One or more of the Following Statements may affect this Document
• This document has been reproduced from the best copy furnished by the
organizational source. It is being released in the interest of making available as
much information as possible.
• This document may contain data, which exceeds the sheet parameters. It was
furnished in this condition by the organizational source and is the best copy
• This document may contain tone-on-tone or color graphs, charts and/or pictures,
which have been reproduced in black and white.
• This document is paginated as submitted by the original source.
• Portions of this document are not fully legible due to the historical nature of some
of the material. However, it is the best reproduction available from the original
Produced by the NASA Center for Aerospace Information (CASI)
NASA Technical Memorandum 81222
..aWe und« NASVf^'l>
in the interest oleaily and
semination of Earth
Program inlorraation am) without iatritim
tor any use made thereof.’
in a Large Agricultural Field
John P. Millard, Robert C. Goettelman,
and Mary J. LeRoy
(E80-10331) INFRARED-TEMPEfiATURE N80-3wi822
VARIABILITY IN A LARGE AGRICULTURAL FIELD
(NASA) 26 p HC A03/J1F A0 1 CSCL 02C
National Aeronautics and
NASA Technical Memorandum 81222
in a Large Agricultural Field
John P. Millard, Ames Research Center, Moffett Field, California
Robert C. Goettelman, LFE Corporation, Richmond, California
Mary J. LeRoy, DCA Corporation, Palo Alto, California
National Aeronautics and
Ames Research Center
Moffett Field California 94035
Inf rated* temperature variability in a large agricultural field
JOHN P. MILLARD
Ames Research Center, NASA, Moffett Field, California 94035, U.S.A.
ROBERT C. GOETTELMAN
LFE Corporation, Richmond, California 94804, U.S.A.
and MARY J. LeROY, DCA Corporation, Palo Alto, California 94303, U.S.A.
Abstract. Airborne thermal imagery of a large varying*terraln comaer-
cial barley field was acquired over a full growing season. The data were
analyzed to determine temperature variability within the field and the
percentage of area within various size Instantaneous fields of view
(ifov's) that would be within 1®, 2®, 3®, and 5® C of the mean. There
appears to be no great advantage in utilizing a small ifov instead of
a large one for remote sensing of crop temperatures.
! f «;urch?sed from;
Send correspondence to: . i L2i- Oj"'*'.'!'
John P. Millard 'i OuX Falls, SD S'?
Ames Research Center, NASA, M/S 240-6
Moffett Field, California 94035
1 . Introduction
The measurement of crop canopy temperature Is increasingly being
suggested as a tool to be used in agricultural crop management and assess-
ment; for example, Jackson et al. (1977), Idso et £l. (1977), and
Hatfield (1979). However, the accuracy with which crop tenq>era*-iires
must be measured and the acceptable temperature variations within a given
field have not been determined. Knowledge of the required accuracy and
variability would make it possible to determine appropriate Instantaneous
fields of view (ifov's) for remote sensors.
The Dunnigan Agro-Meteorological Experiment (DAME) airborne thermal
scanner results provide Insight into the temperature variability question.
DAME was a combined airborne and ground field measurement program which
was conducted over an entire barley growing season in support of a Heat
Capacity Mapping Mission spacecraft experiment. It was performed by Ames
Research Center, USDA/SEA, and the University of California at Davis.
Measurements of crop temperature, soil moisture, and meteorological param-
eters were acquired over the growing season.
This paper is concerned with using the airborne thermal scanner
results of the experiment and the analysis of data to define (1) the
temperature variability (coefficient of variation) that may occur within
various instantaneous fields of view, and (2) the percentage of the area
within various size ov*s that would be within I®, 2®, 3®, and 5® C
of the mean. Because of the extreme variability in slope of the DAME
site, the results may represent a worse-case condition and thus a very
conservative estimate on which to base future calculations.
