Furrow-irrigated chufa crops in Valencia ( Spain ) . II : Performance analysis and optimization

Chufa (Cyperus esculentus L. var. sativus Boeck.) is a traditional crop in the Mediterranean region of Spain, where it is only furrow irrigated. This article analyzes the irrigation performance for this crop, conducting field studies over three consecutive seasons in Valencia (Spain). Irrigation schedule was based on the volumetric soil water content, which was measured with capacitance sensors. Infiltrability was measured with blocked-furrow infiltrometers. An area velocity flow module measured the water flow, the cross-sectional geometry of furrows was determined using furrow profilometers, and times for advance and recession were recorded. WinSRFR software was used to analyze every irrigation event, determining the application efficiency (AE) and distribution uniformity of the minimum (DUmin), and to optimize the combination of furrow inflow (q) and cut-off time (Tco). Average values obtained for AE were 30.1%, 25.6%, and 26.7% in 2007, 2008, and 2009, respectively, and the corresponding DUmin values were 0.54, 0.61, and 0.67. Optimized results showed that it is possible to reach AE and DUmin values up to 87% and 0.86, respectively. However, understanding the q-Tco relationship that maximizes both AE and DUmin is more important than knowing the specific values. A function that related q and Tco was obtained for the typical plot dimensions, and this was validated in 2011. Therefore, this function can be used in most of the plots in the cultivation area. Additional key words: application efficiency; cut-off time; distribution uniformity; furrow inflow rate; irrigation management; vegetable crop. * Corresponding author: bpascual@prv.upv.es Received: 31-07-12. Accepted: 09-01-13. Abbreviations used: AE (application efficiencies); Dapp (average depth of the water applied to the field); Dav (average depth of the water infiltrated to the field); Dmin (minimum depth of the water applied to the field); Dreq (average depth of the water required to fill the root zone); DOY (day of the year); DU (distribution uniformity); DUmin (distribution uniformity of the minimum); EY (experimental year); FC (field capacity); n (roughness coefficient); Q (water flow); q (furrow inflow); RP (refill point); Tco (cut-off time); VSWC (volumetric soil water content. Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA) Spanish Journal of Agricultural Research 2013 11(1), 268-278 Available online at www.inia.es/sjar ISSN: 1695-971-X http://dx.doi.org/10.5424/sjar/2013111-3384 eISSN: 2171-9292


