Site index determination techniques for southern bottomland hardwoods

Abstract

density (Gingrich 1967, Lanner 1985). A site index curve is developed by plotting height and age data for a given species and forming a height-age growth curve (Avery 1975). A series of harmonized curves is then plotted, usually at 10-ft intervals at the base age, to generate a series of anamorphic site index curves (Schnur 1937). Users can then plot tree height and age data collected from the stand onto species-specific curves and determine site index. Husch et al. (1982) reinforced that site index varies according to species on a given site and that site index curves are prepared for individual species. Furthermore, index trees, or the trees from which total height and age are measured, were not suppressed from an overstory canopy during their lifetime; otherwise, the productive potential of the site for a particular species would be underestimated. Several problems exist when using index trees for site index determination. Spurr and Barnes (1980) noted that assigning dominant and codominant crown classes to individual trees for site index determination is subjective and not necessarily repeatable, though Meadows et al. (2001) developed a crown classification system to quantify bottomland hardwood tree crowns on the basis of crown position and condition. Furthermore, tree canopy status changes over time as individuals drop from a dominant or codominant crown class to lower crown classes through competition, reducing the repeatability of measuring the same index trees (Raulier et al. 2003). Determining tree heights accurately also may be difficult because of dense leaf layers in tree tops during the growing season. McNab (1989) noted that age is particularly difficult to measure accurately in hardwoods. Furthermore, questions arise as to how many index trees are needed to accurately determine site index (Mailly et al. 2004). Finally, there is a temptation to use site index curves beyond the range of data used to develop the height-age relationships. Additional problems have been noted using anamorphic site index curves (Tesch 1980). First, an assumption is made in developing anamorphic curves that height development of a species is similar across a variety of age classes and site conditions. Many authors have showed this to not be true (e.g., Carmean 1956, 1972, Hilt and Dale 1982). Curtis (1964) and Carmean (1972) indicated that harvesting was more likely to occur on good sites than poor sites because trees would reach merchantable size earlier on good sites. Therefore, older age classes in site index curve development would mostly be represented by trees growing on poor sites. Second, Carmean (1972) found different patterns of height growth for upland oak species, based on stem analysis data, compared with height development based on anamorphic site index curves; therefore, he used polymorphic height growth patterns to better explain differences in height growth within a species. Anamorphic site index curves are proportional curves that have a constant ratio between heights regardless of tree age (Carmean 1972). Polymorphic site index curves are not proportional; therefore, they will have different height ratios depending on the age at which heights are compared. For example, Carmean (1972) showed differences in height growth in black oak (Quercus velutina Lam.) in later years depending on site quality. These differences would not appear in anamorphic site index curves, but they do appear in polymorphic site index curves. Carmean et al. (1989) published one of the most comprehensive sets of hardwood site index curves, with 53 curves for 31 species or species groups. Table 1 lists all known site index curves for southern bottomland hardwood species. Other curves are available for several of these species but were developed for sites outside of the southern United States. Of the 16 site index curves available for 9 species, only 3 have been published since 1976. Clatterbuck (1987) published site index curves for cherrybark oak (Quercus pagoda Raf.) and sweetgum (Liquidambar styraciflua L.) as part of a stand development study on old-field floodplain sites, and Cao and Durand (1991) published site index curves for eastern cottonwood plantations growing in the Lower Mississippi Alluvial Valley. The reasons for little site index information for southern bottomland hardwood species are four-fold. First, development of site index curves is expensive. Considerable effort is required to locate stands that fit the assumptions for site index determination: relatively undisturbed, even-aged stands. Second, minor changes in elevation strongly influence flooding, sedimentation, soil texture, and soil pH. These changes strongly influence species composition and productivity (Hodges 1997, Wall and Darwin 1999), making site index determination difficult. Third, bottomland hardwood forests have been greatly altered through harvesting practices such as indiscriminate high grading, leading to few stands containing suitable index trees. Finally, changes in hydroperiod from levee/road construction and urban development have resulted in bottomland sites that are constantly changing, making site index a fluid rather than a static value. Site index curves will remain the most popular technique to determine site index, which is a well established measure in the literature and in forest managers’ training. Furthermore, site index is a key variable in many of today’s growth and yield models (Hilt 1985). Forest managers must be aware, though, of the limitations in using site index curves. Soil-Site Index Equations Soil-site index equations provide a quantitative means to estimate site index based on soil and site variables that directly or indirectly influence tree height growth. Variables may include depth of topsoil, soil texture, presence or absence of pans, topography, and aspect. Several reviews of soil-site index equations based on older literature have been published (Carmean 1975, 1977). Soil-site index equations have been developed for several hardwood species on upland and bottomland sites. Equations for upland hardwoods have greater correlation with site index determined from index curves (r 0.61 to 0.86) than those for bottomland hardwoods (r 0.38 to 0.68), primarily because of the strong influence Table 1. Published site index curves for southern bottomland hardwood species (updated from Francis 1984). Species Source area (states) Reference American sycamore MS, LA Briscoe and Ferrill (1958) Cherrybark oak AR, AL, LA, MS, TN Broadfoot (1961) MS Clatterbuck (1987) Eastern cottonwood IL, IN, KY, MO Neebe and Boyce (1959) AR, KY, LA, MS, TN Broadfoot (1960) LA, MS Alexander (1976) MS Cao and Durand (1991) Green ash AR, AL, LA, MS, TN Broadfoot (1969) Nuttall oak AR, AL, LA, MS, TN Broadfoot (1969) Swamp blackgum GA Applequist (1959) Sweetgum AL, FL, MS, SC Winters and Osborne (1935) AR, LA, MS, TN Broadfoot and Krinard (1959) AL Lyle et al. (1975) MS Clatterbuck (1987) Water tupelo GA Applequist (1959) Water oak AR, AL, LA, MS, TN Broadfoot (1963) a Nyssa biflora Walt. b Nyssa aquatica L. 6 SOUTH. J. APPL. FOR. 37(1) 2013 of slope position and aspect on upland tree height growth (e.g., Trimble 1964, Graney 1977, Woolery et al. 2002). Aspect is less important on bottomland sites because of relatively level topography. However, Beaufait (1956) found that differences of only a few inches in elevation strongly influenced silt and clay content, flooding depth and duration, and soil aeration, thus correlating these topographic features with site index. Broadfoot (1969) advocated against using soil-site index equations in bottomland hardwood forests, especially for investment purposes. These equations have little practical application because of high variability in bottomland soils, and some soil attributes, such as chemical properties, are difficult to measure in the field. Soil-site equations do allow for a better understanding of the influence of various soil and site characteristics on tree height growth. For example, Broadfoot (1969) developed the following site index equation for water oak (Quercus nigra L.): SI50 87.11 13.34(X1) 1.558(X2) 0.0264(X3)