Limited on-demand water delivery as a managerial strategy to enhance operational flexibility in pressurized irrigation networks Authors: Younes Aminpour, Nicola Lamaddalena, Eisa Maroufpoor Published: 2026 ABSTRACT Implementing fully on-demand irrigation in existing rotational networks often requires extensive and costly infrastructure rehabilitation, limiting its practical applicability in many irrigation systems. Therefore, identifying operational strategies capable of improving delivery flexibility without major structural modification remains a critical challenge. In this study, limited on-demand operation was evaluated as an intermediate operational strategy for two pressurized drip irrigation networks in Iran operating under rotational water distribution systems: the gravity-fed Ziviyeh network and the pump-supplied Lali network. Hydraulic performance under different operational scenarios was simulated using the COPAM modeling framework, while network flexibility was evaluated using the Delivery Flexibility Index (DFI). The results showed that neither network could reliably support fully on-demand operation under existing conditions because the minimum required hydrant pressure was satisfied in only 10% and 13% of operational combinations in the Ziviyeh and Lali networks, respectively. In addition, pressure-deficit conditions affected approximately 26% and 27% of hydrants in the Ziviyeh and Lali networks, respectively, under a 90% probability of occurrence. However, under limited on-demand operation, pressure deficits occurred in less than 2% of hydrants for DFI values up to 3 in the Ziviyeh network and 4.5 in the Lali network, while hydraulic reliability remained within acceptable limits. These findings demonstrate that DFI-based operational management can substantially improve delivery flexibility without requiring physical infrastructure upgrades. The proposed approach provides a practical and economically feasible transitional modernization strategy for pressurized irrigation networks operating under water-scarcity and financial constraints. Introduction Growing pressure on limited freshwater resources, coupled with increasing competition among water-use sectors, has intensified the need to reassess conventional water delivery and distribution practices in irrigation networks. Traditional systems, including continuous and rotational distribution, rely on fixed schedules and predetermined delivery volumes, limiting their ability to accommodate the spatial and temporal variability of crop water requirements. Consequently, irrigation efficiency, distribution uniformity, and overall water productivity are often reduced1,2. Such rigid operational structures constrain the responsiveness of irrigation networks to fluctuations in water availability and shifts in cropping patterns and, under water-scarcity conditions, can exacerbate resource competition, trigger social conflicts, and undermine system sustainability3,4. Enhancing water delivery flexibility and expanding the decision-making autonomy of water users have therefore emerged as central strategies for improving the performance and resilience of irrigation systems. Allowing farmers to actively determine the timing and volume of water delivery improves the alignment between irrigation practices and actual crop water demands, leading to higher crop productivity, increased water-use efficiency, and reduced social inequities within irrigation schemes3,5. In this context, on-demand irrigation systems, which deliver water in response to real-time user and crop requirements, have been shown to outperform fixed-schedule systems in terms of water loss reduction and operational adaptability6. By enabling participatory and adaptive management, on-demand systems can significantly enhance the technical, economic, and social sustainability of irrigation networks, particularly in regions facing chronic water scarcity3. Despite these advantages, full-scale implementation of on-demand operation in existing irrigation networks typically requires substantial upgrades to conveyance, control, and measurement infrastructure. Such interventions involve considerable financial investment and are frequently beyond the practical capacity of both water users and managing authorities3. Accordingly, there is a critical need for alternative management strategies that enhance delivery flexibility while minimizing the need for large-scale physical modifications5,6. Although delivery flexibility is widely recognized as a key performance attribute of irrigation systems, its conceptualization and quantitative assessment remain insufficiently developed7. For example, Merriam et al.7 introduced the concept of canal “rigidity” to describe unused capacity and non-operational periods, whereas Burt et al.8 proposed a descriptive flexibility index based on discharge, irrigation frequency, and duration. Together, these frameworks provide valuable foundations for comparative evaluation of irrigation system performance and support the development of decision-support tools for water allocation management. Within this framework, limited on-demand operation, first introduced in FAO Irrigation and Drainage Paper No. 449, represents an intermediate approach between rotational and fully on-demand systems. Under this regime, water users may request irrigation at preferred times in accordance with actual crop water requirements, while a predefined maximum discharge is imposed at each turnout to maintain operational control3,10. By balancing efficiency, equity, and managerial oversight, limited on-demand operation offers a cost-effective and rapidly implementable transitional pathway toward fully on-demand irrigation systems. On-demand irrigation systems, in particular, introduce significant hydraulic complexity due to the random nature of user withdrawals and the resulting stochastic demand patterns, which directly affect system reliability and pressure stability11–13. Consequently, several studies have emphasized the importance of stochastic demand analysis and probabilistic hydraulic modeling for evaluating system performance under peak-demand conditions and ensuring reliable operation of pressurized irrigation networks14,15. Simultaneously, the increasing adoption of pressurized irrigation systems has strengthened the coupling between water distribution and energy consumption, highlighting the importance of the water–energy nexus in modern irrigation infrastructure16,17. Energy efficiency, hydraulic losses, and pumping requirements have