Agricultural Water Management 199 (2018) 157–174

Contents lists available at ScienceDirect

Agricultural Water Management

journal homepage: www.elsevier.com/locate/agwat

Advances in designing drip irrigation laterals

Giorgio Baiamonte ∗
Dipartimento di Scienze Agrarie, Alimentari e Forestali, Università degli Studi di Palermo, viale delle Scienze edificio 4, 90128, Palermo, Italy

a r t i c l e i n f o a b s t r a c t

Article history: It is known that using paired laterals, in which two distribution pipes extend in opposite directions
Received 20 June 2017 from a common manifold, contribute to increasing water use efficiency (WUE). Recently, an analytical
Received in revised form procedure to optimally design paired drip laterals on uniform slopes was proposed. More recently, this
23 November 2017
design procedure was simplified by deriving simple explicit relationships, as a function of 16 calibration
Accepted 18 December 2017
constants, with relative errors that were less than 2%. In this paper, further simple design relationships
are derived that require only 3 calibration constants, thus more readily obtainable results are produced
Keywords:
and the influence of the flow rate and diameter exponents of resistance equation are made more evident.
Microirrigation
Sloped/horizontal drip laterals Simple monomial relationships were also extended for the cases in which the lateral is laid on flat fields,
Design relationships on upward fields, and also considering layouts in which the manifold is located at the boundary, instead of
Optimal lateral length inside, of the irrigation unit. For the five considered layouts, evaluations of the effects of design choices in
terms of energy saving and comparisons between optimal lateral lengths, are carried out. Finally, simple
linear relationships linking lateral pressure head tolerance (␦), coefficient of variation of pressure heads
(CV) and emission uniformity coefficient of Keller and Karmeli (EU) are presented. Results showed that
for any design solutions associated with the considered layouts and for ␦ < 0.1, the proposed procedure
allows optimal design of the lateral, optimizing WUE, assuring low values of pressure head variability
(CV < 6.1%) and high values of emission uniformity coefficient (EU > 95%).
© 2017 Elsevier B.V. All rights reserved.

1. Introduction The initial investment equipment cost, which may be higher

Microirrigation systems saves water by allowing water to drip
slowly to the roots plants, either onto the soil surface or directly
near the root zone, through a network of valves, pipes, tubing,
and emitters. It is done through narrow tubes that deliver water
directly to the base of the plant. Thus, they are considered conve-
nient and effective means to supply water directly along individual
crop rows and maintain the current levels of crop production; as a
consequence, microirrigation can reduce water requirements and
operation costs, in respect to other irrigation systems (Keller and
Bliesner, 1990). A lot of factors influence the diffusion of microir-
rigation, such as the water price, the cost of the equipment, and
the crop price. Some factors are related to the characteristics of the
environment, as the precipitation regime (Agnese et al., 2014) and
the soil hydrological characteristics, which may present high spa-
tial variability even though the same land management is assumed
(Genius et al., 2012; Baiamonte et al., 2017).
∗ Corresponding author.
E-mail address: giorgio.baiamonte@unipa.it
than that of traditional methods, can make farmers unwilling to
choose the microirrigation system. However, in the long term, costs
can be reduced through a substantial reduction in labor cost; water
only needs to be regulated, which usually can be done via automat-
ically timed devices. Costs can also be reduced by optimizing water
use efficiency (WUE) and energy saving, which can be attained
when the submain and lateral units are well designed so as to
maximize the emitter uniformity.
Wu et al. (2010) developed a simple gravity-fed drip irrigation
system design procedure for low-cost, single-manifold subunits
with multiple pressure sections, by assuming that the laterals are
laid along elevation contours so that the ground slope along the
lateral is equal to zero. Moreover, it required iterative calculation
process for determining the number of emitters per lateral and
the lateral length. Keshtgar et al. (2013) took emitter uniformity to
extremes for a drip irrigation system, presenting a design procedure
based on variable length microtubes as emitters along the laterals,
thus attaining total emission uniformity. However, the suggested
criterion might be difficult to use in practice.
In designing drip irrigation laterals with traditional emit-
ters, a well-accepted practice consists in limiting the variation of
the pressure head to about ± ␦ of its nominal value along the lateral,

https://doi.org/10.1016/j.agwat.2017.12.015
0378-3774/© 2017 Elsevier B.V. All rights reserved.
158 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Nomenclature ␤ [−] Curvature parameter

␥n Stieltjes constants
A(i) Function of the emitter i ␦ [−] Pressure head tolerance
B01 Numerical constant  [m2 s−1 ] Kinematic water viscosity
B02 Numerical constant ␨ Riemann zeta function
B0 Numerical constant
B1 Numerical calibration constant
B2 Numerical calibration constant where the pressure head tolerance, ␦, can be assumed to be
Cv Manufacturing coefficient of variation around 10%, depending on each design and, in particular, on the
CVmin Minimum value of the coefficient of variation accepted flow rate variability of the emitters installed along the
D [m] Internal pipe diameter laterals.
EU Emission uniformity coefficient This design criterion is also considered in the manufacturers’
F0 Numerical calibration constant catalogues, which usually provide the maximum lateral lengths, for
g [m/s2 ] Acceleration of gravity fixed nominal emitter pressure heads, flow rates, emitter spacing
hi [m] Pressure head of the generic emitter i and pipe diameters for a manifold that is located at the boundary
hi (u) [m] Pressure head of the i-th emitter in the uphill lateral of the irrigation sector and for flat fields, so that the pressure head
hi (d) [m] Pressure head of the i-th emitter in the downhill stays in a certain admitted range of variability.
lateral However, in natural fields, differences in elevation are usual
h*min (d) [m] Minimum normalized pressure head at the in laterals of significant length, so that the values suggested in
downhill end of the lateral the catalogues that manufacturers provide cannot be used with-
h*max (d) [m] Maximum normalized pressure head at the out compromising the water distribution uniformity. Sometimes
downhill end of the lateral in the catalogues, a note warns that “the suggested lengths are sub-
h*min (u) [m] Minimum normalized pressure head at the ject to change due to changes in the conditions in which the tests were
uphill end of the lateral performed” and that “the lengths are valid for drip laterals laid in
h*max (u) [m] Maximum normalized pressure head at the flat fields”. Some companies suggest using pressure-compensating
uphill end of the lateral emitters (PCE), in order to overcome the effect of the slope on
hn [m] Nominal emitter’s pressure head the lateral length “of sloping fields and for very long drip later-
h*n [m] Normalized nominal emitter’s pressure head als where the pressure at the beginning and at the end of the drip
H(.,.) Generalized harmonic number line presents considerable differences that would considerably affect
i [−] Generic emitter of the lateral counted from the man- the scope of the individual emitters”. However, compared to non
ifold connection pressure-compensating emitters, such emitters are more expan-
imin [−] Number of emitters in downhill lateral, from the sive, more complicated structurally and the working mechanism,
manifold connection to the section with minimum which causes difficulty in their research and development, is not
pressure head clear (Chu et al., 2015). Moreover, although in the last decades PCE
K (−) Friction head loss gradient parameter technology was improved, in the past showed contraindications
L [m] Maximum length of the lateral suggested by a man- (Madramootoo et al., 1987; Madramootoo, 1988) and are not sug-
ufacturer (Table 3) gested for low pressure heads (Perea et al., 2013). Generally, it is
Ld [m] Optimal length of the downhill lateral agreed that they will not last as long as non-pressure compensating
Lopt [m] Optimal length of the paired lateral emitters because of the life span of the elastomer.
Lu [m] Optimal length of the uphill lateral The design criterion to control pressure heads of traditional non
mCV [−] Angular coefficient of ␦ − CV relationships pressure-compensating emitters has been widely applied in litera-
mEU [−] Angular coefficient of ␦ − EU relationships ture also for sloping laterals, it has been considered in many papers
n [−] Total number of emitters in the entire lateral describing the basic hydraulics of drip irrigation design (Karmeli
nd [−] Number of emitters in the downhill lateral and Keller, 1975; Wu and Gitlin, 1975; Wu et al., 1983; Baiamonte
nopt [−] Optimal number of emitters in the entire lateral et al., 2015; Baiamonte, 2016a). Moreover, disposing paired laterals
np [−] Number of emitters per plant so that two distribution pipes extend in opposite directions from
nu [−] Number of emitters in the uphill lateral a common manifold contributes to provide more uniform pres-
nx [−] Generic emitter of the lateral counted from the sure to all laterals in the system, increasing water use and energy
uphill end of the lateral efficiency of microirrigation, and maximizing the lateral length (Al-
qav [m3 s−1 ] Average emitter flow rate in the lateral Samarmad, 2002).
qmax [m3 s−1 ] Maximum flow rate in the lateral Keller and Bliesner (1990) stated that the best position of the
qmin [m3 s−1 ] Minimum flow rate in the lateral manifold (BMP) is achieved, when the same minimum pressure
qn [m3 s−1 ] Nominal emitter flow rate head in the uphill and in the downhill lateral occurs and developed
RE [%] Relative error graphical and numerical solution methods, while previously Wu
S [m] Emitter spacing et al. (1983) had developed design charts allowing to determine
S0 [−] Slope of the lateral it. For horizontal paired laterals, the uphill lateral length equals
S0 max [−] Maximum slope value of the lateral, designed as the downhill lateral length (BMP = 0.5). Whereas, for any other field
the field is consider flat slope, the manifold has to be shifted uphill, so to balance the dif-
S0 CVmin [−] Slope of the lateral designed as the field is con- ferences in elevation and pressure losses on both laterals’ side.
sider flat, for which the minimum coefficient of Kang and Nishiyama (1996) used the finite element method to
variation CVmin occurs analyze the pressure head and discharge distribution along lat-
x [−] Exponent of the flow rate-pressure head relation- eral and submain pipe. The golden section search (Gill et al., 1989;
ship Pinthong et al., 2013) was applied to find the operating pressure
heads of lateral and the BMP corresponding to the required unifor-
mity of water application.
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174
More recently, according to the aforementioned definition of
BMP provided by Keller and Bliesner (1990); Jiang and Kang (2010)
used the energy gradient line approach (Wu 1975; Wu and Gitlin
1975, Wu et al., 1986) to derive analytical solutions to deter-
mine BMP. Pinthong et al. (2013) developed a model to analyze
the steady-state hydraulics of branching pipe networks as found
in on-farm pressurized irrigation systems with multiple outlets
at sprinklers or emitters. The model used a new methodology to
identify all flow paths based on a shortest path algorithm. It also
uses elevation, obtained by high-resolution topographical data, and
hydraulic parameters to determine which outlets are flowing and
which are not. An iterative solution approach resulted in the cal-
culated flow rates and pressures at all nodes in the network with a
single source node, which can have a fixed head. The model is based
on a methodology that starts at the downstream end of a branching
pipe system and works in the upstream direction to determine flow
rates and pressures at each node.
The forward Step-by-Step (SBS) procedure, as well-known, rep-
resents the most affordable method to evaluate the pressure heads
and the actual flow rates that correspond to all of the emitters in
the lateral. However, when it is applied from the uphill end to the
downhill end of the lateral, time consuming iterations and trials
are needed to find optimal solutions (Al-Amoud, 1995; Kang et al.,
1996; Kang, 2000).
The design process of the drip lateral can be simplified with the
assumption that local losses can be neglected, i.e. in all of the cases
for which variations of flow velocity due to the emitters is limited
and the morphology of emitter connections does not produce sig-
nificant reductions of the lateral cross section. This could be the
case of flat in-line coextruded drippers that are installed parallel
to the pipe within the manufacturing process with relatively small
profile to minimize flow disorders in the drip lateral. For in-line
emitters characterized by an ␣ fraction of kinetic head flow ≤ 0.3
and emitter spacing equal to 1 m, local losses less than 10% of the
total losses were found by Provenzano and Pumo (2004).
Under the Blasius resistance equation, Baiamonte et al. (2015)
presented an analytical approach to evaluate the optimal length of
paired drip laterals placed on uniformly sloped grounds by assum-
ing that: i) the variations of emitters’ flow rate along the lateral,
as well as ii) the local losses due to emitters’ insertions could be
neglected. Although the latter assumption can be accepted only in
some cases, and can be fully relaxed in the exact SBS procedure
which already provide complete and exact solutions, the useful-
ness of analytical solutions is also important to consider. They don’t
need iterations to be coded in computer programs, give easy cog-
nition of how the factors affecting the studied feature influence the
output variables and serve a useful pedagogical purpose (Anderson
and Woessner, 1992; Giraldez and Woolhiser, 1996; Baiamonte and
Singh, 2016a, 2016b; Baiamonte, 2016c).
The analytical solution of Baiamonte et al. (2015) provided the
maximum number of emitters in the uphill and downhill sides of
the lateral and, therefore, to the optimal lateral length by impos-
ing constant emitter flow rates and a pressure head tolerance,
␦ = 10%, of its nominal value along the entire lateral. More recently,
Baiamonte (2016a, 2016b, 2018) simplified the analytical design
procedure introduced by Baiamonte et al. (2015), which required
solving a system of four implicit equations, by deriving simple
explicit relationships as a function of 16 calibration constants. This
simplified procedure derived the design variables required for the
optimal paired lateral, with a relative error less than 2%. More-
over, the author identified exactly the best position of the manifold
(BMP) associated with the optimal paired lateral design in a sloping
field, which resulted for flat fields equal to 0.24, whereas for com-
plex field topography, this value differs from 0.24 depending on the
concavity or the convexity shape of the paired lateral (Baiamonte,
2017a). It is interesting to observe that BMP = 0.24 (for uniform
159
sloped laterals) and BMP = 0.5 (for horizontal laterals) match with
those derived by the extended design procedure for complex field
topography. In particular, both BMP values are attained, for partic-
ular values of the curvature parameter of the field where the lateral
lays, ␤ = 1 and ␤ ∼ ∞, respectively (see Fig. 8 of Baiamonte, 2017a).
For very irregular slopes, only the SBS procedure should be used, by
locating the manifold at a position that results in same minimum
pressure at uphill and downhill laterals.
In this paper, for uniformly sloped paired drip laterals, even sim-
pler design relationships will be derived that only require a few
constants (3), and that show the influence of the flow rate and
pipe diameter exponents of the resistance equation on the design
variables. As the procedures previously introduced (Baiamonte
et al., 2015; Baiamonte, 2016a), this proposed procedure is a par-
tial approach since it is valid only for one-lateral units and cannot
be used to irrigation units of branching pipe networks, with many
laterals and one manifold. On the other hand, because of its sim-
plicity it is also usable by the general public, and addresses the
main interest of the farmer, i.e. the uniformity of water distribution,
without requiring the classic SBS hydraulic design of the irrigation
system and the use of a personal computer. It reproduces the results
of the analytical approach introduced by Baiamonte et al. (2015)
well, with relative errors that are a bit greater than 2%, and extends
design relationships to i) any imposed pressure head tolerance and
ii) to the Hazen-Williams resistance equation, which is also widely
accepted for smooth-wall Polyethylene Pipe, especially when the
Reynolds number exceeds 105 .
Relationships will be provided for downward paired laterals
(BMP = 0.24), for the most common case in which the paired later-
als are laid on flat fields (BMP = 0.5), for both corresponding layouts
(downward and flat), in which the manifold is located at the bound-
ary of the irrigation unit (BMP = 0) and finally, for upward laterals,
for which only BMP = 0 can be. Thus, there are five considered
layouts all together that are summarized in Table 1. Evaluations
highlighting the effects of design choices in terms of energy sav-
ing, which could have a certain impact by an economic point of
view are discussed. For the considered five layouts, by neglecting
manufacturing coefficient of variation, comparisons between opti-
mal lateral lengths and analyses of lateral pressure head variability
and emission uniformity coefficient, are carried out. Then, for the
five layouts we present and discuss applications of the proposed
relationships, considering different design parameters.
2. The optimal paired drip lateral in a downward sloping
field (Layout # 1) and the best manifold position (BMP)
In Fig. 1, for each of the considered five layouts, a lateral pressure
head distribution line is qualitatively illustrated. Table 1 reports the
slope’s sign (negative downward), the number of emitters depend-
ing on the best manifold position, BMP, the corresponding design
equations and the average coefficient of variation of pressure head,
CV, as illustrated in the following.
The optimal design of paired drip laterals on uniform slopes
(Baiamonte et al., 2015; Baiamonte, 2016a) is moved up by the
motion equations, expressing pressure head variations of the i-th
emitter, hi , along the uphill side of the lateral, hi (u) , as well as along
the downhill side of the lateral, hi (d) . The motion equations are valid
under the assumption that we can neglect: i) the variations of emit-
ters’ flow rate along the lateral (Christiansen, 1942; Zayani et al.,
2001), as well as ii) the local losses due to emitters’ insertions, so
that the morphology of emitter connections does not produce sig-
nificant reductions of the lateral cross section. Accordingly, these
equations are rewritten here:
(u)  
hi (u)
hmax (−r) (−r)
= −K Hnu − Hn + i S0 (1a)
S S u −i
160 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Table 1
The table illustrates the considered five lateral layouts, characterized by different values of BMP, slope and number of emitters. For each layout, the design relationship
equations, a layout sketch and the coefficient of variation of pressure heads, CV, associated with three values of pressure head tolerance, ␦, are also indicated. Two design
relationships are required for sloping laterals, whereas only one is needed for flat laterals. For sloping laterals (Layout #1, #2, #5), depending on the unknown variable, D or
qn , Eq. (20a) can be conveniently substituted with Eq. (21a) or Eq. (21b), respectively. For any layout a pressure head at the inlet equal to (1 + ␦) hn is required.

