BGBiogeosciencesBGBiogeosciences1726-4189Copernicus PublicationsGöttingen, Germany10.5194/bg-13-6183-2016Changes in soil carbon and nutrients following 6 years of litter removal and addition in a tropical semi-evergreen rain forestTannerEdmund Vincent Johnevt1@cam.ac.ukSheldrakeMerlin W. A.TurnerBenjamin L.Department of Plant Sciences, University of Cambridge, Downing St, Cambridge CB2 3EA, UKSmithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Ancon, Republic of PanamaEdmund Vincent John Tanner (evt1@cam.ac.uk)17November201613226183619025May201629July201627October20161November2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://bg.copernicus.org/articles/13/6183/2016/bg-13-6183-2016.htmlThe full text article is available as a PDF file from https://bg.copernicus.org/articles/13/6183/2016/bg-13-6183-2016.pdf
Increasing atmospheric CO2 and temperature may increase forest
productivity, including litterfall, but the consequences for soil organic
matter remain poorly understood. To address this, we measured soil carbon
and nutrient concentrations at nine depths to 2 m after 6 years of
continuous litter removal and litter addition in a semi-evergreen rain
forest in Panama. Soils in litter addition plots, compared to litter removal
plots, had higher pH and contained greater concentrations of KCl-extractable nitrate (both to 30 cm); Mehlich-III extractable phosphorus
and total carbon (both to 20 cm); total nitrogen (to 15 cm); Mehlich-III
calcium (to 10 cm); and Mehlich-III magnesium and lower bulk density (both to 5 cm).
In contrast, litter manipulation did not affect ammonium, manganese,
potassium or zinc, and soils deeper than 30 cm did not differ for any
nutrient. Comparison with previous analyses in the experiment indicates that
the effect of litter manipulation on nutrient concentrations and the depth
to which the effects are significant are increasing with time. To allow for
changes in bulk density in calculation of changes in carbon stocks, we
standardized total carbon and nitrogen on the basis of a constant mineral
mass. For 200 kg m-2 of mineral soil (approximately the upper 20 cm
of the profile) about 0.5 kg C m-2 was “missing” from the litter
removal plots, with a similar amount accumulated in the litter addition
plots. There was an additional 0.4 kg C m-2 extra in the litter
standing crop of the litter addition plots compared to the control. This
increase in carbon in surface soil and the litter standing crop can be
interpreted as a potential partial mitigation of the effects of increasing
CO2 concentrations in the atmosphere.
Introduction
Tropical forests and their soils are an important part of the global carbon
(C) cycle because they contain 692 Pg C, equivalent to 66 % of the C in
atmospheric CO2 (Jobbagy and Jackson, 2000). Carbon in tropical forest
soils is dynamic: Schwendenmann and Pendall (2008) reported a turnover time
of 15 years for the “slow” pool of soil C, comprising 38 % of the total
soil C, in the top 10 cm of soil in semi-evergreen rain forest on Barro
Colorado Island, Panama (61 % of total soil C was “passive” with a
turnover time of the order of a 1000 years). Turner et al. (2015)
reported an approximate 25 % increase in soil C from one dry season to the
next wet season in the top 10 cm of soil on the Gigante Peninsula in Barro
Colorado Nature Monument, Panama, at a site close to the current litter
manipulation experiment. Thus, there is the potential for the amount of C in
tropical soils to change over only a few years, with potentially important
consequences for atmospheric CO2 concentrations.
Atmospheric CO2 concentrations have been steadily increasing for
decades and one of the effects of this could be widespread increases in
forest growth (Nemani et al., 2003) and, as a result, increased litterfall.
There are few experimental studies of the effects of elevated CO2 on
forest growth. Körner (2006) reported that elevated CO2 caused
increased litterfall in one of three studies in steady-state tree stands in
temperate forests, but there have been no such studies in the tropics. Thus, the potential exists for increased CO2 to increase forest growth and
litterfall – though we do not know how widespread and how large any
increase in litterfall might be, especially in the tropics.
Soil C has been shown to respond to experimental changes in litter inputs.
In three studies in temperate forests in the USA, litter removal always
resulted in lower soil organic carbon, but litter addition had much more
variable effects, increasing in one (Lajtha et al., 2014a), not changing in
the second (Bowden et al., 2014) and decreasing in the third (Lajtha et al.,
2014b). The single study from the tropics, in lowland rain forest in
southwestern Costa Rica, reported decreased soil C in litter removal plots
and increased soil C in litter addition plots (Leff et al., 2012). It is
therefore likely that soil C will increase in many, but not all, forests as
a result of increased litter input.
The relative importance of above-ground or below-ground inputs as sources of
soil organic matter has been reassessed in the last decade (Schmidt et al.,
2011). Recently it was shown that 50–70 % of the soil organic matter in
boreal coniferous forest is from roots and root-associated microorganisms
(Clemmensen et al., 2013). The origin of the soil organic matter is
thus a question of the relative contributions of above-ground and
below-ground inputs. Litter manipulation experiments can provide insights
into this issue by controlling one source of C input – above-ground
litterfall.
