Explaining Variation: Extractable P in Test Pit Samples Alexandre Tokovinine

1
Explaining Variation Extractable P in Test Pit
Samples Alexandre Tokovinine
PROJECT GOALS
1) Soils samples to determine P
concentration in soils were taken with soil
coring probes. Due to soil compaction and
root/rock interference, most cores were about
35-45 cm long what corresponded to 100 cm of
probe length. In most cases, only two samples at
fixed intervals of 0-10 cm and 15-25 cm from the
top of the coring probe where taken from each
core to be analyzed for P concentration.
Therefore, one goal of pit samples was to
compare P concentrations at actual depths as
well as at A, Ap, and B soil horizons with
average core sample values. 2) Given that test
pit provided an opportunity for a more refined
sampling Strategy, another goal was to find out
whether there is a correspondence between soil
depth and P concentration

• PROJECT METHODS
• A 1X1X1 m test pit was excavated in each of the
  four main land use
• areas (Map 1a-d) cultivated, mowed-improved
  pasture, unimproved
• pasture, and woodlot. Pit profiles reflecting the
  sequence of soil
• horizons were drawn (Figure 2)
• 2) One set of soil samples were taken at regular
  10 cm intervals
• from the top of each pit down to the full depth
  of the pit (Figure 1)
• 3) Another set of samples were taken from A, Ap
  (if present), and B
• soil horizons
• 4) Samples were dried and stored following same
  procedures as
• for core samples
• 5) Extractable P concentrations were determined
  following same
• procedures as for core samples
• 6) Total P concentrations were determined for
  some of the samples

Figure 2 Test Pit Profiles
Samples by Depth The situation with P
concentration values in samples by
depth resembles the case of horizon samples
outlined in the previous box. Average core values
and pit values for 0-10 cm samples are very
similar for each land use type (Figure 5). There
is more variation for 15-25 cm average core
sample values and potentially corresponding 20
and 30 cm pit samples, with cultivated area pit
samples being the most divergent. Again, the
explanation likely lies in high variation in
cultivated area samples with the pit
representing just one extreme
set of P concentration values
also attested in nearby cores. In other
words, pit samples do not
provide information Figure 5 P
concentrations (mg P / kg soil) by depth (by
depth that would invalidate core
samples vs. average core values)
sampling technique. Another important
observation that one can make even without
relying on sophisticated statistics is
that the variance with depth for test pit
samples in woodland, improved pasture,
and pasture is greater than the variance Figure
6 box plots of P concentration for pit
between the groups of pit samples for each of
these land use areas (Figure 6). samples by
depth from different land use areas The only
set of samples that really stands apart is from
cultivated area.
Samples by Soil Horizon P
concentration values (mg P / kg soil) in pit
samples by soil horizon were compared to
average P concentration values in 0-10 and 15-25
cm core samples of corresponding land use
areas (Figure 1). There is high
similarity in P values in A horizon 0-10
cm and B horizon - 15-25 cm samples
respectively for unimproved pasture and
woodlot. The situation with
cultivated area and improved pasture
is more complicated. Given the depth
and extent of Ap horizon (Figure 2), one
would expect 0-10 cm values to correspond
to Figure 3 P concentrations (mg P / kg soil)
by Ap horizon and 15-25 cm to an depth (horizon
samples vs. average core values) average of Ap
and B values. However, the P values for
cultivated area pit horizon samples are much
lower than 0-10 and 15-25 cm averages, while
horizon values of improved pasture are lower for
A/Ap compared to 0-10 and much higher for B
compared to 15-25. A possible explanation is that
cultivated and improved pasture areas
core samples have much higher variation than
woodlot/unimproved pasture samples (S. D. of 123
and 83 mgP/kg soil vs. 56 and 20 mgP/kg soil ).
Indeed, comparing P values of horizon samples
with those of nearby cores (Figure 4 Map 1)
shows greater similarity for cultivated area
between core 6 and pit samples suggesting
Figure 4 P concentrations (mg P / kg soil) by
depth that there are zones of different P
concentration within land use areas.
