International Commission on A25

Irrigation and Drainage

1st Asian Regional Conference
Seoul, 2001

CROPWAT MODEL TO EVALUATE CROP WATER REQUIREMENTS

IN TAIWAN

Sheng-Feng Kuo1, Bor-Jang Lin2 and Horng-Je Shieh3

Abstract

An irrigation management model simulating the complicated on-farm “crop-soil-climate”

phenomena will facilitate the estimate of the crop evapotranspiration and irrigation schedule,
and, agricultural water requirements with different cropping patterns for irrigation planning.

The field experimental data from the HsuehChia Experimental Station of ChiaNan Irrigation
Association in Taiwan were collected and analyzed, then input the results to the CROPWAT
irrigation management model that was developed by the Food Agricultural Organization
(FAO). The results from CROPWAT model show that the annual potential evapotranspiration
and effective rainfall in HsuehChia area are 1,444 mm and 897 mm, respectively. In the
paddy fields, the crop water requirements and deep percolation are respectively 962 mm and
295 mm for the first rice crop, and 1,114 mm and 296 mm for the second rice crop.

Based on the single rice cropping pattern with 65% planted area of paddy fields and irrigation
management data of HsuehChia Working Station, the CROPWAT model simulated results
show that the annual crop water demands are 1,019 mm, and the peak monthly water
requirements happened on the January with 192 mm. For the double rice cropping pattern, the
annual crop water demands are 1,019 mm, and the peak monthly water requirements
happened on the January with 192 mm.

The research shows that the irrigation management model can effectively and efficiently
estimate the crop water requirements. Yet it still needs further study to fit the model to the
complicated situation of cropping patterns, for upgrading the ability of irrigation management
of Irrigation Associations in Taiwan.
1
Division Chief, Chiseng Water Management R&D foundation, Taiwan & Associate Professor, Dept. of
Civil Engineering , National Ilan Institute of Technology, Taiwan. 3F. No. 18, Lane 90, Sec. 6, Min-Chuan East
Rd., Taipei, Taiwan 114, R.D.C. e-mail:kuosf@chiseng.org.tw, Tel:+886-2-27928810, Fax:+886-2-27918674
2
Senior Specialist, Irrigation and Engineering Division, Forestry Dept, Council of Agriculture, Taiwan &
3
Researcher, Water Resource Bureau, Ministry of Economic Affairs, Taiwan & Assistant Professor, Dept. of
Hydraulic gineering, FengChia University, Taiwan
Keywords: Irrigation Management, Crop Water Requirements, Evapotranspiration,
Cropping Pattern

Introduction

For effectively and efficiently using the available water sources to meet the possibly variation
of cropping pattern, irrigation management plays an important role. To facilitate the
management practice, experimental data based the irrigation management model can be
applied to estimate the crop water demand and upgrading the capability of irrigation
management in 17 Irrigation Associations in Taiwan.

Many existing models (CRPSM, UCA, COMMOD, RIMMOD, ISOM and CROPWAT)
conduct on-farm water demands based on climate-soil-plant systems. Hill, et al. (1987)
developed the crop yield and soil management simulation model (CRPSM) to estimate crop
yield as a function of soil moisture content, crop phenology and climate during the growing
periods. Keller (1987) developed the USU Unit Command Area (UCA) model based partly on
the concepts of CRPSM model. The UCA model consists of two submodels: the on-field
submodel and the UCA water allocation and distribution submodel to calculate of the daily
aggregate UCA water demand and updating of the soil water balance on a daily basis. Chong
(1992) developed the Rice Irrigation Management Model (RIMMOD) to increase potential
crop yield by providing water for transplanting and rice growing at appropriate times and in
appropriate amounts and by reducing water stress with improved system operation and
management. Smith (1991) developed the CROPWAT to evaluate the crop water
requirements for rice and upland crops. Prajamwong (1994) developed the Command Area
Decision Support Model (CADSM) based on the concepts of the WCA model and consists of
three main sub-models: (1) weather and field generation; (2) on-field crop-soil water balance
simulation; and, (3) water allocation and distribution. Kuo, Sheng-Feng (1995) developed the
Irrigation Simulation and Optimization Model (ISOM) based on the implantation of genetic
algorithm (GA) method to on-farm irrigation simulation for optimizing the allocation of
irrigated area to alternative crops for maximum net benefit of irrigated project.

