Plant Soil (2015) 390:61–76

DOI 10.1007/s11104-014-2371-7

REGULAR ARTICLE

Partitioning NEE for absolute C input into various ecosystem

pools by combining results from eddy-covariance, atmospheric
flux partitioning and 13CO2 pulse labeling
M. Riederer & J. Pausch & Y. Kuzyakov & T. Foken

Received: 17 March 2014 / Accepted: 22 December 2014 / Published online: 9 January 2015
# Springer International Publishing Switzerland 2015

Abstract respiration through the use of a C flux partitioning

Background and aims The complexity of ecosystem model. Resultant CO2 assimilation served as absolute
processes, especially under continuously changing en- C input into the ecosystem and was further partitioned
vironmental conditions, requires high-resolution insight by applying the relative C distribution in subsidiary
into ecosystem carbon (C) fluxes. It is essential to gain pools, gained by 13C pulse labeling and tracing.
not only information about relative C balance and fluxes Results The results form eddy-covariance measure-
(common for partitioning studies), but also to obtain ments showed that the extensively managed grass-
these in absolute mass units. land was a significant net C sink of −91 g C m−2 a−1 in
Methods To evaluate absolute fluxes in belowground C 2010.
pools, the results of 21-day eddy-covariance and stable The mean daily assimilation of −7.1 g C m−2 d−1 was
isotope labeling experiment in summer 2010, were com- partitioned into fluxes of 2.5, 0.8, 0.5, 2.3 and 1.0 g C
bined. Eddy-covariance based net ecosystem exchange m−2 d−1 into shoots, roots, soil, shoot respiration and
was measured on extensively managed grassland and CO2 efflux from soil, respectively.
separated into underlying assimilation and ecosystem Conclusions We conclude that the combination of EC
measurements with isotope labeling techniques allowed
Responsible Editor: Zucong Cai. determining the absolute C input into several ecosystem
M. Riederer : T. Foken
pools. Hence, the study demonstrates an approach to
Department of Micrometeorology, University of Bayreuth, expand atmospheric flux measurements and to gain
95440 Bayreuth, Germany insight into the importance of individual ecosystem
pools for soil C cycling.
J. Pausch : Y. Kuzyakov
Department of Soil Science of Temperate Ecosystems,
University of Göttingen, Keywords Stable isotope pulse labeling . Net ecosystem
37077 Göttingen, Germany exchange . Carbon flux . Extensively managed grassland
T. Foken (*)
Member of Bayreuth Center of Ecology and Environmental
Research (BayCEER), University of Bayreuth,
Introduction
95440 Bayreuth, Germany
e-mail: thomas.foken@uni-bayreuth.de
Currently, two dominant approaches in ecosystem sci-
M. Riederer (*) ences are used to gain access to the carbon (C) cycle of
Regensburg Center of Energy and Resources, Regensburg
terrestrial ecosystems. Micrometeorological methods
University of Applied Sciences,
93049 Regensburg, Germany like the eddy-covariance (EC) technique provide a top
e-mail: riederer.michael@googlemail.com view from the atmosphere (Aubinet et al. 2000, 2012;
62 Plant Soil (2015) 390:61–76

Baldocchi 2003; Baldocchi et al. 2001; Moncrieff et al. pulse labeling allows tracing the absolute input of C into
1997), whereas leading isotopic methods used nowa- various ecosystem pools.
days in agricultural and soil science allow a more inte- Previous discussions in the literature about combin-
rior view of the ecosystem (Kuzyakov and Domanski ing stable isotope methods with eddy-covariance tech-
2000; Yakir and Sternberg 2000). Both are occasionally nique were aimed at, for example, acquiring natural
combined with chamber methods to facilitate and ex- atmospheric iso-fluxes (Yakir and Sternberg 2000;
pand investigation of CO2 fluxes (Goulden et al. 1996; Bowling et al. 2001; Wichura 2009) or, in the case of
Davidson et al. 2002; Dore et al. 2003; Subke and pulse labeling, evaluating and comparing the C cycle of
Tenhunen 2004; Rochette and Hutchinson 2005). various ecosystems (Gavrichkova 2009).
While EC methods have the advantage of barely Today, European grasslands are predominantly con-
disturbing ecosystem processes during the experi- sidered as C sinks but there are uncertainties: the IPCC
ment, isotopic methods are mostly destructive due did not agree with this opinion and ascribed a potential
to the necessity of taking plant and soil samples. role of either source or sink to grassland ecosystems
Another difference is that isotopic labeling ap- (IPCC 2007). Janssens (2003) found a certain sink ca-
proaches are largely point measurements, while pacity but with an uncertainty that was larger than the
EC integrates the signal throughout a large flux- sink itself. Also Ciais et al. (2010) could not sufficiently
footprint (Vesala et al. 2008). prove the detected sink capacity. Future climate change
EC is generally the favored technique on grasslands will even increase this uncertainty by affecting C cy-
for measuring the C balance in terms of the net ecosys- cling in temperate grasslands due to increasing temper-
tem carbon exchange (NEE), i.e., the proportion of C atures (Luo 2007), varying precipitation amounts and
released and taken up by the ecosystem (Wohlfahrt et al. patterns (Knapp 2002; Chou et al. 2008), heat waves
2012). To evaluate underlying processes and responses and droughts (Ciais et al. 2005; Joos et al. 2010), and
of the ecosystem to environmental change, NEE has to rising atmospheric CO2 concentrations (Luo et al.
be separated into its components: ecosystem respiration 2006).
(RECO) and gross primary production (GPP), by flux The present study was conducted at an extensively
partitioning models (FPM; Falge et al. 2002; Stoy et al. managed grassland site in Central Europe during the
2006; Desai et al. 2008; Lasslop et al. 2010; Reichstein main vegetation period 2010. Besides addressing the
et al. 2012). These are also used to gap-fill missing or question whether grassland ecosystems function as C
rejected data (Stoy et al. 2006; Ruppert et al. 2006; sink or source, the main aim of the current experiment
Desai et al. 2008; Papale 2012; Falge et al. 2001; was to determine the absolute C input into various
Moffat et al. 2007). By determining temporal variations ecosystem pools.
and the absolute amount of assimilated and released C For these reasons eddy-covariance measurements
for a certain period, the atmospheric approach reaches and a 13CO2 pulse labeling experiment were conducted
its limits. To our knowledge, this is the first study combining
Further partitioning of total CO2 efflux or C input results of EC measurements and of a CO2 pulse labeling
(GPP) into various ecosystem pools is not possible experiment to determine the absolute amounts of C
based on EC, but can be achieved using isotopic tech- transferred to various pools of a grassland ecosystem
niques (Buchmann 2000, 2002; Kuzyakov 2006). in Central Europe.
Thereby, natural continuous (C3 plants grow after C4
plants or vice versa), artificial continuous and artificial
pulse labeling approaches have to be differentiated. Methods
Advantages and disadvantages of the different
labeling approaches were discussed in several Study area
publications (Whipps 1990; Nguyen 2009; Werth
and Kuzyakov 2008). Pulse labeling provides the The experiment was conducted during summer 2010
relative distribution of recently assimilated C into from June 16th (DOY 167) to July 6th (DOY 187) on
various above and below ground pools. a submontane grassland site at the edge of the low
EC delivers the absolute C fluxes above the ecosys- mountain range “Fichtelgebirge”, 624 m a.s.l. (50°05′
tem. Combining the results of EC with that of 13CO2 25″N, 11°51′25″E) in northeast Bavaria, Germany. For
Plant Soil (2015) 390:61–76 63

