Guide To Monitoring Carbon Storage in Agroforestry Projects

A Guide to
Monitoring Carbon Storage

in Forestry and Agroforestry Projects
K.G. MacDicken
Forest Carbon Monitoring Program
Winrock International Institute
for Agricultural Development
+YMHI XS 1SRMXSVMRK *SVIWX 'EVFSR 7XSVEKI
ii
A Guide to
Monitoring Carbon Storage
in Forestry and Agroforestry Projects
K.G. MacDicken
Winrock Internationl Institute for Agricultural Development
Forest Carbon Monitoring Program
October 1997
iii
Contents
Acknowledgments v
Summary 1
1. Introduction 3
1.1 Defining objectives 3
1.2 Factors in inventory design 4
1.3 Effects of product end use 5
1.4 Inventory outputs 7
2. Measuring carbon pools 8
2.1 Inventory design 8
2.2 Inventory timing 12
2.3 Measurement procedures 12
3. Designing monitoring packages for specific land uses 16
3.1 General requirements by land-use 16
3.2 Setting the economic limits 18
4. Tools 20
4.1 Equipment 20
4.2 Using models for interpolation 21
5. Reporting and verification 23
5.1 Reporting 23
5.2 Baseline 23
5.3 Reporting carbon changes 24
5.4 Verifying carbon monitoring estimates 26
Appendix 1 Carbon Inventory Data Form (CIDF) 28
Form A - Level of precision specifications 28
Form B - Project site description 29
Form C - Sampling design 30
Form D - Satellite images 31
Form E - Permanent plot locations 32
iv
Form F - Biomass measurements 34
Form G - Anticipated disposition of biomass 38
Form H - Laboratory methods 39
Form I - Inventory costs 40
Appendix 2 Calculating sample size 41
Appendix 3 Inputs required for carbon modelling 44
Appendix 4 Measuring woody biomass 53
Appendix 5 Field procedures for herbaceous vegetation, soils and litter 65
Appendix 6 Measuring carbon in agroforestry plantings 75
Appendix 7 Estimating root biomass 84
List of Tables
Table 1 Examples of carbon inventory characteristics at three levels of effort 7
Table 2 Suggested allowable limits for measurement error 14
Table 3 Procedures required for reference vs. project case comparisons 17
Table 4 Plot radii for carbon inventory plots 54
Table 5 Comparison of methodologies for determining organic C in soils 69
Table 6 Organic and carbonate carbon mass in soils of the world 70
Table 7 Comparison of methods used for determining total C in soils 73
List of Figures
Figure 1 Flowchart for monitoring major carbon pools 6
Figure 2 Large-scale topographic map with a grid of sample plots 11
Figure 3 Sample point layout using quarter-point plotless method 55
Figure 4 Proper use of a diameter tape 63
Figure 5 Examples of aluminum sampling frames 67
Figure 6 Location of the plot reference point 78
Figure 7 Plot layout for agroforestry and farm forestry inventory plots 79
v
Acknowledgments
This guide is the product of the work of many individuals, and so there are many to thank.
Financial support was provided by the Center for Environment, U.S. Agency for International
Development and the Winrock International Institute for Agricultural Development. Versions
of this guide were reviewed for technical content by the following specialists:
Greg Biging, Dept. of Forestry, University of California at Berkeley, USA
C.B. Briscoe, Silviculturist, Turrialba, Costa Rica
H.E. Burkhart, Dept. of Forestry and Wildlife Resources, VPI, Blacksburg, VA, USA
David E. Chandler, Plantation specialist, Brasilia, Brazil
Noel Cutright, Senior Ecologist, Wisconsin Electric Power Company, USA
Paul Faeth, World Resources Institute, Washington, D.C., USA
Gina Green, The Nature Conservancy, Arlington, VA, USA
Tom Hanson, International Forestry Consultants, Inc., Bellevue, WA, USA
J.P. Kimmins, Dept. of Forest Science, University of British Columbia, Canada
H.G. Lund, U.S. Forest Service, Washington, D.C., USA
Vicente P.G. Moura, EMBRAPA, Vicosa, Brazil
Francisco de Paula Neto, Dept. of Forestry, Universidade Federal de Vicosa, Brazil
Michelle Pinard, University of Florida, Gainesville, Florida, USA
Maria das Gracas Ferreira Reis, Dept. of Forestry, Universidade Federal de Vicosa, Brazil
Jim Roshetko, Forestry/Natural Resources Management Program, Winrock International,
USA
John Rombold, College of Forest Resources, University of Washington, USA
Roger Wilson, Programme for Belize, Belize City, Belize
Anthony Young, University of East Anglia, UK
Special thanks are due to Antonio Claret de Oliveira and Peter Althoff of the Mannesmann FI-
EL Florestal Ltda. (Brazil), A. Joy Grant and Roger Wilson of the Programme for Belize,
Geoffrey Blate and Johann Zweede of the Tropical Forest Foundation for their cooperation
and support for field tests of these methods. Loren Ford, Ross Pomfrey and Mike Benge of
USAID provided key administrative and technical support during the planning and
implementation stages. John Kadyszewski, Sinnammal Souppaya and Fay Ellis provided
essential and much appreciated support through the Winrock Renewable Energy and
Environment Program.
K.G. MacDicken, Senior Forestry Specialist
1
Summary
As the international Joint Implementation (JI) program develops a system for trading carbon
credits to offset greenhouse gas emissions, project managers need a reliable basis for
measuring the carbon storage benefits of carbon offset projects.
Monitoring and verifying carbon storage can be expensive, depending on the level of scientific
validity needed. This guide describes a system of cost-effective methods for monitoring and
verification on a commercial basis, for three types of land use: forest plantations, managed
natural forests and agroforestry. Winrock Internationals Forest Carbon Monitoring Program
developed this system with its partners as a way to provide reliable results using accepted
principles and practices of forest inventory, soil science and ecological surveys. Perhaps most
important, the system brings field research methods to bear on commercial-scale inventories,
at levels of precision specified by funding agencies.
Winrocks system assesses changes in four main carbon pools: above-ground biomass, below-
ground biomass, soils and standing litter crop. It aims to assess the net change in each pool for
project and non-project (or pre-project) areas over a specified time period.
Carbon monitoring efforts require specialized equipment, methods and trained personnel that
can be expensive for individual organizations to procure and maintain. This is particularly
true since most monitoring activities are likely to be performed infrequently once every
two to five years. In developing its monitoring system, Winrock has recognized these costs
and aimed to minimize them. The system is therefore designed for collaboration between an
organization with specially-trained personnel and local organizations at each project site.
The system involves the following components:
baseline determination of pre-project carbon pools in biomass, soils and standing litter
crop
establishment of permanent sample plots for periodic measurement of changes in
carbon pools
plotless vegetation survey methods (quarter point and quadrat sampling)1 to measure
carbon stored in non-project areas or areas with sparse vegetation
calculation of the net difference in carbon accumulated in project and non-project land
uses

1In woody savannah areas, the quarter point method helps in laying out measurement units by using the
distance between a systematic sampling point and the nearest tree or shrub. Quadrat sampling involves the
use of a portable sampling frame to delimit an area for measurement.
2
use of SPOT satellite images as gauges of land-use changes, and as base maps for a
microcomputer-based geographic information system
software for calculating minimum sample size, assigning sample unit locations (either
in a systematic grid or randomly), determining the minimum spacing for plots and
optimizing site-specific monitoring plans
computer modelling of changes in carbon storage for periods between field
measurements
a database of biomass partitioning (roots, wood and foliage) for selected species
A companion volume, entitled Field Tests of Carbon Monitoring Methods in Forestry
Projects, will describe field experience with the methods contained in this guide.
3
1. Introduction
This guide describes methods and procedures for measuring the organic carbon stored by
forestry and agroforestry land uses over time. Such a monitoring effort assesses the net
difference in organic carbon stored in soil2 and forest biomass for project and non-project (or
pre-project) sites over a specified period of time. The difference in carbon stored is the
amount of carbon sequestered, or fixed, by the project.
Carbon sequestration is thought to be a promising means for reducing atmospheric carbon
dioxide, an important greenhouse gas. To offset carbon emissions, at least 15 utility
companies and other organizations are involved in Joint Implementation (JI) land-use projects
involving plantations, improved forest management and natural forest preservation.3
A major constraint to successful forestry-based carbon offset programs is the lack of reliable,
accurate and cost-effective methods for monitoring carbon storage. If carbon becomes an
internationally-traded commodity, as it appears likely, then monitoring the amount of carbon
fixed by projects will become a critical component of any trading system. Current efforts at
the site level represent two extremes: either they are based on preliminary assumptions, or
they involve intensive research efforts that are too expensive for widespread use. The system
described in this guide is intended to provide a cost-effective, precise and accurate accounting
of carbon storage in projects. To the extent possible, the methods are standard approaches to
mensuration and analysis of biomass and carbon.
Quantitative monitoring of carbon sequestration over time, and to a lesser extent verification
of estimates, requires a series of carbon inventories. For practical reasons, these inventories
should employ permanent sample plots, with periodic measurement of these sample plots in
baseline and project cases.4 Furthermore, the economic realities of the costs and benefits of a
carbon inventory should be considered in the design stage, so that expectations of precision
are in harmony with the resources available for the monitoring effort.
1.1 Defining objectives
To maximize the utility of information collected in an inventory and to reduce monitoring
costs, forest managers should define the various objectives for an inventory early, in advance
planning.

2 In general, these methods will be used to monitory only organic carbon. Inorganic carbon stored in
carbonate materials is not likely to change with the vegetation changes anticipated in most forestry projects.
3 Faeth, P., R. Livernash and C. Cort. 1993. Evaluating the carbon sequestration benefits of sustainable
forestry projects in developing countries. Washington, DC: World Resources Institute.
4 The baseline case is defined as on-site conditions without project activities. The project case includes on-
site changes in soil and biomass carbon that occur due to project activities.
4
This guide assumes that the primary objective of carbon monitoring and verification is to
produce sound estimates of carbon sequestered by projects, for use in the trade of carbon
credits (even though a system for trading carbon credits has not yet been finalized). Precision
is an important aspect of this objective. When an international carbon trading system is
established, it will likely set precision standards for monitoring carbon in land-use systems.
Until then, decisions on precision level must be made on a project basis as the monitoring
objectives are defined, so that the inventory can be designed to supply the desired precision.
It is possible and often desirable for a carbon inventory to have several concurrent objectives.
The technical rigor required for carbon monitoring can drive the collection of other data on
forest management, and make the process of forest monitoring more cost effective.
Additional objectives might be to track important wildlife populations, or measure biological
diversity or timber species growth. The use of permanent sample plots also provides
opportunities to study nutrient flows, production sustainability, and other trends.
1.2 Factors in inventory design
The use of permanent sample plots is generally regarded as a statistically superior means of
evaluating changes in forest conditions.5 Permanent plots allow efficient assessment of
changes in carbon fixation over time, provided that the plots represent the larger area for
which the estimates are intended. This means that the sample plots must be subject to the
same management as the rest of the project area. The use of permanent plots also allows the
inventory to continue reliably over more than one rotation. Finally, permanent plots permit
efficient verification at relatively low cost: a verifying organization can find and measure
permanent plots at random to verify, in quantitative terms, the design and implementation of a
projects carbon monitoring plan. To achieve the same level of verification with temporary
sample plots or other inventory approaches would require substantially more time and
expense.
The decision on which carbon pools to measure is critical to inventory design. In general, all
pools that are large and subject to substantial change over the project life should be measured.
Those that are small or very slow to change may not need to be measured. It is likely that an
international carbon trading regime will require the project monitoring of all pools that are
likely to decrease over time.
For pools that are likely to increase, a key factor in the design of an inventory is the cost of
measurement and analysis of each component relative to the economic value of fixed carbon.
For example, if carbon credits are worth US$2 per ton, it does not make economic sense to
spend $2.50 per ton on measurements that include root biomass. However, it probably does
make sense to spend $1.00 per ton on measurements that quantify roots.
In the current pilot phase of Joint Implementation (JI), it is assumed that carbon will have
economic value in the future.6 In terms of costs, most carbon sequestration projects have

5 For more details on this, see Forest Measurements, 3rd edition, eds. T.E. Avery and H.E. Burkhart (New
York: McGraw-Hill, 1983) or Forest Mensuration, 3rd edition, eds. B. Husch, C.I. Miller and T.W. Beers
(New York: John Wiley and Sons, 1982).
6 Joint Implementation refers to cooperative development projects that seek to reduce or sequester
greenhouse gas emissions as described in the U.N. Framework Convention on Climate Change.
5
already calculated the cost of carbon fixed on a per ton basis. Inventories that follow the
system described in this guide should be designed based on calculations of the cost per ton of
carbon, as specified by the project implementor. Figure 1 outlines the overall process for
inventory design and implementation.
1.3 Effects of product end use
The long-term effectiveness of carbon sequestration depends in part on the end-uses of the
wood produced through project activities. The more durable the wood product, the greater the
projects carbon storage effect in the medium and long term. However, carbon stored in wood
is obviously not stored permanently; organic compounds eventually decay and some will
ultimately reappear as greenhouse gases. The impacts of carbon sinks are directly
proportional to the ton-years of storage (that is, tons of carbon multiplied by the number of
years for which the carbon is stored). The methods described in this guide do not cover the
storage of carbon in post-harvest sinks. Anticipated disposition of biomass can be recorded in
Form G.
Questions of leakage7 and off-site baseline changes are also important to the overall Joint
Implementation process. Off-site leakage may determine the success or failure of forest
preservation projects, but it is extremely difficult to quantify. Such off-site impacts are
perhaps best estimated using the Land Use and Carbon Sequestration (LUCS) model available
from the World Resources Institute.8

7 In this context, leakage is the loss of carbon (primarily woody biomass) in non-project areas due to
project activities. For example, leakage occurs if a natural forest area that was previously used locally for
timber and firewood, is closed due to a preservation project, causing fuelwood and timber to be harvested
elsewhere.
8 For information on LUCS, contact the World Resources Institute, 1709 New York Ave., NW,
Washington, DC 20006, USA.
6
Determine Herbs and small Number of live trees, Soil carbon Litter green
diameter stem by species or concentration weight
green weight, by plot species group, and dbh (0-30 cm)
Bulk density Moisture
content
Analyze Moisture content Green weight per tree, Literature
by diameter and species: search
total, bole, crown, roots
Soil carbon
Calculate Total weight of herbs, Total biomass of Litter dry
seedlings, saplings trees > 5cm dbh, weight
by species groups
Multiply by 0.5 or
detailed conversion
factors
Total carbon
(t C)
Figure 1. Flowchart for monitoring changes in major carbon pools in forestry and agroforestry projects.
7
Table 1. Examples of three levels of effort for carbon inventory
Level of effort General description
Basic This provides a very general, low-cost estimate of carbon stored in
plantations. Less intensive sampling keeps costs low, but provides estimates
of mean carbon fixation with accuracy approaching 30% of the estimated
mean. Permanent sample plots are measured only twice: at plot
establishment and at final harvest. Modelling produces interim estimates of
carbon fixation in vegetation and soils.
Moderate This level provides carbon storage estimates that are generally within 20%
of the mean. Sampling intensity is greater, resulting in substantially more
precise estimates than the basic inventory. Permanent plots are monitored
every 2-3 years and at final harvest. Predictive models can be used to
provide estimates of annual carbon fixation but would not be used in most
applications.
High This option produces estimates that are accurate within 10-15% of the
amount of carbon sequestered, due to increased sampling and reduced
reliance on models. Permanent sample plots are measured on an annual
basis.
1.4 Inventory outputs
Carbon inventories of land-use projects can provide two general types of information: 1)
carbon inventory reports that document changes in the quantities of carbon fixed due to
project activities, and 2) commercial timber inventories that document quantities of
merchantable timber. This guide describes only the first type, i.e., reports that indicate
changes in carbon that result from project activities. Inventories of commercial timber,
requiring slightly different data collection methods and analysis, can be done at relatively little
additional cost.9
The inventory process usually yields two general types of outputs: baseline reports that
describe carbon pool sizes at the beginning of the project and periodic reports that describe
changes in these pools based on repeated measurement. The initial baseline carbon report
provides an estimate of the quantity and distribution of carbon in vegetation and soils. This
baseline would be produced before project activities begin and would serve as the benchmark
from which future changes in carbon pool size would be calculated. The baseline report
would be produced only once per site.