become central design considerations, and recent studies have demonstrated that significant opportunities exist for energy recovery within irrigation networks, particularly through pressure dissipation and micro-hydropower integration18,19. Furthermore, the emergence of smart irrigation systems based on IoT technologies, real-time monitoring, and advanced control algorithms has enabled improved pressure management and operational efficiency20,21. These systems allow dynamic regulation of irrigation demand and supply, thereby enhancing both water-use efficiency and energy performance22,23. Overall, the integration of smart control strategies with hydraulic optimization frameworks represents a key direction for future irrigation system modernization. Despite the extensive body of research on pressurized irrigation systems and on-demand water delivery, most previous studies have primarily focused on the hydraulic design of newly developed on-demand networks or on structural rehabilitation measures required for converting existing rotational systems into fully flexible operation schemes. Consequently, limited attention has been devoted to operational and management strategies capable of improving delivery flexibility within the existing hydraulic capacity of irrigation networks. In particular, the application of Delivery Flexibility Index (DFI)-based operational management as a practical approach for regulating hydrant discharge and evaluating the feasibility of limited on-demand operation in existing rotational drip irrigation systems remains insufficiently investigated. Furthermore, only a limited number of studies have evaluated the variation of hydraulic reliability and pressure-deficit conditions under progressive increases in delivery flexibility in large-scale pressurized irrigation networks. Addressing these research gaps is essential for developing cost-effective and operationally feasible transition strategies toward more flexible irrigation operation under water-scarcity conditions. In this study, the feasibility of limited on-demand operation is evaluated for two drip irrigation networks in Iran: the gravity-fed Ziviyeh network and the pump-supplied Lali network. Both networks were originally designed based on peak crop water demand and operate under rotational distribution. The results provide insights applicable to similar irrigation systems in arid and semi-arid regions of Iran and beyond. The specific objectives of this study are to: Assess the potential for limited on-demand operation in drip irrigation networks with traditional rotational distribution;Identify the desirable range of the Delivery Flexibility Index (DFI) for the studied systems; and.Analyze seasonal variations in delivery flexibility during the irrigation period. Assess the potential for limited on-demand operation in drip irrigation networks with traditional rotational distribution; Assess the potential for limited on-demand operation in drip irrigation networks with traditional rotational distribution; Identify the desirable range of the Delivery Flexibility Index (DFI) for the studied systems; and. Identify the desirable range of the Delivery Flexibility Index (DFI) for the studied systems; and. Analyze seasonal variations in delivery flexibility during the irrigation period. Analyze seasonal variations in delivery flexibility during the irrigation period. Materials and methods General characteristics of the studied irrigation networks Modernized Ziviyeh drip irrigation network The Ziviyeh irrigation network is located in Kurdistan Province, Iran, within the geographical coordinates of 34°07′ to 35°02′ N latitude and 46°35′ to 46°45′ E longitude (Fig. 1b,c). This network represents a modernized irrigation system developed under national policies aimed at improving water-use efficiency, in which the original sprinkler irrigation system was converted to a drip irrigation system. Accordingly, the main conveyance pipelines were designed based on peak water demand under sprinkler irrigation conditions, whereas the secondary network was designed to meet peak water demand under drip irrigation conditions. The cropping pattern of the network mainly consists of vegetable crops. Fig. 1Location of the studied irrigation networks (https://www.qgis.org/ and https://www.google.com/earth/): (a) location of the networks on the map of Iran; (b) location of the Ziviyeh network in Kurdistan Province; (c) Ziviyeh drip irrigation network; (d) location of the Lali network in Khuzestan Province; (e) Lali drip irrigation network. Location of the studied irrigation networks (https://www.qgis.org/ and https://www.google.com/earth/): (a) location of the networks on the map of Iran; (b) location of the Ziviyeh network in Kurdistan Province; (c) Ziviyeh drip irrigation network; (d) location of the Lali network in Khuzestan Province; (e) Lali drip irrigation network. The net irrigated area of the network is 1088.62 ha, with land elevations ranging from 1260 to 1490 m.a.s.l. Water is supplied directly from the reservoir lake. Key reservoir elevations include the minimum water level at 1555.2 m.a.s.l., the normal water level at 1585 m.a.s.l., and the network intake elevation at 1544 m.a.s.l. The design discharge of the network is 850 L s⁻¹. Overall, the Ziviyeh network comprises 755 pipeline sections and 512 hydrants. Network pressure is supplied entirely by gravity, with no pumping stations involved. The piezometric head at the network set point (upstream boundary) was assumed to be equal to the minimum reservoir water level, representing the most critical operating condition. Lali drip irrigation network The Lali irrigation network is located in Khuzestan Province, Iran, within the geographical coordinates of 32°18′ to 32°21′ N latitude and 49°15′ to 49°19′ E longitude (Fig. 1d,e). The net irrigated area of the network is 790.57 ha. The cropping pattern primarily consists of orchard crops, and irrigation is performed using a drip irrigation system. The required pressure is supplied through a pumping station. The design discharge of the network, considering future expansion of irrigated areas, is 786.5 L s⁻¹. This discharge is supplied by a pumping station equipped with ten WKL150/4 electro-pumps. In total, the network includes 944 pipeline sections and 747 hydrants. An overview of the studied irrigation networks is shown in Fig. 2. In both networks, the minimum required pressure at the hydrants is 20 m, and the daily operation duration is 22 h. The continuous specific discharge during the