Introduction
In a companion paper (Pascual-Seva et al., 2013, this issue) the interest in chufa (Cyperus esculentus L. var.sativus Boeck., also known as tigernut) cultivation has been reported, both in Spain and in the world, as well as the characterization of the irrigation on the Huerta of Valencia irrigation district, and the productive response of chufa crop to two irrigation strategies was studied.
The increase of irrigation water use efficiency can be achieved by both the increase of crop yield (that has been studied in the companion paper) and the reduction of gross water use through improvements in irrigation efficiency (Playán & Mateos, 2006).In turn, according to Playán et al. (2000) and Neira et al. (2005), an analysis of current irrigation performance must precede any attempt to improve irrigation efficiency, since, detecting specific problems that could affect the water management, enables to recommend solutions to achieve higher efficiencies.
Irrigation system evaluation by field tests in normal conditions determines the parameters that are involved in water application, such as efficiency (average measure of water losses) and uniformity (water distribution in different parts of the plot).This information can help in making decisions to improve the irrigation system both from the economical and design point of view (Merriam et al., 1980).Irrigation efficiency and distribution efficiency terms can be ambiguous depending on the considered inputs and outputs (Burt et al., 1997).The present study used the application efficiency (AE) term as described by Lord & Ayars (2007) as well as the distribution uniformity of the minimum (DU min ) term as described by Strelkoff & Clemmens (2007).
The furrow irrigation design has been traditionally based on intuition and empirical data taken from successful projects rather than on theory.The prediction of surface irrigation behavior is complicated due to many analytical problems.However, nowadays, it is possible to design and provide operation recommendations based on simulations (Strelkoff & Clemmens, 2007).
To improve irrigation performance, mathematical models of the surface irrigation processes have been developed during the last decades.These models consider several variables, including plot dimensions, slope, hydraulic roughness, furrow discharges, and irrigation time.The interaction among these variables determines the advance times, recession times and infiltrated water depth as well as the corresponding efficiencies and uniformities (Jurriens et al., 2001).In the 1970s and 1980s, models included many diagrams and formulas, which made them difficult to be applied by non-specialized people.However, this all changed with advances in computer programming and the availability of personal computers.Thus, irrigation engineers started to make computer programs for the mathematical models so that more users could apply these models without practical problems (Jurriens et al., 2001).
Diverse research groups of different research institutes and universities have developed various computer programs for surface irrigation analysis.Two of the remarkable software are WinSRFR (ALARC, 2009;Bautista et al., 2009a), which was developed by the US Arid Land Agricultural Research Center, and SIRMOD (Walker, 2003), developed by Utah State University.Both are one-dimensional models and they are similar in many aspects; their accuracy has been tested in a wide range of conditions (McClymont et al., 1999).
Several studies have shown that infiltration from furrows is not a one-dimensional process but a two-dimensional, existing other models, as TRIDISUL (Tabuada et al., 1995), HYDRUS-2D (Simunek et al., 1999) and FURINF (Schmitz, 1993), that could guarantee higher accuracy.But, due to their complexity, as in other complex model systems, their main use tends to be for comprehensive researches (Wöhling & Schmitz, 2007).In the present study a small water depth is required and given that it is difficult to obtain high efficiencies in sandy soils, the precision of one-dimensional models (with lower complexity than two-dimensional models) is considered enough to describe the irrigation performance parameters.The accuracy obtained with WinSRFR has been considered suitable for the present analysis.
The aim of this study was to determine the actual volumes of water used in chufa cultivation and the corresponding application efficiency (AE) and distribution uniformity of the minimum (DU min ) values.An additional aim of this study was to propose irrigation parameters [furrow inflow (q) and cut-off time (T co )] to considerably improve both AE and DU min values.

Field experiments
Field studies were conducted over three consecutive years (2007, 2008, and 2009) in two adjacent plots of a commercial farm near the Valencia Polytechnic University campus in Spain (39°29' N and 0°20' W).This farm is representative of the plots in the region.One plot was used in 2007 and 2009, and the other plot was used in 2008 to avoid soil exhaustion problems related to crop repetition.According to the Papadakis' agroclimatic classification system (Elías & Ruiz, 1977), the climate is subtropical Mediterranean with hot, dry summers and an average annual rainfall of approximately 450 mm, which is irregularly distributed throughout the year with 40% of the rainfall occurring in autumn.
The soil textural type is sand in the 2007 and 2009 plot and loamy sand in the 2008 plot.The soils are deep, and they are classified as anthropic torrifluvents according to the USDA's Soil Taxonomy (Soil Survey Staff, 2010).The analyses performed in this study indicated that the soils had slightly or moderately alkaline pH levels and that they were fertile (moderate organic matter content and high available phosphorous and potassium concentrations).
In this area, water comes from the Rascanya Canal, which flows from the Turia River (EC = 1.38 dS m -1 ; SAR (adjusted) = 1.21; pH = 7.2).The irrigation water has no restrictions in terms of salinity for non-sensitive crops, as chufa, or permeability (Ayers & Westcot, 1994).
The standard cultivation practices described by Pascual et al. (1997) were followed during the crop period.Planting was performed on May 8 th in 2007 (128 th day of the year; DOY), April 11 th in 2008 (101 th Performance analysis and optimization of furrow-irrigated chufa crops DOY), and April 27 th in 2009 (117 th DOY).Tubers (120 kg tubers ha -1 ) were planted in ridges, which were 0.20 m high, and the ridge top center spacing was 0.60 m.The furrow length was 82.5 m in 2007 and 2009, and was 66 m in 2008.The number of furrows simultaneously irrigated was 26, 24, and 40 in 2007, 2008, and 2009, respectively.The average furrow slope was approximately 0.3%.This study was designed to identify improvements that growers could easily apply without large investments.Therefore, precision land leveling was not considered in this study although its importance in achieving high performances and water savings is known (Pereira et al., 2002;Horst et al., 2005).In this sense, laser technology is used to achieve a leveled field surface after chufa tuber harvesting.The land leveling in chufa cultivation depends on the crop management of plants grown between chufa in the crop rotation.
Three ECH 2 O EC-5 capacitance sensors with ECH 2 O Utility software (Decagon Devices Inc., Pullman, WA, USA) were placed at a depth of 0.10 m, and they were connected to an Em50 data-logger (Decagon Devices Inc., Pullman, WA, USA) to monitor the volumetric soil water content (VSWC).The irrigation was scheduled so that each event began (refill point; RP) when the VSWC reached approximately 60% of the field capacity (FC; Pascual-Seva et al., 2013, this issue).Variations of VSWC were used to determine the in situ values corresponding to FC, which is defined as the amount of water held in the soil after excess water has drained away and the rate of the downward movement of water has materially decreased (Veihmeyer & Hendrickson, 1931), and therefore reaching a plateau in VSWC variations.
In each event, irrigation management was representative of the practices in the region.The end of the chufa irrigation period is considered as the time when the straw is burned, which takes place before tuber harvesting.In this experiment, the dates of straw burning were November 6 th in 2007, November 7 th in 2008, and October 26 th in 2009.

