You are here

HF122 Detailed Metadata

Printer-friendly version

1830 Massachusetts Documentation

These data are provided for internal use only and are not for distribution or publication; all requests for use of the data should be directed to Harvard Forest.

Please cite these datalayers as follows:

Harvard Forest. 2002. 1830 Map Project. Harvard Forest Archives, Petersham, MA.

Hall, B., G. Motzkin, D. R. Foster, M. Syfert, and J. Burk. 2002. Three hundred years of forest and land-use change in Massachusetts, USA. Journal of Biogeography 129: 1319-1135.

Acknowledgements:

Financial support for this project was provided by the BioMap Project of the Massachusetts Natural Heritage and Endangered Species Program, the National Science Foundation, and the Andrew Mellon Foundation. This project is a contribution of the Harvard Forest Long Term Ecological Research Program.

Background and Data Limitations:

The Massachusetts 1830 map series represents a unique data source that depicts land cover and cultural features during the historical period of widespread land clearing for agricultural. To our knowledge, Massachusetts is the only state in the US where detailed land cover information was comprehensively mapped at such an early date. As a result, these maps provide unusual insight into land cover and cultural patterns in 19th century New England. However, as with any historical data, the limitations and appropriate uses of these data must be recognized: (1) These maps were originally developed by many different surveyors across the state, with varying levels of effort and accuracy. (2) It is apparent that original mapping did not follow consistent surveying or drafting protocols; for instance, no consistent minimum mapping unit was identified or used by different surveyors; as a result, whereas some maps depict only large forest blocks, others also depict small wooded areas, suggesting that numerous smaller woodlands may have gone unmapped in many towns. Surveyors also were apparently not consistent in what they mapped as 'woodlands': comparison with independently collected tax valuation data from the same time period indicates substantial lack of consistency among towns in the relative amounts of 'woodlands', 'unimproved' lands, and 'unimproveable' lands that were mapped as 'woodlands' on the 1830 maps. In some instances, the lack of consistent mapping protocols resulted in substantially different patterns of forest cover being depicted on maps from adjoining towns that may in fact have had relatively similar forest patterns or in woodlands that 'end' at a town boundary. (3) The degree to which these maps represent approximations of 'primary' woodlands (i.e., areas that were never cleared for agriculture during the historical period, but were generally logged for wood products) varies considerably from town to town, depending on whether agricultural land clearing peaked prior to, during, or substantially after 1830. (4) Despite our efforts to accurately geo-reference and digitize these maps, a variety of additional sources of error were introduced in converting the mapped information to electronic data files (see detailed methods below). Thus, we urge considerable caution in interpreting these maps.

Despite these limitations, the 1830 maps present an incredible wealth of information about land cover patterns and cultural features during the early 19th century, a period that continues to exert strong influence on the natural and cultural landscapes of the region.

Methods for the 1830 Map Project:

The original 1830 maps for towns in Massachusetts are bound in a series of volumes that are located in the State Archives in Boston. We obtained copies of the 1830 maps from several sources, using an index to the volumes and page numbers of the 1830 series that was obtained from the UMASS library. During the course of the study, several omissions from or errors in the index were identified; a corrected index is permanently archived at HF.

Copies of the 1830 maps were obtained from: 1) the State Archives (primarily photostat copies); 2) photocopies from the microfilm version of the 1830 series maps located in the basement of the Dubois ('Tower') library at UMASS, Amherst; 3) towns, historical societies, or other local sources for a few towns (e.g., Dartmouth, Littleton, Plymouth, Taunton, etc.). For a few towns where we had difficulty locating or obtaining good reproductions of the 1830 maps (e.g., Gloucester, Littleton, Lowell, Lunenburg, New Bedford, Norton, Plymouth, Scituate, and a portion of Chelsea), we used copies of lithographs produced in the 1830s that were apparently based on the 1830 maps, but may occasionally have involved some re-surveying*. In a few cases where our copies of 1830 maps were of poor quality, digital photographs were also taken of the originals in the State Archives. Numerous trips to the State Archives were also necessary to compare our copies of each 1830 map with the originals and to annotate our copies with notes or features that had not reproduced well.

* Many of these lithographs were part of the 'Pendleton Lithograph' series; an index to the Pendleton (and other) Lithograph collection of The American Antiquarian Society in Worcester is archived at HF.

