THE DUAL CONSTRAINT STRATEGY (FIGURE 3) IMPLIES A RANGE OF OBSERVATIONS REQUIRED TO PROVIDE INFORMATION ON THE SPATIAL AND TEMPORAL DISTRIBUTION OF CARBON SOURCES AND SINKS. Various groups have previously considered these requirements. Similarly to the information requirements, the workshop employed a three-step process to converge on an integrated set of requirements:
Review of observation requirements by existing programmes;
Synthesis of bottom-up and top-down requirements;
Identification of gaps and observation issues.
Various observation requirements for existing programmes were considered (refer to Appendix 3): global programmes and international conventions (Appendix 3 Cihlar and Brown, Gommes), ecosystem modelling at national to global levels (Appendix 3 Ahern; Chen and Cihlar; Potter; and Wickland); and atmospheric modelling (Appendix 3 Raupach, Gerbig et al.) perspectives.
In principle, atmospheric inversions require two types of observations. The first type are those needed to characterize and model the behaviour of the physical climate system. These are similar to the observations and data sets employed in numerical weather prediction or in general circulation models, and presently various national and global observing systems acquire them. The second type, observations on atmospheric concentrations of CO2 and other gases are required to predict the spatio-temporal distribution of sources and sinks. The workshop did not consider these requirements in detail because they had been subject of special meetings (see e.g. Francey, 1997).
Requirements for bottom-up modelling were assembled in table format in a breakout discussion. In preparing the table, several factors were considered for each variable, with the intent to ascertain their observational implications:
Reasons for using the variable and its role in modelling carbon fluxes:
as a driver (thus always required, usually as a gridded data set), or for model validation (needed for a sample of sites/conditions);
as an input (needed for the computations), output (final result of the computations, where the role of the observation is to validate the final results), or internal variable (intermediate product of model computations, where the role of the observation is to check internal model consistency);
The required spatial resolution of the observation;
Measurement method: in situ, remotely sensed, or modelled.
Table 1 contains the list of observation requirements required for the bottom-up approach. The 'Type' column characterizes the nature of the variable as external forcing (thus observations are needed as model input), internal status (as input or for model validation), or output variable (for output validation). The 'Spatial' and 'Temporal' columns refer to the desired spatial coverage of an observation. The 'Method' column describes the expected approach to obtaining the result: through in situ measurement, remote sensing, inventory, or modelling.
It should be noted that the observations required in each setting need consideration not only individually, but also in relation to one another. For example, eddy flux measurements should be associated with suites of ecological measurements, and for the most part should not be done in isolation.
Table 1. Observation requirements for bottom-up approach
|
Variable |
Type (a) |
Spatial (b) |
Temporal (c) |
Method (d) |
Comments |
|
1. DRIVING VARIABLES (for model application/upscaling, required at every grid point) |
|||||
|
ATMOSPHERE |
|
|
|
|
|
|
Air temperature |
1 |
3 |
1,6 |
1,2,3 |
daily maximum, minimum, mean |
|
Precipitation |
1 |
3 |
1,6 |
1,2,3 |
|
|
Photosynthetically active radiation |
1 |
3 |
1,6 |
1,2,3 |
|
|
Relative humidity |
1 |
3 |
1,6 |
1,2,3 |
|
|
Wind speed |
1 |
3 |
1,6 |
1,2,3 |
|
|
Net radiation |
1 |
3 |
1,6 |
1,2,3 |
|
|
Snow water equivalent |
1 |
3 |
1,6 |
1,2,3 |
|
|
Aerosols |
1 |
3 |
1,6 |
1,2,3 |
for atmospheric corrections of optical data |
|
Integrated atmospheric water vapour |
1 |
1 |
6 |
1,2,3 |
for atmospheric corrections of optical data |
|
ECOSYSTEM |
|
|
|
|
|
|
Vegetation cover class |
2 |
1 |
4 |
3 |
physiognomic classes, dominant species (overstory, understory) |
|
Biota biomass |
2 |
1 |
4 |
3 |
may be used to drive decomposition models |
|
Soil moisture |
3 |
1 |
1 |
2,3 |
|
|
LAI |
2 |
1 |
4 |
3 |
|
|
Foliage N |
2 |
1 |
4 |
3 |
needed to drive decomposition rates |
|
Chlorophyll |
2 |
1 |
4 |
3 |
to drive canopy photosynthesis in some models |
|
Natural disturbance history |
1,2 |
1 |
4 |
1,4 |
includes biomass burning and insect-induced mortality |
|
Management history |
1,2 |
1 |
4 |
4 |
includes forest harvest, thinning, fertilization, etc. |
|
Topography |
2 |
1 |
3 |
3,4 |
influences radiation and surface water |
