The methodology used to identify and carry out the 
monitoring of areas affected by drought consists of the 
following steps: 
a) Radiometric and atmospheric correction of channels 1 
and 2 of the AVHRR sensor. 
b) Daily calculation of the Normalised Vegetation Index 
(NDVI) from channels ! and 2 of the AVHRR sensor. 
C) Geometric correction of NDVI images. 
d) Calculation of the composite ten day image 
corresponding to the maximum value of the NDVI in 
the ten day period considered. 
€) Creation of a multitemporal archive with the 
Maximum NDVI values (MNDVI) corresponding to 
the temporal series for each ten day period from 
March to October 1993-1997. 
f) Calculation for each ten day period of the average 
value of the MNDVI corresponding to the period 
1993-1997. 
g) Calculation of the ratio between the ten day value of 
the MNDVI of the year under study and the average 
value of the MNDVI for the period 1993-1997. 
h) Expression of this ratio as a percentage and grouping 
in different intervals, associating them to a specific 
colour scale. 
3. RESULTS 
In order to be able to evaluate the methodology 
developed, 1995 and 1996 have been taken as examples. 
As was previously mentioned, 1995 was the driest year in 
the period 1993-1997 and 1996 was a fairly humid year 
over the whole of Spain and very humid in western 
Andalusia. In the Spanish Levant it was classified as 
normal. 
Figure 1 shows the anomalies with respect to the 
average MNDVI corresponding to the period from March 
to October, 1995. 
It can be easily observed that in the majority of 
the ten day periods yellow prevails which is identified 
with values below the average in Spain as a whole except 
in the north. This is perfectly in keeping with the pattern 
reflected in the anomalies which occurred in the rainfall 
for this year which in turn were the cause of the 
tremendous losses in agriculture and in the lack of 
drinking water in towns and cities as was mentioned in 
the introduction to this work. 
Figure 2 represents the anomalies with respect to the 
average MNDVI value corresponding to the months from 
March to October, 1996. In this case it may be observed 
that the anomalies are clearly positive for most of Spain 
after the third ten day period of March. These anomalies 
were positive even in the summer months and led to a 
drastic reduction in the number of forest fires in Spain for 
the year in question. In Andalusia, Extremadura and 
Castilla La Mancha, vegetation recovery was excellent as 
is clearly expressed in the images and put an end to a 
five-year drought period. 
4. CONCLUSIONS 
The results obtained in this work reflect the 
viability of carrying out the identification and monitoring 
of areas affected by drought using the information 
supplied by the AVHRR sensor of the NOAA satellite 
series. 
The operational application of this method in 
future will permit the rapid identification of those areas 
which might be affected by the phenomenon of drought 
and thus permit measures to be taken to palliate the 
effects on agriculture and forests. 
Having long temporal series of NDVI values 
available will permit more and more representative 
average values to be obtained and we will therefore be in 
a position to determine with more accuracy and 
robustness the anomalies which might occur in the 
evolution of the vegetation development in a given year. 
S.- ACKNOWLEDGEMENTS 
This work has been carried out within the 
framework of the SC96-096 research project financed by 
the Sectorial Research Programme of the Fisheries, 
Agriculture and Food Ministry in Spain. 
6.- BIBLIOGRAPHY 
Gutman, G., 1990. Towards monitoring droughts from 
space. Journal of Climate, 2: 282-295. 
Liu, W.T., Massambani, O. and Nobre, C.A., 1994. 
Satellite recorded vegetation response to drought in 
Brazil. International Journal of Climatology, 14: 343-354. 
Teng, W.L., 1990. AVHRR monitoring of US crops 
during the 1988 drought. Photogrammetric Engineering 
and Remote Sensing, 56:1143-1146. 
Tucker, C.J. and Choudhry, BJ. 1987. Satellite remote 
sensing of drought conditions. Remote Sensing of 
Environment, 23: 243-251. 
710 Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998