Field evaluation procedures
This study determined AE (Lord & Ayars, 2007) and DU min (Strelkoff & Clemmens, 2007).WinSRFR 3.1 (ALARC, 2009;Bautista et al., 2009a) is the software package used to determine these parameters and also to optimize them.For this optimization, the program compared DU min values, so this value was the uniformity parameter presented for the entire study.The AE and DU min values are described by the following equations: where D req is the average depth (mm) required to fill the root zone; D app is the average water depth (mm) applied to the field; and D min is the minimum depth (mm) of water infiltrated in the field.In this study, furrows were blocked on the downstream ends.Thus, there was no runoff and, consequently, D app was equal to the average infiltrated water depth (D av ).
The Merriam & Keller's (1978) post-irrigation volume balance method was used by WinSRFR 3.1.This method estimates the field-averaged infiltration function from the field-measured geometry, inflow, advance times and recession times, thereby, requiring the determination of different measurements, including furrow discharges, furrow cross-sections, infiltration, advance times, recession times, hydraulic roughness, and furrow slopes (Horst et al., 2005).
The water flow (Q) was continuously measured by a flow meter (ISCO 2150 area velocity flow module; Teledyne ISCO Inc., Lincoln, NE, USA) placed on the water entrance to the experimental plots.Discharge data were measured and stored at 15 s intervals, which allowed calculating the average, minimum and maximum inflow rates as well as the variation of inflow rates.The inflow rate was considered uniformly distributed along the different furrows simultaneously irrigated, so q was calculated as the average Q divided by the number of irrigated furrows.
The furrow cross-sectional geometry was determined at a distance half of the length of the furrow before each irrigation event using furrow profilometers (Walker & Skogerboe, 1987).
Among the different options provided by the program for calculating the infiltrated depth, the Kostiakov formula was used in this study.The corresponding a exponent was determined through in situ infiltrability trials using the blocked furrow infiltrometer method, which was developed by Bondurant (1957).For construction of the blocked furrow infiltrometer, a furrow segment (length of 1 m) was isolated by two metal plates (0.6 m × 0.5 m; thickness of 2 mm), which were intro- •100, and duced into the soil to a depth of 0.30 m.A constant water level (approximately 50 mm; similar to the level during the irrigation events) was maintained in this furrow segment by a float valve arrangement.To simulate natural conditions, a section of a furrow was used as a buffer to both sides and extremes.The water level in the supply reservoir was measured every 1 min for at least 3 h, and the infiltration rate was computed from these data points.The infiltration rate may be affected by soil changes, such as surface sealing and crusting (Smith & Warrick, 2007).Therefore, the infiltration rate was determined before each irrigation event in furrows (of the same plot) that were not monitored during the irrigation events to avoid alterations in the furrow bed.
Advance and recession times were measured at 16m and 20-m intervals along the monitored furrows in the plots with length of 66 and 82.5 m, respectively.Recession times were considered at the times when water completely infiltrated the soil at the observed points.However, when irregularities of the furrow bed caused the water to pool for a long time, the recession time was considered when water disappeared from the furrow bed in the areas close to the measurement point (Horst et al., 2005).
Manning's roughness coeff icient, n (s m -1/3 ), for each irrigation event was based on the recommended values by the NRCS (USDA-SCS, 1991b), and this coefficient was fine-tuned considering the furrow bed roughness and the presence of lodged plants.Simulations with WinSRFR 3.1 were conducted to find the n values that minimized the difference between observed and simulated advance and recession times, which were used for analyzing the irrigation performance (Bautista et al., 2009a).
Infiltrability (a exponent of the Kostiakov formula), and advance and recession times were determined in triplicate, and their average values, for each irrigation event, were entered into the WinSRFR 3.1 software program.