peak demand period is 0.78 L s⁻¹ ha⁻¹ for the Ziviyeh network and 0.83 L s⁻¹ ha⁻¹ for the Lali network. The general characteristics of both networks are summarized in Table 1. Table 1Some specifications of the Ziviyeh and Lali drip irrigation networks.InformationZiviyeh irrigation networkLali irrigation networkArea (ha)1088.62790.57Continuous specific discharge (Peak period) (l/s/ha)0.780.83Maximum discharge of each turnout* (hydrant) (l/s)3.02-62.002.02-12.00Setpoint elevation (m.a.s.l)1555.2646.21Number of sections755944Number of turnouts (hydrant)512747* Maximum hydrant discharge under simultaneous irrigation of all associated plots. Table 1Some specifications of the Ziviyeh and Lali drip irrigation networks.InformationZiviyeh irrigation networkLali irrigation networkArea (ha)1088.62790.57Continuous specific discharge (Peak period) (l/s/ha)0.780.83Maximum discharge of each turnout* (hydrant) (l/s)3.02-62.002.02-12.00Setpoint elevation (m.a.s.l)1555.2646.21Number of sections755944Number of turnouts (hydrant)512747* Maximum hydrant discharge under simultaneous irrigation of all associated plots. Some specifications of the Ziviyeh and Lali drip irrigation networks. * Maximum hydrant discharge under simultaneous irrigation of all associated plots. Fig. 2Overview of the studied irrigation networks and their conveyance pipelines: (a) Ziviyeh drip irrigation network, Kurdistan Province, Iran; (b) Lali drip irrigation network, Khuzestan Province, Iran. Overview of the studied irrigation networks and their conveyance pipelines: (a) Ziviyeh drip irrigation network, Kurdistan Province, Iran; (b) Lali drip irrigation network, Khuzestan Province, Iran. Methodology To simulate and analyze the hydraulic behavior of the irrigation networks under different on-demand operational scenarios, the COPAM software package was employed. COPAM was specifically developed for the hydraulic analysis and design of pressurized irrigation networks operating under on-demand conditions and has been widely applied in both research studies and irrigation modernization projects worldwide11,24. The model is particularly appropriate for the present study because it integrates probabilistic demand generation with hydraulic network simulation, enabling evaluation of hydraulic performance under stochastic hydrant operation conditions. Unlike deterministic design approaches based exclusively on peak demand assumptions, COPAM enables assessment of pressure reliability, hydrant performance, and hydraulic adequacy under different operational configurations, which is essential for evaluating limited on-demand operation and DFI-based operational management strategies. To evaluate and optimize the performance of the studied irrigation networks under limited on-demand conditions, the DFI was applied. This index represents the ability of a network to respond to unpredictable water demands from users9. The DFI is defined as the ratio of the allocated discharge at a hydrant to its duty discharge. The duty discharge represents the minimum discharge required for optimal irrigation of the area served by a hydrant and is calculated as the product of the continuous specific discharge and the net irrigated area. In this study, the duty discharge and the maximum allocated discharge were first calculated for each hydrant. The maximum allocated discharge was determined assuming simultaneous irrigation of all plots supplied by the hydrant. In the Ziviyeh irrigation network, the maximum allocated discharge ranged from 3.02 to 62.00 L s⁻¹, whereas in the Lali network it ranged from 2.02 to 12.00 L s⁻¹. Consequently, the maximum DFI for each hydrant was obtained as the ratio of the maximum allocated discharge to the duty discharge. During the peak demand period, the mean maximum DFI values were 12.00 and 11.63 for the Ziviyeh and Lali networks, respectively, which would be required under fully on-demand operation. Under limited on-demand conditions, hydraulic analyses were conducted for DFI values ranging from 1 to 6, with increments of 0.5, to determine the optimal DFI range for each network during the peak demand period. To identify permissible DFI values during other months of the irrigation season, a wider range of delivery flexibility values was examined. According to FAO guidelines9,11, the recommended minimum DFI ranges from 1.5 to 2, while the maximum recommended DFI varies between 6 and 8. Performance evaluation indices The hydraulic performance of the irrigation networks under various operational scenarios was evaluated using the ICARE and AKLA analysis modules integrated within the COPAM framework. These models are commonly applied for assessing the hydraulic reliability of pressurized irrigation systems operating under on-demand conditions. In the ICARE model, a “combination” represents a particular operational configuration of simultaneously operating hydrants corresponding to a given upstream discharge condition. For each generated combination, the hydraulic adequacy of the network is evaluated by comparing the available pressure head at each operating hydrant with the minimum required pressure head. A hydrant combination is considered hydraulically adequate when the following condition is met for all operating hydrants:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\left( {{H_j}} \right)_r} \geqslant {H_{\hbox{min} }}$$\end{document} where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${H_{j,r}}$$\end{document} is the pressure head available at hydrant j under operational combination r, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${H_{\hbox{min} }}$$\end{document} is the minimum required pressure head at each hydrant, respectively. Although the ICARE model identifies whether hydraulic deficiencies exist within the network, it does not determine the spatial distribution of malfunctioning hydrants. Therefore, the AKLA model was additionally applied to evaluate outlet-level hydraulic performance through the Relative Pressure Deficit (RPD) and Reliability Index (RI). The Relative Pressure Deficit (RPD) quantifies the degree to which the available pressure at a hydrant deviates from the required minimum pressure:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$RPD=\frac{{{H_{j,r}} - {H_{\hbox{min} }}}}{{{H_{\hbox{min} }}}}$$\end{document} The Reliability Index (RI) represents the probability that a hydrant receives sufficient pressure across all simulated operational combinations:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$RI=\frac{{\sum\limits_{{R=1}}^{C} {I{h_{j,r}}I{p_{j,r}}} }}{{\sum\limits_{{R=1}}^{K} {I{h_{j,r}}} }}$$\end{document} where