36
Another group of double sampling techniques
uses characteristics of the structure of the community or its individuals. Good results have been obtained using a pin that passes through the vegetation
and assesses canopy interception (Frank and
McNaughton 1990). The r for the relationship
between canopy interception and biomass ranged
between 0.831 and 0.956 but the relationship
changed with leaf size and thickness, indicating the
need to use new calibration curves when there are
changes in plant growth forms in space, time, or
treatment (e.g., grazed vs. ungrazed plots). At the
level of individuals, there are good correlations
(r = 0.997) between stem diameter and aboveground biomass for a large number of species spanning orders of magnitude of biomass (Freedman
1983).
Methods to Estimate ANPP
in Slow Turnover Ecosystems
Ecosystems with marked differences in the turnover time of their aboveground components represent a challenge for effectively assessing primary
production. For example, forests, woodlands, and
shrublands have two separate components of
aboveground primary production: leaves that are
produced in a given time interval, and growth increment of woody material. Each has a different
turnover time and the methodology used to estimate
its primary production is different. As a result of
these characteristics of ecosystems with woody
vegetation, proximate measurements of biomass
are used more frequently than in the case of systems
dominated by herbaceous vegetation.
Production estimates for leaves and small twigs
in forests usually involve the use of baskets or litter
traps distributed in the forest understory (Whittaker
and Marks 1975). The objective is to collect leaves
and small twigs over a time course of at least a year,
with the litterfall representing the aboveground production in the given time interval. Seasonality of
litterfall, rate of litter decomposition, and climatic
factors all must be taken into account when designing a sample scheme for litterfall collections. Litter
traps should be located at random within a defined
plot area, with care taken to avoid edge effects of
the plot. The number of litter traps necessary will
Osvaldo E. Sala and Amy T. Austin
be determined largely by the heterogeneity of the
system, with a minimum of 10 traps per plot area.
The traps can be constructed of mesh-screen or
nylon fabric, attached to a plastic or wooden frame,
and elevated off the ground to avoid contamination
with soil and waterlogging. Depending on the dynamics of the forest under study, pickups of litter
may need to be completed weekly in the case of
some tropical forests (e.g., Wright and Cornejo
1990), at a fixed time interval (e.g., Ogden and
Schmidt 1998), or at the end of the season in temperate deciduous forests. Once taken from the field,
litter is sorted by type or species and dried in an
oven at 70°C for 48 hours for determinations of dry
mass.
Additionally, a double sampling technique has
been developed to estimate leaf production. Leaf
area index (LAI), which is the leaf area in square
meters of leaf area per square meter of ground area,
is estimated by measuring light intensity at the bottom and top of the canopy using a commercial device called a plant canopy analyzer (e.g., Li-Cor
1992). Leaf area index is then modeled using the
following equation: ILIIO = e - kLAI(L), where IL is
incident light at the bottom of the canopy, 10 is
incident light at the top of the canopy, and k is the
light extinction coefficient, which varies according
to the type of vegetation and angle of inclination of
the leaves (Aber and Melillo 1991). Changes in
LAI over time can be converted to leaf biomass and
used to represent leaf production in forests.
The development of methods for measurements
of the woody component of ANPP in forest ecosystems came primarily from yield assessment of
lumber for silviculture and forestry. As a result,
there is allometric information on many economically important tree species with regression equations that relate tree height and diameter at breast
height (DB H) to biomass (e.g., Curtin 1970; Curtis
and Reukema 1970). For species that do not have
published yield tables, however, which is the case
for most nonmanaged forest ecosystems, it is necessary to measure species-specific allometric relationships (i.e., the shape of the trunks and how
trunk diameters change with height) in order to estimate biomass and tree production. Because of the
logistic difficulty of multiple measurements of all
components of woody biomass, dimensional analysis is often used (e.g., Whittaker and Marks 1975;
Whittaker and WoodweIl1968). This approach in-
Another group of double sampling techniques
uses characteristics of the structure of the community or its individuals. Good results have been obtained using a pin that passes through the vegetation
and assesses canopy interception (Frank and
McNaughton 1990). The r for the relationship
between canopy interception and biomass ranged
between 0.831 and 0.956 but the relationship
changed with leaf size and thickness, indicating the
need to use new calibration curves when there are
changes in plant growth forms in space, time, or
treatment (e.g., grazed vs. ungrazed plots). At the
level of individuals, there are good correlations
(r = 0.997) between stem diameter and aboveground biomass for a large number of species spanning orders of magnitude of biomass (Freedman
1983).
