(3a) Plants and Soils: "Biophysical" measurements and models
Need and Objectives for Enhanced Measurements and Models
Regional and continental scale carbon fluxes between the soil-plant system, the
atmosphere and the ocean are the product of diverse ecosystems responding to
the interactive effects of climatic and edaphic factors, natural disturbance, air
pollution, land use and management. Soil- and plant-associated sources and
sinks of carbon are intimately coupled, but on short and medium time scales they
may respond in different ways, and even in different directions, to climate and/or
human perturbations. Land/atmosphere carbon exchange is highly variable
because of these factors, both temporally and spatially. In this document, we use
the term "biophysical" to refer to the ensemble of physiological, ecological,
climatic, and other environmental factors that regulate fluxes of major carbon
gases from terrestrial ecosystems, including soils.
There is a relatively good knowledge base of the size and distribution of soil and
vegetation carbon stocks of the major terrestrial ecosystem types in the U.S.
However, there are still very limited data on changes of soil and vegetation
carbon stocks (i.e. whether ecosystems are functioning as net sinks or sources
for CO2). In non-forested ecosystems (e.g.. grasslands, croplands, wetlands),
comprising 2/3 of the U.S. land surface, variation and change in soil carbon is the
overriding control on net ecosystem C flux. Therefore understanding and
quantifying soil C fluxes and C stock changes are imperative for understanding
the continental-scale C balances. Changes in live and dead above-ground
biomass is important in forests, and in some non-forest systems. Change in soil
C may be important over the longer term for forests.
Ecosystem C flux is poorly documented for some ecosystems. For example,
“woody encroachment”, which refers to the expansion of sparse woodlands into
grasslands as a result of grazing and fire suppression practices, has been
identified as a likely significant carbon sink, yet there are few estimates available.
Comprehensive, systematic measurements of carbon stocks and fluxes in soils
and vegetation will be a major product of the NACP.
Peatlands are particularly important among currently non-inventoried
ecosystems. Though peatlands cover only 12% of the surface of North America,
and total ecosystem productivity rates are low, stocks of soil carbon are huge.
[Harden et al. 1992]. An amount of C (~455 Pg) equal to about 60% of the C
pool in the atmosphere is stored within meters of the surface. Peatlands and nonforested wetlands are also significant sources of atmospheric CH4 [e.g. Crill et
al., 1999]. Peatlands are especially vulnerable to climatic warming. Peat remains
stable only while frozen or saturated with water and changes in temperature,
precipitation or surface hydrology can quickly change a peatland from a small
sink for CO2 and a source of CH4, to a strong source of CO2 and a small sink for
CH4.
Recent studies point to the large role that disturbance regimes (recovery,
frequency) play in the long-term pattern of net carbon uptake over North
America. The current C sink in U.S. forests appears to be primarily the
consequence of land management activities of the 19th and 20th centuries,
bringing into focus the important role of historical legacies in regulating current
balances in major ecosystems. These studies are based on forest and
agricultural inventory data, historical rates of agricultural clearing and
abandonment, historical rates of wood harvest, wild fire statistics, and growth and
decay rates derived from the ecological literature. Because of inconsistent and
incomplete sampling through the historical period, and lack of process
understanding, the uncertainty in attributing the estimated C sink to past
disturbance is high. Natural disturbances, particularly fire and insect epidemics,
may be playing an increasingly important role in regional C fluxes. The impacts
of these disturbances on ecosystem processes and C transfers among different
C pools need to be understood and quantified.
The past and current role of other factors affecting terrestrial C, such as
increasing atmospheric CO2 and N deposition, is poorly understood at large
spatial scales. Better quantification of the relative contribution of different factors
is critical for monitoring and verifying the effect of C sequestration activities, and
attributing observed changes to the appropriate cause. Carbon sequestration is
likely to become part of a suite of policy measures to reduce greenhouse gases
in the atmosphere, and verification of the effectiveness of these measures will be
needed.
A variety of data collected at different temporal and spatial scales, coupled with a
rigorous estimation process, will lead to improved regional- and continental-scale
estimates of land/atmosphere carbon exchange. Data sources include
reconstruction of land use/land cover history from statistical records, compilation
of past and ongoing resource inventories, a variety of remote sensors, and many
different kinds of intensive ecosystem monitoring and process studies such as
those conducted at Long-Term Ecological Research (LTER) sites and a network
of direct flux measurement sites (AmeriFlux). Temporal resolution of land use
records and inventory data is low (5-10 years), but spatial resolution can be very
high (county-level to individual tracts of land). Data from CO2 flux towers, remote
sensing, and enhanced atmospheric sampling can provide information with much
higher temporal resolution, complementing the high spatial resolution of more
traditional land-based data. Observations at intensive ecosystem monitoring
sites provide validation for more extensive measures, and process-level
understanding for interpreting larger-scale phenomena.
The various data sources are input to a variety of modeling approaches, from
bookkeeping to biophysical/biochemical process models. Advances in modeling,
such as newly emerging dynamic global vegetation models (DGVMs) and high
resolution biophysical models, will play a major role in integrating the land data
with atmospheric monitoring. Models are also required to integrate the
biophysical estimates with socioeconomic models to provide the critical link
between the NACP and the needs of policy makers and land managers.
Despite the variety of available land-based and satellite data, and continuing
efforts to improve modeling capabilities, estimates of land/atmosphere C
exchange are unacceptably imprecise, and not uniformly available for the land
surface of North America (or anywhere else in the world). A critical concern is
that, with the exception of measurements at CO2 flux towers, none of the land
surface measurements are designed to monitor changes in C stocks or fluxes.
