Evapotranspiration and Irrigation Water
Evapotranspiration and Irrigation Water
Evapotranspiration and Irrigation Water
ABSTRACT
INTRODUCTION
Studies of projected water demand and supply in Jordan have shown that the
water deficit is increasing with time, with demands on a finite quantity of good
quality water ever increasing. Per capita availability of renewable water today is
less than 175 cubic meters, already far below the projected regional average for
the year 2025, and will decrease to 90 cubic meters by the year 2020 if water
projects are not implemented (Taha and El-Nasser, 2002). In order to meet
minimum current water demands for basic uses, over-pumpage of groundwater
resources in Jordan is estimated to be at 100% above the safe-yield. Of the total
groundwater supplied to all uses in 1998 (485 MCM), irrigated agriculture
consumption accounted for about 53%, at some 258 MCM. Nearly 80% of the
1
Engineer, Ministry of Water and Irrigation, Amman, Jordan,
Suzan_Taha@MWI.gov.jo
2
Professor, University of Idaho, 3793 N. 3600 E., Kimberly, Idaho 83341, 1 208
423 6601, RALLEN@kimberly.uidaho.edu
3
AHT International GmbH, Essen, Germany
4
Chief Engineer, Ministry of Water and Irrigation, Amman, Jordan
5
Ministry of Water and Irrigation
1
2 USCID/EWRI Conference
safe yield of the renewable groundwater resources and 40% of the non-renewable
groundwater are currently used for irrigated agriculture throughout the country.
One of the most important tools implemented for the calculation of water demand
at MWI is the Irrigation Model. The Irrigation Model consists of 3 separate
modules for pre-processing climatic data, calculation of reference ET (ETo) and
computing monthly net irrigation requirements of crops (NIR). These
calculations are combined to estimate present and future irrigation demands for
the whole country and selected planning/development regions. The Irrigation
Model is GIS-based (Jacobi, 2001) and is linked to a Relational Database
Management System under Oracle thus allowing for the updating of data and
visual presentation of the results.
The Irrigation Model is implemented under the Microsoft Access database, and is
linked to a central Oracle database and GIS databases containing Arc View shape
files. The software system models irrigation water demand in the future based on
existing information regarding cropping patterns, irrigated areas, information on
crop water requirements, in combination with application and conveyance
methods, and leaching requirements given water quality. Prediction of future
6
German Agency for Technical Cooperation
7
Conceptual Design was done by Ministry of Water and Irrigation, whereas
software development was by AHT International GmbH in close cooperation with
GTZ.
Irrigation and Drainage 3
water demand is tied to a reference year demand assessment. This permits the
evaluation of the current irrigation demand situation in the spatial unit under
consideration, with respect to water availability and water quality, prior to
proceeding with assumptions regarding the future. Irrigation demand data are
aggregated to demand centers such as towns and villages, developed irrigated
areas and project areas that are geographically represented via point information
in ArcView GIS.
The Irrigation Model requires entry of various data from databases, including:
Irrigated areas for each crop group and demand center
Distribution of irrigation methods for irrigated areas under the various
application methods
Leaching requirements for every crop group and salinity class
Monthly net water requirements (NIR) for each crop group, given the
agroclimatic zone in which it is grown
Application efficiency tables
The NIR_Calculator was originally coded in a Lotus 123 data base during the
early 1990’s UNDP work, but has been since converted to an Access application
using Visual Basic programming (Jacobi, 2001). Various information compiled
by NIR_Calculator includes:
4 USCID/EWRI Conference
Length of the crop development stage; initial, development, mid and late
Crop Factors:
Crop coefficients are organized and specified by crop and irrigation method
(surface, sprinkler, drip), and cultural environment (open field, greenhouse, open
field with plastic mulch, and time of year). This organization allows the values
for Kc to be customized to reflect effects of the specific irrigation system type and
environment on soil evaporation.
Irrigation and Drainage 5
Two separate means for processing rainfall and ET were applied by the Ministry
and GTZ: a) the calculation of effective rainfall and ET for historic years and b) a
statistical evaluation for dry, median and wet years calculated for agro-climatic
zones rather than for individual stations.
The calculations in the NIR_Calculator are subdivided into 5 steps (Figure 2):
Step 1: Calculation of the crop coefficient during the initial growth period, Kc ini
Step 2: Calculation of crop evapotranspiration using ETo and crop factors
Step 3: Calculation of net effective precipitation during non-growing season
Step 4: Calculation of NIR
Step 5: Export of NIR results to the relevant database table.
PERTINENT CALCULATIONS
This section describes some of the ET and NIR procedures that were developed
during the study for use in NIR_Calculator. Several enhancements have been
made to the NIR_Calculator to account for effects of late season depletion, off-
season precipitation, and preplant irrigation on annual NIR.
The NIR for any month “i” is calculated for outside environments as:
NIRi Maximum ( ETc i - R eff i ) Daysi , - 0.5 RAWi PPIi LSD i (1)
where ETc i and Reff i are daily crop ET and effective rainfall for the month (or
period) i, Daysi is the number of days in the month, RAWi is the maximum
6 USCID/EWRI Conference
readily available moisture for month i, PPIi is any net preplant irrigation
requirement for month i (0 unless immediately prior to planting), and LSD is late
season depletion during month i (0 unless during the month or months
immediately prior to harvest).
The maximum of net ETc i and -0.5 RAWi in (1) insures any excess precipitation
during rainy months (where ETc i - Reff i is negative) does not exceed the average
ability of the soil to retain the excess precipitation in the root zone. This average
ability (or capacity) is estimated to be one half of the maximum allowable
depletion of the soil for any month. This way, the total NIR includes the benefit
and effect of useable excess precipitation during the negative value months.
When negative values of NIR are encountered that are caused by effective
precipitation exceeding ET requirements, the NIR sub-module calculator
compares the negative values to those for the following month to insure that the
sum of negative NIR for two or more consecutive months is never more negative
than the –RAW. This prevents more storage in the soil than is possible. Excess
storage is assumed to become deep percolation from the root zone, and ultimately,
ground-water recharge. The RAW used in (1) is based on the maximum rooting
depth for the crop. Negative values for cumulative CNIR at the end of the
growing season are discarded and set to zero. Table 1 illustrates calculation of
NIR with the –RAW limit invoked during wet months.
Table 1. Example NIR calculation for citrus in zone 5 of Jordan for a wet season,
months Dec. - Feb.
In the case of the citrus example in Table 1, the RAW equals 102 mm, with only
one-half of this allowed to be added to soil storage during any month. Beginning
with December (the first month having a negative NIR), one would have, by the
end of January, a summed negative NIR = -102 mm, which is just OK (i.e., less
than or equal to RAW). However, by the end of February, which is the third
consecutive month having negative NIR, one would have a summed negative NIR
over the three months = -137 mm, which is more "carryover" than the soil can
hold. Therefore, the NIR for February is set to 0 mm.
This example presumes that the soil is depleted to RAW at the time of harvest.
This is a reasonable assumption for many field crops. For some crops, however,
Irrigation and Drainage 7
management may keep the soil moisture at higher levels, for example for fruit
quality or to make harvesting easier, so that RAW is too large a value to use in
estimating month to month storage carryover. One may tend to use a smaller
value for RAW in those instances.
