Data Set Citation:
When using this data, please cite the data package:
NCEAS 12218: Woods: Slow systems and complex data-sets: Multi-decade permanent plots permit address of recalcitrant questions about late-successional forests , National Center for Ecological Analysis and Synthesis , and Woods K. 2012.
Multi-decade population studies of Michigan old-growth forests
nceas.980.2 (
General Information:
Title:Multi-decade population studies of Michigan old-growth forests
Established in 1935, a regular grid of 256 permanent plots includes about 20% of a 100-ha old-growth forest at the Dukes Research Natural Area in northern Michigan, USA. Woody stems have been remeasured 3�7 times providing extensive quantitative records of population and community dynamics over periods of up to 72 years. Woody stems in upland hemlock�northern hardwood stands, about half of the study plots, have been mapped and individually tracked since about 1990. Remaining plots are in swampy stands dominated by Fraxinus nigra and Thuja occidentalis. Detailed, long-term demographic data for late-successional forests are rare in general; this data set is both of exceptional duration and unusual in spatial intensity and detail. Because sample plots are in a regular array over the stand, they can support analyses of spatiotemporal pattern at various scales. A major wind disturbance in 2002 provides a unique opportunity to compare disturbance response to baseline dynamics. Several publications based on this data set have already provided new insights into late-successional processes, but general availability of the data set with metadata should permit a range of further comparative and integrative analyses. The study is ongoing, and new measurements will be added to the archived data set. Several ancillary data sets are available from the author. Data presented here are the same from the published Ecological Archives paper found here:
  • Acer saccharum
  • Betula alleghaniensis
  • Fagus grandifolia
  • Fraxinus nigra
  • long-term studies
  • northern hardwood forest
  • old-growth forest
  • permanent plots
  • succession
  • Thuja occidentalis
  • tree mapping
  • Tsuga canadensis
Publication Date:2012-02-06
Data Table, Image, and Other Data Details:
Metadata download: Ecological Metadata Language (EML) File
Data Table:All plots 1935 to 1948.txt ( View Metadata | Download File download)
Data Table:All plots 1974 to 1980.txt ( View Metadata | Download File download)
Data Table:Upland plots 1989 to 2007.txt ( View Metadata | Download File download)
Data Table:Swamp all modern.txt ( View Metadata | Download File download)
Data Table:Sampling history.txt ( View Metadata | Download File download)

Involved Parties

Data Set Creators:
Organization:NCEAS 12218: Woods: Slow systems and complex data-sets: Multi-decade permanent plots permit address of recalcitrant questions about late-successional forests
Organization:National Center for Ecological Analysis and Synthesis
Individual: Kerry Woods
Organization:Bennington College
Natural Sciences Bennington College,
Bennington, Vermont 05201 USA
Email Address:
Data Set Contacts:
Individual: Kerry Woods
Organization:Bennington College
Natural Sciences Bennington College,
Bennington, Vermont 05201 USA
Email Address:
Associated Parties:
Individual: Kerry Woods
Metadata Providers:
Individual: Xueying Han

Data Set Characteristics

Geographic Region:
Geographic Description:Dukes Research Natural Area in northern Michigan, USA
Bounding Coordinates:
West:  -87.15  degrees
East:  -87.15  degrees
North:  46.0333  degrees
South:  46.0333  degrees
Time Period:
Taxonomic Range:
Rank Name:Species
Rank Value:Acer saccharum
Rank Name:Species
Rank Value:Betula alleghaniensis
Rank Name:Species
Rank Value:Fagus grandifolia
Rank Name:Sp
Rank Value:Fraxinus nigra
Rank Name:Species
Rank Value:Thuja occidentalis
Rank Name:Species
Rank Value:Tsuga canadensis

Sampling, Processing and Quality Control Methods

Step by Step Procedures
Step 1:  

Research methods

In 1935, U.S. Forest Service personnel established 246 0.2-acre (809 m2) continuous forest inventory (CFI) plots on a 2 × 5 chain (40.2 × 100.6 m) grid (2 chain spacing on north-south lines) over the entire RNA (Fig. 2). Plot centers were marked with steel pipes. Stems >5 in (12.7 cm) diameter at 1.4 m height (diameter at breast height, DBH) were tallied by species in 1-inch (2.54 cm) diameter classes for most plots (n = 236). In 1948, alternate plots along N-S lines were censused in the same manner (n = 123).

In 1974–1980, under the direction of Dr. Frederick Metzger of the U.S. Forest Service, all CFI plots were recensused, with all woody stems >0.5 in (1.27 cm) DBH recorded by species and DBH to nearest 0.1 inch using Biltmore 'cruising sticks' (Frederick Metzger, personal communication). Standing dead stems > 10 in (25.4 cm) DBH and with bark intact were also recorded, but probably not consistently (see below). Data for three plots have not been relocated (n = 243). Active Forest Service research at the RNA was suspended following closure of the Forest Service Northern Hardwoods Laboratory in Marquette, MI in the early 1980s. Some species, occurring primarily as large shrubs, may not have been consistently recorded (Acer spicatum, Alnus sp, Amelanchier sp, Cornus sp, Corylus cornuta).