Optimized irrigation management
Water savings can be achieved by improving irrigation performances, i.e., increasing the AE and DU parameters, which depend upon many factors, such as q, T co , n, infiltration characteristics of the soil, cross-sectional characteristics of the furrow and furrow slope (Horst et al., 2005).However, because this study focused on the field assessment of water saving potential by adop-ting easily accessible technologies related to furrow irrigation, only factors that can be easily adapted by growers were considered as follows: q and T co .
To optimize these factors, the Operation Analysis World in the WinSRFR 3.1 program was used.This program supplies performance contours that depict the variation of AE and DU min (among others) as a function of q and T co .These contour plots are generated by interpolation from simulation results computed at discrete grid points on a rectangular solution region, and the limits are defined by the user (Bautista et al., 2009a).These plots also present a dotted line representing the solutions that satisfy the irrigation requirement everywhere (i.e., minimum depth equals the required depth).The optimal q-T co combination, which maximizes both AE and DU min , does not fall exactly on this line, but it is near to the line and can be easily found by trial and error (Bautista et al., 2009b).
In this study, the optimal q-T co combination was determined for every irrigation event until the first of September.After this date, lodged plants disrupted the water flow advance because the plants provoked flooding and made it impossible to adequately control the irrigation.

Validation
The q-T co relationship has to be validated before it can be used by the growers.Therefore, this relationship was tested in the same plot that was used in 2007 and 2009 (which length was 82.5 m) in the first three irrigation events in 2011.This agrees with Walker (1989) who suggested evaluating q and T co values for the first irrigation event following planting, when roughness and intake are maxima, and for the third or fourth irrigation event, when these conditions have been reduced by previous irrigations.For the test in 2011, planting was performed on April 5 th .As in 2007, 2008, and 2009, standard cultivation practices were followed during the crop period in 2011.Irrigation management was carried out under normal grower control.The RP, and therefore, the D req value for each irrigation event were similar to the value used in the analyzed experiments.In 2011, the irrigation events were on May 26 th , June 16 th , and June 27 th .
Before the irrigation was started and once Q was roughly stabilized, q was calculated and it was considered to determine T co .Irrigation performance was tested similarly to the previous experiments.

