credit: National Geographic
**EDIT: I've updated these and redone the model to have more consistent assumptions and methods. See it here: http://statchatva.org/2014/03/27/turning-population-projections-into-development-projections/
The Weldon Cooper Center at UVA recently released a series of population projections for Virginia. According to the Center, "Empirical studies show the average error for 30-year projections at the county level is 36%." Some of this uncertainty comes from the fact that it's impossible to predict major events or trends - like the growth of the Federal government in the second half of the 20th century that dramatically changed Northern Virginia.
I'm not particularly concerned with the accuracy of the projections here. My goal is to see what Virginia might look like under this growth - not in terms of numbers of people in a county - but in terms of urbanized area. How far will the sprawl go? See the notes on methodology at the bottom if you're curious about how I went about this.
Here is a land cover raster of Virgina today, showing rural areas in light green (forest) and light yellow (agriculture). Developed areas of any kind - roads, buildings, parking lots, etc. are shown in red. It's a great image for seeing the exact extent of metropolitan areas.
Now - what will this map look like in 25 years if:
A) The Weldon Cooper Center's projections (at the regional level) are roughly correct and
B) We continue to build at the same low suburban densities that have been dominant for the past several decades and
C) We don't invest in any major game-changing roads (other than the Western Bypass in Charlottesville)
Here's Virginia in 2040 (click on the images to see them as a slideshow for easy comparison):
I've tried to distribute the development among metro areas roughly according to the Center's projections. My weighting is still off I'm sure (see methodology below). But rather than show it increasing by a fixed amount in each county, I've shown it emanating out from each urban core in rings that represent driving time.
You could call this map the "New Jersey-fication of Virginia." New Jersey also sprawled along a major interstate corridor between large metro areas to become what it is today. This map shows us joining the Boston-to-Washington megalopolis of continuous urban areas.
Here's another thought experiment... in the past 50 years, Virginia's population doubled and that rate of growth has continued. What if it were to double again in the next 50 years at current densities?
The second reason I don't think this is likely is that we seem to have finally turned a corner when it comes to demand for low-density suburbs. Municipalities are realizing the infrastructure doesn't pay for itself. People are rediscovering the joys of living "in town," and for the first time since World War II, we are driving less than we did five years ago. All that said, it's helpful to see what the place could look like.
Methodology:
I first generated a random raster to approximate the randomness of development and used it as my base. I then used service areas and the VGIN road network to add weights to the random raster based on distance from a metro area. I eyeballed this part to try to match the proportions of the Weldon Cooper Center projections and the general economic status of each area. The weights are by no means exact. I then used a shapefile of conservation easements and publicly conserved lands to subtract these areas from the development raster, along with all currently developed areas. What I ended up with was a raster of undeveloped but developable land with each pixel weighted according to its likelihood. I then used averages from other developed areas to estimate the number of pixels each new person would add. Then I overlayed the development raster onto the original and turned the highest scoring pixels to red and the others to invisible. My threshold for which pixels to show was based on the number of pixels that cleared the threshold and the approximate number that would be developed in each scenario.