The observations therefore lack features needed to attain higher precision and
reliability. Key deficiencies include lack of complete ecosystem C measurements
(particularly below-ground C), gaps in spatial coverage, inconsistent procedures
with time and location, and lack of sufficient temporal resolution (re-measurement
intervals as long as 15 years in important areas). Model interpolations have
been used to fill in the missing information, but evidently rigorous field sampling,
traceable to established long-term benchmarks, is needed.
The goals for NACP development of plant/soil ("biophysical") measurements and
models are to:



reduce the uncertainty in land-based monitoring of changes in carbon
stocks;
fully integrate land-based measurements with atmospheric
measurements; and
provide the mechanistic foundation for inverse modeling and data
assimilation.
Several research objectives will support these goals:



To reduce the uncertainty of ongoing inventory and monitoring of national
greenhouse gas emissions from land, and improve ability to attribute
observed changes to natural and human disturbances;
To develop well-quantified large-scale estimates of C exchange with the
atmosphere, for independent validation of estimates derived primarily from
atmospheric measurements; and
To provide the information on ecosystem-level soil and plant carbon fluxes
necessary to understand and interpret larger-scale regional and
continental flux estimates that will be obtained in the NACP.
A long-term observational strategy should include several key elements:


Identification of gaps in current sampling strategies (land classes,
locations, potential importance in terms of C pool sizes and fluxes)
Enhance established networks or begin new activities to fill gaps, including
in situ sampling and remote sensing.





More complete and comprehensive, longer-term soil-plant-atmosphere
exchange data for representative ecosystems.
Full exploitation and efficient management of existing data (in situ and
satellite): better acquisition, assimilation, analysis, and dissemination.
Assemble and distribution ancillary data sets needed for interpretation: soil
characteristics, hydrology, meteorology/climatology, etc.
Analysis and modeling activities leading to resolution of discrepancies
between predictions of biogeochemical models and atmospheric inverse
models.
Development and testing of new plant-soil-atmosphere models for
integrating land and atmospheric measurements.
Hierarchical Approach to Integrating Biophysical Measurements
The integrating strategy for biophysical measurements and models includes a
number of key features:









Linking observations across space and time using a nested design
Linking observations with understanding from process studies
Careful selection and definition of parameters for integration
Representative of major land cover and land-use types
Estimation of critical variables for understanding and quantifying C fluxes
Measurement of common variables across tiers using standard protocols
QA/QC at all sample phases, and quantification of estimation errors
Dynamic coupling of atmospheric, biospheric, and human systems
Advances in integrated modeling technology and analysis
Large-scale land monitoring programs can be implemented efficiently using a
tiered (multi-phase) sampling strategy. Some data elements are identically
defined and collected at each tier to provide links among tiers, while other
variables may be unique to one or a subset of tiers. Spatial and temporal
resolution of data collection will be unique and complementary among tiers. The
combination of remote sensing, extensive inventory, medium intensity sampling,
and intensive observations comprise a powerful, flexible, and potentially efficient
data collection system. The sample tiers can be linked statistically so that
inferences about the entire population within cover classes can be made. The
observation system should have the capability to integrate with atmospheric
monitoring, but should also stand alone to provide independent estimates of C
fluxes for validation and as a contribution to ecosystem science.
Multi-tiered sampling and analysis systems have been designed and
implemented in the U.S. for land inventories and more recently for linking new
remote sensors with field measurements. For the NACP, the first tier involves
wall-to-wall mapping and remote sensing of cover class at the continental scale.
Middle tiers include (1) existing extensive land inventories composed of a large
number of sample plots, and (2) a proposed new set of approximately 1000
medium intensity plots with more resolved direct measurements of C flux than
land inventories, also selected to represent typical conditions across the
landscape. The final tier includes the existing and potentially new intensive
observation sites where the most detailed observation are made, such as LTER
and AmeriFlux sites.
The following table illustrates the multi-tier concept with a listing of few of the
variables likely to be central to the land observation system:
Example
data elements
Land cover
class
Leaf area
index
Live
Biomass
Land cover
change
Wildfire
disturbance
Climate
variability
Extensive
inventory (FIA
and NRI)
X
Mediumintensity
sample (new)
X
Intensive
observation
sites (LTER
and Ameriflux)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Mapping and
remote
sensing
X
X
X
X
X
X
X
Soil CO2 flux
Methane flux
Dissolved
organic C
Ecosystem
CO2 flux
X
X
A complete and well-defined set of variables will be developed as this plan
becomes implemented. The role of modeling and analysis is central to the
evolution of an efficient data collection system. Appropriate estimators will be
defined through modeling and analytical studies, and recommendations made for
enhancing current observations for North America to produce an efficient,
continuing multi-tier network that is optimized for estimating C flux at multiple
temporal and spatial scales. An efficient way to integrate across scales is not
apparent for some critical variables such as soil CO2 and methane flux.
Preliminary studies and pilot implementation tests will be undertaken to develop
a strategy for these variables.
Enhancements to Ongoing Terrestrial Monitoring Networks
Remote Sensing
Current land inventory systems in the U.S. use a combination of high-altitude
aerial photography and Landsat Thematic Mapper (TM) data for land
classification and for area change detection. A global observation system using
the EOS-MODIS sensor and a network of ground observations has been
deployed for estimating productivity and land cover change. LANDSAT-TM and
EOS-MODIS are already acquiring data that provide land cover estimates but in
the case of Landsat, processing of continental-scale observations into useable
land cover change products is not currently routine or systematic. The MODIS
land cover and land cover change products are providing systematic but coarse
observations at continental scales, but the linkages between MODIS land cover
change estimates and carbon stock changes have not yet been sufficiently
developed or tested to evaluate capability to provide carbon flux estimates.