9 For timber inventory methods compatible with those described in this guide, see Forest Measurements,
3rd edition, eds. T.E. Avery and H.E. Burkhart (New York: McGraw-Hill, 1983) or Forest Mensuration,
3rd edition, eds. B. Husch, C.I. Miller and T.W. Beers (New York: John Wiley and Sons, 1982).
8
Periodic inventory reports based on recurring measurement of permanent sample plots provide
the basis for determining changes in carbon pools. These reports will describe measured
quantities and distribution of organic carbon pools in soils and vegetation in project and non-
project lands and calculate the net carbon stored by project activities and will verify project
area and changes in biomass and soil carbon.10 In order to measure carbon change due to
project activities, both the project and non-project cases must be monitored over time.
The methods in this guide have been field-tested under various conditions in Belize (Orange
Walk District), Brazil (Minas Gerais and Para), Guatemala (La Union), the Philippines
(Isabela Province), and the United States (Washington and Oregon). The experiences of these
field tests are recorded in a companion volume entitled Field Tests of Methods for Monitoring
Carbon in Forestry Projects.
2. Measuring Carbon Pools
Carbon inventories are in effect snapshots of carbon stored at the time of the inventory. To
ensure these snapshots can be usefully compared with each other, it is important for the
inventory team to be consistent in its use of measurement techniques and methods between
different sites, stands, and inventory periods.
The following four carbon pools can be inventoried using the methods outlined in this guide:
1. Above-ground biomass/necromass
2. Below-ground biomass (tree roots)
3. Soil carbon
4. Standing litter crop
Appendices 4 - 6 describe the methods for measuring each of these carbon pools.
As mentioned earlier, permanent sample plots have two main advantages for carbon
monitoring: (1) they provide more reliable data on trends in vegetation development than
temporary plots do; and (2) they are more easily verified than other methods, since permanent
plots can be revisited and remeasured by an external verifier.
The remainder of this guide refers to methods and procedures to be used with permanent
sample plots that will be periodically monitored.
2.1 Inventory design
For the carbon inventories described in this guide, the sample unit is the permanent sample
plot. The sample frame (i.e., the listing of all the sample units) is the projects land area,
excluding buffer zones and areas that are not carbon sinks for project purposes.

10 While the largest largest proportion of carbon changes usually occur in biomass, soil carbon is also
likely to change due to cultivation and conversion to new species. For a more detailed description of the
effects of land-use changes on soil carbon in the tropics, see Dynamics of soil organic matter in tropical
ecosystems, eds. D.C. Coleman, J.M. Oades and G. Uehara (Honolulu: University of Hawaii Press, 1989).
9
Sampling design
There are four options for sampling design: complete enumeration, simple random sampling,
systematic sampling and stratified random sampling. For carbon inventory, stratified random
sampling generally yields more precise estimates for a fixed cost than the other options.
Stratified random sampling requires stratification, or dividing the populations into non-
overlapping subpopulations. Each stratum (or subpopulation) can be defined by vegetation
type, soil type, or topography. For carbon inventory, strata may be most logically defined by
estimated total carbon pool weight. Since that largely depends on above-ground biomass,
stratification criteria that reflect biomass are generally most appropriate.
Useful tools for defining strata include satellite images, aerial photographs, and maps of
vegetation, soils or topography. These should be combined with ground measurements for
verifying remotely-sensed images (or ground truthing). The key to useful stratification is to
ensure that measurements are more alike within each stratum than in the sample frame as a
whole. A geographic information system (GIS) can automatically determine stratum size and
the size of exclusions or buffer zones (e.g., village sites, non-project areas within the larger
project area, stream buffers, archeological sites). Areas can also be determined manually
using a planimeter or dot grid.
Sample size
The level of precision11 required for a carbon inventory has a direct effect on inventory costs
and, as noted earlier, needs to be carefully chosen by those who will use the inventory report.
Once the level of precision has been decided upon, sample sizes must be determined for each
stratum in the project area and for each carbon pool to be measured. Carbon inventory is more
complicated than traditional forest inventory in that each carbon pool may have a different
variance (amount of variation around the mean). So, while the standard error of the mean for
above-ground biomass may be 20% of the mean, if the same sample sizes are used for each
carbon pool the standard error for soil carbon may be 40%, and that for root biomass may be
80% or more. To simplify sampling design and the understanding of the precision presented
in an inventory, sample sizes for each carbon pool should be determined separately. After
that, the inventory manager can decide how many samples to collect for each pool.
Appendix 2 describes a spreadsheet for inventory decisions that will calculate sample size
using standard formulae based on measured variation for the carbon pool to be sampled. The
appendix describes two alternatives for determining sample size and allocating sample plots
among strata: 1) sample plot allocation based on fixed precision levels and, 2) optimum
allocation of plots among strata given fixed inventory costs.
Permanent plots cannot always be relocated (or reoccupied) for a variety of reasons (e.g., plot
markers are overgrown or are removed by people, plots are burned or records are lost). To
help ensure a minimum number of plots are available for remeasurement, it is prudent to
increase the number of plots above the minimum in the initial sampling design. Increasing the
minimum number of plots for the baseline by 10-20% provides a cushion that helps to

11Precision is the degree of agreement in a series of measurements. Accuracy is the closeness of a
measurement to a true value.
10
ensure that the minimum precision requirements will be met even if there are missing plots in
subsequent inventories.
If biomass or soil carbon data are not available for a site, preliminary samples should be taken
from 10 plots of equal area, perhaps as a training exercise for technicians. These data should
be used to estimate the variance for calculating sample size.
Selection of sample units
The sample units will almost always be fixed-area permanent plots. Permanent plot locations
can be selected either randomly or systematically. If stratified random sampling is used,
sample units for each stratum can still be selected systematically. If little is known about the
population being sampled, random selection of sample units is generally safer than systematic
selection. If plot values are distributed irregularly in a random pattern, then both approaches
are about equally precise. If some parts of the strata have higher carbon content than others,
systematic selection will usually result in greater precision than random selection.
Map preparation
Once the sampling design, sampling sizes, and method for selecting sample units have been
determined, the locations of the permanent sample plots must be marked on a map and/or the
satellite image. Accurate, well-annotated maps are essential for finding permanent sample
plots in the field. Using a GIS or desktop mapping system such as MapInfo can help automate
this process and reduce the possibility for error. In the topographic map in Figure 2, 25
permanent sample plot locations in the eastern half of the map were determined using a
systematic grid; sample units in the western half were selected randomly. Utilities for plot
location allow precise descriptions of plot center locations and help crews to readily find any
given plot with the use of a DGPS receiver.
A major advantage of mapping or GIS software is the ability to produce maps at many
different scales quickly, and therefore customize the scale for each set of users. For example,
funding agencies may be interested in small-scale maps (e.g., 1:50,000) that provide an
overview of the project site. Project managers, on the other hand, may find larger-scale maps
(1:25,000) more useful for viewing details of the project site that help them plan and manage
all project components together. Field crews will generally want the largest-scale maps
(1:10,000) to help them navigate. With a well-designed system for collecting and entering
data, mapping software can automate most of this process.
A software package such as MapInfo, together with digital SPOT satellite images, can help an
inventory manager to establish a systematic grid of sample plots at regular spacings and
produce a list of coordinates for each plot. If desired, the program can list distances and
compass bearings for navigating from plot to plot. However, these would normally be
included in a route entered into the GPS receiver before the field crew begins each days trip
to find and measure plots.
11
12
2.2 Inventory timing
Carbon inventories are likely to be infrequent. Unless they involve continuous monitoring
(and substantially greater expense), inventories cannot account for seasonal fluctuations in the
size of carbon pools. Because inventories measure carbon at just one point in the seasonal
cycle, it is crucial to consider the seasonal timing of the inventory carefully before any other
planning. In most cases, the inventory should take place during the season when field crews
can work most efficiently and safely. This will usually be the part of the dry season with the
most favorable temperatures for strenuous field work. For smaller projects or those that
require fewer sample plots, selection of the season for inventory fieldwork can be more
flexible, since they will require less time in the field than larger projects.
To eliminate seasonality as a source of variation in inventory results, subsequent inventories
must be scheduled for the same season as the first inventory, preferably in the same month.
2.3 Measurement procedures
Permanent sample plots should be remeasured at an interval determined jointly by the
inventory sponsor and manager, based on the desired level of precision. The only exception to
the use of permanent sample plots might be the case of a low-intensity inventory employing a
single assessment of biomass at the end of the tree-crop rotation. In this case, a conventional
inventory using either fixed plots, strips or 3P sampling12 can be used in place of permanent
inventory plots.
If managers periodically inventory biomass (or wood volume) for commercial purposes, a
second alternative to measuring permanent plots may be available. In this case, data collection
for non-timber carbon (e.g., litter, soil carbon, understory vegetation) could be added to the
timber inventory procedures during cruises in the project area.
Locating plots in the field
Two options exist for establishing and marking plot locations on the map and in the field:
Preferred option: Global Positioning System (GPS)
The use of GPS receivers to mark plot locations enables efficient and accurate placement and
reoccupation of plots, particularly in projects with few roads. For natural forest projects or
projects in dense vegetation, sample plot locations should be established using differential
correction (to correct for systematic errors that result from the Defense Departments practice
of selective availability).13 Differential correction ensures that plot centers are located as
accurately as possible. Initial plot location can be done using post-processing differential
correction with receivers capable of accuracy within 5 m. For crews to revisit plots in dense
vegetation, the project will require real-time differential correction capacity (i.e., either carrier
phase DGPS to correct signals transmitted over radio waves, the use of differential beacons, a
satellite correction system such as Omnistar or the use of radio-modems for both base station
and field crews).

12 3P sampling is sampling with probability proportional to prediction.
13 A recent decision by the U.S. Government to remove selective availability will at some future time
obviate the need for differential correction for many forestry applications.
13
Alternative: Compass bearing and distance
Relative bearings can be taken from known landmarks for approach lines, distances and
reference points for each plot. GPS coordinates can also be used as the basis for compass
bearing and distance. This is particularly useful when steep topography or very dense canopy
cover at the plot site prevent reliable GPS readings. This information should be recorded on
the CIDF (Appendix 1, Form E).
Above-ground biomass in project plots
Measure above-ground project biomass using a timber cruise of permanent inventory plots
and biomass tables (Appendix 4). Measure the diameter of all woody vegetation of a
minimum diameter and greater (e.g., > 2 cm) in dbh (diameter at breast height, 1.3 m). Paint a
mark on each stem at dbh to ensure correct measurement for the next inventory. Take
subsamples of smaller diameter woody vegetation and herbaceous plants using small quadrats
or circular plots. Convert individual dbh values for each plot to biomass using single-entry
biomass tables. Where single-entry tables do not provide adequate estimates of biomass, use
double-entry tables based on dbh and height (i.e., length). Appendix 6 describes how to
estimate below-ground biomass.
For tree species for which biomass tables do not exist, the project may need to develop the
biomass tables. It is generally preferable to use biomass tables directly rather than to use
wood-density values to convert stem wood volume tables to biomass, because wood density
varies significantly among trees within a species. Biomass tables can be constructed using a
minimum of 30 well-selected trees or with a mean tree approach (Appendix 4). Appendix 5
describes methods for herbaceous vegetation.
A last alternative to developing biomass tables is to use the general biomass equations found
in Appendix 4. However, this alternative is not generally recommended due to the high
variability between species.
Above-ground vegetation on non-project sites
Quantifying changes in non-project (reference case) vegetation is, in most cases, essential for
quantifying a projects net carbon accumulation. Non-project vegetation will most likely
change
during the project period, so the only way to quantify project benefits reliably is to monitor
vegetation on both project and non-project sites and calculate the difference in carbon stored.
Most non-project (reference case) areas will not be in heavy forest cover and may not require
permanent sample plots. For example, in savannahs that are regularly burned or agricultural
lands subject to tillage, plotless sampling methods can yield acceptable levels of precision at
lower cost. The plotless quarter point method is described in Appendix 4.
Below-ground biomass
Even at moderate levels of precision, measuring root biomass is time consuming and
expensive due to the wide variability in the way that roots are distributed in the soil. For many
projects, it might be best to estimate root biomass using a conservative ratio for shoot:root
14
biomass as the basis for claiming carbon credit. For example, the lowest shoot:root ratio ever
reported for Species X is 5:1. To develop a conservative estimate without measuring roots, an
inventory could calculate root biomass as not less than 10 or 15% of above-ground biomass.
However, for cases in which more accurate estimates of below-ground biomass are
economically feasible, Appendix 6 describes measurements using pit, auger/core sample and
pinboard monolith methods.
Soils
Soils are often large storage pools for carbon, both organic and inorganic. Soil carbon can be
determined effectively using composite samples that represent multiple plots. This helps to
reduce costs of data collection and analysis, yet provides a reasonable estimate of soil
properties.14 The sample size calculator described in Appendix 2 can also be used to
calculate the number of samples required per composite soil sample. Appendix 5 describes
methods for sampling and measuring soil carbon.
Measurement standards and check cruising
Measurement standards define the maximum allowable error in measurements. Table 2
provides suggested allowable limits of error. Measurements with error that exceed these
standards should be rejected as unacceptable.
Table 2. Suggested allowable limits for measurement error
________________________________________________________________________
Measurement Allowable error
________________________________________________________________________
Tie lines
Bearing 2
o
of the true bearing
Distance 2
o
of the true horizontal distance
Permanent plots
Missed or extra trees No error within the plot
Tree species or groups No error
Breast height 5 cm of the true height (1.3 m)
D.B.H. 0.1 cm or 1% whichever is greater
Circular plot radius 1% of horizontal
________________________________________________________________________

