Lyford Mapped Tree Plot

• Investigators: Audrey Barker-Plotkin, David Foster, Ann Lezberg, Walter Lyford
• Contact: Audrey Barker Plotkin
• Start date: 1969-01-01
• End date: ongoing
• Location: Prospect Hill Tract (Harvard Forest)
• Latitude: +42.53
• Longitude: -72.18
• Elevation: 345 meters
• Taxa:
• Keywords: forest dynamics, soils, tree maps
• Release date: 1999
• EML version: knb-lter-hfr.32.2
• Revisions: data updated 2003-06-24
• Abstract:

  Permanent forest plots provide an empirical understanding of forest change over time, and are an invaluable part of forestry and ecological research. Walter Lyford began measurements of a 2.9 ha red oak-red maple forest on the Prospect Hill Tract of Harvard Forest in 1969. All trees over 2 inches (5 cm) were mapped on very large-scale (1 inch = 5 feet) hand-drawn maps, and included live and dead trees, stumps, windthrows and other features such as stone walls, boulders, soil moisture and a damage boundary from the 1938 hurricane. All living and dead trees have been re-located and measured (diameter at breast height, canopy class for live trees; condition, decay class, diameter, bole length and stem orientation for fallen dead trees) in 1969, 1975, 1987-1992 and 2001. In 2001, the original, hand-drawn maps were digitized using ArcView GIS.

  Overall trends for the site show typical patterns for maturing forests. Basal area rose from 108 ft2/acre in 1969 to 144 ft2/acre in 2001, while stem density declined from 517 stems/acre in 1969 to 346 stems/acre in 2001. Red/black oak and red maple are the dominant species, plus 21 other tree species. Red/black oak comprised more than half of the total basal area over all years, while red maple comprised about half of all stems.

  In part of the stand that was severely damaged by the 1938 hurricane, basal area and stem density trends indicate that this area is at an earlier stage of stand development than the part of the stand only lightly to moderately damaged. In the severely damaged area, red oak is less important and birch species are more important. Densities of windthrown trees and stumps originating from 1938 are much higher in the severely damaged area.

• Methods:

  Sampling History

  Detailed tree maps were prepared by Walter Lyford over this 2.9ha area between 1962-1969. Map accuracy is within 1-2 feet. Maps cover 32, 100x100 ft. blocks; each block is marked in the field with labeled wooden stakes. The entire site was censused in 1969 and 1975. Between 1987 and 1992, another survey was completed.

  Blocks censused in 1987: SE61, SW60, SW61, SW62, SW70, SW71. Blocks censused in 1991: SE11, SE21, SE31, SE41, SE51, SW10, SW11, SW20, SW22, SW30, SW40, SW50, SW51. Blocks censused in 1992: SW00, SW01, SW21, SW31, SW32, SW33, SW34, SW41, SW42, SW43, SW44, SW52, SW53. Most recently, the entire area was censused in 2001.

  2001 Field Methods

  Each grid block was temporarily flagged around its boundary, and staked flags were used to mark 10 ft. intervals along the perimeter of the block. We did not find it necessary to divide the block into subsections.

  To locate individual trees, we created maps for each block using tree coordinates mapped in ArcView. Trees were color-coded by species, and also coded whether the individual was living or gone at the last measurement. We found it extremely helpful to label each stem on the map with an ID number, corresponding with an ID number on the data sheets. This made it much easier to match previous tree data with an individual on the map and on the ground. We did not bring originals or copies of the large-scale hand-drawn maps into the field. We found that we could confidently navigate the plots with our smaller GIS maps.

  For each tree, data on dbh, canopy position (dominant, codominant, intermediate or suppressed) and condition was recorded. Diameter was measured with a diameter tape to the nearest 0.1cm, and later converted to inches. For dead stems, diameter was measured to the nearest cm. For fallen trees (snaps, root breaks, some crown snaps), length (ft.) and orientation (true N) of the fallen stem was recorded. [Note: we noted in some plots that fallen trees that we relocated were consistently about 15 degrees lower than orientation recorded in 1991-1992. We figured that directions measurements had not been adjusted for declination in these plots]. Diameters for fallen stems were measured with calipers to the nearest cm. Dead trees (standing dead, snapped, stumps, etc.) were given a decay class (D1 = firm/solid, fresh, small twigs or stubs intact; D2 = partial decay, bark may be partially missing, bole may be broken but shape still evident; D3 = decaying into ground/soil, crumbled). If a tree could not be relocated (almost always previously dead trees that had presumably rotted away), the tree was recorded as "gone." Comments included things such as "leaning SW" or "measured diameter above scar." Trees were marked with chalk to track progress.