# Layout Slope, S0 Symbol of # emitters† BMP Design relationships Lateral’s sketch‡‡ CV

␦ = 0.1 ␦ = 0.07 ␦ = 0.3

1 <0 nu + nd 0.24 Eqs. (19a) and (20a) or Eqs. (19b) and (20b) 6.1% 4.3% 1.8%

2 <0 nd 0 Eqs. (22a) and (20a) or Eqs. (22b) and (20b) 6.1% 4.3% 1.8%

3 0 2 nd (=2 nu ) 0.5 Eq. (26b)‡ 5.9% 4.1% 1.7%

4 0 nd 0 Eq. (24b) or Eq. (27b) 5.9% 4.1% 1.7%

5 >0 nu 0 Eqs. (25a) and (20a) or Eqs. (25b) and (20b) 5.7% 4.0% 1.7%

†
The subscript u or d refers to the flow direction, ‡ Lateral length can also be obtained by doubling lengths obtained for layout #4, ‡‡ Dot indicates the best manifold position,
BMP = 0, 0.24 or 0.5. For layout # 1, BMP = 0.24 is replaced by BMP = 0.25, when applying Hazen-Williams resistance equation (see Fig. 3).

Fig. 1. Sketch of the submain units for the five considered lateral layouts (Table 1), for a manifold located inside the unit (a, c), or at the edge of the unit (b, d, e), for different
slopes (S0 < 0, downward lateral, S0 = 0, S0 > 0). In each case, for one lateral, the pressure head distribution line, in between hmin = (1 − ␦) hn and h max (1 + ␦), with ␦ pressure
head tolerance, is also indicated.

hi
(d) (u)
hmax
  
n
(−r) (−r) (−r)
= −K Hnd − Hn − i S0 (1b) -r, truncated at n corresponding to Hn = 1/kr /; n ∈ N, S0
S S d −i
k=1
[−] is the field slope (negative downhill), i [−] is the generic emitter
in which r denotes the flow rate exponent of the flow resistance installed along the lateral, nu and nd are the number of emitters
(−r)
equation, Hn denotes the generalized harmonic number of order
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174 161

along the uphill and the downhill sides of the lateral, hmax (u) /S is the very well. This solution is based on 4 calibration parameters for
maximum pressure head at the manifold connection normalized each unknown variable, nu , nd , imin , S0 (16 calibration parameters
with respect to the emitter spacing, S [m]: in total).
To test the capability of Eq. (4–7) to estimate (nu , nd , imin , S0 )
qrn
K =k (2) and the associated nopt , necessary for the optimal design of drip
Ds
paired laterals, Baiamonte (2016a) calculated the relative errors RE
in which qn [m3 s−1 ] is the average (nominal) emitter discharge, D (%) as a function of K, for h∗n = 5, 10, 20 and 40. RE were calculated
[m] is the inside pipe diameter, and k, r and s are assumed values by assuming that the corresponding analytical values, obtained by
depending on the adopted formula. According to the well-accepted solving the system of Eqs. (4–7), were true. The Author showed
resistance equations for smooth walled Polyethylene Pipe, Blasius’s that RE was less than ± 2%, demonstrating the applicability of the
and Hazen-Williams’s equations (Keller and Bliesner 1990), follow: proposed procedure.
√ In order to show the effect of h∗n on the approximated proce-
0.316 · 4 2 0.25
k=  r = 1.75, s = 4.75 Blasius (3a) dure, Fig. 2a, for five values of K (K = 10−7 , 10−6 , 10−5 , 10−4 and
g r
10−3 ), and under the assumption of Blasius resistance equation,
10.675 reports the relative errors RE (%) on the estimate of nopt as a func-
k= r = 1.852, s = 4.871 Hazen − Williams (3b)
Cr tion of h∗n by the procedure suggested by Baiamonte (2016a). The
8 results show that relative errors rapidly diminish as h∗n increases,
k= r = 2, s = 5 Darcy − Weisbach (3c) for h∗n > 3 and in any case endure within a very narrow range. In
r g
the same figure, the only slightly higher relative errors RE (%) for
where g [m s−2 ] is the acceleration of gravity,  = 1.008 10−6 m2 s−1 the estimate of nopt (solid lines) for another simpler and much
is water’s kinematic viscosity which, according to Blasius’s equa- more manageable design relationship (presented below) are also
tion, for a standard water temperature at 20 ◦ C, gives k = 0.000779, reported.
whereas according to Hazen-Williams’s equation, C is a friction The optimal paired drip lateral, designed according to
coefficient, which is a function of the pipe material’s characteristics. Baiamonte (2016a) associated with Eqs. (4–7), requires a particular
Baiamonte et al. (2015) analytically solved the design problem of position of the manifold denoted as the “best” manifold position,
paired drip laterals in a uniformly sloping field by fixing r = 1.75 and BMP, defined as the ratio between the number of the emitters in
a pressure head tolerance ␦ = 0.1. The hypothesis i) to neglect the the uphill lateral, nu , and the optimal number of the emitters in the
variation of emitters’ flow rates agrees with the latter assumption of entire lateral, nopt .
fixing pressure head tolerance ␦. In fact along the lateral, it is under By introducing approximated and explicit design relationships
limited pressure head variations that the ratio between the emit- to derive the optimal nu and nd , Baiamonte (2016a) showed that
ters’ flow rate variation (q max − qmin ) and its average, qn , is small the BMP is slightly affected by h∗n and K, in their usual range of
(it should be less than 10%), providing a good approximation of this variability, approaching a value very close to 0.24, for r = 1.75. Under
hypothesis (Wu, 1997; Vallesquino and Luque-Escamilla, 2002). the Blasius resistance equation (Eq. (3a)), the Author emphasized
The system of 4 equations leading to the optimal paired drip lat- that BMP = 0.24 is valid if the number of the uphill and downhill
eral design is rewritten here by generalizing for any ␦, and for any emitters are the optimal values, so that h∗min (u) = h∗min (d) = (1- ␦)
r, i.e. for any exponent of the resistance equation (Eq. (2)), and by h∗n and h∗max (u) = h∗max (d) = (1+ ␦) h∗n .
highlighting the scale role of the emitter spacing in lateral design- In this paper, BMP will denote any other manifold position
ing, by mean of the normalized pressure head in respect to the satisfying the latter condition, such as the well-known BMP = 0.5
emitter spacing S [m], denoted here as h∗ (dimensionless): (Fig. 1c), for a flat lateral, as well as BMP = 0, if the laterals are not
nu S0 = K Hnu
(−r)
− 2ı h∗n (4) paired (Fig. 1b,d and e).
Much simpler design relationships providing errors only a bit
(1−r) S0 larger than those derived by Baiamonte (2016a), and that also
Hn =  (1 − r) − (5)
d −imin rK require an imposed BMP, can be derived by solving by substitu-
  tion the Eqs. system (4–7) and by forcing the relationships between
(−r) (−r)
2 ı h∗n = K Hnd − Hn + imin S0 (6)
d −imin the design variables to follow a linear or power function, described
(−r)
below. After a little algebra, for a fixed “distance” between the min-
nd S0 + K Hnd =0 (7) imum pressure at the downhill lateral and its top end, (nd − imin ),
in which i min is the emitter at the downhill lateral for which the and for a fixed r, the system of Eqs. (4–7) can be reduced to only
minimum value of the pressure head occurs, ␨ (1-r) is the Rie- two dimensionless equations for the two unknowns, nu and nd :