Soil nutrients as well as C can change as a result of increasing or
decreasing litter inputs and are important because they will potentially
affect soil fertility. In Panama, mineralization of organic phosphorus (P)
(inferred from the decrease in the concentration of organic P) in the top 2 cm of soil during 3 years of litter removal was calculated to be
sufficient to supply 20 % of the P needed to sustain forest growth –
there were corresponding increases in organic P in litter addition plots,
and total nitrogen (N) showed a similar pattern (Vincent et al., 2010).
“Available” nutrients, including KCl-extractable ammonium (NH4) and
nitrate (NO3) and Mehlich-III extractable P, potassium (K), calcium
(Ca), magnesium (Mg) and micronutrients, all changed over 4 years in the
upper 2 cm of soil as a result of litter manipulation (Sayer and Tanner,
2010). After 6 years of litter manipulation, surface soils (0–10 cm) had
lower NO3 and K in litter removal plots and higher NO3 and Zn in
litter addition plots; other nutrients were not significantly affected
(Sayer et al., 2012). In Costa Rica after 2.5 years of litter manipulation,
surface soils (0–10 cm) had lower net nitrification in both litter removal
and addition treatments, while NH4 concentrations were significantly
lower in litter removal plots (NH4 was 83–91 % of the extractable
N; Wieder et al., 2013). Thus, several soil nutrients in surface soils change
following litter manipulation, but there is no consistent pattern for N,
very little data for P or cations (the latter were not reported for the
Costa Rican experiment), and no data for soils deeper than 10 cm.
Here we report results from the Gigante Litter Manipulation Plots (GLiMP)
experiment over a much greater soil depth (0–200 cm) for total C, N and P and for extractable (“plant-available”) N, P, K, Ca, Mg, manganese (Mn) and
zinc (Zn), measured after 6 years of continuous litter transfer. In
addition, we present a new way of expressing soil C (relative to the
unchanging mineral mass), which allows us to calculate overall changes in
soil C and other elements independently of changes in bulk density. Our
objective was to describe changes in C and nutrient concentrations in the
full soil profile and to calculate C budgets to discover the fate of the
increased C input in litter addition plots. In particular, we aimed to
calculate the proportion of the added C that remains in the soil and the
litter standing crop and can thus be considered as partial mitigation of
atmospheric CO2 accumulation through increased forest productivity
due to increased atmospheric CO2 and temperature – mitigation
because C that is not in the soil will be in the atmosphere as extra
CO2. No other study has tried to quantify the fate of C in organic
matter added to tropical forest soils, though a study of agricultural soil
in the temperate UK calculated that about 2.4 % of organic matter in annual
additions of farmyard manure was still in the soil after 120 years (Powlson
et al., 2011).
Materials and methods
The litter manipulation experiment is located in old-growth semi-evergreen
lowland tropical forest on the Gigante Peninsula (9∘06′ N, 79∘54′ W), part of the Barro
Colorado Nature Monument in central Panama. The experiment is located on the
upper part of the landscape, where soils are Oxisols (Typic Kandiudox).
Surface soils have a pH of 4.5–5.0 and low available P concentrations but
high base saturation and cation exchange capacity. Annual rainfall on nearby
Barro Colorado Island (ca. 5 km from the study site) is 2600 mm and average
temperature is 27 ∘C. There is a strong dry season from January to
April, with approximately 90 % of the annual precipitation during the
rainy season.
The experiment consists of fifteen 45 m × 45 m plots within a 40 ha area of
old-growth forest. In 2001 all 15 plots were trenched to a depth of 0.5 m to
minimize lateral nutrient and water movement via the root–mycorrhizal
network; the trenches were double-lined with plastic and backfilled.
Beginning in January 2003, litter (including branches < 20 mm in
diameter) was raked up once a month in five plots, resulting in low, but not
entirely absent, litter standing crop (litter removal plots). The removed
litter was immediately spread on five further plots (litter addition plots),
with five plots left as controls (CT plots). Treatments were assigned on a
stratified random basis using total litterfall per plot in 2002 (i.e. the
three plots with highest litterfall were randomly assigned to treatments,
then the next three and so on) (Sayer et al., 2007). The plots were
geographically blocked, litter from a particular litter removal plot was
always added to a particular litter addition plot and there was a nearby
control plot.
Soils samples were collected in January 2009, the early dry season, using a
7.6 cm diameter constant volume corer for the top 20 cm of soil and a 7 cm
diameter auger for 20–200 cm. Fresh soils were extracted for NO3
and NH4 within 2 h of sampling in a 2 M KCl solution, with
detection by automated colorimetry on a Lachat Quikchem 8500 (Hach Ltd,
Loveland, CO). Phosphorus and cations were extracted within 24 h in Mehlich
III solution and analysed by inductively coupled plasma optical emission
spectrometry (ICP-OES). Soil pH was measured on a 1:2 fresh soil solution in
distilled water.
