onary
and
or all
data
N
more
ates.
1935,
iment
ludes
patial
deral
ems,
tions
Data
rs to
patial
both
bined
data
aphic
and
a is
and
n of
Jards
34
aw -
ty of
95
aphic
1998
ating
rolina
and
Data
and
olina,
>.uS/,
ating
ation
4
rolina
ation
.net/
MONITORING OF LAND USE AND CROP GROWTH CONDITIONS IN EUROPE
G.J.A. Nieuwenhuis, C.A. Mücher and A.J.W de Wit
DLO Winand Staring Centre (SC-DLO), P.O. Box 125, 6700 AC Wageningen, The Netherlands
ABSTRACT
A pilot study has been performed to map land cover for selected
regions of Europe using multi-temporal NOAA/AVHRR satellite
images. Based on the classification results the meaning of a 1-km
land cover database at the European scale for agricultural and
environmental applications has been illustrated. ¢
In a recent study for the Space Applications Institute (SAI) of the
Joint Research Centre (JRC, Ispra) the integrative use of high and
low resolution satellite data and crop growth models have been
studied by SC-DLO. Vegetation indices as derived from both high
and low resolution satellite data were used to estimate crop
parameters. Especially if the actual situation deviates considerably
from the mean situation, more realistic yield figures were obtained.
Evapotranspiration maps have been composed using both satellite
systems. In general arcal patterns as obtained with high and low
resolution satellite data were in agreement with each other. This
means that such images are highly relevant to detect alarming
situations concerning crop water supply during the growing
season. A quantitative analysis should be further elaborated,
integrating satellite observations and crop growth model
simulations.
1. INTRODUCTION
Since the seventies several initiatives haven been taken to set up a
monitoring system to obtain actual and reliable information on
agricultural statistics at national, continental and global level.
NASA initiated projects resulting in the well known Large Area
Crop Inventory Experiment (LACIE). Landsat data were applied
to estimate regional acreage of crops while simple
agro-meteorological models were used to obtain yield estimates
(NASA, 1978).
More recently in Europe an initiative has been taken by the Space
Applications Institute (SAI) of the Joint Research Centre (JRC,
Ispra). During the period 1990-1994 the Crop Growth Monitoring
System (CGMS) has been developed (Vossen and Rijks, 1995).
Historical yield statistics are used to estimate the technology
driven trend in crop yield. Based on crop growth model
simulations inter-annual variations due to meteorological
conditions are accounted for. The developed system is completely
operational now for regions and countries of the European Union
and will be extended to other regions in Europe and the
Mediterranean region.
To improve the accuracy of yield forecasting and crop growth
monitoring at a regional scale, methods must be sought for to
reduce uncertainties in crop growth simulation. In this respect,
remote sensing offers techniques that can be useful due to its
synoptic overview giving high spatial information and its frequent
measurement of actual situations in the field. Both observations in
the optical and thermal range of the electromagnetic spectrum
supply specific crop parameters (e.g. soil cover, Leaf Area Index
and evapotranspiration). Also observations in the microwave
range can play a role, especially in addition to observations with
optical sensors.
In a recent study for the SAI (JRC, Ispra) the integrative use of
high and low resolution satellite data and crop growth models
have been studied by SC-DLO (Nieuwenhuis et al., 1996).
Results of this study will be presented. Vegetation indices as
derived from both high and low resolution satellite data can be
used to estimate crop parameters. Especially if the actual situation
deviates considerably from the mean situation, more realistic
yield figures are obtained. Evapotranspiration maps have been
composed using both satellite systems. In general the areal
patterns as obtained with the high and low resolution satellite data
are in agreement with each other. This means that such images are
highly relevant to detect alarming situations concerning crop
water supply during the growing season. A quantitative analysis
should be further elaborated, integrating satellite observations and
crop growth model simulations.
To be able to focus the research on agricultural regions, one needs
actual and up-to-date information on land cover. Therefore in this
paper results of land cover and land use mapping at the European
scale are presented. Afterwards the integrative use of remote
sensing observations and crop growth modelling is shown for a
test site in southern Spain to obtain actual and reliable yield
estimates.
2. EUROPEAN LAND COVER MAPPING
2.1 European land cover data bases
Land use and therewith land cover are rapidly changing. To gain
better understanding of the nature of these changes and their
consequences, up-to-date and reliable information on Land
Use/Land Cover (LU/LC) is urgently needed.
In the framework of the CORINE programme, funded by the
European Union, land cover in Europe is mapped at a scale of
1:100.000 by interpreting Landsat-TM colour composites.
However, due to the high spatial resolution of the imagery, the
completion of the project for the whole of Europe will take a long
time and due to the costliness the project does not include future
monitoring on a regular basis. Therefore the applicability of low
resolution satellite images as supplied by NOAA/AVHRR has
been investigated.
Geo-referenced LU/LC databases that cover the whole European
continent are rare. The Pan-European Land Use Database (PELU)
of RIVM (Van de Velde et al., 1994) is a first step to meet the
demands of environmental monitoring on a European scale. Per
unit cell (about 10 by 15 kilometres), statistics on land use are
given, concerning arable land, grassland for agricultural use,
permanent crops, forest, urban areas and inland water.
Because this database is based on statistical data and various
analogue sources, the accuracy varies considerably throughout
Europe. Especially for Eastern European countries it is difficult to
obtain accurate and up-to-date information. Therefore, a more
accurate and consistent method nceds to be developed to solve the
existing imperfections of this database.
2.2 Experience with NOAA/AVHRR satellite imagery
Images | supplied by meteorological satellites, like
NOAA/AVHRR, have a low spatial resolution in comparison
with Landsat-TM and SPOT-HRV. They cover the whole of
Europe on just two images. Another advantage of the NOAA
satellite in comparison with SPOT and LANDSAT is its high
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 143