Late season depletion accounts for the practice of utilizing moisture stored in the
root zone near the end of the growing season, without replacing it. This has the
effect of reducing the within-season NIR. End of season soil moisture depletion
is assumed to be later replenished by precipitation during off-season or by a future
pre-plant irrigation. Late season depletion is computed in the NIR Module as the
minimum of RAW and a user-specified maximum allowed depletion (USMD) due
to cultural or other requirements (Allen, 1991):
RAW AM
MAD
R z max (3)
100
where AM is the total available water stored in the root zone (one or two days
after irrigation) represented by the soil water content between field capacity and
wilting point, and Rz max is the maximum rooting depth. AM varies with the type
of soil and averages about 150 mm/m soil depth. The MAD represents the level of
maximum allowed soil water depletion tolerated to maintain potential crop growth
and it varies with the type of crop. Values for MAD and Rz max are found in
Doorenbos and Pruitt (1977) and Allen et al., (1998).
The USMD was set to 10 mm for root crops such as carrots, potatoes and onions
to assure moist soil during harvest to facilitate extraction of the crop from the soil
without damage. For other crops, such as wheat and barley, which tolerate higher
depletion levels, USMD was set to 100 mm.
CNEPO R eff i K c ngs i ETo i Daysi
n
(4)
1
where n is the number of months or periods outside the growing season, and
Kc ngs i is Kc during month i of the nongrowing season. Beginning season
depletion is calculated as LSD – CNEPO, with a lower limit of 0.
ADI is the moisture required to moisten the seed bed and is computed as:
ADI AM
MAD
Rz I (5)
100
where Rz I is root depth during the initial period. Rz I for annual crops is
estimated as the planting depth of the seed plus 5 to 10 cm to represent upward
movement of moisture toward the seed and rapid root development. A typical
value is 0.15 m. For perennial crops, a typical value is 0.7 m.
MPID is the minimum depth of water that can physically be added to the soil due
to constraints in the irrigation application system. A surface system may have to
apply 40 mm average irrigation depth to push enough water across the basin or
along the furrow to just infiltrate 5 mm at the furthest point. Because drip and
sprinkler irrigation systems are better controlled, the MPID for these systems is 0.
The Kc during midseason (Kc mid) and during the late season (Kc end), i.e., the
second and third anchor points for the FAO style of Kc curve, are adjusted for
predicting ET inside plastic greenhouses by multiplying by 0.75 based on work by
Mazahrih (2001) in Jordan Valley.
Irrigation and Drainage 9
Kc during the initial period, Kc ini, is calculated using Cuenca (1987) for
reproducing Fig. 6 of Doorenbos and Pruitt (1977). This equation was employed
for consistency with past usage by the Ministry. However, the application can
also be applied using Kc ini equations of FAO-56. Wetting events include both
rainfall and irrigation. In NIR_Calculator, the Kc ini method was applied to both
the initial period of crops and to periods between crops (nongrowing periods).
Otherwise, if Kc ini ETo <= Reff, then Iirrig is set to an arbitrary 50 days. The Iirrig
is considered to be necessary, and followed by farmers, to maintain adequate
moisture in the seedbed for germinating and establishing the crop. In this manner,
impacts of Iirrig on evaporation are incorporated.
An average wetting interval is calculated to consider wetting by both rainfall and
irrigation, based on a geometric mean (Allen, 1991). The interval mimics an
irrigation schedule during the initial period that maintains sufficient moisture in
the upper soil layer conducive to seed germination and root development:
I M Int 1
1
1
(10)
I rain I irrig
1
This new estimate for IM is reinserted into (8) for Kc ini and a new value for Kc ini
is calculated. The product Kc ini ETo is again recomputed and compared to Reff
10 USCID/EWRI Conference
and a new Iirrig is computed. The process is repeated until Kc ini and IM have
stable values.
For perennials, the calculation of Kc ini requires modification, since the ground
surface is not entirely bare, so that the basal condition, Kcb > 0.15. The
calculation of Kc ini perennials considers evaporation from both rainfall and
irrigation. However, the evaporation is reduced due to the shading effects of
vegetation. The “optimal” irrigation frequency is determined to retain sufficient
soil moisture in the initial root zone, which for a perennial is assumed to be
substantially close to maximum rooting depth. In the following computations for
Kc ini perennials, the values for ETo, Irain, and Reff are averages over all months in
the initial period. This procedure, based on Allen (1991 and 2001), uses the
Cuenca equation to retain consistency with early work within the Ministry, but is
applied to evaporation from a partially vegetated surface from FAO-56, and can
be applied with the Kc ini equations of FAO-56.
with limits of 0.01 < few < 0.99, where Kc ini_residual is the same as the basal Kcb
for a perennial crop, and represents ET for the crop when it has a dry soil surface.