Field experiment serves to provide the basic data essential for irrigation management.
Therefore, the HsuehChia Experiment Station of ChiaNan Irrigation Association in Taiwan
conducted field experiments to evaluate the crop water requirements and crop coefficients for
corn, sorghum and soybean from 1985 to 1995, respectively. Shih (1986, 1997) stated the
experiment procedures and results from HsuehChia Experiment Station of ChiaNan Irrigation
Association in Taiwan.

The purposes of this study could be summarized as follows:

(1) To collect and analyze the field experiment data from HsuehChia experiment station.
(2) To collect the on-farm water management data from ChiaNan Irrigation Association for
model calibration;
(3) To apply CROPWAT irrigation water management model for evaluating the crop water
requirements; and

2
(4) To evaluate and discuss the agricultural water requirements under different cropping
patterns.

Methodology

The CROPWAT model was originally develop by the FAO in 1990 for planning and
management of irrigation projects. The newest version, namely CROPWAT4W under
windows interface, was jointly formulated by the FAO, Southampton University of UK, and
National Water Research Center (NWRC) of Egypt. Figure 1 shows the flowchart of
CROPWAT model.

In Fig. 1, it can be seen that the input data cover crop, meteorology, and soil. The
meteorology data include: (1) maximum and minimum temperature; (2) wind speed; (3)
sunshine hours; (4) relativity humidity; and (5) rainfall. The Penman Monteith explicit
equation was used to calculate the potential evapotranspiration (ETo).

There are four common empirical methods for calculating effective rainfalls (Smith 1991) as
follows: (1) fixed percentage of rainfall; (2) dependable rainfall; (3) empirical formula; and
(4) USDA Soil Conservation Service Method. The USDA Soil Conservation Service method
was used in this study. These four empirical equations are briefed as follows (Smith 1991):
1. Fixed Percentage of Rainfall:

PEeff = a * Ptot (1)

where PEeff is the effective rainfall; a is a fixed percentage coefficient (specified by the model
user), with a typical range of values from 0.7 to 0.9; and Ptot is the measured (or generated)
total daily rainfall.
2. Dependable rainfall: This empirical formula was developed by the FAO to estimate
dependable rainfall. This method may be used for design purposes where 80%
probability of exceedance is required.

For Ptot < 70 mm; PEeff = 0.6 Ptot - 10

For Ptot > 70 mm; PEeff = 0.8 Ptot - 24 (2)

3. Empirical formula: This formula will determine the effective rainfalls based on analysis
of local climatic records.

For Ptot < z mm; PEeff = a Ptot + b

For Ptot > z mm; PEeff = c Ptot + d (3)
where a, b, c, d and z are empirically-derived correction coefficients.
4. USDA Soil Conservation Service method:

3
 125 − 0.2Ptot
PEeff = Ptot * for Ptot < 250mm
125
PEeff = 125 + 01
. * Ptot for Ptot > 250 mm (4)
where,
PE: effective rainfall (mm)
Ptot: total rainfall (mm)

Given the input of the requirement data, the CROPWAT model can be used to calculate crop-
related data in each decade of a month, such as: (1) crop coefficient, (2) crop leaf index, (3)
crop evapotranspiration, (4) percolation, (5) effective rainfall, and (6) crop water
requirements. Also, the model can be applied to estimate the irrigation schedule for each crop
with 5 different options: (1) each irrigation defined by irrigation manager, (2) irrigation at
below or above critical soil depletion (% RAM), (3) irrigation at fixed interval per crop
growing stage, (4) deficit irrigation, and (5) no irrigation. Afterwards, the CROPWAT model
can simulate the on-farm crop water balance, including : (1) irrigation times, dates and depths,
(2) soil moisture depletion, (3) amount of percolation, (4) actual crop evapotranspiration, and
(5) crop yield. The on-farm water balance was based on the theory of Eq (5) below:

SMD t = SMD t −1 + ETc − PE − IR + RO + DP (5)

where:
t: time (decade of month)
SMDt, SMDt-1 : soil moisture depletion at t and t-1 decade (mm)
ETc : actual crop evapotranspiration (mm)
PE : effective rainfall (mm)
IR: irrigation depth (mm)
RO: runoff (mm)
DP: deep percolation (mm)

The crop yield reduction in each stage was evaluated based on the degree of soil moisture
depletion due to supply of the crop evapotranspiration requirements. Eq. (6) calculates the
crop yield reduction in each stage and the cumulative crop yield reduction represented by Eq.
(7).

 Y   ETa 
1 − a  = K y 1 −  (6)
 Ymax   ETmax 
 Y  Y  Y  Y 
1 − a  = 1 −  a  ×  a  × ........ ×  a  (7)
 Ymax  i  Ym 1  Ym  2  Ym i

where,
i : crop growth stage
Ky : crop yield reduction factor
Ya, ETa : crop actual yield and evapotranspiration, respectively
Ymax, ETmax : maximum crop yield and potential evapotranspiration, respectively

4
After finishing the simulation of irrigation schedule for each crop, the CROPWAT model
could furthermore be used to estimate the monthly agricultural water requirements of an
irrigation scheme, based on different cropping patterns as expressed in the equation below:

 Acrop 
× 0.116 × Ascheme × ∑ (ETcrop − Peff )×
n
1
Q gross =  (8)
ep × t  i =1 Ascheme 
where,
Qgross: monthly agricultural water requirement of irrigation scheme (l/s)
ep : irrigation efficiency (<=1, dimensionless)
t : time operational factor (<=1, dimensionless)
i : crop index of the cropping pattern for an irrigation scheme
Acrop: crop planted area (ha)
Ascheme: total area of irrigation scheme (ha)
ETcrop: crop evapotranspiration (mm/day)
Peff: effective rainfall (mm/day)

Field experiments

The experiment site was at HsuehChia Experimental Station in southern Taiwan, which is
located at 23o13N, 120o11E and 4 m above MSL in sub-tropic area. The experiment station’s
monthly mean temperatures vary from 18.2oC in February to 28.5oC in June, monthly mean
relative humidity ranges from 76.9% in February to 75.5% in June, and monthly mean solar
radiation varies from 221.2 cal/cm2.day in February to 283.1 cal/cm2.day in June. A weather
station set up at the experiment site recorded regularly the air temperature, solar radiation, as
well as wind speed and direction from 1964 to 1999.

Crop evapotranspiration was measured both in the field and lysimeters, the same crops were
also planted in both sites in the same growing seasons. The lysimeters were constructed with
concrete of 3m * 5m * 3 m in size and layered with filters at the bottom, to measure runoff
and deep percolation water. The soil textures of the packed soil were of silty loam and loam.
In Table 1 are listed the soil textures and soil moisture constants at different soil layers. The
overall system consisted of twelve lysimeter blocks, and 10 plots in the field which were
supplied irrigation water with pipes.

All irrigation was carried out by the corrugation method for upland crop. Irrigation water was
applied to maintain the soil moisture at 50% of maximum available soil moisture during the
crop growth season. The gravimetric method was employed to determine the soil moisture
content. Soil moisture was determined by weighing the soil sample before and after irrigation
using oven drying for 24 hours. The soil sampling was made at 10 cm intervals to a depth of
60 cm.

The crop coefficient (Kc) was determined by dividing the measured actual ET by reference
crop ETo, as computed from the modified Penman equation. Table 2 shows the experimental
results of actual evapotranspiration and crop coefficient for corn.

