GIS - joining parcels layer with zoning layer

[jammers]

So what I'm trying to do is join the municipal parcel layer with the zoning layer (using ArcEditor). So far I've been able to join the two layers using the Spatial Join function (one to one) using the "Intersects" matching option. The problem I'm having is the fact that many parcels have multiple zones (i.e. many zones are zoned Residential but are also zoned Environmental Protection where watercourses exist). The resulting attribute table for the joined layer lists parcels and their corresponding zones and if they have more than one zone, they are listed multiple times along with each different zone. My problem is that bordering zones are also included when they should not be. This is most likely due to the fact that the spatial join was based on an "Intersects" fuction. However, I've tried it with the "Contains" and "Is_Within" match option but neither produce results better than the Intersects option.

I've also thought about the possibility of converting the parcel layer into points, but that would not address the issue of multiple zones present in certain parcels since only the zones where the point is contained in would be joined.

I do not want to end up filling in the zones manually for obvious reasons. I have a feeling it should be a lot easier than I'm making it out to be. Does anyone have any suggestions? Please, I'm dying here...hah

[Queen B]

Well you do have quite a dilema that appears all the time in zoning.
I only know enough about GIS to be dangerous.
Zoning and Parcel boundaries are deceptive, one of my first lessons about zoning. The zoning goes with the land so therefore can't really rely on parcel boundary lines. For example Lot 1 is zoned B1 Lot 2 is zoned B2 but is vacant. the person owning Lot 1 buys 1/2 of Lot2 that does not change the zoning to B1 but the new owner has a property that is zoned B1 and B2.

As for protection areas, the only thing I can recommend would be a different layer for each area.
Your parcel line layer is a perfect place to start but it is not the end all to keep you from alot of work. I know that is not what you wanted to hear and may have confused you even more.

[jammers]

Originally posted by Queen B

Well you do have quite a dilema that appears all the time in zoning.
I only know enough about GIS to be dangerous.
Zoning and Parcel boundaries are deceptive, one of my first lessons about zoning. The zoning goes with the land so therefore can't really rely on parcel boundary lines. For example Lot 1 is zoned B1 Lot 2 is zoned B2 but is vacant. the person owning Lot 1 buys 1/2 of Lot2 that does not change the zoning to B1 but the new owner has a property that is zoned B1 and B2.

As for protection areas, the only thing I can recommend would be a different layer for each area.
Your parcel line layer is a perfect place to start but it is not the end all to keep you from alot of work. I know that is not what you wanted to hear and may have confused you even more.

Thanks for the response QueenB. Man, its going to be a bitch trying to clean up the data. The fact that the zoning does follow parcel boundaries is a major pain.
Regardless of what I'll end up doing, what do you think would be the best way to design the attributes to address multiple zones for a single parcel? I was thinking of having a column called "Zone 1", "Zone 2", "Zone 3", etc. instead of listing all the zones in one column using commas. Of course not all parcels will have 3 zones but I figure this will be the most efficient way to create the database so that it will make future analysis based on zoning attributes alot easier and more effective.

[JimPlans]

First, an assumption: I assume that you are not talking about overlay districts for your environmental or other zones, but split parcel zoning. Meaning that, for example, the first 80 feet of a 120 foot deep parcel from the lot line might be Residence-A, but the last 40 feet would be Preservation (or whatever).

Second, I assume that you are running into simple split parcel issues, like what is described above, and not having problems with slivers caused by geocoding errors in one or another of the data sets.

If the above two assumptions are correct, then I'm not sure what the problem is. The only way to properly manage the zoning issues that you are discussing is to track zoning by parcel, and split parcels by definition will have two or more records per parcel to reflect the different zoning districts covering the parcel (and how much of each parcel is covered by each zone). Keeping the parcel/zoning database in this form will ensure that all of your data is in one place.

If you want to view parcels only, just dissolve the layer based on parcel ID. For zoning only, dissolve on zoning ID. Think of the parcel/zoning layer as raw data in a database that generally needs to be processed before it can be used.

