Friday, December 16, 2016

GIS I Lab 4: Mini-Final Project

Introduction
This lab was intended to utilize a variety of tools to provide an answer to a geospatial question. The question to be answered was "Where can I live in Aitkin County, Minnesota? I want to live near a park and a hospital, but away from highways and railways". The specific criteria used were: within 10km of a park, at least 2 km away from railways and highways, and 25 km or less from the nearest hospital. The intended audience for this map is other GIS users and anyone looking for a home in Aitkin County, Minnesota.


Data Sources
To meet the criteria for this map, data concerning park land area, hospital locations, railway and highway locations, and the county border of Aitkin county. All data was obtained from the ESRI database server ( Data concerns include: age of the data (and therefore reliability), whether or not "hospitals" includes ER's and general health clinics, and whether or not railways are active or not.

Methods
To begin creating this map, a new file geodatabase was created to store and easily access all feature classes used in the process. For the Study Area locator map, three data layers were added: a Minnesota boundary layer, Minnesota counties layer, and study area layer. To create the study area layer, a query in the Minnesota counties layer was used to create a new layer containing only Aitkin County.

For the Suitable Living Area map, the three layers used above were added as well as these data feature classes: US Hospitals, US railways, US highways, and US parks. The parks layer was given a buffer of 10 kilometers and then dissolved. The park land layer was then erased from this buffer in order to produce a new layer representing living area within 10 km of park land. This layer was then clipped to the Aitkin County layer. The railway and highway layers were merged and clipped to the study area. A buffer of 2 kilometers was then created and dissolved. Since the criteria demanded that suitable living area exclude any land within 2 km of railways and highways, the buffer was erased from the layer representing living area near parks. This layer was then clipped to the Aitkin County layer. Finally, the hospitals layer was buffered to 25 km and then dissolved and clipped to the feature class near parks and away from rails and highways, creating a final result of a suitable living area including all of the requested criteria.

Below is a data flow model as a visual representation of the steps that were taken and the tools that were used.


Results
The map below represents suitable places to live in Aitkin County for a person looking to live near a park and hospital but away from railways and highways. There are four hospitals in Aitkin County to choose from. Much of the county contains park land, so it is not difficult to live near parks. Avoiding highways and railways is more difficult but can be accomplished. There is ample space to find an appropriate living area in Aitkin County even when considering the criterium.

Source: Esri - GIS Mapping Software, Solutions, Services, Map Apps, and Data.

Evaluation
This project was enjoyable because I got to pick the type of map I wanted to create. It was fun to put my skills to use in a way that brought GIS into a real life application. Something that I struggled with was creating the data flow model, however after developing such an elaborate one for this project I feel very comfortable with that skill. If I were to repeat this project, I would choose a different county. Preferably a larger one. I would also explore something different like where to put a golf course or a community pool. I kept my research question simple to ensure that I could properly display the skills that I have learned without making it too challenging and elaborate.

Sunday, December 11, 2016

GIS I Lab 3: Vector Analysis with ArcGIS

Goal
To utilize the various geoprocessing tools available in ArcGIS for vector analysis and to run script commands using Python to perform geoprocessing operations

Background
The DNR of Michigan wanted help determining where the most suitable bear habitats are located in Marquette County. This was found using these specific criteria: proximity to streams, the most common type of land cover in which bears are found, all within DNR management areas, and at least 5 km away from urban or built up lands.

Methods
Objective 1: Mapping an Excel file
Black bear locations of Marquette County, Michigan, were downloaded in an excel file containing XY coordinates. Since these data are in a non-spatial database, the coordinates were added as an "event theme" and after they were mapped on the NAD 1983 HARN Michigan GeoRef (Meters) coordinate system. They were then exported as a feature class to be brought into the geodatabase properly.

Objective 2: Determining preferred forest type of black bears
All feature classes from the bear management geodatabase were added to the map. A spatial join was performed between the bear locations layer and the land cover type layer. The MINOR_TYPE field was then summarized in order to find the three land cover types which contained the largest count of bears. These 3 forest types were considered the suitable bear habitats and therefore a criteria in determining the most suitable bear habitats. Using "select by attributes", these 3 forest types were selected and made into a new layer.

Objective 3: Determining correlation between streams and black bear habitation
In order to determine if distance to streams is an important factor in bear habitat selection, a buffer of 500 feet was created around all streams in Marquette County, and then the dissolve tool was used to connect the buffers. Next, the clip tool was used on the streams buffer layer with the bear locations layer to discover if most bears (at least 30%) were located near streams.

Objective 4: Finding suitable habitat (near streams in top 3 forest types)
An intersect was performed to find habitats near streams within the top 3 forest types. This provides the best suitable habitat for bears.

Objective 5: Finding suitable habitat within DNR management land
Once the best suitable habitat was determined, the DNR management land needed to be added to the criterion. A clip was performed to limit the DNR land to the study area being researched. The clip was then dissolved to eliminate the individual divisions inside of the DNR management lands. After this, the intersect tool was used to determine where the best suitable habitats were found within DNR management lands in the study area.

Objective 6: Eliminating areas near urban or built up areas
Urban and built up areas are hazardous to bears, so those needed to be eliminated from the possible study locations. To accomplish this, a spatial query was used in the land cover layer to create a new layer containing only urban and built up areas. In this new layer, a buffer of 5 km was created around all urban and built up areas, and then an erase was performed to eliminate these areas from the most suitable habitat on DNR management land layer. This provided the most suitable study area for bears in Marquette County, Michigan (Figure B).

Objective 7: Generating a data flow model
A data flow model was created including each step that was taken and the tools which were used to accomplish each task (Figure A)

Objective 8: Using Python
A few simple command scripts for geoprocessing tools were run using Python to demonstrate a more efficient way of running tools in ArcGIS (Figure C).

Results
The top 3 suitable forest types for bears were Mixed Forest Land, Forested Wetlands, and Evergreen Forest Land. It was also found that approx.  72% of bears lived within 500 meters of a stream, so streams were designated an important criterion in selecting suitable bear habitats. Figure B shows the most suitable bear habitat locations in Marquette County based on these two criterion, as well as the limitations of being within DNR management land and at least 5 km away from urban or built up lands.

Figures

Figure A. Data Flow Model

Figure B. Map of Suitable Bear Habitat in Marquette County, Michigan

Figure C. Python script


Sources
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