Beginning in 1989 and continuing over the next several years, I initiated remeasurements, initially for upland plots only. For upland plots (n = 134), all woody stems > 5 cm DBH were measured to nearest 0.1 cm and mapped in polar coordinates from the plot center. Smaller stems were similarly mapped and measured in a central sub-plot 8 m in radius: in 1989 the minimum size for recording stems in the sub-plot was 3 cm; in 1992, minimum size was 2.5 cm; for all subsequent years all stems > 1 cm DBH were included. In 1989–1992, distances were measured using tapes and azimuth using Suunto site-through compasses; from 1993 onward, distances were measured using Sonin ultrasonic rangers. Subsequently, woody stems in upland plots have been remeasured and newly established stems mapped and measured at, generally, five-year intervals (three or four measurements for most); since 1999, these plots have been consolidated into two groups sampled at staggered intervals such that the most recent measurements are for 2004 and 2007. For some species of large shrubs (Acer spicatum, Corylus cornuta, Cornus rugosa) stems were tallied without mapping. All dead stems > ca. 20 cm (including down stems and stumps) were included in initial samples of upland plots. Dead stems were identified to species where possible, and rated on an 8-point decay scale (Table 1), and classed as standing, tipped up, or broken (with height of break recorded); DBH was measured or estimated for dead stems where possible (for decay codes of 7–9, diameter measurements were typically impossible). In subsequent remeasurements of upland plots, stems dying since previous measurements were similarly coded (generally using only the first two grades of the decay scale) where possible (dead smaller stems could not always be relocated); after initial plot measurements, DBH was not recorded for dead stems.

Since 1994, with major efforts beginning in 2001, I have remeasured a total of 65 additional plots in wetlands on peat and muck soils. As for upland plots, all stems > 5 cm DBH were measured for the entire plot and stems from 1–5 cm for an 8-m radius sub-plot, but stems were not mapped in these stands. Dead stems were not consistently recorded, and are not included here.

Sampling Area And Frequency:
The 100-ha Dukes Research Natural Area (RNA) (Fig. 1) is part of the Hiawatha National Forest in northern Michigan, USA (46º02'N, 87º09'W). Lake Superior lies ca. 15 km to the north, and influences a generally continental, cool-temperate climate. At Marquette, Michigan (30 km NW from the study site), annual average temperature 1970–2000 was 4.2 C (average January temperature = -10.9 C, average July temperature = 18.3 C), and annual precipitation averaged 92 cm. Snowfall is high, generally exceeding 400 cm per winter. Total relief within the RNA is < 10 m, but soil drainage, texture, and chemistry vary greatly. Soils of mesic upland areas (about half the study area) are developed on thick glacial till over Paleozoic sedimentary bedrock, and range from sandy to silty loams (Munising and Shoepac-Trenary series silt-loams, Munising fine sandy loam; Schwenner 2007); these are classified as haplorthods and fragiorthods, with cation exchange capacities classed as 'active' to 'superactive' (Soil Survey Staff 1999). These areas are typically strongly hardwood documented (especially Acer saccharum). Somewhat poorly drained silt-loams (Sundell series; hapludolls with superactive CEC) frequently have well-developed shallow hard-pans and support mixtures of northern hardwood species and Tsuga canadensis. Soils in the remainder of the RNA are poorly drained organic mucks and peats (Nahma mucks, Carbondale and Tawas peats and mucks; humaquepts and haplosaprists respectively) that are generally of relatively high pH (5.8–7.0) and CEC. Forests of upland areas are typical of 'hemlock-northern hardwood' late-successional forests of the region, dominated by mixtures of highly shade-tolerant Acer saccharum, Fagus grandifolia, and Tsuga canadensis, along with varying amounts of Betula alleghaniensis, Tilia americana, and Acer rubrum. Fagus has been present locally <1000 yr (Woods and Davis 1989), and its regional range-limit is < 10 km to the west. Peaty wetlands are dominated by Thuja occidentalis, Picea mariana and P. glauca, Fraxinus nigra, Acer rubrum, and, formerly, Ulmus americana along with admixtures of species more prominent on uplands (Woods 2000b, 2004, 2007). There is no record or evidence of logging within the RNA. Increment cores confirm that populations of shade-tolerant canopy dominants are mixed-age, with maximum ages approaching 400 yr (unpublished data), indicating that there have been no stand-initiating disturbances for several centuries, probably > 500 yr. Stand-initiating disturbances, primarily extreme winds, are estimated to occur in this region with return times of several millennia (Canham and Loucks 1984, Frelich and Lorimer 1991, Zhang et al. 1999) Only a handful of unlogged mesic forests remain in the larger region, but it is reasonable to assume that, in terms of successional status, the RNA is representative of a significant fraction of the pre-settlement landscape. The RNA is buffered on all sides by never-cleared stands experimentally managed with long cutting cycles and high residual basal area. In 2002, the stand was affected by a local but intense wind-storm. Immediate consequences for upland areas were consistent with interpretations of stand dynamics above. Canopy destruction was patchy, with some plots losing over 30% of basal area. Disturbance intensity was greatest in areas with well-developed hard-pan and highest canopy diversity, generating gaps averaging about 800–1000 m2 (large enough to favor regeneration of less tolerant species). Mortality risk varied with species; shade-tolerant competitive dominants Fagus and Acer suffered much higher mortality than Betula (Woods 2004).
Sampling Description:
Permanent plots were established throughout the RNA in 1935, and subsequent remeasurements show compositional trends towards increasing dominance of shade-tolerant species (especially Fagus) and commensurate declines in less tolerant dominant species (especially Betula alleghaniensis in uplands, Fraxinus nigra in swampy forests). Even in centuries-old stands, non-equilibrial, essentially successional dynamics dominate compositional trends. Compositional trends appear to depend on local habitat, primarily soil properties, and are convergent within relatively uniform habitat patches. (Woods 2000, 2007).

Data Set Usage Rights

no restrictions
Access Control:
Auth System:knb
Metadata download: Ecological Metadata Language (EML) File