly, the rail station located the closest to the county is next to the western border (Figure 4). Another element, which caused the primary suitability, is the slope of the county. Figure 5 shows that most of the center of the state has a lot of elevation change which is not productive for mining. The last feature that I would like to high light is the locations of residential, populated, and school areas (Figure 9. The southeastern portion of the county is almost entirely filled with populated exclusion zones. This pushes any available into the northwestern part of the state, exactly like we see in figure 10.