∞
(−r) (−r)
mann Zeta function  (1 − r) = − 1r + (−1)n
n (−r)n
(Agnese et al., Hn + Hnu
d −imin
n! = r (nd − imin − nu ) (8a)
(1−r)
n=0 Hn −  (1 − r)
2014), where ␥n are the Stieltjes constants. It is noteworthy that for d −imin

a fixed K, only the discharge exponent of the resistance equation (−r)

affects the system solution. For r = 1.75 and r = 1.852, ␨ (1- r) is equal
to −0.1336 and −0.1108, respectively.
For a fixed r, ␦, K (i.e. qn and D), h∗n and S, the system of the
four Eqs. (4–7) can be solved obtaining the analytical solution (nu ,
nd , imin , S0 ) for which operating pressure heads along the entire
lateral are in the desired range ((1- ı) h∗n ≤ h∗i ≤ (1 + ␦) h∗n ). Thus,
the knowledge of nu and nd and the related optimal number of the
emitters installed in the entire lateral, nopt = nu + nd , yielding the
optimal length, Lopt = S nopt , requires the solution of the four Eqs.
(4–7). Baiamonte (2016a, 2016b, 2018) simplified the solution to
design paired drip laterals in sloping fields, by using exponential
laws and power functions that approximate the analytical solution
Hnd
(1−r)
Hn =  (1 − r) + (8b)
d −imin r nd
The analytical BMP = nu /(nu + nd ) cannot be expressed by Eqs. (8a
and 8b) in an explicit form, however for any optimal (nd − imin ) sat-
isfying Eqs. (8a and 8b), it can usefully be represented as a function
of the number of emitters in the uphill lateral, nu , with the dis-
charge exponent r of the resistance equation as a parameter, as
illustrated in Fig. 3. The considered r values in Fig. 3 are those of Eqs.
(8a and 8b), r = 1.75 and r = 1.852, and, just for r = 2 that corresponds
to the exponent of the well-known Darcy-Weisbach formula. The
figure shows that for an nu greater than around 10, BMP becomes
constant assuming values 0.2396, 0.2478, 0.2560, for r = 1.75, 1.871
162 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Fig. 2. Relative errors RE (%) on the estimate of nopt as a function of h∗n (dashed lines) for different values of K (K = 10−7 , 10−6 , 10−5 , 10−4 and 10−3 ), obtained by using the
design relationships of Baiamonte (2016a) (Fig. 2a) and by using Eqs. (19) and Eqs. (20) (Fig. 2b). In both cases, the RE were calculated by assuming the analytical values of
nopt obtained by solving the system of Eqs. (4–7) as true.

two decimal places, which is advisable in practice. Thus, by fixing

BMP = 24, 25 and 26% of the entire paired drip lateral, for r = 1.75,
1.871 and 2, the corresponding B0 constant can be evaluated, yield-
ing B0 = 0.316, 0.333 and 0.351, respectively. Therefore, B0 cannot be
considered a calibration parameter of the actual design procedure,
but simply a consequence of the choice to fix BMP.
The optimal number of the emitters of the paired lateral, nopt ,
can be linked to the normalized emitter pressure head, h∗n , and to
the lateral slope S0 , by considering the system of Eq. (4), (7) and
(10):
ı h∗n
nopt = nu + nd = B1 (12)
|S0 |
in which B1 is a numerical constant that only depends on r, and that
will be determined by calibration. Therefore, under the hypothe-
sis, nu > 10, which is widely accepted in practice, the number of
the emitters of the uphill and downhill lateral, nu and nd , can
be obtained by simply considering the imposed BMP, which only
Fig. 3. Analytical BMP = nu /(nu + nd ) is represented as a function of the number of
slightly depends on the selected flow rate exponent of the resis-
emitters in the uphill lateral, nu , with the discharge exponent, r, of the resistance
equation (Eq. (2)) as a parameter. The corresponding constant and round BMP values, tance equation (Fig. 3):
assignable in practice, are also indicated.
ı h∗n
nu = B1 BMP (13a)
|S0 |
and 2, respectively and, confirms the BMP value close to 0.24 that
Baiamonte (2016a), for r = 1.75, obtained by the approximated rela- ı h∗n
nd = B1 (1 − BMP) (13b)
tionships. |S0 |
An application of Eqs. (8a and 8b), for fixed exponent r, showed Once the optimal number of emitters has been derived, the
that an almost linear and simpler relationship occurs between nu , hydraulic and geometric characteristics of the emitters of the drip
nd and (nd − imin ): lateral have to be determined. In other words, for the optimal design
nu = B0,1 (nd − imin ) (9a) of the lateral, the relationship between the K constant, expressed
by Eq. (2), and the optimal number of the emitters of Eq. (13) is
nd = B0,2 (nd − imin ) (9b) necessary. Analogously to the derivation of Eq. (12), an application
in which B0,1 in B0,1 are numerical coefficients, which only depend of Eqs. (4–7) fit a power-function:
on r. By combining Eqs. (9a) and (9b), the relationship between  |S | 1/r
0
nd and nu , under the condition of the optimal (maximum length) nopt = nu + nd = B2 (14)
K
lateral design, we can derived:
in which, like B0 and B1 , B2 only depends on the exponent r and,
B0,1
nu = n (10) together with B1 , will be determined by calibration. Thus, analo-
B0,2 d
gously to Eq. (13), the number of the emitters of the uphill and
which shows that the ratio B0 = B0,1 /B0,2 can be expressed as a downhill lateral can be obtained as a function of K:
function of the BMP according to the following relationship:  |S | 1/r
0
BMP nu = B2 BMP (15a)
B0 = (11) K
1 − BMP  |S | 1/r
0
Interestingly, B0 (and BMP) does not depend on the pressure nd = B2 (1 − BMP) (15b)
K
tolerance, ␦, and only depends on the discharge exponent, r, of the
pipe resistance formula. According to the results illustrated in Fig. 3, Finally, by combing Eqs. (12) and (14), the relationships between
BMP in Eq. (11) can reasonably be fixed as a constant rounded to the main dimensionless parameters, h∗n and K, for the verification
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174 163

Table 2
Maximum length of drip lateral, L (m), laid on flat fields, suggested by manufacturer A), for BMP = 0, for inside pipe diameter, D = 14.2 mm, for 3 bar inlet pressure head, for
different values of emitters’ spacing, S, and flow rates, qn .

Emitter spacing, S (cm)

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

qn (l/h) L (m)

1.25 95 137 176 213 248 281 313 344 373

2 70 101 130 157 182 207 230 253 275
2.9 55 79 102 123 143 162 182 199 216
4.4 41 60 77 94 109 124 138 152 165

Table 3
Maximum length of drip lateral, L (m), laid on flat fields, suggested by manufacturer B), for BMP = 0, for different values of inside pipe diameter, D, nominal emitter pressure
head, hn , emitters’ spacing, S, and flow rates, qn .

D (mm) hn (m) Emitter spacing, S (m)

0.3 0.4 0.5 0.6 0.8 1

16 15 89 70 53 115 91 68 140 110 84 162 128 97 205 162 123 245 194 148
25 113 89 68 147 115 87 177 140 106 206 163 124 262 207 158 312 247 188
35 131 103 78 168 133 101 204 161 122 238 187 143 302 239 181 360 284 216
20 15 156 122 94 196 155 118 233 184 140 267 212 160 336 266 203 388 307 234
25 198 157 119 250 197 150 297 235 179 341 270 205 430 340 260 496 392 300
35 229 180 137 288 228 173 343 271 206 294 311 238 497 393 301 573 454 347
qn (l/h) 1.6 2.3 3.5 1.6 2.3 3.5 1.6 2.3 3.5 1.6 2.3 3.5 1.6 2.3 3.5 1.6 2.3 3.5

problem (K is known, Eq. (16b) as well as for the design problem irrigation system manufacturer catalogues were shown. Tables 2–4
(h∗n are known, Eq. (16a), we can derive: report the value of the maximum lengths for drip laterals laid on
 B r |S |1+r flat fields, suggested by three manufacturers, denoted A), B) and
2
K=
B
 0 r (16a) C), for different flow rates, qn , inside pipe diameters, D, emitter’s
1 ı h∗n spacing, S, and nominal pressure heads, hn , values, according to
the corresponding catalogues, which are available on the manufac-
B2 |S0 |1+1/r turer’s website. For these design parameters, K values are in the
h∗n = (16b)
B1 ı K 1/r range 7.00·10−7 –2.09·10−5 .
It is noteworthy that because of Eq. (16b), the pressure head To extend the calibration to other design parameters, which can
distribution line of the optimal paired drip lateral, expressed by Eq. usually be selected for sprinkler irrigation systems (where higher
(1), only depends on the K value, on the lateral slope S0 , and on B1 flow rates are used), the K range was extended to higher K values
and B2 . (10−3 –10−7 ). For a very high number of K values (>10000) in such
By considering Eq. (2), Eq. (16b) can be rewritten, depending on range, the optimal slope and the B1 and B2 constants were deter-
mined by satisfying the optimization procedure according to the
objective functions expressed by the system:

⎧  
⎪ h∗n = B2 /B1 |S0 |1+1/r / K 1/r ı
⎪
⎨      r   r (−r)
B1 ı h∗n / |S0 | BMP − 2 ı h∗n / |S0 | + B2 / B1 |S0 |/ ı h∗n H =0 (18)
B1 BMP (ı h∗n /|S0 |)
⎪
⎩ B ı h / |S | (1 − BMP ) − B / B r  |S |/ ı h r H (−r)
⎪
=0
1 ∗n 0 2 1 0 ∗n B (1−BMP) (ı h / |S |)1 ∗n 0

For a selected pressure head tolerance, ␦, Eq. (18) satisfies the

the design problem of the lateral, by defining i) the internal pipe
occurrence i) of the maximum admitted pressure head, (1 + ␦)h∗n ,
diameter, D, (if qn is fixed) and ii) the average emitter discharge, qn
at the inlet (e.g. at the connection laterals-manifold) and at the top-
(if D is fixed):
end downward lateral, and ii) of the minimum admitted, (1- ı)h∗n ,
 B r/s qn ı h∗n r/s at the top-end upward lateral and for a certain emitter (i = imin ) in
1/s 1
D=k (17a) the downhill lateral.
B2 |S0 |(1+r)/s The optimization procedure was carried out for Blasius, Hazen-
Williams and Darcy-Weisbach resistance equation (BMP = 0.24,
B2 Ds/r |S0 |(1+r)/r
qn = (17b) 0.25, and 0.26 corresponding to B0 = 0.316, 0.333 and 0.351). In each
B1 k1/r ı h∗n
case, the calibration gave values of the B1 and B2 variables in a very
Eqs. (17a and 17b) appear much simpler in respect to those pre- narrow range, verifying the functional links expressed by Eqs. (12)
viously derived (Baiamonte, 2016a), and better show the influence and (14), with the variation coefficients equal to about 0.07% and
of the exponents r and s on the optimal lateral design, further- 0.16%, respectively.
more, they are monomial and their use requires only 2 calibration The calibration provided the values B1 and B2 , reported in
constants, in respect to the previous 16. Table 5, which can be used for practical applications, for the Bla-
B1 and B2 constants were determined for a set of common values sius, Hazen-Williams and Darcy-Weisbach resistance equation. In
of pipe diameters, flow rates, pressure heads and emitter’s spacing. the following, the optimal number of the emitters for the optimal
To refer to usual values of these parameters, some examples of drip paired drip lateral design, according to the Blasius equation (r = 1.75
164 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Table 5

Maximum length of drip lateral, L (m), laid on flat fields, suggested by manufacturer C), for BMP = 0, for inside pipe diameter, D = 14 mm, for different values of nominal emitter pressure head, hn , emitters’ spacing, S, and flow

156
242
293
327
352
373
Exponents of Blasius, Hazen-Williams and Darcy-Weisbach resistance equation,

4
constant B0 associated with the best manifold position. BMP, and values of the
corresponding calibration constants B1 , B2 and F0 .

294
356
398
427
453
190

3
Resistance equation r s BMP B0 B1 B2 F0

324
393
439
471
209

500
2.4
Blasius 1.750 4.750 0.24 0.316 7.35 2.34 1.85
Hazen-Williams 1.852 4.871 0.25 0.333 7.07 2.34 1.83
Darcy-Weisbach 2 5 0.26 0.351 6.84 2.33 1.81

235
364
441
493
529
562
2
252
391
474

569
530

604
100

1.8
and k = 0.000779) obtained by substituting of B0 = 0.316, B1 = 7.35
and B2 = 2.34 into Eqs. (12) and (14), are reported:
125
194
235
263
282
299
h∗n

4
nopt = 7.35 ı (19a)
|S0 |
152
235
286
319
342
366  |S | 0.571
3 0
nopt = 2.34 (19b)
168

315
352
378
260

401

K
2.4
rates, qn . The manufacturer catalogue reports that drip lateral lengths are calculated by imposing a minimum value of pressure head at the end of the lateral equal to 0.7 bar.

Whereas Eq. (16), which links the main dimensionless parameters,

188
292
354
395
424
450

h∗n and K, for the verification problem as well as for the design
2

problem, becomes:
314
381
425
456
484
202

1.8
75

|S0 |2.75
K = 0.135  1.75 (20a)
163
197

236
251
105

220

ı h∗n
4

|S0 |1.571
127
197

267
287
240

305

h∗ = 0.319 (20b)
3

ı K 0.571
141
218
264
295
317
336
2.4

For the Blasius resistance equation, in the following the design

relationships (Eq. (17)) for the internal pipe diameter, D, and for
158
245
297
331
356
378

the average emitter flow rate, qn , are also reported:

2

 0.368
qn ı h∗n
263
319
356
383
170

406
1.8
60

D = 0.338 (21a)
|S0 |0.579
217
140
170
190
204
91

D2.714 |S0 |1.571

4

qn = 19.03 (21b)
ı h∗n
231
248
263
110
170
207

As the following will show later, each of the calibrated Eqs. (19),
121
188
228
254
273
290

(20) and (21) can be applied to find the design variables required to
2.4

design optimal paired drip laterals or to produce very simple design

charts.
136
211
256
286

326
307

For r = 1.75, relative errors, RE, for nopt estimates associated with
Eqs. (19), (20) and (21) are graphed in Fig. 2b, by assuming as true
126
227
274

351
307
330

1.8
50

the corresponding analytical values obtained by solving the sys-

tem of Eqs. (4–7) as a function of h∗n (solid lines), for different
117
142
158
170
180
75

K values in between the range already considered for calibration

(K = 10−7 , 10−6 , 10−5 , 10−4 and 10−3 ). RE for nopt can be compared
142
172
192

219
206
91

with the corresponding ones obtained in Fig. 2a, by using the sim-
plified equations previously introduced (Baiamonte, 2016a). The
156

212
227
241
101

190

2.4

figure shows that, in respect to the latter, Eqs. (19), (20) and (21)
provide errors that are only slightly larger. RE becomes less than
113
176
213
238
255
271

2.5%, for h∗n > 10, and decreases more and more as the normalized
2

pressure head increases. For a fixed h∗n , RE increases with K, prob-

122
183
229
256
274
291
1.8
40

ably because of the fixed B0 (BMP) values. As a matter of fact, the

increasing of RE with K, shown in Fig. 2b, agrees with the differ-
111
124
133
141
59
91

ences between BMP and the fixed BMP that also increase with K, as
4

previously observed, for r = 1.75 (Baiamonte, 2016a).

111
135

161
171
150
72

However, the simplicity of the newer procedure definitely

Emitter spacing, S (cm)

rewards the associated, slightly greater, RE for the nopt estimate,

122
148
166
178
189
2.4

with respect to those associated with the analytical (Baiamonte

79

et al., 2015) or the better approximated procedure (Baiamonte,

137
167
186

212
200

2016a).
89

This described procedure can be applied to both drip lateral irri-

147
179

215
228
200

gation systems as well as to sprinkler irrigation systems without

1.8
95
30

diminishing its usefulness. Furthermore, the method can be used

hn (bar)

to extend the application of the drip lateral lengths suggested by

qn (l/h)
Table 4

the manufacturers, which generally are only valid for drip laterals
1.7
2.4

3.5
1

laid on flat fields, to drip laterals laid on sloping fields.

 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174
3. Application examples
The choice of Eqs. (19), (20) and (21) to use for practical appli-
cation depends on the verification or design problem that has to be
solved, if it is a new or an already implemented irrigation system,
and on the knowledge of the involved parameters. In the following,
the verification and design problems are described by assuming, as
is usually the case, that the length of the irrigation unit (i.e. the num-
ber of the emitters to be installed along the paired lateral, nopt , and
the emitter spacing, S) are already assigned. Moreover, because a
trickle irrigations design is strongly influenced by the internal pipe
diameter, and there are only a few values of available internal pipe
diameters, the latter will be considered as assigned. Since nopt , S
and D are known, in this context the “verification problem” con-
ventionally denotes the case in which the emitters’ pressure head
has to be determined. Whereas for the “design problem”, the emit-
ter pressure head is known, and the flow rate has to be established.
In the following, we apply the procedure to solve a design problem
in such terms.
In Fig. 4a, by assuming the Blasius equation, for a pressure tol-
erance ␦ = 0.1 and for a fixed diameter D = 11.6 mm, the optimal
lateral length, nopt , expressed by Eq. (19a) is graphed as a func-
tion of |S0 |, with h∗n as a parameter. As expected, the figure shows
that as the lateral slope increases, the optimal emitter number
decreases. Furthermore, for a fixed slope, the higher the normal-
ized pressure head is, the “longer” the lateral lengths are. In the
same figure for the same normalized pressure head values as the
parameter, the secondary axis shows the flow rate vs. the lateral
slope, expressed by Eq. (21b). Interestingly, the latter shows that,
for a fixed h∗n and as expected, an increase in slope determines
on the one hand, as already observed, a decrease in nopt , and on
the other hand an increase in flow rate. The same figure also illus-
trates how a numerical application showing how the combined
use of Eqs. (19a) and (21b) leads to the solution. As example, for
a fixed optimal number of emitters, nopt = 145, and a mean lat-
eral slope |S0 | = 13.5%, corresponds a normalized pressure head
h∗n = 26.7, which can be associated to different possible pairs (hn ,
S) (40, 1.5; 33, 1.24; 20, 0.75 or 10, 0.375). The corresponding opti-
mal length of the lateral, Lopt = S nopt , will be 217.5, 181.3, 108.8
or 54.4 m, depending on the selected pressure head value. Choos-
ing an emitter spacing S = 1.5 m and hn = 40 m, the optimal design
of the paired drip lateral results in the flow rate qn = 6.2 l/h (cor-
responding to K = 9.96·10−5 ), Lopt = 217.5 m, the length of the uphill
lateral Lu = (BMP nopt ) S = (35)*1.5 = 52.5 m (and the downhill lateral
Ld = (1-BMP) Lopt = 165 m).
In Fig. 4b, for the above-mentioned parameters (Layout #1),
by applying Eq. (1), i.e. by considering the exponent of the flow
rate-pressure head relationships (q = k hx ), x = 0, the pressure head
distribution line and the hydraulic grade line were graphed (white
dots and black dots, respectively). By using the iterative forward
step-by-step (SBS) procedure, starting from the manifold connec-
tion to the end of both the downhill and the uphill sides of the
lateral, the pressure head distribution line was also graphed by
setting the exponent of the flow rate-pressure head relationships,
x = 0.5 (dashed line).
According to previous results (see Baiamonte et al., 2015, Fig. 7,
and Baiamonte 2016a, Fig. 11), the figure shows that differences
in pressure head distributions, under the two assumptions x = 0
and x = 0.5 are very moderate. By neglecting the manufacturing
coefficient of variation, Cv , and following SBS procedure, the pres-
sure head coefficient of variation, CV, and the emission uniformity
coefficient, EU (Keller and Karmeli, 1975), associated with the two
assumptions, were also calculated. They resulted CV = 6.25% and
the obvious EU = 100%, for x = 0, and CV = 6.04% and EU = 95.8%, for
x = 0.5.
165
Thus, Fig. 4b allows the proposed design procedure to be verified
once the drip lateral length extreme pressure heads are satisfied
(h∗min (u) = h∗min (d) = (1- ı) h∗n and h∗max (u) = h∗max (d) = (1 + ı) h∗n , with
␦ = 0.1), providing the maximum energy and hydraulic efficiency
in the paired drip lateral design, for both assumptions (x = 0 and
x = 0.5) and thus including not pressure compensating emitter.
In order to extend the proposed procedure to any values of the
involved parameters of the optimal paired drip lateral, Fig. 5a charts
the design problem, and Fig. 5b charts the verification problem,
graphing Eqs. (19a) and (20a) with h∗n as a parameter, and Eqs.
(19b) and (20b) with K as a parameter, respectively. In both figures,
the previous numerical example is also represented. For the design
problem, the chart application gives K = 1·10−4 (close to the solu-
tion that was obtained analytically), whereas the verification chart
application provides h∗n = 26.7. Of course, for these K and h∗n , any
congruent optimal pairs for qn and D and, as mentioned for hn and
S, can be selected to provide an optimal paired lateral in a sloping
field.
Furthermore, both graphs can be used by reversing the sec-
ondary vertical axis to the main axis. For example, this would be
useful if the maximum length of the lateral is not fixed but is instead
to be determined.
In the next section, by introducing a third last calibration con-
stant, the optimal lateral lengths are derived for the case in which
the manifold is located at the top-end of the drip laterals (BMP = 0)
in a flat field (S0 = 0), for which manufacturer catalogues already
provide values, and for sloping on uphill and downhill fields, which
are particular cases of the previous paired lateral systems.
4. Manifold at the top-beginning of the laterals (BMP = 0,
layout # 2, # 4 and # 5)
The procedure previously described allows the optimal design
of a paired drip lateral in sloping fields to be elaborated by locating
the manifold, according to the Blasius equation, at 24% of the entire
optimal length of the lateral (BMP = 0.24, Fig. 1a). For a fixed inlet
pressure head at the manifold, e.g. 1.1 hn , it is evident that the solu-
tion BMP = 0.24 provides lateral lengths greater than 24% in respect
to those obtained by locating the manifold at the top-end of the lat-
eral, i.e. by considering only the downhill lateral (nu = 0). Also, for
flat terrain (S0 = 0), for which the lengths L suggested by the manu-
facturer catalogue can be considered, the manifold connection can
be conveniently located in the center of the laterals (BMP = 0.5) so
that, for a fixed inlet pressure heads at the manifold, ␦ hn , provides
doubled lengths 2 L.
Notwithstanding the fact that the following solutions (BMP = 0),
for flat fields as well as for sloping uphill and downhill fields (layout
# 2, # 4 and # 5 of Table 1), are not convenient from an irrigation
efficiency point of view, they are described in the following because
sometimes these solutions are preferred by the farmers. They facil-
itate the underground manifold pipe’s repair by making it easier to
excavate since it is not located in between crop rows (Fig. 1b, d and
e). Furthermore, depending on the chosen manifold material, the
BMP = 0.24 or BMP = 0.5, requires a double number of connections
of the paired drip laterals, which can have an important role on the
amount of costs that the farmers have to sustain.
4.1. Lateral laid on downward sloping fields (BMP = 0, nu = 0,
S0 < 0, Layout #2)
Drip lateral design equations for downward sloping fields, when
the manifold is located at the boundary of the irrigation sector
(BMP = 0, Fig. 1b), is a particular case of the previously illustrated
166 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Fig. 4. For Layout # 1, optimal number of emitters, nopt , as a function of the mean slope |S0 |, with h∗n as a parameter (Eq. (19)a) and, for the same parameter, h∗n , flow rate,
qn , as function the mean slope |S0 | (Eq. (21)b), are reported. Black dots (Fig. 4a) indicate the design variables, |S0 |, nopt , h∗n , and qn , corresponding to the application illustrated
in Fig. 4b, where the pressure head distributions, for x = 0 (Eqs. 1) and for x = 0.5 (SBS), both satisfying the condition (1- ı) hn ≤ hi ≤ (1- ı) hn , with ␦ = 0.1, are illustrated.