The following is a list of towns for which we have been unable to locate any 1830 map or related lithograph:

  1. A portion of the City of Boston
  2. Carver
  3. A portion of Chelsea (we have a lithograph of Mt. Bellingham)
  4. Plympton
  5. Revere
  6. Winthrop
  7. Aquinnah (=Gay Head; see Chilmark 1830 map)
  8. Northern section of Sturbridge (missing from the original in State Archives)
  9. Northern section of Holyoke (part of Mt. Tom, not shown on any town map)
  10. Western porton of Leominster (unincorporated in 1830)
  11. Western part of Fall River and Westport (part of Rhode Island until 1861)

Although most 1830 maps (or Lithographs) indicate woodland cover, the following do not:

Andover (included North Andover and part of Lawrence on 1830 map) *
Bernardston *
Boston Corner
Boylston (included West Boylston on 1830 map)
Buckland *
Charlemont
Charlestown
Chelsea (Lithograph of Mt. Bellingham indicates one small grove of trees)
Chesterfield
Chicopee (See Springfield map) *
Colrain
Dalton
Dracut *
East Longmeadow (See Longmeadow map)*
Florida *
Goshen *
Hancock
Hardwick
Hawley
Heath
Hinsdale
Hull
Huntington ('Norwich' in 1830)
Lawrence (see Andover and Methuen maps) *
Lenox
Longmeadow (included East Longmeadow on 1830 map) *
Marblehead (see Salem map)
Marlborough
Middlefield
Middleton *
Monterey
New Ashford *
New Braintree *
New Marlborough
North Andover (See Andover map) *
Oakham
Peru
Rowe
Rutland
Salem (included Marblehead on 1830 map)
Sandisfield
Savoy
Sheffield
Shrewsbury
South Hadley (Hadley map shows woodlands for a small portion of S. Hadley)
Springfield (included Chicopee on 1830 map) *
Southwick *
Sturbridge
Sudbury
Tyringham
West Boylston (See Boylston map)
Windsor

* Woodlands were not mapped on original 1830 maps, but 'point' data or more general references to woodlands (e.g., marginal notes indicating "75% of the town is wooded") were found on these 1830 maps and entered into the "Notes" field in the "1830townbounds" and in "1830miscellaneousinfo" ArcView GIS shapefiles.

The following information from the 1830 maps was geo-referenced to USGS topographic maps (1:25,000 or 1:24,000) using a zoom transfer scope: woodlands, roads, meadows, wetlands (swamps, fresh and salt marshes), water bodies (see below), some buildings (meeting houses, mills, taverns, alms house, jails, schools, light houses, windmills, etc., but not residences), cemeteries/graveyards, and town boundaries. In instances where stand types (hardwood, conifer, mixed) or species (e.g., hemlock, chestnut, or maple) were mapped, we retained this information in the "1830LULC" shapefile table "Notes" field. References to special ecological features (cedar swamps, pine plains, magnolia swamp, scrub oak, etc.) were indicated as polygons in the "1830LULC" shapefile or points in the "1830MiscellaneousInfo" shapefile, with comments listed as well. Hills indicated on the 1830 maps were only occasionally zoom transferred and digitized, if there was some reason to suspect that they may have been wooded (with comments in the "1830LULC" shapefile table). In cases where 1830 town boundaries differed from modern town boundaries, 1830 boundaries were zoom transferred to the USGS base maps. However, in several instances, maps of abutting towns differed in their common boundaries; we arbitrarily selected one of these for digitizing and indicated this on the USGS base maps. In cases where 1830 roads clearly followed modern roads but were drafted incorrectly on the 1830 maps, we followed the modern roads. In general, we placed greater attention on woodlands and other land cover features than on human structures, roads, or town boundaries. For instance, although some towns mapped individual houses, we did not transfer and digitize this information. Similarly, some salt works may not have been transferred. 1830 maps (or portions of maps) that are highly distorted were noted as such on the USGS base maps and this information was added to the "1830miscellaneousinfo" shapefile to warn users.

In zoom transferring, we attempted to use as much of the information from the original 1830 maps as possible to geo-reference features. Although we typically used roads surrounding a feature for geo-referencing, if a woodland was clearly indicated as occurring on a hill that was within a polygon of roads on the 1830 map (or along the shores of a mapped pond or stream), we would fit the woodland polygon to that topographic or geographic position as best as possible, even in cases where this represented considerable distortion relative to nearby roads or other features. (See detailed methods below for treatment of water bodies and wetlands).

