A Guide to Soil Quality Monitoring for Long Term Ecosystem Sustainability on Northern Region National Forests

SOLO HOME > HABITAT TYPES > FOREST HABITAT TYPES OF NORTHERN IDAHO > CHARACTERIZATION AND DISTRIBUTION OF HABITAT TYPES

Forest Habitat Types of Northern Idaho
PRODUCTIVITY/MANAGEMENT AND SOILS EXCERPTS

[Excerpted from: Cooper, Stephen V.; Neiman, Kenneth E.; Roberts, David W. Rev. 1991. Forest habitat types of northern Idaho: a second approximation. Gen. Tech. Rep. INT-236. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 143 p.]

CHARACTERIZATION AND DISTRIBUTION OF HABITAT TYPES

Soils

Our soils data vary in completeness. Some researchers contributed much vegetation data but had not recorded information on soils; others had recorded characteristics of only the upper 4 to 8 inches (10 to 20 cm) of the mineral soil; whereas the authors described soil profiles to control depth or to an impermeable layer on every sample plot (totaling more than 600 plots). Preliminary results for characteristics of the upper horizon(s) and other important factors (for example, effective rooting depth) are presented in appendix D (for syntaxa with three or more samples) and as a paragraph following most habitat type descriptions.

Previous soil sampling strategies for the Northern Region and Intermountain Region have been designed to simply characterize the surface soil for each h.t. rather than investigate detailed soil-vegetation relationships. Our intention in acquiring complete profile descriptions is to build a data base for future soil-vegetation-site productivity research and to correlate our studies with those of Northern Region soil scientists. Based even on our limited data, some habitat types appear strongly controlled by edaphic or topoedaphic factors and express a narrow range in soil values; other h.t.'s are found on a wide range of soil conditions. The most salient associations between h.t.'s and soil conditions occur on sites with seasonally or permanently high water tables (THPL/OPHO, THPL/ATFI, ABLA/STAM, ABLA/CACA); these sites have less gravel (except THPL/OPHO), finer textures, higher pH, and deeper litter accumulations than upslope, well-drained sites. Neiman (1986) identified four soil physical characteristics that appear to be highly useful for differentiating between the extremely similar ABGR/CLUN, ABGR/ASCA, THPL/CLUN, THPL/ASCA, TSHE/CLUN, and TSHE/ASCA h.t.'s. His study indicates that data stratification into narrowly defined parent material groups within a restricted geographic area is a requisite to meaningful analysis.

The influence of parent material on vegetation patterns, which is so evident in Montana (east of the Continental Divide [Pfister and others 1977]) and northwestern Wyoming and adjacent Idaho (Steele and others 1983), is not manifested in obvious ways under the Inland Maritime climatic regime of northern Idaho. Of the commonly occurring parent material groups, the most strongly contrasted pair in terms of their influence on vegetation in Montana, calcareous versus noncalcareous, are rarely represented in northern Idaho owing to a lack of calcareous substrates (<1 percent of plots). Valley locations and loess-derived soils are underrepresented in the dataset because these lands are largely both privately held and deforested. Due to the extensive level of this study, insufficient representation within contrasting parent materials precludes demonstrating significant contrasts in terms of vegetation response.

The soil moisture regime and depletion rate studies of McMinn (1952) and Daubenmire (1968b) have demonstrated that these rates differ significantly and predictably among types, but these differences are not necessarily demonstrable by conventionally inventoried soil properties (for soil classification, USDA 1975). The often-advanced hypothesis that vegetation types (h.t.'s) are predictable from a standard set of soil characteristics was discredited in northern Idaho by the Daubenmires (1968). Even on a local scale (Priest River Experimental Forest), Daubenmire (1973) demonstrated poor correlation between habitat types and soils at even the Order level of soil classification. This result is not unexpected because (1) soil classification systems are not designed to reflect primarily those properties influencing vegetation development; (2) vegetation response to climatic fluctuation is relatively rapid, whereas development of equilibrium in soil properties requires a relatively long time; (3) many factors, of which soil characteristics are only one, influence vegetational development, and through factor compensation species are able to grow on a wide range of substrates (soils); and (4) a rather coarse vegetation classification containing large within-type variation was used. Thus, one must exercise caution when attempting to “shortcut” inventories of either vegetative potentials or soils through a process of assumed correlations; relationships must be objectively and adequately tested and cautiously extrapolated. The results of Neiman (1986) indicate that with adequate data and proper stratification, certain of the physical soil characteristics can be correlated with identifiable vegetation patterns.

Timber Productivity

Timber productivity is one of the key management concerns for which data were accumulated during the study. Our data base for estimating productivity by habitat type is much reduced compared to that for characterizing the vegetational composition because (1) near-climax stands frequently have no suitable site trees; and (2) we have borrowed many plots from studies interested primarily in vegetation classification, in which they did not record associated productivity features.

All site trees were relatively free-growing, of dominant and dominant crown classes. We have expanded the number of trees sampled, from one per species per plot, as done in previous regional studies (Steele and others 1981, 1983; Pfister and others 1977), to three to five (or more) trees per species per plot to better represent the within stand variation. Preliminary indications from work in western Montana for four major seral species are that a one-tree sample will be within ±9 to 10 site index units (50-year base) of the true mean 90 percent of the time; five tree samples will be within ±3 to 5 site index units of the true mean 90 percent of the time (Fiedler 1983).