RPD represents the relative pressure deficit at hydrant j; RI denotes the reliability of hydrant j; \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I{h_{j,r}}=1$$\end{document} if hydrant j is open in combination r; \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I{h_{j,r}}=0$$\end{document} if hydrant j is closed in combination r; \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I{p_{j,r}}=1$$\end{document} if the pressure head at the open hydrant j in combination r exceeds the minimum allowable pressure head; \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I{p_{j,r}}=0$$\end{document} if the pressure head at the open hydrant j in combination r is lower than the minimum allowable pressure head; and C is the total number of generated combinations. An RPD value greater than or equal to zero indicates adequate pressure supply at the hydrant. Similarly, RI values approaching unity indicate higher hydraulic reliability and more satisfactory network operating conditions. According to FAO guidelines11, RI values below 0.8 are considered hydraulically unacceptable. Although the total number of theoretically possible hydrant operation combinations in large-scale pressurized irrigation networks is extremely high, probabilistic hydraulic assessment methods are generally based on representative stochastic sampling rather than exhaustive evaluation of all possible combinations. In the present study, 475 operational scenarios were generated to represent the dominant probabilistic characteristics of hydrant simultaneity and discharge variability within the networks. Previous applications of COPAM and similar probabilistic hydraulic models have shown that hydraulic performance indicators, including pressure reliability and pressure-deficit occurrence, generally converge beyond a representative sampling threshold, whereas computational demand increases substantially with additional simulations11. Therefore, the selected number of operational scenarios was considered sufficient to provide a statistically representative assessment of network hydraulic performance within practical computational limits. Results and discussion Maximum probable discharge required by the irrigation networks under full and limited on-demand operation Figure 3 illustrates the variation in the maximum probable discharge required by the irrigation networks during the peak demand period as a function of the Delivery Flexibility Index (DFI). As DFI increases, the required network discharge increases correspondingly in both irrigation systems. However, the rate of increase gradually diminishes at higher DFI values, indicating a nonlinear response of the system to increasing demand levels. Fig. 3Variation trend of the maximum probable discharge required for the Ziviyeh and Lali irrigation networks under full on-demand conditions with different values of the Delivery Flexibility Index (DFI). Variation trend of the maximum probable discharge required for the Ziviyeh and Lali irrigation networks under full on-demand conditions with different values of the Delivery Flexibility Index (DFI). The average slope of the curve per unit increase in DFI was estimated at approximately 20.6 L s⁻¹ for the Ziviyeh irrigation network and 13.7 L s⁻¹ for the Lali irrigation network. These results are consistent with the first Clément equation11, which describes the relationship between the number of hydrants and the percentage increase in network discharge. According to this relationship, as the number of hydrants increases, the percentage increase in probable discharge decreases. Given that the Ziviyeh network comprises 512 hydrants, whereas the Lali network includes 747 hydrants, the lower percentage increase in discharge observed for the Lali network is therefore expected. At DFI = 1, the discharge allocated to each outlet is equal to its nominal (duty) discharge. Consequently, the discharge predicted by the COPAM model should correspond to the design discharge of the network under a rotational distribution system. The design discharges of the Ziviyeh and Lali networks are 849 L s⁻¹ and 656 L s⁻¹, respectively (excluding future land development considerations). In contrast, the maximum probable discharges estimated under on-demand operation were 996 L s⁻¹ for Ziviyeh and 747 L s⁻¹ for Lali. In other words, under on-demand conditions at DFI = 1, the required network discharge exceeds the design discharge by approximately 17% in the Ziviyeh network and 14% in the Lali network. At DFI = 1, since all hydrants are assumed to be open, the network discharges estimated by the two approaches; peak-period-based design and the COPAM probabilistic model were expected to be identical. This discrepancy reflects the fundamental difference between deterministic network design procedures and probabilistic hydraulic simulation approaches. In conventional rotational irrigation design, network discharge is generally determined based on fixed peak-period assumptions and uniform outlet operating conditions. In contrast, the COPAM framework evaluates stochastic operational configurations generated through probabilistic hydrant activation patterns and outlet simultaneity analysis. Therefore, even under DFI = 1 conditions, localized temporal overlap among hydrant operations may produce short-duration peak discharge conditions exceeding the minimum required network discharge. Under full on-demand operation (maximum DFI), the maximum probable discharges required by the Ziviyeh and Lali networks were calculated as 1125 L s⁻¹ and 864 L s⁻¹, respectively. Considering the net irrigated areas of 1088.20 ha for Ziviyeh and 790.57 ha for Lali, these values correspond to continuous specific discharges of 1.03 L s⁻¹ ha⁻¹ and 1.09 L s⁻¹ ha⁻¹, respectively. The main network of Ziviyeh was originally designed based on the continuous specific discharge of a sprinkler irrigation system during the peak demand period (1.14 L s⁻¹ ha⁻¹), and therefore possesses sufficient physical capacity to support a transition toward full on-demand operation. In contrast, the secondary network, which was designed based on the peak-period continuous specific discharge of drip irrigation (0.78 L s⁻¹ ha⁻¹), faces hydraulic limitations. Similarly, in the Lali irrigation network, the design continuous specific discharge (0.83 L s⁻¹ ha⁻¹) is lower than that required under full on-demand operation (1.09 L s⁻¹ ha⁻¹). Overall, the maximum probable discharge required