eLearning

Over the past few decades, more than a dozen
studies have reported rates of N accretion in forest
soils that are beyond our current expectations of
rates of atmospheric N deposition, some by more
than an order of magnitude. The strength of evi-
dence varies among these studies from question-
able, unreplicated chronosequences to very care-
fully designed resampling of individual plots. These
reports of “occult” (as in: hidden from view, unex-
plained) N accumulations have not been examined
carefully relative to the prevailing paradigm that
assumes the major components of forest N cycles
are well understood. Possible explanations for oc-
cult N inputs include unexpectedly high rates of
nitrogen fixation; unexpectedly high rates of atmo-
spheric N deposition; analytical discrepancies; and
unknown processes.
We think that large, occult inputs of N to forests
are unlikely. Reports of anomalously high inputs
have not been critically examined, and it is possible
that the scientific community has overlooked a real
phenomenon. In this review, we examine reports of
N accretion in forest soils and evaluate the strength
of the experimental designs and overall evidence
for unexpectedly large accretion rates for N.
THREE APPROACHES TO ESTIMATING
RATES OF NCHANGE
Five decades of intensive research in forest soils and
forest nutrition have produced surprisingly few case
studies of rates of change in total soil N content
(Fisher and Binkley 2000). Most studies of changes
in soil N focused on documenting changes from
nitrogen-fixing species, including chronosequence
studies (see Chapin and others 1994 for the best,
replicated example), and comparisons of soils under
N-fixing and non–N-fixing trees (reviews by Bin-
kley and others 1994; Binkley and Giardina 1997).
In this review, we omit case studies that included
symbiotic N-fixing plants and focus instead on sit-
uations where N inputs would be expected to be
relatively minor.
Three approaches have been used to assess N
accretion in forest soils: paired plots, chronose-
quences, and remeasurement of individual sites.
Paired plots contrast soil N content, at a single sam-
pling time, under the influence of different vegeta-
tion types. Any difference in soil N should reflect
differences in vegetation N content, inputs, or out-
puts from the system. The chronosequence ap-
proach evaluates a variety of locations that differ in
age of the vegetation (or soil). If an assumption that
other site factors do not differ substantially is valid,
then differences in forest N content at one time