the last 10 years the experimental site was used as Utah USA) and collected daily by a computer system as
extensively managed grassland without fertilization or a backup.
grazing, but with sporadic mowing once or twice a year.
The soil type is gleysol (IUSS Working Group WRB Data acquisition and analysis
2007), with a thickness of at least 70 cm. The average
annual temperature and precipitation are 5.8 °C and The raw data for the turbulent CO2 fluxes were post
1066 mm, respectively (Foken 2003). The “Großer processed and quality controlled based on micrometeo-
Waldstein” (877 m a.s.l.) lies north of the study site rological standards, applying the software package TK2
and the “Schneeberg” (1051 m a.s.l.) is to the south. developed at the University of Bayreuth (Mauder and
These two mountains generate a channeled wind field Foken 2004). This still evolving software (TK3 is now
on the site with East and above all West as dominating available; Mauder and Foken 2011) includes all neces-
wind directions (prevailing wind direction 263°). The sary data correction and data quality tools (Foken et al.
plant community can be described as Molinio– 2012), was proved in comparison with six other com-
Arrhenatheretea R. Tx. 1937 – economic grassland. monly used software packages (Mauder et al. 2008) and
With 48 species, the biodiversity is quite high. The most successfully applied in numerous major field campaigns
dominant species are Alchemilla monticola, Juncus (Mauder et al. 2006, 2007; Eigenmann et al. 2009). The
filiformis, Polygonum bistorta, Ranunculus acris and included quality flagging system evaluated stationarity
Trifolium repens. These species were considered when and turbulence during the averaging interval of 30 min
to decide the exact location of the labeling plots to gain and marked the resulting flux data with quality flags
best possible comparability with the whole ecosystem. from 1 (very good quality) to 9 (very low quality; Foken
Except for single larger individuals, the canopy height and Wichura 1996; Foken et al. 2004). The flux data
was about 0.4 m at the date of labeling. were then filtered according to these flags and only data
with quality 3 or better were used during the whole
experiment. In addition to that, footprint analysis was
Micrometeorological determination of absolute C input performed (Göckede et al. 2004, 2006). It could be
assured that the signal measured by EC originated ex-
Experiment setup clusively from the target land use type grassland
(Rannik et al. 2012). Due to the channeled wind regime,
An automated weather station provided 10 min averages two clubbed footprints evolved in western and eastern
of a range of climate data to evaluate short term effects, directions. Thus, disturbances of the turbulent fluxes
but also to provide the input parameters for the measured by EC could be avoided by installing the other
partitioning of the NEE into its source and sink compo- experimental devices directly adjacent to the EC mast
nents. The most important collected parameters were but perpendicular to the main wind direction.
up– and down welling short– and long wave radiation,
air and soil temperature, humidity and soil moisture and NEE flux partitioning
precipitation. High frequency (20 Hz, 2.5 m above
ground) data were collected to determine turbulent In order to finally gain absolute C input into the ecosys-
fluxes, such as NEE by eddy-covariance. Water vapor tem from the NEE data, two tasks were performed: Due
and CO2 concentration were measured by an open–path to rejection of outliers and low quality data, small gaps
gas analyzer (LI–7500, LI–COR Biosciences, Lincoln, occurred within the 30 min NEE time series that had to
Nebraska USA) and wind vector and sonic temperature be filled and the NEE had to be partitioned into its
(TS) by a 3D sonic anemometer (CSAT3, Campbell underlying fluxes, assimilation (GPP) and respiration
Scientific, Inc., Logan, Utah USA). CSAT3 and LI– (RECO). To parameterize temperature dependant RECO,
7500 were pointed in a northerly direction, normal to equal to nighttime NEE due to missing assimilation, the
the prevailing wind direction of 263°. Thus, disturbance Lloyd–Taylor function was applied (Lloyd and Taylor
of the flux by the instruments was minimized (Li et al. 1994; Falge et al. 2001; Ammann et al. 2007; Reichstein
2013). Tower shading could be avoided completely due et al. 2005). Light response regression on the basis of the
to the channeled wind regime. Data were stored on a Michaelis–Menten function (Michaelis and Menten
data logger (CR3000, Campbell Scientific, Inc., Logan, 1913) was used to parameterize daytime solar radiation
64 Plant Soil (2015) 390:61–76

dependant GPP (Falge et al. 2001; Ruppert et al. 2006). of the labeling. The IR–sensor of this device detects
For both, the flux-partitioning model used a time–win- only about 30 % of the 13CO2 (McDermitt et al. 1993),
dow scheme instead of the conventional temperature but the concentration increased at the beginning to a
binning approach that was suitable for sites with distinct value of about 1400 ppm and a concentration next to
seasonal variation (Ammann et al. 2007). zero after the 3 h was measured indicating the complete
uptake of the tracer. Shortly before the labeling the CO2
13
C pulse labeling for determination of relative concentration within the chamber dropped down to zero
proportion of C partitioning due to assimilation. It is expected that this very short
lack of CO 2 had no noticeable influence on the
Experiment setup experiment.

Five stainless steel soil frames (each 1 × 1 m) with a u- Data acquisition and analysis
shaped bar at the upper end were inserted up to 10 cm
depth, 3 weeks prior to labeling in order to reduce Translocation of the assimilated 13C was analyzed dur-
disturbances. For 13CO2 pulse labeling the upper part ing a 21-day period in shoots, roots, soil and soil CO2
of the chamber, consisting of aluminum frames (base of efflux on all 5 plots. Samples were taken immediately
the frame 1 × 1 m, height 0.5 m, cross section 2 × 2 cm) (0), 1, 2, 4, 9 and 21 days after the labeling. Shoots were
were placed into the u-shaped bar, which was filled with sampled from a circular area of 10 cm diameter. Soil
water (containing a small amount of H2SO4) to ensure samples were taken in the middle of this area from 0 to
sealing of upper and lower parts of the chamber. The 30 cm depth using a soil corer (inner diameter: 4.6 cm).
aluminum frames were covered with transparent LDPE– Afterwards, the holes in the soil were plugged with
foil (thickness: 0.2 mm; total light transmission: ~ 90 %) PVC–tubes to avoid changing conditions around the
shortly before the tracer addition To minimize the influ- holes.
ence of the chamber on the tracer uptake, five cooling In addition, samples from unlabeled plots were taken
aggregates (EZetil Iceakku, 220 g) arranged in parallel in the same way close to each of the labeled plots to
were installed in each chamber. A fan positioned behind determine the δ13C natural abundance for calculations.
the aggregates guaranteed turbulent mixing of the cham- All samples were frozen (−20 °C) until further anal-
ber air and forced the air to pass the cooling aggregates. ysis. Roots were carefully separated from the soil sam-
High temperatures were thereby avoided and the humid- ples with tweezers. All shoot, root and soil samples were
ity was reduced by condensation of the water vapor at dried, weighed and homogenized by ball milling.
the cooling aggregates’ surfaces. Hence, the condensa- Total C and the δ13C (‰) signatures of the samples
tion at the chamber walls was reduced and better light were determined using an element analyzer – isotope
conditions for the plants were assured. For more detailed ratio mass spectrometer (EA–IRMS, Delta Plus;
information about the chamber construction see Drösler Thermo Fisher Scientific, Bremen, Germany, interfaced
(2005). A flask, containing the 13C tracer as Na213CO3 to an elemental analyzer (NC 2500; CE Instruments,
(5 g 99 % 13C-eniched Na2CO3), was placed behind the Milano, Italy) and calibrated with reference to the inter-
fan to assure homogenous distribution of the labeled national standard VPDB (Vienna Peedee Belemnite).
CO2. An excess of 5 M H2SO4 was added to the tracer The total CO2 efflux from soil was determined on all
solution from outside the chamber through the transpar- labeled and on unlabeled (natural abundance) plots with
ent LDPE–foil with a syringe. The puncture holes in the the static alkali (NaOH) absorption method
foil were afterwards sealed with tape. The labeling was (Lundegardh 1921; Kirita 1971; Singh and Gupta
done almost simultaneously for all five chambers with 1977) After cutting the vegetation to avoid any fraction-
only short time shifts of some minutes. Plants were ation of the isotopic signal by photosynthesis and shoot
labeled for 3 h to assure complete uptake of the 13CO2. respiration, a stainless steel soil collar (i.D. 11 cm;
To avoid noon depression of photosynthesis, labeling height 10 cm) was placed 5 cm into the soil. It has to
was conducted from 2:30 to 5:30 pm. In one of the be considered that cutting aboveground vegetation may
chambers the CO2 concentration was monitored with cause decrease in root respiration and increased turnover
an infrared gas analyzer (LI–820, LI–COR Biosciences, of dead root biomass. A jar with 1 M NaOH were placed
Lincoln, Nebraska USA) at the beginning and at the end into each collar and the collar was closed with a dark lid.
Plant Soil (2015) 390:61–76 65

Soil CO2 efflux was calculated using the following These standard errors may not completely characterize
equation: the uncertainty (Zobitz et al. 2006).
By multiplication with the total C amount (n(C)P, g C
xð C Þ P m−2) of the pool, the 13C amount (n(13C)P, g 13C m−2) of
F CO2 ;soil ¼ ; ð1Þ
A⋅△t the pool was calculated:
with the total amount of C captured x(C)P, the closed


n 13 C P ¼ xE 13 C ⋅nðC ÞP : ð3Þ
time of the collar △t and the area enclosed A.
Shortly after the labeling a NaOH trap was placed in Since all calculations were carried out with area units
each chamber. NaOH was exchanged at each sampling (m−2) it has to be mentioned that in the case of soil and
date and additionally on the 12th day after labeling. The roots all results referred to the sampled soil layer from 0
amount of NaOH was adjusted to the period by increas- to 30 cm. To gain a reference value for the recovered
ing from 40 ml at the beginning up to 80 ml at the end, to amount of 13C during the sampling period, the total
be sure that the neutralization did not exceed one–third amounts of 13C found immediately after the labeling
of the capacity of the NaOH (Gupta and Singh 1977).