Fig. 5. The optimal number of emitters, nopt , vs. the mean slope of the optimal lateral length, |S0 |, expressed by Eq. (19a), with h∗n , as a parameter. For the same h∗n parameter,
in the secondary axis, K vs. |S0 | is also reported (Eq. (20a) (Fig. 5a). The optimal number of emitters, nopt , vs. the mean slope of the optimal lateral length, |S0 |, expressed by
Eq. (19b), with K as a parameter. For the same K parameter, in the secondary axis, h∗n vs. |S0 | is also reported (Eq. (20b)) (Fig. 5b). For both figures, black dots, for nopt = 145,
indicate the mean slope, |S0 | = 18.5%, and the associated, h∗n = 26.7 and K = 9.83 10−5 , corresponding to the application illustrated in Fig. 4b.

equations. For the case nu = 0, S0 < 0 only Eqs. (13b) and (15b) need obtained for the case (BMP = 0, Fig. 1d), the latter needs to be for-
considering, giving: mulated. As it is well known, in the simplest case in which the
lateral laid on a level field (S0 = 0), the minimum pressure head is
ı h∗n
nopt = 5.58 (22a) located at the downstream end of the lateral (i min ∼ = nd = nopt ), for
|S0 |
any K value. Thus, in this case the admitted minimum of the pres-
 |S | 0.571 sure head ((1- ␦) hn ) needs to be imposed at the downhill end of
0
nopt = 1.778 (22b) (−r) ∼
K the laterals (i = nd ) into Eq. (6), which considering Hn −i = 0 and
d min

As for layout #1, for Eq. (22a) or Eq. (22b) to be applied, it needs S0 = 0, yields:
to be coupled to Eqs. (20) or Eq. (21). Fig. 6a, for r = 1.75, for ␦ = 0.1, 2 ı h∗n (−r)
for qn = 5 l/h and D = 32 mm (K = 5.52 10−7 ), and for hn = 22.5 m and = Hnd (23)
K
S = 1.5 m (h∗n = 15), illustrates the pressure head distribution for an
To facilitate the designing of laterals laid on flat fields, a power-
optimal lateral designed according to Eqs. (22a) and (21b) giving
law can be used in place of Eq. (23), obtaining a relationship
nd = 590 and S0 = − 1.42%.
between the main dimensionless parameters for the verification
By comparing Eqs. (19a) and (22a) and by considering Eq. (21a),
problem (h∗n vs. nopt , if K is known), and for the design problem
it is interesting to observe for sloping laterals how much less conve-
(h∗n is known and K or nopt have to be determined):
nient the choice BMP = 0 is in respect to BMP = 0.24. In particular, for
a fixed lateral length, the emitters’ flow rate and slope, and for any n1+r
opt K
pressure head tolerance, the corresponding optimal design requires h∗n = (24a)
the emitters’ pressure head and pipe diameter to be greater than F01+r ı
32% and 11%, respectively (hn #2/hn #1, D #2/D #1 in Table 6). ı h∗n
K = F01+r (24b)
n1+r
opt
4.2. Lateral laid on flat fields (BMP = 0, S0 = 0, Layout #4)
 1/(1+r)
ı h∗n
To detect the effect of the field slope on the optimal (maximum) nopt = F0 (24c)
K
lateral lengths for the case BMP = 0.24 or 0.5 (Fig. 1a) with those
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174 167

Fig. 6. For qn , = 5 l/h, D = 32 mm (K = 5.52 10−7 ), hn = 22.5 m, S = 1.5 m (h∗n = 15), Fig. 6a shows the optimal pressure head distribution ((1 − ı) hn < hi < (1 − ı) hn , with pressure
tolerance ␦ = 0.1) obtained by designing the drip lateral according to different slopes, S0 , and lengths, S nd , as indicated in the box. For a set of design variables (S0 , qn , D, S,
h∗n ), Fig. 6b illustrates the relative error RE (%) on the estimate of nopt by Eq. (24c), associated to the optimal design of a drip lateral laid on a level field (S0 = 0), as a function
of the optimal number of the emitters (dashed line). Fig. 6b also illustrates relative errors associated to the pair drip lateral on uniform slope and BMP = 0.24 (Eqs. 19a and
20a, solid line).

Table 6
Matrix of the 25 combinations of the hn ratios for the five considered layouts. Values in round brackets indicate the corresponding inside diameter ratio. For example, the
bold ratio m21 = hn #2/hn #1 = 1.32, under the main diagonal of the matrix, indicates that for the same design variables (nopt , qn , D, ␦, S) layout #2 requires a nominal pressure
head and a lateral diameter 32% and 11% greater than those corresponding to layout #1, respectively.

Layouts #1 #2 #3 #4 #5

#1 1 0.76 (0.90) 1.12† 0.17 0.24 (0.59)

#2 1.32 (1.11) 1 2.40 0.36 0.32 (0.65)
#3 0.89 0.42 1 0.15 0.02
#4 5.98 2.80 6.73 1 0.12
#5 4.17 (1.69) 3.17 (1.53) 57.01 8.47 1
†
If the value in round brackets is not indicated, the inside diameter ratio is the unity, i.e. the comparing laterals have the same inside diameters.

where F0 is the third calibration constant for designing the five con-
sidered layouts (Table 1), in addition to B1 and B2 , valid for sloped
laterals (Eq. (18)). For horizontal laterals, F0 was evaluated for the
Blasius, Hazen-Williams and Darcy-Weisbach resistance equation,
giving 1.85, 1.83, and 1.81, respectively.
For a set of design variables (S0 , qn , D, S, h∗n ), Fig. 6b illustrates
the relative error RE (%) for the estimate of the optimal number of
the emitters, nopt , by Eq. (24c), associated to the optimal design of
a drip lateral laid on a level field (S0 = 0) and to the Blasius resis-
tance equation, as a function of nopt (dashed line), by assuming
the analytical values obtained by Eq. (23), as the true. RE markedly
diminishes with increasing nopt , and becomes negligible, less than
1%, for nopt > 30, suggesting that Eq. (24) is suitable for optimally
designing drip laterals on flat fields. The standard error estimate SEE
associated to Eq. (24c) is very small, equal to 2.89. As Fig. 6b shows,
the relative errors RE (%) associated to Eq. (24c) are lower than
those associated to Eq. (19), valid for sloped laterals and BMP = 0.24.
Very similar errors were obtained by assuming the Hazen-William
resistance equation.
Fig. 6a illustrates an example application for the same design
variables considered for layout #2, the pressure head distri-
bution obtained by designing the drip lateral with Eq. (24)
that, for nopt = 590, results in between the admitted range
(0.9 hn < hi < 1.1 hn). Of course, for flat fields, the results can be easily
extended to the case in which the manifold is located in the middle
of the lateral (BMP = 0.5) by simply doubling the obtained optimal
lengths.
By considering Eq. (24c), it is interesting to observe that in
respect to BMP = 0.5, the choice BMP = 0 is much less convenient
because, for fixed lateral length, the emitters’ flow rate and pipe
diameter, and for any pressure head tolerance, the corresponding
optimal design requires the emitters’ pressure head to be 6.7 times
greater (hn #3/hn #4 in Table 6).
4.3. Lateral laid on upward sloping fields (BMP = 0, nd = 0, S0 > 0,
Layout #5)
Here, we consider the design of drip lateral relationships for
upward sloping fields, when the manifold is located at the boundary
of the irrigation sector (BMP = 0, Fig. 1e). Even though we know
that this condition is not to be preferred, we will consider it in
the following because this position may be necessary if the overall
structure of the irrigation system necessitates this design. Thus, for
this case, nd = 0, S0 > 0 only Eqs. (13a) and (15a) yielding:
ı h∗n
nopt = 1.763 (25a)
|S0 |
 |S | 0.571
0
nopt = 0.562 (25b)
K
Analogously to the previous cases, as an application example, by
using the same design variables of layout #2 and #4 again, Fig. 6a
illustrates the pressure head distribution obtained by designing
the drip lateral with Eq. (25a), which with the expected smallest
value (nopt = 186), is the best choice since it is between the admit-
ted range (0.9 hn < hi < 1.1 hn). The typical almost linear behavior of
168 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174
Fig. 7. Comparison between the number of the emitters obtained by the catalogue,
as the ratio length L – spacing, S, provided by 3 manufacturers (Table 2–4), for drip
laterals laid on flat fields (S0 = 0), and the optimal number of the emitters derived
by the proposed procedure (Eq. (24c) or Eq. (27b). The line of perfect agreements is
also indicated.
the pressure heads along the upward lateral is also qualitatively
indicated in layout #5 of Fig. 1e, which strongly differs from that of
the corresponding downward lateral (layout #2, Fig. 1b).
A comparison in respect to layout #2 can be carried out for lay-
out #5. The choice of an upward lateral, for a fixed lateral length,
emitters’ flow rate and pipe diameter, and for any pressure head
tolerance, as expected, is much less convenient because the cor-
responding optimal design requires the emitters’ pressure head
and pipe diameter to be 3.2 times and 1.5 greater, respectively,
in respect to the downward lateral (see Table 6, 5.58/1.763 and
(5.58/1.763)r /4.75 , with r = 1.75).
5. Energetic evaluations for the considered five layouts
Considering the increasing demand and price for energy are
affecting all economic sectors including irrigated agriculture
(Fernández García et al., 2015), an inspection on the energetic per-
formance of each considering layouts was in the following carried
out.
To show and compare the energetic advantages of each of the
considered layouts, Eqs. (19), (20) and (21) for layout #1 and Eqs.
(22), (24) and (25) for those corresponding to the other layouts,
were reformulated to derive the hn ratios for the 20 combina-
tions of the five considered layouts. 3 main hn ratio relationships
were derived. One for comparing both sloped lateral layouts (e.g. hn
#2/hn #1 = 7.35/5.58), one for a mixed comparison of sloped and
1+r
horizontal laterals (e.g. hn #4/hn #1 = 19.03r (7.35/25) ) and the
last one for comparing both horizontal laterals (e.g. hn #4/hn
1+r
#3 = (50/25) ). The remaining 7 hn ratios can be determined
by the three mentioned relationships, by substituting the design
relationship coefficients of the corresponding layout, with the
exception of the 19.03 coefficient (Eq. (21)b), which is constant for
any sloped layout (#1, #2 or #5).
Table 6 reports the hn ratios for the 20 possible combinations.
In the table, the hn ratios corresponding to the mentioned compar-
isons already carried out in the previous section, for BMP = 0 can be
observed (hn #2/hn #1 = 1.32, hn #4/hn #3 = 6.7, hn #5/hn #2 = 3.17).
Values in round brackets indicate the corresponding ratio referring
to the inside diameter. Thus, from an energetic point of view, the
hn ratios lets us predict the economic advantages or disadvantages
of each layout. For example, the value 1.32 (1.11), under the main
diagonal of the matrix, as already observed, indicates that for the
same design variables (nopt , qn , ␦, S) layout #2 requires a nominal
pressure head and lateral diameter 32% and 11% greater than that
corresponding to layout #1. The hn ratios over the main diagonal
of the matrix are, of course, the reciprocals of those reported under
it. The hn #3/hn #1 and hn #3/hn #2 ratios are noteworthy, 0.89
and 0.42 respectively, because they back up what it is well-known
in irrigation design, i.e. BMP = 0.5 in flat fields is the best efficiency
choice for the manifold.
It is interesting to observe that by excluding the cases hn #2/hn
#1, hn #5/hn #1 and hn #5/hn #2 #2, for which the pipe diameters
for the comparing layouts are not the same (values in brackets in
Table 6), the remaining pressure head ratios are valid for any lateral
length, emitter spacing, flow rate, pressure head tolerance, pipe
material and diameter too. Thus, it is expected that the same h
ratios can be extended to the corresponding energy ratios (since
the latter two are linearly related), with similar implications by an
economic point of view.
Finally, the value hn #4/hn #3 = 6.7 is close to 8 that Baiamonte
(2017b) found for the horizontal layouts, for BMP = 0 and for
BMP = 0.5, by using close-form solutions of the required pressure
heads derived under the Darcy-Weisbach resistance formula.
6. Comparing optimal lateral lengths for sloping fields and
for flat fields
Under the assumption of the Blasius resistance equation, in
order to compare optimal lateral lengths for flat fields and for slop-
ing fields, as well as the two different manifold positions, i.e. inside
the irrigation sector, layout #3 (BMP = 0.5, for flat fields, Fig. 1c) and
layout #1 (BMP = 0.24 for sloping fields, Fig. 1a) and at the boundary
of the irrigation sector, layout #2, #4 and #5 (BMP = 0, for both slop-
ing fields, Fig. 1b,e and flat fields, Fig. 1d), the associated optimal
lateral length relationships are derived below.
For sloping fields and BMP = 0.24, by combining Eqs. (19a) and
(21), the following relationship of the optimal length, Lopt |S0 |<0 , was
obtained:
1/(1+r)
S <0 h∗n ı Ds/(1+r)
0
Lopt = 207.4 S for BMP = 0.24 (Layout # 1) (26a)
r/(1+r)
qn
Eq. (26a) could misrepresent the influence of the slopes on the
optimal lateral length, which explicitly does not figure into it. In
fact, it must be noted that Eq. (26a) is only valid if the lateral slopes
satisfy Eqs. (21), which in turn must also satisfy the optimal number
of the emitters expressed by Eq. (19b) and (26a).
Optimal lengths for a paired drip lateral laid on a flat field can be
expressed by combining Eq. (24c) and Eq. (2) with the k constant, for
r = 1.75, equal to 0.000779 and by considering BMP = 0.5, yielding:
1/(1+r)
S =0 ı1/(1+r) h∗n Ds/(1+r)
0
Lopt = 50 S for BMP = 0.5 (Layout # 3)
r/(1+r)
qn
(26b)
By assuming as usual ␦ = 0.1, a comparison between Eq. (26a)
and Eq. (26b) states that for optimal lateral laid in flat fields, for
which BMP = 0.5, and for optimal lateral for which BMP = 0.24, the
associated optimal lateral lengths are very close, nevertheless dif-
ferences in term of hn were higher.
Analogously to Eqs. (26), for the manifold located at the bound-
ary of the irrigation sector (BMP = 0, Fig. 1b), similar equations can
be derived. For sloping uphill and downhill fields, to determine
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174 169

