Running Kernel Density Estimation in ArcGIS Pro

in this video we're going to use ArcGIS Pro to perform a kernel density estimation on a point data set that has an attribute of a catch per unit effort value from some insect trapping for this analysis we can see we have two vector files a polygon of a ranch or a study boundary and a point data set indicated here in red and each of those locations was where a Fly Trap was run and individual flies were collected and the number of flies per unit time was calculated our goal is to convert these discrete data into a continuous data set using kernel density estimation so we can look at the concentration of biting flies across this study ranch kernel density estimation requires that we provide several key parameters first we need a point data set here the red dots traps CPU e and we need to set several parameters about the continuous surface we want to create first we need to define a grid cell size that's user-defined for this example we're going to use 100 meters we also have to define the search radius or how far we want our kernel density estimation to search before calculating that kernel function to calculate this we're going to use something defined as a chopped or the optimum bandwidth and that I'll show you the math for in a different video but it requires that we calculate the standard distance or a measure of dispersion of those trap locations and then use that in a calculation to estimate how far our search radius should extend we can calculate standard distance in ArcGIS Pro using the standard distance tool which is in the spatial statistics tools will first populate our input feature class the traps our output standard distance feature class I have going directly into a geo database so we can call this traps CPU e s dist this is going to create a new feature class or a new vector file and the standard distance is going to be a shape file which is a complete circle one standard deviation will be left as the default and in this case we want to weight those locations by the catch per unit effort value you can see our lab videos on our YouTube webpage that describe the difference between weighted and unweighted standard difference at another time and the case field is going to be left blank because we only need to calculate standard distance for this one single dataset if this were a categorized point dataset we could calculate on different groups which we defined by the case field and we'll run this analysis when this test runs successfully we get a green checkmark and our box is highlighted in green that this is run successfully and we get the shapefile added here or the feature class added here to our map window so here's the standard distance or the measure of dispersion for the trap locations the H opt calculation requires two pieces of information first it requires the sample size 42 it requires the standard distance is calculated in the analysis to derive H opt for us and you can learn about the H opt calculation in greater detail in one of our other videos briefly the a chopped is the calculation of 2/3 times in that calculation raised to the 1/4 power multiplied by the standard distance and map units these data are in UTM s so those are meters an H apt is calculated as 1230 3.5 meters or for this analysis we'll round up to 1234 meters let's look quickly at how we can derive those two pieces of information first to arrive at the number of samples in the traps from ArcGIS probe we can open the table for the attributes of trap CPU in here at the bottom you can see we have 42 events so there are 42 trap locations that corresponds to the in if we open the attribute table for the trap CPU e standard distance our standard distance value here three four seven five point one nine yes the standard distance that I used in the excel table you saw moments ago so those are what we get those two pieces of information to calculate a chopped which is going to allow us to populate our kernel density estimation there are several ways to find tools in ArcGIS Pro first we could come here and find tools here I have kernel density is the first tool that comes up when I start searching that term likewise it's one of the tools built in to the analysis tab in ArcGIS pro by default both of those are going to launch the geoprocessing tool here for kernel density estimation now we can populate our kernel density function first we're going to need our input which is the trap CPU e and in this case we're not interested in the concentration of the trap locations but rather the Flies that were captured at each trap site so we're going to populate this with that catch per unit effort value I described at the beginning of the video next we have to define an output location for this I'm going to call this KDE CPU e as my reminder of what function we're applying this to the CPU II data my output cell size is going to be 100 100 meters by default that's a user selection and my search radius is going to come from that table we populated moments ago 1234 meters for a chopped my area units are going to be in square map units and my output cell values by default will be density we could also select the expected count and remember here the expected count would be the Flies per unit effort we can then click on the environments tab we can select that the output coordinate system match the same as the traps that's our UTM and we can define a mask and we can say mask this kernel density to the output of the ranch by default that kernel density if it runs successfully again will complete and be in green you can read the statistics on that if you're interested when you run it and the map itself is going to be added automatically we can see that here in the background here ArcGIS Pro has assigned default light purple to dark purple color ramp and we can see very quickly that we have a high concentration of those biting flies captured here in the northeastern portion of the ranch and across the northern part of the ranch we also see shades of purple and then as we move out towards the western and southern part of the ranch we can see that very very low or no flies were collected so there you have it kernel density estimation in ArcGIS pro good luck

Channel: SEER LAB UNIV OF FLORIDA

Views: 2,553

Rating: 5 out of 5

Keywords:

Id: eQ8ZNmtImRk

Length: 8min 52sec (532 seconds)

Published: Wed Feb 05 2020