Spatial patterns of greenhouse gas emission in a tropical rainforest in
Indonesia
Shigehiro Ishizuka
1,
*
, Anas Iswandi
2
, Yasuhiro Nakajima
3
, Seiichiro Yonemura
3
,
Shigeto Sudo
3
, Haruo Tsuruta
3
and Daniel Muriyarso
4
1
Forestry and Forest Products Research Institute, Sapporo, Hokkaido 062-8516, Japan;
2
Department of Soil
Science, Bogor Agricultural University (IPB), Jl. Raya Pajajaran, Bogor 16143, Indonesia;
3
National Institute
for Agro-Environmental Sciences, Tsukuba, Ibaraki 305-8604, Japan;
4
Department of Geophysics and
Meteorology, Bogor Agricultural University (IPB), Jl. Raya Pajajaran, Bogor 16143, Indonesia;
*
Author for
correspondence (tel.: +81-11-851-4131; fax: +81-11-851-4167; e-mail: ishiz03@ffpri.affrc.go.jp)
Key words: Flux measurement, Geostatistics, Greenhouse gas, Semivariogram, Spatial variability, Sumatra of
Indonesia
Abstract
Spatial patterns of CO
2
, CH
4
, and N
2
O flux were analyzed in the soil of a primary forest in Sumatra, Indonesia.
The fluxes were measured at 3-m intervals on a sampling grid of 8 rows by 10 columns, with fluxes found to be
below the minimum detection level at 12 points for CH
4
and 29 points for N
2
O. All three gas fluxes distributed
log-normally. The means and standard deviations of CO
2
and CH
4
fluxes calculated by the maximum likelihood
method were 3.68 1.32 g C m
–2
d
–1
and 0.79 0.60 mg C m
–2
d
–1
, respectively. The mean and standard
deviation of N
2
O fluxes using a maximum likelihood estimator for the censored data set was 2.99 3.26 gN
m
–2
h
–1
. The spatial dependency of CH
4
fluxes was not detected in 3-m intervals, while weak spatial dependency
was observed in CO
2
and N
2
O fluxes. The coefficients of variation of CH
4
and N
2
O were higher than that of
CO
2
. Some hot spots where high levels of CH
4
and N
2
O were generated in the studied field may increase the
variability of these gases. The resulting patterns of variability suggest that sampling distances of 10 m and
20 m are required to obtain statistically independent samples for CO
2
and N
2
O flux in the studied field,
respectively. But because of weak or no spatial dependency of each flux, a sampling distance of more than 10 m
intervals is enough to prevent a significant problem of autocorrelation for each flux measurement.
Introduction
CO
2
, CH
4
and N
2
O have been targeted as greenhouse
gases Prather et al. 2001. It is very important to
know the precise variation in the intensity of these gas
fluxes in tropical regions. Bouwman et al. 1993 es-
timated the annual global N
2
O emission from soils as
6.8 Tg N, 80% of which was derived from the trop-
ics. The global CH
4
uptake rate by soils was
estimated as 15 to 35 Tg y
–1
, and humid tropical for-
est accounts for 10 to 20% of this Potter et al. 1996.
The annual CO
2
flux from soils to the atmosphere is
estimated to be 76.5 Pg C y
–1
globally Raich and
Potter 1995, and the CO
2
flux from tropical moist
forest is the highest of all terrestrial ecosystems
Raich and Potter 1995.
Most ecosystems are spatially heterogeneous, but
little attention has focused on quantifying the field
scale variability of greenhouse gas emissions, espe-
cially in the tropical regions. For quantitative assess-
ment, representative field measurements and informa-
tion on spatial structure are needed. Geostatistical
analysis is useful for examining the structure of the
spatial variability e.g., Oliver and Webster 1991.
Geostatistical analysis has been applied in various
soil studies, for example, about soil chemical proper-
ties Yost et al. 1982; Paz-González and Taboada
2000, litter components Gourbiere and Debouzie
Nutrient Cycling in Agroecosystems 71: 55–62, 2005.
DOI 10.1007/s10705-004-5284-7 © Springer 2005