Harvard Forest

Data Archive

HF315

Conservation and Economic Data for New England Towns 1990-2015

Related Publications

Data

• hf315-01: panel data (preview)
• hf315-02: GIS shapefiles

Overview

• Lead: Katharine Sims, Jonathan Thompson, Spencer Meyer, Christoph Nolte, Joshua Plisinski
• Investigators:
• Contact: Information Manager
• Start date: 1990
• End date: 2015
• Status: ongoing
• Location: New England
• Latitude: +41.0 to +47.5 degrees
• Longitude: -74.0 to -67.0 degrees
• Elevation: 0 to 1917 meter
• Datum: WGS84
• Taxa:
• Release date: 2023
• Language: English
• EML file: knb-lter-hfr.315.4
• DOI: digital object identifier
• EDI: data package
• DataONE: data package
• Related links:
• Study type: historical
• Research topic: conservation and management; historical and retrospective studies; regional studies
• LTER core area: disturbance patterns, human-environment interactions
• Keywords: census, conservation, geographic information systems, land cover, land use
• Abstract:

  Land protection, whether public or private, is often controversial at the local level because residents worry about lost economic activity. We used panel data and a quasi-experimental impact-evaluation approach to determine how key economic indicators were related to the percentage of land protected. Specifically, we estimated the impacts of public and private land protection based on local area employment and housing permits data from 5 periods spanning 1990-2015 for all major towns and cities in New England. To generate rigorous impact estimates, we modeled economic outcomes as a function of the percentage of land protected in the prior period, conditional on town fixed effects, metro-region trends, and controls for period and neighboring protection. Contrary to narratives that conservation depresses economic growth, land protection was associated with a modest increase in the number of people employed and in the labor force and did not affect new housing permits, population, or median income. Public and private protection led to different patterns of positive employment impacts at distances close to and far from cities, indicating the importance of investing in both types of land protection to increase local opportunities. The greatest magnitude of employment impacts were due to protection in more rural areas, where opportunities for both visitation and amenity-related economic growth may be greatest. Overall, we provide novel evidence that land protection can be compatible with local economic growth and illustrate a method that can be broadly applied to assess the net economic impacts of protection.

• Methods:

  This dataset was used to quantify the impacts of land protection on local economic activity and is associated with Sims et al. (2019), which described the methods for all analyses. Here we describe the sources of data used in those analyses and the methods used to build the dataset. The data correspond to 1501 county sub-divisions (CSDs) from the Integrated Public Use Microdata Series (IPUMS) National Historic Geographic Information System (NHGIS). We chose to use the IPUMS-NHGIS shapefiles because they ensure accurate matching of census data with the correct vintage shapefile. These CSDs were selected as a subset of all New England CSDs based on the following two criteria: (1) reported populations greater than 100 in the 1990 census; (2) no substantial boundary changes. The data include information for five equally spaced time periods, spanning 1990 to 2015.

  Time series analysis of county sub-divisions is complicated by the fact that they are subject to changes in name and shape over time. CSD boundary shapefiles are updated and re-released periodically by the Census Bureau due to advances in GIS mapping techniques, legal changes in political boundaries, or the incorporation/un-incorporation of towns due to population changes. For datasets where we were free to control the collection unit (land cover and protected areas) we chose to use the 2015 NHGIS CSD Boundary Shapefiles to represent the spatial extent of each CSD. For the Census data, we were constrained by the boundary files used at the time of collection. The default collection unit for the ACS 2011-2015 matched our choice of the 2015 NHGIS CSD Boundary Shapefile however the ACS 2006-2010 and Nominal Decadal Time Series are based on the 2010 NHGIS CSD Boundary Shapefile, thus necessitating a methodology to accurately incorporate data based on the 2010 shapefile into the 2015 shapefile. Within the NHGIS Boundary Shapefiles, the GISJOIN code is used as the unique identifier for each CSD. Fortunately, there are only four CSDs that change GISJOIN codes between the 2010 and 2015 NHGIS CSDs shapefiles. While their GISJOIN codes change, visual inspection of the shapefiles verified that their boundaries do not change. This makes it possible to simply update/replace the 2010 GISJOIN codes in the ACS 2006-2010 and Nominal Decadal Time Series with the matching 2015 code. Marshall Island UT Maine changes from G230009043475 to G230009043578, Sanford city Maine changes from G230031065760 to G230031065725, Enosburgh town Vermont changes from G500011023875 to G500011024050 Alburgh town Vermont changes from G500013000700 to G500013000860, Penobscot Indian Island Reservation Maine changes from G230019057937 to G230019057936. Thus result is an analysis ready dataset that incorporates historical census data into the most current 2015 representation of the CSD shapefile.

