Monday, March 26

New Census Geographies Tell an Ambiguous Urban Story

The U.S. Census Bureau announced today that the U.S. population is more urbanized than ever before. In 2000, 79.0% of the population lived in an Urban Area, as defined by the Census Bureau, and in 2010 that number grew to 80.7% of all Americans. While this finding might suggest that Americans are moving into the kinds of urban places featured on this blog – walkable neighborhoods with a diversity of uses in close proximity - the full story is much more nuanced. "Urban" can mean many things.

While the urbanized population grew between 2000 and 2010, the area of land considered urbanized grew by even more. This requires some explanation of the terms. The Census determines two types of Urban Areas, using a complex set of mapping criteria based on neighborhood-scale population density, impervious surfaces, and other factors. There are “Urbanized Areas” of 50,000 people or more, and “Urban Clusters” of 2,500 people or more. These are important geographies because, unlike metropolitan areas or incorporated cities, they are based on the results of the census. They are not arbitrary lines on the map but a description of the facts on the ground.

Before looking at some results, I'd like to note that the minimum density threshold to qualify an area as urban is not all that high. An area must have core Census Tracts with at least 1,000 people per square mile, and the urbanized area can spread out indefinitely as long as there are contiguous areas with 500 people per square mile. It can even "jump" and "hop" over sparsely populated land to continue along a corridor. As a point of reference, a neighborhood of single-family houses on two acres lots would qualify as "urban" under this definition, even considering roadways and other leftover spaces.

Urbanized Areas (the cities, not the towns) increased in population by 13.3% between 2000 and 2010, but they also increased in land area by 18.5%. This means the population density of all Urbanized Areas in the United States dropped by 4.4%. This fact is particularly troubling, because Urbanized Area population densities actually increased by 1.9% between 1990 and 2000 (the methodology for computing urban areas changed considerably between 1990 and 2000, but this number comes from applying the 2000 methodology back to 1990). It’s probably more accurate to say that Americans didn’t become more urbanized in the last decade, but that we witnessed a convergence. The Urban Areas grew, but became less urban in the process.

Of course, the changes in density happened differently throughout the country, and neighborhood-scaled density arranged itself much differently inside each urban area. Here are some top ten lists from the new census geographies. In each case, density is measured in People per Square Mile for the whole Urbanized Area:


Top Ten Densest Urbanized Areas (over 50,000) in 2010

Rank
Name
Population
Area
Density
Change
2000 Rank
1
Los Angeles--Long Beach--Anaheim, CA
12,150,996
1,736.02
 6,999
 (69)
1
2
San Francisco--Oakland, CA
3,281,212
523.62
 6,266
 136
3
3
San Jose, CA
1,664,496
285.98
 5,820
 (94)
4
4
Delano, CA
54,372
9.92
 5,481
(1,182)
2
5
New York--Newark, NY--NJ--CT
18,351,295
3,450.20
 5,319
 10
6
6
Davis, CA
72,794
14.12
 5,155
 312
8
7
Lompoc, CA
51,509
10.70
 4,814
 3,888
487
8
Urban Honolulu, HI
802,459
170.17
 4,716
 56
10
9
Woodland, CA
55,513
12.20
 4,550
(1,095)
5
10
Las Vegas--Henderson, NV
1,886,011
416.84
 4,525
 (73)
12

 The California cites remain the densest in the county by this measurement. Either because of water supply, topography, federal land, or local land-use regulations, many these cities have sharp edges that tend to maximize the density of development within the urban boundaries. We see here New York lagging behind LA in urban area density, although, as I've posted before, the average person in NY lives in a neighborhood about two and a half times as dense as the average person in LA. The scale of the density measurement will give different results.



Top Ten Decreases in Density (over 50,000) between 2000 and 2010

Rank
Name
Pop
Change
Area
Change
Density
Change
1
New Orleans, LA
899,703
(109580)
251.39
53.55
 3,579
 (1,523)
2
Delano, CA
54,372
14860
9.92
3.99
 5,481
 (1,182)
3
Woodland, CA
55,513
6345
12.20
3.49
 4,550
 (1,095)
4
Port Arthur, TX
153,150
38494
105.56
59.70
 1,451
 (1,049)
5
Fairfield, CA
133,683
21237
39.46
13.64
 3,388
 (967)
6
Grand Forks, ND--MN
61,270
4697
24.44
7.68
 2,507
 (869)
7
Lewiston, ME
59,397
8830
35.40
14.76
 1,678
 (772)
8
Vallejo, CA
165,074
6107
42.01
8.06
 3,929
 (753)
9
Provo--Orem, UT
482,819
179139
168.97
83.97
 2,857
 (715)
10
San Luis Obispo, CA
59,219
5721
20.41
5.58
 2,901
 (706)

New Orleans, of course, suffered dramatic population loss in the wake of hurricane Katrina, but note that the city also expanded in land area by about 25% over the decade. New Orleans was the 7th densest Urbanized Area in the country in 2000, right behind New York, but a natural disaster interacting with suburban sprawl has knocked it down to 31st, just ahead of Denver. We could be seeing the results of a different kind of disaster in some of the California cities on this list, many of which were hit hard by foreclosures. Extensive home construction over the decade may have increased the urbanized area without a corresponding increase in people to live in those houses. Provo-Orem, Utah has a much more straight-forward explanation: lots of growth, all spreading outward at very low density.