5
Case study

The ChiaNan Irrigation Association (IA) is the biggest IA among the 17 IAs in Taiwan. It has
a total service area of 78,422 ha as of 1999. The Association, to service 182,903 farmers
members, has employed 678 staff members and assigned to 72 working stations located in the
service area of subtropical zone. The area has an average annual temperature of 21o-24oC, and
annual rainfall of 2500 mm of which 80% are concentrated in the wet season from May to
September.

Based on the field experiments and water management data, the CROPWAT model simulated
the on-farm water balance in the first paddy crop’s field. Figures 2 and 3 show the physiology
data for rice and corn. The results are summarized as shown in Table 3 and Figure 4. Table 3
shows the percentage of percolation for the single rice field which is 30.7% (295/962 *100) of
the crop water requirements, and 26% (296/1140 *100) for the double rice field. The average
of deep percolation in paddy rice fields is about 28.4% of the total crop water requirements.
Therefore, it is evident that the paddy fields function to recharge groundwater, and hence
remarkably preserve of ecological environment. CROPWAT model was also applied to
simulate the upland crops, including corn, sorghum and soybean, to evaluate the on-farm
water balance which results are as shown in Table 4 and Figure 5.

For evaluating the irrigation scheme’s monthly water requirements in HsuehChia Working
Station of the Chianan IA, two types of cropping patterns were defined based on the 1999
water management data as shown in Table 5. Meanwhile, Table 6 and Figure 6 show the
simulated results of monthly water requirements for the HsuehChia Working Station area. The
CROPWAT model simulated results show that the annual agricultural water demand is 537
mm, and the peak monthly scheme water requirements of 126 mm happened in the October.
For the double rice cropping pattern, the annual agricultural water demand is 1019 mm, and
the peak monthly scheme water requirements of 192 mm happened in the January.

Table 1 Soil texture and moisture constants in the experimental station area

Items Soil depth from ground surface (cm)

0-10 10-20 20-30 30-40 40-60 Average
Soil texture Silty loam Silty loam Silty loam Silty loam Loam
Saturation content 31.4 31.9 33.05 32.8 31.7 32.17
(% of dry weight)
Field capacity 24.8 20.2 20.4 21.9 19.5 21.4
(% of dry weight)
Wilting point 5.98 6.31 6.37 5.9 5.84 6.08
(% of dry weight)
Bulk density (g/cm3) 1.37 1.54 1.6 1.56 1.52 1.52

6
 Table 2 Actual ET and crop coefficient (Kc) at four growing stages of corn

( ) ( ) ( ) ( )

Seasons ETa Kc ETa Kc ETa Kc ETa Kc

(mm) (mm) (mm) (mm)

1986 40.4 0.58 122.6 0.94 169.5 1.02 92.2 0.75

1987 20.6 0.27 101.0 0.85 166.1 0.89 61.0 0.65
1988 21.4 0.36 61.1 0.54 130.7 0.77 87.3 0.74
Avg 27.5 0.40 94.9 0.78 155.4 0.89 80.2 0.71

Table 3 CROPWAT model simulating the on-farm water balance of paddy fields in ChiaNan Irrigation
Association’s area

Items Single rice (planted: Jan. 15) Double rice (planted: June 15)
Crop evapotranspiration (mm) 488 639
Percolation (mm) 295 296
Effective rainfall (mm) 202 584
Crop water requirements (mm) 962 1140
Crop water requirements (mm) 760.8 541.4

Table 4 CROPWAT simulating the on-farm water balance in upland crop fields in irrigation service area

Items Corn Sorghum Soybean

Spring Autumn Spring Autumn Spring Autumn
Planting date Feb 17 Sep 23 Feb 28 June 4 Mar 7 Sep 20
Crop ET (mm) 358.1 272.9 332.1 366.2 350 264.8
Effective rainfall (mm) 328.2 74.6 293.8 517.9 323.1 73.9
Crop water requirement (mm) 69.5 203.6 75.3 0 38.4 198.2

Table 5 Cropping patterns in service area in 1999

items Crop name Planting date Harvest date Area (%)

types
Corn Feb 15 Jun 14 45
Single rice Sorghum Feb 15 May 31 45
cropping pattern Rice July 15 Nov 12 65
Sorghum July 1 Oct 14 25

