Spatial patterning of terrestrial ecosystems: assessment of the impact of Global Change on land cover distribution using satellite data
start on 01/10/2000_____end on 30/09/2003
Research done at:
Universitaire Instelling Antwerpen (UA-UIA)
= University of Antwerp-UIA campus (UA-UIA)
Department of Biology
Phone: +32-(0)3-820.22.56 ---------- Fax: +32-(0)3-820.22.71
Official language: Dutch
more information on this research:
The aim of this project is to assess the effects of human disturbance (e.g. deforestation, agricultural development, fragmentation) and global warming ("greening trend") on the global presence and distribution of terrestrial ecosystems.
Global land cover change is recognized as a major aspect of the large-scale environmental perturbations known as "global change". The drivers of this land cover change are (i) climatic change, caused by greenhouse gasses, and resulting in a global temperature increase (atmospheric component), and (ii) the expanding human population, increasing pressure on the remaining natural vegetation (biospheric component). Change in presence of ecosystems is expected to influence the functioning of the entire biosphere, including biodiversity decline, and impact on hydrology, climate, and global biogeochemical cycles (e.g. carbon cycle).
The objectives can be summarized as (1) to study global land cover change as a component of "global change"; (2) to develop quantitative measures to detect and describe land cover pattern; (3) to develop quantitative measures to evaluate the ecological consequences of pattern change; and (4) to validate the use of satellite data (AVHRR,MODIS) for "global change" studies.
In the short term, climate change can alter the mix of plant species in land ecostystems such as grasslands. In the long term, climate change has the potential to dramatically alter the geographic distribution of major vegetation types - savannas, forests, and tundra. Biomes are expected to shift to northerly latitudes ("greening trend"). Due to human disturbance, natural vegetation is replaced by agriculture or infrastructure (e.g. housing, industry). The presence of biomes, and their spatial arrangement (e.g. cumulative area, connectivity) are therefore expected to change.
The development of indices to detect and quantify pattern in the spatial heterogeneity of landscapes ismoreover necessary because of the influence of ecosystem patterning on its dynamics and functioning, e.g. by generating edge effects. A strong link is accepted to exist between ecological pattern and ecological function and process. Two types of pattern descriptors are considered. First, geometrical patch features like size, shape, perimeter and number are calculated, reflecting the general patterning of the present biomes. Second, indices describing the ecological consequences of theobserved geometrical pattern are calculated, e.g. the patch interior-to-edge ratio and the degree of isolation of patches representing similar habitats. These indices enable a more insightful interpretation of the observed spatial pattern. Twolevels of spatial pattern assessment are considered: the isolated patch and the entire landscape scene. Application and development of indices is dependent upon (i) the characteristics of the pattern of land cover, subjected to change, and (ii) the nature of the spatial proces transforming the land cover. Landscapes characterized by large continuous vegetations (e.g. amazonian tropical forests) can be altered by a different process (i.c. perforation) than fragmented landscapes, which are expected to be rearranged by patch dissection, shrinkage, and attrition. The use of a minimum number of indices to describe maximum variation of pattern change features, and the use of reference values to facilitate index value interpretation, will be explored. The effect of spatial resolution on the index calculation results is evaluated using spatial aggregation.
The starting point of the quantification methodology are the procedures developed for fragmentation measurement.
Remote sensing is accepted to be the most useful technique to monitor global land cover change, e.g. to quantify deforestation and fragmentation, or to detect changes of photosynthetic activity of terrestrial vegetation and the concomitant shift towards higher latitudes of vegetation occurrence. Two data sets will be used to asses land cover change, abbreviated as AVHRR data and MODIS data. The AVHRR data refer to the Advanced Very High Resolution Radiometers (AVHRRs) on board of the National Oceanic and Atmospheric Administration (NOAA) series of meteorological satellites (NOAA-7, -9 and -11). From daily observations of channel 1 (wavelengths 0.58-0.68 10-6 m) and channel 2 (0.72-1.1 10-6 m) reflectances, global sets of normalized difference vegetation index (NDVI) have been produced. NDVI data are strongly correlated with the fraction of photosynthetically active radiation (0.4-0.7 10-6 m) absorbed by vegetation, that is, to the photosynthetic activity of vegetation canopies. This data set has been used over the years to study vegetation dynamics, and is calibrated for intra- and inter-sensor variations, and partial atmospheric correction for gaseous absorption and scattering; stratospheric aerosol effects (volcanic eruptions, El Checon, Mount Pinatubo) were corrected. The AVHRR data are available from July 1981 to December 1999, and there is continued processing of data. Spatial resolution is 8 km; 15-days composite global images are available for land cover interpretation. The AVHRR data will be used to analyze the "greening trend", i.e. emphasis is laid on large scale (global) landcover shift processes. The following questions are addressed: How does the spatial distribution of biomes change during the shift toward higher latitudes? Can large-scale changes in ecological functioning be detected or derived? The MODIS (Moderate Resolution Imaging Spectro-radiometer) is the key instrument aboard the Terra (EOS AM-1) satellite; it was successfully lanched on December 19, 1999 (Vandenberg Air Force Base, CA). Terra's sensors will begin collecting first images roughly 30 days after launch (c. February 2000). The MODIS instruments represent the finest in engineering of space flight hardware for land remote sensing. The instruments cover 36 spectral bands which enables the calculation of enhanced vegetation indices; this improves detection of land cover distribution. Spatial resolution is as fine as 250 m. Temporal resolution is set to 8 days (global). MODIS is characterized by better calibration, explicit atmospheric correction, and by improved geolocation. Using MODIS data, small scale spatial processes like deforestation, shifting agriculture and fragmentation can be studied. The effect of spatial rearranging of ecosystems on the ecological functioning, expressed as edge effects and isolation, is a main topic for the MODIS data.
Cooperation is previewed with (i) Prof. Dr. R.B. Myneni (Boston University, Climate and Vegetation Research Group), member of the MODIS science team, and (ii) Dr. C.J. Tucker ( NASA Goddard Space Flight Center, Maryland), who processes the AVHRR data.
Publications related to this research:
Academic Bibliography of the University of Antwerp