Lab 4: To what extent do cattle and agriculture affect water bodies in Wisconsin?

Goal

The goal of this lab was to answer the question: To what extent do cattle and agriculture affect water bodies in Wisconsin?

Background

With a substantial amount of phosphorus runoff from cattle farms and conventional agriculture, Wisconsin lakes and streams can bear a heavy load of this nutrient. Usually a limiting nutrient in water, large inputs of phosphorus put precious bodies of water at risk of an overgrowth of algae and weeds, anaerobic conditions, and even anoxia. A GIS risk assessment of such eutrophication in terms of cattle inventory and fertilized acres per county could make this relationship clear to those who may be unfamiliar with it.

Procedure

First, counties were clipped to the state of Wisconsin using the ESRI2013 database provided by UWEC. Next, data was gathered from the USDA QuickStats online index. This included the amount of fertilized acres and the total amount of cattle inventory by county in Wisconsin. Data were entered into a spreadsheet and imported into ArcMap as a data table

This table was joined to the county attribute table, using “NAME” of the counties layer and “County” of the spreadsheet data. Graduated color maps were then made of both cattle inventory and fertilizer acres, which was normalized by total square miles of each county.  

The two highest levels, based on the Jenks natural breaks classification method, of cattle inventory (>31000 total) and fertilizer use (>161 acres per mile²) were selected and new layers were created for each criteria. In addition, a layer of counties that were both high in cattle inventory and fertilizer use was created using an intersect between them. All layers were put onto a third data frame.

The DNR database offered map data of both impaired lakes and impaired streams. Streams and lakes whose impairment was high phosphorus levels were selected and new layers were created from those selections. Phosphorus-impaired water bodies were added to the all three data frames.

Eight separate select by location tools were used to determine how many lakes and streams that are known to have high phosphorus concentrations intersect the highly fertilized counties, high inventory cattle counties, and counties with both criteria.

Figure 1: Data flow model for Lab 4

Results

The map shows cattle inventory and fertilizer use in Wisconsin counties, as well as locations of streams and lakes with high levels of phosphorus. 

Figure 2: Map of Wisconsin counties and phosphorus-impaired rivers

A table was also created to show how the amount of streams and lake affected by high levels of phosphorus is correlated with high cattle inventory (>31000), a high proportion of fertilized acres (>161 acres per mile²), and both criteria together. As shown, cattle and fertilizer use alone increased the likelihood of a stream having high phosphorus levels, while the likelihood for counties high in both criteria was over three times that of all counties in general. This shows a significant correlation between cattle, fertilizer use, and lake and stream phosphorus levels.

Counties	Total Area (mile2)	Affected Streams	Streams/ mile2	Affected Lakes	Lakes/ mile2
All	68423.81	259	0.0037	119	0.0017
High Cattle	17136.67	146	0.0085	51	0.0030
High Fertilizer	21236.1	192	0.009	67	0.003
Cattle + Fertilizer	12485.24	131	0.011	45	0.0036

Figure 3: Streams and lakes with high phosphorus levels

Sources

2012 cattle and fertilizer data was gathered from:

https://www.nass.usda.gov/Quick_Stats/index.php

County data:

ESRI2013 database

Lake and Stream Data:

DNR database

Evaluation
I thought this was a fun lab. Although I probably could have made it easier by doing something more similar to Lab 3, land conservation is interesting to me so I pursued it. I think this made it difficult to come up with creative ways to use different tools. 

Lab 3: Finding Suitable Bear Habitat in Marquette County, MI

Goal: The goal of this lab was to use tools in a data flow model to find suitable habitat for bears.
Background: Using multiple criteria including land type, stream vicinity, distance from urban areas, and DNR management, suitable habitat for bears was found in a study area of Marquette County. Specifically, land type criteria and stream vicinity criteria were assessed by looking at the proportion of bears inhabiting those areas. Detailed methods are listed below:

1. Add locations as XY event theme.
1. Go to the file menu to add XY data.
2. Locate the bear location table and assign X Field and Y field based on table columns.
3. Choose the correct coordinate system (NAD_1983_HARN_Michigan_GeoRef_Meters)
4. Export data points and save in a “lab3” geodatabase.
2. Determine bear habitat
1. Add and arrange all feature classes from the bear_management_area feature dataset. 
2. Symbolize landcover by “minor type”
3. Intersect landcover and bear_locations to create a table in which the bear ID numbers correspond to the type of landcover they were found in.
4. Summarize “Minor Type” to get a count of the total number of bears found in that type of landcover. The top three sites are Mixed Forest Land, Forested Wetlands, and Evergreen Forest Land.
3. Determine if bears are found most often within 500 meters of stream.
1. Create a 500 meter buffer for the “streams” layer.
2. Intersect this stream buffer with the bear locations. Although the resulting area is only about 10% of the total land cover in this study, there are 49 total bear sightings within 500 meters of a stream, of a total of 68 sightings. This is 72%, which is well above 30%, so streams are clearly important habitat for bears and should be used for criteria for suitable habitat.
4. Find suitable areas for bear habitat.
1. Using previous research, it is known that bears prefer to be within 500 meters of a stream and inhabit the Mixed Forest Land, Forested Wetlands, and Evergreen Forest Land more than other land cover types. 
2. Create a new layer of the preferred land cover types. 
3. Intersect the stream buffer with the preferred landcover types. This results in several polygons, so the dissolve tool was used to result in a single, albeit disjointed, shape containing suitable bear habitat in the study area. 
5. Make recommendations to the Michigan DNR for bear habitat on their land. 
1. Add the dnr_mgmt feature class and clip to only include the study area.
2. Because the units of DNR managent are not important, their boundaries were dissolved.
3. An intersection was used to find the common areas.
6. Eliminate Urban or Built Up Areas from the DNR Managed Habitat Area
1. A new layer from landcover was selected to possess only the Urban or Built Up Areas
2. A 5 km buffer was made around the area.
3. The DNR Managed Habitat was erased where it was within that buffer. The result is all areas that are suitible for bears, managed by the DNR, and more than 5 km from an urban or built up area.
7. Create a cartographically pleasing data flow model and map.
1. Throughout the process, queries were made and tools were used in a single data flow model that was kept orderly. This made the process easy to visualize and a simple capture of the model shows the data flow in its entirety.
2. A map was created using bright colors to signify various suitable habitats for bears in the study area. The basic map elements, including legend, title, author, source, scale, and north arrow were added. A locator map showing the location of the study area within Marquette County was also include
8. Perform geoprocessing commands using Python
1. After opening the Python window, “arcpy” was imported by typing “import arcpy.”
9. Code for all tasks, including a buffer analysis for within 1 km of streams, an intersect between this buffer and suitable land types, and an erase of areas within 5 km of urban areas was written and/or selected from the drop-down menu as follows:

Results: The resulting map shows different levels of suitable bear habitat in the study area of Marquette County, Michigan. First, it was found that bears prefer forested areas that are also within 500 meters of a stream. Because the DNR is interested in a management plan for bears, areas of suitable habitat within the DNR management area were identified and even further reduced to exclude areas within 5 km of an urban or built-up area.

Sources:

• All of the data were downloaded from the State of Michigan Open GIS Data
• http://gis.michigan.opendata.arcgis.com/
• Landcover is from USGS NLCD
• http://www.mcgi.state.mi.us/mgdl/nlcd/metadata/nlcdshp.html
• DNR management units
• http://www.dnr.state.mi.us/spatialdatalibrary/metadata/wildlife_mgmt_units.htm
• Streams from
• http://www.mcgi.state.mi.us/mgdl/framework/metadata/Marquette.html