production,
ick" hard
ts larger
results
coeffi-
ithin
endency
ed. These
an the good
or the
fallow (smal
ring wheat
ificant
] a ten-
f spring
ays the
ared to the
)nometric
à II and
sat from
in esti-
an. Also,
iat spring
am spring
nages
15.12 1552 1564
To [Rov on
rte en
236 234 234
[er [oct [Rov [ort
em
463.7 4613 4606
3567 3629 366.4
E
LseP T'ocr [Nov ]
. Spring
acquired during the ripening stage of
wheat. The area estimates were less than
1% under (as compared to the 10.7% under-
age experienced in Phase I and the 14%
underage of Phase II). If the major
differences between the spring wheat re-
gions of the yardstick area and the
U.S.S.R. are taken into consideration,
the yardstick results are supportive of
what was observed in the U.S.S.R. results
in Phases II and III. LACIE investigators
are optimistic that with Landsat D con-
siderable improvement will be possible in
the regions where strip fallow fields and
confusion crops have made classification
difficult.
Figure 21 compares the LACIE segment
wheat proportion estimates with ground
truth prepared after the analysis of
Landsat data was completed. These data
indicate a significant improvement
in the proportion estimates using the
Phase III procedures. These data also
support the improvement in the aggre-
gated results for the total region,
already described. The actual contact
time required to analyze a Landsat seg-
ment, manually select training fields,
compute training statistics, and computer
process the nearly 23 000 elements of a
80
150 U.S. segments
70 |" = .82
Slope = .67
60
3 50 "id
&
Í ao yo dice
$ t nt.
ul . 2° . :
9 30 i .* P
= SAM. LV ;
- : ..
20] = Pati
jl w^.
i ei 1 A A A 1
0 10 20 30 40 50 60 70 80
Ground—truth wheat percent
PHASE II
LACIE wheat percent
LACIE image was reduced from 10 - 12
hours in Phase I, to 6 - 8 hours in
Phase II, to 2 - 4 hours in Phase III.
Investigators concluded that the time-
liness goal of 14 days from acquisition
to report could be realized in a future
operational system.
FUTURE OUTLOOK
The totality of results from the
three crop years of LACIE strongly indi-
cates that current technology and know-
how do provide the "where with all" to
develop a global crop monitoring system.
Already the nations of the world share
their weather information through the
World Meteorological Organization. For
the first time on a global scale, Land-
sat provides a way to observe directly
the appearance of crops and to estimate
objectively the production at harvest.
In the future investigators will learn
how to combine the occurrence of weather
with crop appearance to make better
production forecasts. Improved equip-
ment, such as Landsat D with more
channels and better spatial resolution,
will make possible improved predictions
in the more difficult situations.
80
143 U.S. segments
70 r= ‚86
Slope = .74 7
60 |
50 | eu
.. f
40[ ; t
30 | 7 o7.
^ + 24
* 4 Fo …
go 5 e
. of ZA 4, s
iol. 7 ns
M CI
A
0 10 20 30 40 50 60 70 80
Ground—truth wheat percent
PHASE II
Figure 21.— Comparison of LACIE total wheat estimates with ground
truth - U.S. yardstick test sites.