For real fun, try scanning a 40 year old paper zoning map colored with markers (with erasures, changes, white-out, and all that), geocoding it, overlaying parcel lines, and going parcel-by-parcel to enter the zoning info (including parcel splits). Fun.

This is the problem with GIS. Entering and creating data can be incredibly tedious and time-consuming. It's the nature of the beast.

[jammers]

Originally posted by JimPlans

First, an assumption: I assume that you are not talking about overlay districts for your environmental or other zones, but split parcel zoning. Meaning that, for example, the first 80 feet of a 120 foot deep parcel from the lot line might be Residence-A, but the last 40 feet would be Preservation (or whatever).

Second, I assume that you are running into simple split parcel issues, like what is described above, and not having problems with slivers caused by geocoding errors in one or another of the data sets.

If the above two assumptions are correct, then I'm not sure what the problem is. The only way to properly manage the zoning issues that you are discussing is to track zoning by parcel, and split parcels by definition will have two or more records per parcel to reflect the different zoning districts covering the parcel (and how much of each parcel is covered by each zone). Keeping the parcel/zoning database in this form will ensure that all of your data is in one place.

If you want to view parcels only, just dissolve the layer based on parcel ID. For zoning only, dissolve on zoning ID. Think of the parcel/zoning layer as raw data in a database that generally needs to be processed before it can be used.

For real fun, try scanning a 40 year old paper zoning map colored with markers (with erasures, changes, white-out, and all that), geocoding it, overlaying parcel lines, and going parcel-by-parcel to enter the zoning info (including parcel splits). Fun.

This is the problem with GIS. Entering and creating data can be incredibly tedious and time-consuming. It's the nature of the beast.

JimPlans, your assumptions are correct. Just to clarify, the records in my Zoning layer attribute table are made up of polygons that represent different (but also identical) zones scattered throughout the Municipality. As a result, I had to use a Spatial Join rather than a join based on a common attribute between the two layers (since one was based on Parcel ID and the other based on Polygon ID that has nothing to do with Parcel ID).

The problem with the Spatial Join is that yes, it does split parcels with two or more records per parcel to reflect different zoning districts covering the parcel, BUT it also includes neighbouring zones that border the parcel (as a result of using the Intersects matching option). As a result, I have to go to each parcel and delete the neighbouring zoning codes that are included with the parcel after the join function.

I don't understand what dissolving by parcel/zoning ID would accomplish. Would you be able to clarify?

Hopefully I have made some sense trying to explain this confuffle...

[JimPlans]

Originally posted by jammers

JimPlans, your assumptions are correct. Just to clarify, the records in my Zoning layer attribute table are made up of polygons that represent different (but also identical) zones scattered throughout the Municipality. As a result, I had to use a Spatial Join rather than a join based on a common attribute between the two layers (since one was based on Parcel ID and the other based on Polygon ID that has nothing to do with Parcel ID).

The problem with the Spatial Join is that yes, it does split parcels with two or more records per parcel to reflect different zoning districts covering the parcel, BUT it also includes neighbouring zones that border the parcel (as a result of using the Intersects matching option). As a result, I have to go to each parcel and delete the neighbouring zoning codes that are included with the parcel after the join function.

I don't understand what dissolving by parcel/zoning ID would accomplish. Would you be able to clarify?

Hopefully I have made some sense trying to explain this confuffle...

First, while I am a Maptitude user first and an ArcGIS user second (or third), I'm pretty sure that in ArcGIS you want to do a Union and not a Spatial Join. You want to actually split each of the parcel polygons into multiple polygons based on the zoning polygons and not just append data from the zoning layer to the parcel layer while leaving the parcel layer's polygons intact, as I understand the task.

http://webhelp.esri.com/arcgisdeskto...erlay_analysis

If you're just appending data, I would think that setting the Search Radius on the Spatial Join tool to 0 and setting the proper units (I assume meters) should stop adjacent zoning polygons from "infecting" the results.

As far as dissolving is concerned, if one parcel polygon is split into two because of zoning, if you want to present your data as whole parcel polygons in the future you will need to dissolve them back together. You wouldn't really have to do that except for display reasons, as you can always perform analysis with the parcel/zoning polygon layer without dissolving it back together as one or the other.