(horizon samples vs. cores 6 and 14)
a
b Map 1 Test Pits
Location (Courtesy of David Diaz) (a) cultivated
area (red arrow) and mowed/improved pasture
(black arrow) (b) woodlot (blue arrow) (c)
unimproved pasture (yellow arrow) Woodlot is
not distinguished from unimproved pasture on the
map c
P Concentration by Depth Observable Trends A
brief look at the distribution of P values by
depth (Figure 5) suggests two trends while
cultivated area test pit P values rise sharply
between 40 and 60 cm, P values for other test pit
show a rise in P values with depth. The latter
trend is statistically significant and can be
described as linear regression (r20.432
plt0.001) explaining slightly less than a half of
total variation (Figure 7). Given small
sample size, it is pointless to provide a more
sophisticated statistical model to explain the
relationship between depth and P values for the
improved/unimproved pasture and woodland. No
model for the relationship between P values and
depth for cultivated area pit samples can be
proposed for the same reason. Figure 7
scatter plot of P concentrations vs. depth
(except Higher concentrations of extractable P at
greater depth, especially in the cultivated area)
with regression line and 1 S.D. area shown area
of woodlot and unimproved pasture (where these
values have clearly nothing or little to do with
manure application) are of special interest. This
trend was not observed for the only full set of
total P concentration samples taken from the
woodlot area pit. Therefore, it seems plausible
to suggest that higher P concentrations have
something to do with different soil properties at
greater depth.
P in soils Pools and Factors There are three
different pools of P in soils (Busman et al.
2002 Filipelli 2002 McGechan and Lewis 2002
McGechan 2002). When solution P (H2PO4- in
acidic conditions) enters soil as a result of
organic matter decomposition or manure
application, some of it is immediately taken by
plants and some is adsorbed onto soil particles.
In the latter case, it becomes part of the
active (non-occluded) P pool from which P can be
easily released when the particles are dissolved
in water. However, P can be adsorbed onto Fe,
Al, and Mg oxyhydroxides (occluded P) and as a
result become much harder to dissolve. Reducible
oxyhydroxides Figure 8 P sorption
depending on soil lower P provide large P
storage capacity because of their large surface
area and availabilityhigher sorption onto Fe
and Al oxides numerous delocalized positively
charged sites. Soil pH is the key factor for
(Busman et al. 2002 Fig. 3) P sorption onto
oxyhydroxides (Busman et al. 2002) since sorption
will occur within particular pH ranges lt3 for
Fe and 3-6 for Al oxides (Figure 8). For the
current project, acid extraction of P means that
extractable P comprises both P adsorbed onto
surface particles and P adsorbed onto Fe and Al
oxyhydroxides. Therefore, variation in mineral
content (mineral vs. organic content and
availability of Fe and Al oxides) and pH may be
the factors determining variation in P
concentrations in soils.
Harvard Forest Soils pH S Com
pton and Boones (2001) research on the impact of
past agricultural practices on the C, P,
and N content of New England soils provides
some information about the pH of the
topsoil and the mineral soil at a depth of
0-15 cm for the local soil types (Figure
10). Figure 9 soil types at Prospect Hill
(Compton and Boone 2001Tab.1) Most soils
of Prospect Hill farm track studied by the
authors are Canton series, while some are
Scituate (Figure 9). According USDA and New
England Soil Survey data, Canton soils
range from pH 4.5 to 6.0 with change in pH
with depth, while Scituate range from pH 3.5 to
6.0 with higher pH in substratum. These
data fit the mean pH values for mineral
soil as reported by Compton and Boone
(Figure 10) although the authors do not
provide any data on entire spectrum of pH
values. Therefore, given Al oxide pH 4-6
sorption window (other things being Figure 10
Soil properties for lt5.6 mm and 015 cm depth
equal) one may expect a more pronounced
increase in occluded P (Compton and Boone
2001Tab.2) with depth in Scituate soils.