(n 13 C Pt , g 13C m−2) were summed up over all inves-
The amount of collected C was determined by a C/N 0

analyzer (Multi N/C 2100, AnalytikJena, Germany). To tigated pools. Then the 13C amounts of every single pool


obtain δ13C (‰) values, SrCO3 was precipitated with at every point of time (n 13 C Pt , g 13C m−2) could be
SrCl 2 , neutralized and dried for the EA–IRMS related to this total value and the recovery (R, %) of the
measurements. tracer could be calculated using the equation:
For the calculation of the relative proportion of 13C

input into various pools (shoots, roots, soil and CO2
13  n 13 C Pt
R C Pt ¼ 4 ð4Þ
efflux were investigated) after 13CO2 pulse labeling X

several calculation steps were necessary. The enrich- n 13 C Pt ðiÞ
0
ment of 13C in a C pool (xE(13C), atom%) was derived i¼1

by subtracting the naturally abundant amount of 13C

where t represents any date of sampling and t0 the
(x(13C)std, atom%) from the amount of 13C in the labeled
point of time immediately after the labeling, when sam-
pool P (x(13C)P, atom%):
ples were taken for the first time. These calculations



 were conducted similarly for all pool types i with one
xE 13 C ¼ x 13 C P −x 13 C std ð2Þ
exception. In contrast to the other pools, where sampling
where E marks the excess on 13C of the atom fraction was destructive and therefore spatially distributed, the
x (= amount of an isotope of a chemical element, divided
13
C amount (n(13C)P, g 13C m−2) within the CO2 efflux
by the total amount of atoms of this element; Coplen (F CO2 ;soil ) was always sampled at the same position.
2011). This was compensated by finally summing all values
The natural abundance δ13C value of soil CO2 efflux, of the single sampling dates. Hence, the complete
measured beside the labeling plots, was determined by amount of 13C was considered in that pool as well.
correcting the measured δ13C values for the admixture The losses of 13C by shoot respiration were not
of atmospheric CO2, based on the Miller/Tans model measured, but could be estimated by the following
(Miller and Tans 2003; Pausch and Kuzyakov 2012). equation:
Therefore, measured δ13C values multiplied by the re-
spective CO2 concentrations were plotted against the
 X
4


R 13 C Pt ¼ 100%− R 13 C Pt ðiÞ ð5Þ
CO2 concentrations. The slope of the regression line is Shoot
i¼1
equivalent to the δ13C value of soil CO2 efflux purified
from atmospheric CO2 (Miller and Tans 2003). The Translocation to deeper soil layers was excluded by
Miller/Tans model was applied in combination with a taking and analyzing samples between 30 and 50 cm
geometric mean regression (GMR), as suggested for soil depth. There no noteworthy amount of tracer could be
CO2 by Kayler et al. (2010). The standard errors for the detected. Consequently, it is assumed that shoot respi-
slope of the GMR were taken from the respective ordi- ration is the only relevant missing sink of 13C within the
nary least square regression (Sokal and Rohlf 2008). considered system, the 13C recovered (%) of all four
66 Plant Soil (2015) 390:61–76

measured pools i could be summed, and then subtracted events were measured in August. The fluxes at the
from 100 % (Hafner et al. 2012). However, a slight labeling day and during the chase period (CP) are shown
overestimation of the soil respiration might occur due in Table 1. The mean daily sum of GPP at the labeling
to missing of small amounts of carbon leaching during day was −6.0 g C m−2 d−1) whereas a mean GPP of −7.1
the rainfall events. To assure that the 13C recovered no ±0.4 g C m−2 d−1 was determined for the whole chase
longer changed in time, i.e., that the allocation did reach period. Figure 1 provides a general view of the intra-
a steady state, the 13C recovery in all pools was checked annual variability of the ecosystem fluxes, indicating
by applying a repeated measures ANOVA with a post that a number of pulse labeling experiments would be
hoc Bonferroni test. Means and standard errors of the necessary to achieve detailed seasonal partitioning of
means (SEM) are presented in the figures and tables. absolute carbon fluxes. The labeling experiment was
To finally gain absolute C input into the particular conducted within a long time period with a quite uni-
ecosystem pools, labeling and eddy-covariance results form assimilation flux that did not end until the first
were combined, i.e., the relative proportion of the 13C cutting (Fig. 1).
recovered at the end of the C allocation was combined
13
with the total C input into the system C dynamics and allocation


nðC ÞP ¼ GPP⋅R 13 C Pt ð6Þ The sampling immediately started after the 13CO2 tracer
end

was completely assimilated. The 13C recovery in the

where n(C)P (g C m−2 s−1) is the absolute C input of
shoot biomass strongly decreased from 72.3 % imme-
the respective pool.
diately after labeling to 46.6 % 1 day after labeling,
Note that chamber conditions and CO2 concentra-
mainly due to shoot respiration (Fig. 2). About 14.7 %
tions during labeling may have influenced the photo-
of 13C was translocated from shoots into roots directly
synthetic rate. Hence, total CO2 uptake during labeling
after labeling. The 13C recovery of roots did not change
presumably differed from that measured by EC.
significantly over 21 days. In contrast to the roots, the
However, we assume that the impact of the chamber
maximum 13C amount of the soil pool was detected
conditions on relative 13C partitioning within the plant-
1 day after the labeling. Thereafter, the 13C recovery in
soil system were negligible because after the short la-
the soil slightly decreased and reached 6.4 % 21 days
beling period (3 h) the plants were again exposed to
after labeling (Fig. 2b). Similar to shoot respiration, 13C
natural conditions.
in soil CO2 efflux was highest during the first day and
then declined over time.
The allocation of 13C tracer was mostly completed
Results after 9 days and the 13C recovery in all pools did not
change significantly between the last two samplings.
Absolute atmospheric CO2 fluxes Therefore, the precondition for the partitioning of the
absolute C input, the steady state, was fulfilled.
Plants started to growth already at the end of February, Figure 2b illustrates the final percentage at the end of
and the growth period ended in mid-October (Fig. 1). At the translocation process. The C flux back into the
the beginning, the biomass growth was decelerated by a atmosphere, consisting of shoot respiration and soil
frost period in March, and during summer the assimilat- CO2 efflux, dominates the proportion by accounting
ing biomass was harvested by two cutting events, (DOY for almost half (46.7 %) of the assimilated 13C. About
188 and 265, marked with ‘c’ in Fig. 1) which became one third (34.9 %) remains in the shoots, while roots and
apparent in the GPP and NEE time series. soil obtain, with 12 and 6.4 %, respectively, compara-
The isotopic pulse labeling was conducted on June tively small proportions of 13C. Overall about 32 % of
16th (DOY 167, left edge of grey dashed box in Fig. 1) assimilated 13C were allocated to below–ground pools
and the subsequent chase period (CP, grey dashed box in
Fig. 1, Table 1), where samples were taken to investigate Partitioned absolute C allocation
13
C dynamics and translocation, ended on July 06th
(DOY 187) with the last sampling, shortly before the The absolute amount of total assimilated C (GPP) by the
first meadow cutting. The most extreme precipitation ecosystem during the chase period (CP in Table 1) was
Plant Soil (2015) 390:61–76 67

Fig. 1 Cumulative annual fluxes of NEE, GPP and RECO (flat pulse labeling and comprises the chase period (CP), beginning
lines), daily sums of precipitation (black bars), daily means of with the pulse labeling and ending shortly before the first mowing
global radiation (grey filled circles) and daily mean temperatures event (c). Time on x–axis in day of year (DOY)
(black filled circles). The box with dashed outline begins with the

partitioned for absolute C allocation into individual independently of the labeling by separating the NEE
pools based on the 13C recovery of the respective pool. by the FPM, the equal results verify our approach
The 13C recovery rates could only be applied to the GPP
from the chase period (Fig. 1), since the transferability
beyond this period was not validated by accounting for, Discussion
for example plant physiological factors. On average, 2.5
±0.2 g C m−2 d−1 were incorporated into the shoot and Discussion overview
0.8±0.3 g C m−2 d−1 into the root biomass. 0.5±0.1 g C
m−2 d−1 remained in the soil, whereas 2.3±0.3 g C m−2 By combining the results of atmospheric CO2 flux mea-
d−1 were released to the atmosphere as shoot respiration surements and 13CO2 pulse labeling, a new approach for
and 1.0±0.1 g C m−2 d−1 as soil CO2 efflux. The sum of partitioning ecosystem C fluxes was introduced. In the
the soil CO2 efflux and shoot respiration (3.3±0.4 g C following, the results will be discussed in detail, i.e.,
m−2 d−1) is in accordance with the RECO of 3.5±0.2 g C absolute atmospheric CO2 fluxes will be compared to
m −2 d −1 (Fig. 3). Since R ECO was determined further flux measurements under similar environmental

Table 1 Annual (g C m−2 a−1)

and daily (g C m−2 d−1) C fluxes 2010 (365 days) Chase period (21 days) Labeling day
(±SEM) for the chase period (CP) annual sum mean of daily sums daily sum
and the day of labeling (June
16th) in 2010 NEE −249 −3.5±0.4 −1.8
GPP −1097 −7.1±0.3 −6.0
RECO 849 3.5±0.2 4.1
Harvest 158
Balance −91
68 Plant Soil (2015) 390:61–76