Dried (22 ∘C × 10 d) and ground soil was analysed for total C and N by
combustion and gas chromatography on a Flash 1112 analyzer (Thermo, Bremen,
Germany). Total P was determined by ignition at 550 ∘C for 1 h and
extraction for 16 h in 1 M H2SO4, with detection by automated molybdate
colorimetry at 880 nm using a Lachat Quikchem 8500 (Hach Ltd, Loveland, CO).
Nutrient data were analysed using mixed-effects models, with “litter
treatment”, “depth” and their interaction as fixed effects and “plot” as a
random effect. Where nutrient concentrations varied non-linearly with depth,
we used splines with two or three knots. Some nutrients showed severe
heteroscedasticity, and we accounted for this in the model by using
“variance covariates”, which model the variance as a function of one or more
of the effects in the model (Pinheiro and Bates, 2000; Zuur et al., 2009). For
all nutrients, depth was modelled as a numeric predictor and log transformed
prior to analysis. We performed model selection based on likelihood ratio
tests and the Aikake information criterion with correction for small sample
sizes (AICc; Burnham and Anderson, 2002). We derived P values for fixed
effects by comparing null models to full models using likelihood ratio
tests. Final models were refitted using restricted maximum likelihood
estimation (REML) (Zuur, 2009). Where the treatment × depth term was
significant, we refitted the model omitting either the litter addition
treatment or the litter removal treatment to assess the contribution of each
of the treatments (litter addition and litter removal) to the overall
interaction term. Analyses were done in R version 3.1.2.
Amounts of soil total C and N were also calculated relative to soil mineral
mass to allow comparisons between the treatments where bulk density and soil
depth was changing due to removal and addition of litter; soil in litter
removal plots was shrinking and had increasing bulk density; soil in litter
addition plots was increasing in depth and had lower bulk density.
Expressing potentially changing elements relative to unchanging mineral mass
allows for change to be expressed against an unchanging reference; it is
analogous to expressing soil water relative to soil dry mass rather than
soil fresh mass. Soil organic C with depth was calculated for each plot by
fitting a line to cumulative soil organic C (Y) against cumulative soil
mineral mass (X). Bulk density data were measured for each plot only in the
top 0–5 cm for soil. Below that we used bulk density data for one pit only.
Bulk density below 10 cm depth does not vary much across the site; data for
four soil pits (not in any of the plots) have a coefficient of variation of
about 10 % for soils from 10–20 cm depth and 3 % for soils
from 20–50 cm depth), whereas coefficients of variation of bulk densities in
surface 0–5 cm soils were higher: control 12 %, litter addition 15 %
and litter removal 4.9 %. Bulk density data were used to estimate
approximate soil depth for control plots in Figs. 3 and 4. Statistical
comparisons of modelled cumulative total C against cumulative mineral matter
were compared by bootstrapping, using R version 3.1.2.
Results
Soils in litter addition plots, compared to litter removal plots, had
significantly lower bulk density (both to 5 cm) and higher NO3 and pH
(to 30 cm), PMeh and total C (both to 20 cm), total N (to 15 cm), Ca
(to 10 cm), and Mg (to 5 cm) (Figs. 1 and 2 and Supplement Tables S1 and S2). There
were fewer differences when compared to control soils: litter addition soils
had higher concentrations of PMeh (to 20 cm), NO3 (to 15 cm), Ca
(to 10 cm) and pH (to 10 cm). Nutrient concentrations in litter removal
soils were not significantly lower than those in controls. Nutrient
concentrations in soils > 30 cm deep did not differ significantly
for any nutrient. Thus, in some way total C, total N, NO3, PMeh,
Ca and Mg were significantly affected by litter removal or addition, but K,
Mn, NH4, Zn and were not; effect sizes (log response ratio for 0–5 cm
soils) decreased from 0.81 for NO3 to 0.39 for Ca, 0.27 for Zn, 0.20
for PMeh, 0.20 for Mg, 0.15 for Ctot and 0.11 for Ntot.
Concentrations of soil C, N, P (various fractions) and cations
(Mehlich extractions), plotted against the midpoint of the soil layers
sampled (Zn values should be divided by 1000 to obtain actual means); control points are displaced below treatments. Data are fitted values of the
mixed-effects models with 95 % confidence intervals (see the “Material and methods” section) in
litter removal •, control ∘ and litter addition
▾ plots.
Mean concentrations of ammonium and nitrate plotted against the
midpoint of the soil layers sampled; control points are displaced below
treatments. Data are fitted values of the mixed-effects models with 95 %
confidence intervals (see the “Material and methods” section) in litter removal •, control ∘ and litter addition
▾ plots.