[jammers]

Alrighty, so I've performed a Union which produced better results than the Spatial Join. I've also done a Spatial Join by setting the Search Radius to 0 as well as -1, -10, just to see if there would be any difference. Unfortunately I still get adjacent zones "infecting" the results. I've also set the XY Tolerance to 0, -1, and -10 for the Union function, but the results are the same.

The only difference between the two methods I found was that the Union was more accurate in picking up zones that did not intersect the boundaries of the parcels. Unfortunately I don't have the ArcInfo license which would allow me to use the Dissolve tool. Would dissolving the parcels back together also append all the appropriate zones to one record in the attribute table instead of multiple records for the same parcel? Ideally, I'd like to have one unique parcel record and multiple zoning designations as additional columns. I'm doing it manually right now and cannot figure out a way to do it more efficiently. Maybe this is as far as I can go?

Thanks for all your help JimPlans. Definately made my life a lot easier! You rock!

[JimPlans]

Originally posted by jammers

The only difference between the two methods I found was that the Union was more accurate in picking up zones that did not intersect the boundaries of the parcels. Unfortunately I don't have the ArcInfo license which would allow me to use the Dissolve tool. Would dissolving the parcels back together also append all the appropriate zones to one record in the attribute table instead of multiple records for the same parcel? Ideally, I'd like to have one unique parcel record and multiple zoning designations as additional columns. I'm doing it manually right now and cannot figure out a way to do it more efficiently. Maybe this is as far as I can go?

I thought everyone who had an ArcGIS license had dissolve. I have ArcMap 9.2 and I have it in my toolbox. Unfortunately, dissolve doesn't concatenate records together, it just chooses the first or last value (or sums or averages, etc.).

This is one of my most hated problems: Concatenating multiple fields into one field in a database. I only have to do it once every couple of years, and the only way I have ever done it correctly is by using SQL. I don't process any of my data in ArcGIS, so I don't know how it would be done using that program. I would do it in MS-Access (ugh), and since I haven't had to do it in a couple of years it would take quite a while to figure it out again.

If there are only a limited number of zones, you could work around it by making a new variable (column) for each zone, and making a simple formula that would test the zone category that you have written to the database with the Union query, and create a simple matrix. Then, you could dissolve the layer back together and have a column for each zoning type with an indicator in it if that zone is represented on the property. That might even be a more useful final product, because then you would be able to manipulate the data more easily.

[BrianVDB]

Jammers,

What is the goal of the project? Is it simply so you can identify a parcel and know the zoning without turning on the zoning map? If so, and this is probably the most obvious solution and you've already considered it, but do you have access to an assessing program with zoning district attibutes? You would simply export the information to a database and use the parcel identification numbers to join the polygons to the database. The assessing program we use allows input of multiple zoning fields for split zones.

If that's not an option, and you can't create a geoprocessing model, have you experimented with using the "Select By Location" tool to identify where the zoning districts fall? For example, select the "Commercial" zoning district within your zoning layer. Open the Select By Location tool, and tell it that you want to select features from Parcels that .... (have their centroid in, are within, etc.) ... the features in Zoning, and check the "Use Selected Features" box. Run the query, open up the Parcel attribute table, and use the Field Calculator tool to assign attributes to the Zoning field of the selected parcels. Obviously, this is still going to be labor intensive, because you'll have to be conscious not to overwrite data you've already assigned (just sort the results based on values in the zoning field to consolidate). On parcels where you've already assigned data to the Primary Zoning field, assign it to a Secondary Zoning field, and so on. Then again, you might still run into problems with the GIS selecting parcels that share borders, but it's something you could play around with. I'd start with a small area where you've already identified conflicts and see if it works.

That said, and I don't know the nature of the project, but I don't know if I would want to permanently assign zoning data to my primary parcel layer, because when a zoning map amendment occurs you're going to have to edit each parcel. Maybe that's not a big deal, but it's a much simpler process to include the zoning attributes in your assessing database, and export and update periorodically along with the owner information, mailing addresses, tax values, etc.