Additional remote sensors are becoming operational to provide more direct
estimates of above ground biomass stocks (e.g. LIDAR and RADAR from both
airborne and space platforms). Hyper-spectral airborne measurements may be
useful for distinguishing between living and dead biomass. These technologies
hold promise for estimates of carbon inventories at continental scales, but the
resources to support systematic observations are not currently available.
Several specific needs have been identified for improving estimates of carbon
stocks with the help of remote sensing products: (1) timely systematic and
routine processing of satellite data from the North American Continent into land
cover and land cover change products, including both natural and human
disturbances (2) integration of satellite observations with in situ measurements of
carbon stocks and existing inventories (3) augmentation of satellite and in situ
estimates of carbon stocks with airborne and surface measurements (4)
development appropriate estimation models.
Additional details about remote sensing contributions to the NACP are included
in the Appendix.
Extensive Inventories
Current large-scale land inventories conducted primarily by USDA (FIA and NRI)
employ multi-tier sampling strategies involving remote sensing and ground
measurements. These continuous inventories provide baseline information
about land cover, management intensity, productivity, and disturbance that can
be used to estimate carbon stock changes over 5-10 year periods. A very high
sampling intensity allows detailed description of some of the causes of observed
carbon stock changes, such as the effects of vegetation growth, mortality, and
harvesting. Historical data are available to trace land use and management
history.
The ability of current land inventories to provide true monitoring of C stock
changes and estimates of C flux is limited in several key ways: lack of complete
ecosystem C measurements; lack of sufficient temporal resolution; and lack of
easily available and usable historical data. Models based on data from
ecosystem process studies and intensive monitoring sites are used to fill in data
gaps, but uncertainty is high.
Gaps in spatial coverageMajor gaps in the U.S. include some “reserved” areas of the U.S.; lightly sampled
areas of the Intermountain West, the Pacific Coast, and Alaska; developed lands
especially urban and suburban areas; and large areas of public nonforest land
(mostly grazing land in the U.S. West). Large areas of Canada and Mexico, all
cover types, have been hardly sampled at all with field plots, or if sampled, the
data is currently inaccessible. Although enhancements to ongoing inventories
are filling some of these gaps, especially in forested areas, it is unlikely that
these improvements in coverage will be fully implemented with repeated
measurements during the early stages of the NACP. Therefore an interim
strategy is needed to increase the use of current and historical remote sensing
data to identify land cover status and changes, coupled with selected new field
measurements to estimate biomass and other ecosystem C stocks and rates of
change for undersampled areas.
Sparsely measured C poolsEnhancements to measurements in all major cover classes include stumps, live
and dead roots, mineral soil, litter, and coarse woody debris. Limited
measurements of complete ecosystem C stocks and fluxes are available from
intensive sites, and some pilot efforts are underway to modify extensive
inventories, but representative spatial coverage is spotty. An aggressive field
campaign early in the NACP is required to collect data on these poorly measured
C pools particularly in areas that have been poorly represented in past data
collection efforts. This new information will facilitate the development of
ecosystem carbon budgets that represent the major conditions on the landscape,
both disturbed and undisturbed. Some new direct measurements of C fluxes will
be needed during the intensive field program and at a new medium-intensity
sampling network (discussed below). These new measurements may include
litter production and soil CO2 and CH4 fluxes, which can be combined with
remote sensing estimates of land cover and land cover change to provide
validation data for ecosystem flux estimates from atmospheric measurements
and for input to model estimates of the North American C budget. The methods
to convert relatively simple allometric field measurements to estimates of mass
need improvement. Estimation methods (mathematical models) will be
developed to relate the quantity of C is different ecosystem C pools to the
measurements typically taken with the extensive inventory system, such as tree
diameter, height, and size of dead wood. The applicability of biomass equations
will be reviewed and supplemental equations derived from new field studies if
necessary. Special studies will be required to improve estimation of some C
pools, for example, the density of decaying wood for various species and
regions. Improved estimates of the translocation of harvested agricultural and
forest products are needed both at the national scale (exports) and for regional
studies in order to match the land accounting with the atmospheric accounting for
the same regions.
Improved temporal resolution –
Current land inventories and estimation methods have been carried out to
provide a “rolling average” estimate with a temporal resolution of 5-10 years,
which is sufficient for many applications, but incompatible (though
complementary with) with the higher temporal resolution attainable from
atmospheric measurement approaches. New designs for forest inventories
involve continuous sampling of all land areas using successive sample “panels”.
Each panel is sampled with a re-measurement period of 5-10 years. It is feasible
to develop and apply advanced statistical techniques to estimate annual changes
in C stocks from sample panels, if supplemental data are used to estimate the
major causes of variations in C flux: productivity, mortality, harvest, and land use
change. Attention to integration of disturbance monitoring from satellite data is
critical to improving the time resolution of estimates from inventories. Sources of
higher temporal resolution data include flux towers (productivity and trace gas
dynamics), aerial and satellite disturbance surveys (land use change, damage
and mortality), and timber and agricultural product surveys (harvest quantities).
Each of these sources of information can be related to the sample panels
through statistical models and appropriate consideration of the differences in
what is measured.
