![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
What is urban
It's been a very long time since I did a post about a geography problem, so here we go. Here at work I was asked to assist a project wanting to highlight how much work our agency does in urban areas. Sounds great, urban agriculture is a thing, and our agency supports it, so lets highlight it. So, how much do we do in urban areas? Well that depends on what exactly an urban area is. Urban is one of those "I know it when I see it" things that doesn't have a commonly agreed upon definition. We could go with the city limits of our largest cities, but that would leave out the surburbs which I think most people would consider part of the urban area.
Many government reports use the Census Bureau's Metropolitan Statistical Areas (MSA), which count any county where at least 25% of the population commutes in to a central city that is large enough. The trouble with MSAs is that they use whole counties as their building blocks and in reality most counties are a mix of urban and rural so as a result you get a very broad definition that includes wide swaths of what people would normally think of as rural. Our first cut was using MSAs, but we found it was indeed too broad, most of what it was capturing was not stuff that would be considered urban agriculture. Fortunately we have another source to help us out, the National Land Cover Database (NLCD) which is a satellite imagery classification scheme that sorts land use patterns visible in the imagery into various categories.
So let's take a look, we have the MSA which is too broad, the city limits which is too narrow. We can see on the Land Cover background in general what we'd like to capture - we want the pink-reddish stuff. But the problem with the NLCD is there is no actual boundary there, we could use it to eyeball a particular farm and say whether it's urban or not, but we want a criteria that we can apply nationwide.
The U.S. Census Bureau to the rescue, they have developed what they call "urbanized areas" which they define as "a continuously built-up area with a population of 50,000 or more." Mapping the one for Clevelend it looks pretty promising and so we run some reports using that boundary. Seeing the results our team decides that it is still too broad, it's mostly capturing farms along the edge of the boundary, not really "urban" enough for what we want to be reporting.
So maybe we can use the NLCD directly. We can still use that Census urbanized area boundary to limit our area of interest. Then on the NLCD we can highlight the "Developed, Medium intensity" and "Developed, High Intensity" areas, which generally translates to commercial, industrial, and multi-unit residential areas that are the core of "urban." So do we have any agriculture on those areas? A little, but it turns out that we've gone too narrow again, since even urban farms usually come up as either the "Culivated Crops" or "Hay/Pasture" in the NLCD. What we really want are farms that are right next to those medium-high intensity developed areas.
Fortunately we can do buffers, which are perfect for this job. Here I've drawn a 500 meter buffer around those most developed areas, and we're still using the Census urbanized area as an outer boundary to keep us as urban as possible. We can and did play around with exactly how wide you want to make that buffer, but this is what our team ended out going with and it turns out for the Cleveland area we had 201 projects in that buffer in the last 5 years. Bringing up Google maps satellite imagery we check out some of them to make sure we're capturing what we want to be capturing and it looks good. I can't show you the farm locations because our customer information is private, but I can show the resulting national map.
Not too bad.
Many government reports use the Census Bureau's Metropolitan Statistical Areas (MSA), which count any county where at least 25% of the population commutes in to a central city that is large enough. The trouble with MSAs is that they use whole counties as their building blocks and in reality most counties are a mix of urban and rural so as a result you get a very broad definition that includes wide swaths of what people would normally think of as rural. Our first cut was using MSAs, but we found it was indeed too broad, most of what it was capturing was not stuff that would be considered urban agriculture. Fortunately we have another source to help us out, the National Land Cover Database (NLCD) which is a satellite imagery classification scheme that sorts land use patterns visible in the imagery into various categories.
So let's take a look, we have the MSA which is too broad, the city limits which is too narrow. We can see on the Land Cover background in general what we'd like to capture - we want the pink-reddish stuff. But the problem with the NLCD is there is no actual boundary there, we could use it to eyeball a particular farm and say whether it's urban or not, but we want a criteria that we can apply nationwide.
The U.S. Census Bureau to the rescue, they have developed what they call "urbanized areas" which they define as "a continuously built-up area with a population of 50,000 or more." Mapping the one for Clevelend it looks pretty promising and so we run some reports using that boundary. Seeing the results our team decides that it is still too broad, it's mostly capturing farms along the edge of the boundary, not really "urban" enough for what we want to be reporting.
So maybe we can use the NLCD directly. We can still use that Census urbanized area boundary to limit our area of interest. Then on the NLCD we can highlight the "Developed, Medium intensity" and "Developed, High Intensity" areas, which generally translates to commercial, industrial, and multi-unit residential areas that are the core of "urban." So do we have any agriculture on those areas? A little, but it turns out that we've gone too narrow again, since even urban farms usually come up as either the "Culivated Crops" or "Hay/Pasture" in the NLCD. What we really want are farms that are right next to those medium-high intensity developed areas.
Fortunately we can do buffers, which are perfect for this job. Here I've drawn a 500 meter buffer around those most developed areas, and we're still using the Census urbanized area as an outer boundary to keep us as urban as possible. We can and did play around with exactly how wide you want to make that buffer, but this is what our team ended out going with and it turns out for the Cleveland area we had 201 projects in that buffer in the last 5 years. Bringing up Google maps satellite imagery we check out some of them to make sure we're capturing what we want to be capturing and it looks good. I can't show you the farm locations because our customer information is private, but I can show the resulting national map.
Not too bad.
no subject
no subject
no subject
no subject
I'm in Havre de Grace: northeast corner of the map, just before the Susquehanna River. I believe (you can correct me if I'm wrong) that the southern boundary of the urban area on that map is the railroad, the middle portion is Route 40, and the northern boundary is I95. So this confirms what I'd always suspected: that it was the combination of the train and the roads that shaped the southeastern portion of this county.
Here's Harford County in 1795. None of the Harford communities are marked yet, but if you look at the close-up, you can see that there was already a road going from Baltimore over the Susquehanna, through Havre de Grace. There was a ferry at that time.
Here's an 1822 map, sixteen years before the railroad was built. Again, a road through Havre de Grace. (Neither of these maps appears to be entirely accurate about the location of the road, incidentally.) Bel Air is on the map now.
And here's a much more accurate map of Harford County in 1858. There's still no bridge over the Susquehanna; the trains used a ferry at that time. But you can see the effects of the railroad: Aberdeen is starting to grow into a town, and other little towns are growing along the line. The map shows the network of roads connecting the railroad corridor with Bel Air. That hasn't changed, and the railroad still runs along the same tracks.
Dusk (who writes historical pieces, so I spend a lot of time with old maps)
no subject
no subject