14 Peterson, R.G. and L.D. Calvin. 1982. Sampling. In Methods of soil analysis, part 1. Agronomy
Monograph no. 9 (2nd Edition). ASA-SSSA, Madison, Wisconsin.
15
The following general standards are required for carbon inventories using permanent sample
plots:
1. Describe sample locations accurately enough to enable a crew to revisit the sample
plots.
2. Keep adequate records of all data.
3. Specify standards for stratification and sampling design for every inventory, and
adhere to these standards carefully.
4. Take all measurements carefully, using properly adjusted instruments of proven
accuracy. Make every effort to eliminate personal bias by using well-understood
instructions and factual observation in the field.
5. Calculate sampling errors.
6. Inadequate marking, measurement or recording of data, or the sloppy location of
plot centers, may indicate errors or biased location of sample plots. This may cause a
sponsoring agency to reject the inventory.
Check cruising is the verification of field measurements, and involves remeasuring a
percentage of plots to ensure reliable, accurate data of known quality. Check cruising is
necessary for all cruise-based inventories. In general, check cruises should remeasure 1-5% of
all plots within two weeks after initial measurement. The crew performing remeasurement
should not include any members who participated in the initial measurements. Check cruising
should be done with greater intensity (e.g., check cruising of 15% of the plots) for the first
weeks plots. Measurements for the second week might be monitored using a check cruise of
10%, and 5% for the third weeks plots. This provides crews with direct feedback on their
performance and helps to correct procedural errors before they become expensive to correct.
As confidence increases in the crews abilities to collect reliable data of known precision, the
check cruise intensity might be reduced to random remeasurement of 1% of the plots.
16
3. Designing Monitoring Packages
for Specific Land Uses
Project-specific monitoring must meet the specifications of the inventory sponsor and at the
same time use methods and procedures that are appropriate for the site. Different forest types
or land-uses may require different sets of the methods in this guide. Table 3 lists some of the
comparisons and procedures required for the kinds of land-use likely to be included in Joint
Implementation projects.
Packages of methods/procedures should be assembled to meet both the practical and technical
requirements of a site (and the institution(s) conducting the monitoring and verification) and
the cost of using these methods. Section 3.1 discusses general aspects of monitoring design by
land use. Sections 3.2 and 3.3 describe some of the important economic issues relevant to
monitoring designs for specific projects.
3.1 General requirements by land-use
Natural forest preservation
Natural forest preservation projects provide perhaps the greatest amount of fixed carbon in the
early years of a JI project because biomass density is high and deforestation at non-project
sites often releases large portions of stored carbon in biomass due to clearing, wood removals
and burning. The primary carbon comparisons required for these types of projects are
between the areas being preserved and the land-use(s) the forest would be converted to if the
forest were not protected.
Satellite images are important tools in the monitoring of preservation areas because they
provide a clear record of land-use change. Permanent plots can provide reliable trend data on
carbon pool changes and are valuable in protected forest areas. If forest lands outside the
protected area are being converted to agricultural fields, the only measurements required in the
reference case will be soil carbon. If the reference case lands contain woody biomass or
grasslands, then either permanent plots or transect methods are suggested for the measurement
of biomass carbon.
Natural forest management
Monitoring the changes in carbon due to management of natural forests requires paired
comparisons between forests managed with improved regimes and comparable areas using
reference case management. The differences in stored carbon are likely to be smaller in
natural forest management projects than in forest preservation or plantations. This may mean
that for larger sample sizes will be required in forest management projects to attain the same
level of precision as forest preservation or plantation projects.
17
Table 3. Procedures required for reference vs. project case comparisons
Land use Comparison Procedures required
Natural forest
preservation
Reference case:
Adjacent land converted from
natural forest to agriculture or
other uses
Temporary plots for soils, transect methods
for perennial crops
Project case:
Preserved natural forest
Periodic satellite photos of project area,
permanent sample plot measurements
Natural forest
management
Reference case:
Existing management practices
paired permanent sample plot
measurements.
Project case:
Introduced management
practices
Plantations Reference case:
Pre-project vegetation
Transect methods for above-ground woody
biomass with sampling quadrats for herbs,
soils and litter
Project case:
Plantations
permanent sample plot measurements
Agroforestry/farm
forestry
Reference case:
Existing land use systems
Transect methods and temporary plots for
above-ground woody biomass with
sampling quadrats for herbs, soils and litter.
Inter-
Project case:
Improved/expanded
agroforestry or farm forestry
areas
active survey methods are used to solicit
farmer input and to provide information to
farmers about project monitoring.
Layout of the paired plots is also critical to successful monitoring of improved forest
management. Each plot in a pair should be in the same vegetation type with no readily
discernible difference in site quality, stand morphology or population density. Plots should be
located as close to one another as possible, provided adequate buffer or border space. A rule
of thumb for determining border space is two times the height of dominant trees in the stand.
Carbon should be calculated as the sum of the net carbon differences between each pair.
18
Forest plantations
Plantations are often the easiest projects to monitor because when compared to natural forests
they usually have higher road densities, better records and easier access to plots. Permanent
plot methods are appropriate in forest plantations, although the easy access and greater
management intensity introduce a potentially high risk that the plots will be managed
differently from surrounding project areas. To minimize that risk, plots should be marked
inconspicuously using markers that are far from the actual plot centers coupled with a buried
iron pipe or special marking magnets and a specialized metal detector. Plot locations should
be kept confidential to avoid the intentional application of additional inputs to the plots.
Satellite imagery provides a clear picture of plantation size and location. When analyzed with
mapping software it also provides a ready means of calculating strata and total project areas.
Agroforestry and farm forestry
Spatial variability and farm to farm differences in management are the greatest challenges for
monitoring agroforestry or farm forestry areas. Most projects of this type will include farms
that are widely dispersed and managed in different ways. Many of the methods used to
measure carbon sequestration in natural forests can be directly applied to agroforestry
plantings, but there are some important differences. These include:
agroforestry plantings require intensive labor inputs, and are typically small in size.
agroforestry plantings are often widely scattered over the landscape. Broad expanses
of non-project vegetation may separate individual plantings.
trees in agroforestry plantations are often widely spaced to provide light for associated
crops. As a result, the tree canopy is discontinuous and may be highly variable.
in some agroforestry systems, trees are arranged in regularly spaced rows. This could
introduce bias into systematic sampling schemes arranged in linear grid-like patterns.
Agroforestry plantings are usually established and maintained by small landholders.
Thus, any measurement of an agroforestry plantation necessarily involves professional
interaction with farmers that may not occur in other types of land-use projects.
3.2 Setting economic limits
Some potential JI projects do not fix enough carbon for a monitoring effort to be economically
worthwhile. Although this should be evaluated prior to project approval, it may be necessary
to make preliminary estimates of monitoring costs at the proposal development stage. When
designing a monitoring system, the cost of measuring each component should be estimated
and compared to the value of carbon. During the JI pilot phase there is no actual trading value
for carbon, so the relative value of each measurement should be calculated and packaged in a
way that fits the project budget.
The monitoring package design should depend on how much carbon will be fixed per unit
area. For example, projects that fix less than 2 or 3 t C per hectare per year can not likely be
monitored in a cost-effective way because the costs of measuring these quantities are nearly
19
the same as the cost of monitoring 10 or 15 t C per hectare per year. When preparing a
monitoring plan and budget, these economies of scale are important.
Project area is also an important factor in determining both the economic feasibility of
monitoring and the cost per ton of carbon. In general, because fixed costs are a large part of
the total monitoring cost, the larger the project area, the lower the unit costs for monitoring. It
is likely that projects smaller than 1,000 ha will be very difficult to monitor in a cost-effective
way with reasonable precision.
20
4. Tools
Because a carbon inventory might be estimating carbon worth millions of dollars, the tools for
these inventories need to be accurate, rugged and durable to withstand the rigors of field use
under adverse conditions. They should also contribute to efficient planning, data collection,
analysis, and reporting. This section describes equipment for field work and software for
modelling. Models can be useful tools for estimating changes in carbon pools for periods
between inventories; but for traded commodities, models are not adequate substitutes for
measurements.
4.1 Equipment
In order to perform an inventory accurately, reliably and at minimum cost, an inventory team
must have good-quality equipment. Anything less can result in higher labor costs, greater
safety risks and unreliable carbon estimates. The following list of field tools and equipment
continues to evolve as methods are refined and new equipment becomes available. To some
extent, equipment needs will vary with inventory objectives, available labor and skills, terrain,
and vegetation or soil type. Figures 3 to 5 show examples of some of the equipment needed.
The following equipment and supplies are recommended for field crews:
Equipment
compass/clinometer combination for navigation, plotting on the map, bearing and
slope measurements
diameter tape for measuring dbh
loggers tape for measuring dbh and as a back-up for plot radius measurements
Hagloff distance measure, tripod and extra threaded tripod adapter for measuring
distance to trees for diameters, and for measuring fixed plot boundaries
calculator for calculating height, diameter
cruisers vest for carrying cruising equipment
entrenching tool, folding shovel or soil corer for taking soil samples
precision spring scales (e.g., Pesola 1kg + 300g) + weighing bags, e.g., Tyvek 10 x
17" bags (25.4 x 43.2 cm)
sampling frames (2), round or square, hinged
pruning saw and shears (e.g., Felco 60 and Felco 8)
sheet holder
two-way radio and extra batteries
GPS (2) with differential correction capability and remote antenna mounted on fixed
frame backpack (one as base station and one as remote), differential correction
software, extra battery pack and charger
notebook computer for database use and differential correction, and for generating
maps
21
Supplies
laminated maps or photos, with plot locations and coordinates
pencils, marking pens, map scales
ribbon (flagging) and painted high-grade PVC pipe for marking plot centers that are
conspicuously marked
plot cards, field aids and instructions
rain gear
safety equipment such as a first aid kit, hard hat, space blanket, waterproof matches,
candle, insect repellent
flashlight
50 cm x 50 cm piece of 5-mm mesh screen and small plastic tarp for screening and
mixing soil samples
sampling bags for soil, vegetation and litter, e.g., Tyvek 5 x 7" bags (12.7 x 17.8 cm)
Additional equipment for non-destructive biomass table measurements:
Spiegel relaskop - metric scale for measuring tree diameters, heights and slope
Jacob staff or monopod and ball joint adapter for use with relaskop
bark gauge
compact binoculars for use with relaskop
Additional equipment for verification measurements:
hand-held data collection or pen-based computer
Additional requirements for real-time DGPS use in permanent plot reoccupation:
radio modem transceivers (2)
GPS base station capable of transmission
4.2 Models for interpolation
The size of carbon pools can be estimated for periods between inventories by using prediction
models. Changes in soil and biomass carbon can be modeled using software packages such as
Soil Changes Under Agroforestry (SCUAF), CENTURY and LUCS, using baseline survey
data and estimates of biomass growth. SCUAF Version 2.0 is recommended for use with the
methods described in this guide, mainly for the following reasons:
Ease of use. SCUAF is menu-driven and relatively user-friendly.
Relevance. SCUAF was designed for use in agroforestry and incorporates
both soil and biomass predictions for carbon and nitrogen.
Presence of a default data set (for basic-level monitoring). Data on a number
of variables that are expensive to collect (e.g., root biomass, feedback factors,
erosion rates) are provided in a default data set that has been carefully selected
from the literature.
22
Cost and availability. SCUAF costs less than US$50 and is readily available
from the International Centre for Research in Agroforestry.15
Documentation. The SCUAF manual is well written and provides good
information on the softwares theoretical basis and operations.
Appendix 3 describes the data collection requirements for SCUAF.
The prospect of saving on measuring costs by using computer models to predict future carbon
storage may be tempting. However, the costs of accurate, verified modelling are likely to be at
least as high as the costs of actual measurement, and perhaps greater. Modelling should only
be used for interpolation, i.e., when investors require an estimate of carbon storage at a point
in time between two actual measurements.