the optimal lengths of drip laterals, Eqs. (22a) and (21) have to be However, this discrepancy must be due to the fact that, as reported
combined, yielding: in the catalogue, the lateral length suggested by the manufacturer
1/(1+r)
does not agree with the optimal design criteria assumed in all of the
S <0 h∗n ı Ds/(1+r) proposed design relationships, which should be preferred, to fix at
0
Lopt = 157.6 S for BMP = 0 (Layout # 2) (27a)
r/(1+r)
qn the last emitter (the most disadvantaged in flat fields) to assure a
pressure head equal to ␦ hn . As a matter of fact, (see Table 4, foot-
For laterals laid on flat fields with BMP = 0, the optimal drip lat-
note) a pressure head equal to 0.7 bar is assumed, independently of
eral length can be simply obtained by halving the optimal length
the nominal pressure head of the emitter. In other words, the last
expressed by Eq. (26b):
emitter for the suggested length operates as an example for a nomi-
1/(1+r) nal pressure head pn = 4 bar; according to Eq. (27b), it could operate
S =0 ı1/(1+r) h∗n Ds/(1+r)
0
Lopt = 25 S for BMP = 0 (Layout #4) (27b) with a pressure near to the nominal one pe = 0.9 hn = 0.9 * 4 = 3.6 bar,
r/(1+r)
qn
whereas according to the catalogue, it should operate with a pres-
Whereas for upward sloping laterals, by combining Eqs. (25a) sure head much lower than the nominal one (pe = 0.7 bar). Thus,
S0 >0 , can be obtained:
and (21), the corresponding optimal length, Lopt drip lateral lengths suggested in the catalogue are greater than
those corresponding to the suggested design relationships to the
1/(1+r)
S >0 h∗n ı Ds/(1+r) detriment of flow rate uniformity.
0
Lopt = 4.97 S for BMP = 0 (Layout # 5) (27c)
r/(1+r)
qn
8. Analysis of lateral pressure head variability (CV) and
It is apparent that the coefficients appearing into Eqs. (26) and emission uniformity coefficient (EU) for the considered five
(27) only depend on the BMP value (B0 ), on the calibration coeffi- layouts
cients B1 , B2 , F0 , and on k.
A comparison between Eqs. (27a, and 27b) states that, for Interesting analysis and discussion about the coefficient of vari-
BMP = 0, the optimal design on a downward sloping fields provides ation of water distribution and emission uniformity, were carried
lateral lengths that are significantly higher than those for flat fields, out by Royuela-Tomás et al. (2005) and Juana et al. (2004), who pro-
S0 <0 = 1.46 LS0 =0 . Thus, manufacturer catalogues can significantly
Lopt opt vided analytical solutions. In this section, under the assumption of
underestimate the lateral length if the slope is ignored. On the con- the mentioned hypothesis, and for the selected five layouts, rela-
trary, for uphill fields, a comparison between Eqs. (27a, and 27c) tionships to determine the pressure head coefficient of variation
yields LoptS0 >0 = 0.05 LS0 =0 , indicating the well-known great incor-
opt (CV) and emission uniformity coefficient according to Keller and
rectness of locating the manifold at the boundary of the irrigation Karmeli (1975), are derived.
unit, when the fields are upwards. As observed, because of Eq. (20b), for sloping laterals, and Eq.
It is noteworthy, that the compared coefficients in Eqs. (26), 0.96, (24a), for horizontal laterals, the pressure head distribution of the
and Eqs. (27), 1.46 and 0.05, are influenced by the pressure head optimal lateral only depends on the K parameter and on the slope
tolerance ␦· As an example, they increase to 1.24 and, 1.88 and value, S0 .
0.06, as the pressure head tolerance, ␦, increases from 0.1 to 0.15. By neglecting manufacturing coefficient of variation, for the K
values ranging between 10−3 –10−7 , pressure head uniformity has
7. Validating optimal lateral lengths for flat fields been investigated by considering the drip laterals of the five lay-
outs reported in Table 1. By considering Eq. (1), the coefficient of
variation, CV, was calculated for each layout and for each K value,
according to the derived relationship:
n 
 u    2  nd    2
     
 (nu + nd ) 1 + ı h∗n − K
(−r)
Hnu − Hn −i
(−r)
+ i S0 + 1 + ı h∗n − K
(−r) (−r)
Hnd − Hn −i − i S0

 u d

CV = 
i=1 i=1
n  − 1 (28a)
 u      nd     2
     
 1 + ı h∗n − K
(−r)
Hnu − Hn −i
(−r)
+ i S0 + 1 + ı h∗n − K
(−r)
Hnd − Hn −i
(−r)
− i S0
u d
i=1 i=1

in which h∗n is expressed according to Eq. (20b), for sloped later-

Design relationships obtained for flat fields (Eqs. (24c) or (27b) als, or to Eq. (24a) for horizontal laterals. For layout #3, for which
even if already tested by the application illustrated in Fig. 6a (solid BMP = 0.5 (nu = nd ) and S0 = 0, and for layouts #2, #4 and #5 for
line), can be used as a verification via a comparison with the lat- which BMP = 0 (nu = 0 or nd = 0), Eq. (28a) reduces to:
eral lengths, L, suggested by the manufacturers (Table 2–4), since
 n    2
the same hypotheses are valid (S0 = 0 and BMP = 0). Fig. 7 reports    
n 1 + ı h∗n − K
(−r) (−r)
Hn − Hn−i ∓ i S0
the number of emitters suggested by the catalogues in abscissa 
(computed as L/S), for three different manufacturers A), B) and C)  i=1
CV =   2 − 1
and the corresponding number of the optimal emitters in ordinate    
 n  
calculated by Eq. (27b), by assuming ␦ = 0.1 and nopt = LoptS0 =0 /S.
 1 + ı h∗n − K
(−r) (−r)
Hn − Hn−i ∓ i S0
It can be observed that, for the manufacturers A) and B), only a
i=1
few pairs are far from the line of perfect agreement. This is probably
(28b)
due to the fact that, as the catalogue says, “the suggested lengths are
subject to change due to changes in the conditions in which the tests where n = nu or nd depending on the layout and whether the sign
were performed”, or because the considered resistance equation in of S0 is “-” or “ + ”, depending on whether the field is downhill or
the catalogue is different from the well-accepted Blasius equation uphill, respectively.
considered here. Thus, for manufacturers A) and B), the pairs are For layout #1 (downward sloping laterals), for a fixed K value
almost next to the line of perfect agreement, whereas for the man- (K = 10−3 , 10−4 and 10−5 ) and for different pressure head tolerances
ufacturer C) Eq. (27b) underestimates the number of the emitters. (␦ = 0.1, 0.07, 0.05 and 0.03), Fig. 8 illustrates CV as a function of |S0 |.
170 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Fig. 8. Coefficient of variation of the lateral pressure head, CV, evaluated for layout #1 by Eq. (28), as a function of the lateral slope, |S0 |, for 3 values of K parameter (K = 10−3 ,
10−4 , 10−6 ) and for 4 values of pressure head tolerance, ␦ (␦ = 0.1, 0.07, 0.05 and 0.03).
For the highest K value (Fig. 8a), CV seems to vary with |S0 |. Vari-
ability over a constant CV value tends to vanish with a decreasing K,
until it completely disappears for the lowest K value (long laterals,
Fig. 8c).
This CV behavior can be justified for the low K values for which,
for any ␦, the number of the emitters, especially in the uphill side,
can be low (e.g. <10–15), as can be observed by Eq. (14) and Eqs.
(15). In fact, for a low number of emitters, the calibrated parameters
may not result in the working range (see Fig. 3), pushing the lat-
eral farther from the optimal operating pressure head range. Some
dispersion over the constant CV value can also be justified by the
necessary rounding to the integer in applying Eq. (28).
Table 1 reports the mean value, CV, for 3 values of the pressure
head tolerance (␦ = 0.1, 0.07 and 0.03) which, as expected, increases
with ␦ and varies in a very narrow range. As illustrated in Fig. 9a,
CV and ␦ are linearly correlated. The angular coefficients, mCV , also
reported in the figure, can be used to set the desired CV value, in
designing optimal laterals, once the corresponding pressure head
tolerance, ␦, is calculated by the equation indicated in the figure.
Analogously to CV, also the familiar design emission uniformity
coefficient, EU, of Keller and Karmeli (1975), commonly used for the
evaluation of the performance of drip irrigation systems, associated
with the proposed procedure, was derived (Baiamonte, 2016b):
⎛ ⎞x
that, as expected, with increasing of ␦, CV increases whereas EU
decreases. Moreover, (Layout #1 and #2, downward sloped fields)
and (Layout #3 and #4, horizontal fields) have the same behavior in
terms of CV and EU. For any layout, it is also interesting to observe
that the design procedure allows a lateral to be designed with the
maximum length, optimizing water use efficiency (WUE), assur-
ing low values of pressure head variability (CV < 6.1%, for ␦ = 0.1)
and high values of emission uniformity (EU > 95%, for ␦ = 0.1). Inter-
estingly, CV and EU values, corresponding to Layout #1, inversely
derived by the relationships indicated in Fig. 9ab, for ␦ = 0.1, match
with those calculated for the example of the application illustrated
in Fig. 4b, where the exact SBS procedure was applied: CV = 6.2%
and EU = 96.0%. The latter showed that the considered assumption
x = 0, also in computing the average (for CV, Eq. (28), and for EU,
Eq. (29)) and the standard deviation (for CV, Eq. (28)) of emitters’
pressure heads provide satisfactory results.
Further investigations are carried out below, which aim to detect
variations of CV as a function of the slope, for laterals designed
according to the proposed procedure, under the assumption that
the lateral is laid on a flat field, for BMP = 0.5 and for BMP = 0. In
respect to Fig. 8, the role of the slope is different; in Fig. 8, the slope
corresponds to the optimal design, whereas it varies in the actual
context.