Error Checking:

Mindy Syfert was responsible for most zoom transferring and digitizing for this project. After zoom transferring but prior to digitizing, all base maps were error-checked by a different person (typically G. Motzkin) against our copies of the original 1830 maps in order to identify and correct any omissions or incorrect interpretations. The corrected geo-referenced data were then digitized in ArcView in UTM projection (NAD 27, UTM zones 18 and 19), and subsequently converted to MA Mainland State Plane projection (NAD 83). The digitized data layer was subsequently checked by Brian Hall against the originals and corrected as necessary. A final round of error checking was conducted (by A. Meister and B. Hall) on the electronic data for the entire State (e.g., to ensure that roads are connected across town and county boundaries, to make sure that all notes from USGS base maps regarding distortion and inaccuracies have been recorded in the electronic files, etc.), comparing with the originals where necessary.

The 1830 maps for portions of the State (e.g., towns in north-central MA, Cape and the Islands) were previously zoom transferred and digitized for other projects. These maps were checked to ensure that interpretations were fully consistent with the remainder of our work. In instances where minor errors or differences of interpretation were noted, these were corrected. In cases of major differences of interpretation or for several towns where roads had not previously been transferred (primarily on Cape Cod and the Islands), we re-did the entire map following the standard protocols outlined in this document.

We have permanently archived in the Harvard Forest Archives:

  1. Copies of all 1830s maps (or associated lithographs) that were used for this project;
  2. 'Working maps' that were used for digitizing and that have the zoom transferred information, including errors prior to correction of electronic data, and notes on USGS base maps.
  3. A corrected index to the 1830 maps series.
  4. A microfilm copy of the 1830 map series.

Water Bodies:

General Considerations:

Water bodies on the original 1830 maps were inconsistently and inaccurately mapped. In many instances, natural water bodies were omitted on the original 1830 maps or there were substantial errors in the location, size and shape of water bodies. In the interest of developing a series of maps that depict the 1830 landscape as accurately as possible, we developed the following protocol for water bodies:

All ponds mapped on the original 1830 maps are included in the HF GIS layer. In addition, ponds that appear to be natural water bodies but that were omitted from the original 1830 maps are included in the HF GIS layer. Ponds that were not mapped in 1830 and are apparently artificial (e.g., 20th century reservoirs) were omitted from the "1830LULC" shapefile.

Note: modern wetlands and water bodies represent ~3.1% and ~ 3.4% of the land cover in Massachusetts. The HF 1830 data layer should not be used for investigations of wetland, waterbody, and coastline change.

Specific Methods:

The following specific protocols were applied to all water bodies:

Water bodies were extracted from the MassGIS land use data layer that was compiled from 1:25,000 scale aerial photos taken in 1985 and 1990. Minimum mapping unit was 1 acre.

  1. If a pond was mapped in 1830 and does not appear on the MassGIS data layer, the pond was zoom transferred to a modern topographic map and then digitized. Such ponds are coded '7' and the notes field is left blank.
  2. If a pond was mapped in 1830 and the general size and shape are similar to modern water bodies (suggesting that the pond has not been substantially altered), we assume that the modern size and shape are correct. For the HF GIS layer, these pond polygons were copied directly from the MassGIS datalayer, coded "7", and indicated in the notes field as "from MassGIS modern LULC".
  3. If a pond was mapped in 1830 but it's size and shape are substantially different from the modern water body indicated in the MassGIS datalayer (suggesting the pond has been substantially altered since 1830), the outline of the pond from the original 1830 map was zoom transferred to a modern topographic map and then digitized. Such ponds were coded '7' and the notes field is left blank. These ponds are often indicated as 'enhanced' or 'artificial' in "An Inventory of the Ponds, Lakes, and Reservoirs of Massachusetts".
  4. If a pond appears in the MassGIS data layer but was not mapped in 1830, we checked the USGS 1880 topographic maps to determine whether it was mapped in 1880. Ponds that were not mapped in 1830 or 1880 are assumed to be more recent, artificial ponds, and are not included in the "1830LULC" shapefile. The only exception to this is that very small ponds (<~1.5 ha) that are indicated in the MassGIS data layer, do not appear to be artificial, and were not mapped in 1880 (or 1830), were retained. This seemed reasonable since many small, apparently natural ponds were not mapped in 1830 or 1880 (1880 map scale is 1:62,500). These ponds are noted as "from MassGIS modern LULC" and coded "700".
  5. Ponds that were not mapped in 1830 but were mapped in 1880 and the MassGIS data layer were generally included in the "1830LULC" shapefile as they are likely to have existed in 1830. These ponds are noted "from MassGIS modern LULC" (Code = 700). This represents a potential source of error, as some ponds may have been included that were constructed between 1830 and 1880. To minimize this error, we did not include ponds that were mapped in 1880 but are clearly artificial.
  6. "An Inventory of the Ponds, Lakes, and Reservoirs of Massachusetts" was occasionally used to help distinguish natural or 'enhanced' ponds from artificial ones, although many ponds are not included in this book.
CodeNotesDefinition
7 pond ZTS from 1830 then digitized
7'from MassGIS modern LULC'pond indicated in 1830 but mapped from MassGIS data layer
700'from MassGIS modern LULC'pond not mapped in 1830 but retained in HF GIS from MassGIS