Fig. 2 Cumulative 13C label-incorporation into the various x–axis of (a) intersects at y=1 % for a better illustration. Only one
C-pools; a 13C dynamics during the chase period; b relative value remains below 1 %, the 13CO2 efflux immediately after the
proportion of 13C recovered, i.e., final distribution by percentage labeling accounting for 0.1 %. Error bars represent standard errors
at the last day of sampling (day 21) in the ecosystem C pools; the of the mean (±SEM)

conditions, and relative assimilate distribution will be carbon sink in relation to other comparable extensively
compared to those of other 13C labeling experiments. managed grasslands. In Table 2, recent studies dealing
Since there are no studies referring to comparable efforts with atmospheric CO2 fluxes on such grasslands at
in determining partitioned absolute C allocation in the elevations from 375 to 1770 m a.s.l., with mean annual
plant–soil–atmosphere system, on-hand results are com- temperatures from 5.5 to 9.5 °C and annual precipitation
pared to studies in which these quantities were sums from 655 to 1816 mm, were reviewed. Although
estimated. the sites were chosen in a range which was as narrow as
possible in terms of these parameters, there are notable
Atmospheric C fluxes differences in the NEE. However, the NEE of the pres-
ent study lies in the middle of those of the reviewed
NEE was directly measured by eddy-covariance in 2010 studies (Table 2). In general, the role of grasslands in the
(−249 g C m−2 a−1). After subtraction of the harvest global carbon cycle is still uncertain, as recently de-
output (158 g C m−2 a−1), −91 g C m−2 a−1 still scribed by Gilmanov et al. (2010). There a mean NEE
remained, identifying the site as being a relatively big of 70 g C m−2 a−1, but also maximum C sources up to

Fig. 3 Average daily absolute

input (GPP), output (RECO) and
partitioned absolute C distribution
after assimilation (g C m−2 d−1,
±SEM) during the chase period of
the labeling experiment. Please
note that for illustration all values,
even GPP, have a positive sign
Plant Soil (2015) 390:61–76 69

temperature (T, °C); all sites were managed extensively, some with temporary light grazing instead of cutting. Harvest means harvested C yield from field. All fluxes are presented in g C m−2
Table 2 Atmospheric C fluxes, determined on European grassland sites with comparable parameters: elevation (m a.s.l.), annual sum of precipitation (RR, mm), sorted by annual mean

(Wohlfahrt et al. 2008)

(Gilmanov et al. 2007)

(Gilmanov et al. 2007)

(Gilmanov et al. 2007)

(Gilmanov et al. 2007)
481 g C m−2 a−1 and maximum C sinks up to −366 g C

(Ammann et al. 2007)

(Hussain et al. 2011)
(Allard et al. 2007)
m−2 a−1 were reviewed for extensively managed grass-
lands all over the world.
References

This study
Separating NEE into underlying assimilation (GPP)
and respiration (RECO) fluxes using the short time win-
dow scheme was certain to capture the dynamics of this
fast changing ecosystem (Ammann et al. 2007;
Wohlfahrt et al. 2012), because it sufficiently accounted
RECO

1586
1089
1592
851

559
849
1160
640
1440
for seasonal parameter variability (Lasslop et al. 2010).
Total annual sums in 2010 (RECO: 849 and GPP:
−1097 g C m−2 a−1) are within the range of those
−1303
−1856
−1128

−1568
−606
−1097
−1235
−1083
−1514
GPP

reviewed in Table 2. It is therefore important to note

that the results of this study match best to sites with
−214
−254
−278
−75
18
−47
−249
−75
−443 certain restrictions relating to ecosystem productivity,
NEE

e.g., low annual temperature means, combined with

high elevations (site No. 8 and 9, Table 2). There is also
Harvest

good agreement with another low elevation site (No. 3,

~147
311

~317

158

Table 2), but in that case GPP is probably limited by a

lack of precipitation. With that exception, the grassland
in the present study is more comparable to higher ele-
extensive, 2 and 3 cuts
extensive, cut / grazed

vation sites due to its cold climate. This is also con-

extensive, grazed
extensive, grazed

extensive, grazed
extensive, 3 cuts

extensive, 3 cuts

extensive, 2 cuts

firmed by RECO, which is on average smaller than that of

extensive, 1 cut
Management

the warmer sites with low elevation, but higher than that
of high elevation sites. Ammann et al. (2007), who even
applied a similar flux partitioning model on an exten-
sively managed grassland in Switzerland, found C
fluxes more than one third higher, despite similar
1816
1189
655
1200
1109

1064
852
1234

1066

elevation and precipitation, but with a 3.2 K higher

RR

mean annual temperature.

In a global context, European extensively managed
5.5
5.5
7.9
9.0

7.0
6.5
6.1
9.5

5.8
T

grasslands are outstandingly productive. While

Gilmanov et al. (2010) reviewed a worldwide GPP of
Elevation

−154±463 g C m−2 a−1, Schulze et al. (2010) found an

1699
1550
375
450

1040
970
1770
900

624

average GPP for Europe that is almost ten times higher:

−1343±269 g C m−2 a−1. This in turn is within the range
of the GPP of grassland sites reviewed in Table 2, which
2003
2004
2003/2004
2002–2004

2002–2004
2001–2006
2004

2003/2004
2010

are obviously representative for European extensively

managed grasslands.
Year

Relative 13C allocation

Oensingen, Switzerland
Grillenburg, Germany

Monte Bondone, Italy

Malga Arpaco, Spain
Voitsumra, Germany

Isotopic pulse labeling, the most frequently applied

Laqueille, France
Neustift, Austria

tracer method, was used to quantify the input of 13C to

Amplero, Italy

Alinyà, Spain

diverse ecosystem C pools. At first view, pulse labeling

reveals the relative distribution of assimilated C at the
moment of labeling and not the distribution of total
Site

unlabeled C in different plant parts (Kuzyakov and

Domanski 2000). However, by observing 13C allocation
No.
a−1

over a certain period, up to a steady state within the

1
2
3
4
5
6
7
8
9
70 Plant Soil (2015) 390:61–76

whole plant-soil-atmosphere system, a representative above-mentioned night–time fluxes. However, shoot

proportion for total C is finally found (Saggar et al. respiration dynamics seem feasible and the final propor-
1997; Saggar and Hedley 2001; Wu et al. 2010). The tion of 30 % lies within the range found in the literature
end of the chase period was defined as occurring when (Table 3).
the amount of 13C recovered in the last two samples of The proportion of below–ground C input (32 %) into
each pool no longer changed significantly (Saggar et al. roots (12 %), soil (6.4 %) and CO2 efflux (13.6 %) is
1997). That happened after 21 days (cf. Keith et al. also in line with Kuzyakov and Domanski (2000);
1986; Swinnen et al. 1994). Depending on the pools (Table 3). The relatively low allocation to below–
considered and the sampling frequency, the end of ground pools, especially to the root system, may be
the 13C (14C) allocation period was defined as explained on the general steadiness of long–established
being between 4 and 28 days (Domanski et al. grassland root systems (Saggar et al. 1997). This was
2001; Wu et al. 2010; Hafner et al. 2012; Ostle confirmed by the biomass data and can be also an
et al. 2000; Saggar et al. 1997). While numerous explanation for the non–significant changes of the 13C
pulse labeling studies address the back diffusion of recovery during the chase period (Fig. 2). However,
tracer to soil pore space occurring during the la- results of other studies are quite heterogeneous, but
beling (Subke et al. 2009; Bahn et al. 2009; these found mostly higher amounts (Table 3) and, be-
Staddon 2003; Leake et al. 2006), dealing with yond that, diverse patterns in C allocation to roots. The
isotopic steady state (after 21 days) allows this maximum amount of tracer reached the roots one
difficulty to be disregarded, as it is only relevant (Johnson et al. 2002) or 2 days (Ostle et al. 2000;
for the first 2 days after the labeling (Gamnitzer Staddon 2003), or even weeks later, but then mostly
et al. 2011; Biasi et al. 2012). without significant differences (Rangel–Castro et al.
In accordance with Wu et al. (2010), the percentage 2004; Leake et al. 2006; Hafner et al. 2012). A slight
of 13C recovered – rather than the isotope fraction – was peak at the fourth day as in the current study is a realistic
used to determine the overall proportion. Calculation of result if it is considered that Kuzyakov and Domanski
the 13C recovered was achieved by referring to summa- (2000) suggested a period of hours to days after the
tion of 13C in all measured pools (Kaštovská and labeling. CO2 efflux from soil exhibits the same pattern
Šantrůčková 2007; Hafner et al. 2012) in order to not presented by Staddon (2003) and Hafner et al.
underestimate the initial fixation by considering only (2012): An initial peak, an exponentially decreas-
13
C found in shoots directly after labeling. About one ing recovery of 13C over time and a decreasing
third of the C remains in the shoot biomass as reviewed slope in the cumulative 13CO2 efflux (Fig. 2). This
by Kuzyakov and Domanski (2000) for numerous pas- pattern of the soil CO2 efflux indicates fast trans-
ture plant studies (Table 3). In contrast, agricultural location of recently assimilated C through the sys-
plants like wheat or maize incorporate a lager proportion tem, probably released by root-derived respiration
(50–60 %) into the shoot (Jones et al. 2009; Table 3). (Kuzyakov et al. 2001). However, CO2 efflux from
During the chase period the amount of tracer decreased soil was determined with the static alkali (NaOH)
by 48 % within the shoots, which is quite close to the absorption method. This method is useful but has
32–51 % of Johnson et al. (2002) and 55 % of Butler also disadvantages as e.g., scrubbing CO2 from the
et al. (2004) and Wu et al. (2010). Higher rates are also chamber headspace of missing atmospheric turbu-
possible for grasslands, e.g., 77 % (Ostle et al. 2000) and lence. Although those two are opposite effects, on
70 % (Leake et al. 2006), even during the first day after the whole the flux rate might be overestimated.
the labeling. In this study the maximum decline also Compared to the other pools, 13C enrichment of
took place between first and second sampling, including bulk soil after pulse labeling was relatively low.
the first night after the labeling, caused mainly by night– The amount of 13C recovered in the soil (6.4 %) is
time shoot respiration and allocation to roots (Butler comparable to other studies, especially those sum-
et al. 2004; Leake et al. 2006). Shoot respiration dynam- marized in the reviews (Table 3). A slightly higher
ics agree with this finding, by increasing after the first amount of 13C was found after 1 day, but just as
sampling, which took place in the late afternoon at the the weak peaks of Staddon (2003) after 12 and
labeling day. The much higher percentage of 13C was Rangel–Castro et al. (2004) after 7 days, it was
recovered at the second sampling resulted from the not significant (Fig. 2).
Table 3 Comparable partitioning studies related to species, methods and investigated compartments. All values are presented in % of recovery or % of (net) assimilated tracer
Plant Soil (2015) 390:61–76