Furrow characteristics
Table 1 presents the cumulative infiltration curves [a and k values and corresponding coefficients of determination (R 2 )], obtained with the blocked furrow infiltrometer and fitted to the Kostiakov equations.For every curve, the R 2 value was large (> 0.83), and the significance level was p ≤ 0.01.All determined values of a ranged from 0.40 to 0.69, which agreed with the values reported in the literature for furrows (ranging from 0.19 to 0.69) (Bautista & Wallender, 1993;Clemmens & Bautista, 2009).
Simulations were conducted to test the effect of n values on the advance and recession times for each irrigation event (Bautista et al., 2009b) resulting in that 0.08 minimizes the difference between observed and simulated advance and recession curves for the first event and 0.04 for the following event.The n values were increased according to the furrow bed conditions (lodged plants) up to 0.12 in 2007 and 2008, and 0.15 in 2009.These values were in the range considered by the program (ALARC, 2009), which were in accordance to those recommended by the NRCS (USDA-SCS, 1991a, b).
Furrow cross-section data obtained by means of profilometers were introduced into the program which adjusted them to a trapezoid shape.

Actual irrigation performances
The FC values were 0.27, 0.28, and 0.27 m 3 m -3 for 2007, 2008, and 2009, respectively, with corresponding RP values of 0.16, 0.17, and 0.16 m 3 m -3 .Considering that the root zone depth was estimated as the maximum development of root masses (20 cm based on phenological measurements; Pascual-Seva, 2011), D req was considered to be 22 mm for all irrigation events.There were 15, 13, and 12 irrigation events in 2007, 2008, and 2009, respectively, with a total of 40 irrigation events for the experiment.Each irrigation D av value was determined by integrating q and T co .The seasonal D av value was 1201, 1320, and 1140 mm in 2007, 2008, and 2009, respectively, and the seasonal rainfall water input was 498, 438, and 320 mm in 2007, 2008, and 2009, respectively.Table 2 shows the Q values (average, minimum, and maximum) and the corresponding coefficient of variation (CV) in each irrigation event of 2007, 2008 and 2009.Among and within the different irrigation events, Q varied considerably.In each irrigation event, the CV ranged from 3.03 (August 14 th , 2008; Q = 64.2L s -1 ) to 26.43% (May 18 th , 2009; Q = 66.8 L s -1 ).The average seasonal Q value of 2007, 2008, and 2009  flow rate in the canal, as the consequence of the delivery policy and to the greater number of furrows to simultaneously irrigate.
All the irrigation events could be satisfactorily modeled with the WinSRFR 3.1 program given that the mass balance error was lower than 1% and the rootmean-square errors of the advance and recession times were lower than 30% of the corresponding values in all cases (data not presented), thresholds stated by Mahesshawari & McMahon (1993), therefore the analysis can be considered valid and the AE and DU min values could be determined.Table 3 shows the q, T co , D av , AE, and DU min values in each irrigation event as obtained by the WinSRFR 3.1 software program.
The average q values in 2007, 2008, and 2009 were 1.6, 1.9, and 1.9 L s -1 , respectively.The minimum and maximum average q registered values were 1.1 and 2.5 L s -1 , respectively, which ranged among the limits considered for gradient furrows (0.4-2.5 L s -1 ) by Hart et al. (1980) and among the limits of inflow (0.9-2.4 L s -1 ) used by Oyonarte et al. (2002).The erosion was negligible when comparing the cross-sections before the consecutive irrigation events.In this sense, the maximum flow velocity (that depends on a large variety of factors, among others soil type, furrow geometry, slope, roughness, and intake) was approximately 6 m min -1 , which was lower than the maximum flow velo-city value (13-15 m min -1 ) suggested in sandy soils to avoid erosion (Walker, 1989).
The T co value of the first irrigation event of each season was greater than those for the following events due to the small flow velocity resulting from the low soil compaction and high roughness (Walker, 1989) for tillage operations.In general, the T co value decreased with increasing q values, and the T co value increased when plants were lodged at the end of the cultivation cycle.It should be noted that, to avoid a cut-off of water due to canal maintenance, the fifth irrigation in 2008 was moved forward to July 10 th , presenting a small T co value because the VSWC at the beginning of this event was higher than usual, thus, resulting in a faster flow and, consequently, a smaller T co value.
In the first event, the D av value was (as T co ) considerably greater than the values in the remaining events with a value of 146, 178, and 142% greater than the average values of the other events in 2007, 2008, and 2009, respectively.For this reason the average values in Table 3 do not include the corresponding values of the first irrigation events.
The average AE values (Table 3) were 30.1, 25.6, and 26.7 for 2007, 2008, and 2009, respectively (20.6, 13.5, and 17.3% for the first irrigation event of each season).This variability was not infrequent, with larger differences among AE values being obtained in diffe- rent years in the same plots (Kruse & Heermann, 1977;Khatri & Smith, 2006).These values were considered low when compared to potential values previously reported (Bos, 1980;Elliot & Walker, 1982;Solomon et al., 2007).Nevertheless, these AE values were similar or only slightly lower than those obtained under normal grower control as others have reported (Smith et al., 2005).These small values were partially justified by the small D req value due, in turn, to the shallow chufa plant root system (20 cm).Keller (1965) showed the effect of stored water depth on field irrigation efficiency for surface irrigation methods and presented (with actual field practices for downslope furrows in studies carried out in Idaho by C. H. Pair) efficiency values close to 10% for a stored water depth of 25 mm.The efficiency value reported by Keller (1965) was less than the values obtained in the present study.
When considering the average seasonal data, the D av value increased and the AE value decreased with increasing q values.However, these relationships were not significant (p ≤ 0.05).
The average DU min values of the 2007, 2008, and 2009 irrigation events were 0.54, 0.61, and 0.67, respectively.The minimum and maximum DU min registered values were 0.46 (July 10 th irrigation event in 2008 coincided with the smallest D av value) and 0.90 (May 18 th irrigation event in 2009 coincided with the largest D av value), respectively.The DU min values obtained in the first events of each season were high (0.74, 0.81, and 0.90) as a consequence of the corresponding high D av values (114, 164, and 128 mm).