All nutrients decreased in concentration with increasing soil depth. In
control soils, concentrations at 50–100 cm compared to 0–5 cm were as follows:
NH4 50 %, Mg 37 %, Ptot 36 %, K 32 %, PMeh 25 %,
NO3 24 %, Ntot 12 %, Ca 11 % and Ctot 11 %;
NO3 was only 24 % of the total inorganic N in controls (mean over
all depths) (Figs. 1 and 2 and Table S1). Concentrations of most elements
continued to decrease below 100 cm depth in the soil; those from 150–200 cm
were about half those from 50–100 cm (ranging from 14 % for Ca to 81 % for
NH4; Table S1).
Soil bulk density in the top 5 cm was significantly lower in litter addition
than litter removal, though neither was significantly different from the
controls. Soil C stocks standardized to a consistent mineral mass (i.e. that in
the control plots) was significantly greater in litter addition compared to
litter removal to about 10 cm depth in the soil (Figs. 3 and 4). Total N per
mineral mass of soil was also significantly greater in litter addition than
litter removal in approximately the top 10 cm of soil. In contrast, C : N
ratios changed little with depth. In control soils, C : N was about 10.5 near
the surface and 10.0 at 150–200 cm; in litter removal plots, C : N was 10.5
at the surface and 10.3 at depth, while litter addition soils were more
variable, with C : N being 11.7 at the surface and about 10.0 at 150–200 cm
depth.
Soil carbon content and mineral content in litter addition, control
and litter addition, expressed as kg C m-2 cumulatively from 0 to 30 cm
soil depth. Values are means for five plots per treatment ± SE; litter
removal •, control ∘ and litter addition ▾.
Differences in soil carbon content relative to control soils (mean
and SE; n=5), after 6 years of litter manipulation, plotted for
successive soil layers: 0–100 kg (mineral matter) m-2, plotted at 100 kg m-2
on right y axis; 100–200 kg m-2, plotted at 200 kg m-2;
and so on to 900–1000 kg m-2, plotted at 1000 kg m-2 in
litter removal • and litter addition ▾
plots. We calculated the soil C in the litter removal and litter addition
plots at the mineral mass equal to that at various depths in the control
plots (0–5, 5–10 cm, etc.). We then calculated the difference in C between
each litter removal (or litter addition) and its control plot for the same
mineral mass. Approximate depth for cumulative soil mineral mass in control
plots is shown on left y axis.
DiscussionSoil carbon dynamics
The amount of C “missing” from litter removal and “extra” in the litter
addition over about the top 20 cm of soil (from calculations based on C per
mineral matter), 6 years after (January 2009) litter removal and addition
started, was about 0.5 kg C m-2 (Fig. 3). These changes are ca. 1 %
per year; in contrast if we calculate the change based on a fixed
depth of 20 cm, ignoring changes in bulk density, we get a change of about
2 % per year. Thus, ignoring the changes in bulk density results in a
misleading doubling of the estimated rate of change. The similarity of the
losses from litter removal and gains in litter addition probably has
different causes: we speculate that losses from the soil in the litter
removal plots are due to respiration being greater than additions; we did
not physically remove organic matter from the mineral soil. We further
speculate that increases in C in the mineral soil in the litter addition
plots are a result of infiltration of dissolved and particulate organic
matter draining from the litter standing crop and/or changes in root
exudates; increases in root growth are not the explanation – root growth
was lower in litter addition plots (Sayer et al., 2006).
In addition to the extra soil C in the litter addition plots, the litter
standing crop was also larger in litter addition plots. In September 2005
(2.8 years after litter manipulation started), there was an additional
0.4 kg C m-2 in the Oi and Oe layers compared to control plots (Sayer and
Tanner, 2010), and data from 2013 show that litter standing crop was at about
this level (C. Rodtassana, University of Cambridge, unpublished data).
Together this extra 0.9 kg C m-2 in the litter addition soil and litter
standing crop is about 30 % of the 3 kg C m-2 in litter added to the
litter addition plots over 6 years (litterfall is ca. 1 kg m-2 yr-1,
ca. 45 % is C, times 6 years). This increase in C in
surface soil and the litter standing crop could be interpreted as potential partial mitigation of the effects of increasing CO2 concentrations in
the atmosphere, though any increases in litterfall due to increased CO2
will be less than our experimental doubling. For example, a free-air CO2 experiment in a 13-year-old loblolly pine plantation in North
Carolina USA reported a 12 % increase in litterfall over 9 years (Lichter
et al., 2005, 2008).