under full on-demand operation is approximately 30% higher than that required under rotational distribution. The increase in maximum probable discharge observed under full on-demand operation is consistent with the general hydraulic trend reported in previous studies on pressurized irrigation systems. For example, Aminpour et al.25 reported discharge increases of approximately 58–62% in Italian drip irrigation networks operating under on-demand conditions. This difference likely reflects variations in the original network design basis, hydraulic safety margins, and operational assumptions. In many irrigation systems in developing countries, network components are commonly designed with conservative hydraulic criteria to accommodate uncertainties associated with cropping patterns, unauthorized water withdrawal, and future network expansion. Consequently, part of the hydraulic capacity required for flexible operation may already be available within the system, even when the network was originally designed for rotational water distribution. Increasing network discharge typically requires physical infrastructure upgrades, such as enlarging pipe diameters, improving junction capacity, or reinforcing pumping stations, as well as managerial interventions, including redefining the network set point, adjusting abstraction schedules, and revising the DFI. Since structural interventions are generally costly and time-consuming, managerial strategies based on demand limitation, namely Limited on-demand operation, represent a more efficient and cost-effective approach for enhancing the flexibility of existing irrigation networks. Performance assessment of the irrigation networks under full on-demand operation (DFI ≈ 12) Analysis of indexed characteristic curves The indexed characteristic curves for both irrigation networks, generated based on 475 random configurations, are presented in Fig. 4. These curves depict pressure variations at different network nodes relative to the set point and constitute the primary tool for assessing the hydraulic stability of pressurized irrigation networks under random and on-demand operating conditions11. Fig. 4Indexed characteristic curves under full on-demand distribution: (a) Ziviyeh irrigation network, (b) Lali irrigation network. Indexed characteristic curves under full on-demand distribution: (a) Ziviyeh irrigation network, (b) Lali irrigation network. In the Ziviyeh irrigation network, the set point corresponds to the minimum reservoir water level, whereas in the Lali network it is defined at the pumping station outlet. In the Ziviyeh network, the set point lies below the 10% indexed characteristic curve, while in the Lali network it coincides with the 13% curve. These results indicate that, in approximately 90% of the random outlet combinations in the Ziviyeh network, the pressure at least one outlet falls below the required design pressure (20 m). A similar condition was observed in the Lali network, where this proportion reaches approximately 87%. The results indicate that under current conditions, full on-demand operation is not hydraulically feasible in either network without pressure correction measures. The convergence of results between the two networks, despite their different pressure supply mechanisms (gravity-fed in Ziviyeh and pump-fed in Lali), suggests that pressure limitations are not solely operational but also structural in nature. In other words, pressure deficits under simultaneous outlet operation are inherently embedded in the network configuration and are likely attributable to uneven hydraulic load distribution in secondary pipelines, excessive longitudinal head losses, or limited capacity at terminal nodes. These results highlight the necessity of redefining the network set point and identifying an optimal range of the Delivery Flexibility Index (DFI) to improve hydraulic performance. Accordingly, prior to any transition toward full on-demand operation, the implementation of limited on-demand operation is recommended as an intermediate, technically feasible, and cost-effective solution. Relative Pressure Deficit Index (RPD) and Reliability Index (RI) Figure 5 presents the variation of the Relative Pressure Deficit Index (RPD) at individual hydrants under full on-demand operation (DFI ≈ 12) for the set points of both networks. In the Ziviyeh network, with a 90% probability of occurrence, 134 hydrants (approximately 26% of all hydrants) exhibit RPD < 0, indicating insufficient pressure. In the Lali network, this condition was observed for 203 hydrants (approximately 27% of all hydrants). The minimum RPD values were − 30% for Ziviyeh and − 12% for Lali, representing the most critical pressure-deficit locations. Fig. 5Variations of the Relative Pressure Deficit index (RPD) at each hydrant, based on 475 generated operational scenarios: (a) Ziviyeh irrigation network (minimum reservoir level) under full on-demand conditions (DFI ≈ 12), (b) Lali irrigation network under full on-demand conditions (DFI ≈ 11.63). Variations of the Relative Pressure Deficit index (RPD) at each hydrant, based on 475 generated operational scenarios: (a) Ziviyeh irrigation network (minimum reservoir level) under full on-demand conditions (DFI ≈ 12), (b) Lali irrigation network under full on-demand conditions (DFI ≈ 11.63). The Reliability Index (RI), shown in Fig. 6, reveals a consistent pattern. In the Ziviyeh network, 41% of hydrants (209 hydrants) exhibit RI = 1, indicating full pressure supply, while 33% (169 hydrants) fall within the range 0.9–1.0, 10% (53 hydrants) within 0.8–0.9, and 16% (81 hydrants) have RI < 0.8. Four hydrants show RI values below 0.2, indicating severe hydraulic unreliability. In the Lali network, 40% of hydrants have RI = 1, 33% fall within 0.9–1.0, 13% within 0.8–0.9, and 14% exhibit RI < 0.8, with a minimum RI value of 0.45. Fig. 6Variations of the Reliability Index (RI) at each hydrant, based on 475 generated operational scenarios: (a) Ziviyeh irrigation network (minimum reservoir level) under full on-demand conditions (DFI ≈ 12), (b) Lali irrigation network under full on-demand conditions (DFI ≈ 11.63). Variations of the Reliability Index (RI) at each hydrant, based on 475 generated operational scenarios: (a) Ziviyeh irrigation network (minimum reservoir level) under full on-demand conditions (DFI ≈ 12), (b) Lali irrigation network under full on-demand conditions (DFI ≈ 11.63). An evaluation of the Ziviyeh network at the normal reservoir water level indicates that the set point is located above