Clear-cutting is without question the most controversial timber harvesting technique. While the forests of northern New England have undergone timber harvesting for well over a century, the rise of large-scale mechanized clear-cutting in the 1970s brought increased attention to the cutting practices employed by large commercial forestland owners.

Beginning in 1993, the AMC (in cooperation with researchers from the Dartmouth College Department of Earth Science) sought to map the extent and distribution — both temporal and spatial — of forest clearing across northern Maine, New Hampshire, and Vermont since 1973. This was one of the first major projects undertaken by researchers in the AMC's newly established Northern Forest mapping program. The critical information:

• Allows an assessment of changing timber-harvesting practices over time.
  
  
• Helps determine the effectiveness of clear-cutting regulations that have been enacted.
  
  
• Contributes to an understanding of current forest conditions, especially the distribution of younger forest stands and large unfragmented blocks.

Until recently, it was difficult to track the location and extent of forest clearing across the region. However, with the launch of the first Landsat satellite in 1973, information on the condition of large landscapes became readily available. These satellites, orbiting several hundred miles above the Earth's surface, take electronic images of the Earth's surface at several different wavelengths (both visible light and infrared). Each image covers an area of more than 100 by 100 miles. Each individual data unit (or "pixel") in the image covers an area of about 250 by 250 feet. Every part of the Earth is "photographed" about every two-and-a-half weeks. (View sample image.)

Methodology
The first phase of the AMC's two-part survey used Landsat Multi-Spectral Scanner (MSS) imagery from the summers of 1973, 1979, 1985, and 1991. (The six-year intervals were used because within seven to ten years many cleared forest areas regenerate to the point that they are indistinguishable from surrounding uncleared forests.) Visible band data (which looks like a low-resolution black-and-white aerial photograph) was used to classify each pixel as either "forest" or "open". Open areas include everything without tree cover, such as recent clear-cuts, fields, developed areas, open wetlands, and rock. They appear as white to light gray. Forest areas show up as darker gray. Water was classified separately using the infrared data.

The data was then digitally overlaid to create a composite data set containing information from all four time periods, allowing the determination of which pixels had changed classification from one time period to the next. This allowed the composite data set to be classified as follows:

• Forest — Areas classified as forest during all four time periods and thus not cleared between 1973 and 1991. They may have undergone partial harvesting but not significantly enough to show up on the imagery.
  
  
• Permanent Open — Areas classified as open during all four time periods — developed areas, agricultural lands, non-forested wetlands and naturally barren areas.
  
  
• Cleared — Areas that changed from forest to open between subsequent time periods.
  
  
• Regrowth — Areas that changed from open to forest between subsequent time periods.

The information gathered in this survey has some limitations. The procedure used to classify the individual images was conservative, and only identified areas that had been almost totally cleared. In other words, areas that contained some residual overstory, even if heavily harvested, were not identified. In addition, the nature of the clearing (whether timber harvesting, clearing for development, or natural disturbance) could not be identified. However, across this region by far the greatest source of forest clearing was timber harvesting. Most of the study area consisted of large commercial forest ownerships where no development was taking place, and the rate of severe natural disturbance in these forests is very low.

The Findings
The results of this first phase show varying temporal patterns of forest clearing across the three New England states (see graph). Throughout the entire region, forests were cleared at a rate of about 0.54 percent of the total area per year. To put that in perspective, natural disturbance in Northern Forest ecosystems creates large openings at a rate of much less than 0.25 percent per year.

Clearing rates were highest in Maine because of:

• A higher concentration of industrial landowners (who were the most likely to make widespread use of clear-cutting).
  
  
• Extensive salvage harvesting during a severe spruce budworm epidemic, which killed large areas of spruce-fir forest between the early 1970s and the mid 1980s.
  
  
• The more gentle topography, which allows more extensive use of mechanized harvesting equipment.

Clearing rates in Maine increased significantly in the 1979-85 period as budworm salvage harvesting reached its peak, and stayed high as industrial landowners continued to make use of large-scale clear-cutting. Clearing in New Hampshire followed a similar pattern but at a lower rate because of the more rugged landscape and the limited area affected by spruce budworm. In Vermont, initially low rates of forest clearing rose sharply after 1985. Though Vermont escaped the spruce budworm epidemic, there was a significant increase in large-scale liquidation harvesting, which led to the establishment of regulations controlling heavy harvesting in the mid-1990s.

Study's Second Phase
The second phase of this study, begun in 2000, is still in progress. It has involved updating the previously developed information with more recent satellite imagery from 1996 to 1998. Among other things, this information will allow an assessment of the effects of the Maine Forest Practices Act of 1989, which set regulatory limits on the allowable size of clear-cuts. This law, along with other factors, such as the end of the spruce budworm epidemic and increased public attention to timber harvesting practices, has led to a significant decline in both the size and extent of clear-cutting across the state. Interestingly, data from the Maine Forest Service indicates that the total amount of wood being harvested has not declined — landowners are either harvesting more acres or removing more wood during partial harvests.

While no statistical analysis is available, a sample map of the data developed to date indicates that the pattern of clear-cutting in Maine has changed significantly. Areas cleared between 1991 and 1997 are much smaller in size and in conformance with the new regulations.