The patterns and causes of temporal changes in forest production and standing biomass are fundamental questions for both basic and applied ecology (Peet 1981, Ryan et al. 1997). Although long-term observation of permanent plots is the most appropriate means to gather information on these questions, relevant data and analyses are few (Peet 1981, Ryan et al. 1997). Even where permanent plot records exist, in most cases these data do not cover the entire duration of stand development. In the Pacific Northwest, for example, many plots established early in the 20th century to quantify timber yield of second-growth stands continue to be measured (Williamson 1963, Acker et al. 1998). Although these plots were established before the peak of timber production, given standards of timber utilization at the time, the plots were established after the peak of production of total tree volume or biomass (Munger 1946, Curtis 1992). Thus, it would be valuable to have some means of inferring trends prior to establishment of such plots based on and constrained by the existing measurements.
In this note I describe an objective means of fitting curves of bole production and mortality to such datasets. This enables one to reconstruct the early dynamics of stands with long-term plot records, from the time of stand establishment to the initiation of plot measurements. In this way it is possible to explore the logical implications of long-term observations for the earlier stages of stand dynamics. In addition to observations of live trees, our approach depends on measurement of coarse woody debris (CWD) mass at some point during the interval of live-tree measurements. Information on the quantity of CWD produced since the last stand-replacing disturbance is critical in constraining the fitted curves, inasmuch as at any point in time, the sum of standing biomass and CWD represents the integrated trends of tree production, tree mortality, and decomposition of CWD.
This approach is by no means a mechanistic treatment of the processes of growth, mortality, and decay. Rather, it is a logical framework for exploiting long-term data on forest dynamics to explore probable patterns of productivity and biomass accumulation during the early decades after stand-replacing disturbance.
The basis of the approach is a set of simple assumptions that are, for the most part, based on published studies and review papers:
To illustrate these assumptions (and also to allow for testing of the approach, see below), I developed hypothetical time trends of biomass, mortality, CWD, and bole production based on asymptotic accumulation of biomass towards an upper limit along a sigmoid curve (Peet 1981; Fig. 1). Illustrated are the assumptions of a lag before appreciable bole production (Fig. 2), a subsequent steep increase of bole production followed by a more-or-less linear decline (Fig. 3), a lag before bole mortality (Fig. 4), and a subsequent approximately linear increase in mortality (Fig. 5).
The five assumptions underlying the approach have been assembled in the attached QuickBASIC program ( "PRODCURV.BAS"). As an example, we used the program to fit curves of bole production and mortality to two sets of plots in western hemlock-Sitka spruce forest on the Oregon coast (see Harcombe et al. 1990, Greene et al. 1992, Acker et al. in prep.). We set the time lag before appreciable bole production to 10 years (cf. Binkley and Greene 1983) and the decay rate of CWD to 3% per year (Sollins 1982, Spies et al. 1988). We then iteratively modified the time lag before tree mortality to match observed CWD for the two plot sets. Then we iteratively modified the timing and magnitude of the peak in bole production to match standing biomass. We started this step by testing whether the first observed value of bole production could be the peak, and then increased peak production and shifted it earlier as necessary.
The resulting fitted and observed curves of bole production, mortality, standing biomass, and CWD are displayed in Figs. 6-9. According to the reconstruction, production of boles has been higher on the "CHEF" plots (Cascade Head Experimental Forest) than on "NCRNA" plots (Neskowin Crest Research Natural Area) for most of the history of the stand. The difference in production during the period when both sets of plots were measured was not large enough to account for the current difference between the plot sets in standing biomass. Furthermore, CWD amounts were similar on the two sets of plots. As a consequence, reconstructed mortality trends are similar for the two plot sets. Differences in standing biomass are mostly due to much higher production on the CHEF plots from about stand age 20 to 100, an interval almost entirely without measurements. This should be viewed as a plausible, parsimonious reconstruction. Other scenarios would require invoking more complicated sets of assumptions.
Test datasets
The data files for the CHEF and NCRNA plots are available for testing this approach using PRODCURV.BAS (the files are "chefrate.txt" and "ncrnrate.txt", respectively). The header lines of these files contain measurement information to enter interactively: stand age and biomass at first measurement, and the stand age at which CWD was measured and the observed value.