No. Plant / Conditions Method Days after Shoot Below ground Root Soil RECO Shoot resp. CO2 References
labeling efflux

14
1 Lolium perenne / controlled cond. C pulse, % of recovery 7 49.8 40.7 21.8 1.8 16.6 9.5 17.1 (Rattray et al. 1995)
14
2 Lolium perenne / controlled cond. C pulse, % of recovery 2 40.0 60.0 14.6 30.0 15.4 (Bazot et al. 2006)
13
3 Festuca, controlled cond. C pulse, % of recovery 2 43.9 54.9 39.7 4.1 11.1 (Allard et al. 2006)
14
4 White clover / controlled cond. C pulse, % of recovery 2 56.9 43.0 9.0 7.2 26.8 (Todorovic et al. 1999)
13
5 Grassland / field conditions C pulse, % of recovery 32 28.9 58.7 34.2 7.3 29.6 12.4 17.2 (Wu et al. 2010)
13
6 Grassland / field conditions C pulse, % of recovery 27 38.0 20.0 0.5 10.4 51.0 42.0 9.0 (Hafner et al. 2012)
14
7 Pasture / field conditions C pulse, % of recovery 35 26.4 34.7 2.1 36.8 (Saggar et al. 1997)
14
8 Pasture / field conditions C pulse, % of recovery 35 31.0 27.0 5.2 37.0 (Saggar and Hedley 2001)
14
9 Lolium perenne / controlled cond. C continuousa, % of 47.8 52.1 39.7 2.6 9.8 (van Ginkel et al. 1997)
recovery
14
10 Brome grass / controlled cond. C % rep. pulseb, % of 27.0 5.0 14.0 54.0 (Davenport and Thomas 1988)
assimilatedc
11 21 agric.plants Review % of net assimilatedd 60 36 19 5 12 (Jones et al. 2009)
12 Pasture plants, Review % of assimilatedc 30 40 20 5 45 30 15 (Kuzyakov and Domanski 2000)
13
13 Grassland / field conditions C pulse, % of recovery 21 34.9 32.0 12.0 6.4 46.7 33.1 13.6 this study
a
Continuously labeled with 14 C during the whole time of growing
b
Weekly repeated pulse labeling, as suggested to cover whole growth period
c
Recovery related to assimilated amount of tracer (shoot respiration is not considered, may underestimate final result by up to 30 %)
d
Recovery related to net assimilated amount of tracer (shoot respiration is considered)
71
72 Plant Soil (2015) 390:61–76

Partitioned absolute C fluxes Concluding remarks

Up to now, partitioned absolute amounts of allocated C Application of EC showed that the extensively managed
were only roughly estimated, although in most studies grassland was a significant net carbon sink of −91 g C
addressing to C balance and turnover, total masses are m−2 a−1 in 2010. The NEE flux-partitioning model
important. Kuzyakov and Domanski (2000) calculated revealed a mean underlying assimilated amount of car-
mean absolute values for below-ground translocated C bon of −7.1±0.3 g C m−2 d−1 during the 21 days of the
by grasses and cereals from the literature: 179 g C m−2 13
C pulse labeling experiment. Pulse labeling and trac-
for all studies and 220 g C m−2 for studies longer than ing provided relative partitioning of 13C input into dis-
100 days (i.e., 2.2 g C m−2 on average). Absolute C tinct ecosystem C pools. First-time combining the re-
inputs found for an alpine Kobresia humilis pasture (Wu sults of these methods to an integrative approach
et al. 2010) were about one third smaller than in the allowed partitioning of absolute C input by assim-
present study in all compartments except the roots, when ilation into absolute C fluxes into shoots, roots
taking the length of the growth period into account. This and soil and the contributions to the respiration
results from the generally lower turnover rates in high fluxes CO2-efflux and shoot respiration. Two dif-
altitude grasslands (Budge et al. 2011). In addition to ferent areas benefit from this innovation: labeling
that, the percentage of root biomass is considerably approaches are upgraded by finally dealing with
higher in these regions (Ammann et al. 2009; Leifeld absolute instead of relative C allocation and fur-
et al. 2009; Unteregelsbacher et al. 2011). One further ther separation of the NEE beyond assimilation
comparison allows the rough estimation of total C in- and respiration fluxes is provided. Moreover, indi-
and outputs for pasture plants. Kuzyakov and Domanski vidual reactions of sensitive subsidiary ecosystem
(2000) measured fluxes that are on average 1.5 times pools and processes can be detected and evaluated
lower than that of this study, but the input into the root on the basis of mass units.
system matches very well. However, under the currently changing environmen-
Obviously, there is a lack of studies presenting abso- tal conditions, both approaches benefit from the reduc-
lute values of C input to distinct ecosystem compart- tion of uncertainties by the detection and evaluation of
ments. Coupling of atmospheric C flux measurements individual reactions of sensitive subsidiary ecosystem
with 13C pulse labeling provides partitioning of absolute pools and processes on the basis of mass units. The
C fluxes. In general, the combination of methods works results of this study are in line with the available litera-
well and allows a more detailed insight into the C cycle ture and should encourage combining methods of atmo-
of grasslands. One limitation is that the expansion be- sphere, plant and soil science also in future studies. The
yond the chase period has to be checked independently suggested method can be also applied to C pools such as
using other methods. Whereas the atmospheric fluxes microbial biomass and dissolved organic carbon. Also
are mostly representative, at least as long as weather for ecosystem modelers dealing with C pools and fluxes,
conditions and management activities are within a cer- it provides data on C incorporation in pools in absolute
tain range, plant physiological parameters – and thereby units.
partitioning patterns – vary too much to allow transfer of
the result of a single pulse labeling to the whole growth Acknowledgments The project “Investigation of carbon turn-
over of grasslands in a northern Bavarian low mountain range
period (Gregory and Atwell 1991; Kuzyakov and
under extreme climate conditions” was funded within the joint
Domanski 2000). In contrast, a series of labeling pulses research project “FORKAST” by the Bavarian State Ministry of
at regular intervals (Keith et al. 1986; Swinnen et al. Sciences, Research and Arts. The authors wish to acknowledge the
1994; Kuzyakov et al. 1999, 2001; Kuzyakov and support of the participants of the FORKAST project, especially
Prof. Dr. Anke Jentsch and Alexander Ulmer for evaluation of
Schneckenberger 2004; Davenport and Thomas 1988)
biodiversity and species determination on our site. Finally, the
could provide reasonable estimates of the relative authors want to thank Prof. Dr. Gerhard Gebauer and his team of
partitioning for the whole growth period, to be applied the Laboratory of Isotope Biogeochemistry for the abundance
to the more easily available time series of C input, analysis of the carbon isotopes, all colleagues, technicians and
research assistants who took part in the field and laboratory work,
obtained by micrometeorological flux measurements.
especially Martin Rimmler, Martin Pannek, Ilse Thaufelder,
This way, mowing events or grazing could also be Johannes Olesch and last but not least Peng Zhao for his support
considered. with the flux partitioning model.
Plant Soil (2015) 390:61–76 73