  The percentage of land protected in each town in each time period comes from a regionally aggregated dataset of protected areas maintained by the Highstead Foundation and the Harvard Forest. Sources for this spatial data include:

  1. The Nature Conservancy. 2015. U.S. Eastern Division Secured Lands. easterndivision.s3.amazonaws.com/SecuredAreas2015public.zip

  2. The Trust for Public Land and Ducks Unlimited. 2018. National Conservation Easement Database. https://www.conservationeasement.us/downloads/

  3. Conservation Biology Institute. 2012. Protected Areas Database of the US (CBI Edition) Version 2. https://consbio.org/products/projects/pad-us-cbi-edition

  4. Vermont Center for Geographic Information. 2017. Vermont Protected Lands Database. http://geodata.vermont.gov/datasets/072bb8ad3c454b0e9cb0f517e9a296a3_10

  5. MassGIS and Executive Office of Energy and Environmental Affairs. 2018. Protected and Recreational OpenSpace. https://docs.digital.mass.gov/dataset/massgis-data-protected-and-recreational-openspace

  6. Connecticut Department of Energy and Environmental Protection and Applied Geographics, Inc. 2011. Protected Open Space Data. ftp://ftp.ct.gov/dep/gis/shapefile_format_zip/Protected_Open_Space_shp.zip

  7. Earth Systems Research Center, University of New Hampshire. 2018. GRANIT: New Hampshire Conservation/Public Lands at 1:24,000 Scale. http://www.granit.unh.edu/data/search

  8. MEGIS. 2018. Maine Conserved Lands. https://geolibrary-maine.opendata.arcgis.com/datasets/a6797f12a07b4229bc2501d3741c98d4_0

  9. U.S. Geological Survey, Gap Analysis Program (GAP). May 2016. Protected Areas Database of the United States (PAD-US), version 1.4 Combined Feature Class. https://gapanalysis.usgs.gov/padus/data/download/

  10. RIGIS. 2014. Municipal and Non-Governmental Organization Conservation Lands; locCons14. Rhode Island Geographic Information System (RIGIS) Data Distribution System, Environmental Data Center, University of Rhode Island, Kingston, Rhode Island. http://rigis.org

  11. RIGIS. 2014. State Conservation Lands; staCons14. Rhode Island Geographic Information System (RIGIS) Data Distribution System, Environmental Data Center, University of Rhode Island, Kingston, Rhode Island. http://rigis.org

  All land cover area tabulations were made using the 2015 version of the Census County Sub Divisions shapefile. Land cover information is derived from two Landsat derived data sources: Continuous Change Detection and Classification (CCDC) data (Olofsson et al 2016) and National Land Cover Database (NLCD) data (Homer et al 2015). CCDC is an annual product covering the majority New England excluding northwest Vermont and northeast Maine. We considered it to be the more accurate data source and used it first where available. Where CCDC data was not available, we filled the remainder of the study area with the NLCD product that best matched our target year. Additionally, CCDC data contains dispersed NoData values for areas where the land cover classification algorithm was unable to determine a stable land cover class. These areas were also filled with NLCD values from the closest NLCD product date.

  Data on building permits comes from the United States Census Bureau, Building Permits Survey https://www2.census.gov/econ/bps/Place/Northeast%20Region/. We have aggregated the data into five-year increments. Places are identified by name but geographic identifiers were not always consistent across time. This was resolved by matching on name and then hard coding several places where the names of towns changed and verifying that boundaries remained consistent across time or when smaller reporting entities were combined into larger entities.