the upper envelope of the indexed characteristic curves with a 90% probability, ensuring that the required pressure is supplied to all hydrants. Under this condition, all hydrants exhibit acceptable RI values, and the network can be considered hydraulically secure. The hydraulic reliability conditions identified in the present study are consistent with previous investigations indicating that secondary and terminal pipeline sections generally exhibit the highest hydraulic sensitivity under flexible irrigation operation. Similar conclusions were reported by Lamaddalena and Sagardoy11, who demonstrated that increasing hydrant simultaneity substantially increases pressure fluctuations in downstream sections of pressurized irrigation networks. Overall, the results indicate that, under current conditions, neither network possesses sufficient capacity to reliably support full on-demand operation. The high proportion of unsupplied outlet combinations reflects the strong sensitivity of secondary networks to discharge fluctuations and outlet operation patterns. The combined RPD and RI results confirm that pressure limitations are fundamentally structural. Even in the Ziviyeh network, which has relatively adequate main network capacity, more than one-quarter of hydrants fail to receive the minimum required pressure under various operating scenarios. Although increasing pipe diameters is often proposed as the primary solution for enabling full on-demand operation11,24,25, this approach requires extensive redesign, excavation, and substantial financial investment, making it impractical in many real-world applications. In contrast, limited on-demand operation, based on managerial control of the Delivery Flexibility Index, offers a cost-effective, rapidly implementable alternative. By adjusting DFI, a significant proportion of hydraulically deficient operating combinations can be converted into acceptable ones without costly structural modifications. Moreover, unlike local pressure-boosting solutions (e.g., auxiliary pumping stations), this approach avoids long-term energy, maintenance, and operational costs, making it a more sustainable option. Feasibility of limited on-demand operation through DFI management Indexed characteristic curves at different DFI values Figure 7 illustrates the evolution of indexed characteristic curves for 475 operating scenarios across different DFI values. In both irrigation networks, at DFI = 1, the COPAM model slightly overestimates the maximum network discharge compared to the rotational distribution requirement. Nevertheless, both networks remain hydraulically stable, with the set point located within the upper envelope of the 90% indexed curves, indicating sufficient capacity to accommodate low levels of demand flexibility. Fig. 7Indexed characteristic curves under limited on-demand distribution for different values of the Delivery Flexibility Index (DFI): (a) Ziviyeh irrigation network, (b) Lali irrigation network. Indexed characteristic curves under limited on-demand distribution for different values of the Delivery Flexibility Index (DFI): (a) Ziviyeh irrigation network, (b) Lali irrigation network. As DFI gradually increases reflecting greater delivery flexibility and increased farmer autonomy the hydraulic behavior of the networks changes. Higher DFI values lead to increased instantaneous network discharge and, simultaneously, to reduced dynamic pressure in certain parts of the network due to altered flow patterns and increased friction losses in secondary pipelines. At DFI = 6, the proportion of operating combinations in which the required pressure is fully supplied to all hydrants is approximately 31% in the Ziviyeh network (minimum reservoir level) and 57% in the Lali network. In contrast, at DFI = 3, which may be considered an intermediate flexibility level, the proportion of successful combinations increases to 72% in Ziviyeh and 92% in Lali. These results indicate that, while both networks exhibit stable behavior at low DFI values, increasing the Delivery Flexibility Index progressively exposes structural pressure limitations within secondary networks. These limitations are particularly pronounced in the Ziviyeh network due to its gravity-fed pressure supply and the absence of pumping stations. Accordingly, a detailed outlet-level pressure analysis is required to identify critical locations and define safe operating ranges for limited on-demand operation, forming the basis for managerial decision-making regarding the optimal DFI. RPD and RI analysis under different DFI values Figure 8 shows the variation in the Relative Pressure Deficit Index, while Fig. 9 presents changes in the Reliability Index for different DFI values. Simulation results indicate that, in both irrigation networks, at DFI = 1, the minimum required pressure is supplied to all hydrants, and no outlet exhibits RI < 0.8, confirming complete hydraulic stability. Fig. 8Percentage of pressure-supplied (RPD ≥ 0) and unsupplied (RPD < 0) hydrants with a 90% occurrence probability for different values of the Delivery Flexibility Index (DFI): (a) Ziviyeh irrigation network (minimum reservoir level), (b) Lali irrigation network. Percentage of pressure-supplied (RPD ≥ 0) and unsupplied (RPD < 0) hydrants with a 90% occurrence probability for different values of the Delivery Flexibility Index (DFI): (a) Ziviyeh irrigation network (minimum reservoir level), (b) Lali irrigation network. Fig. 9Variations of the Reliability Index (RI) (%) for different values of the Delivery Flexibility Index (DFI): (a) Ziviyeh irrigation network (minimum reservoir level), (b) Lali irrigation network. Variations of the Reliability Index (RI) (%) for different values of the Delivery Flexibility Index (DFI): (a) Ziviyeh irrigation network (minimum reservoir level), (b) Lali irrigation network. As DFI increases, simultaneous outlet operation intensifies, leading to greater dynamic pressure fluctuations and the emergence of pressure-deficient hydrants. At DFI = 6, with a 90% probability of occurrence, 49 hydrants (approximately 9.6%) in the Ziviyeh network (minimum reservoir level) and 43 hydrants (approximately 5.7%) in the Lali network exhibit RPD < 0. Under these conditions, 344 hydrants (67.2%) in Ziviyeh and 527 hydrants (70.5%) in Lali maintained fully acceptable reliability. The observed reduction in hydraulic performance with increasing DFI is primarily associated with the nonlinear interaction between hydrant simultaneity and friction-induced head losses within secondary