References Buchmann N (2002) Plant ecophysiology and forest response to

global change. Tree Physiol: 1177–1184
Budge K, Leifeld J, Hiltbrunner E, Fuhrer J (2011) Alpine grass-
Allard V, Robin C, Newton P, Lieffering M, Soussana J (2006) land soils contain large proportion of labile carbon but indi-
Short and long–term effects of elevated CO2 on Lolium cate long turnover times. Biogeosciences 8:1911–1923
perenne rhizodeposition and its consequences on soil organic Butler JL, Bottomley PJ, Griffith SM, Myrold DD (2004)
matter turnover and plant N yield. Soil Biol Biochem 38: Distribution and turnover of recently fixed photosynthate in
1178–1187 ryegrass rhizospheres. Soil Biol Biochem 36:371–382
Allard V, Soussana J, Falcimagne R, Berbigier P, Bonnefond J, Chou WW, Silver WL, Jackson RD, Thompson AW, Allen–Diaz
Ceschia E, D’hour P, Hénault C, Laville P, Martin C, Pinarès– B (2008) The sensitivity of annual grassland carbon cycling
Patino C (2007) The role of grazing management for the net to the quantity and timing of rainfall. Glob Chang Biol 14:
biome productivity and greenhouse gas budget (CO2, N2O 1382–1394
and CH4) of semi–natural grassland. Agric Ecosyst Environ Ciais P, Reichstein M, Viovy N, Granier A, Ogée J, Allard V,
121:47–58 Aubinet M, Buchmann N, Bernhofer C, Carrara A,
Ammann C, Flechard CR, Leifeld J, Neftel A, Fuhrer J (2007) The Chevallier F, de Noblet N, Friend AD, Friedlingstein P,
carbon budget of newly established temperate grassland de- Grünwald T, Heinesch B, Keronen P, Knohl A, Krinner G,
pends on management intensity. The greenhouse gas balance Loustau D, Manca G, Matteucci G, Miglietta F, Ourcival JM,
of grasslands in Europe. Agric Ecosyst Environ 121:5–20 Papale D, Pilegaard K, Rambal S, Seufert G, Soussana JF,
Ammann C, Spirig C, Leifeld J, Neftel A (2009) Assessment of Sanz MJ, Schulze ED, Vesala T, Valentini R (2005) Europe–
the nitrogen and carbon budget of two managed temperate wide reduction in primary productivity caused by the heat
grassland fields. Agric Ecosyst Environ 133:150–162 and drought in 2003. Nature 437:529–533
Aubinet M, Grelle A, Ibrom A, Rannik Ü, Moncrieff J, Foken T, Ciais P, Soussana JF, Vuichard N, Luyssaert S, Don A, Janssens
Kowalski AS, Martin PH, Berbigier P, Bernhofer C, Clement IA, Piao SL, Dechow R, Lathière J, Maignan F, Wattenbach
R, Elbers JA, Granier A, Grünwald T, Morgenstern K, M, Smith P, Ammann C, Freibauer A, Schulze ED (2010)
Pilegaard K, Rebmann C, Snijders W, Valentini R, Vesala T The greenhouse gas balance of European grasslands.
(2000) Estimates of the Annual Net Carbon and Water Biogeosci Discus 7:5997–6050
Exchange of Forests: The EUROFLUX Methodology. Adv Coplen TB (2011) Guidelines and recommended terms for expres-
Ecol Res: 113–175 sion of stable–isotope–ratio and gas–ratio measurement re-
Aubinet M, Vesala T, Papale D (2012) Eddy covariance. Springer, sults. Rapid Commun Mass Spectrom 25:2538–2560
Dordrecht Davenport JR, Thomas RL (1988) Carbon partitioning and
Bahn M, Schmitt M, Siegwolf R, Richter A, Brüggemann N rhizodeposition in corn and bromegrass. Can J soil Sci:
(2009) Does photosynthesis affect grassland soil–respired 693–701
CO 2 and its carbon isotope composition on a diurnal time- Davidson EA, Savage K, Verchot LV, Navarro R (2002)
scale? New Phytol 182:451–460 Minimizing artefacts and biases in chamber–based measure-
Baldocchi DD (2003) Assessing the eddy covariance technique for ments of soil respiration. FLUXNET 2000 Synthesis. Agric
evaluating carbon dioxide exchange rates of ecosystems: For Meteorol 113:21–37
past, present and future. Glob Chang Biol 9:479–492 Desai AR, Richardson AD, Moffat AM, Kattge J, Hollinger DY,
Baldocchi DD, Falge E, Gu LH, Olson R, Hollinger DY, Running Barr A, Falge E, Noormets A, Papale D, Reichstein M,
S, Anthoni P, Bernhofer C, Davis K, Evans R, Fuentes J, Stauch VJ (2008) Cross–site evaluation of eddy covariance
Goldstein AH, Katul GG, Law BE, Lee XH, Malhi Y, Meyers GPP and RE decomposition techniques. Agric For Meteorol
TP, Munger JW, Oechel W, Paw UKT, Pilegaard K, Schmid 148:821–838
HP, Valentini R, Verma S, Vesala T, Wilson K, Wofsy S Domanski G, Kuzyakov Y, Siniakina SV, Stahr K (2001) Carbon
(2001) FLUXNET: a new tool to study the temporal and flows in the rhizosphere of ryegrass (Lolium perenne). J Plant
spatial variability of ecosystem–scale carbon dioxide, water Nutr Soil Sci 164:381–387
vapor, and energy flux densities. Bull Am Meteorol Soc 82: Dore S, Hymus GJ, Johnson DP, Hinkle CR, Valentini R, Drake
2415–2434 BG (2003) Cross validation of open–top chamber and eddy
Bazot S, Ulff L, Blum H, Nguyen C, Robin C (2006) Effects of covariance measurements of ecosystem CO2 exchange in a
elevated CO2 concentration on rhizodeposition from Lolium Florida scrub–oak ecosystem. Glob Chang Biol 9:84–95
perenne grown on soil exposed to 9 years of CO2 enrichment. Drösler M (2005) Trace gas exchange and climatic relevance of
Soil Biol Biochem 38:729–736 bog ecosystems, Southern Germany. PhD–Thesis,
Biasi C, Pitkämäki AS, Tavi NM, Koponen HT, Martikainen PJ Technische Universität München, 179 pp. Online publica-
(2012) An isotope approach based on 13C pulse–chase label- tion: urn:nbn:de:bvb:91–diss20050901–1249431017
ling vs. the root trenching method to separate heterotrophic Eigenmann R, Metzger S, Foken T (2009) Generation of free
and autotrophic respiration in cultivated peat lands. Boreal convection due to changes of the local circulation system.
Environ Res 17(3/4):184–192 Atmos Chem Phys 9:8587–8600
Bowling DR, Tans PP, Monson RK (2001) Partitioning net eco- Falge E, Baldocchi DD, Olson R, Anthoni P, Aubinet M,
system carbon exchange with isotopic fluxes of CO2. Glob Bernhofer C, Burba G, Ceulemans R, Clement R, Dolman
Chang Biol 7:127–145 H, Granier A, Gross P, Grünwald T, Hollinger DY, Jensen N,
Buchmann N (2000) Biotic and abiotic factors controlling soil Katul GG, Keronen P, Kowalski AS, Lai CT, Law BE,
respiration rates in Picea abies stands. Soil Biol Biochem Meyers TP, Moncrieff J, Moors E, Munger JW, Pilegaard
32:1625–1635 K, Rannik Ü, Rebmann C, Suyker A, Tenhunen JD, Tu KP,
74 Plant Soil (2015) 390:61–76