The increases in soil C in our litter addition plots (ca. 1 % per year, of
total C to ca. 20 cm depth) are much smaller than those reported in the other
study of litter manipulation in tropical forest (lowland rain forest in
southwestern Costa Rica), where 2 years of litter removal reduced soil C
concentration in the top 10 cm of soil by 26 % and doubling litter
increased soil C by 31 % (Leff et al., 2012). In three temperate forest
studies, rates of change in soil C were small, but they were measured over
much longer periods. In the north central USA, soil C content decreased by 44 %
in litter removal plots and increased by 31 % in double litter plots
over a 50-year period (Table 2 in Lajtha et al., 2014a). In Pennsylvania, USA,
20 years of removing litter reduced soil C by 24 %, although the
corresponding litter doubling had no effect (Bowden et al., 2014). In a
deciduous forest in Massachusetts, USA, 20 years of litter removal also
reduced mineral soil C (by 19 %), but litter addition also resulted in
lower mineral soil C (by 6 %, Lajtha et al., 2014b). Differences between
forests in the effect of litter addition on soil organic matter could be
partly due to differences in priming of pre-existing soil organic C
resulting in no, or small, increases in soil C in double litter plots.
Priming might be greater in N-limited temperate forests remote from
atmospheric N pollution because one cause of priming is mining of soil
organic matter for N by microbes stimulated by additions of litter with low
N concentrations (relative to soil organic matter) (e.g. Nottingham et al.,
2015). It is therefore likely that many, but not all, forests will show
increased C in soils as a result of increased litter input.
Soil C might on average originate predominantly from roots rather than
shoots (Rasse et al., 2005), and that may be the case in our soils in Panama
because although changes in litter inputs have caused changes in soil C, they
are small – approximately 1 % of total soil C per year – compared to the
“normal” turnover of C of 25 % (0–10 cm soil) within 6 months (as
calculated from changes in C concentration from wet season to dry season;
Turner et al., 2015) and an annual turnover of about 7 % based on
incorporation of 13C into soils over decades (Schwendenmann and Pendall,
2008). Turnover rates of soil C are also high in other tropical forests; for
example, in eastern Brazil 40–50 % of the C in the top 40 cm of soil had
been fixed in about 32 years (Trumbore, 2000). In Panama the much greater
rates of turnover of soil C as compared to changes caused by litter removal
and addition suggest that the main source of soil organic matter (over
months to a few years) is roots, root exudates and mycorrhizal fungi.
Nevertheless, changes in above-ground litter input are still important because they have resulted in overall decreases and increases in soil C.
Litter manipulation – depth of effects
Effects of litter removal and addition differed among nutrients and were
strongest near the soil surface, with no significant differences below 30 cm.
The strength of the effects and the depth to which they were significant
are increasing with time. Four years after the start of litter manipulation, six nutrients showed significant effects in the upper 2 cm of soil
(NO3, NH4, PMeh, K, Ca, Mg), whereas only NO3 and Ca
showed significant effects from 0–10 cm (Sayer et al., 2010). After 6 years,
in the early dry season 2009 (current paper), effects were seen to greater
depths: NO3 was higher to 30 cm and PMeh to 20 cm in litter
addition plots. Over time significant differences have become apparent for
more nutrients and to greater depth in the soil; these differences were
caused by differences in litter input.
The concentrations of NH4 and NO3 are usually only measured in
surface soils in tropical rain forests, perhaps because N is generally
thought not to limit growth in such forests. However, fertilization with N
and K together increased growth of saplings and seedlings in the Gigante
Fertilization Project, which is adjacent to our litter manipulation
experiment in Panama (Wright et al., 2011). Relevant concentrations of
NH4 and NO3 are also difficult to measure since they change
rapidly over only a few hours (Turner and Romero, 2009); extractions
for the current paper were done within 2 h of collecting soils. In
our litter manipulation experiment, NH4 accounted for 76 % of the
sum of NH4 and NO3 (mean over all depths in controls plots)
and decreased less with depth than NO3 (at 50–100 cm NH4 was about
50 % of surface values, whereas NO3 was about 25 %). In the
nutrient addition experiment, Koehler et al. (2012) reported that NH4
also decreased less with depth (at 200 cm it was 41 % of surface
soils) than NO3 (to 17 % of surface soils) and that NH4
was the dominant form of total inorganic N (about 80 %) – the same
patterns as in our litter manipulation experiment. Nitrogen dynamics in
soils have also been measured in a litter manipulation experiment in Costa
Rica (Wieder et al., 2013), where nitrification rates were lower in both
litter removal and litter addition plots and extractable NH4 was
significantly lower in litter removal plots. This contrasts with our results
of greater NO3 in litter addition compared to litter removal and no
effect on NH4; the differences between the experiments might be due in
part to different soils and a wetter climate in Costa Rica (ca. 5 m rain per
year, cf. 2.6 m in Panama). Thus, soil N dynamics differ somewhat between the
only two tropical litter manipulation experiments, but in both NH4
was the dominant form of inorganic N and in both total inorganic N
decreased in litter removal plots and increased in litter addition plots
(though differences were not always statistically significant).