  Data on local area unemployment comes from the United States Bureau of Labor Statistics, Local Area Unemployment Statistics https://www.bls.gov/lau/lauov.htm. We aggregate the annual versions of the datasets into five year averages from 1990 to 2015. Data were joined to the 2015 CSD shapefile using the ‘geoid’ field, with some hard coding corrections for towns in CT.

  Literature Cited:

  Homer C.G., Dewitz J.A., Yang L., Jin S., Danielson P., Xian G., Coulston J., Herold N.D., Wickham J.D., and Megown K., 2015. Completion of the 2011 National Land Cover Database for the conterminous United States-Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v. 81, no. 5, p. 345-354

  Olofsson P., Holden C.E., Bullock E.L., and Woodcock C.E. 2016. Time series analysis of satellite data reveals continuous deforestation of New England since the 1980s. Environmental Research Letters. IOP Publishing.

  Sims, K. R. E., Thompson, J. R., Meyer, S. R., Nolte, C., Plisinski, J. S. 2019. Assessing the local economic impacts of land protection. Conservation Biology 33: 1035–1044.

• Organization: Harvard Forest. 324 North Main Street, Petersham, MA 01366, USA. Phone (978) 724-3302. Fax (978) 724-3595.

• Project: The Harvard Forest Long-Term Ecological Research (LTER) program examines ecological dynamics in the New England region resulting from natural disturbances, environmental change, and human impacts. (ROR).

• Funding: National Science Foundation LTER grants: DEB-8811764, DEB-9411975, DEB-0080592, DEB-0620443, DEB-1237491, DEB-1832210.

• Use: This dataset is released to the public under Creative Commons CC0 1.0 (No Rights Reserved). Please keep the dataset creators informed of any plans to use the dataset. Consultation with the original investigators is strongly encouraged. Publications and data products that make use of the dataset should include proper acknowledgement.

• License: Creative Commons Zero v1.0 Universal (CC0-1.0)

• Citation: Sims K, Thompson J, Meyer S, Nolte C, Plisinski J. 2023. Conservation and Economic Data for New England Towns 1990-2015. Harvard Forest Data Archive: HF315 (v.4). Environmental Data Initiative: https://doi.org/10.6073/pasta/a43538dd84cf4ed74bd080f56f042fdf.

