DEMOGRAPHIC IMPACTS OF NATURAL DISASTERS

Transcription

1 DEMOGRAPHIC IMPACTS OF NATURAL DISASTERS Originally presented at the Urban Affairs Association Annual Conference Pittsburgh, PA April 2012 Shannon Van Zandt, AICP Walter Gillis Peacock Painting by Debra Hicks

2 Disasters interrupt normal forces of population change 4 STUDY PURPOSE 1. Develop a valid method for assessing the current population a. Identify and assess useful secondary data sources b. Develop quick and efficient methods of primary data collection when needed 2. Assess demographic changes resulting from the disaster

3 Literature Existing studies depend on: Decennial census data (e.g., Wright et al. 1979) Ad-hoc and incomplete third-party estimates (Geipel 1989; Haas et al. 1977; Kimball & Bolton 1994) Qualitative interviewing (McCarthy et al. 2006) Administrative data (Plyer, Bonguro, and Hodges 2010) We improve on seminal work by demographer Stanley K. Smith and colleagues (1996) by: Exploring additional secondary and spatial data sources Disaggregating data spatially and by housing type Considering secondary sources to assess demographic composition

4 Typical methods are rendered inappropriate Hurricane Ike damage on Galveston Island, by simminch at flickr.com Housing Unit Method Population = ( (HU* OCC) *PPH)+ GQ Housing Units may be destroyed or rendered uninhabitable Occupancy rates are altered by displacement and dislocation Persons Per Household may be altered by doubling up during evacuation and recovery Group Quarters may include long-term shelters

5 Understanding Galveston s Geography West End In 2000, contained less than 10% of population 6,551 (18%) of the island s housing units Occupancy Rate = 47% Persons Per Household (PPH) = 2.22 Urban Core Considerably more multi-family housing 25,375 housing units Occupancy Rate = 85% PPH = 2.34

6 Adaptation of the HU Method

7 Data sources evaluated Secondary: 1990 and 2000 Decennial Census American Community Survey US Postal Service delivery data Galveston ISD enrollment data (for demographic composition only) Primary: Damage assessments from TAMU s Hazard Reduction & Recovery Center ( ) Multi-family survey conducted by students

8 Single-family Housing Units (HU) Source American Community Survey (ACS) ACS , after having applied HRRC Destruction Rate Urban Core West End Citywide Census ,270 4,412 32,510 Census ,375 5,473 30,848 n/a n/a 33,439 n/a n/a 32,523 US Postal Service (USPS) 4Q ,858 4,041 31,899 Average of USPS Quarters since Ike Housing unit data shows good consistency among data sources USPS appears to be a sound source for the number of housing units 27,752 3,937 31,689

9 Occupancy Rates (OCC) Source Urban Core West End Citywide U.S. Census, % 43.6% 75.6% U.S. Census, % 47.2% 76.1% ACS, n/a n/a 67.9% USPS, mean of 4Q 2008 to 4Q % 87.7% 92.7% USPS, 4Q % 88.0% 92.8% Primary Damage Assessment, 12/ % 38.0% 45.7% Primary Damage Assessment, 01/ % 58.0% 66.6% USPS vacancy rates proved unreliable, because it does not differentiate between owner-occupied and renter-occupied units nor does it identify seasonal or recreational units Primary data collection includes single-family only ACS data was consistent with findings from primary data collection

10 Persons per household (PPH) Source Urban Core West End Citywide U.S. Census, U.S. Census, U.S. Census ACS Survey, n/a n/a 2.21 Primary Damage Assessment, 12/ Primary Damage Assessment, 01/ Data shows evidence for doubling up in the months after the storm Evidence suggests differences in PPH from West End to Urban Core

11 Multi-family survey Nature of damage (flooding) affected multi-family properties differently No reliable source for multi-family occupancy rates was found Graduate students (or volunteers) were able to complete the work Housing Unit Units with Average Occupied Category known Occupancy Units Occupancy Rate 3 or 4 units % 5 to 9 units % 10 to 19 units % 20 or more units 6,579 4,900 74%

12 Final population data Sources Single-Family Units Multi-Family Units Scenario HU OCC PPH HU OCC PPH ACS Primary Primary Low In February (1) of 2011, final census MF survey counts MF for survey ACS (adjusted) survey survey Galveston ACS indicated Primary a population Primary of ACS 47,743, just Medium (2) MF survey 630 fewer (adjusted) residents survey than survey our estimate (adjusted) ACS Primary Primary ACS High (3) USPS MF survey survey survey (adjusted) ACS Population Estimate Single Family Estimate Multi-Family Estimate Total Population Estimate Scenario 1 32,059 14,007 46,066 Scenario 2 32,059 16,314 48,373 Scenario 3 32,131 16,314 48,445

13 Distribution of Enrolled Students Bolivar Galveston Mainland Location Findings: little evidence to suggest that racial and ethnic 1% 25% 39% 46% 42% groups 51% had shifted from neighborhood to 25% neighborhood on the island 19% overall loss of population appears to have hit some neighborhoods harder than others loss of public housing units on the island s far east Number of Students Enrolled (Percent) 35% Students that Qualify for Free or Reduced Lunch (Percent) end (one of the most physically vulnerable areas) appears to have displaced large numbers of African- American households Bolivar 896 (15%) 748 (17%) Galveston 4,670 (80%) 3,569 (79%) Mainland 283 (5%) 174 (4%) Total 5,849* 4,491 * Other category not shown, and 1% of enrollment unable to map

14 Conclusions Secondary data sources are widely available, but vary in their reliability and validity Spatial disaggregation of data can result in more accurate estimates Disaggregation of data by housing type can uncover important differences due to difference in the nature of the damage Impacts of sudden and dramatic declines in population can change community dynamics Disadvantaged populations are likely to be disproportionately impacted by natural disasters, due to an interaction between increased exposure to damage and higher levels of social vulnerability

15 Acknowledgements