The available forms of P are also not often reported for the deeper
horizons of tropical forest soils, despite the fact that P is usually
regarded as the most likely limiting nutrient in such forests (Tanner et al.,
1998 and Cleveland et al., 2011) and has been shown to limit fine-litter
production in the adjacent nutrient addition experiment (Wright et al.,
2011). Mehlich P and total P both decreased with depth in control soils in
our litter manipulation experiment (at 50–100 cm concentrations were 25 and
29 % of those at 0–5 cm); in litter removal soils the decrease was less
steep (37 and 36 %). Litter addition increased Mehlich P in
the surface soils (though total P was not significantly greater), indicating
increased P availability, which is consistent with the finding that litter
addition decreased the strength of phosphate sorption in these soils
(Schreeg at al., 2013). Thus, for P, potentially the most commonly limiting
nutrient in tropical rain forest soils, 6 years of continuous removal and
addition of litter in our experiment has reduced and increased available P
down to 20 cm in the soil.
The relative amounts of exchangeable cations and their change with depth in
the control plots of the Panamanian litter manipulation soils are similar to
patterns in other tropical forest soils. In our experiment, Ca
concentrations (in centimoles of charge) are about twice those of Mg in
surface soils (though below 30 cm Mg-to-Ca ratios exceed 1); K
concentrations are usually less than 5 % of the total exchangeable bases.
With increasing depth, Ca, Mg and K concentrations all decrease, with Ca
decreasing more than Mg or K. Other tropical forest soils are
similar: in 19 profiles throughout Amazonia the sum of base cations (Ca, Mg,
K) was usually dominated by exchangeable Ca (11 cases) or Ca was equal to Mg
(4 cases), and both Ca and Mg mostly decreased with depth, while K was in
low or in trace concentrations in all profiles (Quesada et al., 2011). In
Hawaii (Porder and Chadwick, 2009), much younger soils (11 000 BP on lava),
with much higher concentrations of Ca, Mg and K than in Panama and Amazonia,
showed similar patterns: Ca was the dominant cation, K was usually less than
5 % of the sum of exchangeable Ca, Mg and K, and all cations decreased
with depth at the wetter sites (but not at the drier sites). Thus, in most
wet tropical forest soils, Ca is the most abundant cation and most cations
decrease with depth. Litter addition in Panama increased Ca and Mg
concentrations in the surface soils and thus steepened the depth gradient,
whereas litter removal decreased Ca and Mg and therefore decreased the
gradient; K was at much lower concentrations (as in Amazonia and Hawaii) and
was not affected by litter addition and litter removal even in 0–5 cm soils.
Design of litter manipulation experiments
The design of litter manipulation experiments needs to be carefully
considered when evaluating their results. The strength of the effect of
litter manipulation on soil C in Panama was much less than that in Costa
Rica, but the Panama and Costa Rica experiments are very different in
spatial scale. Plots in Panama are large (45×45 m); those in Costa Rica are
small (3×3 m). The small plots are “hot” and “cold” spots relative to large
individual tree crown areas (and likely tree root areas); crowns of the
largest trees in lowland rain forests are commonly 25 m in diameter, so a 3×3 m plot is 2 % of that area. These differences in experimental design
and their effects on the pattern of the results should be considered when
trying to understand ecosystem level processes; small hot and cold spots may
not represent what would happen in plots on the scale of the large trees, as
pointed out by Leff et al. (2012).
Conclusions
The increase in C in the mineral soil and the litter standing crop following
litter addition was statistically significant in the top 20 cm of the soil,
suggesting that any increased litterfall as a result of increased
atmospheric CO2 and/or temperature could result in a substantial
increase in soil C and therefore partially mitigate the increase in
atmospheric CO2. However, the current experiment added much more litter
than might be produced by an increase in CO2 of, say, 200 ppm and
added more nutrients than might occur even in temperate polluted sites. Thus,
new experiments are required to investigate the effects of more realistic
increases in litterfall using litter with low nutrient concentrations.
Data availability
The data used in this paper can be found in the tables in the Supplement.
The Supplement related to this article is available online at doi:10.5194/bg-13-6183-2016-supplement.
Acknowledgements
We thank J. Bee, L. Hayes, S. Queenborough, R. Upson and M. Vorontsova for
surveying the plots, J. Bee for setting up the experiment in 2000 and 2001;
E. Sayer for running the experiment from 2001–2009; A. Vincent for helping to
maintain the experiment from 2003–2005. T. Jucker did the statistics to
compare the effect of treatment on soil C relative to mineral matter.
Funding for the project was originally from the Mellon Foundation
(1999–2002); ongoing costs were paid for by the Gates Cambridge Trust (E. Sayer);
The University of Cambridge Domestic Research Studentship Scheme and
the Wolfson College Alice Evans Fund (A. Vincent); and The Drummond Fund of
Gonville and Caius College and Cambridge University (E. Tanner). The whole
of the experiment depended on the continuous raking of litter, which was
done by Jesus Valdez and Francisco Valdez. We thank D. Agudo and T. Romero
for doing the laboratory work and J. Rodriguez for collecting the samples in
the forest. S. J. Wright has been a frequent source of help for many aspects
of the experiment.