References Busman, L., Lamb, J., Randall, G.,
Rehm, G., and M. Schmitt 2002 The nature of
Phosphorus in Soils. www.extension.umn.edu/distrib
ution/cropsystems/DC6795.html Compton, Jana E.
and Boone, Richard D. 2001 Long-Term Impacts of
Agriculture on Soil Carbon and Nitrogen in New
England Forests. Ecology Vol. 8123142330.
Eghball, B., G.D. Binford, and D.D.
Baltensperger 1996 Phosphorus movement and
adsorption in a soil receiving long-term manure
and fertilizer application. Journal of
Environmental Quality 251339-1343 Filippelli,
Gabriel M 2002 The Global Phosphorus Cycle. In
Phosphates Geochemical, Geobiological, and
Materials Importance, Mathew J. Kohn, John
Rakovan, and John M. Hughes, eds., pp. 391-425.
Reviews in Mineralogy and Geochemistry, Vol. 48.
Kleinman, P. A., Needelman, B. A., Sharpley,
A. N., and McDowell, R.W. 2003 Using Soil
Phosphorus Profile Data to Assess Phosphorus
Leaching Potential in Manured Soils Soil Science
Society of America Journal 67215-224 McGechan,
M. B., D. R. Lewis 2002 Sorption of Phosphorus by
Soil, Part 1 Principles, Equations, and Models.
Biosystems Engineering 821-24 McGechan,
M.B. 2002 Sorption of Phosphorus by Soil, Part
2 Measurement Methods, Results, and Model
Parameter Values. Biosystems Engineering
82115-130 United States Department of
Agriculture http//ortho.ftw.nrcs.usda.gov/ Soil
Survey Data for New England States
http//nesoil.com/
Conclusions 1) Horizon and by-depth soil samples
taken from test pits in four different land use
areas suggest that the results of sampling by
soil coring probes are representative of
variation in corresponding depth and soil
horizons despite compression and resulting
averaging of P values during sampling 2) An
increase in P concentration values with depth was
documented for the samples taken from pits in the
areas of woodlot, mowed/improved, and unimproved
pasture, but not in the cultivated area. This
difference was tentatively attributed to P
adsorption onto Al oxides depending on the
variation in pH with depth in different soil
types (Canton vs. Scituate). The most obvious
theoretical caveat of this interpretation is that
P leaching may follow preferential flow pathways
creating P distributions that have little to do
with pH or Al oxihydroxide concentrations. In
general, more research is needed to verify this
hypothesis, especially a better identification of
soil types and a study of Al oxide concentration
and pH through the entire profile in each of the
land use areas, preferably for more than one test
pit per area.
Speculations and Caveats Interestingly, the depth
and color of the soil horizons in the soil
profile for unimproved pasture (Figure 2, 11)
resemble Scituate as described by USDA (note
thick A and strong-brown Bw1, yellowish brown
C1), whereas cultivated area pit profile, (Figure
1, 2), despite the fact that upper horizons are
disturbed, fits Canton description (Bw1-Bw1 from
yellowish brown to light yellowish brown, olive
grey C1). This interpretation is speculative
since other important parameters such as soil
texture were not considered. However, the
possibility that variation in P concentration
with depth may be caused by variation in pH
depending on soil type is intriguing. There are
two additional caveats. First, the role of tree
roots that can slowly dissolve occluded P by
directly excreting phosphatase or with the help
of mycorrhizae (Filippelli 2002) has not been
taken into consideration. It is worth noting that
areas of high uniformity (40-60 cm) or variation
in P concentrations (above and below) may be
caused by tree roots. Secondly, the very
mechanism of P leaching remains debatable. Some
researches argue for a kind of matrix leaching of
P into subsoil evidenced in P concetrations
(Eghball et al. 1996). However, actual leaching
experiments suggest that P is transported
downwards via preferential flow pathways
resulting in highly variable P concentrations
(Kleinman et al. 2003). The latter interpretation
seems probable in our case because of attested
high variation in P concentrations, although
Kleinman et al. experimented with alkalic, not
acidic soils, where sorption onto Fe and Al
oxides is not supposed to be as important because
of high pH
Figure 11 Test Pit 4 profile (unimproved pasture)