Optimized irrigation management
The optimization results are presented in Table 4.All of the optimized q values ranged from 0.60 to 2.40 L s -1 , which were between the limits for gradient furrows (0.40 to 2.50 L s -1 ) as considered by Hart et al. (1980), except for the q value (2.75 L s -1 ) corresponding to the April 28th irrigation event in 2008.In the first irrigation event of each season, the optimized q values were higher than the actual values used in the irrigation events, which was contrary to what occurred in the remaining events.
The AE and DU min values followed similar trends [AE (%) being approximately 100 times DU min (-)].Therefore, all of the reported statements related to AE in this study are also applicable to DU min .In the first irrigation event of each season, the maximized AE values (70.5, 66.4, and 67.2% for 2007, 2008, and 2009, respectively) were similar to those obtained by Camacho et al. (1997) and by Mateos & Oyonarte (2005), which were considered reasonable and satisfactory results.Average values (Av) and standard error (SE) correspond to all irrigation events, except for the first event.
In the remaining events, better results were reported (on average 78.4, 81.9, and 77.5% for 2007, 2008, and 2009, respectively), which were considerably higher than those (57.4% in a sloping furrow for a D req of 80 mm, and 76.1% in a sloping furrow for a 50% cutback flow and the same D req ) reported by Clemmens et al. (2007).
Given that each irrigation event is unrepeatable, more important than knowing the specific values is to know the relation between the optimized q and T co values.
For each plot these relationships (without the data of the first irrigation after planting in each year) are represented in Fig. 1 and fitted with a power law function, obtaining the following equations that related optimized T co and q: Plot with length of 66 m (2008): T co = 17.603 • q -0.920 (R 2 = 0.9045; p ≤ 0.01) Plot with length of 82.5 m (2007 and 2009): T co = 22.868 • q -0.920 (R 2 = 0.9880; p ≤ 0.01).
Given that the curves were similar, the following single equation was also considered (a large R 2 value should permit its generalization and its utilization in a typical plot length of 60-90 m in the cultivation area): T co = 21.542• q -0.998 (R 2 = 0.8996; p ≤ 0.01).
These equations permit T co to be selected according to the q used in each event.In these cases, high AE and DU min values would be obtained reaching values as large as those shown in Table 4.With regard to the first irrigation events, q would be selected on the basis of the corresponding results shown in Table 4.