We adhered to the criteria of Pfister and others (1977) for recognizing suppressed trees by increment core analysis, rejecting those with 10 or more years of suppression. Preliminary onsite counting of tree rings, examination of their pattern, and tree height computation enabled us to recognize and delete from the data set those trees whose radial growth reflected a questionable or atypical height growth rate. Some sites with high incidence of root rot or Indian paint rot (Echidontium tinctorium) required coring of 10 to 12 trees of one species to attain an acceptable three- to five-tree sample. For any mature stand the probability is high that all site trees have incurred some degree of growth reduction caused by insects, disease, or short-term climatic phenomena (drought years of late 1930's are readily detectable). Though such damage can be detected by ring width measurements (Carlson and McCaughey 1982) for determining the frequency and severity of past pest population eruptions, these measurements are not currently applicable for assessing the degree to which tree height or stand productivity has been reduced. It is also possible that some habitat types are more subject to higher incidences or greater degrees of growth reduction. We mention these factors as a cautionary note to uncritical acceptance or comparison of the timber productivity of various habitat types.

Each species requires a specific algorithm for computing its site index. Most site index computations require total age and height to utilize the curves; the age to breast height (4.5 ft [1.4 m]), which varies with stand history and h.t., must be estimated or measured. For species not having a specific site curve, a curve from a species hypothesized to have a similar growth curve is substituted. Table 2 summarizes the criteria used to determine total age and the source of site index curves; some of our curve selections differ from those employed in previous regional classifications for the reasons outlined below. See Steele and Cooper (1986) for a compilation of applicable site curves.

Pseudotsuga menziesii site indexes were determined using Monserud's (1984a) curves, which are derived from stem analysis of stands throughout our study area and based on breast-height age.

For Picea engelmannii we employed Clendenen's (1977) 50-year age base modifications of Alexander's (1967) curves. These curves are not very different from Brickell's (1970) and have the advantage of being based on breast-height age and have yield data associated with them (Alexander and others 1975). We also used Clendenen (1977) for Abies lasiocarpa and Tsuga mertensiana, reasoning that all three have similar ecological roles in subalpine ecosystems and may possess similar growth curve form.

We used Barnes' (1962) curves derived for coastal Tsuga heterophylla for both inland populations of T. heterophylla and Thuja plicata, based on the similarity of their growth forms and ecological roles.

The site index data (50-year base age) and basal areas have been summarized by species within types and phases.

We have computed an average site index whenever values exist for three or more stands; for five or more values a 90 percent confidence interval (CI) for estimating the true population mean was computed. The population sampled for site index (S.I.) is a species as it occurs on a given h.t. (or phase). An example for interpretation: average S.I. = 50, CI = ±10 (90 percent), n = 15, means that if one were to take 15 S.I. readings (sample 15 stands for a given species and h.t.), then 90 percent of the time the true S.I. mean for that species on the specified h.t. will lie between 40 and 60. The CI narrows with both increased sample size and decreased variability.

For many of the same reasons that considerable vegetational variation occurs within a given h.t., so, too, do we expect variability in productivity estimates, within and between species, within types. Even though all species are related to a 50-year age base, the difference in estimates based on actual breast-height age versus adjusted basal age means the species' S.I. values are not strictly comparable. On less productive h.t.'s the time to reach breast height may be as much as two to five times greater than that on more productive sites. The shortcomings outlined by Daubenmire (1961) in using anamorphic site index curves are still with us and cast a degree of doubt on using index values for comparative purposes; Monserud (1984b, 1985) is yet more critical of anamorphic curves and the whole site index concept, from outmoded curve construction techniques to erroneous assumptions made in predicting volume productivity.

Pfister and others (1977) have noted productivity differences between eastside and westside (of Continental Divide) stands for a given h.t. We recognize the desirability of regionally stratifying our site index data, but our data are presently insufficient for this approach. Preliminary indications are that h.t. productivities for the study area tend to be higher than for comparable types of contiguous areas to the south or east. The most appropriate, meaningful stratification could be made by individual forests or divisions thereof, drawing upon their inventory data.

Previous regional h.t. publications have presented net yield capabilities (in ft³/acre/year [m³/ha/year]) by h.t. as being more meaningful than S.I. as estimates of site productivity. Brickell (1970) states “yield capability, as used by Forest Survey, is defined as mean annual increment of growing stock attainable in fully stocked natural stands at the age of culmination of mean annual increment.” We agree that yield capability is a more meaningful comparison, but only when some estimate of confidence can be placed on the values. The curves presented by Pfister and others (1977) for relating S.I. to yield capability by species have relatively low R² and unspecified standard error of estimate values. Given the relatively low or uncertain confidence to be placed in these curves, the fact that these curves do not apply to S.I. values greater than 80 (commonly exceeded in northern Idaho), and the misuse to which the values generated have been applied, we feel it is best to refrain from their presentation at this time; again this is an area for which individual National Forests (or other agencies) could profitably generate their own conversions. Research is sorely needed to develop the relationships between site variables, h.t., species site index, basal area stockability, and productivity.

<< Back to top

<< Back to Index