15 International Centre for Research in Agroforestry, United Nations Avenue, Gigiri, P.O. Box 30677,
Nairobi, Kenya.
23
5.0 Reporting and Verification
The format and frequency of reports will depend in part on the inventory design, resources and
the reporting requirements of the sponsoring agency. Use the reporting formats described in
this section to present the results of carbon monitoring. Reports intended for use outside the
projects technical staff should always include a summary explaining the consequences of the
findings.
5.1 Reporting
The way in which a project reports carbon credits will likely be determined by governmental
regulations or intergovernmental agreements. Until such guidelines are in place, the following
two types of reporting might be considered.
1. Report mean values for carbon stored along with confidence limits (at p=0.05). The
formula for confidence interval calcuations is:
CI = X ts
X
where : t = a two-sided t value for a probability level of 0.05
s
X
= the standard error of the mean from the carbon inventory
2. Report the Reliable Minimum Estimate (RME) as a conservative measure of the
minimum quantity expected to be present with its probability.16 The formula for this
calculation is:
RME = X ts
X
where : t = a one-sided t value for a probability level of 0.05 (i.e., use p=0.10
in a two-tailed t table)
s
X
= the standard error of the mean from the carbon inventory
For most current uses, reporting mean values with confidence intervals is probably most
appropriate given the need for maximum incentives to potential investors in carbon offset
projects.
5.2 BaselineThis report provides an estimate of organic carbon as it is distributed in
vegetation and soils before start of the project. It is derived from data summarized in the
Carbon Inventory Data Form (Appendix 1).
Project description

16 For more details see Dawkins, H.C. 1957. Some results of stratified random sampling of tropical high
forest. Seventh British Commonwealth Forestry Conf. Item 7 (iii).
24
Site name:
Contact person:
Project sponsor(s):
Project manager:
Local name of project site:
Address, State, Country:
Latitude:
Longitude:
Elevation (m):
Project species:
Accuracy level specifications
Baseline carbon distribution
Carbon pool
Area
(ha)
Mean carbon
density
(Mg ha
-1
)
Total carbon
(Mg)
Confidence
interval (Mg)
Reference case
Above-ground
Below-ground
Forest floor
Soil to depth of 30 cm
Total - reference case
NA
Project case
Above-ground
Below-ground
Forest floor
Total - project case
ha
NET CARBON STORED
THROUGH PROJECT
ACTIVITY
5.3 Reporting carbon changes
This format is for use in reporting changes in carbon stored due to project activities.
Dates:
Date of previous measurement:
Primary person responsible for monitoring:
25
Description
Site name:
Contact person:
Sponsors:
Project manager:
Local name of site:
Latitude:
Longitude:
Elevation (m):
Primary species:
Accuracy level specifications
Site history since carbon statement or last inventory
Describe any significant changes in management, pest and disease problems, harvesting or
other mortality.
26
Carbon distribution17
Carbon pool Area
(ha)
Mean carbon
density
(Mg ha
-1
)
Total carbon
(Mg)
Confidence interval
(Mg)
Reference case
Above-ground
Below-ground
Forest floor
Total - reference case
NA
Project case
Above-ground
Below-ground
Forest floor
Total - project case
ha
NET CARBON STORED
THROUGH PROJECT
ACTIVITY
5.4 Verifying carbon monitoring estimates
Verification of carbon offset projects by a third party is similar to an accounting audit
performed by an objective party. For greatest efficiency and the most useful results, the
regular monitoring team and the auditing organization should agree on procedures and
methods before start of the project.
A verification audit of carbon monitoring is a form of quality assurance that is presently
required by the U.S. Initiative on Joint Implementation (USIJI). It is also likely to be required
in future carbon offset land-use programs. Just as periodic audits are required for companies
involved in other types of trade, a system of verification will be necessary in order to avoid
needless litigation over project benefits and credits.
Agencies aiming to verify a forestry projects carbon storage estimates might follow the
general procedures used by auditing firms in accounting. These include:
1. Prior agreement on carbon monitoring methods at the outset. If the verifying agency and
the projects carbon monitoring team agree on a system of methods for measuring carbon
before the project begins, then the process can be evaluated efficiently, with little danger of
problems that would call monitoring estimates into question.

17Carbon estimates are based on samples taken from permanent, fixed-area plots in sites prior to site
preparation for establishment.
27
2. Review of all monitoring records, including field data collection sheets,
spreadsheet/database files, computer model outputs, maps, remote-sensing data, plans,
analyses, and reports.
3. Inspection and calibration of measurement and analytical tools used by the monitoring
team.
4. Reoccupation and measurement of a random sample of the permanent plots used in the
inventory.
5. If satellite imagery was not used to calculate project area for previous inventories, obtain
and process images to verify project area.
28
Appendix 1: Carbon Inventory Data Form (CIDF)
Form A - Level of precision specifications
This form is designed to record instructions from the inventory sponsor regarding the desired
levels of precision. (NOTE: Each carbon pool will likely have a unique variance and will
require a unique sampling intensity to achieve a constant overall level of precision. For
example, root biomass is likely to be more variable than above-ground biomass; foliage
biomass is usually more variable than stemwood biomass.)
Form of decision from inventory sponsors:
____ General level of precision ____ Specific confidence limits (%)
____ Optimum precision for fixed-cost____ Cost based on precision
If a general level of precision is specified, record below the detailed specifications for
modelling vs. field data collection, cost limits from sponsors, and overall desire for precision
(e.g. basic, moderate, high):
Percentage of plots to be established in excess of the calculated minimum
requirement:______%
Other specifications requested by inventory sponsor(s):
29
Form B - Project site description
A complete site description provides enough information to identify and locate the site and to
allow some explanation of performance. Most data required for this form should be available
from the project manager.
Site name:
Contact person:
Local name of site:
Elevation range (m):
Ecological zone or general site type:
Most common slope class (flat or gentle = 0-5
o
; intermediate = 5-10o; steep = 11-45o; very
steep >45o):
Mean annual rainfall (mm):
Rainfall regime (summer, winter, bimodal, uniform):
Maximum length of dry season (months <50mm):
Mean annual temperature (oC):
Surface soil texture (sand, loam, clay):
Sub-soil texture (sand, loam, clay):
Soil depth to impermeable layer (<25 cm, 25-50 cm, 50-100 cm, or >100 cm):
Surface soil pH (A horizon):
Sub-soil pH (B horizon):
Map with at least 3 latitude/longitude points18:

18 Points identified from: _____ local maps _____ known survey points
_____ differentially corrected GPS coordinates _____ uncorrected GPS coordinates
30
Form C - Sampling design
This form should be used in conjunction with the explanation for calculating sample size
(Appendix 2).
Sampling design: Stratified systematic sampling with random start
Basis of stratification:
Source of variance estimates:
Variable used for estimate:
Number of samples used for estimate of sample plot requirements:
Acceptable error (% of treatment mean):
Stratum
number19
Vegetation
type
Area (ha) Mean biomass
(t ha-1)
Coefficient of
variation (%)
Number of
sample plots
required20
TOTAL
Append a map of the project area with sample plot locations marked and and coordinates for
each permanent sample plot location.

19 Coding system: First letter = component (i.e., A or B), Second letter = Treatment (i.e., P =
project case; R = reference case), Number = stratum number from vegetation map
20 As calculated using the Winrock inventory sample size calculator
31
Form D - Satellite images
Satellite imagery, taken on an annual basis, can document the project area in an unbiased
manner. Panchromatic SPOT imagery is recommended for this use due to the availability of
high spatial resolution (10 m) in a 7.5
or 15view. (However, this is a rapidly changing

technology with new products and services constantly emerging.21) SPOT offers the
additional advantages of (1) the ability to program the satellite cameras to cover a specific area
at a specific time and (2) the convenience of images that do not require correction for
topographic displacement in a commonly used map format.
Project managers, sponsors and field crews should receive copies of ortho-corrected,
panchromatic prints. Digital image processing allows the input of geo-referenced images in a
GIS such as MapInfo or ARC/Info.
Land area (including total area and mortality due to fire, clearing, insect and disease pests)
should be determined annually from new satellite photos. In cases where cloud cover
precludes the use of satellite photos, specify plans for either aerial photography or ground-
based verification of the area. When ordering satellite images, the following parameters are
generally needed:
Parameters or qualifiers Value
Spectral mode
Maximum acceptable cloud cover
Scene date window
Site location
Angle range
Imagery must be ordered at least two weeks before the start of the desired viewing window.
Images take approximately four weeks to process after viewing.

21If satellite images are desired for other uses, such as species identification or monitoring of stand health,
options include the more expensive Landsat Thematic Mapper (30 m resolution reflected) and SPOT
multispectral images (20 m resolution). For more information, see Monitoring vegetation change using
satellite data, by B.N. Rock, D.L. Skole and B.J. Choudhury, inVegetation dynamics and global change,
eds. A.M. Solomon and H.H. Shugart (New York: Chapman and Hall, 1993).
32
Form E - Permanent plot locations
It is essential to mark clearly and record the locations of permanent sample plots to ensure
efficient reoccupation of the plots for later measurements. Plot center markers painted with
florescent paint and large quantities of bright colored flagging are recommended for marking
plot centers. The following form can be used to record planned and actual plot locations.
When using a GPS receiver, the actual position will usually differ from the planned position
due to the Defense Departments policy of selective availability (which introduces dithered
satellite signals containing intentional errors) and/or difficult terrain at the planned location. In
any case, it is important to record the actual position, using either the best average fix from the
GPS receiver (if differential correction is not used) or the corrected position after the
permanent plots are established and the corrections made.
Please note: this form must be accompanied by a map indicating permanent plot locations.
33
Plot Locations
Cruiser ________________ Date ___/___/___
Project ________________ Country________________
Plot location method (circle one): GPS Compass
Strat
a
Plot Plot
Landmark
or known
Bearing
from
Planned position Actual position*
Elev.
No. No. size point landmark Latitude Longitude Latitude Longitude (m) Comments
*If different from planned
34
Form F - Biomass measurements
Biomass accumulation should be monitored periodically by measuring vegetation at project
and non-project sites. The plot card can be used to record measurements in permanent sample
plots. The quarter-point data collection sheet can be used for data collected using plotless
methods.
35
PLOT CARD
Carbon Inventory - Winrock International
Strata number___________ Plot number:_________ Waypoint number ______
Vegetation type:_______________ Crew chief ________________ Date __/__/___
Tree
no.
Spp.
Code
DBH
(1.3m)
Tree
no.
Spp.
Code
DBH
(1.3m)
Quadrat sampling
Number of
quadrats
Herb weight
(g)
Litter weight
(g)
1 21
2 22 Sub-samples for moisture content
3 23 Herbs Litter
4 24 Sample no. Weight (g) Sample no. Weight (g)
5 25 H L
6 26 Sub-samples
7 27 Tree spp. Sample no. Soil
8 28 W Sample number
9 29 W S
10 30 W
11 31 W Sampling notes
12 32 W
13 33 W
14 34 Notes: Record the diameter of dead trees here, both
standing and fallen.
15 35
16 36
17 37
18 38
19 39
20 40
Next waypoint number:_____ Bearing to next waypoint:______ Distance:______m
Landmarks?
36
Quarter Point Method Data Collection Form
Line number: Litter samples Vegetation
Point
number
Quarter
number
Species or
Species
group code
DBH
(1.3m)
Diameter
@30cm
Height
(m)
Distance
(m)
Soil
sample
Number
Number
Sample
weight
Number
Sample
weight
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
37
Form G - Anticipated disposition of biomass
To gauge the effectiveness of carbon sequestration projects, it is essential to have an idea of the
intended fate or end-use of the biomass grown.
Check one or more option. If more than one, then indicate approximate percentage of disposition for
each category.
____ Durable timber products (e.g., furniture, construction)
____ Pulp
____ Fuelwood (firewood or charcoal)
____ Foliage uses
____ Other (specify)
Non-project land use
____ Grazing
____ Periodic burning (specify approximate frequency)
____ Crops
____ Other (specify)
38
Form H - Laboratory methods
Soil and biomass carbon testing must be conducted by an established, reputable laboratory capable of
periodic analyses throughout the project period. Total carbon is most often measured by either the dry
combustion or wet combustion methods described by Nelson and Sommers.22 For soils that have
carbonate minerals present, corrections need to be made for inorganic carbon using one of the
methods described by Nelson and Sommers. Total nitrogen will be analyzed using the regular
Kjeldahl distillation method, as described in Bremner and Mulvaney.23
Laboratory: ____________________________________
Address: _______________________________________
_______________________________________
_______________________________________
Telephone: ______________________ Fax number: __________________________
Contact person for analysis: ________________________________________
Carbon analysis methods (check one):
____ dry combustion in a resistance furnace
____ dry combustion in a induction furnace
____ dry combustion using automated methods
____ wet combustion using a combustion train
____ wet combustion using a Van Slyke-Neil apparatus
____ other (specify)
Cost per sample _____
Total nitrogen analysis methods (check one):
____ regular Kjeldahl distillation method
____ modified Kjeldahl distillation method (describe)
____ other (specify)
Cost per sample _____
Notes:

22 Nelson, D.W. and L.E. Sommers. 1982. Total carbon, organic carbon and organic matter. In Methods of soil
analysis, part 2. Agronomy Monograph no. 9 (2nd Edition). Madison, Wisconsin: ASA-SSSA.
23 Bremner, J.M. and C.S. Mulvaney. 1982. Nitrogen - total. In Methods of soil analysis, part 2. Agronomy
Monograph no. 9 (2nd Edition). Madison, Wisconsin: ASA-SSSA.
39
Form I - Inventory costs
The amount of time and money required to collect and analyze baseline and annual carbon data should
be documented in this form. The most important use of this data will be for estimating the cost of
sampling in each stratum to determine the optimum allocation of sample plots.
Cost per Total
Man-day days Cost
Planning
Supervision
Materials
Transportation
Other
Training
Instructors
Materials
Other
Quarter point method survey
Supervision
Labor
Transportation
Other
Permanent plot establishment
Supervision
Labor
Transportation
Other
Permanent plot monitoring
Supervision
Labor
Transportation
Other
Analysis, interpretation and reporting
Personnel
Laboratory analyses
Materials
Other
Total inventory costs
40
Appendix 2: Calculating Sample Size
The spreadsheet that follows includes sample data from non-project stands as an example of how the
calculator can be used. Two approaches are used for determining sample size and sample plot
allocations among strata: 1. optimum plot allocation based on fixed precision levels and, 2. optimum
allocation of plots among strata given fixed inventory costs.24
1. Optimum plot allocation based on fixed precision levels
Use the following formula to calculate the number of sample units required to obtain a desired
standard of precision:
n =
t
A

2
WhSh Ch
h=1
L

WhSh / Ch
h=1
L

where n = sample size (i.e., total number of sample plots required)
t = tabular value of Students t
h = stratum number
L = the number of strata
W
h
= N
h
/N
Nh = number of sample units in stratum h
N = total number of sample units
S = stratum standard deviation
A = allowable error expressed in units of the mean
Ch = the cost of selecting a sample plot in stratum h
Allocation of these sample plots among strata is calculated as: n
h
= np
h
where: n
h
= number of sample plots for stratum h
n = total number of sample plots
and
Ph = WhSh / Ch ( )/ WhSh / Ch
h=1
L