2. DAME site airborne experiment
The DAME site was located on 1 section of land (I x 1 mile) (1.6 x 1.6 km)
located near Dunnigan, California « about AO km (25 miles) NW of Sacramento.
The terrain of this site varied from flat to slopes of about 30 percent;
thu8» almost any bar ley- growing terrain in the world was duplicated.
Barley, variety Briggs, was planted in 1 eceober 1977 and harvested in late
May 1978. The field was not irrigated, but 65 cm (26 in.) of rainfall,
almost twice normal, was received.
Figure I is a topographic map of the site, on which Is superimposed
(1) 16-ha (40-acre) cells; (2) mean slope, m, in percent; (3) standard
deviation, o, of the slope within each cell; and (A) coefficient of varia-
tion, Vg, of the slope in each cell. The NE cells are the most rugged,
and the southern cells are the flattest.
Airborne thermal Imagery of the site was acquired throughout the
growing season, from planting to harvest, except in April when the air-
craft was down for maintenance. Data were acquired both prior to sunup
and about 1 hour after solar noon; these represent minimum and maximum
surface temperatures, respectively. Thermal Imagery was acquired with a
Texas Instruments Model RS-25 infrared scanner operating in the 10.5- to
12.5-ym bandpass region. This instrument has an Ifov of 2 m (6.6 ft) at
the flight altitude of 1.2 km (AOOO ft) and a temperature accuracy of
about 0.2° C. It contains two blackbody calibration sources with platln\im
resistance thermometers for continuous inflight calibration. All thermal
data were digitally processed on an HP 3000 computer. In addition to
thermal data, natural and color-IR photography were acquired on alternate
flight days with a 70-mm Haaaelblad camera. At the completion of each
flight, atmospheric tesq>erature and humidity were measured at various levels
down to near ground level. These were used to correct the thermal->IR data
for water vapor absorption.
Airborne photographs acquired throughout the growing season (fig-
ures 2a-2e) showed that a truly uniform-appearing field never existed.
This nonuniformity was caused by variable slope, soil color, gullies, and
drainage-induced crop growth patterns. Figure 2a Is a natural-color photo-
graph obtained in August prior to planting and when the soil was dry.
The nonuniformity of the soil is the result of past leaching and the
presence of alluvial soil In the gullies. Figure 2b is a natural color
photograph obtained on the 49th day after planting (DAP), after the plants
had emerged. Much soil background is still apparent. Nonuniformities in
appearance were caused by farm equipment tracks, varying growth patterns,
and double-seeded areas. Figure 2c is a color-IR photograph obtained on
DAP 94. Except for gullies, the scene was rather uniform In appearance.
Figure 2d is a natural-color photograph obtained on DAP 98. Although only
4 days after DAP 94, the nonuniform scene appearance is quite striking.
Much bare soil and many gullies are apparent. The reason for this sudden
change between DAPS 94 and 98 is unknown, although It may be wind-induced.
Finally, figure 2e pertains to DAP 154, very close to harvest. This
shows the effect of crop maturity differences, caused by differing soil
moisture-holding capacities. The crops on the upper slopes and the tops
of the ridges matured earlier than those in the gully areas. Thus, there
are many causes of scene nonuniformities throughout the season.
The coeff lcient-o£-vari .tlon of afternoon temperatures, pixel by pixel,
within the DAME site la shown In figure 3. Maximum values of about 0.22
were obtained near planting tine for bare soil and no winds. Throughout
the remainder of the growing season values were less than 0.11, and minimum
values of about 0.02 were reached under wet soil conditions. As an aid to
interpreting these effects, soil moisture, wind conditions, and agronomic
values are presented in figure 3.
Figures 4-8 demonstrate the percentage of area, A^^, within various
size fields of view that would be within 1“, 2", 3®, and 5® C of the mean.
These were computed from the airborne scanner data, which consisted of
equivalent blackbody temperatures for every 2 m (6.6 ft) of the DAME site.