pipelines. As delivery flexibility increases, hydrants operate with higher instantaneous discharge rates over shorter irrigation durations, thereby increasing the probability of simultaneous operation. Under these conditions, flow velocities within downstream pipeline sections increase substantially, resulting in greater friction losses and non-uniform pressure distribution throughout the network. These effects are particularly pronounced in terminal sections characterized by limited available hydraulic head and higher sensitivity to discharge variations. Consequently, even networks with adequate average hydraulic capacity may experience localized pressure-deficit conditions under highly flexible operating conditions. Based on these results, in the Ziviyeh network up to DFI = 3, and in the Lali network up to DFI = 4.5, fewer than 2% of hydrants exhibit pressure deficits at a 90% probability level, and RI values remain within acceptable ranges. Therefore, both networks can support limited on-demand operation up to these DFI values without any structural modifications. According to FAO guidelines, optimal DFI values depend on cropping characteristics, plot size, labor availability, and network management objectives. Higher DFI values are generally recommended for small farms with limited labor and low crop water requirements, whereas lower values are more suitable for large-scale networks with extensive plots and abundant labor. Reference DFI values range from 1.5 to 2 for minimal farmer flexibility to 6–8 for maximum flexibility11. Accordingly, DFI values of 3 for the Ziviyeh network and 4.5 for the Lali network are considered acceptable and provide a desirable level of delivery flexibility. Because the estimated maximum network discharges and friction losses may differ slightly from actual field conditions, final DFI values should be refined through on-farm operational assessments. Although networks are designed based on peak-period crop water requirements, maximum DFI values occur during the peak demand period (DFI ≈ 12 for Ziviyeh and 11.63 for Lali). In other months, reduced crop water requirements lower the nominal discharge, thereby increasing the maximum attainable DFI. Monthly variations in maximum and optimal DFI values, along with corresponding continuous specific discharges and maximum probable network discharges, are shown in Fig. 10. Fig. 10Maximum and optimal values of the Delivery Flexibility Index (DFI) and their corresponding maximum probable discharge on a monthly basis under limited on-demand conditions: (a) Ziviyeh irrigation network (minimum reservoir level), (b) Lali irrigation network. Maximum and optimal values of the Delivery Flexibility Index (DFI) and their corresponding maximum probable discharge on a monthly basis under limited on-demand conditions: (a) Ziviyeh irrigation network (minimum reservoir level), (b) Lali irrigation network. Across the irrigation season, neither network can support full on-demand operation in any month, as allowable DFI values remain below the maximum theoretical values. The most restrictive conditions occur during the peak demand period, when higher water requirements lead to longer outlet operation times, increased simultaneity, and higher instantaneous network discharge, thereby reducing allowable outlet discharge. Implementation of limited on-demand operation requires volumetric water delivery and, therefore, flow measurement devices (flow meters or mechanical water meters) at hydrants11. Such equipment already exists in rotational systems. Additionally, hydrants must be equipped with discharge control valves to enforce maximum allowable discharges. Advances in mechanical and electronic valve technologies now enable precise discharge regulation across a wide operating range, facilitating practical implementation. When discharge control valves are used, the maximum allowable outlet discharge defined based on DFI = 3 and 4.5 during the peak period remains constant throughout the irrigation season. Under the assumption that farmers abstract water at this maximum rate, the resulting seasonal DFI values are shown in Fig. 11. Except for May (9.3 and 8.2 for Ziviyeh and Lali, respectively) and October (11.6 and 11.9, respectively), DFI values remain below the recommended upper limit of 811, confirming the operational feasibility of the proposed approach. Fig. 11Available Delivery Flexibility Index (DFI) values on a monthly basis under limited on-demand conditions with flow control valves. Available Delivery Flexibility Index (DFI) values on a monthly basis under limited on-demand conditions with flow control valves. Although implementation of limited on-demand operation requires installation of discharge-control valves at hydrants, implementation of fully on-demand operation, in addition to such control devices, generally requires large-scale structural rehabilitation measures such as pipe replacement, network resizing, or pumping-station reinforcement. Moreover, localized operational equipment can generally be implemented incrementally according to network priorities and available financial resources, whereas structural rehabilitation commonly requires complete network shutdown, extensive excavation, and substantial capital investment. Therefore, from both economic and operational perspectives, DFI-based operational control may provide a more feasible transitional modernization strategy for irrigation systems operating under financial and institutional constraints. Implications for irrigation modernization and network management The findings of this study have important implications for irrigation modernization strategies in water-scarce regions. In many existing irrigation schemes, direct transition from rotational to fully on-demand operation is economically impractical because of the extensive infrastructure rehabilitation required. Under such conditions, limited on-demand operation may provide an intermediate operational strategy capable of improving operational flexibility for water users while avoiding costly structural modifications. The proposed DFI-based operational framework can support irrigation authorities in identifying acceptable ranges of operational flexibility according to the hydraulic capacity of existing networks. By regulating the maximum allowable hydrant discharge, irrigation system operators can reduce the probability of pressure-deficit conditions while maintaining adequate responsiveness to crop water requirements. This approach may be particularly beneficial in regions