A New Intermediate Sampling Tier
A new medium-intensity sampling tier is proposed to fill in the gap between the
extensive, representative sample from land inventories, and the incomplete
network of intensive research and monitoring sites. This new sampling tier is
similar to the partially implemented “phase 3” of the U.S. forest inventory, but
would cover all land types and include specific measurements of C flux needed
for spatial interpolation between existing extensive and intensive sites.
The number of new medium intensity sample sites needed is approximately 1000
– this estimate will be refined during pilot phases of NACP implementation.
Sample sites may be selected from some combination of two sources: a
subsample of existing extensive network sample sites, or a set of completely
independent sample sites that could be drawn from existing research sites. The
advantage of the first approach is that fewer sites would be needed to attain the
same accuracy of population-level estimates. But there may be disadvantages,
for example, access to make the repeated measurements may be prohibited by
current inventory programs. The advantage of the second approach is that
existing infrastructure and measurement programs may provide a high degree of
leverage as well as facilitate involvement of many collaborators.
Key measurements at the medium-intensity sample locations include (1)
representative measurements of soil C flux, a response to soil respiration which
is strongly regulated by climate and disturbance, and accounts for about 1/3 of
the total land/atmosphere C flux; and (2) methane flux, and important
greenhouse gas particularly in peatlands, wetlands, and agricultural systems. IN
addition, these sites could be sampled more frequently for variables such as
litterfall that vary during the year, and can become sites where special studies
can be conducted to fill in data gaps for estimating ecosystem C flux, such as
decay rates of downed dead wood. The new sites will have value for
parameterizing or validating the predictions of ecosystem process models. The
new sites may also be useful as sites where the impacts of natural disturbances
on ecosystem processes can be studied relative to a previously established
baseline C budget.
Intensive Research and Monitoring Sites
The North American Carbon Program needs estimates of terrestrial-based plantsoil-atmospheric fluxes and C stock changes, from the diversity of our
ecosystems, in order to provide long-term, independent checks on carbon
budgets derived from atmospheric methods, to help interpret atmospheric and
satellite measurements, and fill in data gaps from extensive inventories.
Flux tower sites –
Net ecosystem CO2 exchange is presently measured at more than 30 sites in
North America that are part of the AmeriFlux network (supported largely by DoE
through the Terrestrial Carbon program and the National Institute for Global
Environmental Change), with additional sites planned in the Flux-Canada
network. Summed over the course of a month, season or year, data from these
sites provide direct measures of ecosystem CO2 source or sink strengths. In
contrast to the network of tall towers described earlier, most of the flux towers are
“small” (<60 m) and provide information specific to one ecosystem type.
The data from the flux sites have attracted strong interest beyond carbon cycle
studies, with applications in ecology, weather forecasting, and climate starting to
appear. Many flux sites now have several years of data and are beginning to
quantify inter-annual flux variability. Companion physiological and ecological
measurements help to partition carbon fluxes into plant and soil components and
reveal the mechanisms that are responsible for these fluxes. At some sites,
biomass-based estimates of C storage have been made that validate
micrometeorologically-derived tower estimates (e.g. Curtis et al., 2001). Many of
the flux sites are involved in testing or developing physiological models of C
exchange, relating fluxes to remotely sensed ecosystem characteristics, or other
specialized studies.
Several enhancements are required for this network:
The present network of flux sites must be augmented in both capacity and
number to achieve the goals of the NACP. Sites that cover under-represented
ecosystems, land use history, and current management are needed, including
actively managed cropland, forest, pasture and arid ecosystems, as well as
studies in Mexico and northern Canada.
Data will need to be transmitted rapidly to an available data center. This
development will enable the measurements to be used in data assimilation
activities (see below).
Flux sites will need to carry out continuous, high-precision measurements of
atmospheric CO2, CO, and CH4 concentrations. Presently only a few sites do
these measurements, which provide continuous datasets representing
covariances among the key species in a region.
Flux sites need to include a consistent suite of biophysical measurements to
characterize the sites for synthesis and for linking with other measurement
systems. Some important measurements include biomass stock, species
composition, soil C stock, and annual NPP.
Network structure should be strengthened. The AmeriFlux network has grown on
an ad hoc basis with individual sites funded by a variety of agencies and
programs. Coordination is voluntary, and consists of standard and
recommended measurements, data handling, adherence to quality control
procedures, and deposit of data with CDIAC. A more formal structure, with
defined site selection, QA/QC, review, and analysis procedures, would appear to
be desirable.
Because about 1/3 of North America consists of topography that is too complex
for eddy flux measurements, gaps will be filled in through remotw sensing and
modeling techniques. This topic is addressed more fully under the “data
assimilation, analysis, and modeling” section.
NSF-LTER (Long-Term Ecological Research) sitesLTER sites also have the necessary background monitoring and scientific
capability to contribute to understanding terrestrial C exchange. Some include
unique measurements such as dissolved organic C and particulate C losses.
Existing long-term agricultural experiments provide another major resource. The
present CASMGS (Consortium for Agricultural Soils Mitigation of Greenhouse
Gases) program involves an extensive number of USDA and University
Experiment Station long term sites. This program is focused on soil management
decision-making issues but provides another excellent resource as do ongoing
FACE experiments. Enhancement of the instrumentation on a number of these
sites with tower- and chamber-based flux measurements of CO2 and CH4 fluxes,
isotopic and plant physiology-soil process measurements to support
understanding and modeling the controls on carbon cycling, are envisioned as
parts of the NACP.