Top Ten Increases in Density among Urbanized Areas (over 50,000) between 2000 and 2010
 
Rank
Name
Population
Change
Area
Change
Density
Change
1
Lompoc, CA
51,509
-4,158
10.70
-49.40
 4,814
 3,888
2
Sierra Vista, AZ
52,745
5,804
30.28
-50.16
 1,742
 1,158
3
Santa Maria, CA
130,447
10,150
29.13
-6.42
 4,478
 1,094
4
Victoria, TX
63,683
2,154
29.02
-21.99
 2,194
 988
5
Corvallis, OR
62,433
4,204
21.11
-7.87
 2,958
 948
6
Madera, CA
78,413
20,386
22.39
-0.22
 3,502
 936
7
Lodi, CA
68,738
-14,997
15.91
-7.63
 4,320
 763
8
San Diego, CA
2,956,746
282,310
732.41
-49.88
 4,037
 618
9
Colorado Springs, CO
559,409
93,287
187.84
-9.52
 2,978
 616
10
Lady Lake, FL
112,991
62,270
71.07
21.09
 1,590
 575

Lompoc, CA and Sierra Vista, CA are probably flukes in the counting method. The urban areas for these places in 2000 were very odd, and the 2010 calculations show much more constricted, and probably more accurate, areas. San Diego certainly has a legitimate claim to an increase in density, but the fact that the Camp Pendleton Marine Corp base was dropped from the Urban Area between 2000 and 2010 gives it an artificial boost. The only other big cities to show notable increases in density were Portland, OR (5.6%) and Seattle (6.5%). Although it's just short of making the top ten, the population density of Charlottesville, Virginia grew by 23% over the decade to a respectable 2,672 people per square mile.

Top Ten Densest Urban Clusters (2,500 > = < 50,000) in 2010

Rank
Name
Population
Area
Density
1
Richgrove, CA
2867
0.29
       10,016
2
Parlier, CA
14490
1.58
         9,162
3
Mattawa, WA
4988
0.62
         7,985
4
San Luis, AZ
24091
3.57
         6,755
5
Santa Paula, CA
29742
4.42
         6,735
6
Mountain House, CA
9616
1.43
         6,703
7
Guadalupe, CA
7080
1.06
         6,689
8
Orange Cove, CA
9774
1.51
         6,471
9
Soledad, CA
25943
4.14
         6,269
10
Greenfield, CA
16451
2.74
         5,997

We can't forget the small towns. California towns with heavy agriculture dominate this list, and with some very high levels of density.

Now we'll leave population behind, and look just at the increasing amount of land used for human habitation.

 Top Ten Urbanized Areas (over 50,000) with Greatest Increase in Area between 2000 and 2010

Rank
Name
Area
Change
% Change
1
Atlanta, GA
2,645.35
682.77
34.79
2
Dallas--Fort Worth--Arlington, TX
1,779.13
372.09
26.45
3
Houston, TX
1,660.02
364.75
28.16
4
Phoenix--Mesa, AZ
1,146.57
347.56
43.50


5
Chicago, IL--IN
2,442.75
319.94
15.07
6
Charlotte, NC--SC
741.49
306.58
70.49


7
Austin, TX
523.03
204.90
64.41
8
Raleigh, NC
518.14
198.53
62.11


9
San Antonio, TX
597.10
189.54
46.51
10
Philadelphia, PA--NJ--DE--MD
1981.37
181.86
10.11




Maps of some selected urbanized areas paint a picture of the bookends of urban change over the decade:

Dark purple = in 2000 and 2010 urban areas
Light purple = new for 2010


Atlanta had the greatest expansion of urban land over the decade, with no less than 683 square miles of growth.


Charlotte had the fastest population growth of any urbanized area, but the expansion of land was even faster at a 70% rate of growth.


After several decades of rapid expansion, the Phoenix-Mesa urbanized area increased by another 43% between 2000 and and 2010. Atlanta had much more growth and Charlotte had a much higher growth rate, but Phoenix held it's own on both measures, perhaps adding more ammunition to those who consider this desert oasis to be the world's least sustainable city.


It was a surprise to me to see the older Philadelphia ranked as an expanding urbanized area.


The urbanized area of Austin kept pace with Dallas-Fort Worth and Houston in terms of total land growth. But it's rate of land expansion was far higher than the large Texas metropolises.

And now for some more positive news:

Dark blue = in both 2000 and 2010 urban areas
Grey = new for 2010
Light blue = in 2000, but not 2010 urban areas


All of the Bay Area urbanized areas held steady on expansion, and the urban area even shrunk in Marin County and some spots in Silicon Valley. Population growth in San Francisco-Oakland was not particularly strong, but the San Jose urbanized area grew by a moderate 8.2% over the decade.

San Diego showed little outward expansion, while absorbing an additional 282 thousand people. The official number was even negative, but the drop was caused by the removal of a military base from the urban area.


Colorado Springs is an especially interesting case. The urbanized population grew by an impressive 21% over the decade with very minimal expansion of its footprint.

This decade will remain very tricky to characterize as a whole, because of the sharp divide in development rates that struck in 2008. Because the census was taken in April of 2010 and construction always lags behind financing, most of the growth shown here probably represents the pre-recession reality of homebuilding and commercial expansion.  We only have such thorough snapshots of urban growth once per decade, so discerning the future, in this case, may come down to reading between the lines.