7
 Double rice Rice Jan 15 May 15 90
cropping pattern Rice Jul 1 Oct 29 90
Table 6 Monthly agricultural water requirements in service area

Month 1 2 3 4 5 6 7 8 9 10 11 12 Total
Type
Single rice 0 2 28 48 2 49 121 44 79 126 38 0 537
cropping pattern
(mm)
Double rice 192 89 118 100 7 177 42 36 77 109 0 72 1019
cropping pattern
(mm)

8
 Start

Input irrigation
Management basic
data

Crop Meteorological Soil

1. growth dates 1. temperature 1. soil texture
2. crop coefficient 2. wind speed 2. available soil moisture
3. root zone 3. sunshine hours 3. infiltration rate
4. crop yield coefficient 4. relative humidity 4. initial soil moisture
5. rainfall

1. Calculate reference evapotranspiration

(ETo): Penman Monteith method
2. Calculate effective rainfall: USDA soil
conservation method

Cropping Pattern
Input each crop's planted area, date and harvest date
within an irrigation scheme
C

Crop Irrigation Requirements

Simulate each crop's parameters within growth
seasons, including: (1) crop coefficient, (2) leaf index,
(3) crop evapotranspiration, (4) effective rainfall, (5)
percolation, (6) irrigation requirement

Irrigation Schedule
Five optins to choice:
B (1) Defined times, date, depth by users
(2) Irrigation at percentage of soil moisture
depletion (% RAM)
(3) Irrigation at fixed intervals per stage
(4) Irrigation at given ET or yield reductuoin
(5) No irrigation, only rainfall

Calculate Actual Crop Irrigation Requirement

Estimate the actual crop irrigation requirement,
including: (1) irrigation times, date and depth, (2) soil
moisture depletion, (3) actual evapotranspiration, (4)
deep percolation, (5) irrigation depth, (6) crop yield

Figre 1. Flow chart of CROPWAT irrigation management model

9
 A
B

No
Satisfy irrigsation
results or not ?

Yes

Irrigation requirements of Irrigation Scheme

To estimate the monthly irrigation requirements of irrigation
scheme with different cropping patterns

Changing
Yes cropping pattern
or not ?

No
Printout irrigation
results

End

Figre 1. Flow chart of CROPWAT irrigation management model (continued)

Table 6 Monthly agricultural water requirements in service area

Month 1 2 3 4 5 6 7 8 9 10 11 12 Total
Type
Single rice 0 2 28 48 2 49 121 44 79 126 38 0 537
cropping pattern
(mm)
Double rice 192 89 118 100 7 177 42 36 77 109 0 72 1019
cropping pattern
(mm)

10
 Fig. 2 Rice physiology data in each growth stage for CROPWAT model

Fig. 3 Corn physiology data in each growth stage for CROPWAT model

D o u b le R ice C r o p

180
ETc (mm/dec)

160
140
120
100
80
60
40
20
0
M ay M ay Jun Jun Jun Jul Jul Jul A u g A u g A u g Sep Sep Sep O ct O ct
2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2
D ecade

Evapotranspiration Effective rainfall Irrigation

Fig. 4 CROPWAT simulated the on-farm water balance for double rice crop

11
 Spring Corn
ET, Rain, Irr(mm/dec)
60
50
40
30
20
10
0
Feb Feb Mar Mar Mar Apr Apr Apr May May May Jun 1 Jun 2
2 3 1 2 3 1 2 3 1 2 3
Decade
Evapotranspiration Effective rainfall Irrigation

Fig. 5 On-farm balance for spring and autumn corn

250
water rquirement (mm)

200
Monthly irrigation

150

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12
Month

Single rice cropping pattern Double rice cropping pattern

Fig. 6 Monthly agricultural water requirements of service area of ChaiNan Irrigation Association