New intensive observation sitesSome new intensive observation sites will be required in currently
underrepresented regions and ecosystem types, for example, non-forest noncrop lands, peatlands, wetlands, and coastal marine areas. The networks of
available intensive research and monitoring sites need to be reviewed to identify
major gaps in coverage. For example, there has been a tendency to locate
observations in relatively undisturbed sites, whereas the contribution of disturbed
sites to ecosystem flux is much larger given the extent and impact of disturbance
on ecosystem processes. Filling these information gaps is likely to be one of the
most solid and long-term contributions of the NACP to ecosystem science.
The criteria for selecting new observation sites will need to be carefully
developed. Some initial considerations include (1) a history of collaboration and
ability to work across scales, (2) consideration of the needs of atmospheric
modelers, (3) data consolidation and management capabilities, (4) need to fill
critical gaps in coverage.
Measurements needed at intensive sitesIn addition to the continuous net ecosystem fluxes, chamber measurements of
CO2 flux and a suite of soil C measurements will be needed. Measurements
would include baseline soil C stocks along with diagnostic soil organic C
fractions, radiocarbon, 13C natural abundance and soil respiration. These can be
used to estimate soil organic C turnover, which together with data on net primary
production and C inputs to soil is necessary to estimate net soil C fluxes.
Additional discussion of respiration and soil CO2 flux measurement issues is in
the Appendix.
Measurements of dissolved organic C and particulate organic C are needed at
least at a subset of sites, because these translocations of C may be significant in
some ecosystems, and do represent a significant input of C to oceanic and land
margin systems.
All locations should be georeferenced to allow re-measurement at some time in
the future for independent determination of system net C change and to link with
remote sensing estimates of productivity and land use/land cover change.
Methane (CH4) is also important to measure at some intensive sites – this topic is
discussed in more detail below.
Soil Respiration Issues
One of the most challenging aspects of the NACP will be the estimation of soil
respiration fluxes. The efflux of CO2 from soils includes both plant respiration
(from live roots) and microbial heterotrophic respiration. Together, these CO 2
sources may equal or exceed NPP. Determining the net release or uptake of
carbon by the ecosystem or biome therefore requires not only careful
measurement of soil CO2 efflux, but also evaluation of the specific processes that
cycle and transport belowground carbon. The estimation of soil respiration may
be the weakest component in our ability to understand temporal and spatial
trends in the terrestrial carbon cycle.
Whereas spatially integrated estimates of NPP can now be developed in the
context of robust mechanistic models beginning at the leaf level, there is no
comparable framework for understanding the belowground processes that control
soil respiration fluxes over the range of temporal and spatial scales of interest to
the NACP. Previous work has delineated soil temperature and moisture as
important factors controlling rates of soil respiration. These influences are
mediated by heterotrophic substrate supply, soil texture, landscape setting, and
physiological constraints on microbial and root metabolism. Quantification of
these controls is often based on measurements that are necessarily very local
and short-term. Extrapolation in space and time is extremely difficult. Existing
local studies employ a diverse range of techniques and protocols that are often
difficult to compare. Soils are very heterogeneous – both laterally and vertically –
and soil properties are not readily accessible by remote sensing. Maps of some
soil properties are available, but they cannot yet be related systematically to soil
respiration fluxes. There is no widely accepted mechanistic model linking the flux
of CO2 at the soil surface to belowground soil properties and carbon cycling and
transport processes.
To address these problems, we will need to determine how measurements of
CO2 fluxes at the soil surface can be related both to estimates based on
atmospheric data assimilation (including eddy correlation) and to carbon
dynamics and soil properties below the soil surface. This effort will require
particular attention to development of methods and protocols for soil
measurements and their interpretation at intensive and intermediate study sites.
Significant improvements will be required in our ability to monitor soil-surface and
belowground processes and to assure the comparability of measurements from
place to place and time to time. The optimum strategy will likely combine
selective soil respiration measurements with soil descriptive data (including
texture, organic matter content, and landscape setting), dynamic biophysical data
(including soil temperature and moisture content, as well as land management
activities), and new models linking belowground processes and CO2 fluxes to
ecosystem function (controlling substrate supply and metabolic rates). This
strategy will be particularly important to estimating overall CO2 exchange in areas
of rolling or mountainous terrain, where estimates based on atmospheric
measurements are exceptionally difficult.
Methane-related Issues
Methane is of particular interest because of its importance in the radiation
budget. A molecule of CH4 contributes about 20 as much as a CO2 molecule to
radiative forcing, over a 100 year timescale. It is also a key species in the
chemistry of the atmosphere. Methane has been the most rapidly increasing
greenhouse gas, rising 145% since the beginning of the industrial revolution.
After years of near–steady growth rates, ~12 ppbv/yr in the 1980’s, growth rates
became highly variable, up to 15 ppbv/yr in 1991 and 1998, but 0 ppbv in 2000.
The changes in growth rate for atmospheric concentrations are not well
understood, and we cannot confidently predict future increases or decreases.
Over 70% of CH4 emissions are anthropogenic, dominated by biogenic sources
(e.g. landfills, domestic sewage, rice agriculture, ruminants, animal waste), with
X% associated with fossil energy production and use. The agricultural sector
accounts for ~50% of the human-induced CH4, and ~30% of total methane
emissions in the US. Agricultural sources of methane include concentrated (e.g.,
feedlot) and diffuse (nonpoint source) emissions, which are affected by
production practices such as applications of water, fertilizers, and manures.
Precise determination of agricultural methane emissions is needed to quantify
the North American and global carbon budgets.