 ⎜   ⎟
⎜ (nu + nd ) 1 − ı h∗n ⎟
EU = 100 1 − 1.27 √
Cv ⎜ ⎟ (29a)
np ⎜nu
     
nd
    ⎟
⎝ 1+ı h −K
(−r)
Hnu − Hn
(−r)
+ i S0 + 1 + ı h∗n − K
(−r)
Hn
(−r)
− Hn − i S0
⎠
∗n
u −i d d −i

i=1 i=1

where x is the exponent of the flow rate-pressure head rela-

tionships, np are the number of emitters per plant, Cv is the
manufacturing coefficient of variation, q min and qav are the mini-
mum and the average emitter flow rates in the lateral. For layouts
#2, #3, #4 and #5 Eq. (29a) reduces to:
⎛ ⎞x
 ⎜   ⎟
Cv ⎜ n 1 − ı h∗n ⎟
EU = 100 1 − 1.27 √ ⎜ ⎟ (29b)
np ⎜n 
    ⎟
⎝ (−r) (−r) ⎠
1 + ı h∗n − K Hn − Hn−i ∓ i S0
i=1

where n refers to the total number of the emitters of the lateral,

and the sign of S0 is ‘-’ or “ + ”, depending on whether the field is
downhill or uphill, respectively.
By neglecting the manufacturing coefficient of variation, Cv , Eq.
(29a) and Eq. (29b) are illustrated in Fig. 9b, which shows a linear
relationship between EU and ␦, also indicated in the figure, together
with the corresponding angular coefficient, mEU . As for CV (Fig. 9b),
the latter can be used to determine the laterals pressure head toler-
ance of the, ␦, corresponding to a desired EU value. The figure shows
For r = 1.75, ␦ = 0.1 and for two values of K (7·10−7 and 8·10−6 )
and h∗n (10 and 120), Fig. 10 illustrates the coefficient of variation,
CV, as a function of the field slope, for BMP = 0.5 (Fig. 10a) and for
BMP = 0 (Fig. 10b). In both cases, the number of the emitters is the
optimal, 2 nd and nd , calculated by Eq. (24c), for S0 = 0, for both BMP
values (layout #3 and #4), as reported in Table 1. Fig. 10 shows that
for S0 ∼ 0, CV ∼
= 5.9%. Moreover, depending on the BMP value, it is
interesting to observe a different effect of the lateral slope on CV.
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174 171

Fig. 9. Pressure head tolerance, ␦, as a function of the average coefficient of variation of pressure head CV (Fig. 9a), and as a function of the emission uniformity coefficient of
Keller and Karmeli, EU, (Fig. 9b), for the five considered layouts reported in Table 1. The angular coefficient, mCV and mEU , obtained by the corresponding linear regression,
are also indicated.

Fig. 10. For two different values of h∗n (10 and 120) and K (7·10−7 and 8·10−6 ), corresponding to different set of the design variables, qn , D, hn , S, the coefficient of variation
CV is graphed as function of the field slope, |S0 |, for BMP = 0.5 (Fig. 10a) and for BMP = 0 (Fig. 10b). In both cases, the optimal number of emitters is used, by assuming that the
paired lateral (BMP = 0.5, Eq. (26b) and not paired (BMP = 0, Eq. (27b), lays on a flat field. In Fig. 10b, white circles indicate CVmin , for S0CVmin values, calculated by Eq. (30b).

For BMP = 0.5, for any h∗n and K values, because of the mini- ing from 0.9 hn (also imposed for S0 = 0) because of the imposed
mum admitted pressure head at the two extreme-ends of the lateral maximum pressure head at the manifold.
(0.9 hn ), CV monotonically increases with the lateral slope. In fact, Towards the aim of formalizing the S0 values corresponding to
any |S0 | > 0 values shows a decrease at the top-end pressure head CVmin , S CV
0
min , i.e. the maximum values of |S | in order to increase
0
of the lateral (hi (u) < 0.9 hn ) and an increasing pressure head at the water distribution uniformity, the derivative of Eq. (28b) in respect
to S0 , was carried out:


nd

nd

2 A(i) i A(i)
∂CV i=1 i=1 h∗n. max S0 (−r) (−r)
= nd (1 + nd ) − with A(i) = − i − (Hnd − Hn −i ) (30a)
∂S0 
nd
2
K K d

A(i)
i=1

Once Eq. (30a) was imposed equal to zero, by using a power-law

(R2 = 1, standard error of the estimate, SEE = 22.1), a good fit of Eq.
(30a) was obtained
bottom-end of the lateral (hi (d) > 1.1 hn ). CVmin
|S0 |
 h 0.6364
∗n
Contrarily, for BMP = 0, in a certain low range of |S0 |, CV decreases = 0.2308 (30b)
K K
with increasing |S0 |, showing a minimum CV value, CVmin , equal
to 2%, over which CV, as for BMP = 0.5, increases with |S0 |. This Eq. (30b), much simpler than Eq. (30a), for any h∗n /K, easily
CV behavior appears reasonable, if one considers that as the provides S0CVmin values.
slope increases (from S0 = 0), CV reductions are determined by the In Fig. 11, for ␦ = 0.1 and for the same values of K (7·10−7 and
increasing of the pressure head at the top-end of the lateral start- 8·10−6 ) and h∗n (10 and 120) considered in Fig. 10, the normal-
172 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174
Fig. 11. For two different values of h∗n (10 and 120) and K (7·10−7 and 8·10−6 ), and for an imposed pressure head tolerance, ␦ = 0.1, the normalized pressure head, h∗n , at
the uphill-end of the lateral, i = nu , (for BMP = 0.5), at the downhill-end of the lateral, i = nd , and at the emitter where |S0max | the minimum of pressure head occurs, i = imin , is
graphed as a function of |S0 |, for BMP = 0.5 (Fig. 11a) and for BMP = 0 (Fig. 11b). In both cases, the number of the emitters is the optimal calculated by Eqs. (26b) and (27b)),
by assuming that the field is flat. The value of the minimum (dot lines) and maximum (dash-dot lines) admitted normalized pressure head, and values (Eq. (31b), over which
pressure heads are not in the admitted range, are also indicated.
ized pressure head, h∗n , i) at the uphill-end of the lateral, i = nu , (for
BMP = 0.5), ii) at the downhill-end of the lateral, i = nd , and iii) for
the emitter where the minimum of pressure head occurs, i = i min
(for both BMP = 0.5 and 0), are graphed as function of the field slope.
As for the CV analysis, in order to show the effects of the slope on
operating pressure heads for paired drip laterals, we can consider
the optimal number of emitters obtained by Eq. (26b) and Eq. (27b),
in order to refer the effect of the slope on a lateral designed for flat
fields. The value of the minimum (dot lines) and maximum (dash-
dot lines) admitted normalized pressure heads, which let us detect
the associated admitted slope values, are also indicated.
The figure shows that for BMP = 0.5, h∗n (imin ) is in the admitted
range of normalized pressure head in the downhill lateral, and h∗n
(nd ) is in the admitted range until a maximum slope value |S0max |,
whereas, as expected, any lateral slopes greater than zero provide a
normalized pressure head h∗n (nu ) lower than the admitted 0.9 h∗n .
Thus, for paired drip laterals with BMP = 0.5, on sloping fields the
optimal length has to be established according to the procedure
previously presented.
Contrarily, for BMP = 0 for which no upstream lateral occurs,
Fig. 11b shows that laterals designed with the field assumed as flat
can also be applied to sloping fields, for slopes which do not over-
come an |S0max | value over which pressure heads are not admitted.
(d) (u)
|S0max | can be easily evaluated by putting h∗i = h∗max and i = nd , into
Eq. (1b), obtaining:
 