Meadows and Wetlands

Meadows on the 1830 maps are almost exclusively along streams or rivers (except along the coast). 1830 meadows were zoom transferred to USGS maps, fitting them topographically to the lowlands along streams. MassGIS wetland boundaries were primarily used for extensive wetlands such as the marshes in Essex and Barnstable Counties and in those instances when the general shape and size of 1830 wetlands matched MassGIS wetlands. Occasionally, meadows were indicated on the 1830 maps that could not be easily associated with modern streams, wetlands, or topographic features; these were zoom transferred and digitized with notes indicating that these may represent upland meadows. In addition to meadows indicated on 1830 maps, other wetlands indicated in MassGIS were retained in the HF GIS layer. In all cases except Billingsgate in Wellfleet, the MassGIS shoreline was used since shorelines were frequently distorted on the 1830 maps.

CodeNotesDefinition
20 meadow ZTS from 1830 map then digitized
20'from MassGIS modern LULC'meadow indicated in 1830 but mapped from MassGIS datalayer
500'from MassGIS modern LULC'wetland not mapped in 1830 but retained in HF GIS from MassGIS

Data Files:

All data are in ArcView GIS shapefile format, Massachusetts Mainland State Plane Projection with NAD 83 (units are meters). The shapefiles and description are as follows:

  1. 1830TownBounds - boundaries of the towns in 1830.
  2. 1830Roads - roads, some canals.
  3. 1830MissingDataTowns - town outlines of those towns that are missing data, the shapefile table lists which types of data are missing.
  4. 1830MiscellaneousInfo - various town-wide or point location notes made by either original surveyors or by Harvard Forest personnel (areas of distortion, mountains that were probably forested, etc).
  5. 1830LULC- landcover; the attribute table has a numeric code and string (word) description for each polygon. Codes and Descriptions are as follows:
    LandcoverDescription
    1Openland*
    2Woodland, deciduous
    3Woodland, mixed
    4Woodland, coniferous
    5Wetland, undifferentiated
    6Wetland, swamp
    7Water, shown on 1830 map
    20Meadow, shown on 1830 map
    50Woodland, undifferentiated
    100Special - see notes field
    200Town with woodland data not shown
    500MassGIS wetland
    700MassGIS waterbody
    * for most of the state, openlands form the "matrix" for the other polygons. The openland matrix was not included in this layer but openland polygons were included in some situations such as small "donuts" in large polygons or as true islands in rivers, etc).
  6. 1830Buildings - important public, commercial, or industrial buildings shown on 1830 map.
    CodeDescription
    10Meeting House (Church)
    11Cemetery
    20School
    21College
    22Academy
    25Store
    26Tavern
    27Hotel or Inn
    30Mill or Factory (details in notes field if specified on original)
    31Cloth Related (type specified)
    32Paper Mill
    33Saw Mill
    34Grist Mill/ Corn Mill
    35Iron/Metal related, type specified
    36Saltworks
    37Machine Shop/Machine Factory
    60Town House/Town Hall
    61Post Office
    62Charity - Alms House, Poor Farm, Town Farm, Halfway House
    63Jail house
    64Mine/Quarry, type specified
    70Lighthouse
    71Windmill
    100Special - see notes field
    1000Unknown - see notes field for reason
  7. 1830Lulc.avl - ArcView legend connected to the "Landcover" field in the "1830LULC" shapefile; contains descriptions of LULC types.
  8. 1830Buildings.avl - ArcView legend for 1830Buildings shapefile, "Type" field. The End