1430 Summit Ave, Seattle, WA 98122
Lemur_blog_post1

Population Change Near Nuclear Power Plants

On March 11, 2013 we passed the two year anniversary of the devastating Japan earthquake and subsequent meltdown of the reactors at Fukushima Daiichi. The lessons learned from this tragic event are long and sobering to anyone who cares to read about it. An article in my local newspaper caught my eye, though. The article covered a GAO report questioning the NRC’s understanding of how people who live beyond the 10 mile evacuation zone would react in the event that people within the 10 mile evacuation zone were told to evacuate. The phrase that caught my eye was “Shadow Evacuation”. Would people take flight (on the roads) on their own, clogging vital evacuation routes and potentially delaying or making a quick evacuation impossible?

With five nuclear plants within 50 miles of my house, I noticed some large population change numbers being quoted in the articles. Not that I didn’t believe the numbers, but I decided that this was something I wanted to independently verify using ArcGIS. The question I set out to answer was:

“How has population changed since 1990 within 10, 20, and 50 miles around our nation’s nuclear plants?”

The answer is a lot more complicated than expected due to a number of factors. First, the changing nature of the Tiger/Line census boundary files makes decadal change analysis difficult. Second, data is available on the Census fact finder site but it’s not exactly “user” friendly to extract and helped me understand the value of the datasets published by Esri and others. Third, at what scale can I get the same data more or less in the same place. Fourth, people make a living doing this type of change analysis and now I know why. It has a lot of caveats and always will until the day Census data is published down to the household street address level (never).

In my quest for a comparable census dataset I found the National Historical Geographic information System. I could rant all day about FactFinder, or just save my energy and go here instead. For this analysis I was able to grab population data and GIS boundaries at the tract level for 1990, 2000, and 2010 for the entire US. Thank you very much!

Results

All results are available in this web map published on ArcGIS Online.

The results (xls here) confirm the population around US nuclear plants is growing in absolute terms but not much more than the overall change rate for the entire nation. While the total numbers shouldn’t be used as an exact reference, they are valid in a relative sense (to the census tract level) and likely within a few percentage points if this analysis were done at the census block level. More than half the country lives within 50 Miles of a Nuclear Plant (Table 1). More surprising to me was the big increases in the number of people living within 10 and 20 miles of these nuclear plants.

Table 1: Population Totals
Distance

1990

2000

2010

Percent Change 1990-2010
10 Mile pop

5,387,868

6,107,044

6,854,501

27.22
20 Mile pop

20,551,331

23,204,791

25,628,673

24.71
50 mile pop

154,575,763

169,292,247

180,447,681

16.74
US Total

248,709,873

281,421,906

308,745,538

24.14

Table 2 shows plants where the surrounding population increased by more than 50% (an arbitrary breakpoint). I was then able to take a closer look at the most dramatic population changes that are above the national 24% change since 1990. Surprising to me is that population growth rates exceeding 24% were seen at 28 of the 65 active nuclear plants, and it isn’t just the “Shadow” population that needs review.

Table 2: 10 Mile Population Totals
Name

1990

2000

2010

Percent Change 1990-2010
Shearon Harris

68,720

125,056

171,465

149.51
McGuire 1/2

126,321

209,853

282,210

123.41
Monticello

60,684

77,121

108,969

79.57
Catawba 1/2

162,276

209,505

290,391

78.95
Brunswick 1/2

34,501

52,885

59,982

73.86
Turkey Point 3/4

137,228

153,649

224,043

63.26
Comanche Peak 1/2

36,936

51,327

58,296

57.83
Columbia 2

42,372

61,484

64,444

52.09
St. Lucie 1/2

175,394

219,707

265,886

Looking at the Shearon Harris 10 Mile area (Figure 1) and the 149% growth in population from the 1990 baseline (Table 2), it begs the question of where in the immediate area did this population growth occur?


Figure 1. Shearon Harris 10 Mile population growth.


Figure 2.  Shearon Harris 10 Mile 1990 and 2010 census tracts.

 

To identify areas of growth I symbolized the 1990 census tract polygons in purple outline so we can see the tract boundary changes in comparison to 2010. For the most part, the greatest population changes occurred directly to the east and northeast of the Shearon Harris plant towards Raleigh. Notice all the new and smaller area census tracts? Also, notice the slivers of other census tracts that have been included as part of the 10 mile calculation? I would use the tract centroid if needing higher accuracy on the population numbers since the method I used adds several thousand people. Maybe next time.

Taking a step back to the “Shadow Evacuation” topic, the population change from 1990-2010 within 20 Miles of Shearon Harris increased by 91% (Table 3) or approximately 250K more people, and almost 60% within 50 miles (Table 4) or approximately 1 million more people.

Table 3: 20 Mile Population Totals
Name 1990 2000 2010 Percent Change 1990-2010
Palo Verde 1/2/3 24,169 40,845 72,534 200.11
Shearon Harris 411,287 608,669 786,636 91.26
Dresden 2/3 434,350 593,942 795,598 83.17
St. Lucie 1/2 277,678 351,689 463,953 67.08
North Anna 1/2 133,792 184,030 222,562 66.35
Monticello 192,699 247,854 318,775 65.43
Crystal River 93,712 125,084 152,988 63.25
Brunswick 1/2 134,917 175,283 218,717 62.11
Columbia 2 163,268 201,822 264,463 61.98
Table 4: 50 Mile Population Totals
Name 1990 2000 2010 Percent Change 1990-2010
Palo Verde 1/2/3 1,193,183 1,652,466 2,102,216 76.19
Brunswick 1/2 311,914 413,705 515,682 65.33
Shearon Harris 1,722,699 2,239,310 2,755,530 59.95
Crystal River 760,013 967,017 1,187,323 56.22
St. Lucie 1/2 889,107 1,115,053 1,362,616 53.26
Columbia 2 312,670 394,479 478,384 53.00

Another example of “Shadow” population growth is at the Palo Verde nuclear plant (Figure3). The population grew 200% from 1990 to 2010 within 20 miles (Table 3). While not large in total numbers it still represents a doubling of the population. Looking out to 50 miles, there are more than 1 Million more people in the area (Table 4).


Figure 3.  Palo Verde Nuclear plant 50 mile 2010 census tracts.

 

Conclusion

Checking into one phrase in a newspaper article (“Shadow Evacuation”) led me to a much more interesting and complex analysis than I anticipated. But thanks to some great data provided by NHGIS and some awesome Esri ArcGIS tools (ArcGIS Desktop, ArcGIS Server, ArcGIS Online) I was able to build and share this analysis. I imagine pairing this data up with changes in transportation infrastructure capacity, backed up with studies that analyze maximum traffic flow in these high growth areas. I hope our regulators are doing this kind of analysis and more, planning for the safe evacuation of our citizens in the unlikely event of our own Fukushima Daiichi.