It is assumed that evaporation from rainfall occurs in the exposed portion of the
field, in between beds (i.e., in few), when there is no plastic mulch or when the
plastic mulch is covered by sufficient soil. For surface or sprinkle irrigation, the
total "drying" of the soil is computed as a weighted average based on few. The
irrigation frequency of perennials for drip with plastic mulch along the beds is
less coupled with rainfall. For drip, it is assumed that the bed is mulch covered so
that irrigation does not substantially impact Kc ini. The same basic equation for
Kc ini, i.e. that by Cuenca (1987) is used, but with ratioing according to few.
The evaporation from the few area is added to the basal Kc ini residual
where Kc ini (Eqn. 8) is Kc ini from (8). Limits are placed so that:
K c ini peren Minimum K c ini peren ( Eqn.12) , Maximum( 1.25, K c ini residual ) (13)
Irrigation and Drainage 11
Similar to the condition for bare soil, an “optimal” irrigation interval, Iirrig, is
recommended if predicted evaporation exceeds effective rainfall:
If K c ini peren ETo R eff then I irrig
ADI
K c ini peren ETo R eff
(14)
where ADI is water that is readily available in the rooting zone during the initial
period. If Kc ini peren ETo <= Reff, then Iirrig is set to an arbitrary 50 days. (14) is
not invoked for drip irrigation, where it is assumed that the surface is generally
wetted within the shade of the vegetation, so that no extra evaporation from
irrigation occurs. Therefore, for drip, Iirrig = 50 days. Once a value for Iirrig is
predicted, a geometric mean wetting interval (IM) is calculated using (10). The
process iterates on Kc ini peren and Iirrig and IM until values are stable.
NonGrowing Season ET
The Kc ini function (8) is applied during the nongrowing season to predict total
evaporation and associated effectiveness of precipitation during the nongrowing
season. This calculation is part of the prediction of the net change in soil water
storage during the period from harvest of one crop until the planting of another.
Application of the Kc ini calculation is made on a monthly basis. The calculation
assumes that the soil surface is void of green vegetation. Where green vegetation
exists during the nongrowing season, Kc ini peren from (12) can be used to
approximate Kc ngs. In the application to the nongrowing period, it is assumed
that there is no irrigation, so that IM in (8) is set equal to IRain, and no iteration is
necessary. However, an upper limit is applied, so that:
where Kc ngs is the Kc during the nongrowing season and is from (8). This
conditional limits evaporation to effective rainfall.
Following the calculation of monthly NIR, results are exported and saved in the
relevant tables in the Central Oracle database. Irrigation demand assessment
proceeds using the irrigation demand model, which can be linked on-line to this
database. Figure 3 illustrates the user interface used in the irrigation model to
specify irrigation types and efficiencies, and resulting water demands.
Figure 4 illustrates how monthly water demand data are presented to the user by
main crop group (MCG) and month, in tabular and graphical form, under dry,
median and wet conditions.
12 USCID/EWRI Conference
Comparisons between irrigation demand predicted from the irrigation model and
water diversions and deliveries recorded by the Ministry are providing feedback
concerning any need to modify (i.e., reduce) Kc values in NIR_Calculator to
reflect impacts of water stress, salnity, low plant density, and water management
on total ET. In some areas, the “pristine” assumptions of FAO Kc factors (Allen
et al., 1998) may overpredict total ET from an area. Therefore, reducing factors
may need to be developed.