Verma S, Vesala T, Wilson K, Wofsy S (2001) Gap filling sites. Flux and concentration footprints. Agric For Meteorol
strategies for defensible annual sums of net ecosystem ex- 127:175–188
change. Agric For Meteorol 107:43–69 Göckede M, Markkanen T, Hasager CB, Foken T (2006) Update
Falge E, Baldocchi D, Tenhunen J, Aubinet M, Bakwin P, of a footprint–based approach for the characterization of
Berbigier P, Bernhofer C, Burba G, Clement R, Davis KJ, complex measurement sites. Bound-Layer Meteorol 118:
Elbers JA, Goldstein AH, Grelle A, Granier A, 635–655
Guðmundsson J, Hollinger D, Kowalski AS, Katul G, Law Goulden ML, Munger JW, Fan S, Daube BC, Wofsy S (1996)
BE, Malhi Y, Meyers T, Monson RK, Munger J, Oechel W, Measurements of carbon sequestration by long–term eddy
Paw UKT, Pilegaard K, Rannik Ü, Rebmann C, Suyker A, covariance: methods and a critical evaluation of accuracy.
Valentini R, Wilson K, Wofsy S (2002) Seasonality of eco- Glob Chang Biol 2:169–182
system respiration and gross primary production as derived Gregory PJ, Atwell BJ (1991) The fate of carbon in pulse–labelled
from FLUXNET measurements. Agric For Meteorol 113:53– crops of barley and wheat. Plant Soil 136:205–213
74 Gupta SR, Singh JS (1977) Effect of alkali concentration, volume
Foken T (2003) Lufthygienisch–bioklimatische Kennzeichnung and absorption area on the measurement of soil respiration in
des oberen Egertales (Fichtelgebirge bis Karlovy Vary). a tropical sward. Pedobiologia: 233–239
Bayreuther Forum Ökologie: 70 S Hafner S, Unteregelsbacher S, Seeber E, Lena B, Xu X, Li X,
Foken T, Wichura B (1996) Tools for quality assessment of surface Guggenberger G, Miehe G, Kuzyakov Y (2012) Effect of
based flux measurements. Agric For Meteorol 78:83–105 grazing on carbon stocks and assimilate partitioning in a
Foken T, Göckede M, Mauder M, Mahrt L, Amiro BD, Munger Tibetan montane pasture revealed by 13CO2 pulse labeling.
JW (2004) Post–field data quality control. In: Lee X, Glob Chang Biol 18:528–538
Massman W, Law B (eds) Handbook of micrometeorology: Hussain M, Grünwald T, Tenhunen J, Li Y, Mirzae H, Bernhofer
a guide for surface flux measurement and analysis. Kluwer, C, Otieno D, Dinh N, Schmidt M, Wartinger M, Owen K
Dordrecht, pp 181–208 (2011) Summer drought influence on CO2 and water fluxes
Foken T, Leuning R, Oncley SP, Mauder M, Aubinet M (2012) of extensively managed grassland in Germany. Agric Ecosyst
Corrections and data quality. In: Aubinet M, Vesala T, Papale Environ 141:67–76
D (eds) Eddy covariance: a practical guide to measurement IPCC (2007) Climate change 2007 – the physical science basis.
and data analysis. Springer, Dordrecht, pp 85–131 Contribution of working group I to the fourth assessment report
Gamnitzer U, Moyes AB, Bowling DR, Schnyder H (2011) of the IPCC. Cambridge University Press, Cambridge, 996 S
Measuring and modelling the isotopic composition of soil IUSS Working Group WRB (2007) World reference base for soil
respiration: insights from a grassland tracer experiment. resources 2006 first update 2007. World soil resources re-
Biogeosciences 8:1333–1350 ports no. 103. FAO, Rome
Gavrichkova O (2009) Drivers of soil respiration of root and Janssens IA (2003) Europe’s terrestrial biosphere absorbs 7 to
microbial origin in grasslands. PhD–Thesis, Università degli 12 % of European anthropogenic CO2 emissions. Science
Studi della Tuscia. http://dspace.unitus.it/bitstream/2067/ 300:1538–1542
1104/1/ogavrichkova_tesid.pdf Johnson D, Leake JR, Ostle NJ, Ineson P, Read DJ (2002) In situ
Gilmanov TG, Soussana JF, Aires L, Allard V, Ammann C, 13CO2 pulse–labelling of upland grassland demonstrates a
Balzarolo M, Barcza Z, Bernhofer C, Campbell CL, rapid pathway of carbon flux from arbuscular mycorrhizal
Cernusca A, Cescatti A, Clifton–Brown J, Dirks BOM, mycelia to the soil. New Phytol 153:327–334
Dore S, Eugster W, Fuhrer J, Gimeno C, Grünwald T, Jones DL, Nguyen C, Finlay RD (2009) Carbon flow in the
Haszpra L, Hensen A, Ibrom A, Jacobs AFG, Jones MB, rhizosphere: carbon trading at the soil–root interface. Plant
Lanigan G, Laurila T, Lohila A, Manca G, Marcolla B, Nagy Soil 321:5–33
Z, Pilegaard K, Pinter K, Pio C, Raschi A, Rogiers N, Sanz Joos O, Hagedorn F, Heim A, Gilgen AK, Schmidt MWI,
MJ, Stefani P, Sutton MA, Tuba Z, Valentini R, Williams Siegwolf RTW, Buchmann N (2010) Summer drought re-
ML, Wohlfahrt G (2007) Partitioning European grassland net duces total and litter–derived soil CO2 effluxes in temperate
ecosystem CO2 exchange into gross primary productivity and grassland – clues from a 13C litter addition experiment.
ecosystem respiration using light response function analysis. Biogeosciences 7:1031–1041
Agric Ecosyst Environ 121:93–120 Kaštovská E, Šantrůčková H (2007) Fate and dynamics of recently
Gilmanov TG, Aires L, Barcza Z, Baron VS, Belelli L, Beringer J, fixed C in pasture plant–soil system under field conditions.
Billesbach D, Bonal D, Bradford J, Ceschia E, Cook D, Plant Soil 300:61–69
Corradi C, Frank A, Gianelle D, Gimeno C, Grünwald T, Kayler Z, Ganio L, Hauck M, Pypker TG, Sulzman EW, Mix AC,
Guo H, Hanan N, Haszpra L, Heilman J, Jacobs AFG, Jones Bond B.J (2010) Bias and uncertainty of δ13CO2 isotopic
MB, Johnson DA, Kiely G, Li S, Magliulo V, Moors E, Nagy mixing models. Oecologia: 227–234
Z, Nasyrov M, Owensby C, Pinter K, Pio C, Reichstein M, Keith H, Oades JM, Martin JK (1986) Input of carbon to soil from
Sanz MJ, Scott R, Soussana JF, Stoy PC, Svejcar T, Tuba Z, wheat plants. Soil Biol Biochem 18:445–449
Zhou G (2010) Productivity, respiration, and light–response Kirita H (1971) Re–examination of the absorption method of mea-
parameters of world grassland and agroecosystems derived suring soil respira tion under field conditions III. Combined
from flux–tower measurements. Rangel Ecol Manag 63:16– effect of the covered ground area and the surface area of KOH
39 solution on CO2–absorption rates. J J Ecol: 37–42
Göckede M, Rebmann C, Foken T (2004) A combination of Knapp AK (2002) Rainfall variability, carbon cycling, and plant
quality assessment tools for eddy covariance measurements species diversity in a mesic grassland. Science 298:2202–
with footprint modelling for the characterisation of complex 2205
Plant Soil (2015) 390:61–76 75

Kuzyakov Y (2006) Sources of CO2 efflux from soil and review of McDermitt DK, Welles JM, Eckles RD (1993) Effects of temper-
partitioning methods. Soil Biol Biochem 38:425–448 ature, pressure and water vapor on gas phase infrared absorp-
Kuzyakov Y, Domanski G (2000) Carbon input by plants into the tion by CO2. application note # 116. Li–COR, Lincoln
soil. Review. J Plant Nutr Soil Sci: 421–431 Michaelis L, Menten ML (1913) Die Kinetik der Invertinwirkung,
Kuzyakov Y, Schneckenberger K (2004) Review of estimation of Kinetics of the invertin reaction. Biochemische Zeitschrift:
plant rhizodeposition and their contribution to soil organic 333–369
matter formation. Arch Agron Soil Sci 50:115–132 Miller J, Tans PP (2003) Calculating isotopic fractionation from
Kuzyakov Y, Kretzschmar A, Stahr K (1999) Contribution of atmospheric measurements at various scales. Tellus B: 207–
Lolium perenne rhizodeposition to carbon turnover of pasture 214
soil. Plant Soil 213:127–136 Moffat AM, Papale D, Reichstein M, Hollinger DY, Richardson
Kuzyakov Y, Ehrensberger H, Stahr K (2001) Carbon partitioning AD, Barr AG, Beckstein C, Braswell BH, Churkina G, Desai
and below–ground translocation by Lolium perenne. Soil AR, Falge E, Gove JH, Heimann M, Hui D, Jarvis AJ, Kattge
Biol Biochem 33:61–74 J, Noormets A, Stauch VJ (2007) Comprehensive compari-
Lasslop G, Reichstein M, Papale D, Richardson AD, Arneth A, son of gap–filling techniques for eddy covariance net carbon
Barr A, Stoy PC, Wohlfahrt G (2010) Separation of net fluxes. Agric For Meteorol 147:209–232
ecosystem exchange into assimilation and respiration using Moncrieff J, Valentini R, Greco S, Seufert G, Ciccioli P (1997)
a light response curve approach: critical issues and global Trace gas exchange over terrestrial ecosystems: methods and
evaluation. Glob Chang Biol 16:187–208 perspectives in micrometeorology. J Exp Bot 48:1133–1142
Leake JR, Ostle NJ, Rangel–Castro I, Johnson D (2006) Carbon Nguyen C (2009) Rhizodeposition of organic C by plant:
fluxes from plants through soil organisms determined by field mechanisms and controls. J Sustain Agric: 97–123
13
CO2 pulse–labelling in an upland grassland. Appl Soil Ecol Ostle NJ, Ineson P, Benham D, Sleep D (2000) Carbon assimila-
33:152–175 tion and turnover in grassland vegetation using an in
Leifeld J, Zimmermann M, Fuhrer J, Conen F (2009) Storage situ13CO2 pulse labelling system. Rapid Commun Mass
and turnover of carbon in grassland soils along an Spectrom 14:1345–1350
elevation gradient in the Swiss Alps. Glob Chang Biol Papale D (2012) Data gap filling. In: Aubinet M, Vesala T, Papale
15:668–679 D (eds) Eddy covariance: a practical guide to measurement
Li M, Babel W, Tanaka K, Foken T (2013) Note on the and data analysis. Springer, Dordrecht, pp 159–172
applifcation of planar–fit rotation for non–omnidirectional Pausch J, Kuzyakov Y (2012) Soil organic carbon decomposition
sonic anemometers. Atmos Meas Techn: 221–229 from recently added and older sources estimated by δ13C
Lloyd J, Taylor J (1994) On the temperature dependence of soil values of CO2 and organic matter. Soil Biol Biochem: 40–47
respiration. Funct. Ecol.: 315–323 Rangel–Castro I, Prosser JI, Scrimgeour CM, Smith P, Ostle NJ,
Lundegardh H (1921) Ecological studies in the assimilation of Ineson P, Meharg A, Killham K (2004) Carbon flow in an
certain forest plants and shore plants. Svensk Botaniska upland grassland: effect of liming on the flux of recently
Tidskrift: 46–94 photosynthesized carbon to rhizosphere soil. Glob Chang
Luo Y (2007) Terrestrial carbon–cycle feedback to climate Biol 10:2100–2108
warming. Annu Rev Ecol Evol Syst 38:683–712 Rannik Ü, Sogachev A, Foken T, Göckede M, Kljun N, Leclerc
Luo Y, Hui D, Zhang D (2006) Elevated CO2 stimulates net MY, Vesala T (2012) Footprint analysis. In: Aubinet M,
accumulations of carbon and nitrogen in land ecosystems: a Vesala T, Papale D (eds) Covariance: a practical guide to
meta–analysis. Ecology 87:53–63 measurement and data analysis. Springer, Netherlands, pp
Mauder M, Foken T (2004) Documentation and instruction man- 211–261
ual of the eddy covariance software package TK2, Univ. Rattray EAS, Paterson E, Killham K (1995) Characterisation of
Bayreuth, Abt. Mikrometeorol., ISSN: 1614–89166, 26, the dynamics of C–partitioning within Lolium perenne and to
42 pp the rhizosphere microbial biomass using 14C pulse chase.
Mauder M, Foken T (2011) Documentation and instruction man- Biol Fertil Soils 19:280–286
ual of the eddy covariance software package TK3. Reichstein M, Falge E, Baldocchi D, Papale D, Aubinet M,
Universität Bayreuth, Abt. Mikrometeorologie, Print, ISSN Berbigier P, Bernhofer C, Buchmann N, Gilmanov T,
1614–8916, Univ. Bayreuth, Abt. Mikrometeorol., ISSN: Granier A, Grunwald T, Havrankova K, Ilvesniemi H,
1614–89166, 46, 58 pp Janous D, Knohl A, Laurila T, Lohila A, Loustau D,
Mauder M, Liebethal C, Göckede M, Leps J, Beyrich F, Foken T Matteucci G, Meyers T, Miglietta F, Ourcival J, Pumpanen
(2006) Processing and quality control of flux data during J, Rambal S, Rotenberg E, Sanz M, Tenhunen J, Seufert G,
LITFASS–2003. Bound-Layer Meteorol 121:67–88 Vaccari F, Vesala T, Yakir D, Valentini R (2005) On the
Mauder M, Oncley SP, Vogt R, Weidinger T, Ribeiro L, separation of net ecosystem exchange into assimilation and
Bernhofer C, Foken T, Kohsiek W, Bruin HAR, Liu H ecosystem respiration: review and improved algorithm. Glob
(2007) The energy balance experiment EBEX–2000. Part Chang Biol 11:1424–1439
II: intercomparison of eddy–covariance sensors and post– Reichstein M, Stoy PC, Desai AR, Lasslop G, Richardson AD
field data processing methods. Bound-Layer Meteorol (2012) Partitioning of net fluxes. In: Aubinet M, Vesala T,
123:29–54 Papale D (eds) Eddy covariance: a practical guide to mea-
Mauder M, Foken T, Clement R, Elbers JA, Eugster W, Grünwald surement and data analysis. Partitioning of net fluxes.
T, Heusinkveld B, Kolle O (2008) Quality control of Springer, Dordrecht, pp 263–289
CarboEurope flux data – part 2: inter–comparison of eddy– Rochette P, Hutchinson GL (2005) Measurement of soil respira-
covariance software. Biogeosciences 5:451–462 tion in situ: chamber techniques. In: Hatfield JL, Baker JM
76 Plant Soil (2015) 390:61–76