Natural wetlands account for more than 20% of the global CH4 source, largely
from northern peatlands and tropical wetlands. CH4 exchange from these
environments is intimately linked to hydrology, system productivity and carbon
accumulation and balance. At the regional scale CH4 emission for many
landscapes in N. Amermica are dominated by natural sources (termites,
wetlands, lakes and coastal waters). Different mixes of anthropogenic and
natural sources and sinks determine the net fluxes in different regions. For
example, in New England northern peatland sources dominate CH4 emissions in
Maine but landfills and energy use dominate in Massachusetts and south.
The NACP atmospheric measurements should be complemented by surface
observations at representative sites to enable optimal evaluation of source/sinks.
The challenge is to insure that the combine atmospheric and surface
measurements quantitatively resolve the major elements that produce the net
flux in order to obtain accurate assessments of fluxes and their feedbacks.
Importantly, identifying the sources and developing long time series of
observations (a stated aim of the program) will determine how well we will be
able to develop models that quantify the variability and resolve processes at
inter-annual to decadal timescales.
Data Assimilation, Analysis and Models
The Challenge
In the section on Atmospheric Data Analysis, Modeling and Data Assimilation
(Element 2), we put forward a vision of an integrated data assimilation
framework, built on closely coupled, data-driven models for the atmosphere, soils
and plants. The outline of the atmospheric components has been given above.
This section outlines the development of the soil and plant components.
The core problem is to define a vegetation-soil-biogeochemistry modeling
framework that can interface optimally with input data from diverse sources. In
principle this model could be as complex as nature itself, with countless
parameters. The challenge is to develop a model whose parameters can be
constrained by observations, focusing on quantitatively defining those processes
that regulate the key emergent properties of the ecosystem (fluxes, stocks,
structure) on the relevant time scales (hours, years, decades). Models within this
framework will be used diagnostically, to analyze observations from the NACP by
conventional or full-system assimilation methods, and will function prognostically
when linked to climate models for future scenarios.
The elements must include improved estimates of carbon stocks in soils and in
natural and managed vegetation, as noted in the previous section. The roles of
prior disturbance and land use history, nutrient limitation and inputs, pollution
effects, extreme meteorological events, chronic and acute stress and herbivory,
and invasive species all feature prominently as mechanisms in recent US-wide
estimates of carbon fluxes. Considerable thought should go into the design of
this new class of model, and into the observations and manipulations that test
mechanisms operating over both long and short time scales. This type of
information is needed both to complete the link between process and
atmospheric observation and to select sites and analyze results from the NACP
observing system.
Data Needs for Biophysical Models
Recent studies have highlighted the important underlying roles of historical land
use change and management intensity in explaining current observed terrestrial
CO2 fluxes. Important variables include current age class distribution, past land
use, past disturbance, management intensity (e.g. plantation or cropping
practice), and exposure to air pollution and deposition. An important activity that
will support many aspects of this project is to assemble historical data into a
spatial database for use by various estimation and model elements, so that a
common land history can underlay all of the various flux estimates. A project to
develop a prototype of such a database has begun and will provide guidance for
the larger effort needed to support this proposal. Special attention will be given
to provide detailed historical information about selected intensive study areas.
The historical data provides a starting point for a dynamic database, updated
using the combination of remote sensing and enhanced inventories outlined in
the previous section. The goal is an accurate, high-resolution, time-varying
geographic data base containing the following kinds of information:









Land cover, land use, and management intensity
Current vegetation type, community structure, age classes, LAI
Vegetation biomass, live and dead
Soil physical properties (texture, water/thermal capacities, etc.)
Soil chemical properties (carbon, nitrogen, phosphorus pools in organic
matter; inorganic C)
Topography and geographic boundaries
Climate (Temperature, precipitation, humidity, wind, etc.)
Atmospheric deposition (ozone, N)
Natural disturbances (wildfire, insects, weather)
These data must reflect disturbance and land use history in order to make
meaningful NEE and trace gas flux estimates. The size of the soil C and N pools
and forest stem C directly influences the magnitude of autotrophic and
heterotrophic respiration computed and so impacts the accuracy of the final CO2
balance considerably. In many cases this detail determines whether the CO 2 flux
is a source or a sink.
Biophysical/biogeochemical Models
Biogeochemical models simulating fluxes of mass (CO2, H2O, CH4) and energy,
productivity, respiration, and effects of disturbance will be linked to three sources
of data that will drive models: (1) The link to the map of land cover/land
use/inventories establishes the components (stock, vegetation, soils) at each
gridpoint. (2) The link to remote sensing and in situ biophysical data provides the
current state of the system (sunlight, soil moisture, phenology, recent events
such as drought, wind, ice). (3) The link to atmospheric boundary conditions,
through the assimilated meteorological products, drive the system processes,
given the current state and recent history of the system. Linked models within
this framework can be used in many ways—to analyze small-scale process
studies, to generate forward model results to compare with observations, to
provide enhanced priors for inverse models, and to provide the land-surface
component of the full assimilation approach. One expects in general that the
complexity of these models will decrease as the scale of application increases,
but it is very advantageous to maintain a consistent conceptual framework,
especially the strong links.
The land surface biophysical model will be ideally driven at time steps of ~ 1 hr to
provide temporally and spatially complete estimates of surface CO2 and CH4 flux,
to compare against the aircraft transect data. These models must first accurately
compute energy and water balances under all vegetation, climatic, and
disturbance regimes represented on the continent and coastal ocean. The model
then must compute hourly CO2 and CH4 inputs and outputs, i.e. photosynthetic
uptake and autotrophic and heterotrophic respiration emissions of CO 2 and net
CH4 exchange across different landscapes, again under all vegetation/climate
combinations on the continent. Nutrient cycling and other long-term factors (e.g.
vegetation structure, soil organic carbon, permafrost) must be carefully treated in
these models so that they can evolve over time and provide regulation of the net
exchanges.