Detailed Metadata

hf315-01: panel data

1. town.id: short identification variable for county sub-divisions
2. gis.name: county sub-division name
3. gis.join: identification variable for GIS dataset
4. year: five-year panel period
5. cbsa: metro region
6. state.name: state
7. co.pre100pct: percent of town protected, all types (unit: dimensionless / missing value: NA)
8. co.pub.pre100pct: percent of town protected, public (unit: dimensionless / missing value: NA)
9. co.priv.pre100pct: percent of town protected, private (unit: dimensionless / missing value: NA)
10. co.bwf.pre100pct: percent of town protected, large protected timber land (unit: dimensionless / missing value: NA)
11. nn5.co.pre100pct: percent protected in five nearest neighbors (unit: dimensionless / missing value: NA)
12. nn10.co.pre100pct: percent protected in ten nearest neighbors (unit: dimensionless / missing value: NA)
13. co.pre100: log percent of town protected, public (unit: dimensionless / missing value: NA)
14. co.pub.pre100: log percent of town protected, public (unit: dimensionless / missing value: NA)
15. co.priv.pre100: log percent of town protected, private (unit: dimensionless / missing value: NA)
16. co.bwf.pre100: log percent of town protected, large protected timberland (unit: dimensionless / missing value: NA)
17. nn5.co.pre100: log percent protected, 5 nearest neighbors (unit: dimensionless / missing value: NA)
18. nn10.co.pre100: log percent protected, 10 nearest neighbors (unit: dimensionless / missing value: NA)
19. l5.co.pre100: 5 year lag of log percent protected, all types (unit: dimensionless / missing value: NA)
20. l5.co.pub.pre100: 5 year lag of log percent protected, public (unit: dimensionless / missing value: NA)
21. l5.co.priv.pre100: 5 year lag of log percent protected, private (unit: dimensionless / missing value: NA)
22. l5.co.bwf.pre100: 5 year lag of log percent protected, large protected timberlands (unit: dimensionless / missing value: NA)
23. l5.nn5.co.pre100: 5 year lag of log percent protected, 5 nearest neighbors (unit: dimensionless / missing value: NA)
24. l5.nn10.co.pre100: 5 year lag of log percent protected, 10 nearest neighbors (unit: dimensionless / missing value: NA)
25. unemp.r.pre: average annual unemployment rate for five years prior (unit: dimensionless / missing value: NA)
26. emp.n.pre: average number people employed for five years prior (unit: number / missing value: NA)
27. labf.n.pre: average number people in labor force five years prior (unit: number / missing value: NA)
28. i.units.pre: total housing units permitted, imputed (unit: number / missing value: NA)
29. r.units.pre: total housing units permitted, reported (unit: dimensionless / missing value: NA)
30. ln.unempr.pre: log average unemployment rate for five years prior (unit: dimensionless / missing value: NA)
31. ln.emp.n.pre: log average number people employed for five years prior (unit: number / missing value: NA)
32. ln.labf.n.pre: log average number people in labor force five years prior (unit: number / missing value: NA)
33. ln.r.units.pre: log total housing units permitted, reported (unit: number / missing value: NA)
34. mhhi.ia.pre: median household income, inflation adjusted (unit: number / missing value: NA)
35. ln.mhhi.ia.pre: log median household income, inflation adjusted (unit: number / missing value: NA)
36. agricu.pre: share employment in agriculture/forestry/fishing/hunting/mining (unit: dimensionless / missing value: NA)
37. arts.pre: share employment in recreation/arts/entertainment/accommodation/food (unit: dimensionless / missing value: NA)
38. ln.agricu.pre: log share employment in agriculture/forestry/fishing/hunting/mining (unit: dimensionless / missing value: NA)
39. ln.arts.pre: log share employment in recreation/arts/entertainment/accommodation/food (unit: dimensionless / missing value: NA)
40. popcen.pre: population (unit: dimensionless / missing value: NA)
41. ln.popcen.pre: log population (unit: dimensionless / missing value: NA)
42. for.pct: percent forested land use (unit: dimensionless / missing value: NA)
43. agg.pct: percent in agricultural land use (unit: dimensionless / missing value: NA)
44. hiden.pct: percent high density land use (unit: dimensionless / missing value: NA)
45. loden.pct: percent low density land use (unit: dimensionless / missing value: NA)
46. other.pct: percent other types land use (unit: dimensionless / missing value: NA)
47. ndata.pct: percent of town missing land use data (unit: dimensionless / missing value: NA)
48. urban90: whether or not large urban, more than 128 housing units/sq km and more than 30,000 people in 1990
    • 0: not large urban
    • 1: large urban
49. suburb90: whether or not small urban, more than 128 housing units/sq km and fewer than or equal to 30,000 people in 1990
    • 0: not small urban
    • 1: small urban
50. exurb90: whether or not exurban, 16-128 housing units/sq km
    • 0: not exurban
    • 1: exurban
51. rural90: whether or not rural, few than 16 units/sq km
    • 0: not rural
    • 1: rural
52. ctydist100: distance in km to city with more than 100K people 1990 (unit: kilometer / missing value: NA)
53. ctydist30: distance in km to city with more than 30K people 1990 (unit: kilometer / missing value: NA)
54. ln.ctydist100: log distance in km to city with more than 100K people 1990 (unit: kilometer / missing value: NA)
55. ln.ctydist30: log distance in km to city with more than 30K people 1990 (unit: kilometer / missing value: NA)
56. ln.pop.cen90: log population in 1990 (unit: dimensionless / missing value: NA)

hf315-02: GIS shapefiles

• Compression: zip
• Format: Esri shapefile
• Type: vector GIS