12
Conclusion

(1) The irrigation water requirements for spring and autumn corn are 69.5 mm and 203.6 mm;
and for soybean, 68.4 mm and 198.2 mm, respectively. The great difference of irrigation
requirements between the spring and autumn crops can be explained by the amounts of
effective rainfalls. For instance, the autumn corn could utilize an amount of PE only about 74
mm; while the spring corn, about 328 mm.
(2) The CROPWAT model simulated results show that for the single rice cropping pattern
area, the annual agricultural water demand amounted to 537 mm, which peak of 126 mm
monthly scheme water requirements happened in the October; and for the double rice
cropping pattern area, the annual demand was 1019 mm, which peak of 192 mm monthly
scheme water requirements happened in the January.
(3) The percentage of percolation in the single rice fields reaches to about 30.7% of the crop
water requirements, and in the double rice area, 26%, which give an deep percolation of
28.4% of the totally crop water requirements. Therefore, the paddy fields can recharge
groundwater, and hence have remarkable ecological environment function.
(4) The field experiment data obtained in 1986 up to 1995 from the HsuehChia Experiment
Station were used to calculate the crop coefficients for upland crops, and then applied to
irrigation management model which could exactly evaluate the agricultural water
requirements. Therefore, it is impressive to conclude that the field experiments are very
essential in support of the agricultural water management.
(5) This study hence shows a more important hint that the CROPWAT irrigation management
model could be used to effectively and efficiently to estimate the agricultural water
requirements with different cropping patterns. This will facilitate the IAs in Taiwan to meet
the possible change of irrigation management after Taiwan entering the WTO, due to the
expected impacts to agricultural sector.

Acknowledgements

The authors would like to thank the Council of Agriculture of the Republic of China for financial support and
technical assistance and the HsuehChia Experiment Station for their help in this study.

Reference

Allen, R.G. and W.O. Pruitt. (1991). "FAO-24 Reference Evapotranspiration Factors." J. of
Irrig. and Drain. Engrg., ASCE, 117(5):758-773.
Cheong, B.H. (1992). "Improving Irrigation System Performance for Rice Culture." Ph.D.
Dissertation, Utah State University, Logan, Utah.
Hargreaves, G.L., G.H. Hargreaves and J.P. Riley. (1985). "Agricultural benefits for the
Senegal River Basin." J. Irrig. and Drain. Engrg., ASCE, 111(3):265-275.
Hill, R.W., R.J. Hanks and J.L. Wright. (1987). "Crop Yield Models Adapted to Irrigation
Scheduling Programs," Proc. Irrigation Systems for the 21st Century Proceedings
International Conference on Genetic Algorithms, Portland, OR.

13
Jensen, M.E. and H.R. Haise. (1963). "Estimating Evapotranspiration from Solar Radiation."
J. of Irrig. and Drain. Engrg., ASCE, 89:15-41.
Keller, A.A. (1987). The USU Unit Command Area Model. Water Management Synthesis II
Project, WMS Report No. 71, USAID. Logan, UT
Kuo, ShengFeng (1995). “Decision Support for Irrigated Project Planning Using a Genetic
Algorithm. “ Ph.D. Dissertation, Utah State University, Logan, UT.
Pleban, S., and Israeli, I. (1989). “Improved Approach to Irrigation Scheduling.” J. of Irrig.
And Drain. Engrg., ASCE, 115(4):577-587.
Prajamwong, S. (1994). "Command Area Decision Support System for Irrigation Projects."
Ph.D. Dissertation, Utah State University, Logan, Utah.
Shih, C.C. (1986). “Irrigation water requirements experiment for corn”, Final report, Council
of Agriculture, R.O.C.
Shih, C.C. (1997). “Study on the crop coefficient of soybean.” Taiwan Conservancy, 45(2):1-
11.
Smith, Martin (1991). “CROPWAT: Manual and Guidelines.”, FAO of UN, Rome.
Wright, J.L. (1982). "New Evapotranspiration Crop Coefficients." J. of Irrig. and Drain.
Engrg., ASCE, 108(2):57-74.

14