POST
|
Hi Dan, thanks for the response. There are four polygon zones (A, B, C, D) in my vector layer. These zones are not always contiguous. What I really want to know is how many tree canopy pixels (value of 1) are in each zone. Unfortunately, grouping them together by contiguity doesn't help me do that unless there are some additional steps you would suggest.
... View more
10-22-2019
08:41 AM
|
0
|
1
|
2853
|
POST
|
What I’m trying to accomplish is simple to define: what is the percentage tree cover (raster) across four polygon zones (vector)? When I run my analysis using Zonal Statistics as a Table, I get an equal ratio of tree canopy in each zone (Count/Area). Each zone seems to have 25% of the total tree canopy count, but this can’t possibly be accurate. Screenshot is below and the ratio of Count (i.e. tree pixels)/Area across all four zones is .25. Common sense tells us that the tree canopy wouldn't be equally distributed across any area like this so obviously I'm doing something wrong. HOLC_Grade is the four polygon zones grades where A='good' neighborhoods' and D= 'bad' neighborhoods according to redlining criteria used to establish discriminatory lending practices in Chicago. My hypothesis is that areas that were redlined in the past might have a lower percentage of tree canopy than areas that were not. Data: High-Resolution Land Cover, Cook County 2010 the Chicago Region Trees Initiative https://datahub.cmap.illinois.gov/dataset/high-resolution-land-cover-cook-county-2010 Home Owners Lending Corporation Redlining shapefile for Chicago from here: https://dsl.richmond.edu/panorama/redlining/#loc=9/41.944/-87.771&maps=0&text=downloads This layer has four redlining ‘grades’ that indicate how viable a neighborhood was for lending. These grades were symbolized as polygons of different colors where red='bad' areas denied lending. Here is my work flow: Step 1: Land Cover layer - Use the Extract By Attributes tool to create a new layer that is just Tree Canopy (excluding the other 6 land cover classifications) Step 2: Clip the new Tree Canopy raster to the Redlining layer (which has been symbolized according to the four redlining grades). I used the Data Management Tools>Raster>Raster Processing>Clip tool. *I also tried this step by Extract by Mask instead of Clip and got the same results. I get a clipped layer with an attribute table that looks like this (where Value of 1 is = Tree Canopy). So there are 354,919,880 tree canopy pixels in the redlined polygon zones. Step 3: Run the Zonal Statistics as Table tool using a) Redlining layer as the feature zone data b) ‘holc_grade’ as my Zone Field (these are the four investment zones) c) My new Clipped Tree Canopy raster as the Input Value Raster d) selecting all statistics The resulting zonal statistics table I get looks like this, with the Count/Area ratio by 25% across each polygon zone I'm wondering if a) I'm not understanding what Count and Area really mean? Count is the number of pixels, and Area is square meters... b) There is a projection issue? I set the Data Frame to UTM Zone 15N when I started the project. However, when I check the Source Properties for my Redlined layer it says the GCS is WGS 1984, but doesn't list a projection. I'm also including a screenshot of my TOC and map just for reference. Any help is much appreciated!
... View more
10-15-2019
09:27 AM
|
0
|
3
|
3779
|
Online Status |
Offline
|
Date Last Visited |
11-11-2020
02:25 AM
|