Figure 4 pertains to 4 ha (10 acre) ifov's. Since there are 64 such
4-ha (10-acre) cells in the DAME site, we decided to present only values
for very rugged terrains (the four 4-ha (10-acre) cells in the upper NE
cell of figure 1] and for gently rolling terrains [the four 4-ha (10-acre)
cells in the SE cell of figure 2]. These are adequate to bracket the
Figure 4 shows a large difference in A^^ (percentage of area within
1®, 2®, 3®, and 5® C of the mean) values between level and rough terrain,
and even between adjacent 4-ha (10-acre) cells. The explanation for these
differences is that many 4-ha (10-acre) cells vary in terms of slope,
gullies, and areas of nonuniform appearing vegetation. Vlhere such condi-
tions occur, a wide range of temperatures may exist, resulting in low Aj^
values. Where the scene is uniform and homogeneous, the spread in tempera-
ture is small and high A^^ values result. Figure 4 shows chat most
temperature InViomogeneity occurs for bare soil conditions. As the canopy
cover Increases, so does temperature homogeneity, especially for rough
terrain, thus Indicating a smoothing effect of the canopy. For near-level
terrain, about 95 percent of the data points are within 3* C of the mean
cell temperature during full-canopy conditions; for very rough terrain,
about 80 percent are within 3* C.
The percentage of 16-ha (40-acre) cells within various temperatures
of the mean Is plotted in figures 5 and 6. Two rather distinct families of
covers resulted: one for level and intermediate-slope terrains (figure 5)
and one for hlgh-slope terrain (figure 6). A small amount of crossover,
or Inconsistent data did exist; the reason for this Is not known. In
general, however, the temperature uniformity of the level and Intermediate
terrain cells increased rapidly with crop growth and then remained level
or dropped slightly throughout the remainder of the season. The cells
with high slope (figure 6) showed values that Increased steadily
with crop growth and reached maximum values later In the season. Over
80 percent of the data points are within 3° C of the mean cell temperature
over most of the growing season. Comparing the 4-ha (10-acre) cell size
results of figure 4 with the 16-ha (40-acre) results of figures 5 and 6,
we find identical trends and nearly the same values of A^, but note that
individual A^ values for 4-ha (10-acre) cells can be much more variable,
thus reflecting the uniformity or nonuniformity of the scene.
Finally, figures 7 and 8 pertain to cell sizes of 65 ha (160 acres)
and 259 ha (640 acres), respectively. Basically, the same magnitudes
and trends noted for the previous cell sizes are observed. Approximately
80 percent of the data points are within S'* C of the mean over most of
the growing season; for level terrain, the value is 90 percent.
The reasons for cen^erature variability within an agricultural field
are nany. Variability caused not only by varying topography* but also
by water~carved gullies* varying soil color* nu>isture state of the soil
and crop* nonuniform ph".nology* and bare spots i Althou^ these various
effects were nut separated* the cc.abined effect was measured for commer-
cially grown harlv.y planted on varying terrain. For all but the most rugged
terrain, over 80 percAr.t of the area within 4-* 16-* 65-* and 259-ha cells
(10-, 40- 160- , and C40-acre cells) was at temperatures within 3* C of
the mean cell temperature The result of using relatively small* 4-ha
(lO-acrc) ifov's for remote sensing applications is that either the worst
or the best of conditions is often observed. For example* the observed
temperature uniformity of a homogeneous field containing a stream will
vary considerably depending on whether the ifov contains the stream. If
only the homogeneous field is observed* great temperature uniformity might
be observed, but if the stream is within the ifov* then great nonuniformity
may be observed.
There appears to be no great advantage in utilizing a small ifov
[e.g.* 4 ha (10 acres)] instead of a large one [e.g.* 65 ha (160 acres)
or 259 ha (640 acres)] for remote sensing of crop canopy temperatures.