|
2. CALIBRATION/VALIDATION VARIABLES (required at selected sites) |
|||||
|
ATMOSPHERE |
|
|
|
|
|
|
Air temperature |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
Precipitation |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
Solar radiation |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
Relative humidity |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
Wind speed |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
Net radiation |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
CO2 concentration profile |
1 |
2 |
6 |
1 |
15 to 60 minute averages (continuous) |
|
Integrated atmospheric water vapour |
1 |
2 |
6 |
1 |
for atmospheric corrections of optical data |
|
Snow water equivalent |
1 |
2 |
1,6 |
1 |
15 to 60 minute averages (continuous) |
|
Aerosols |
1 |
2 |
1,6 |
1 |
15 to 60 minute averages (continuous; for atmospheric corrections) |
|
ECOSYSTEM |
|
|
|
|
|
|
SITE |
|
|
|
|
|
|
Natural disturbance history |
1,2 |
2 |
4 |
1,4 |
includes fires and insect-induced mortality |
|
Management history |
1,2 |
2 |
4 |
4 |
includes harvest, thinning, fertilization, etc. |
|
Topography |
2 |
2 |
3 |
3,4 |
influences radiation, and water fields |
|
Spatial pattern |
2 |
1,2 |
3 |
3, 4 |
may assist spatial scaling |
|
VEGETATION |
|
|
|
|
|
|
Vegetation cover class |
2 |
2 |
2 |
1 |
physiognomic classes, dominant species (overstory, understory) |
|
Root carbon |
2 |
2 |
2 |
1 |
coarse and fine |
|
Aboveground biomass |
2 |
2 |
2 |
1 |
stem, branch, foliage |
|
Leaf area index |
2 |
2 |
4 |
1 |
|
|
Foliage N |
2 |
2 |
4 |
1 |
used for canopy photosynthesis modelling |
|
SOIL |
|
|
|
|
|
|
Biota C, N |
2 |
2 |
4 |
1 |
may be used to drive decomposition models |
|
Biota biomass |
2 |
2 |
4 |
1 |
may be used to drive decomposition models |
|
Temperature profile |
1,2 |
2 |
4 |
1,2 |
profiles are useful as a driver and for process studies |
|
Maximum thaw depth |
1,2 |
2 |
4 |
1,2 |
critical for climate impact on permafrost-affected areas |
|
Thermal conductance |
2 |
2 |
3 |
1,2 |
to estimate heat transfer and heterotrophic respiration |
|
Thermal diffusivity |
2 |
2 |
3 |
1,2 |
related to thermal conductance but needs heat capacity information |
|
Soil moisture |
1,2 |
2 |
5 |
1,2 |
affects heat transfer and decomposition |
|
Hydraulic properties |
2 |
2 |
3 |
1,2 |
for vertical and horizontal water exchange |
|
Ground water table depth |
2 |
1,2 |
4,5 |
1,2 |
influences wetland dynamics |
|
Carbon content (org. and inorg.) |
2 |
2 |
3 |
1 |
directly affects heterotrophic respiration |
|
Carbon age |
2 |
2 |
3 |
1 |
needed to improve Rh calculation |
|
N, P content |
2 |
2 |
3 |
1 |
affects gross primary productivity |
|
Bulk density |
2 |
2 |
3 |
1 |
needed for diffusivity estimation |
|
Sand and clay fraction (%) |
2 |
2 |
3 |
1 |
|
|
pH |
2 |
2 |
3 |
1 |
important limitation to growth and soil biology |
|
Macro and micro nutrients |
2 |
2 |
3 |
1 |
these processes affect plant nutrient uptake |
|
Microbial biomass |
2 |
2 |
3 |
1 |
affects decomposition |
|
PHYSIOLOGY |
|
|
|
|
|
|
Foliage N |
2 |
2 |
2 |
1 |
needed to drive decomposition rates |
|
Foliage lignin |
2 |
2 |
2 |
1 |
needed to drive decomposition rates |
|
Chlorophyll |
2 |
2 |
2 |
1 |
needed to drive canopy photosynthesis in some models |
|
Rubisco |
2 |
2 |
2 |
1 |
needed to drive canopy photosynthesis in some models |
|
FLUXES |
|
|
|
|
|
|
Carbon fluxes (above and near ground) |
3 |
2 |
6 |
1 |
critical for model validation |
|
Aboveground NPP |
3 |
2 |
4 |
1 |
C storage flux |
|
Belowground NPP |
3 |
2 |
4 |
1 |
C storage flux |
|
Litterfall N, P, C |
2 |
2 |
2 |
1 |
C flux to soil & litterfall nutrients indicate nutrient availability |
|
H, ET (above stand) |
3 |
2 |
6 |
1 |
important for C flux estimation |
|
CH4 |
3 |
2 |
6 |
1 |
important for wetlands |
|
VOC |
3 |
2 |
6 |
1 |
can be significant in total carbon budget |
|
DOC |
3 |
2 |
2 |
1 |
C exchange can affect stocks and processes |
|
Heterotrophic respiration rate |
3 |
2 |
4 |
1 |
needed to validate NPP and NEP components |
DOC = dissolved organic carbon, VOC = volatile organic carbon
a: 1 = external forcing variable; 2 = internal status variable; 3 = output variable
b: 1 = gridded with a resolution of 1 km or better; 2 = one or more sites for each land cover class; 3 = gridded with a resolution of 0.5-1 degree or better
c: 1, since industrialisation with desirable frequency; 2, periodical measurement once every 5-10 years; 3, one-time measurement; 4: multiple-year continuous measurement; 5, daily in calibrations years; 6, continuous
d: 1 = site measurement (including characterization of its spatial heterogeneity as appropriate); 2 = modelling; 3 = remote sensing; 4 = existing survey or inventory
