FOR ONLINE PUBLICATION ONLY
Appendix A Description of Data Used for Simulation of Decay Class Transitions Across the
Eastern US
The simulation dataset chosen for this analysis was a DWD inventory collected across 23 eastern
US states in 2001 and independent of datasets used in related studies (Woodall and others 2012;
Russell and others 2013). Each plot consisted of four 7.32-m (24.0-ft) fixed radius subplots for a
total plot area of approximately 0.07 ha where tree and site attributes were measured. Downed
woody pieces were defined as DWD in forested conditions with a diameter greater than 7.62 cm
(3.0 in) along a length of at least 0.91 m (3.0 ft) and a lean angle greater than 45 degrees from
vertical. Individual DWD pieces were sampled using a line-intercept sampling method (Van
Wagner 1968) on 18.0-m (58.9-ft) horizontal distance transects radiating from each FIA subplot
center at azimuths of 30, 150, or 270 degrees. Only two transects from the three azimuths were
sampled within each of the four FIA subplots depending on spatial arrangement, totaling 143.6 m
for an entire inventory plot. Data collected for every DWD piece included small-end (DIASM),
and large-end (DIALG) diameters, DC, length (LEN), species, and piece location (that is, plot,
subplot, and transect number; horizontal distance along a sampling transect). Piece LEN was
defined as the total length of the DWD piece in m. For DC 5 pieces, species was not identified
and end diameters were not measured due to inability to assign species with advanced decay and
substantial reductions in taper (Woodall and Westfall 2008). In total, 4,384 DWD pieces from
516 plots were collected from 32 conifer and 87 hardwood species as part of the 2001 inventory
(see Table 1 in manuscript text).
Assigning a length to the DWD pieces to initiate the simulation required the testing of some
important assumptions related to log decay. We hypothesized that for DWD in advanced stages
of decay (for example, DC 3, 4, and 5) length may have shortened through breakage or
fragmentation. Hence, we tested for differences in observed length (that is, the 2001
measurement) and predicted length assuming a DC 1. Length was predicted using equations that
employed observed DC (1 through 5), DIASM and DIALG as independent variables (Woodall and
others 2008). For the purposes of simulating DC dynamics of the DWD population measured in
2001, we considered all DWD pieces to be non-decayed. This included setting their length equal
to their 2001 measurement and assuming no structural loss due to decay. We refit models
predicting length using a multiple linear regression with DC, DIASM, and DIALG as independent
variables (Woodall and others 2008; their Eq. 3) using DWD data measured at recent FIA
inventories in 2002-2010 (Woodall and others 2012) using multiple linear regression. Predictions
of length for conifer (residual SE = 4.03 m; R2 = 0.33) and hardwood species (residual SE = 3.86
m; R2 = 0.37) were tested against observed length using a two-one-sided test (Wellek 2003) in an
equivalence testing framework (Robinson and Froese 2004). Results indicated no dissimilarity
between observed and predicted length, indicating that setting length equal to the observed
length measured in 2001 was appropriate to initiate the simulation. The DD5 variable, MAT, and
additional climate information were obtained by specifying latitude, longitude, and elevation of
each FIA plot location to a spline surface model developed from climate station data across
forests of North America (Rehfeldt 2006; USFS 2012).
References
Rehfeldt, G.E., 2006. A spline model of climate for the western United States. USDA For. Serv.
Gen. Tech. Rep. RMRS-165.
Robinson, A.P., Froese, R.E., 2004. Model validation using equivalence tests. Ecological
Modelling 176, 349-358.
Russell, M.B., Woodall, C.W., Fraver, S., D'Amato, A.W., 2013. Estimates of coarse woody
debris decay class transitions for forests across the eastern United States. Ecological Modelling
251, 22-31.
United States Forest Service, 2012. Research on forest climate change: potential effects of global
warming on forests and plant climate relationships in western North America and Mexico. Rocky
Mountain Research Station, Moscow Laboratory. http://forest.moscowfsl.wsu.edu/climate/ (last
accessed 28 June 2012).
Van Wagner, C.E., 1968. The line-intersect method in forest fuel sampling. Forest Science 14,
20-26.
Wellek, S., 2003. Testing statistical hypotheses of equivalence. Chapman and Hall, London.
Woodall, C.W., Walters, B.F., Westfall, J.A., 2012. Tracking downed dead wood in forests over
time: development of a piece matching algorithm for line intercept sampling. Forest Ecology and
Management 277, 196-204.
Woodall, C.W., Westfall, J.A., 2008. Controlling coarse woody debris inventory quality: taper
and relative size methods. Canadian Journal of Forest Research 38, 631-636.
Woodall, C.W., Westfall, J.A., Lutes, D.C., Oswalt, S.N., 2008. End-point diameter and total
length coarse woody debris models for the United States. Forest Ecology and Management 255,
3700-3706.
Appendix C Decay rate constants k predicting the annual decomposition of conifers and
hardwoods located in four contrasting climate regimes, grouped by their mean annual
temperature (MAT; °C). Curves displayed assume a 100-kg downed woody debris piece after
using a decay class and volume reduction factor approach.
Appendix D Decay rate constants k predicting the annual decomposition for selected species
using a decay class and volume reduction factor approach. Species displayed are for the
minimum, mode, and maximum k values observed for conifers (black) and hardwoods (gray)
assuming a 100-kg downed woody debris piece.