Development of current stand structure in dry fir-pine forests of eastern Washington

Journal of the Torrey Botanical Society, Apr-Jun 2007 by Everett, Richard, Baumgartner, David, Ohlson, Peter, Schellhaas, Richard, Harrod, Richy

Adding time required from germination to the coring height improved the estimate of total tree age and allowed at least a decadal level of resolution in reporting results (Wong and Lertzman 2001). The accuracy in determining the exact mortality date of individual deadwood pieces declined with increasing decay as time intervals within later decay stages increased (Cline et al. 1980). We used the median time since mortality for all stems in a specific decay state; lesser and greater decay times of individual pieces should have balanced out. The median decay time should have remained constant for the population of snags or logs in each decay state at each sampling date. The amounts of deadwood stems and the patterns of occurrence should not have been affected by this approach, but the timing of patterns reflected the median time intervals to reach specific decay states.

Tracing small-diameter deadwood back in time was somewhat difficult and limited the accuracy of describing historic stand structure. However, Ful� et al. (1997) listed several factors that ameliorated this problem: 1) frequent fires limited the number of understory trees present in historical forests, 2) understory in open historical stands was not subject to competitive stress mortality for several decades following the last fire event, and 3) slow decay rates insured that deadwood from trees that died in the late 1800s persist today (Covington and Moore 1994b, Arno et al. 1995). As self-thinning mortality was not significant on our sites until the 1930s (see deadwood accumulation, this paper), any historical trees that died in the early-to-mid 1900s should have been of sufficient diameter to persist until our sampling period. However, undetected mortality of very small trees would have caused an underestimate in tree density.

One or more cross-referenced fire scars previously analyzed (Ohlson 1996, Everett et al. 2003b, Schellhaas et al. 2003) from these plots provided a composite median fire-free interval estimate for each plant association group. Our interest was not to duplicate the already extensive fire regime studies in this forest type, but only to provide information on the general fire regime of each plant association group and to document the last fire event in sampled plots.

ANALYSIS. Peaks and troughs in tree recruitment were presented graphically for the entire structural record (1400-160Os to present) in each plant association group and indicated the source of current live and deadwood structure. Live tree, snag, and log data from all stands within a plant association group were pooled and significant periods of tree establishment and mortality noted for the plant association group as a whole. Quantitative descriptions of tree abundance, basal area, and deadwood stems were provided from the historical period of 1860 to 2000. Also, we estimated the number of understory trees (seedlings and sapling

Because time sequence data was not independent, a paired plot i-test of differences analysis (Little and Hills 1978) was used to test for significant increases and declines in live tree density at four sampling periods (1860 = historical, 1915 = effective fire suppression, 1935-1965 = maximum tree density, and 2000 = current conditions). A paired plot r-test analysis was similarly used to evaluate basal area contributions from trees established preand post- 1860 (historical reference) and preand post-1915 (fire suppression reference). Change in age-class distribution, evenness in tree abundance among age-classes (modified Hills ratio, equation 5, Ludwig and Reynolds 1988: 94-97), and change in the range of tree ages present were analyzed for historical, maximum stand density, and current periods.