The calculations will best be accomplished with an "ecological data assimilation"
approach. At each iteration, updates from ongoing satellite and in situ
observations are used to calibrate the model to ongoing monitoring data. In situ
data will be from AmeriFlux sites, and similar installations developed for the
NACP, where fluxes and forcing factors are measured, and vegetation and soils
monitored, continuously. The data will have to reach a central location in a short
time. This type of procedure is already being attempted in a pilot study for
evaluation of MODIS data products, and on-line data transmission has been
quite successful (weekly downloads). There is clearly much work to be done
however to obtain daily or hourly downloads, and the uniformity and reliability of
the AmeriFlux data products would require significant enhancement.
Some specific remote sensing data is needed and will be available by 2004:







1km snowcover
1km albedo
1km surface evaporation resistance for energy partitioning
LAI and FPAR
1km GPP, for defining regional gradients and phenology
1km fire area coverage and plume dispersion
Surface moisture/wetlands delineation/drainage class
Currently the best attainable spatial and temporal resolution BGC modeling is
daily at 10km resolution, or hourly at 50km resolution. Improvements are needed
to run BGC models hourly at <10km resolution. This will be possible by 2004,
and will also depend on whether the continental U.S. or all of North America is
the study domain. The bottom-up models need to produce an hourly estimate of
surface CO2 flux, at about 10x10km resolution to have compatible spatial
resolution with the aircraft data.
The ongoing driving variables are hourly meteorology from surface weather
observation network and will require spatial interpolation and gridding to the
necessary spatial resolution.
At least one third of the North American continent has topography too complex
for eddy flux towers or low altitude aircraft measurements. These mountainous
regions also are heavily forested , so may have a disproportionate carbon sink
potential. In order to produce regular NEE estimates for these regions without the
flux measurements, we plan a multi-step procedure. Remote sensing data will
cover these areas, providing landcover and a weekly GPP estimate from MODIS
as a first approximation. Daily surface meteorology observations from NWS can
then be extrapolated from topoclimatology principles such as elevational lapse
rates and aspect defined incoming solar radiation to provide a 1km mapping of
the complex surface microclimate. Photosynthesis, auto and heterotrophic
respiration can be computed if initial landcover and carbon content of the
landsurface has been pre-defined. These model based estimates of NEE in the
mountains may best be tested using gauged watersheds to estimate hydrologic
fluxes, and biomass inventories to validate carbon fluxes. Although this
procedure is not as direct a test of surface NEE as the flux measurements, it can
provide complete and consistent NEE estimates for all mountain areas that can
resolve topographic complexity to any resolution desired.
Key model outputs include:




GPP, NPP, NEE, and CH4 flux
Respiration components
Water, Energy balances, hourly and daily.
Albedo, roughness, etc
The final latent heat, NEE and CH4 surface fluxes link to the atmospheric models
to generate atmospheric concentrations comparable to the aircraft data.
Complete Carbon Accounting
All carbon sources and sinks must be fully accounted for. Although this program
is focused on exchanges between the land, atmosphere, and oceans, C may be
transported in or out of an analysis region through erosion/sedimentation or
product harvest (crops, timber). Estimates of C flux for a region may be
misleading unless all sources and sinks are counted. Atmospheric estimates
must be carefully matched with complementary accounting for fluxes on the land.
Deposition or mobilization of inorganic C as carbonates can play a significant role
especially in arid and semi-arid soils, and must be considered in C transfers
between land, atmosphere, and oceans.
Data Management
Many of the necessary data streams have begun today but have not yet been
produced consistently at the time/space resolution needed. Moreover, these data
streams have never been assembled into the integrated modeling package
needed, but could be by 2004. Because of the diversity of data being collected
and generated by models, at multiple temporal and spatial scales, it will be a
continuing challenge to make it available and useful to the inverse/data
assimilation activities and, eventually, to the public.
The Role of Process Research
Process research studies are critical for defining key mechanisms responsible for
carbon exchanges between the atmosphere, oceans, and land. Known and
potentially unknown mechanisms can be identified prior to initiation of major
monitoring activities to guide selection of variables to measure. While
experimental research studies are outside the scope of this monitoring program,
there should be a strong linkage especially considering that the effects of some
mechanisms, such as the effect of increased atmospheric CO2 on C allocation,
may be difficult if not impossible to measure directly. Process research studies
will have direct input into the BGC models.
Intensive Field Campaigns and Phased Approach
There are three broad, consecutive phases to implementation of the biophysical
measurements and modeling part of this program, corresponding to the phased
approach described in the section on atmospheric measurements: (1) preliminary
work, (2) intensive field studies, and (3) long-term implementation. Preliminary
work is to be completed prior to initiation of this program using existing funding
mechanisms.
Phase 1: Preliminary Work
Network design issues – Existing monitoring programs, previous large-scale field
campaigns, and ongoing large-scale research projects may all provide some
guidance for developing the specific design of a North American program. It is
critical to use whatever relevant information is available to speed the attainment
of an efficient monitoring system. It is also important early in the program to
develop the criteria for selecting new intensive observation sites and enhancing
old sites.
Enhancements to extensive inventories – The existing extensive monitoring
networks include gaps and overlaps in geographic coverage. The NACP needs
to work with these programs to identify areas that can potentially be added or
reconciled to provide a complete wall-to-wall coverage of the land surface.
Identified gaps include developed lands (urban and suburban), rangelands,
pinyon-juniper forests, interior Alaska, and West Texas. In addition, an effort is
needed to compile a prototype data base of historical information, including the
current state of land surface (cover, age since disturbance, type of disturbance,
management, biomass, soil carbon stocks). Landsat data collected over the
continent for '80s, '90s, '00 and '02 (in preparation for campaigns beginning in
'04) need to be assembled and processed to provide high spatial resolution land
cover change information for the last two decades.
Measurement issues and intensives: field tests and deployment of new
instruments – New methodology for in situ soil C measurements are becoming
available. They must be tested, and if suitable, used for more background and
benchmark validation measurements. In general, the required biophysical
measurements now occur at only a few sites. Many sites do not have the
instrumentation required to supply the data at appropriate time scales to integrate
with the atmospheric studies. We need to augment capabilities to allow the
measurement of all C gas fluxes and to partition the component fluxes at
resolutions that will allow the quantification of daily, seasonal, and inter-annual
variability at appropriate spatial scales. New low power, high precision, high
frequency detectors for CO2 (ndir), CH4 and CO (GC or ndir-gfc) need to be
developed and deployed as part of a low cost chemical meteorology package in
order to expand our measurement capacity. Related to all proposed
measurements, definitions ans standards need to be developed to prepare for
future implementation.
.
Model development and applications – As far as possible, analysis needs should
be identified early in the program so that model development may provide
appropriate information. Input and output variables needed by a variety of models
need to be specified. Model comparison studies may be useful. Testing of
methods to integrate new measurement protocols with existing intensive and
extensive sampling networks is an ongoing activity.
Evaluation of information capabilities of new remote sensors – Some remote
sensors have been in use long enough that their capabilities are well known (e.g.
Landsat-TM). Others such as EOS-MODIS are relatively new and evaluation
studies are underway. In addition, above ground biomass measurements using
airborne LIDAR, RADAR and hyperspectral techniques should be evaluated
further through field testing. Once a sampling strategy has been identified these
approaches can be used in conjunction with in situ measurements and modeling
to develop continent-wide data on current carbon stocks, fluxes, and changes in
stocks and fluxes.
Phase 2: Intensive Field Studies
Network design issues – Proposed designs from the preliminary phase will be
rigorously tested and evaluated in field conditions. Particular attention must be
given to integrating the designs with atmospheric measurements.
Enhancements to extensive inventories – compilation of historical data and
support for intensive studies is required at this phase. It will be necessary to
specifically test and evaluate field procedures to attain accurate and seamless
extensive data collection. All relevant historical data should be compiled and
available at the end of this phase. A major effort is required to address data gaps
from extensive inventories by applying substitute methods based on remote
sensing, alternative plot networks, and ecosystem models. This supplemental
data should be fully compatible with existing inventory data except that the
sources of information will be different.
Measurement issues and intensives -- Support is needed for the initial intensives
of the atmospheric measurements “proof of concept” exercise (agricultural and
urban test areas), and subsequent phasing of atmospheric measurements over a
variety of cover classes. In addition, intensive field studies will be needed to test
the measurements and design issues for the new medium intensity sites in
representative cover classes. One or more intensive field studies will be needed
in agricultural, urban/suburban, forest, wetland, and rangeland cover classes. If
possible, such studies should be added to areas where significant intensive
process monitoring is already underway, such as areas with AmeriFlux towers
and/or LTER monitoring. This will make it possible to provide data on ecosystem
response to climate change and to supply the required ground truth for a
representative selection of sites.
Modeling and analysis – BGC models, prototype analyses, and data
management systems must be developed in conjunction with the intensive field
campaigns.
Remote sensing activities during intensive field campaigns include:
 monitor disturbance (e.g. fire, forestry, agriculture) during campaigns using
satellite observations in combination with in situ measurements
 monitor phenology over the continent precisely (<= weekly, <= 1km).
 airborne remote sensing of state of vegetation: biomass, stress, and foliar
chemistry using LIDAR, RADAR, hyperspectral, multiangle techniques
 during-campaign data base of biophysical state (lai/fpar, soil moisture,
meteorological conditions, inundation, disturbance such as fire, logging,
other)
 compile appropriate descriptive historical data
Phase 3: Long-term Implementation
Based on preliminary studies and intensive campaigns, modifications to longterm networks will be proposed to broadly enhance ability to monitor fluxes of
major C species, and controlling plant-soil characteristics and processes, for
North America. The full four-tier monitoring system will be operational. Key
limitations described earlier will be resolved: gaps in spatial coverage will be
filled; complete ecosystem C stock changes will be estimated; and temporal
resolution will be high (annual to monthly). New comprehensive data and
analysis tools will facilitate development of predictive models to evaluate policy
scenarios for managing greenhouse gases. Near-real-time, quality controlled
data will be delivered to the sites for the data assimilation activities.
Summary of Deliverables for Plant-Soil-Atmosphere Measurements and
Models





Data bases
o Historical information (various scales)
o Current trends in carbon gas dynamics
o Remote sensing estimates
o Intensive site monitoring data
Monitoring instruments and techniques
o Soil-plant complex
o Direct measurements of atmosphere-biosphere exchange
o Biomass
o Disturbance and land use patterns
o Land cover change
Infrastructure
o New intensive monitoring sites
o Data management and near-real-time dissemination
o Assessment capability
Improved models and analytical tools
o Biophysical/biogeochemical models of processes and
controls
o Statistical methods for correlation, extrapolation, and quality
control
o Integrated models with sufficient resolution and scope for
use as policy tools
Improved C flux estimates
o Continental
o Regional
o Ecosystem