Remote sensing is in general the practice of collecting information from a distance. However, the term “remote sensing” was introduced in the 60’s when the first meteorological satellites were put in orbit. Thus, nowadays it is used to describe the acquisition of spectral information regarding Earth’s surface using satellite or aerial technologies. The major advantage of such technologies over field measurements is that they provide information in large spatial scale. There are satellite instruments that can acquire imagery over nearly the entire Earth’s surface in 24h. The temporal scale of such data is also very valuable. Daily acquisitions of the same region give the opportunity to produce long-term time series of multiple data that can be used to monitor dynamic processes of Earth’s surface. Such processes include vegetation phenological or biochemical behavior.
The spectral information of a satellite instrument is very important and determines its prospective applications. According to the spectral resolution, the sensors can be characterized as multispectral or hyperspectral (Figure 1). Multispectral sensors acquire spectral information usually in few and broad parts of the spectrum (bands), while hypespectral sensors possess the capacity to record detailed spectral attributes of Earth's surface through many continuous narrow spectral bands. A common practice to study vegetation using multispectral or hyperspectral sensors is to use their bands in mathematical formulas (Vegetation Indices - VIs) in order to calculate a value indicative of certain vegetation attributes. The most common VI calculated from satellite instruments is the Normalized Difference Vegetation Index (NDVI):
where RNIR and Rred are the reflectances of the red and the NIR band accordingly. NDVI takes advantage of the vegetation unique spectral characteristic to absorb in the red and reflect strongly in the NIR part. The denser and healthier the vegetation is, the higher NDVI gets. NDVI and other VIs are found to be strongly correlated with Leaf Area Index (LAI) that is a key component in ecosystem modeling. Therefore, satellite remote sensing can provide spatially and temporally extended information that is valuable for ecosystem modeling.
In the Laboratory of Botany satellite remote sensing is used in multiple ecophysiological studies. There is experience in processing hyperspectral and multispectral data. Moreover, an ecosystem primary productivity model (ModSat) that uses multispectral satellite and ground meteorological data is developed and calibrated for several ecosystems in Epirus, Greece (Figure 2).
Figure 1. Reflectance of a Fagus sylvatica ecosystem calculated from multispectral (MODIS) and hyperspectral (CHRIS/PROBA) satellite images.
Video showing the the gross primary productivity of the world's land areas for the period 2000-2009 as calculated from Terra's MODIS instrument (Credit: NASA/Goddard Space Flight Center).
Figure 2. Gross Primary Productivity (GPP) of a Fagus sylvatica ecosystem, as estimated by the ModSat model for a 3-year period.