You are here

Harvard Forest Data Archive


Forest Damage Patterns at Harvard Forest in the 1938 Hurricane

Related Publications



  • Lead: David Foster, Emery Boose
  • Investigators:
  • Contact: Information Manager
  • Start date: 1936
  • End date: 1941
  • Status: completed
  • Location: Tom Swamp Tract (Harvard Forest), Petersham MA
  • Latitude: +42.42 to +42.55
  • Longitude: -72.29 to -72.10
  • Elevation: 160 to 395 meter
  • Taxa:
  • Release date: 1999
  • Revisions:
  • EML file: knb-lter-hfr.9.16
  • DOI: digital object identifier
  • EDI: data package
  • DataONE: data package
  • Related links:
  • Study type: historical, modeling
  • Research topic: ecological informatics and modelling; historical and retrospective studies; regional studies
  • LTER core area: disturbance
  • Keywords: geographic information systems, hurricane damage, landscape, modeling, wind
  • Abstract:

    This study examined landscape-level patterns of forest damage at the Harvard Forest caused by the 1938 New England Hurricane. For details on methods and results, please see the published paper (Foster, D. R. and E. R. Boose. 1992. Patterns of forest damage resulting from catastrophic wind in central New England, USA. Journal of Ecology 80: 79-98). The Abstract from the paper is reproduced below.

    "1. The effect of catastrophic winds on a forested landscape in central Massachusetts was examined to investigate the factors controlling the geographic pattern of damage. The study area, Tom Swamp Tract, Harvard Forest, comprises a valley and adjoining hillslopes supporting second growth hardwood and confer stands. Much of the study used records and maps that were analyzed cartographically with a geographic information system (GIS).

    "2. Areally, forest damage was distributed fairly evenly among different damage classes ranging from no damage to more than 75% of stems broken or uprooted. However, there was a negative exponential size distribution of contiguous areas of the same damage intensity, with a preponderance less than 2 ha; these areas ranged from less than 0.04 ha to more than 35 ha; hurricane damage exhibited a continuum ranging from minor damage of individual trees to extensive blow-down of broad areas of forest.

    "3. The spatial pattern of wind damage was controlled by vegetation height and composition and by site exposure, which is predominantly determined by slope orientation and angle. Approximately 3% of the stands in the study site occupied protected sites, 31% intermediate sites, and 66% exposed sites.

    "4. Forest type susceptibility followed the ranking (from highest to lowest): Pinus strobes, conifer plantations, Pinus strobus-hardwood = Tsuga canadensis-hardwood-Pinus strobes, hardwood-Pinus strobes, hardwood. Damage increased with increasing site exposure to wind and increased approximately linearly with stand height.

    "5. An empirical GIS model of landscape-level response to wind was constructed based on other stands in the same township (not including the Tom Swamp Tract). Hurricane damage in these stands was analysed as a function of site factors (exposure) and vegetational factors (height and composition). Model predictions for the study area agreed well with observed effects, suggesting that a relatively small number of variables can be used to explain the damage in this topographically simple area. Significant variation in the predicted damage under different vegetational scenarios suggests that the landscape-level response to catastrophic wind may be highly sensitive to historical changes in vegetation."

  • Methods:

    The original study was completed using the PC version of the Map Analysis Package (MAP). Files are archived in Idrisi raster binary format (local coordinates, resolution = 20 m). See the Methods section and published paper for more details.

    ASP$9 = Topographic aspect in 9 classes (1 = N, 8 = NW, 9 = level). Derived from ELE-M using the MAP Orient command.

    DAM$5 = Actual 1938 Hurricane damage by stand in 5 classes (1 = no damage, 5 = greater than 75% damage). Recoded from STA!.

    DRA$5 = Soil drainage in 5 classes (1 = very well, 5 = very poor). Recoded from a soils map in 28 classes digitized from a paper map in the Harvard Forest Archives.

    ELE-M = Elevation in meters. Digitized from the USGS topographic map.

    EXP$3C = Predicted topographic exposure to Southeast winds in 3 classes (1 = protected, 2 = intermediate, 3 = exposed). Values were calculated on a cell by cell basis as follows: Protected = W, NW, or N aspect and slope greater than 10 degrees; Exposed = E, SE, or S aspect or slope less than 5 degrees; Intermediate = all others.

    FOR$7 = 1938 forest types by stand in 7 classes. Recoded from STA!.

    HEI$11 = 1938 forest height by stand in 11 classes (1 = less than 3 m, 11 = more than 30 m). Recoded from STA!.

    SLO! = Topographic slope in degrees. Derived from a wider version of ELE-M (to minimize boundary problems) using the MAP Differentiate command.

    STA! = 1938 forest stands (1-239; missing = 9, 69, 131, 132, 164, 194). Digitized from a paper map in the Harvard Forest Archives.

  • 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.

  • Citation:

    Foster D, Boose E. 1999. Forest Damage Patterns at Harvard Forest in the 1938 Hurricane. Harvard Forest Data Archive: HF009 (v.16). Environmental Data Initiative:

Detailed Metadata

HF009-01: Tom Swamp GIS

  • Compression: zip
  • Format: Idrisi raster
  • Type: raster GIS