Rd. Yanai et al., Challenges of measuring forest floor organic matter dynamics: Repeated measures from a chronosequence, FOREST ECOL, 138(1-3), 2000, pp. 273-283
Accurate estimates of the retention of carbon in forest soils following for
est disturbances are essential to predictions of global carbon cycling. The
belief that 50% of soil carbon is lost in the first 20 years after clearcu
tting is largely based on a chronosequence study of forest floors in New Ha
mpshire northern hardwoods (Covington, 1981). We resampled forest floors in
13 stands in a similar chronosequence after an interval of 15 years. The t
hree youngest stands, which were predicted to lose organic matter over this
time, did not exhibit the 40-50% losses predicted by the original chronose
quence. The oldest stands had about twice as much organic mass in the fores
t floor as those cut recently, but this pattern could be explained equally
well by historical changes in the nature of forest harvest as by the age of
the stands. For example, mechanized logging probably causes more mechanica
l disturbance to the forest floor than horse logging, burying more organic
matter into the mineral soil. Markets for forest products and the intensity
of harvest removals have also changed over time, possibly contributing to
lower organic matter in the forest floor in young stands. In any chronosequ
ence study, effects of change in the nature of the treatment over time can
easily be misinterpreted as change with time since treatment; Repeated samp
ling of the chronosequence provides controls for some of these effects. Ln
the case of forest floor organic matter, however, high spatial variation ma
kes it difficult to distinguish whether the observed variation is more like
ly due to changes in treatment over time or to differences in time since tr
eatment. Because of the large amounts of carbon involved, small changes in
rates of soil organic matter storage may be quite important in global clima
te change, but they will remain difficult to detect, even with very intensi
ve sampling. (C) 2000 Elsevier Science B.V. All rights reserved.