24 For a more detailed description of this approach see Forestry handbook (2nd edition), ed. K.F. Wenger (New
York: John Wiley and Sons, 1984) and Sampling techniques (3rd edition), by W.G. Cochran (New York: John
Wiley and Sons, 1977).
41
2. Optimum allocation of plots with fixed costs
If costs for sampling are fixed before sample size or plot allocations are determined, plot allocations
can be assigned to minimize the inventory cost. The second portion of the spreadsheet performs these
calculations using the following formula:
Vc = n ChPh
h =1
L
where: V
c
= variable costs for sampling
The user inputs include site name, measured parameter (e.g. woody biomass), number of preliminary
sample plots used to calculate plot-to-plot variances, desired level of probability (p), allowable error
(in % of the mean), sampling budget (if variable costs for sampling are set before sample size is
calculated), sample plot size, costs for the establishment and measurement of sample plots, and the
proposed number of times measurements will be made after plot establishment and baseline
measurement. The minimum value for planned number of samplings is one (final measurement).
When using the formula, the following conventions must be used:
If the mean and standard deviation values are not calculated from sample plot data in the
spreadsheet, they must also be entered for each stratum.
Cost values must be entered into the column labeled "Cost per plot" near the bottom of the
sheet. Calculate cost per plot using the input values in the box at the top of the sheet, but enter
these manually in the cost per plot cells to allow separate cost values for each stratum.
44
44
Appendix 3: Inputs for Carbon Modelling with SCUAF
The data required for carbon modelling using the Soil Changes Under Agroforestry (SCUAF)
software are listed in the table below. Data on soil organic carbon, tree growth rates, biomass
partitioning, and site descriptions must be entered before running models. All inputs, including
defaults, must be listed in the form. If data other than the default values are used, the source of this
data should be listed in the "Source" column.
The following are inputs required to use SCUAF:25
Input Value Source
1 CYCLE
Cycle selected for modelling
If carbon cycle only is selected, it is not necessary to input
data on nitrogen.
1 Carbon
Cycle
2 Carbon
and Nitrogen
Cycles
2 DOCUMENTATION
File name
Title
Source
Location
Date
Notes

25The source references indicate values that are different from the SCUAF default data set.
45
3 PHYSICAL ENVIRONMENT
Climate:
Soil texture:
Drainage:
Soil reaction:
Slope class:
1 Lowland
humid
2 Lowland
subhumid
3 Lowland
semi-arid
4 Highland
humid
5 Highland
subhumid
6 Highland
semi-arid
1 Medium
textured
2 Sandy
3 Clayey
1 Free
2 Imperfect
3 Poor
1 Strongly
acid
2 Acid
3 Neutral
4 Alkaline
1 Flat
2 Gentle
3 Moderate
4 Steep
4 AGROFORESTRY SYSTEM
Length (years)
Fraction of land under trees
Fraction of land under crop
Is it a cut year? (Yes/No)
What fraction of tree is N-fixing?
What fraction of crop is N-fixing?
Period
1 2 3 4 5 6
46
5 INITIAL SOIL CONDITIONS
DEPTH
Topsoil depth (cm)
Soil depth considered (cm)
Total depth of soil (cm)
CARBON
Initial Carbon, Topsoil (percent)
Initial Carbon, Subsoil (percent)
Bulk density, Topsoil (g/cc)
Bulk density, Subsoil (g/cc)
Initial soil Carbon (kg/ha)
NITROGEN
Initial Nitrogen, Topsoil (percent)
Initial soil Nitrogen (kg/ha)
6 EROSION
Soil Erosion (kg/ha/yr) = Climate Factor * Soil Erodibility Factor
* Slope Factor * Cover Factor * 1000
Enter best estimate for each factor:
Climate factor
Soil erodibility factor
Slope factor
Cover factor under tree
Cover factor under crop
Soil erosion under tree (kg/ha/yr)
Soil erosion under crop (kg/ha/yr)
Tree proportionality factor
Measured soil erosion in Year 1 (kg/ha/yr)
Carbon enrichment factor
Nitrogen enrichment factor
47
7 INITIAL PLANT GROWTH
Tree, Net Primary Production, above-ground (kg DM/ha/yr)
Crop, Net Primary Production, above-ground (kg DM/ha/yr)
Roots as a fraction of above-ground NPP, tree
Roots as a fraction of above-ground NPP, crop
NPP in parts of tree (kg/ha/yr):
Leaf
Fruit
Wood
Root
NPP in parts of crop (kg/ha/yr):
Leaf
Fruit
Wood
Root
Fractions of Tree retained as growth annually:
Leaf
Fruit
Wood
Root
Fractions of Crop retained as growth annually:
Leaf
Fruit
Wood
Root
Proportion of tree roots that are coarse roots
Proportion of crop roots that are coarse roots
Is any part of tree or crop retained as growth in cut year (Yes/No)
If Yes:
Fractions of Tree retained as growth during cut year:
Leaf
Fruit
Wood
Root
Fractions of Crop retained as growth during cut year:
Leaf
Fruit
Wood
Root
Fraction of roots growing below soil depth considered:
Tree roots
Crop roots
48
CARBON
Carbon Fraction in dry mass, Tree
Carbon Fraction in dry mass, Crop
NITROGEN
Nitrogen % in:
or
Carbon:Nitrogen ratio of:
Tree Leaf
Tree Fruit
Tree Wood
Tree Root
Crop Leaf
Crop Fruit
Crop Wood
Tree Root or Crop Root
8 ADDITIONS
Organic (kg DM/ha/yr)
Carbon fraction in organic additions
Nitrogen percent in organic additions
or Carbon:Nitrogen ratio in organic additions
Fertilizer (kg/ha/yr)
Nitrogen fraction in fertilizer
Period
1 2 3 4 5 6
Period
1 2 3 4 5 6
49
9 REMOVALS
A: HARVEST
Fraction of Tree harvested annually:
Leaf
Fruit
Wood
Root
Fraction of Crop harvested annually:
Leaf
Fruit
Wood
Root
Additional fraction of Tree harvested in cut year:
Leaf
Fruit
Wood
Root
Additional fraction of Crop harvested in cut year:
Leaf
Fruit
Wood
Root
B: OTHER LOSSES FROM SYSTEM
Are there any losses of plant material from the system
other than harvest (e.g. burning)? (Yes/No) If Yes:
Fraction of Tree lost annually:
Leaf
Fruit
Wood
Root
Fraction of Crop lost annually:
Leaf
Fruit
Wood
Root
Additional fraction of Tree lost in cut year:
Leaf
Fruit
Wood
Root
Additional fraction of Crop lost in cut year:
Leaf
Fruit
Wood
Root
50
10 SOIL PROCESSES
CONVERSION LOSSES (Litter to Humus)
Fraction of above-ground parts lost through oxidation
Fraction of roots lost through oxidation
Fraction of organic additions lost through oxidation
Fraction of coarse tree roots decaying at least 1 year later
Fraction of coarse crop roots decaying at least 1 year later
Fraction of remaining coarse tree roots decaying 2 years later
Fraction of remaining coarse crop roots decaying 2 years later
51
HUMUS DECOMPOSITION CONSTANTS
Number of humus fractions considered (1 or 2)
LABILE HUMUS
K for the tree
K for the crop
For 2-fraction humus only:
STABLE HUMUS
K for the tree
K for the crop
Fraction of humified litter becoming labile humus
Fraction of labile humus transformed annually to stable
NITROGEN CYCLE
NITROGEN GAINS
Symbiotic Fixation per unit area of N-fixing Tree (kg/ha/yr)
Symbiotic Fixation per unit of N- fixing Crop (kg/ha/yr)
Fraction of symbiotic fixed N entering soil humus
Non-Symbiotic Fixation (kg/ha/yr)
Throughfall and stemflow (kg/ha/yr)
NITROGEN LOSSES
A. Mineral N of organic origin:
Fraction of mineral N leached under tree
Fraction of mineral N leached under crop
Fraction of mineral N lost
- by gaseous losses (dentrification + volatilization)
- by fixation onto clay minerals
(net)
B. Fertilizer N:
Fraction of fertilizer N leached under tree
Fraction of fertilizer N leached under crop
Fraction of fertilizer N lost:
- by gaseous losses
(denitrification+volatilization)
- by fixation onto clay minerations (ncl)
52
11 SOIL/PLANT FEEDBACK FACTORS
CARBON
Rise or fall in soil carbon, relative to initial state, of 1
percent causes increase or decrease in rate of plant growth by x
percent:
For Tree
For Crop
NITROGEN
Rise or fall in soil nitrogen, relative to initial state, of 1 percent
causes increase or decrease in rate of plant growth by x percent:
For Tree
For Crop
SOIL DEPTH
Rise or fall in soil depth, relative to initial state, of 1 percent
causes increase or decrease in rate of plant growth by x percent:
For Tree
For Crop
NOTES
53
Appendix 4: Measuring Woody Biomass
Woody biomass will nearly always be the largest and most easily manipulated carbon pool in carbon
storage forestry projects. The inventory approaches for woody biomass are:
1. Measure the woody biomass of trees larger than a minimum diameter (e.g. >5 cm dbh) in
project and non-project areas using timber cruises of permanent, continuous inventory sample
plots. At the same time, sample the herbaceous biomass, standing litter crop and soil carbon
vegetation, using the methods described in Appendix 5.
2. If permanent plots are not desirable or practical (due to frequent wildfire or grazing, for
example), use the plotless quarter point vegetation survey to evaluate biomass and carbon
content of vegetation on non-project sites. In plantation projects, the quarter-point method
(which uses the distance between a systematic sampling point and the nearest tree or shrub)
can be used to monitor lands left under natural vegetation. For preservation projects, this
method can be used to monitor changes in lands converted to agriculture or other land uses
with low tree population densities.
A. Timber cruising
Use permanent, inventory plots to measure timber on project and non-project sites, following a
stratified random sampling design. In general, most carbon sequestered by project activities will occur
in the largest diameter classes, so the timber cruise part of the inventory requires particular care.
Size and shape of fixed plots
Circular plots established using a well-identified plot center and a digital distance measure are
recommended (Table 4). Optimal plot size can also be calculated using the following formula26:
Size = P
1
t
2
m
where: P
1
= size of plot used in preliminary sample to assess time and variation in any unit of area; t =
average travel time between neighboring plots in minutes; m=average plot measurement time for plot
size P
1
in minutes
To calculate optimal plot size, measure a minimum of three plots of size P1. Plots should be separated
by the same distances anticipated in an actual inventory. Travel speed between plots and the plot
measurement time (m) should be recorded. Calculate t by dividing the distance between neighboring
plots by the travel speed or measure and average travel time between plots. Calculate P using the
formula above.

26 Zeide, B. 1980. Plot size optimization. Forest Science 26:251-257.
54
Table 4. Plot radii for carbon inventory plots
Plot size (m
2
) Plot radius (m) Area per tree (m
2
tree-1) Application
100 5.64 0 - 15 Very dense vegetation, stands
with large number of small
diameter stems, uniform
distribution of larger stems
250 8.92 15 - 40 Moderately dense woody
vegetation
500 12.62 40 - 70 Moderately sparse woody
vegetation
666.7 14.56 70 - 100 Sparse woody vegetation
1,000
or use quarter
point method
17.84 > 100 Very sparse woody vegetation
Relating diameter to biomass
Biomass tables or equations are needed to relate dbh and the number of plants per ha to total biomass.
Biomass tables should therefore be sought for important species of native vegetation. Where these
tables are not available, there are three alternatives:
1. Develop biomass tables for each important tree species using the method described in
section C of this appendix. This is the most precise (and most costly) approach.
2. Develop biomass tables for groups of tree or shrub species. The most useful groupings may
be by morphology class (e.g., single-stemmed trees, multiple-stemmed trees, shrubs).
3. Use one of the general biomass equations found in section C. This is the least precise
approach, but also the least expensive. Given the wide range of species included in these
equations, they should not be used except where the alternatives above are not possible.
B. Non-project woody vegetation
This method should be used to measure natural vegetation prior to project establishment at the same
interval set for the measurement of the permanent inventory plots in the project. To save time in
laying out plots in measuring woody savannah vegetation, use the quarter point method,27 which
uses the distance between a systematic sampling point and the nearest tree or shrub.
The steps for this method are:
1. Establish a series of parallel sample lines 100 m apart. Locate sample points every 10 m
along each line.

27 For a more detailed description of this approach see Methods of sampling lesser vegetation, pp. 58-59 in Forestry
handbook (2nd edition), ed. K.F. Wenger (New York: John Wiley and Sons, 1984).
55
2. At each sample point, divide the immediate area into quarters using the sample line plus a
second line that crosses the sample point perpendicularly to the sample line. Figure 3
demonstrates placement of line and sample point, along with quarter numbers for each sample
point.
3. Collect species and diameter data using the data collection form. For shrubs and small trees
with low branching, measure diameter at 30 cm above the ground; for trees with better tree
form, use dbh at 1.3 m. Also record the distance from each tree or shrub to the sample point.
A minimum of 100 distance measurements are required per stratum.
4. Calculate average distances as follows:
Average the four distances at each point, then average the distances for the entire
sample area. Use this number to calculate the mean area per tree as: M = d
2
, where M
= mean area per tree in m2 and d = average distance over the entire sample area.
For the total sample area, calculate density using the formula: D = 10,000 / M,
where D = trees per hectare.28
Construct a stand table, with appropriate size diameter classes, and estimate biomass
based on a biomass table.