(−r)
K Hnd
|S0max | =
nd
By imposing into Eq. (31a) the optimal number of emitters corre-
sponding to a flat field, provided by Eq. (24c), a good fit of Eq. (31a)
can be obtained, by once again fitting a power-law relationship
(R2 = 1, standard error of the estimate, SEE = 25.9):
|S0max |
 h 0.6353
∗n
= 0.2514
K K
The values for CV min and |S0max | for the two extreme values
of K (7·10−7 and 8·10−6 ) and h∗n (10 and 120) obtained by using
Eq. (30b) and Eq. (31b) are indicated as white and black dots in
Fig. 10b and in Fig. 11a–b, respectively, testing the reliability of
both equations. As an example, for BMP = 0, qn , = 5 l/h, D = 32 mm
(K = 5.52·10−7 ), hn = 22.5 m, S = 1.5 m (h∗n = 15), Fig. 6a illustrates the
pressure head distribution of the lateral designed according to the
proposed procedure, Eqs. (21 and 22a), and according to Eqs, (30b)
and (31b). As expected, the results show that i) for slope values
corresponding to the optimal S0 , the maximum length of the lat-
eral under the admitted pressure heads occurs, ii) for S 0 = |S0CVmin |,
CV = CVmin = 2%, whereas forS0 = |S0max |, the pressure heads at the
top-end of the lateral agrees with the maximum admitted value,
1.1 hn .
9. Conclusions
The importance of an adequate design of drip lateral hydraulics
aimed to find the optimal length or diameter of laterals laid on slop-
ing fields has been emphasized. When designing paired laterals, it
is fundamental to evaluate the best manifold position (BMP), which
usually is defined as the location of the manifold determining the
same minimum pressure in uphill and downhill laterals.
Recently, by deriving simple explicit relationships as a func-
tion of 16 calibration constants, Baiamonte (2016a) identified equal
to 0.24, the best position of the manifold (BMP) associated with
the optimal paired lateral design in sloping fields. Moreover, the
Author simplified the analytical design procedure introduced by
Baiamonte et al. (2015), to optimally design a paired drip lateral,
which required solving a system of four implicit equations.
Following these main finding, the most important advances in
the optimal design of irrigation laterals laid on uniform slopes cov-
ered in this paper can be summarized as follows:
• Simple and monomial design relationships are derived to design
(31a) laterals on uniform sloping fields that satisfies limited pressure
head variations along the entire lateral (± ␦ of the emitters’ nom-
inal value, with ␦ pressure head tolerance), for five considered
layouts (#1. Manifold inside the irrigation unit, for downward-
paired laterals, #2. Manifold at the edge of the irrigation unit for
downward laterals, #3. Manifold inside the irrigation unit for flat
laterals, #4. Manifold at the edge of the irrigation unit for flat lat-
(31b) erals, #5. Manifold at the edge of the irrigation unit for upward
laterals).
• These relationships are valid for uniform laterals’ slope, under the
assumptions of neglecting i) manufacturing coefficient of varia-
tion, ii) the variations of emitters’ flow rate along the lateral, iii)
the local losses due to emitters’ insertions. Thus, the procedure
can be applied when the morphology of emitter connections does
not produce significant reductions of the lateral cross section.
Moreover, the procedure here presented is a partial approach
since it is valid only for one-lateral units and cannot be used to
 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174
irrigation units of branching pipe networks, with many laterals
and one manifold.
• The proposed procedure reduces calibration constants of the sim-
plified procedure of Baiamonte (2016a), from 16 to 3, provides
much simpler relationships, making evident the influence of flow
rate and pipe diameter exponents in the design variables.
• The proposed procedure extends previous design relationships
to i) any imposed pressure head tolerance and ii) to the Hazen-
Williams resistance equation, which is also widely accepted for
smooth-wall Polyethylene Pipe, especially when the Reynolds
number exceeds 105 .
• The proposed procedure is a good approximation of the classical
step-by-step procedure (SBS) with relative errors that are only a
bit greater than 2%, and because of its simplicity it is also usable
for the general public.
• For an assigned slope, the design procedure allows a lateral to
be designed with the maximum length, optimizing water use
efficiency (WUE), assuring low values of pressure head variabil-
ity (CV < 6.1%, for ␦ = 0.1) and high values of emission uniformity
(EU > 95%, for ␦ = 0.1).
• For layout #4 (Manifold at the edge of the irrigation unit for flat
laterals), a positive comparison with lateral lengths suggested in
3 manufacture’s catalogues is carried out.
• For each considered layout, relationships between the pressure
head tolerance, the variation coefficient of pressure head along
the lateral and the emission uniformity coefficient of Keller and
Karmeli, are presented.
• Evaluations highlighting the effects of design choices in terms of
energy saving, comparisons between optimal lateral lengths for
the five layouts, and analysis of lateral pressure heads variability,
are carried out. As an example, for the most common case of flat
fields, results showed that, compared to the case in which the
manifold is at the boundary of the irrigation unit (BMP = 0), the
manifold located in the middle (BMP = 0.5) requires an emitters’
pressure head that is 6.7 times lower.
Acknowledgments
Research was supported by PRIN 2015 project granted by MIUR
(Italian Ministry for University and Research) No. 2015AKR4HX.
Part of the work disclosed here is patent pending, Ministero
dello Sviluppo Economico – Ufficio Italiano Brevetti e Marchi, n◦
102015000017479, May 25, 2015.
References
Agnese, C., Baiamonte, G., Cammalleri, C., 2014. Modelling the occurrence of rainy
days in a typical Mediterranean. Adv Water Resour. 64, 62–76, http://dx.doi.
org/10.1016/j.advwatres.2013.12.005.
Al-Amoud, A.I., 1995. Significance of energy losses due to emitter connections in
trickle irrigation lines. J. Agric. Eng. Res. 60 (1), 1–5.
Al-Samarmad, O.T., 2002. Optimum Dimension of a Trickle Irrigation Subunit by
Using Local Prices. Dept. of Irrig. and Drain. Eng., Coll. of Eng., University of
Baghdad, Baghdad, Iraq (M. Sc. Thesis).
Anderson, M.P., Woessner, W.W., 1992. Applied Groundwater
Modeling—Simulation of Flow and Advective Transport: San Diego, Calif.
Academic Press, Inc (381 p.).
Baiamonte, G., Singh, V.P., 2016a. Analytical solution of kinematic wave time of
concentration for overland flow under Green-Ampt Infiltration. J. Hydrol. E
ASCE 21 (3), 04015072, http://dx.doi.org/10.1061/(ASCE)HE.1943-5584.
0001266.
Baiamonte, G., Singh, V.P., 2016b. Overland flow times of concentration for
hillslopes of complex topography. J. Irrig. Drain. E-ASCE 142 (3), 04015059,
http://dx.doi.org/10.1061/(ASCE)IR.1943-4774.0000984.
Baiamonte, G., Provenzano, G., Rallo, G., 2015. Analytical approach determining the
optimal length of paired drip laterals in uniformly sloped fields. J. Irrig. Drain
Eng. 141 (1), 04014042, http://dx.doi.org/10.1061/(ASCE)IR.1943-4774.
0000768.
Baiamonte, G., Bagarello, V., D’Asaro, F., Palmeri, V., 2017. Factors influencing point
measurement of near-surface saturated soil hydraulic conductivity in a small
Sicilian basin. Land Degrad. Dev. 28, 970–982, http://dx.doi.org/10.1002/ldr.
2674.
173
Baiamonte, G., 2016a. Simple relationships for the optimal design of paired drip
laterals on uniform slopes. J. Irrig. Drain. E-ASCE. 142 (2), 04015054, http://dx.
doi.org/10.1061/(ASCE)IR.1943-4774.0000971.
Baiamonte, G., 2016b. Closure to Simple relationships for the optimal design of
paired drip laterals on uniform slopes by giorgio baiamonte. J. Irrig. Drain.
E-ASCE 142 (12), 07016019, http://dx.doi.org/10.1061/(ASCE)IR.1943-4774.
0001126.
Baiamonte, G., 2016c. Discussion of Analysis of geometrical relationships and
friction losses in small-diameter lay-flat polyethylene pipes by giuseppe
provenzano, vincenzo alagna, dario autovino, juan manzano juarez, and
giovanni rallo. J. Irrig. Drain. E-ASCE 142 (10), 07016013, http://dx.doi.org/10.
1061/(ASCE)IR.1943-4774.0001071.
Baiamonte, G., 2017a. Design of concave and convex paired sloped drip laterals.
Agric. Water Manage. 191, 173–183, http://dx.doi.org/10.1016/j.agwat.2017.
06.015.
Baiamonte, G., 2017b. Closed-form solutions of the energy balance equation for
drip laterals under the Darcy-Weisbach resistance formula. Submit. ASCE J.
Irrig. Drain., E–ASCE.
Baiamonte, G., 2018. Erratum for simple relationships for the optimal design of
paired drip laterals on uniform slopes. J. Irrig. Drain. E-ASCE 144 (2), http://dx.
doi.org/10.1061/(ASCE)IR.1943-4774.0000971, 08217003.
Christiansen, J.E., 1942. Irrigation by Sprinkling. California Agricultural
Experimental Station Bull. No. 670. Univ. of California at Davis, Davis, Calif.
Chu, H., Wei, Z., Yuan, W., Zhou, X., Shengli, M.A., 2015. Fluid-Structure interaction
simulation and visualization experiments for pressure-compensating emitters.
J. Mech. Eng. 51 (12), 170–177, http://dx.doi.org/10.3901/JME.2015.12.170.
Fernández García, I., Montesinos, P., Camacho Poyato, E., Rodríguez Díaz, J.A., 2015.
Energy cost optimization in pressurized irrigation networks. Publ. line Irrig.
Sci., http://dx.doi.org/10.1007/s00271-015-0475-3.
Genius, M., Koundouriy, P., Naugesz, C., Tzouvelekas, V., 2012. Information
transmission in irrigation technology adoption and diffusion: social learning,
extension services and spatial effects. In: Working Paper 1211. Athens
University of Economics and Business, Athens-Greece, http://dx.doi.org/10.
1093/ajae/aat054.
Gill, P.E., Murray, W., Wright, M.H., 1989. Practical Optimization. Academic Press,
Inc., Inc San Diego, Calif, pp. 90–91.
Giraldez, J.V., Woolhiser, D.A., 1996. Analytical integration of the kinematic
equation for runoff on a plane under constant rainfall rate and Smith and
Parlange infiltration. Water Resour. Res. 32 (11), 3385–3389.
Jiang, S., Kang, Y., 2010. Simple method for the design of microirrigation paired
laterals. J. Irrig. Drain. Eng. 136 (4), 271–275.
Juana, L., Losada, A., Rodrigues-Sinobas, L., Sanchez, R., 2004. Analytical
relationships for designing rectangular drip irrigation units. J. Irrig. Drain.
E-ASCE, 47–59, http://dx.doi.org/10.1061/(ASCE)0733-9437(2004)130:1(47).
Kang, Y., Nishiyama, S., 1996. Analysis and design of microirrigation laterals. J.
Irrig. Drain. Eng. 122 (2), 75–82.
Kang, Y., Nishiyama, S., Chen, H., 1996. Design of microirrigation laterals on
nonuniform slopes. Irrig. Sci. 17, 3–14.
Kang, Y., 2000. Hydraulic characteristics and computerized design of
microirrigation systems. In: Ecosystem Service and Sustainable Watershed
Management in North China International Conference, Beijing, P.R. China,
August 23–25.
Karmeli, D., Keller, J., 1975. Trickle Irrigation Design. Rain Bird Sprinkler
Manufacturing Corporation, Glendora, Calif.
Keller, J., Bliesner, R.D., 1990. Sprinkle and Trickle Irrigation. The Blackburn Press,
New York, pp. 652, ISBN-13: 978–1930665194.
Keshtgar, A., Bhuiyan, M.A., Jayasuriya, N., 2013. Design of drip irrigation system
using microtubes for full emission uniformity. Irrig. Sci. 62 (5), 613–623,
http://dx.doi.org/10.1002/ird.1763.
Madramootoo, C.A., Khatri, K.C., Rigby, M., 1987. Hydraulic performances of five
different trickle irrigation emitters. Can. Agric. Eng. 30, 1–4.
Madramootoo, C.A., 1988. Effect of pressure changes on the discharge
characteristics of pressure compensating emitters. J. Agric. Eng. Res. 40 (2),
159–164.
Perea, H., Enciso-Medina, J., Singh, V., Dutta, D., Lesikar, B., 2013. Statistical analysis
of non-Pressure-Compensating and pressure-Compensating drip emitters. J.
Irrig. Drain Eng., 986–994, http://dx.doi.org/10.1061/(ASCE)IR.1943-4774.
0000644.
Pinthong, K., Merkley, G.P., Chittaladakorn, S., 2013. Flow path and hydraulic
analysis for on-farm pressurized irrigation systems. Irrig. Sci. 31 (3), 371–383.
Provenzano, G., Pumo, D., 2004. Experimental analysis of local pressure losses for
microirrigation laterals. J. Irrig. Drain. Eng. 130 (4), 318–324.
Royuela-Tomás, Á., Manzano-Juárez, J., Palau-Salvador, G., 2005. Discussion of
analytical relationships for designing rectangular drip irrigation units by luis
juana, alberto losada, leonor Rodríguez-Sinobas, and Raúl Sánchez. J. Irrig.
Drain. Eng., 484–486, http://dx.doi.org/10.1061/(ASCE)0733-
9437(2005)131:5(484).
Vallesquino, P., Luque-Escamilla, P.L., 2002. Equivalent friction factor method for
hydraulic calculation in irrigation laterals. J. Irrig. Drain. Eng. 128 (5), 278–286.
Wu, I.P., Gitlin, H.M., 1975. Energy gradient line for drip irrigation laterals. J. Irrig.
Drain. Div. 101 (IR4), 321–326.
Wu, I.P., Saruwatari, C.A., Gitlin, H.M., 1983. Design of drip irrigation lateral length
of uniform slopes. Irrig. Sci. 4 (2), 117–135.
Wu, I.P., Gitlin, H.M., Solomon, K.H., Saruwatari, C.A., 1986. In: Nakayama, F.S.,
Bucks, D.A. (Eds.), Design Principles: Trickle Irrigation for Crop Production.
Elsevier Science, Phoenix, pp. 53–92.
174 G. Baiamonte / Agricultural Water Management 199 (2018) 157–174

Wu, P.T., Zhu, D.L., Wang, J., 2010. Gravity-fed drip irrigation design procedure for Wu, I.P., 1997. An assessment of hydraulic design of micro-irrigation systems.
a single-manifold subunit. Irrig. Sci. 28 (4), 359–369. Agric. Water Manag. 32, 275–284.
Wu, I.P., 1975. Design of drip irrigation main lines. J. Irrig. Drain. Div. 101 (IR4), Zayani, K., Alouini, A., Lebdi, F., Lamaddalena, N., 2001. Design of drip irrigation
265–278. systems using the energy drop ratio approach. Trans. ASAE 44 (5), 1127–1133.