Change Detection Using Kernel Density Analysis in ArcGIS Pro

Change detection is an important part of spatial analysis because it allows us to identify and assess alterations in spatial patterns across time. Understanding variations in the distribution and density of phenomena such as vegetation, population, or occurrences is critical for making informed decisions in sectors like as environmental monitoring, urban planning, and others.

ArcGIS Pro offers an interactive collection of spatial analytic capabilities, including the sophisticated Kernel Density tool. Kernel density analysis is a popular technique for transforming point data into continuous surfaces and displaying the spatial distribution of characteristics. This approach is especially good for detecting changes in point patterns over time.

Tutorial for detecting change using ArcGIS Pro

This tutorial describes a complete procedure for detecting changes in ArcGIS Pro using kernel density. Using this methodology, analysts and researchers can acquire useful insights into spatial trends, identify regions of major change, and contribute to better planning and resource management. We will look at how to input and prepare data, perform kernel density analysis, produce change detection surfaces, and visualize the results. The use of kernel density analysis in change detection not only helps identify areas of development or decrease, but it also makes it easier to analyze spatial trends, laying the groundwork for more in-depth analyses and decision assistance.

An overview of this change detection GIS tutorial

For this task, conflict data was accessed from the AmeriGEOSS community portal (Advancing Data Driven Decision Making in the Americas) in MS Excel file format which contains latitude and longitude for importing into ArcGIS Pro. The Excel file was regrouped into the years 1990-1996 and 1997-2003 and saved into .csv format as two separate files for importing into ArcGIS Pro.

Uganda districts shapefile was also downloaded from the operational data portal of the UN Refugee Agency (UNHCR). The conflicts and Uganda data were imported into ArcGIS Pro. Kernel density tool was used to map out the probability density of conflicts (1990-1996 and 1997-2003) in Uganda. The raster calculator was used to determine the change in conflicts from 1990 to 2003 in the area. Figure 1 shows the geospatial flow employed in downloading conflicts and Uganda districts data for import into ArcGIS Pro, using kernel density tool to determine the probability density of conflicts, and utilizing the raster calculator to determine the change between the conflicts.

Downloading and importing conflicts and Uganda districts data. 

1. Visit the United Nations High Commissioner for Refugees’s Operational Data Portal (ODP) to access the Uganda Districts 2020 shapefiles. The data contains boundaries of Districts in Uganda as at 2020. 

2. Retrieve the Uganda – Conflict Events dataset from the AmeriGEOSS Community Platform DataHub.

3. The dataset contains data on conflicts from 1989 to 2021 which we will now filter for conflicts between 1990 to 2003 to use for this tutorial which we will split into two CSV files.
4. Select the year column in the Excel file.
5. From the Home tab, select “Sort & filter” and specify smallest to highest. 
6. From the sort warning, select “Expand the selection.”

3. With a focus on the year field, copy data ranging from 1990 to 1996, paste into a new excel file and save as .csv. 
4. With a focus on the year field, copy data ranging from 1997-2003 and paste into a new excel file. Be sure to save the output of the new excel as .csv for importing into ArcGIS Pro. 

Using kernel density to map out conflicts from (1990-1996) to (1997-2007)

1. Open and create a new project in ArcGIS Pro. 
2. From the map tab, select “Add data.”
3. Toggle to the location of the conflicts data (1990-1996 and 1997-2003) file. 
4. Click OK to add the conflicts datasets onto the map canvas in table formats. 
5. Right-click on the 1990-1996 conflicts data (table format) from the table of contents and select “display X Y data.”
6. From the window that opens, set output feature class to your desired output location. 
7. Set X field to Longitude and Y field to latitude. 
8. Ensure that the coordinate system is set to GCS_WGS_1984.
9. Repeat the steps to display the 1997-2003 conflict data in shapefile format. 

10. To add the Uganda districts shapefile, select “Add data.”
11. Toggle to the location of the districts file. 
12. Click Ok to add the districts dataset onto the map canvas. 
13. From the search bar of the geoprocessing toolbox, type and search kernel density. 
14. Set the “input features” to 1990-1996 conflicts data. 
15. Set population field to none. 
16. Set output field to your desired output name and location. 
17. Click on the Environments tab.
18. Set processing extent to Uganda districts. 
19. Under the raster analysis tab, set mask to Uganda districts. 
20. Click “Run”.

21. Repeat this process for the second dataset. Set the “input features” to 1997-2003 conflicts data. 
22. Set population field to none. 
23. Set output field to your desired output name and location. 
24. Click on the Environments tab.
25. Set processing extent to Uganda districts. 
26. Under the raster analysis tab, set mask to Uganda districts. 
27. Click “Run.”

Using Raster calculator for determining change in conflicts

1. From the search bar of the geoprocessing toolbox, type and search raster calculator. 
2. Set the search bar to “1997-2003 Kernel” – “1990-1996 Kernel.”
3. Click Run.

Next, symbolize the resulting raster layer to symbolize areas that have increased change versus decreased change.

1. Type and search reclassify from the search bar of the geoprocessing toolbox. 
2. Set the number of classes to 5. 
3. Set input features to the results of the processing done with the raster calculator. 
4. Click Ok. 

The resulting symbolization maps areas that have experienced a decrease or an increase in conflict based on the comparison between the two datasets.

More ArcGIS Pro Tutorials

Fonte : National Geographic