28 If only a few species predominate, then accuracy is probably increased by determining biomass and carbon
content by individual species. The method outlined in this section presumes relatively high species diversity and
relatively low return to the additional investment needed to estimate non-project biomass by species.
56
C. Developing biomass tables29Tree biomass weight tables show the average weight of
individual trees for one or more dimensions, usually stem diameter alone (for local tables) or stem
diameter along with height or length. The tables employ data obtained in destructive sampling.
Through regression analysis, wood or foliage weight can be related to dbh, diameter at 0.3 m, and
height. This appendix describes the process for developing standard biomass weight tables.
Developing local weight tables is not recommended in view of their limited application and the great
deal of effort they require, as well as the duplication of effort they involve.
Planning a weight table
Critical decisions must be made before beginning to develop a weight table, including:
Defining the range of tree or shrub dimensions to be covered by the table.
Defining the type(s) of biomass to be included in the table(s); this may include stem wood,
stem and branch wood, or foliage; a table will need to be developed for each biomass type.
Determining what measurements to use in relating biomass to a practical field measurement
(e.g., d, dbh, or h).
Defining size classes throughout the range of tree sizes.
Field sampling
A key issue is the number of trees necessary to develop a weight table for a given species. Estimates
from many recent biomass studies suggest that 30-100 trees are enough for a regional table using
stratified sampling of the population. Sufficient evidence supports the case that if sample trees are
selected in equal or near-equal numbers for each size class, 30 trees for an individual tree biomass
table are adequate. At least 30 well-selected trees should be used per species for individual tree
biomass tables, unless the tables are to be used only for a specific site. For such purposes, as few as
12 trees may be adequate.
For calculating foliar nutrient content, more samples may be needed for foliage and branches than for
stem wood. If scarce funds prevent the measurement of more than 30 sample trees for foliage,
consider reducing the number of samples for stem wood in order to make resources available for
foliage and branch sampling.
Partitioning and weighing
Each individual tree or shrub should be harvested; measured for diameter, length, and height30; and
divided into the major components defined during the planning stage. Length measurements are
recommended, but if vertical height is to be included as an independent variable, the measurement
must be taken before the tree is harvested. Individual components (stem, branches, leaves, etc.)
should be divided into several size classes for convenient handling and sub-sampling. Sub-samples
should be taken to determine moisture content and specific gravity.
Techniques for weighing tree components depend largely on tree size and the availability of
equipment. Take care to keep each tree and its parts separate from other trees. The best way to ensure

29This section is taken largely from Standard research methods for multipurpose trees and shrubs, eds. K.G.
MacDicken, G.V. Wolf and C.B. Briscoe (Arlington, Virginia: Winrock International, 1991).
30Length is the measured distance of a tree or specified portion of a tree following the lean or curvature, not
necessarily vertical or straight; height is the vertical distance between a standing trees apical bud and ground level.
57
this is to partition and weigh only one tree at a time. Measure green weight in the field. To calculate
moisture content, sub-samples should be taken, dried at 80 C and re-veighed.
Although harvesting is the preferred way to develop weight tables, it is not always possible due to
conservation or regeneration considerations. For biomass tables from non-destructive samples,
calculate stemwood volumes and convert them to biomass using specific gravity for wood and
expansion factors for canopy biomass. This requires a device for measuring diameter, such as a
Wheeler pentaprism caliper, Spiegel Relaskop or laser measuring device. A sample form for
recording data to construct biomass tables or volume tables using a metric Relaskop follows.
Analysis
Regression analysis should be performed for each biomass type defined during the planning process.
To make the tables useful for the widest possible range of environments, at least one complete set of
equations and weight tables should include both diameter and height terms.
58
The most common equations for biomass types include:
____________________________________________________________________________
Biomass Type Most Common Equation(s)*
____________________________________________________________________________
Whole Tree B = b
0
+ b
1
D
2
H
B = b
0
+ b
1
D
Woody Biomass B = b
0
+ b
1
D
2
H
Branch, Foliage, B = b
0
+ b
1
D
2
or Crown Weight B = b
0
+ b
1
D
____________________________________________________________________________
*B = predicted total or above-ground biomass, D = diameter (cm) at breast height (1.3 m), H = total
height (m), b
0
and b
1
= regression parameters estimated from the data.
Results should be represented as both regression equation and table to allow the widest possible
application of the work. When tables are unavailable and it is not practical to develop species-specific
biomass tables, the following general equations can be used:
Climate type based on
annual rainfall
Equation R
2
adjusted
Dry (<1500 mm) y = 34.4703 - 8.0671 D + 0.6589 D
2
.67
Moist (1500-4000 mm) y = 38.4908 - 11.7883 D + 1.1926 D
2
y = exp[-3.1141 + 0.9719 ln( D
2
H)]
y = exp[-2.4090 + 0.9522 ln( D
2
HS)]
H = exp[1.0710 + 0.5677 ln D)]
.78
.97
.99
.61
Wet (>4000 mm) y = 13.2579 - 4.8945 D
y = exp[-3.3012 + 0.9439 ln( D
2
H)]
H = exp[1.2017 + 0.5627 ln D]
.90
.90
.74
SOURCE: Brown, S., A.J.R. Gillespie and A.E. Lugo. 1989. Biomass estimation methods for tropical
forests with applications to forest inventory data. Forest Science 35:881-902.
where: exp [...] means " raised to the power of [...]"
y = above-ground biomass in kg
H = height in m
D = diameter at breast height (1.3 m)
S = wood density in units of tons/m3
NOTE: These equations are valid only for stems with dbh >5 cm.
The mean tree technique for biomass estimation
Allometry is an effective method for accurately estimating biomass of trees, tree components and
stands. However, the labor and expense of constructing and validating the necessary equations limit
the application of the allometric approach in biomass sampling. Many of the allometric equations
developed in the past were published in obscure journals, and furthermore have restricted applicability
outside the area of their development.
59
The mean tree technique can be a cost-effective alternative to more time-consuming allometric
methods. The mean tree technique was developed by several investigators during the 1960s and 70s
(Baskerville 1963; Attiwill and Ovington 1968; Crow 1971; Madgwick 1970; Madgwick and Satoo
1975; Madgwick 1981; Satoo and Madgwick 1982). The concept behind the method is that an
average-sized tree will also have an average amount of biomass. The usual approach is to select a
tree or trees of mean basal area. Basal area tends to be a good predictor of total biomass, since
diameter, basal area, and sapwood area all have a similar functional relationship to the quantity of live
foliage and branches in the crown. The selected trees are then destructively sampled to determine
their biomass. Subsampling may be used in the case of large trees (see Satoo and Madgwick 1982 for
detailed applications of subsampling tree components). The mean tree weight is then multiplied by
the number of trees in the stand to obtain an estimate for the total stand biomass. This basic technique
can be modified by including stratified random sampling, the basal area ratio method, or by using
weighted average values (Madgwick and Satoo 1975; Satoo and Madgwick 1982).
Properly used, the mean tree technique has several significant advantages: it is fast, it can be accurate,
and it does not require elaborate computations. It is most appropriately applied in homogenous, even-
aged, and well-spaced stands. The accuracy of this technique declines in diverse stands with a wide
array of bole diameters and tree sizes. Most agroforestry plantings, with their systematically spaced
trees of near-uniform age and size, are well-suited to the mean tree technique. Biomass estimates
within 2-10% of the true value appear realistic based on literature.
The precision of stand biomass estimates obtained by the mean tree technique can be improved by
using the basal area ratio method, and by stratified random sampling. Stratified random sampling
should be considered if the range of stem sizes is large. In stratified designs, approximately five trees
over several diameter classes are sampled. The biomass of each diameter class is calculated
separately, and then the class estimates are combined to derive a biomass estimate for the stand. This
is a less intensive sampling effort than would be required for the development of allometric equations,
but more effort than is needed for stands in which a single mean tree is adequate. If substantially more
than five trees per size class need to sampled, the mean tree technique loses any advantage over
standard allometric approaches.
The main disadvantage of the mean tree technique is that there is no estimate of the error. This leads
to two problems. First, without replication, there is no way to detect a poor estimate. Secondly, there
is no statistical method to compare sequential samples. Another shortcoming is that almost all
applications of the mean tree technique have been on coniferous species, which tend to be more
uniform in shape than deciduous species. Finally, since tree size is related exponentially to diameter,
the mean tree technique tends to be biased towards an underestimation of the actual stand biomass.
This bias becomes more pronounced as the range of tree sizes in the stand increases.
The largest challenge in using the mean tree technique in the field is to select trees for biomass
determination that are truly of average size. This requires careful measurement of stand diameters and
simple (yet crucial) math computations that could be a source of error by inexperienced technicians.
Second, if subsampling or stand stratification are required, the math and technical problems increase,
and more highly-trained field crews are required. However, it should be noted that these problems
apply to any method biomass determination used on trees in the field. Biomass measurement is a
tedious process. If the trees of a stand have a uniform size and structure and reliable allometric
equations for biomass are not available, the mean tree technique may provide rapid estimates of the
biomass of the stand.
60
References
Attiwill, P.M. and J.D. Ovington. 1968. Determination of forest biomass. Forest Science 14: 13-15.
Baskerville, G.L. 1965. Estimation of dry weight of tree components and total standing crop in
conifer stands. Ecology 46: 867-869.
Crow, T.R. 1971. Estimation of biomass in an even-aged: regression and mean tree techniques.
University of Minnesota Agricultural Experiment Station Science Journal Paper Series Paper no.
7487.
Crow, T.R. 1978. Common regressions to estimate tree biomass in tropical stands. Forest Science
24: 110-114.
Crow, T.R. and P.R. Laidly. 1980. Alternative models for estimating woody plant biomass.
Canadian Journal of Forest Resources 10: 367-370.
Madgwick, H.A.I. and T. Satoo. 1975. On estimating the aboveground weights of tree stands.
Ecology 56: 1446-1450.
Madgwick, H.A.I. 1981. Estimating the above-ground weight of forest plots using the basal area ratio
method. New Zealand Journal of Forest Science 11(3): 278-286.
Satoo, T. and H.A.I. Madgwick. 1982. Forest biomass. The Hague: Martinus Nijhoff / Dr. W. Junk,
Publishers.
Snell, J.A.K. and J.K. Brown. 1978. Comparison of tree biomass estimators: DBH and sapwood area.
Forest Science 24(4): 455-457.
61
Biomass table data collection form
Cruiser ________________ Project ________________ Date ___/___/___
Country________________ Tree species:________________ Biomass or volume (circle one)
Volume
Volume
segment no.
Stump
Diameter (cm) or Metric Relaskop
(Circle one)
Length
(m)
Bark
thickness
(cm)
Tree
no.
or biomass
component
diameter
(cm)
DBH
(1.3 m)
D1 D2 D3 Weight (kg) Comments
Metric relaskop diameter measurements
Distance to tree (m) Wide strip value(cm) Narrow strip value (cm) Distance to tree (m) Wide strip value (cm) Narrow strip value (cm)
15 30 7.5 25 50 12.5
20 40 10 30 60 15
62
D. Plot procedures
1. Navigate to plot center coordinates provided from database, map, map table or Form E.
2. Establish plot center by setting a plot center post (preferably PVC pipe painted with fluorescent
paint and marked with the plot number). Flag plot center area to increase visibility.
3. Set DGPS unit up near plot center and collect data for a minimum of 10 minutes (this assumes the
base station is already set and is collecting data simultaneously with the same receiver settings).
4. Set digital distance measurer on tripod over the plot center.
5. If the slope is greater than 10%, use a clinometer, Abney hand level or relaskop to determine slope.
Correct for slope using the following formula:
L
s
= L / cos S
where L
s
is the corrected plot radius, S is the slope angle in degrees, cos is the cosine decimal taken
from the back of the clinometer or from a table, and L is the plot radius.
Note plot dimension corrections on the plot card.
6. The crew chief begins by measuring the distance to the plot edge, flagging the beginning point and
directing a technician to begin taking dbh measurements. Each tree should be marked with bright,
durable paint at 1.3 m. The top edge of the painted mark should be at 1.3 m. Figure 4 shows the
proper placement of the dbh tape. The technician should read out the measurement, which the crew
chief should record and check visually.
7. When all of the trees in the plot have been measured, the crew chief must check to see that all of the
trees have been measured and painted.
63
64
Figure 4 continued.
65
Appendix 5: Field Procedures for Herbaceous
Vegetation, Soils and Standing Litter Crop
The carbon content of soil and litter can be measured and analyzed at relatively low cost if the data is
collected at the same time the inventory is conducted. This appendix describes sampling and analysis
of carbon in herbaceous vegetation, soil and standing litter crop.
In general, data should be collected in the following order:
1. Herbaceous vegetation
2. Standing litter
3. Soil
A. Procedure overview
1. Go to the northern edge of the plot and select a point 1 m inside the outside edge of the permanent
sample plot. This will be the first sampling location for herbaceous vegetation, litter and soils.
2. Lay the quadrat or circular sampling frame on the ground with the outer edge 1 m from the plot
boundary. Include in the sample only the vegetation that originates inside the sampling frame.
Exclude vegetation over-hanging inside the frame if the plant originates outside the frame, but include
vegetation over-hanging outside the frame if the plant originates inside the frame.
3. Clip herbaceous vegetation and small woody vegetation of less than 2 cm dbh, place in the sample
weighing bag, weigh, and record the weight. Select a small random sub-sample (e.g., a handful) of
this vegetation and place in a numbered sample bag for moisture content determination.
4. Before moving on, collect standing litter from the same sample site, place in the sample-weighing
bag, weigh, and record the weight. Mix the sample well and select a small random sub-sample (e.g., a
handful) of this litter and place in a numbered sample bag for moisture content determination.
5. Collect a soil core or slice for soil carbon analysis, place this on a plastic tarp, screen with 5-mm
mesh, mix well with other cores or slices, randomly select a sample, and place in a numbered sample
bag for carbon content analysis.
6. Proceed in a clockwise direction to the next sampling site within the sample plot. From the first
sampling location (i.e., North) this will be East, the next will be South and the last sampling location
will be West.
66
B. How many samples to collect
An inventory effort should calculate the number of samples required for a specified level of precision
before plot measurements begin. However, if data can be analyzed during the inventory (i.e., using
data collected from the previous days collections), fewer samples may be needed. Use Form C to
describe the sampling design for soil organic carbon.
C. Herbaceous vegetation
In permanent sample plots, herbaceous understory vegetation can be sampled using four to six
quadrats or circular sample frames per plot. Figure 5 shows three types of sampling frames useful for
this type of sampling, although many alternatives exist. The main criterion is that the frames be
durable and retain their size and shape over long periods of use. All frames used in a project must be
of the same size. Frames with at least one hinge allow the user to wrap the frame around broad
canopied plants when necessary. Experience suggests that round aluminum frames with hinges on
two sides are more durable than welded square frames, and are also more easily transported.
In plotless quarter-point surveys, one quadrat should be collected at random in each quarter (see
Figure 3).
Cut all vegetation inside each quadrat/circular sampling frame at ground level. Take care to cut at the
same height for each sample. Clip herbs within the sampling frame in a vertical column extending
from inside the sampling frame, so that samples represent the biomass within the frames area. Weigh
biomass for each quadrat and take and weigh a sub-sample for moisture content, and possibly for
determining nutrient concentration.
D. Standing litter crop31
Changes in standing litter crop can be important, particularly when forest soils are converted to land
uses that oxidize organic matter (e.g., crops that require intensive cultivation). It is easy to measure
the standing litter crop, but it requires consistent adherence to pre-defined standards.
Measure the standing litter crop by collecting all litter on the soil surface in each of the sampling
frames used for measuring herbaceous vegetation. Samples can be bulked by plot. Make sure to
record the number of sample frames collected in each plot. Samples should be weighed and sub-
samples collected in the same way as for herbaceous vegetation.

