Tree Growth and Ecosystem Respiration in Central Amazon Forest

1
Tree Growth and Ecosystem Respiration in Central
Amazon Forest
Jeffrey Q. Chambers, Edgard S. Tribuzy, Roseana
P. da Silva, Ligia C. Toledo, Joaquim dos Santos,
Niro Higuchi, and Susan E. Trumbore
2
Calculating Forest Net Primary Productivity
When estimating production from permanent
inventory plots both old and new losses must be
considered. Typically, for above-ground NPP, the
biomass increment and fine litterfall is
considered. Branchfall is usually not considered
in NPP estimates.
Trees are not simple mass accumulators.
3
Allometric Models for Predicting Tree Mass
Tree mass allometry should be based on a
representative sample including damaged and
senescent trees. Accelerating losses for the
largest trees shows up as curvature in the
relationship. We have developed a generic moist
forest allometric model from five tropical forest
sites.
An unbiased relationship followed a log-log-cubic
curve.
4
Why does branchfall need to be included in NPP
estimates? Envision a hypothetical forest where
trees are losing branches as fast as new wood is
being produced so that tree mass over time does
not increase (Mt1 Mt2). The production not
being accounted for here is equal to branchfall
(Lbranch). Because trees both gain and lose mass
as they increase in size, both old and new losses
must be included.
5
Estimates are based on numerous field studies
over large temporal and spatial
scales. Below-ground productivity can be
estimated using a modified Raich and Nadelhoffer
method ...
0.4
4.0 0.3
(?)
2.1 0.2
Units Mg C ha-1 yr-1
Mass losses from damaged trees is a relatively
small but significant flux.
6
Central Amazon Forest Autotrophic Respiration
Total 29.3
BG Assumptions Changes in SOM relatively small
BG production and mortality are about equal and
carbon use efficiencies (CUEs) AG and BG about
equal.
17.5
Rleaf
6.8
5.0
RBG
Rstem
CUE 0.23
Units Mg C ha-1 yr-1
AG respiration is considerably higher than BG
respiration.
7
Values presented for NPP and respiration are
long-term average fluxes. We are also finding
considerable temporal and spatial variability in
these fluxes.
Total 8.6
0.4
4.0 0.3
2.1
2.1 0.2
Units Mg C ha-1 yr-1
AG-NPP is considerably greater than BG-NPP.
8
Partitioning of NPP in Central Amazon Forest
To quantify how carbon storage responds to an
increase in productivity both the allocation of
NPP ...
Woody tissues represent 49 of above-ground NPP
9
Carbon Pools in Central Amazon Forest
and the size of the carbon reservoir determines
the residence time and the capacity for
sequestering carbon.
Only large wood and SOM have a high capacity to
sequester carbon for long periods of time.
10
We have developed a stochastic-empirical
individual-based model to explore the carbon
cycling dynamics of large wood in Central Amazon
forest.
Tree
Stand
Plot
11
Carbon cycling structure of the model
We can use this model to explore how changes
affecting individual trees and canopy gaps
influences ecosystem scale carbon cycling and
storage.
12
Carbon balance with only background mortality
Total large wood (TLW) includes both dead and
live material. Points above the line indicate
TLW is a carbon source, and below the line TLW is
a carbon sink. TLW can act as a carbon sink by
and increase in production or a decrease in
coarse litter respiration.
100 ha model run
Considerable temporal variability but long-term
carbon balance.
13
Carbon balance w/ background and catastrophic
mortality
This simulation shows how large wood carbon
balance responds to a single 50 mortality event
at year 1000. There is a quick and large loss of
ecosystem carbon right after the
event. Afterwards large wood acts as a small
sink for many years.
50 mortality event
Also in long-term carbon balance but this forest
is usually a carbon sink.
14
Carbon balance w/ background and catastrophic
mortality
Catastrophic mortality can play an important role
in temporal variability in ecosystem carbon
balance. It is important to quantify the
frequency and intensity of these events. Large
mortality events referred to as blowdowns are
known to occur throughout the Amazon.
20 mortality events
There is a similar response to 20 mortality
events every 200 years.
15
Ikonos image of 10x10 km area in the Central
Amazon
High resolution satellite imagery is useful for
quantifying the frequency and extent of
catastrophic mortality events such as this
blowdown which can exceed 2,000 ha in size.
200 m
Severe downburst type winds are associated with
late dry season storms.
16
The carbon sequestration potential of large wood
We can also use the model to explore how large
wood carbon balance responds to an increase in
NPP. NPP was increased 0.5 per year for fifty
years (the gray area). Large wood sequestered
carbon for many years but at a low annual rate.
The annual sink (0.5 Mg C ha-1 yr-1) is much
lower than estimate from eddy covariance tower
studies.
17
Precipitation Variability in Manaus from 1910-85
The Central Amazon experiences considerable
intra- and inter-annual variability in
precipitation. Changes in moisture availability
affects a number of physiological processes such
as respiration and photosynthesis. El Niño often
results in a dry early wet season.
18
Seasonal Variability in Forest Wood Production
There is a steep increase in wood production at
the onset of the rainy season. Using linear
gauge tensiometers and TDR we are comparing stem
volume changes associated with moisture with
changes from actual growth. Daily shrinking and
swelling associated with transpiration is small
compared with growth.
Production changes seasonally
19
Precipitation Variables and Growth using
Regression Analysis
Changes in wood production is not only dependent
on the amount of precipitation, but also on how
precipitation is distributed. When the same
amount of precipitation is distributed over many
days, growth rates decline. This is probably
because light availability is reduced on cloudy
days limiting photosynthesis.
20
A Synthesis of Global NPP Studies
A workshop at NCEAS compiled NPP data from
numerous globally distributed studies. Some
empirical relationships were evident that may
help extrapolate NPP estimates from sites to
regions.
21
Conclusions

• NPP estimates based on tree mass allometry must
  consider both old and new litter losses.
• BG-NPP is about 20 of total NPP.
• Only large wood and SOM can sequester large
  amounts of carbon in response to an increase in
  NPP.
• The frequency and extent of catastrophic
  mortality is important for understanding net
  carbon balance.
• Central Amazon forests can sequester a large
  amount of carbon with increasing NPP, but in any
  given year sequestration is limited to about
  0.5-1.0 Mg C ha-1 yr-1.
• NPP is not a constant and exhibits both intra-
  and inter-annual variability with environmental
  changes.

For further information and reprints contact Jeff
Chambers at chambersjq_at_yahoo.com