The percentage of the area within any of these Ifov’s that contributes
temperatures that are within 1“, 2*, 3** and 5* C of the mean is nominally
the same. The two alternatives for design purposes are then either (1) a
very high spatial resolution, of the order of a meter or so, where the
field is very accurately temperature mapped, or (2) a low resolution,
where the actual size seems to make little difference.
The authors v->uld like to tlumk Dr. Robert Regiiuito» USDA/SEA* and
Dr. Jerry Hatfield, U. of California at Davie, for their constructive
criticism of this manuscript and for providing the agroncmic values ahornn
in figure 3.
HATFIELD, J. L. , 1979, Aeron. J . 71, 889-892.
IDSO, S. B., JACKSON, R. D., end REGINATO, R. J., 1977, Science IH, 19-25.
JACKSON, R. D., REGINATO, R. J., end IDSO, S. B. , 1977, Water Res . U,
Figure 1. Topographic map of DAME aite Including mean alope M» atandard
deviation a, and coefficient of variance (c/M), Vg.
Figure 2, Aerial photograph of DAME site: (a) Dec. 9, 1977 (DAP 3)t
natural color film; (b) Jan. 24, 1978 (DAP 49), natural color film;
(c) Mar. 10, 1978 (DAP 94), color-IR film; (d) Mar. 14. 1978 (DAP 98),
natural color film; and (e) May 9, 1978 (DAP 154), color-lR film.
Figure 3. Coefficient of variation of temperatures within the DAME site.
Figure 4. Percentage of 4 -ha (lU-acre) cells (1/64 DAME site) that is
within designated temperature limits of the mean — for the four 4-ha
(10-acre) cells in the upper NE cell of figure 2, and the four 4-ha
(10-acre) cells in the SE cell of figure 1; (a) within 1" C of the
mean; (b) within 2° C of the mean; (c) within 3“ C of the mean; and
(d) within 5” C of the mean.
Figure 5. Percentage of l6-ha (40-acre) cells (1/16 DAME site) that is
within designated temperature limits of the mean - for level and
intermediate slope terrains; those cells in the central and lower
part of figure 1: (a) within 1* C of the mean; (b) within 2* C of
the mean; (c) within 3* C of the mean; and (d) within 5* C of the
Figure 6. Percentage of 16«ha (40-acre) cella (1/16 DAMP, site) that la
within deolgnated teoperature Halts of the mean — t%)r hlgh**slope
terralna; those cella in the upper-right comer of figure 1:
(a) within 1* C of the mean; (b) within 2* C of the aean; (c) within
3* C of the mean; and (d) within 5* C of the aean.
Figure 7. Percentage of 65-ha (160-acre) cells (1/4 DAME site) that Is
within designated temperature limits of the mean: (a) within 1* C
of the mean; (b) within 2** C of the mean; (c) within 3* C of the
mean; and (d) within 5* C of the mean.
Figure 8. Percentage of 259-ha (640-acre) cells (full DAME site) that
is within 1*, 2®, 3®, and S® C of the mean temperature.
'^KIGIXAL page is
•' PGOP QUALITY
DAYS AFTER PLANTING
I 1 1 I 1 I I
0.5 1.1 2.5 3.3 1.2
LEAF AREA INDEX
1 1 1 1 1
5 7 11 8 4
NO, OF GREEN LEAVES/PLANT
1 1 l 1 1 L 1
17 21 43 77 98
PLANT HEIGHT, cm
AREA, percent AREA, percent AREA, percent AREA, percent
— RUGGED TERRAIN
- GENTLY ROLLING TERRAIN
0 I 1 1 1 I I I I I I I I I
-60 -40 -20 0 20 40 60 80 100 120 140 160 180
DAYS AFTER PLANTING
AREA, percent AREA, percent AREA,percent AREA, percent
0 - - - . ____ __ . , ^
-60 -40 -20 0 20 40 60 80 100 120 140 160180
DAYS AFTER PLANTING