12 Years Mediterranean Satellite Data Set and Analysis
169
can be interpreted as the phase of the vegetation wave in different years. The
considerable differences in the phase of the vegetation wave of successive years
indicate such evaluations as useful tools for the analysis of land surface processes.
The deviations of the ten year's mean values from the average of these years is
presented in Fig. 7 for the Iberian Peninsula. It provides evidence that in the
beginning of the nineties there was a trend towards a reduction of the biomass
production while the vegetation recovered in the last years of the century. This is an
example which makes it evident that long data series of the identical kind of data are
necessary to identify real trends and to separate them from interannual variations.
In all examples given here the data retain the full A VHRR resolution.
Even if the twelve year's period is rather short for trend analysis, it is
worthwhile to study the development of NDVI and reflectances. The main
interesting information is given by the overall differences and the local gradients
rather than by the absolute values. Interesting features can be found if the
development is represented by the slopes of regression lines for each individual
pixel position at full resolution for the period under consideration. Fig. 8 shows the
NDVI development for twelve years in the western part of the Mediterranean and
central Europe. The slope of the regression line is given in percent per year and
presented in colours. It is interesting to note that these slopes only in a few positions
(such as the south-east coast of Spain and Ebro valley in opposit directions) depart
significantly from zero which is presented in yellow color. In the upper part of Fig.
9 the development for the period from 1989 to 1995 is shown for the same area and
in the lower part of Fig. 9 for the period 1995 to 2000 in the whole Mediterranean
Basin and central Europe.
Most interesting are the striking inverse developments not only for the Ibearian
Peninsula and northwest Africa but also for western and central Europe. This
example gives an impression of the larger scale synergies of climate variability.
The moderate spatial resolution of about I km furthermore allows to zoom into
local events like forest fires as presented in Fig. 10.
3 The MEDOKADS Data Set As A Substantial Part of A
Remote Sensing Data Network for A Mediterranean
Research and Applications Network
The MEDOKADS data was developed at the Free University of Berlin where the
HRPT A VHRR data receiving station has been working since nearly 20 years. To
ensure the further access to the archived data in case of a failure of the system and
to extend the data set in a near real-time manner, efforts are under way to built up
a network of national institutions working with A VHRR data. On the one hand there
have been made arrangements between the receiving stations at the German Remote
Sensing Data Center (DFD) of the DLR, GKSS and Berlin to supply each other with
169
can be interpreted as the phase of the vegetation wave in different years. The
considerable differences in the phase of the vegetation wave of successive years
indicate such evaluations as useful tools for the analysis of land surface processes.
The deviations of the ten year's mean values from the average of these years is
presented in Fig. 7 for the Iberian Peninsula. It provides evidence that in the
beginning of the nineties there was a trend towards a reduction of the biomass
production while the vegetation recovered in the last years of the century. This is an
example which makes it evident that long data series of the identical kind of data are
necessary to identify real trends and to separate them from interannual variations.
In all examples given here the data retain the full A VHRR resolution.
Even if the twelve year's period is rather short for trend analysis, it is
worthwhile to study the development of NDVI and reflectances. The main
interesting information is given by the overall differences and the local gradients
rather than by the absolute values. Interesting features can be found if the
development is represented by the slopes of regression lines for each individual
pixel position at full resolution for the period under consideration. Fig. 8 shows the
NDVI development for twelve years in the western part of the Mediterranean and
central Europe. The slope of the regression line is given in percent per year and
presented in colours. It is interesting to note that these slopes only in a few positions
(such as the south-east coast of Spain and Ebro valley in opposit directions) depart
significantly from zero which is presented in yellow color. In the upper part of Fig.
9 the development for the period from 1989 to 1995 is shown for the same area and
in the lower part of Fig. 9 for the period 1995 to 2000 in the whole Mediterranean
Basin and central Europe.
Most interesting are the striking inverse developments not only for the Ibearian
Peninsula and northwest Africa but also for western and central Europe. This
example gives an impression of the larger scale synergies of climate variability.
The moderate spatial resolution of about I km furthermore allows to zoom into
local events like forest fires as presented in Fig. 10.
3 The MEDOKADS Data Set As A Substantial Part of A
Remote Sensing Data Network for A Mediterranean
Research and Applications Network
The MEDOKADS data was developed at the Free University of Berlin where the
HRPT A VHRR data receiving station has been working since nearly 20 years. To
ensure the further access to the archived data in case of a failure of the system and
to extend the data set in a near real-time manner, efforts are under way to built up
a network of national institutions working with A VHRR data. On the one hand there
have been made arrangements between the receiving stations at the German Remote
Sensing Data Center (DFD) of the DLR, GKSS and Berlin to supply each other with