Validation
Table 5 shows the measured Q, the determined q and T co , and the performance analysis parameters (AE and DU min ) of the three irrigation events carried out in the 2011 validation experiment.In the first irrigation event, the q value was equal to 1.64 L s -1 , which was considerably lower than the optimized values for the first irrigation events presented in Table 4 (given that Q was the maximum available flow in the canal).Therefore, the T co value was proportionally increased compared to the values shown in Average values (Av) and standard error (SE) correspond to all irrigation events until the first of September, except for the first event.
Table 5. Measured flow (Q; L s -1 ), determined furrow inflow rate (q; L s -1 ) and cut-off time (Tco; min), and performance analysis parameters [application efficiency (AE; %), and distribution uniformity of the minimum (DU min ; -)] of the three irrigation events carried out in the 2011 validation experiment In the remaining events T co was determined using the previously obtained general equation.Performance analysis resulted in large AE and DU min values (AE = 75.5% and DU min = 0.87 in the second event; AE = 73.8% and DU min = 0.78 in the third event), which were similar to those obtained with the optimization in the same plot in 2007 (AE values ranged from 73.2 to 82.5% and DU min values ranged from 0.73 to 0.83) and in 2009 (AE values ranged from 70.6 to 82.9% and DU min values ranged from 0.70 to 0.83).These results validated the q-T co relationship.Therefore, this relationship can be extended to most of the plots on the cultivation area as well as to other crops cultivated as part of the crop rotation with shallow root masses.This q-T co relationship can improve the application efficiency in these irrigation systems, thus, leading to water savings.
As final conclusions, with the usual irrigation management of chufa crops, AE values are low mostly due to the shallow chufa plant root system.Low AE values occurred especially in the first irrigation event of each season due to both low soil compaction and roughness of furrow beds as a consequence of the tillage operations.Using the WinSRFR 3.1 program, the q-T co combination was optimized.By considering the optimized q and T co values, a relationship between these parameters was obtained.This relationship was validated in 2011, which allows its use in most of the plots and crops in the cultivation area.By applying this relationship, AE and DU min values would increase.Further research is needed to evaluate the water saving potential and productive response by applying the optimized q-T co combination as an irrigation management strategy.N. Pascual-Seva et al. / Span J Agric Res (2013) 274N.Pascual-Seva et al. / Span J Agric Res (2013)  11(1), 268-278

Table 1 .
The a and k parameters of the Kostiakov equation and corresponding R 2 values from dates obtained with the blocked furrow infiltrometers before each irrigation event In all events, p ≤ 0.01.

Table 2 .
Performance analysis and optimization of furrow-irrigated chufa crops273 Date, flow rate (average, Q; minimum, Q min ; and maximum, Q max , L s -1 ), and coefficient of variation (CV; %) in each irrigation event Average values (Av) and standard error (SE) correspond to all irrigation events.

Table 3 .
Average furrow inflow rate (q; L s -1 ), cut-off time (T co ; min), average depth of water infiltrated (D av ; mm), application efficiency (AE; %), and distribution uniformity of the minimum (DU min ; -) in each irrigation event

Table 4 .
It is advisable to usePerformance analysis and optimization of furrow-irrigated chufa crops 275

Table 4 .
Performances of irrigations [application efficiency (AE; %) and distribution uniformity of the minimum (DU min ; -)] carried out until the first of September when the furrow inflow (q; L s -1 ) and the cut-off times ( Tc o; min) were optimized