I have also generated a hypothetical time series of biomass, mortality, CWD, and bole production so that reconstructed time trends can be compared to a known, complete time trend. The hypothetical trends were constructed so that biomass would asymptotically approach a plateau on a sigmoid curve, and so that bole production would peak relatively early and then decline noticeably. The shape of the biomass trend was accomplished with a Chapman-Richards function (Richards 1959); parameter values were adjusted to produce a peak and decline of bole production. Tree mortality was set as a constant fraction of bole biomass (1% per year). Decomposition of CWD was also set as a constant (3% per year).
The resulting simulated data are available in the file IDLBIOM1.TXT (see also Figs. 1-5). As a test, small segments of the data can be used as "measurement values" (e.g., from stand age 80 to 120 years). One can then attempt to reconstruct the earlier phases of stand dynamics using PRODCURV.BAS as described above. Finally, the reconstructed curves of bole production, mortality, and biomass can be compared to the "real" data.
The objective approach to curve-fitting presented in PRODCURV.BAS enables one to explore the implications of observations of coarse woody debris and tree growth and mortality on the temporal patterns of bole production and mortality during the aggrading phase of stand dynamics. The method makes use of a small number of explicitly stated assumptions that are amenable to empirical testing. The values of time lags and decay rates used in the example could be altered as needed for application to other systems.
Acker, S.A., T.E. Sabin, L.M. Ganio, W.A. McKee 1998. Development of old-growth structure and timber volume growth trends in maturing Douglas-fir stands. Forest Ecol. Manag. 104: 265-280.
Acker, S.A., P.A. Harcombe, M.E. Harmon, and S.E. Greene. Biomass accumulation over the first 150 years in coastal Oregon spruce-hemlock forest. in prep.
Curtis, R,O. 1992. A new look at an old question--Douglas-fir culmination age. W. J. Appl. For. 7(4): 97-99
Binkley, D., and S. Greene. 1983. Production in mixtures of conifers and red alder: the importance of site fertility and stand age. Pp. 112-117 in R. Ballard and S.P. Gessell tech. eds. IUFRO sympoisum on forest site and continuous productivity. Gen. Tech. Rep. PNW-163, USDA For. Serv. PNW Forest & Range Exp. Sta. Portland, OR
Greene, S.E., P.A. Harcombe, M.E. Harmon, and G. Spycher. 1992. Patterns of growth, mortality and biomass change in a coastal Picea sitchensis-Tsuga heterophylla forest. J. Veg. Sci. 3: 697-706
Harcombe, P.A., M.E. Harmon, and S.E. Greene. 1990. Changes in biomass and production over 53 years in a coastal Picea sitchensis-Tsuga heterophylla forest approaching maturity. Can. J. For. Res. 20: 1602-1610
Munger, T.T. 1946. Watching a Douglas-fir forest for thirty-five years. J. Forestry 44 (10): 705-708.
Peet, R.K. 1981. Changes in biomass and production during secondary forest succession. Pp. 324-338 in H.H. Shugart, D.C. West, and W.R. Emanuel, eds., Forest succession: concepts and applications. Spring-Verlag, New York.
Richards, F.J. 1959. A flexible growth function for empirical use. J. Exper. Biol. 10:290-300.
Ryan, M.G., D. Binkley and J.H. Fownes. 1997. Age-related decline in forest productivity: pattern and process. Adv. Ecol. Res. 27: 213-262
Sollins, P. 1982. Input and decay of coarse woody debris in coniferous stands in western Oregon and Washington. Can. J. For. Res. 12: 18-28
Spies, T.A., J.F. Franklin, and T.B. Thomas. 1988. Coarse woody debris in Douglas-fir forests of western Oregon and Washington. Ecology 69: 1689-1702
Williamson, R.L. 1963. Growth and yield records from well-stocked stands of Douglas-fir. Res. Pap. PNW-4. USDA Forest Service Pacific Northwest Forest and Range Exp. Sta., Portland, OR. 24 p.