Edited by: S. Zaehle
Reviewed by: W. Wieder and one anonymous referee
References
Bowden, R. D., Deem, L., Plante, A. F., Peltre, C., Nadelhoffer, K., and
Lajtha, K.: Litter input controls soil carbon in temperate deciduous forest,
Soil Sci. Soc. Am. J., 78, s66–s75, 2014.
Burnham, K. P. and Anderson, D. R.: Information and likelihood theory: a
basis for model selection and inference, in: Model Selection and Multimodel
Inference, edited by: Burnham, K. P. and Anderson, D. R., Springer New York,
49–97, 2002.
Clemmensen, K. E., Bahr, A., Ovaskainen, O., Dahlberg, A., Ekblad, A.,
Wallander, H., Stenlid, J. Finlay, R. D., Wardle, D. A., and Lindahl, B. D.:
Roots and associated fungi drive long-term carbon sequestration in boreal
forest, Science, 339, 1615–1618, 2013.
Cleveland, C. C., Townsend, A. R., Taylor, P., Alvrez-Clare, S., Bustamente,
M. M. C., Chuyong, G., Dobrowski, S. Z., Grierson, P., Harms, K. E.,
Houlton, B. Z., Marklein, A., Parton, W., Porder, S., Reed, S. C., Sierre,
C. A., Silver, W. L., Tanner, E. V. J., and Wieder, W. R.: Relationships
among net primary productivity, nutrients and climate in tropical rain
forest: a pan-tropical analysis, Ecol. Lett., 14, 939–947, 2011.
Jobbagy, E. G. and Jackson, R. B.: The vertical distribution of soil organic
carbon and its relation to climate and vegetation, Ecol. Appl., 10, 423–436,
2000.Koehler, B., Corre, M. D., Steger, K., Well, R., Zehe, E., Sueta, J. P., and
Veldkamp, E.: An in-depth look into a tropical lowland forest soil:
nitrogen-addition effects on the contents of N2O, CO2 and CH4
and N2O isotopic signatures down to 2-m depth, Biogeochemistry, 111,
695–713, 2012.Körner, C.: Plant CO2 responses: an issue of definition, time and
resource supply, New Phytol., 172, 393–411, 2006.
Lajtha, K., Townsend, K. L., Kramer, M. G., Swanston, C., Bowden, R. B., and
Nadelhoffer, K.: Changes to particulate versus mineral-associated soil
carbon after 50 years of litter manipulation in forest and prairie
experimental ecosystems, Biogeochemistry, 119, 341–360, 2014a.
Lajtha, K., Bowden, R. D., and Nadelhoffer, K.: Litter and root manipulations
provide insights into soil organic matter dynamics and stability, Soil Sci.
Soc. Am. J., 78, s261–s269, 2014b.
Leff, J. W., Wieder, W. R., Taylor, P. G., Townsend, A. R., Nemergut, D. R.,
Grandy, A. S., and Cleveland, C. C.: Experimental litterfall manipulation
drives large and rapid changes in soil carbon cycling in a wet tropical
forest, Glob. Change Biol., 18, 2969–2979, 2012.Lichter, J., Barron, S. H., Bevacqua, C. E., Finzi, A. E., Irving, K. F.,
Stemmler E. A., and Schlesinger, W. H.: Soil carbon sequestration and
turnover in a pine forest after six years of atmospheric CO2
enrichment, Ecology, 86, 1835–1847, 2005.Lichter, J., Billings, S. A., Ziegler, S. E., Gaindh, D., Ryals, R., Finzi,
A. C., Jackson, R. B., Stemmler, E. A., and Schlesinger, W. H.: Soil carbon
sequestration in a pine forest after 9 years of atmospheric CO2 enrichment,
Glob. Change Biol., 14, 2910–2922, 2008.
Nemani, R. R., Keeling, C. D., Hashimoto, H., Jolly, W. M., Piper, S. C.,
Tucker, C. J., Myneni, R. B., and Running, S. W.: Climate-driven increases in
global terrestrial net primary production from 1982 to 1999, Science, 300,
1560–1563, 2003.
Nottingham, A. T., Turner, B. L., Stott, A. W., and Tanner, E. V. J.:
Nitrogen and phosphorus constrain labile and stable carbon turnover in
lowland tropical forest soils, Soil Biol. Biochem., 80, 26–33, 2015.
Pinheiro, J. and Bates, D.: Mixed-effects Models in S and S-PLUS, Springer,
New York, 548 pp., 2000.
Porder, S. and Chadwick, O. A.: Climate and soil-age constraints on nutrient
uplift and retention by plants, Ecology, 90, 623–636, 2009.