31Standing litter crop is the total weight per unit area of litter on the soil surface at the time of sampling. Litter is
organic debris on the soil surface, and is usually freshly fallen or slightly decomposed vegetation. Measurement of
the standing litter crop does NOT require monitoring of litterfall.
67
E. Soil sampling In general, soil samples should be taken when the permanent plots are
established and measured. Use either a soil corer of 30 cm in length or hand-dug pits of 30 cm in
depth. A soil corer may provide greater efficiency where soils are not excessively stony, although a
folding entrenching shovel (military type) is usually lighter and more versatile. Due to charcoals
high carbon content, it is important to take special care to remove bits of charcoal from samples at any
sites that have been burned prior to sampling.
Soil samples should be collected from the 0-30 cm horizon unless otherwise specified.32 To collect
soil samples, remove all vegetation and litter from the soil surface prior to sampling. Place the soil
core or slice on the plastic tarp and remove coarse fragments using a 5-mm screen. If multiple
subsamples are to be taken per plot, screen all samples on the plastic tarp and mix thoroughly to a
uniform color and consistency. Place a sample in a clearly labeled sample bag (preferably a cloth or
Tyvek oil sand bag). The quantity of soil required may depend upon the laboratory and analysis to be
used; discuss sample needs thoroughly with laboratory technicians beforehand, to ensure that samples
are properly prepared and labelled in the field.
To convert total or organic carbon concentrations into total quantities, bulk density of soils is required.

32 The greatest changes in soil organic carbon in non-humic tropical soils are often found between the 0-30 cm and
>30 cm horizons. For a summary of organic C contents in a wide range of tropical soils, see Table 5.2 in Properties
and management of soils in the tropics, by P. Sanchez (New York: John Wiley and Sons, 1976).
68
Bulk density is considered to have relatively low spatial variability,33 with coefficients of variability
of less than 10%. For a uniform soil type, four samples should be sufficient to estimate mean bulk
density to within 10% of the true value 95% of the time. The following procedure can be used to
determine bulk density with a Modified Uhland soil corer:
1. Identify tin sample boxes and tops, weigh and record as W1 (g).
3. Prepare a smooth surface at a sampling depth of 5 cm.
4. Drive sampler into the soil to fill inner core without compression (use mineral oil if soil-metal
adhesion occurs).
5. Trim ends, remove core. If core does not completely fill the cylinder, use glass bead adjustment. If
it does fill the cylinder, push contents into sample tin, close tin, mark and record tin number.
6. Place samples in an oven set to 100
o
C for about 72 hours. After drying, record the weight of the
tin + dry soil as W2 (g).
7. Calculate bulk density as: BD (g cm
-3
) = (W2-W1)/344.77
Soil C content (t ha-1 for the 0-30 cm soil depth) = BD * 300 kg m-2 * C concentration (%) * 10
F. Deciding what type of soil carbon analysis to do
Soils can contain two types of carbon: organic and inorganic (carbonate). All agricultural soils
contain some organic carbon, but not all soils contain inorganic carbon. In most cases, soil organic
carbon will be the most important source of soil carbon, although this is not true in arid soils
(Aridisols) and several other soil types. Most changes in soil carbon due to project activities are
assumed to be in organic matter34, and not in inorganic carbonate.
Many laboratories routinely use the Walkley-Black procedure for determining soil organic carbon,
although it is known to have a number of important limitations. However, because it is commonly
used, rapid, and simple, this method is recommended for analysis of soil organic carbon where total
carbon analysis is not required. Table 5 compares methodologies for determining soil organic carbon,
including Walkley-Black.
If soils are known to contain substantial quantities of inorganic carbonate and the inorganic carbonate
fraction is likely to change (e.g., if an arid soil is irrigated), then total carbon methods are necessary.
Table 6 provides a summary of the relative quantities of organic and carbonate soil carbon in soils, by
suborder, using Soil Taxonomy. Table 7 summarizes methodologies used for determining total soil
carbon.

33 Warrick, A.W. and D.R. Neilson. 1980. Spatial variability of soil physical properties in the field. In D. Hillel,
ed., Applications of soil physics. New York: Academic Press.
34 The average C content of soil organic matter ranges from 48 to 58%.
69
Table 5 Comparison of methodologies for determining organic C in soils.
Method Principle Advantages Disadvantages
Difference between total
C and inorganic C
Total C and inorganic C are determined on
separate samples: Organic C = Total C -
inorganic C.
Useful if total C and inorganic C
are routinely determined
Two separate analyses are required.
Total C determination requires special
equipment. Organic C calculated by
difference has some inherent error.
Determined as total C
after removal of
inorganic C
Total C is determined in soil sample after
removal of inorganic C with an acid
pretreatment:
Organic C = Total C
Accurate if dolomite is absent
from soil
Not all dolomite in soil may be removed
by acid treatment. Specialized equipment
needed.
Dichromate oxidation
without external heat
Dichromate oxidizes organic C to CO
2
in acid
medium. Amounts of Cr
2
O
7
2-
reduced is
quantitatively related to organic C present.
Not all organic C in samples is oxidized when
external heat is omitted, and a correction
factor is required.
Very rapid and simple. No
special equipment required
Incomplete oxidation of organic C
necessitates use of correction factors,
which often results in erroneous values.
Chloride, Fe
2+
and MnO
4
interfere with
method. It assumes soil organic C has an
average valence of 0.
Dichromate oxidation
with external heat
This is the same as the dichromate method
above except that all organic C in the sample
is oxidized, and no correction factor is
required.
Rapid and simple. Complete
oxidation of organic C occurs
Chloride, Fe
2+
, and MnO
2
interfere with
method. Some specialized equipment is
needed. It assumes soil organic C has an
average valence of 0>
+YMHI XS *SVIWX 'EVFSR 7XSVEKI
70
Table 6. Organic and carbonate carbon mass in soils of the world35
Organic Carbon Carbonate Carbon Total Carbon
Suborder/Order
---------------- Gigatons of carbon (Petagrams or 1 x 10
15
g) ----------------
Folists 1 0 1
Fibrists 250 0 250
Hemists 68 0 68
Saprists 71 0 71
Histosols 390 0 390
Aquands 1 0 1
Cyrands 18 0 18
Torrands 1 1 2
Xerands 2 0 2
Vitrands 1 0 1
Ustands 13 0 13
Udands 33 0 33
Andisols 69 1 70
Aquods 4 0 4
Ferrods 0 0 0
Humods 41 0 41
Orthods 53 0 53
Spodosols 98 0 98
Aquox 1 0 1
Torrox 0 0 0
Ustox 41 0 41
Perox 16 0 16
Udox 92 0 92
Oxisols 150 0 150
Aquerts 1 1 1
Xerets 5 1 6
Torrets 12 14 26
Uderts 5 0 5
Usterts 15 9 24

35 Source: Eswaran, H., E. Van den Berg, P. Reich and J. Kimble.1995. Global soil carbon resources in R. Lal, J.
Kimble, E. Levine and B.A. Stewart (Eds.). Soils and Global Change, CRC Lewis Publishers, Boca Raton, FL.
71
Table 6 continued.
Suborder/Order Organic Carbon Carbonate Carbon Total Carbon
Vertisols
Salids 5 113 118
Gypsids 3 12 15
Calcids 17 407 424
Durids 0 1 1
Argids 38 112 150
Cambids 47 399 446
Aridisols 110 1044 1154
Aquults 5 0 5
Humults 4 0 4
Udults 50 0 50
Ustults 40 0 40
Xerults 2 0 2
Ultisols 101 0 101
Albolls 2 0 2
Aquolls 1 1 2
Rendolls 0 1 1
Xerolls 13 23 36
Borolls 15 29 54
Ustolls 17 32 49
Udolls 24 53 49
Mollisols 72 139 139
Aqualfs 6 0 6
Boralfs 35 0 35
Ustalfs 45 71 116
Xeralfs 14 0 14
Udalfs 36 56 92
Alfisols 136 127 236
Aquepts 67 12 79
Plaggepts 0 0 0
Tropepts 20 26 46
Ochrepts 135 247 382
Umbrepts 45 0 45
72
Table 6 continued.
Suborder/Order Organic Carbon Carbonate Carbon Total Carbon
Inceptisols 267 285 552
Aquents 20 0 20
Arents 0 0 0
Psamments 21 30 51
Fluvents 3 6 9
Orthents 62 81 143
Entisols 106 117 223
Rocky land 13 0 13
Shifting sand 5 0 5
Misc land 18 0 18
TOTAL 1555 1738 3293
73
Table 7. Comparison of methods used for determining total C in soils.
Method Principles CO
2
Determination
Advantages Disadvantages
Dry
combustion
(resistance
furnace)
Sample is mixed with CuO and
heated to 1000 C in a stream of O
2
to
convert all C in sample to CO
2
.
Gravimetric
Titrimetric
Reference method widely used
in other disciplines Variable
sample size
Time-consuming; leakfree O2
sweep train is required. Slow
release of CO
2
from alkaline
earth carbonates
Dry
combustion
(induction
furnace)
Sample is mixed with Fe or
accelerators and rapidly heated to
>1,650 C in a stream of O
2
to convert
all C in sample to CO
2
Gravimetric Rapid combustion
High temperature ensures
conversion of C to CO
2
Leakfree O
2
sweep train
required, induction furnace is
expensive
Dry
combustion
(automated
methods)
Sample is mixed with catalysts or
accelerators and heated with
resistance or induction furnaces in a
stream of O
2
to convert all C in
sample to CO
2
Gas
chromatography
Gravimetric
Conductrimetric
Rapid and simple, good
precision
Expensive equipment. Slow
release of CO
2
from alkaline
earth carbonates with
resistance furnace
Wet
combustion
(combustion
train)
Sample is heated with K
2
Cr
2
O
7
-
H
2
SO
4
-H
3
PO
4
mixture in a CO
2
-free
air stream to convert all C in sample
to CO
2
Gravimetric,
Titrimetric
Equipment readily available,
good accuracy, easily adapted
to analysis of solutions,
titrimetric analysis of CO
2
less
subject to operator error
Time-consuming; gravimetric
determination of CO
2
requires
careful analytical techniques,
titrimetric determination of
CO
2
is less precise
74
G. Sample preparation
This section describes procedures for preparing samples of soil, litter and vegetation for analysis after
they have been collected in the field.
Soils
Soils should be air-dried, but not exposed to direct sunlight. Check with the laboratory for detailed
arrangements.
Litter and vegetation
After weighing the samples, take sub-samples of litter and vegetation to determine moisture content
and nutrient concentration. The following guidelines are suggested:
Moisture content: Mix the sample and collect one random sub-sample of approximately one handful
of litter of vegetation per quadrat/circular sample plot. Bulk these subsamples by permanent plot or
transect when using the plotless method. For moisture content, collect at least five sub-samples for
each vegetation type. Sub-samples should be weighed in the field then returned to the laboratory for
oven-drying at 70-80
o
C to a constant weight and reweighed to determine dry-matter.
Nutrient concentration: This is necessary only if the decision is made to use actual C concentration
data for vegetation, or if actual data are to be used to predict carbon pool changes using a computer
model. A minimum of five samples for each partition is suggested for C (for carbon calculations
only) or C and N analysis (for modelling purposes). This means five core samples of wood (collected
at dbh), five foliage and litter samples.
75
Appendix 6: Measuring Carbon in Agroforestry
These methods are designed for use in agroforestry and farm forestry plantings as described in Section
3. They have been only preliinarily field tested, and should be used or cited with caution (see Field
Tests of Methods for Monitoring Carbon in Forestry Projects).
Map the project area.
1) Using current aerial photos, determine the number, size, and location of agroforestry plantings in
the project area. If aerial photos are not available, obtain a list of farmers who have associated
agroforestry plantings in association with the project and plot the locations of their farms to the
extent possible on a topographical map.
2) Use a soils map to stratify the plantings into groups if soil types vary significantly within the
project area. If a soils map is not available for the area, consider stratification if there are major
differences in topography, drainage, or parent material which affect the suitability of soil for crops
or trees. Avoid establishing more than three strata, if possible.
3) Assign a number to each planting within the project area.
Select a preliminary sample of plantings for the determination of required sample size.
1) Using the assigned numbers, randomly select three farms from each stratum. Measure
accumulated above- and below ground-biomass in the agroforestry plantings using the methods
outlined below. Calculate the variance in the data, and use this value to estimate the number of
farms in each stratum that must be sampled to estimate carbon accumulation at the desired level of
precision (see Appendix 2).
2) If aerial photos are not available for the region and a preliminary estimate of average
agroforestry plantation size was not calculated, then at least six farms must be selected per stratum.
The size of agroforestry plantings will be measured on each of these six farms, while three will be
selected at random for the measurement of accumulated biomass. As with biomass, the variance in
planting size will be calculated to determine the number of farms that must be sampled to estimate
the average area of agroforestry plantings at the desired level of accuracy.
Select a sample of plantings for the determination of accumulated biomass.
1) Using the numbers assigned to each farm and the desired sample size calculated in the preceding
step, randomly select a sample of farms from each stratum. Make the sample list large enough so
that it includes several alternates.
2) The sampling scheme must be nested if aerial photos were not used to provide a preliminary
estimate of farm size. First, establish a random list of farms in each stratum that will be sampled
for agroforestry planting size. Next, from this list randomly select a sub-sample. Agroforestry
plantations on these farms will be sampled for both size and accumulated biomass.
76
3) Circle the location of the farms to be sampled on an aerial photo or topographic map. Use this
map to efficiently organize the sampling effort. Include the names of each farmer on both the map
and list of sample farms, if available.
Farmer contact
1) Before initiating field measurements, first make contact with local community leaders and
officials. Be prepared to present identification, official letters, or other forms of authorization.
2) Contact a farmer on the sample list. Formally introduce the carbon inventory crew: the names of
each crew member, where they come from, their professional titles, and the names of their
respective organizations.
3) Describe clearly the purpose of the carbon inventory. Do not avoid explaining why carbon
sequestration is important, if asked. Describe the types of measurements that will be conducted.
4) Ask permission to inventory the accumulated biomass of the farmers agroforestry planting. If
permission is not granted, thank the family for their time, and move on to the next farm. Replace
the sample from the list of alternates.
Conduct farmer interview (see interview form, this Appendix).
Plot reference point location
1) Prepare a sketch map of the farms agroforestry plantings on the reference point location form.
2) Walk around the perimeter of the agroforestry planting to determine the location of each corner
of the planting with the GPS unit. Record these values on the sketch map.
3) Estimate the approximate length and width of the agroforestry planting. Record the values on
the reference point location form.
4) Divide the estimated length and width of the agroforestry planting by two. Record the values on
5) Locate the southeast corner of the agroforestry planting. Label the location of the southeast
corner on the sketch map.
6) Starting at the southeast corner and proceeding along the long side of the agroforestry plantation,
measure out a line exactly equal to one half the value of the previously estimated length of the
plantation (Figure 6). Use a 100 meter tape to measure this distance precisely. Record the exact
distance and bearing of the line on the reference point location form. The endpoint of this line shall
be referred to as the turn point.
7) Paint a blue ring at DBH on a tree located along the perimeter of the plantation proximal to the
turn point.
8) At the turn point, turn exactly 90 relative to the direction of travel and towards the interior of the
planting. Proceed a distance exactly equal to one half of the previously estimated width of the
agroforestry planting. Use the 100 meter tape to measure out this distance precisely. The endpoint
77
of the line shall be referred to as the plot reference point. Record the exact distance and bearing of
the line on the reference point location form.
9) Mark the plot reference point with a 40cm section of rebar. The bar shall be driven 30cm into
the soil, and the last ten cm of the rebar above the soil surface shall be painted blue.
10) Measure the exact distance and bearing to the plot reference point from two reference trees.
Record these values on the reference point location form. The difference between the bearings
from the two reference trees should be approximately 90 . Paint a blue ring around each reference
tree at DBH, and record the species and DBH of each reference tree on the reference point location
form.
11) Determine the coordinates of the plot reference point with the GPS. Record the coordinates on
12) Once a plot size has been selected, employ the same plot size on all plots throughout the
duration of the biomass inventory, regardless of the tree spacing encountered on a particular
agroforestry planting.
Plot location
1) Refer to Figure 7. If the inventory plot size is 1/20th ha or greater, the following distances and
bearings shall be used to locate four plots in relation to the plot reference point (RP):
plot 1 is located 60.0 m from the RP at a bearing of 45 NE
plot 2 is located 20.0 m from the RP at a bearing of 135 SE
plot 3 is located 60.0 m from the RP at a bearing of 225 SW
plot 4 is located 20.0 m from the RP at a bearing of 315 NW
78
79
80
2) If the inventory plot size is 1/40th ha or less, use the following distances and bearings to locate
four plots in relation to the plot reference point (RP):
plot 1 is located 45.0 m from the RP at a bearing of 45 NE
plot 2 is located 15.0 m from the RP at a bearing of 135 SE
plot 3 is located 45.0 m from the RP at a bearing of 225 SW
plot 4 is located 15.0 m from the RP at a bearing of 315 NW
3) If either the plot center or more than 25% of the plot area is located outside the perimeter of the
agroforestry planting, do not establish a plot at that point. Instead, attempt to install a plot at an
alternate location. The first alternate is plot 5, falls outside of the agroforestry planting, then the
next alternate is plot 6, then plot 7, and finally plot 8.
4) Install the plot rope stake firmly at the center of the plot.
5) Mark the plot center with a wire flag labeled with the plot number.
Conduct an inventory of woody stems >5.0 cm DBH
1) Starting at north and moving clockwise, record the total height, DBH, and species of all woody
stems > 5.0 cm DBH that fall within the plot. Record that data on the large stem biomass form.
2) For borderline trees, if more than half the stem falls within the plot, the tree is in; if more than
half the stem falls outside the plot, the tree is out. If the plot boundary coincides exactly with the
center point of the tree, flip a coin. If heads, the tree is in; if tails, the tree is out.
3) The corrected slope distance should be calculated for borderline trees or for trees just outside the
plot if the slope is greater than ~20%. To do this, determine the slope angle from the plot center to
the tree in question with a clinometer. Next, multiply the cosine of the angle (provided by the table
printed on the side of the clinometer) by the apparent distance. The resulting value is the true
horizontal distance. Use this value to determine if the tree is in or out of the plot.
Conduct inventory of herbs, litter, soil and woody stems < 5.0 cm DBH (See Appendix 5)
Limitations
The methods outlined here should be adequate for a wide array of agroforestry system configurations,
but they are not appropriate for all types of plantings that may be encountered. These methods are
intended for agroforestry systems which: are predominantly square, rectangular, or round in shape; at
least 0.25 hectares in size; and have trees as the predominant cover type. Narrow strips of trees, such
as windbreaks, would require a different inventory method; so would silvopastural systems in which
the trees are widely scattered over a dominant matrix of grass. However, for most situations in which
an objective of the agroforestry project is to sequester carbon, this inventory scheme should prove
adequate.
81
Woody stem inventory form
mo / day / year
Crew:____________________________________________ date:___/_____/____
Farm number:_______________ Stratum number:________________________________
Plot number: _______________ Plot radius: _________________
Woody stems >5.0 cm DBH
tree
number
species
code
height DBH tree
number
species
code
height DBH
82
Farmer interview
Crew:____________________________________________ date:___/___/___
Farmers name: ________________________________________________________________
Farm number:_______________ Stratum number:________________________________
Farm location:_________________________________________________________________
Approximate date when agroforestry planting was established:__________________________
Land use of the plot before the planting was established:
_ fallow___ years _ pasture___ years _ crop___ years _ forest
Approximate size of agroforestry planting:___________________________________________
Reasons for establishing the agroforestry planting:_____________________________________
_____________________________________________________________________________
Tree component of the agroforestry planting:
species spacing number
planted
growth rate problems products/ yield
Crop component of the agroforestry planting:
species spacing planting date harvest date problems products/ yield
83
Reference point location form
mo / day / year
Crew:____________________________________________ date:___/_____/____
Farm number:_______________ Stratum number:_________________________________
Agroforestry plantation sketch map
est. length of agroforestry plantation = ______ 2 = ______ bearing to turn pt. = _______
est. width of agroforestry plantation = ______ 2 = ______ bearing to reference pt. = _______
species DBH distance to
reference pt.
bearing to
reference pt.
1st reference tree
2nd reference tree
GPS coordinates of the plot reference point = ________________
84
Appendix 7: Estimating Root Biomass36
Estimating root biomass is expensive. Yet, root biomass is an important carbon pool because
it often represents 10 to 40% of total biomass. Two general approaches for claiming carbon
credit for root biomass are possible: 1) use conservative, non-controversial estimates of root
biomass based on literature values for similar vegetation types, and; 2) measure root biomass.
The only advantage to measuring biomass for carbon credit is that in most cases, actual root
biomass will likely be substantially greater than the conservative estimates. The decision of
whether or not to measure should be based on the price of carbon compared to the cost of
collecting the additional data required to claim credit.
A. Estimating root biomass using the literature
The data available on root biomass are limited due to the high costs of sampling and
measuring roots. However, the literature does contain root biomass values for a wide range of
vegetation types. Unfortunately, the methods used vary greatly and the very limited
information on vegetation type by site class does not allow high confidence in using actual
literature values in most cases. For example, limited root biomass data from tropical forests
suggest that the root:shoot ratio varies from 0.03 to 0.49, with below-ground biomass ranging
from 11 to over 130 t/ha. The fact that some root biomass exists below living above-ground
biomass is undisputable, but the question is, How much? How conservative should estimates
using literature values be? In a literature-based approach, the key is to use estimates that are
conservative enough so that they are not easily refuted. In the example of the root:shoot ratio
in tropical forests, a value of 0.10 or 0.15 would generally suit this purpose.
The level of conservatism required to pass minimum criteria for carbon credit is still
undefined. A reasonable approach might be to use the lowest above-ground:below-ground
biomass ratios to estimate below-ground biomass, based on actual inventory data of above-
ground biomass.
B. Measuring root biomass
If such data are not available, root biomass can be estimated by sampling and measurement
using the methods described in the following paragraphs. In measuring root biomass, note the
following points:
Samples should be taken from representative volumes of soil usually 0-30 cm soil depth
unless otherwise specified.

