Line vs belt transect

USDA Forest Service Gen. Tech. Rep. PSW-GTR-181. 2002. 665 Assessment of the Line Transect Method: An Examination of the Spatial Patterns of Down and Standing Dead Wood1Duncan C. Lutes2 Abstract The line transect method, its underlying assumptions, and the spatial patterning of down and standing pieces of dead wood were examined at the Tenderfoot Creek Experimental Forest in central Montana. The accuracy of the line transect method was not determined due to conflicting results of t-tests and ordinary least squares regression. In most instances down pieces were randomly distributed along transect segments. Down pieces generally had a clumped distribution of their directional orientation. Standing pieces were usually found to be randomly distributed within belt transects. Consistent clumping scale of down or standing pieces was not found when studied using the paired quadrat variance method. Introduction Coarse woody debris (CWD), principally logs and snags, plays a key role in a wide range of ecological processes in conifer forests. It is important for wildlife, plant regeneration, nutrient cycling, water quality, fire fuels and more (Harmon and others 1986, Maser and others 1979, Maser and others 1988). A number of researchers have used the line transect method (Brown 1974, Howard and Ward 1972, Van Wagner 1968, Warren and Olsen 1964) for quantifying the down component of CWD. Most studies have assumed the accuracy of the line transect method when quantifying down debris. In a literature search for this paper, studies were not found that compared volume measured on a fixed area to the volume estimated on that same area with the line transect method, in the natural setting. Few studies have tested two important assumptions of the line transect method for large debris: random piece orientation and spatially random piece distribution. Correcting for nonrandom piece orientation requires additional measurement effort in the form of additional transect lines or a mathematical correction. Identifying a random distribution may reduce the number of line segments and thus the effort required to reach some desired level of precision. If pieces are spatially clumped at some consistent scale on the forest floor, the identification of that scale 1 An abbreviated version of this paper was presented at the Symposium on the Ecology and Management of Dead Wood in Western Forests, November 2-4, 1999, Reno, Nevada. 2 Consultant, Systems for Environmental Management, P.O. Box 17101, Missoula, MT 59808-7101 (e-mail: dlutes@fs.fed.us)
Assessment of the Line Transect Method—Lutes USDA Forest Service Gen. Tech. Rep. PSW-GTR-181. 2002. 666 may be used to reduce or eliminate the bias introduced by nonrandom spatial distribution. The objectives of this study were to 1) compare the volume of down debris (logs) measured on fixed-area plots against estimates of down debris made with the line transect method on the same area; 2) test two assumptions of the line transect method: pieces need to be distributed randomly in their orientation and at random, spatially, on the forest floor; 3) identify the spatial distribution of standing pieces (snags); and 4) examine for a consistent scale of clumping of down and standing pieces. Methods Study Site This study was conducted at the Tenderfoot Creek Experimental Forest (TCEF) in central Montana (fig. 1). The forest extends in elevation from approximately 1900 to 2400 m and covers an area just less than 3,700 ha. Lodgepole pine (Pinus contorta) is the predominant cover type with scattered patches of subalpine fir (Abies lasiocarpa) and Engelmann spruce (Picea engelmannii) dominating older stands. Whitebark pine (Pinus albicaulis) and limber pine (Pinus flexilis) are also found throughout the TCEF. Figure 1―Study map of the Tenderfoot Creek Experimental Forest in central Montana, showing the locations of the13 sampling areas. Sampling areas were established in five stands chosen from a fire history map of the TCEF (Barrett 1993). Four of the stands had last burned in 1873, 1845, 1765 and 1726, respectively. The fifth stand had burned in 1845 and again in 1873 (two-burn stand). None of the sampled stands showed any significant, identifiable sign of disturbance since the time of the last fire event. The 1873, 1845 and the two-burn