The IGBP Transect Approach

Although new satellite images allow a more and more detailed observation of the earth surface (e.g., Defries et al., 2000), the ground truth remains essential. This was recognized by IGBP as a significant problem because it remains impossible to study processes with global coverage. Continental transects were suggested that represent the major climatic regions of the globe and allow repeated observations at the same time (Walker et al., 1999). Continental transects were suggested for the boreal region (Alaska, West and East Siberia), the temperate arid regions (USA, China, Europe), and the subtropical climate (West Africa, Patagonia, South America, Australia). It was recognized that land-use change and not climate is the main driver in the humid tropics, and thus the "transect" consists of different land use types in that region (Steffen et al., 1999).

The transect approach was just one of many solutions to cope with the problem of spatial variability of processes and their integration to larger scales. Figure 1 summarizes a whole suit of approaches that were used to understand and to integrate highly variable processes at the landscape scale (Schulze et al., 1999).

In the following we will present ecological process data, to demonstrate the variability or constancy of processes which may or may not correlate with vegetation, plant functional types or species.

2.1 The Patagonian Transect

A range of plant parameters were studied in Patagonia including vegetation types ranging from tall forest to desert (Fig. 2, Schulze et al., 1996). Only root biomass and density in the top soil decreased linearly with rain fall. All other parameters, either changed in a threshold manner (such as LAI) or remained constant. <513C


Gopvright 200! ■ Acklemic Press. AH rights of reproduction in Liny form reserveel.



Treatment Experiment


Driving variable

Re-evaluation of literature


Time series

Long-term studies

Continental transects


Natural C02 springs

Free-air carbon dioxide exposure

Open top chambers

Branch bag f

CBL flask C I sampling




Boundary layer

Convective boundary layer integration

Watershed (horizontal & vertical fluxes)

Eddy covariance flux tower

Leaf level chamber experiments


Species X


1 1



_L 1


Block I II III

Field experiments

FIGURE 1 A summary of approaches to integrate highly variable processes of ecosystems at the plot and at the landscape scale in the context of global change (Schulze et ill., 1998).

Patagonia 1995

Patagonia 1995

~ 2.2 The Australian Transect

Total blomass - [3

Root biomass

500 Rainfall [mm]

FIGURE 2 A range of plant parameters studied in Patagonia along a transect of decreasing rainfall. Vegetation types range from tall forest to desert (Schulze et al„ 1996).

Total blomass - [3

Root biomass

L 0

- 50

ONI '£





- 25

fc 0







- 0

- -30




- -20

O £_,

CO „

- -10 £



L 0

FIGURE 2 A range of plant parameters studied in Patagonia along a transect of decreasing rainfall. Vegetation types range from tall forest to desert (Schulze et al„ 1996).

was — 27%o independent of rainfall, and leaf nitrogen remained at 13 mg g-1. There was no distinct relation to vegetation type or functional types of the vegetation. Assuming that a correlation exists between biomass and carbon storage in soils (Schulze et al., 2000), it would seem to be impossible to infer from S,3C on the sink capacity of the underlying vegetation (Ciais et al., 1995). The constancy in t)l3C and leaf nitrogen concentration is caused by a change in species composition which maintains intrinsic water use constant at decreasing rainfall due to changes in leaf morphology, independent of productivity.

~ 2.2 The Australian Transect

The North Australian Transect extends about 300 km from Darwin to the interior of Australia. A study was undertaken (Schulze et al., 1998) in which this transect was extended to the higher (1800 mm) and lower (216 mm) rainfall regions along a transect extending to about 1000 km. The most striking observation was that the S,3C value remained essentially constant at — 28.1%o (Fig. 3). Only when rainfall decreased below 400 mm, an effect was seen on the carbon isotope ratio. Again, this conservative response of the carbon isotope ratio was associated with a decrease in specific leaf area, and this leaf property was associated with a fivefold change in leaf nitrogen concentrations, depending on plant functional types. Plant species that were potentially fixing atmospheric N2 had higher N concentrations and a higher specific leaf area than spinescent species. However, also the classification of plant functional types did not describe the functional process involved. The constant <5I3C value was associated with, and is most likely the result of, a change in species within each functional type; only when the diversity of species decreased to a single species in the dry region, the <5I3C values changed. At any one site, the local variability between species was as large as the continental variability along the transect.

The Australian transect extended from tall monsoon forest (21 m canopy height) to scattered dwarf trees (6 m height) in a subtropical grassland. We expect that the atmospheric isotope signal (not measured) would increase due to the increasing proportion of C4 grass photosynthesis, despite the fact that the total sink capacity of the region would decrease. This would make it very hard to infer from the isotopic signal on the sink capacity at the land surface.

2.3 The European Transect

Forest sites were selected across Europe ranging from North Sweden to Central Italy in order to study the effect of nitrogen deposition on ecosystem processes in coniferous and deciduous plantations. This study was designed such that the edaphic conditions were maintained as constant as possible, i.e., acid soils were chosen when available in order to detect effects of nitrogen. Figure 4 shows that the nitrogen concentration in needles and leaves were remarkably constant for conifers and deciduous trees, although Feigns sylvatica had a higher N concentration than Picea ahies. Also, the SI3C and <5I:,N concentrations were remarkably constant despite the large variation in climate and plant species (deciduous vs conifer). There was no obvious relation to NPP or leaf area index. Besides the fact that the carbon isotope ratios were increasing from north to south despite increasing NPP, the most remarkable observation was,

(} decid.+N7fixing 4 decid. -N2fixing A evergr.+N2 fixing A evergr. -N2fixing u ev. plantation -N2flx. ■ Aiiocasuarina -N2 fix. o Splnescent plants o Adansonia o-

O 15

too OTH


Latitude Latitude

FIGURE 3 The Northern Australian Transect: Latitudinal changes of specific leaf area, leaf nitrogen concentration, S"C-iso-tope discrimination and of the S'^N-isotope ratio in the plant functional types: potentially N2 fixing deciduous and evergreen trees (d. + N2, ev. + N2) and non-N2 fixing deciduous and evergreen trees (d. — N2, ev. — N2), spinescent species (spine), bottle tree Adansonia (Ad.), Aiiocasuarina (Allocas.), and evergreen cultivated fruit tree plantations (ev. pl.) (Schulze et al„ 1998).

that by selection of specific sites, the N concentration in the foliage was responding neither to N deposition nor to climate. However, at each site, the local variation in N concentration showed a range larger than concentrations along the whole transect. For instance, at the German "Waldstein" site, N concentrations in needles vary between habitats from 0.54 to 2.12 minol g~', while the whole continental transect varied between 0.5 and 1.1 mmol g-1. Again, inferring from isotope ratios or from N concentrations or LAI or C-fixation would be difficult, and it would not recover the local mosaic type variation. In addition, there was no significant difference between conifers and deciduous trees with respect to NPP despite the difference in foliage N concentration. The complex basis for the observed homeostatic response is explained by Schulze et al. (2000).

0 0

Post a comment