Use of Remote Sensing technique in scheduling Irrigation

Use of Remote Sensing technique in scheduling Irrigation

D.K. Das
Division of Agricultural Physics
Indian Agricultural Research Institute
New Delhi-110 012, India

Abstract
Increased interest in the utility of Remote Sensing technique for efficient water management has necessitated study of the possibility for its use in irrigation scheduling for crops over large areas. The paper discusses the different remotely sensed canopy temperature derived indices and their implications in scheduling irrigations for different agricultural crops. Research results based on experiments conducted in semi-arid region of Delhi, India are presented to show that in wheat irrigation based on standard deviation of midday canopy temperature (CT-SD) of ±-10.30°C produced yields of 98.30 and 97.00 q ha-1 of dry matter, 42.92 and 41.62 q ha of grains with a saving of 100 and 60mm of irrigation water over the plots frequently irrigated at critical growth stages. Irrigation based on canopy temperature derived indices also resulted in lowering of leaf diffusion resistance (LDR) and was associated with an increase in leaf water potential (LWP), transpiration rate and lowering of canopy temperature over the water stressed wheat crop thus maintaining favourable canopy environment in the former. The results demonstrated the potentiality of irrigation scheduling based on remotely sensed canopy temperature to attain higher yield and water use efficiency of crops.

Introduction
Efficient water management is key to success in augmenting crop production. In developing countries such as in India, which has created the highest irrigation potential in the world, the use efficiency of irrigation water resource is quite low ranging from 20 to 40 percent owing mainly to improper irrigation schedules, over irrigation, excessive seepage and percolation. The conventional methods of scheduling irrigation based on either per cent soil water depletion, irrigation water-pan evaporation ratio (IW/CPE) or crop growth stages often result in greater number and amount of irrigation than actually required for specific crops. Under conditions of low evaporative demand, water extraction by a plant may balance the transpiration demand over a relatively greater depletion of soil moisture but under high evaporative demand, even a high soil moisture level may result in yield reduction level crop water deficit. Moreover the point measurements have limited application have over large areas. Hence increase in irrigation water use efficiency necessitates improved irrigation scheduling techniques based on integrated effects of climate-soil-crop characteristics. In recent times there has been increasing interest in the use of newly developed Remote Sensing techniques for efficient management of economic and costly water resources (Sahai, 1990; Da et al. 1990). The use of crop canopy temperature measured through remotely sensed sensors (by satellite, air-born or ground based sensors) had opened up new vistas for control of crop water supply, in proper scheduling of irrigation and better utilization of water resources (Jackson et al., 1977; IDSO et al., 1977; Idso et al., 1977; Das and Kalra, 1990). This report briefly discusses the thermal indices for detection and quantification of plant water stress, some research results concerning the use of these indices in scheduling irrigation and the implications in adopting the frontier technology for farm water management.

Canopy Temperature Based Indices
Thermal-IR techniques can be used to detect and quantify plant water stress. The methods are associated with increased leaf temperature variability in a cropped field with restricted transpiration because of a deficit in water supply (Tanner, 1963). Four canopy temperature based indices have been developed for detecting plant water stress and scheduling irrigation; (i) canopy-air temperature difference (CATD) and stress degree days (SDD), (ii) canopy temperature variability (CTV), (iii) temperature stress day (TSD) and (iv) crop water stress index (CWSI) (Jackson et al. 1986).

Stress degree day is the cumulative difference between the canopy temperature (T) and air temperature (T) measured post-noon near the time of maximum heating (Idso et al., 1977; Jackson et al., 1977). It is assumed that the canopy temperatures would account for the effect of environmental factors such as vapour pressure, net radiation and wind. The SDD increases with increasing plant water stress (Fig. 1, 2). A crop is considered stressed if the value is high and positive and unstressed if it is negative. This change over is, however, arbitrary and may not be valid for all environments.
The canopy temperature validity (CTV) is the variability of temperatures encountered in a field during a particular measurement period. It is expressed as the standard deviation of mid-day canopy temperature within a field. The basis for CTV index is that soils are inherently non-homogeneous. Some areas within the field becomes stressed earlier than others. As water limiting in the former, the canopy temperature would show a greater variability. This variability can be used to signal the onset of deficit and schedule irrigation (Gardner et al., 1981)
The temperature stress day (TSD) is the difference in temperature between a stressed plot and a well irrigated plot (Gardner et al., 1981). Use of well watered plot as reference compensates for environmental effects. It needs to be in the vicinity of the field to be irrigated.
The Crop water stress index (CWSI), defined as CWSI= (1 – AET/PET) is generally accepted as a global quantitative assessment of the water stress (Jackson, 1982). AET is the actual evapotranspiration (Water effectively used as a daily or large time scale equivalent of LE) and PET is the estimate of potential or maximum evapotranspiration. By using the steady state energy balance for a crop canopy CWSI was computed for irrigation scheduling purposes based on relationships between canopy air temperature differences and vapor pressure deficit (Idso et al., 1981; Jackson et al., 1981). It is computed as the ratio of the observed (T_c – T_a) for the given condition of air saturated deficit e (in mb or Kp) at the time of observation to the maximum possible (T_c – T_a)max for a fully dry crop in the same conditions (Fig. 3).