Irrigation and Drainage 13
Figure 4: Example computation of monthly historic demand for main crop groups
(MCG) (from Taha and El-Naser, 2001).
The irrigation demand model is one of several tools that have been developed to
enable the management of the present and future water balance in Jordan. The
model aims at prediction and management of irrigation demand given certain
development scenarios and based on recent data. As such, the model can be used
to evaluate present irrigation demands with respect to actual water use and
availability, prevailing cropping patterns, irrigation methods and efficiencies. The
tool’s reporting flexibility allows digital output of various types of data for
additional processing. Being GIS-based, spatial examination and analysis of both
the present and future irrigation demands can be made. In addition, the digital
nature of the model and its database dependency allows for easy updating of the
tool. These features permit more flexible responses to merging new realities and
changing conditions.
The interactive feature of the model allows testing the impact of various
development scenarios on irrigation demand and subsequently on water
balancing, thus allowing for review of water strategies. In addition, scenarios
reflecting various water sector strategies and policies can be examined using the
model, and as such, the tool can be used as one of several elements to support
decision making with regards to demand and supply management strategies, and
hence the optimization of water resources.
14 USCID/EWRI Conference
Acknowlegements
Mr. Lothar Nolte of GTZ, Amman, provided valuable coordinating and
interfacing. Dr. Peter McCornick of ARD, Inc., Amman, provided background
and support. Engineer Rasha Sharkawi of MWI assisted with Access data base
operation and NIR_Calculator debugging. Work by R.G. Allen was supported by
ARD, Inc. with funding by US-AID.
REFERENCES
AHT International GMBH (2001). Water Sector Planning Support at the Ministry
of Water and Irrigation, Digital Balancing Model for Water Resources
Management, 2nd Phase. PN 92.22.38.1-01.100. Final Report, February 2001.
Allen, R.G. 1991. Net Irrigation Requirements Calculator using Spreadsheets.
Final Report to UNDP and Ministry of Water and Irrigation. Amman, Jordan.
Allen, R.G., L.S. Pereira, D. Raes, and M. Smith. 1998. Crop
Evapotranspiration: Guidelines for Computing Crop Water Requirements.
United Nations FAO, Irrig. and Drain. Paper 56. Rome, Italy. 300 p.
Allen, R.G. 2001. Review and support of developments in the Net Irrigation
Requirement Calculator, Jordan Water Policy Support, Crop Water
Requirements Specialist, Report to ARD, Inc., March, 2001.
Cuenca, R.H. 1987. Irrigation System Design: An Engineering Approach.
Prentice Hall, Englewood Cliffs, New Jersey. 552 p.
Doorenbos, J. and Pruitt, W. O., 1977. "Crop water requirements." Irrigation
and Drainage Paper No. 24, (rev.) FAO, Rome, Italy. l44 p.
FAO, 1990. Report on the Expert consultation on revision of FAO methodologies for
crop water requirements. Annex V, Rome Italy. 26 p.
Taha S., and H. El-Naser. (1999). Water Resources Management in Jordan: The
Use of Interactive Digital Planning Tools based on GIS applications.
Presented to the Seminar on Information Technology applied to resources
management, May 1999, Petra, Jordan
Taha, S. and H. El-Naser. 2001. Projection of Irrigation Demand in Jordan:
Introduction to GIS-based Model for the Projection and Management of
Irrigation Demand. Paper pres. at MedAqua Conf., Amman, Jordan, 6/2001.
Jacobi, K. 2001. Calculation of Net Irrigation Requirements (NIR) with Access
Application NIR_calc.mdb. Report submitted to Jordan Ministry of Water
and Irrigation, Amman, Jordan. 18 p.
Mazahrih, Naem Thiyab. 2001. Evapotranspiration Measurement and Modelling
for Bermuda Grass, Cucumber, and Tomato Grown Under Protected
Cultivation in the Central Jordan Valley. Ph.D. Dissertation, Univ. Jordan,
Amman. 168 p.