1021
ESTIMATING CROP WATER DEFICIENCY USING THE RELATION BETWEEN
SURFACE MINUS AIR TEMPERATURE AND SPECTRAL VEGETATION INDEX
T. R. Clarke 1 , M. S. Moran 1 , Y. Inoue 2 , and A. Vidal 3
‘USDA-ARS U.S. Water Conservation Laboratory, 4331 E. Broadway Rd., Phoenix, AZ 85040 USA
NIAES, Division of Information Analysis, Laboratory of Agro-Biological Measurements, Tsukuba, JAPAN
3 CEMAGREF-ENGREF Remote Sensing Laboratory, B.P. 5095, 34033 Montpellier, FRANCE
ABSTRACT:
A new concept termed the Vegetation Index/Temperature (VIT) Trapezoid combines a spectral vegetation index
with the composite surface min us air temperature for all possible crop water conditions to form a two-
dimensional space, the limits of which are theoretically derived. In addition to measurements of surface
reflectance and temperature, an estimate of net radiation as well as air temperature, vapor pressure, wind speed,
and some values specific to crop type are required. Potential applications for irrigation management are
discussed, including a Water Deficit Index (WDI) which compares actual to potential évapotranspiration, and
the estimation of canopy temperature from populations of composite soil/vegetation temperatures.
KEY WORDS: Evapotranspiration, Reflectance, Temperature, Water Deficit, Crop Water Stress, SAVI,
Irrigation Scheduling.
1 - INTRODUCTION
The advent of commercially available high resolution thermal imaging systems has generated interest in the use
of this kind of instrument as an airborne remote sensing tool for irrigation scheduling in arid climate farmlands.
This is especially important as subsurface drip irrigation gains popularity, since traditional physical sampling
methods are less effective in determining soil moisture in the root zone of drip irrigated plants. Farmers can
currently deter min e their crop water needs using hand-held infrared thermometers and applying the measured
canopy temperatures to the crop water stress index or CWSI (Idso et al, 1981), which correlates water stress
to the foliage-minus-air temperature difference. Jackson et al. (1981) used a theoretical approach, writing the
energy balance equation in terms of foliage-minus-air temperature,
(T 0 -TJ=[r.(R B -G)/CJ[y(l +ryrJ/{A+ 7 (l + r 0 /rJ}]-[VPD/{A+7(l+ryO}], (1)
where T„ is the crop foliage temperature (°C), T, the air temperature ( 8 Q, r. the aerodynamic resistance (s m' 1 ),
R, the net radiant heat flux density (W m' 2 ), G the soil heat flux density (W m' 2 ), C, the volumetric heat
capacity of air (J °C‘ m' 3 ), r c the canopy resistance (s m 1 ) to vapor transport, y the psychrometric constant (kPa
“C- 1 ), A the slope of the saturated vapor pressure-temperature relation (kPa “C 1 ), and VPD the vapor pressure
deficit of the air (kPa). Eq. (1) was then solved for the ratio tjt, which was used in the relation
CWSI = l-I7ET p = [ 7 (l+ryrJ-7l/[A+7(l+rA)], (2)
where y" = y(l -t-r^rj; r^ being the canopy resistance at potential évapotranspiration and
r A = {trr.R n /CJ-[(T 0 -TJ(A+y)]-VPD}/{y[(T 0 -TJ-r.(R I -G)/C,]}, (3)
to obtain the ratio of transpiration T to potential évapotranspiration ET p . The CWSI defined by Eq. (2) was
used as an index of crop stress and thus an indicator of water deficiency.
The measurement process can be tedious and time consuming, for the user must be careful not to include
any soil background in the thermometer’s field of view, as the warmer soil temperature would produce
erroneous results. The same problem occurs when airborne thermal imagers are employed, as a well-watered,
immature crop with soil background temperatures influencing every pixel would be indistinguishable from a
water-stressed, full cover crop. It was felt that the inclusion of some measure of vegetative cover, such as a
vegetation index, would assist in the interpretation of thermal data. While the striking relationship between
surface temperature and vegetation index has been noted previously (Nemani and Running 1989; Price 1990),