Testing The Indices for Scheduling Irrigation
CATD and SDD: Ehrler (1973) suggested that leaf-air temperature differences could be used as a guide irrigation scheduling. In field experiment on wheat, profile water depletion had been correlated with SDD (fig. 2) and it had been suggested that irrigation to whet might be applied when the positive SDD value was 10 or below (Jackson, 1977). Scheduling irrigation for snap beans based on SDD resulted in similar grain yield to that of irrigation based on soil water potentials or growth stage (Bonano and Mack, 1983). Geiser et al. (1982) developed an irrigation scheduling model in corn using CATD on the dependable variable and net radiation, relative humidity and available soil water as independent variables. The water balance and resistance methods called for additional water application of 39 and 18 per cent, respectively when compared with temperature difference method.

CTV : The average canopy temperatures of a water stressed plot might no typify water stress. On the other hand the canopy temperature variability indicated areas of adequate or inadequate water in a field. A measure of temperature variability, therefore, might be used in irrigation management (Nixon et al., 1973). Fully irrigated plots of corn exhibited standard deviation of CT ± 0.3°C whereas in non-irrigated plots the S.D. was as high as ± 4.2°C. It was, therefore, concluded (Gardner et al., 1981) that crops exhibiting S.D. above ± 0.3°C were in need of irrigation.

TSD : In testing the concept of temperature stress days (TSD) for irrigation scheduling, irrigation was applied to corn when the average of all CT measured in stressed plot during a time period 1°C warmer than the well irrigated plot (Calwson and Blad, 1982). Comparison of CTV and TSD indicated that the former be used to signal onset of plant stress but the severity of stress was better indicated by the magnitude of the elevation in average CT above that of a well watered reference plot.

CWSI : Computation of crop water stress index based on CATD and vapour pressure deficit, indicated that with CWSI of 0.3, a reduction in growth rate was imminent. If it reaches 0.5, net growth would cease and might decrease. Irrigation should, therefore, be applied when CWSI is between 0.3 and 0.5 (Jackson, 1983). Originally the approach was developed and verified under warm and dry climatic conditions, and only limited work (Keener and Kirchner, 1983; Pennington and Heathery, 1989) has been conducted using the approach under humid conditions.

Experiments under semi-arid climatic conditions of delhi
In a field experiment on sandy loam soil (Typic Ustochrept) wheat cv. HD 2285 was grown with 5 different levels of irrigation (Table 1) during 1986-87 and 1987-88 rabi seasons. Irrigation applied at canopy temperature standard deviation (CT-SD) of ±0.3°C (I₃) resulted in higher water use efficiencies (12.04 and 10.84 kg grain ha mm^-1 of water) of wheat. This treatment (I₃) produced 98.3 and 97.00 q ha^-1 of dry matter and 42.92 and 41.62 q ha^-1 of grains in the two years which were statistically at par with this obtained in frequently irrigated plot (I₄). But with treatment I₃ there resulted a saving of 100 and 60 mm of irrigation water and a reduction in water use by 60.4 and 47.0 mm over the I₄ treated plot with only marginal reduction in grain yield of the order of 4.4 and 3.7 per cent in the two years. Irrigation based on CTV (I3) also resulted in lowering of leaf diffusion resistance and canopy temperature and was associated with an increase in transpiration rate and leaf water potential (Table 2) over the unirrigated water stressed and limited irrigation treatments (I₀ to I₂).

In a second experiment irrigation schedule to wheat was based on temperature stress day (TSD) i.e when the average canopy temperature of the stressed minus well watered plots was 0.5°C (table 3). The two cultivars sonalika and HD 2329 yielded 49.14 and 45.71 q ha^-1 of grains in the TSD based treatment which were significantly higher than that of the control (A, unirrigated) and IW/CPE ratio of 0.6 (B) and were similar to the yields (50.28 and 46.85 q ha^-1) obtained in frequently irrigated plot (D). The amount of soil water depletion was, however, considerably lower in the TSD based treatment (C) (261 and 277mm) as compared with the later treatment (301 and 308mm). The results demonstrated the potentiality of irrigation scheduling based on canopy temperature variability and temperature stress days.

Fig 1: Seasonal progression of cumulative stress-degree days in wheat under differnent irrigation treatments (1986-87)
The relative variation and correspondence between crop water stress index (CWSI) based on canopy temperature, and fraction of available water in the soil exhibits a definite pattern under conditions of limited water supply (fig. 4). As the fraction of available water varied from 0.43 to 0.20, the CWSI increased from 0.32 to 0.41, respectively. As CWSI incorporates the integrated effects of water and atmospheric environment, it emerges as a better indicator of plant water stress than fraction of available water which is soil dependent index (Chopra et al. 1990).