36 Portions of this appendix come from Appendix D: Roots: Length, biomass, production and mortality, by
M. Van Noordwijk, in Tropical soil biology and fertility: A handbook of methods, eds. J.M. Anderson and
J.S. Ingram (London: CAB International, 1992) and Methods of studying root systems, by W. Bohm
(Berlin: Springer-Verlag, 1979).
85
Samples should be taken during the time when expected standing root biomass is highest
(e.g., avoid the late part of the growing season).
The methods for sampling, storing, and washing samples will always lead to some loss of
dry weight and nutrients. A correction factor of 1.25 - 2.0 should be applied to the final
data, with the correction factor based on the estimated losses due to sampling and
processing.
Two types of sampling may be required: 1) core sampling to determine root biomass in the 0-
30-cm soil depth; 2) monolith sampling to determine relative root distribution beyond 30 cm
soil depth. Decisions about the types of sampling required must be site specific and include
consideration of precision needs, the availability of data on root distributions for the species
being inventoried, soil depth, texture and stoniness.
Core sampling
A soil corer removes a known volume of soil from a known depth in the profile, without the
need for digging a soil pit. A core of 50 - 80 mm diameter is satisfactory, and the corer can be
inserted either manually or mechanically. Manual coring is difficult at depths greater than 50
cm and in clay or stony soil. In dry sandy soil a smaller core diameter may be needed to
reduce losses of soil when extracting the core. In very stony soil, or where there are many
woody tree roots, coring may not be possible. In these cases, regular, known volumes of soil
(monoliths) can be taken from the face of a pit and treated in the same way as cores.
A commercially available split-core corer, such as the AMS split core sampler kit with core
tip, is recommended.
Ideally the profile should be sampled to the limits of rooting depth. At that depth, however,
rooting intensity is low and spatial variability high. A meaningful lower limit can be set based
on initial observations of the profile wall. In some cases a linear relationship of the log of root
mass versus depth (a negative exponential root distribution) may help to extrapolate root
densities in the soil beyond sampling depth. All soils must be sampled to a minimum depth of
30 cm.
Root extraction
The best approach to root extraction is to wash roots from the cores immediately upon return
from the field. Core samples can be stored in sealed polyethylene bags in a refrigerator for a
few days or deep freeze until processed. If deep freeze facilities are not available, samples can
be stored air-dried and re-wetted before washing. Losses of dry weight due to the methods
used for storage should be checked.
Soil texture, structure, degree of compaction and organic matter content greatly influence the
precision and time required to extract roots from cores. The simplest method involves gently
washing a presoaked sample over a large diameter sieve of 0.3 - 0.5 mm mesh. The work can
be simplified by washing over a combination of sieves: one with 1.1 and one with 0.3 mm
mesh. The first sieve will contain mostly roots, the second mostly debris. The material
removed from the sieve(s) can then be mixed in water and the suspended material decanted
(live roots of most species have a specific gravity of about 1.0). This residue should then be
hand sorted in shallow dishes under water to remove fragments of organic matter and dead
roots; normally it is better to pick live roots from the sample and leave debris behind in the
dish.
86
Presoaking samples overnight in 5% sodium hexametaphosphate expedites the process of
washing roots from clay soils, but the chemical discolors the roots (particularly in soils with
high organic matter content) and may disrupt the tissue, making subsequent identification of
live roots more difficult. Such pretreatment will also interfere with chemical analyses. Any
lengthy washing procedure may alter the element content of root tissue; only a subsample
hand sorted with a minimum of water and processed on the day of sampling should be used
for analysis.
Classifying the roots
Fine roots are the most important part of the root system for water and nutrient uptake, as they
form the largest part of total root length or root surface area. For woody perennial vegetation
there is a fairly obvious distinction between the more or less permanent, secondarily thickened
roots and the ephemeral, unthickened roots. This functional distinction usually falls
somewhere between 1 and 3 mm root diameter. Roots above 10 mm diameter are not
adequately sampled by coring. For herbaceous perennial and short-lived vegetation, roots
should be separated into <2 mm and > 2 mm classes. In mixed vegetation, separation of roots
of different species is difficult and is not necessary.
Sampling intensity
Even in the most homogeneous soils, spatial variability of root density will be high, with
coefficients of variation in root weight commonly in excess of 40%. On heterogeneous soils
the C.V. may be much higher. This variability implies that many replicate samples are needed
if estimates of root weight need to be precise.
It is advisable to obtain reliable information at one or two well chosen situations, rather than
non-reliable data on many. Within each treatment plot take at least 3 cores. Within each plot
the samples can be pooled. In natural vegetation where there is no obvious strategy for sample
stratification, take the cores on random coordinates. Where patterns are likely to occur (e.g,.
row crops, alley cropping) stratification should use within row vs. between-row strata.
Monolith sampling
Monolith samples can be obtained with pinboards made by inserting U-shaped, stainless steel
pins or bolts in plywood. The size of the pinboard is determined by the vegetation type, based
on previous observations, such as rooting depth and distribution and practical considerations.
Soil collected with a pinboard is heavy (a sample of 100 x 60 x 10 cm of soil will weigh about
100 kg), so pinboard size should be matched to the means of supporting and moving a full
pinboard. Washing away the soil exposes the roots for observation. If a coarse mesh screen is
put on the pins before the board is pushed into the soil, this screen can help to keep the roots in
their original location while washing the sample. Washing the sample can be facilitated by
soaking overnight in water, deep freezing (for clay soils), soaking in oxalic acid (for soils with
free calcium carbonate) or soaking in hexametaphosphate, preferably under vacuum.
Whatever the pretreatment used, gentle washing must follow.
After washing away the soil: lift the root system on the mesh screen; photograph it (on a black
cloth as background); and/or cut it according to soil layers (indicated by string between the
87
pins while washing the sample), depth zones and/or distance to the plant, in order to obtain
root biomass and/or root length (see below for root length). To estimate total biomass per
plant, root weight density per zone and depth has to be integrated over the relevant volume.
Although the pinboard method is more time-consuming than other methods, it gives more
information per unit of effort spent. The methods major weakness is that roots may break or
be displaced during washing. It is easier to distinguish between live and dead roots via
pinboard sampling than in methods where the root system is not sampled in its entirety.
Assessment of root mass
Washed root samples can be stored in sealed polyethylene bags for a short time in a
refrigerator, but deep-freeze storage is preferable. Oven-dry the roots and weigh. Next the
dried samples should be combusted for 5 hr. in a muffle furnace in 550

C and the residue
weighed. Results should be expressed as ash-free oven-dry mass per unit volume of soil.

Guide To Monitoring Carbon Storage in Agroforestry Projects

Uploaded by

Copyright:

Available Formats

Guide To Monitoring Carbon Storage in Agroforestry Projects

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Guide To Monitoring Carbon Storage in Agroforestry Projects

Uploaded by

Copyright:

Available Formats

A Guide to

Monitoring Carbon Storage

or 15view. (However, this is a rapidly changing

You might also like