Methods to Estimate ANPP
in Slow Turnover Ecosystems
Ecosystems with marked differences in the turnover time of their aboveground components represent a challenge for effectively assessing primary
production. For example, forests, woodlands, and
shrublands have two separate components of
aboveground primary production: leaves that are
produced in a given time interval, and growth increment of woody material. Each has a different
turnover time and the methodology used to estimate
its primary production is different. As a result of
these characteristics of ecosystems with woody
vegetation, proximate measurements of biomass
are used more frequently than in the case of systems
dominated by herbaceous vegetation.
Production estimates for leaves and small twigs
in forests usually involve the use of baskets or litter
traps distributed in the forest understory (Whittaker
and Marks 1975). The objective is to collect leaves
and small twigs over a time course of at least a year,
with the litterfall representing the aboveground production in the given time interval. Seasonality of
litterfall, rate of litter decomposition, and climatic
factors all must be taken into account when designing a sample scheme for litterfall collections. Litter
traps should be located at random within a defined
plot area, with care taken to avoid edge effects of
the plot. The number of litter traps necessary will
Osvaldo E. Sala and Amy T. Austin
be determined largely by the heterogeneity of the
system, with a minimum of 10 traps per plot area.
The traps can be constructed of mesh-screen or
nylon fabric, attached to a plastic or wooden frame,
and elevated off the ground to avoid contamination
with soil and waterlogging. Depending on the dynamics of the forest under study, pickups of litter
may need to be completed weekly in the case of
some tropical forests (e.g., Wright and Cornejo
1990), at a fixed time interval (e.g., Ogden and
Schmidt 1998), or at the end of the season in temperate deciduous forests. Once taken from the field,
litter is sorted by type or species and dried in an
oven at 70°C for 48 hours for determinations of dry
mass.
Additionally, a double sampling technique has
been developed to estimate leaf production. Leaf
area index (LAI), which is the leaf area in square
meters of leaf area per square meter of ground area,
is estimated by measuring light intensity at the bottom and top of the canopy using a commercial device called a plant canopy analyzer (e.g., Li-Cor
1992). Leaf area index is then modeled using the
following equation: ILIIO = e - kLAI(L), where IL is
incident light at the bottom of the canopy, 10 is
incident light at the top of the canopy, and k is the
light extinction coefficient, which varies according
to the type of vegetation and angle of inclination of
the leaves (Aber and Melillo 1991). Changes in
LAI over time can be converted to leaf biomass and
used to represent leaf production in forests.
The development of methods for measurements
of the woody component of ANPP in forest ecosystems came primarily from yield assessment of
lumber for silviculture and forestry. As a result,
there is allometric information on many economically important tree species with regression equations that relate tree height and diameter at breast
height (DB H) to biomass (e.g., Curtin 1970; Curtis
and Reukema 1970). For species that do not have
published yield tables, however, which is the case
for most nonmanaged forest ecosystems, it is necessary to measure species-specific allometric relationships (i.e., the shape of the trunks and how
trunk diameters change with height) in order to estimate biomass and tree production. Because of the
logistic difficulty of multiple measurements of all
components of woody biomass, dimensional analysis is often used (e.g., Whittaker and Marks 1975;
Whittaker and WoodweIl1968). This approach in-