(eds) Micrometeorology in agricultural systems. American Unteregelsbacher S, Hafner S, Guggenberger G, Miehe G, Xu X,

Society of Agronomy; Crop Science Society of America; Soil Liu J, Kuzyakov Y (2011) Response of long–, medium– and
Science Society of America, Madison short–term processes of the carbon budget to overgrazing–
Ruppert J, Mauder M, Thomas C, Lüers J (2006) Innovative gap– induced crusts in the Tibetan Plateau. Biogeochemistry
filling strategy for annual sums of CO2 net ecosystem ex- van Ginkel J, Gorissen A, van Veen JA (1997) Carbon and
change. Agric For Meteorol 138:5–18 nitrogen allocation in Lolium perenne in response to elevated
Saggar S, Hedley C (2001) Estimating seasonal and annual carbon atmospheric CO2 with emphasis on soil carbon dynamics.
inputs, and root decomposition rates in a temperate pasture Plant Soil 188:299–308
following field 14C pulse–labelling. Plant Soil 236:91–103 Vesala T, Kljun N, Rannik Ü, Rinne J, Sogachev A, Markkanen T,
Saggar S, Hedley C, Mackay AD (1997) Partitioning and translo- Sabelfeld K, Foken T, Leclerc M (2008) Flux and concentra-
cation of photosynthetically fixed 14 C in grazed hill pas- tion footprint modelling: state of the art. Environ Pollut 152:
tures. Biol Fertil Soils 25:152–158 653–666
Schulze ED, Ciais P, Luyssaert S, Schrumpf M, Janssens IA, Werth M, Kuzyakov Y (2008) Root-derived carbon in soil
Thiruchittampalam B, Theloke J, Saurat M, Bringezu S, respiration and microbial biomass determined by 14C and
13
Lelieveld J, Lohila A, Rebmann C, Jung M, Bastviken D, C. Soil Biol Biochem 40(3):625–637. doi:10.1016/j.
Abril G, Grassi G, Leip A, Freibauer A, Kutsch W, Don A, soilbio.2007.09.022
Nieschulze J, Börner A, Gash JH, Dolman AJ (2010) The Whipps JM (1990) Carbon economy. In: Lynch JM (ed) The
European carbon balance. Part 4: integration of carbon and rhizosphere. John Wiley, Chichester, pp 59–97
other trace–gas fluxes. Glob Chang Biol 16:1451–1469 Wichura B (2009) Untersuchungen zum Kohlendioxid–Austausch
Singh J, Gupta SR (1977) Plant decomposition and soil respiration über einem Fichtenwaldbestand auf der Grundlage von
in terrestrial eco–systems. Bot Rev: 450–511 Hyperbolic–Relaxed–Eddy–Accumulation Messungen für
Sokal RR, Rohlf FJ (2008) Biometry, 4th edn. W.H. Freeman, das stabile Kohlenstoffisotop 13C und von
New York Wavelletanalysen. Dissertation. Bayreuther Forum
Staddon PL (2003) The speed of soil carbon throughput in an upland Ökologie, Band 114, 324 S
grassland is increased by liming. J Exp Bot 54:1461–1469 Wohlfahrt G, Hammerle A, Haslwanter A, Bahn M, Tappeiner U,
Stoy PC, Katul GG, Siqueira MB, Juang J, Novick KA, Uebelherr Cernusca A (2008) Seasonal and inter–annual variability of
JM, Oren R (2006) An evaluation of models for partitioning the net ecosystem CO2 exchange of a temperate mountain
eddy covariance–measured net ecosystem exchange into grassland: effects of weather and management. J Geophys
photosynthesis and respiration. Agric For Meteorol 141:2–18 Res 113:1–14
Subke J, Tenhunen JD (2004) Direct measurements of CO2 flux Wohlfahrt G, Klumpp K, Soussana JF (2012) Eddy covariance
below a spruce forest canopy. Agric For Meteorol 126:157– measurements over grasslands. In: Aubinet M, Vesala T,
168 Papale D (eds) Eddy covariance: a practical guide to mea-
Subke J, Vallack HW, Magnusson T, Keel SG, Metcalfe DB, surement and data analysis. Springer, Dordrecht, pp 333–344
Högberg P, Ineson P (2009) Short–term dynamics of abiotic Wu Y, Tan H, Deng Y, Wu J, Xu X, Wang Y, Tang Y, Higashi T,
and biotic soil 13CO2 effluxes after in situ 13CO2 pulse Cui X (2010) Partitioning pattern of carbon flux in a Kobresia
labelling of a boreal pine forest. New Phytol 183:349–357 grassland on the Qinghai–Tibetan Plateau revealed by field
Swinnen J, van Veen JA, Merckx R (1994) 14C pulse–labelling of 13
C pulse–labeling. Glob Chang Biol 16:2322–2333
field–grown spring wheat: an evaluation of its use in rhizosphere Yakir D, Sternberg LSL (2000) The use of stable isotopes to study
carbon budget estimations. Soil Biol Biochem 26:161–170 ecosystem gas exchange. Oecologia 123:297–311
Todorovic C, Nguyen C, Robin C, Guckert A (1999) 14C–assim- Zobitz J, Keener JP, Schnyder H, Bowling DR (2006) Sensitivity
ilate partitioning within white clover plant–soil system: ef- analysis and quantification of uncertainty for isotopic mixing
fects of photoperiod/temperature treatments and defoliation. relationships in carbon cycle research. Agric For Meteorol:
Eur J Agron 11:13–21 56–75