'e con- 
under- 
sts are 
ntrasts 
ist. be- 
es can 
crease 
ifferen- 
iduous 
ERS-1 
'curacy 
dorschkamp 
flevo 
robusta 
zeeiand 
oxford 
' pseudoplatanus 
nus excelsior 
'cus robur 
s sylvestris 
tion of 
dorschkamp 
fievo 
robusta 
zeeland 
oxford 
pseudoplatanus 
»us excelsior 
cus robur 
; sylvestris 
tion of 
wn in 
> back- 
or. This 
longer 
> cano- 
its, i.e. 
] inter- 
be the 
; Popu- 
latively 
by the 
n-verti- 
ised by 
lower 
contribution from the trunk-ground interaction. In 
contrast to ERS-1, backscatter levels, in general, incre- 
ase between February and March and show a gradual 
decrease when leaves develop. The latter may indicate 
that contributions from bigger branches are attenuated 
by the leaf layer, while the leaves themselves give a 
relatively low return. The patterns for Pinus sylvestris 
and Quercus robur are different, which, following the 
same line of reasoning, is logical because Pinus keeps 
its needles during winter and Quercus starts develo- 
ping leaves after the May observation, in contrast to 
the other deciduous species present here. The drop in 
the signature of Pinus in August may be explained by 
the increase of undergrowth, and, consequently, redu- 
ced trunk-ground interaction. In summer it may be well 
possible to distinguish between coniferous and decidu- 
ous tree species in L-band. 
A maximum likelihood classification using four JERS-1 
SAR images resulted into an overall classification 
accuracy of 72%. 
6. CONCLUSIONS 
As the contrast of the JERS-1 images appeared to be 
rather well when compared with the ERS-1 images 
visually, classification for agricultural crops was expec- 
ted to give good results. However, the dynamic range 
of the JERS-1 time series during the growing season 
was not as large as found before for L-band in other 
campaigns. The backscatter signatures of JERS-1 
during the growing season were difficult to compare 
with those of ERS-1 during the growing season. For 
the leafy crops such as potato or sugar beet, there 
were some similarities in backscatter behaviour. 
When comparing the results of a multitemporal (three 
dates) per field classification based on ERS-1 data with 
the results for JERS-1 data, varying results were obtai- 
ned. For instance, grassland and potatoes were classi- 
fied most accurately using ERS-1 data. On the contra- 
ry, sugar beet, maize and barley, for instance, were 
classified more accurately using JERS-1 data. With an 
optimal multi-temporal data set agricultural crops can 
be classified with an overall accuracy of 80% using 
ERS-1 SAR data. Although in this experiment we had 
only a restricted number of JERS-1 and ERS-1 images 
for the growing season of 1993, several crops could 
be classified with an accuracy of more than 80%. 
Results showed that JERS-1 is better suited to discri- 
minate forest species than ERS-1. A multi-temporal 
approach, i.e. at least one observation in winter and 
one in summer/spring, is recommended. A visual inter- 
pretation of images revealed that the discrimination 
between deciduous and coniferous species is well 
possible with a single L-band summer image. In winter 
this is not possible and in C-band it is not possible at 
151 
all. These results are in agreement with results of 
previous studies for this test area. 
ACKNOWLEDGEMENTS 
This report describes a study that was carried out in 
the framework of the NRSP-2 under responsibility of 
the Netherlands Remote Sensing Board (BCRS). NAS- 
DA is acknowledged for providing JERS-1 SAR images 
and ESA is acknowledged for providing ERS-1 SAR 
images and for financing the ground truth collection. 
Martin Vissers is acknowledged for doing the field 
work and the preprocessing of all the radar images. 
This study was performed partly in parallel to ESA- 
ESTEC study contract no.1154/94/NL/NB(SC), which 
was managed by Synoptics. 
REFERENCES 
Janssen L.L.F. 1994. Methodology for updating terrain 
object data from remote sensing data. PhD thesis, 
Wageningen Agricultural University, the Netherlands, 
173'pp: 
Kleijweg, J.C.M & J.S. Groot, 1995. JERS-1 Calibrati- 
on. Physics and Electronics Laboratory TNO-FEL, The 
Hague. 
Laur, H., 1992. ERS-1 SAR calibration; derivation of 
backscattering coefficient o° in ERS-1 SAR PRI pro- 
ducts. ESA, ESRIN, 16 pp. 
Leeuwen, H.J.C. van, 1996. Methodology for combi- 
ning optical and microwave remote sensing in agricul- 
tural crop monitoring. PhD thesis, Wageningen Agri- 
cultural University, the Netherlands, 233 pp. 
Linden, M. van der, 1995. Monitoring van bossen met 
ERS-1 en de JPL/AIRSAR. MSc Thesis, Wageningen 
Agricultural University, Dept. of Water Resources. 
Nieuwenhuis, G.J.A. & W.W.L. van Rooij (eds), 1994. 
Application of ERS-1 SAR data in agriculture and fores- 
try. Report 101, SC-DLO, Wageningen, the Nether- 
lands, 182 pp. 
Schotten, C.G.J., W.W.L. van Rooij & L.L.F. Janssen, 
1995. Assessment of the capabilities of multi-temporal 
ERS-1 SAR data to discriminate between agricultural 
crops. Int. J. Remote Sensing, vol. 16, no. 14, 2619- 
2637. 
Vissers, M.A.M., 1994. ERS-1 / JERS-1 Ground truth 
data collection in Flevoland (NL) 1993. Wageningen 
Radar Surveys, Wageningen, The Netherlands, 44 
pages & 3 appendices. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996