where financial resources, energy availability, and institutional capacity limit the feasibility of large-scale modernization projects. Furthermore, the methodology presented in this study can be extended to other pressurized irrigation systems operating under rotational or semi-rotational delivery conditions, providing a practical framework for evaluating transition strategies toward more adaptive irrigation operation. Although increasing delivery flexibility can improve irrigation efficiency and operational adaptability, pressurized on-demand operation may also increase energy consumption, particularly in pump-supplied irrigation systems such as the Lali network. Under highly flexible operating conditions, maintaining acceptable pressure conditions throughout the network may require greater pumping energy and more advanced pressure-regulation strategies. At the same time, the present results indicate that several hydrants operate under excess pressure-head conditions (RPD > 0), suggesting the presence of recoverable hydraulic energy within portions of the network. Recent studies have demonstrated that excess pressure in water distribution and irrigation systems can potentially be utilized through energy-recovery technologies such as pumps-as-turbines (PATs), micro turbines, and pressure-regulating energy-recovery devices26–29. These approaches may simultaneously improve pressure regulation and offset part of the operational energy demand associated with pressurized irrigation modernization. Therefore, integration of DFI-based hydraulic management with energy-recovery and smart pressure-regulation technologies may represent a promising direction for sustainable modernization of irrigation networks under water- and energy-scarcity conditions. Study limitations and future perspectives Although the present study provides useful insights into the feasibility of limited on-demand irrigation operation, several limitations should be acknowledged. First, the hydraulic analyses were based on probabilistic operational scenarios generated within the COPAM framework and therefore may not fully represent real-time water-user operation and temporal variability in water abstraction patterns. Second, transient hydraulic effects, including pressure surges and rapid valve operations, were not considered in the present analysis. Third, the study primarily focused on hydraulic performance indicators and did not explicitly evaluate economic, energy-consumption, environmental, or socio-institutional aspects associated with implementation of limited on-demand operation. Future studies should therefore integrate field operational data, economic feasibility assessment, energy optimization, and evaluation of water-user operational behavior into DFI-based irrigation management frameworks. In addition, integration of smart flow-control devices, real-time monitoring systems, and adaptive pressure-regulation technologies may further improve the operational applicability of limited on-demand irrigation systems under water-scarcity conditions. Conclusion Rotational irrigation systems are inherently limited in their ability to respond to the spatial and temporal variability of crop water requirements, often resulting in reduced irrigation efficiency, limited operational flexibility, and inadequate adaptability to variable irrigation demand conditions. Although fully on-demand operation can substantially improve irrigation system performance, implementation of such operation in existing pressurized irrigation networks frequently requires extensive and costly infrastructure rehabilitation. The present study demonstrated that limited on-demand operation, combined with Delivery Flexibility Index (DFI)-based operational management, can provide a practical transitional operational strategy for improving delivery flexibility without major structural modification of existing irrigation networks. Hydraulic simulations performed using the COPAM framework showed that neither of the investigated networks could reliably support full on-demand operation under existing conditions because of pressure instability and hydraulic limitations within secondary pipelines. However, under limited on-demand operation, both networks maintained acceptable hydraulic reliability within specific DFI ranges (up to 3 for the Ziviyeh network and 4.5 for the Lali network). These findings indicate that operational regulation of hydrant discharge can significantly improve delivery flexibility while avoiding the substantial financial and operational costs associated with pipe replacement, pumping reinforcement, and large-scale network rehabilitation. From a practical perspective, implementation of the proposed approach requires volumetric water-delivery systems, flow-measurement devices, and discharge-control valves at hydrants to enforce allowable discharge limits. Such operational requirements are already available in many pressurized irrigation systems, making the proposed strategy operationally applicable and economically feasible for irrigation modernization programs. The proposed methodology may assist irrigation authorities and water-management agencies in identifying suitable operational flexibility ranges according to the hydraulic capacity of existing networks. In addition, the framework developed in this study can support operational planning and decision-making for transitional irrigation modernization in water-scarce regions where financial, institutional, and energy limitations restrict complete infrastructure redevelopment. Nevertheless, several limitations should be acknowledged. The hydraulic analyses were conducted based on simulated operational scenarios generated within the COPAM environment and did not incorporate real-time water-use behavior, temporal variability in actual irrigation scheduling, or transient hydraulic effects. In addition, the analysis primarily focused on hydraulic performance indicators and did not explicitly evaluate economic, energy-consumption, or socio-institutional aspects associated with implementation of limited on-demand operation. Future research should therefore focus on integration of real operational data, economic feasibility assessment, water-user operational response, and energy optimization into DFI-based irrigation management frameworks. Further investigation of real-time pressure regulation and smart hydraulic control systems may also improve the operational applicability of limited on-demand operation in large-scale irrigation modernization projects.