Fig. 2 : Cumulative SDD and soil water depletion as a function of time after sowing for two wheat canopies (Jackson et al., 1977)

Fig. 3 Canopy air temperature difference as affected by vapour pressure deficit at maximum (upper line) and minimum (lower line) water stress. Points A and C represent canopy Air temperature differences at maximum and minimum stress, respectively, and B is a measured value (Jackson, 1982)

Table - 1 : Effect of irrigation scheduling based on canopy temperature variability on dry matter, grain yield, water use and water user efficiency (WUE) of wheat.

Fig. 4 : Variation of crop water stress index (CWSI) in wheat with fraction of available soil water (a) 60 days after sowing, (b) 90 days after sowing.

Table 2: Effect of irrigation scheduling based on canopy temperature variability on leaf diffusion resistance (LDR), transpiration rate (TR), leaf water potential (LWP) and canopy temperature (CT) at 85 days of growth of wheat.

Table 3: Efect of irrigation scheduling based on temperature stress days (TSD) on yield and soil water depletion of two wheat cultivars.

Implications
The thermal infra red measurements allowing access to the surface temperature of crop canopies, indicate the onset of degree of stress at a particular and thus could be used in scheduling irrigation. However, use of satellite based surface temperature data for quantitative stress assessment and schedule irrigation is still speculative. Operational systems need to be developed. Secondly, the stress day (TSD) index requiring a well irrigated crop in the vicinity as a standard comparison with canopy temperature of stressed crop would not be suitable in majority of the rainfed and dryland areas where such condition does not exist. Hence a simpler approach for scheduling irrigation needs to be developed. While potentially of CWSI has been demonstrated, the critical limits for different crops for irrigation application status and schedule irrigation, particularly under humid climatic conditions, is complicated by the fact that relatively small temperature differences need to be measured within a narrow range of vapour pressure deficit and under conditions of fluctuating global radiation and wind speed (Jensen et al., 1990).

Conclusion
Assessing crop water status and scheduling irrigation based on canopy temperature measurement through ground based Remote Sensing technique, such as use of radiation thermometer, holds great promise. However, there is a need to compare the canopy temperature based indices with different soil and plant based criteria to find out best suited irrigation schedule under specific soil and climatic conditions to attain maximum yield and water use efficiency.

References
Bonnao, A.R. and Mack, H.S. 1983. J.Am. Soc. Hort. Sci., 108 : 826-831

Chopra, U.K. Srivastava, S.K. and Das, D.K. 1990. Proc. Natl. Symp. On Remote Sensing for Agricultural Applications, IARI, New Delhi; Dec. 6-8, 1990. pp. 197-203.

Clawson, K.L. and Blad, B.L. 1982 . Agron. J. 74: 311-316

Das, D.K. Singh, G and Tomar, K.P. 1986. Proc. Int. Sem. On Water Management in Arid and Semi-arid Zones; HAU, Hisar, Nov. 27-29, 1986, 264-269.

Das, D.K. And Kalra, N. 1990. in ‘Technology Blending and Agrarian Prosperity’. Eds. J.P. Verma and A. Varma; Malhotra Publishing House, New Delhi; pp. 98-118

Ehrler, W.I. 1973. Agron J., 65 ; 404 – 409

Gardner, B.R. Blad B.L. and Watts, D.C. 1981a. Irrig. Sci. 2: 213-224

Gardner, B.R. Blad B.L. and Watts, D.C. 1981b. Agric. Meteorol., 28 : 207-217

Geiser, K.M. Slock. D.C., Allred, E.R. and Stange, K.W. 1982, Trans A. SAE., 25 : 689-694.

Idso, S.B., Jackson, R.D. and REginato, R.J. 1977 Science, 196 : 19-25

Idso, S.B., Reginato, R.J. Reicosky, D.C. and Hatfield, J.L. 1981, Agron J., 73 : 826-830.

Jackson, R.D. 1982, In Advances in Irrigation (Ed. D. Hillel). Academic Press, N.Y., 1: 43-85

Jackson, R.D. Idso, S.B> Rginato, R.Jadn Pinter, Jr. P.J. 1981, Water REso. Res. 17 : 1133-1138

Jackson, R.D., Pinter, Jr. P.J. Reginato, R.J. and Idso, S.B. 1986. IEEE Trans. Geosci. Remote Sensing 24: 99-106.

Jackson, R.D. Reginato, R.J. and Idso, S.B 1977. Water REso. Rec., 13 : 651-656.

Jensen, H.E. Svendsen, H., Jensen S.E and Mogensen, V.O. 1990, Irrig. Sci. 11: 181-188.

Keener, M.E. and Kirchner, P.L. 1983. Agric. Meteorol. 28 : 339-349

Nixon, P.R., Namken, L.N. and Wiegand, C.L. 1973. in Remote Sensing of Earth Resources (Ed. F. Shahroki). Univ. of Tennessee. Tullahoma, T.N> : 643-657.

Pennington, D.A Heatherly, L. 1989. Agric. Forest Meroerol. 46 : 1-14.

Sahai, B. 1990. In Proc. Natl. Symp. On ‘Remote Sensing of Agricultural Applications’. I.A.R.I., New Delhi; Dec. 6-8, 1990, pp. 333-345.

Tanner, C.B. 1963, Agron J., 55