  New trees (stems that had grown above 5cm/2in. since the block was last measured) were added; diameter, canopy class, condition and species were noted. Coordinates were determined either by measuring from a very nearby tree with known coordinates, from corner posts, or most commonly, by the INTERPNT method of triangulation (Boose et al. 1998, Ecology 79:819-827). This method is based on recording distances (in meters) from the new tree to three reference points - in this case, usually three trees with known coordinates. We followed the basic methods given in the program documentation, with a few modifications:

  We converted reference tree coordinates from feet to meters, to meet the program assumption that tree diameters are 100X less than distance measurements.

  Since our reference trees had known coordinates, all of these stems were included in a large benchmark file (we just included all 5000+ stems; this worked fine). In the few cases in which we used a grid post as a reference point, we gave the post an arbitrary number and added its coordinates to the benchmark file.

  For more than half of the plots, we followed the recommended method of measuring from outside bark to outside bark from the target (new) tree to the reference tree. Since the reference trees were all considered benchmarks, we had to adjust distance measurements to the center of the reference tree (i.e., distance + 1/2diameter). This was rather cumbersome, so we later switched to measuring from outside bark of the target tree to the midpoint of the reference tree. We did not find any additional level of error by using this shortcut.

  Since the locations of the reference trees were originally mapped only to the nearest 1-2 feet, we adjusted the default tolerance levels to allow a higher error. We used 0.3 as the first tolerance level (about 1 foot) and 1.0 as the second tolerance level (about 3 feet - trees that definitely had some problem that should be fixed).

  We remeasured all trees with an error greater than 0.3. Many were improved with the second try at triangulation. In our final set of coordinates, 15 trees exceed the 0.3 tolerance level (within original error level in mapping), and one tree exceeded the 1.0 tolerance level, but not by much. We felt this level of error was acceptable considering the error level of the original mapping, and the purposes we envision for having known tree locations.

  2001 GIS Methods

  Our objective was to translate most of the information on Lyford's hand-drawn maps to digital format. We used ArcView 3.2 software, and spatial information from the original 22x22 inch maps was digitized using a digitizing tablet (Summagraphics Microgrid II). Most themes are polygons, including stonewalls and roads, since the map scale is so large. The following themes were created:

  Block corners (just the coordinates where the grid posts are located - a point theme); block polygons were created by connecting the dots and modifying SW01 (incompletely surveyed) and SW44 (incomplete as part of block is in swamp and was not surveyed).

  Trees (point theme derived from coordinates in data table). Footpath (line theme). Intermittent streams (line theme). Soil moisture type (polygon). Area damaged by 1938 hurricane (polygon). Road (polygon). Woods road (polygon - not very evident on ground anymore). Stonewalls (polygon).

  Manipulations (polygon). Some experimental sites are still evident on the ground. Nitrogen fertilization treatments done by Lyford in separate theme.

  Rocks (patchy coverage - Lyford drew in boulders only on some maps (greater than 2ft. diam.) and both boulders and rocks (1-2ft. diam.) on others).

  Fallen stems (line). Endpoints of each fallen stem were calculated from length and bearing measurements, using an extension called "bearing.avx" written by Yingming Zhou. A line theme was created from these points using an ArcView extension called "drawline2.ave" written by Rodrigo Nobrega. In some cases, we noted that the bole was displaced from the tree base - the X,Y coordinates were adjusted accordingly only if both amount and direction of displacement were noted.

• Use:

  This dataset is released to the public and may be freely downloaded. Please keep the designated Contact person informed of any plans to use the dataset. Consultation or collaboration with the original investigators is strongly encouraged. Publications and data products that make use of the dataset must include proper acknowledgement. For more information on LTER Network data access and use policies, please see: http://www.lternet.edu/data/netpolicy.html.

• Related datasets: