An initial core was collected using a steel slide-hammer with a barrel
6.3 cm in outside diameter, to a depth of 20cm (+/- 10%); soil was
discarded outside the plot. Three 6-mm aluminum rods were inserted
around the circumference of this core, and the volume was filled with
air-dry soil collected from locations outside the study plots, which
had been twice passed through a 1.4 mm sieve to remove intact roots.
The core volume was filled slowly and gently compacted to prevent
excessive settling during incubation. Separate mineral and organic
root-free soils were prepared, and each core was filled with these to
approximately match the surrounding soil horizons. Leaf litter was
pinned to surface using the aluminum rods, to prevent erosion of the
loose soil and to allow soil moisture to reach an equilibrium with the
surrounding intact soil. Cores were generally installed in the fall
after trees became dormant but before the first substantial snowfall,
and harvested at the same time the following year to represent one
full growing season. The center of each core ingrowth was harvested
with a sharpened PVC pipe, which was 4.05 cm in inside diameter,
immediately placed on ice, and then frozen until lab processing could
occur.
Reference ingrowth cores were installed in each of the three root
sampling subplots associated with each of the CCASE reference plots.
Due the need to avoid destructive soil sampling within the CCASE
reference plots, these root sampling subplots were located outside
(generally to the S and W) of each of the CCASE reference plots.
Beginning in 2019, reference plots 9 and 10 were used instead, as it
had become clear by this point that plots 1 and 2 had different soil
properties than 7 and 8.
Within the DroughtNet plots, ingrowth cores were installed in subplots
reserved for destructive soil sampling. Within each of these, during
years with 50% throughfall exclusion, a pair of cores was installed at
each time step, one directly under a trough (labeled “C” for
“covered), and one directly under a gap between troughs (labeled “O”
for “open”). In reference plots or in TFE plots during 95% treatment
years, this level of replication was simply labeled “A” and “B”.
Note that the 2015 data do not represent a full year of ingrowth
(cores installed in May and harvested in late October). The other two
years were complete 12-month incubations (usually October - October).
No cores were installed for the 2018 growing season. For the 2019
growing season, only reference plots 9 and 10 were used for ingrowth
cores. For the 2020 growing season, ingrowth cores were mostly moved
out of the ingrowth-core subplots, in cases where no undisturbed area
remained that would accommodate a 20 cm deep core. Cores were instead
installed in comparable microtopography elsewhere in the plot, and
labeled according to the nearest tagged tree.
The number of samples successfully harvested each year varied due to
occasional disturbance by small mammals. Cores that lost more than 10%
of their volume (i.e. 2cm of depth on average within the central 4cm
of the core) were not collected.
Samples were frozen for storage and thawed shortly before processing.
Samples were gently wet-sieved on a 2 mm sieve and intact root
branches were gently rinsed free of soil. Water and soil that passed
the sieve were carefully checked for root fragments >5mm in length
(smaller fragments were considered more likely to be a component of
the sieved soil than true ingrowth). Root samples were then dried at
60ºC. Dried roots were sorted into “fine” and “thick” size classes
based on whether they were greater or less than 1mm in diameter (after
drying). The large variance in thicker roots makes the fine size class
more appropriate for the statistical detection of differences. Data
are expressed as grams of dry matter per square meter per year, based
on the inside area of the sampling corer, which was 12.88 cm2.