Powlson, D. S., Whitmore, A. P., and Goulding, K. W. T.: Soil carbon
sequestration to mitigate climate change: a critical re-examination to
identify the true and the false, Eur. J. Soil Sci., 62, 42–55, 2011.Quesada, C. A., Lloyd, J., Anderson, L. O., Fyllas, N. M., Schwarz, M., and
Czimczik, C. I.: Soils of Amazonia with particular reference to the RAINFOR sites,
Biogeosciences, 8, 1415–1440, 10.5194/bg-8-1415-2011, 2011.
Rasse, D. P., Rumpel, C., and Dignac, M.-F.: Is soil carbon mostly root
carbon? Mechanisms for a specific stabilisation, Plant Soil, 269, 341–356,
2005.
Sayer, E. J. and Tanner, E. V. J.: Experimental investigation of the
importance of litterfall in lowland semi-evergreen tropical forest nutrient
cycling, J. Ecol., 98, 1052–1062, 2010.
Sayer, E. J., Tanner, E. V. J., and Cheesman, A. W.: Increased litterfall
changes fine root distribution in a moist tropical forest, Plant Soil, 281,
5–13, 2006.Sayer, E. J., Powers, J. S., and Tanner, E. V. J.: Increased litterfall in
tropical forests boosts the transfer of soil CO2 to the Atmosphere,
PLoS ONE, 2, e1299, 10.1371/journal.pone.0001299, 2007.
Sayer, E. J., Heard, M. S., Grant, H. K., Marthews, T. R., and Tanner, E. V.
J.: Soil carbon release enhanced by increased tropical forest litterfall,
Nature Clim. Change, 1, 304–307, 2010.
Sayer, E. J., Wright, S. J., Tanner, E. V. J., Yavitt, J. B., Harms, K. E.,
Powers, J. S., Kaspari, M., Garcia, M. N., and Turner, B. L.: Variable
responses of lowland tropical forest nutrient status to fertilization and
litter manipulation, Ecosystems, 15, 387–400, 2012.Schmidt, M. W. I., Torn, M. S., Abiven, S., Dittmar, T., Guggenberger, G.,
Janssens, I. A., Kleber, M., Kogel-Knabner, I., Lehmann, J., Manning, D. A.
C., Nannipieri, P., Rasse, D. P., Weiner, S., and Trumbore, S. E.:
Persistence of soil organic matter as an ecosystem property, Nature, 478,
49–56, 2011.
Schreeg, L. A., Mack, M. C., and Turner, B. L.: Leaf litter inputs decrease
phosphate sorption in a strongly weathered tropical soil over two time
scales, Biogeochemistry, 113, 507–524, 2013.
Schwendenmann, L. and Pendall, E.: Response of soil organic matter dynamics
to conversion from tropical forest to grassland as determined by long-term
incubation, Biol. Fert. Soils, 44, 1053–1062, 2008.
Tanner, E. V. J., Vitousek, P. M., and Cuevas, E.: Experimental
investigation of nutrient limitation of forest growth on wet tropical
mountains, Ecology, 79, 10–22, 1998.
Trumbore, S. E.: Age of soil organic matter and soil respiration:
radiocarbon constraints on belowground C dynamics, Ecol. Appl., 10, 399–411,
2000.
Turner, B. L. and Romero, T. E.: Short-term changes in extractable inorganic
nutrients during storage of tropical rain forest soils, Soil Sci. Soc. Am.
J., 73, 1972–1979, 2009.
Turner, B. L., Yavitt, J. B., Harms, K. E., Garcia, M., and Wright, S. J.:
Seasonal changes in soil organic matter after a decade of nutrient addition
in a lowland tropical forest, Biogeochemistry, 123, 221–235, 2015.
Vincent, A. G., Turner, B. L., and Tanner, E. V. J.: Soil organic phosphorus
dynamics following perturbation of litter cycling in a tropical moist
forest, Eur. J. Soil Sci., 61, 48–57, 2010.
Wieder, W. R., Cleveland, C. C., Taylor, P. G., Nemergut, D. R., Hinkley,
E.-L., Philippot, L., Bru, D., Weintraub, S. R., Martin, M., and Townsend, A.
R.: Experimental removal and addition of leaf litter inputs reduces nitrate
production and loss in a lowland tropical forest, Biogeochemistry, 113,
629–642, 2013.
Wright, S. J., Yavitt, J. B., Wurzburger, N., Turner, B. L., Tanner, E. V.
J., Sayer, E. J. Santiago, L. S., Kaspari, M., Hedin, L. O., Harms, K. E.,
Garcia, M. N., and Corre, M. D.: Potassium, phosphorus, or nitrogen limit
root allocation, tree growth, or litter production in a lowland tropical
forest, Ecology, 98, 1616–1625, 2011.
Zuur, A. F., Leno, E. N., Walker, N., Saveliev, A. A., and Smith, G. M.: Mixed
Effects Models and Extensions in Ecology with R, Springer, New York, 574
pp., 2009.