E ﬀ ects of Biochar on Irrigation Management and Water Use E ﬃ ciency for Three Di ﬀ erent Crops in a Desert Sandy Soil

: This paper aimed at investigating if the application of biochar ( BC ) to desert sand ( DS ) from the United Arab Emirates (UAE), characterized by a very poor soil-water retention ( SWR ) and by a very low value of the maximum water available for crops ( AW max ), could positively a ﬀ ect soil water balance, by reducing the irrigation needs ( V IRR ) and improving the irrigation water use e ﬃ ciency ( IWUE ) and the water use e ﬃ ciency ( WUE ). The analysis was performed for three crops, i.e., wheat ( Triticum aestivum ), sorghum ( Sorghum vulgare ) and tomato ( Lycopersicon esculentum ). BC was applied to the DS at di ﬀ erent fractions, f BC ( f BC = 0, 0.091, 0.23 and 0.33). Drip irrigation was adopted as a highly e ﬃ cient water saving method, which is particularly relevant in arid, water-scarce countries. Soil water balance and irrigation scheduling were simulated by application of the AQUACROP model, using as input the SWR measured without and with BC addition. The e ﬀ ect of BC was investigated under either a no-water stress ( NWS ) condition for the crops or deﬁcit irrigation ( DI ). The results showed that the application of BC made it possible to reduce the predicted V IRR and to increase the IWUE under the NWS scenario, especially for wheat and sorghum, with less evident beneﬁts for tomato. When a deﬁcit irrigation ( DI ) was considered, even at the lowest considered f BC (0.091), BC counterbalanced the lower V IRR provided under DI , thus mitigating the yield reduction due to water stress, and improved the WUE . The inﬂuence of BC was more pronounced in wheat and tomato than in sorghum. The results evidenced that the application of BC could be a potential strategy for saving irrigation water and / or reducing the e ﬀ ects of drought stress in desert sand. This means that biochar could be used a management option to promote local production and reduce the dependency on food import, not only in the UAE, but also in other countries with extremely arid climatic conditions and large extensions of sandy soils similar to the considered DS . Soil data —number of layers (maximum 5) and soil layer depth, d (m), and for each layer volumetric water content at saturation, θ s (m 3 m − 3 ), ﬁeld capacity, θ fc , and wilting point, θ wp . Irrigation —irrigation system; scheduling parameters (time and depth criteria), in case of irrigation scheduled by the model, or time and volume of each water supply in case of user-imposed irrigation depth. The irrigation system can be selected as “sprinkler”, “surface” or “drip”. In the latter case, an adjustment coe ﬃ cient (%) for partial soil wetting can be used to take into account the actual irrigated area. irrigation system; scheduling (time 0.23 and Z DS + BC = 30 cm, the reductions in the V IRR values were equal to 10.7%, 10.2% and 6.9% for wheat, sorghum and tomato, respectively, with the minimum e ﬀ ect on tomato and comparable e ﬀ ects on wheat and sorghum. These results indicated that, under the climatic conditions explored in this investigation, the addition of BC to the DS made it possible to save water, showing the positive role of BC in improving irrigation management for the considered DS and for the three crops.


Introduction
Fresh water is an indispensable natural resource, which plays a vital role in the development of any country. Water scarcity originates not only from the physical constraints of available fresh water, but also from competition between different uses and from inefficient use and poor management (FAO, 2008). As a result, water allocation has become one of the most vexing problems faced by policy makers.
Agriculture accounts for roughly 70% of total freshwater withdrawals and for over 90% in the majority of the least developed countries [1]. Without improved efficiency strategies, agricultural water consumption is expected to increase by about 20% by 2050 [2]. As a consequence, the world could face a 40% global water deficit by 2030 under a business-as-usual scenario [3]. Water withdrawal for agricultural purposes accounts for about 75% of all usages in developing countries, and irrigation and had positive effects on soil physical properties such as porosity (θ s ), soil water retention and the maximum water available for crops, AW max [44,45]. Several studies also showed that the application of BC was effective in reducing salinity stress by improving the soil's physico-chemical and biological properties directly related to Na removal, Na adsorption ratio and electrical conductivity [46]. Sandy soils lacking sufficient organic matter to sustain microbial activity are widespread in various arid and semi-arid regions of the world, including India, Tunisia and Saudi Arabia, as well as the United Arab Emirates (UAE) [47][48][49]. These soils have a low capacity of water retention [50] and low values of the maximum water available for crops, AW max , i.e., the difference between field capacity (θ fc ) and wilting point (θ wp ). Without any amendment, the agricultural production capacity of these soils is weak, and therefore a considerable amount of irrigation water is necessary to meet the crop water requirements.
Baiamonte et al. [51] found that the application of BC to desert sand (DS) from the UAE (Al-Foah area, border with Oman) increased soil porosity and enhanced the formation of textural and structural pores, especially increasing the amount of storage pores (pores with diameter D = 9.09 µm). BC also modified the portion of the soil pore size distribution associated with aggregation [52], increasing the soil aggregate stability [53,54]. The specific surface area (SSA) measured by Nitrogen adsorption measurements confirmed that BC modified the internal soil pores' structure, increasing the number of structural, connected pores [55]. As a consequence, BC increased the values of saturated volumetric water content, θ s , field capacity, θ fc , and wilting point, θ wp , enhancing the soil water retention and increasing the AW max from a value of 19 mm/m for the DS without BC to values of 62, 103 and 109 when BC fractions, f bc , equal to 0.091, 0.23 and 0.33, respectively, were mixed with the DS.
These results suggested that the application of BC could be a management option improving the soil's structural and retention characteristics, as well as the AW max of the DS. This could be of considerable potential interest not only for the Al-Foah area but also for other arid and/or semi-arid regions characterized by large extensions of sandy soils similar to that considered, with poor structural and retention characteristics.
The objective of this paper was to explore the hitherto little investigated application of BC to DS, as an amendment option improving soil water balance and irrigation management for three different crops, i.e., wheat (Triticum aestivum), sorghum (Sorghum vulgare) and tomato (Lycopersicon esculentum). Soil water balance and irrigation scheduling were simulated by AQUACROP using as input the soil's hydrological parameters previously measured at f bc = 0, 0.091, 0.023 and 0.33 [51]. The specific objective was to investigate if under drip irrigation, the application of BC to DS improved the irrigation water use efficiency (IWUE) and water use efficiency (WUE), under conditions of no water stress (NWS) for the crops and under deficit irrigation (DI).

Study-Area and Soil Characteristics
The sandy soil was taken from the dunes near Al Foah (long 55 • 50 E and lat 24 • 30 N), located north of Al Ain, a city in the Eastern Region of the Emirate of Abu Dhabi, which is one of the seven States of the UAE, with an area of about 64,000 km 2 , most of which covered by Aeolian sand. Agricultural development is limited to a few small-scattered oases, such as the Liwa group and those at Al Ain. The latter, seven in Abu Dhabi State and two, including Buraimi, in Muscat territory, have always been the most important centers of agriculture in the region.
Climate is arid, with very high summer temperatures. The interior desert region has hot summers with temperatures rising up to about 50 • C. Annual rainfalls are very scarce and vary between 15 mm and 78 mm, with January and February being the rainy months.
The desert sand (DS) considered in the AQUACROP application was characterized by a content in clay, silt and sand equal to 2.6, 1.9 and 95.6%, respectively, a low organic matter content (0.54 g/kg) and a Ca and Mg content equal to 12 and 4 mg/kg, respectively. A weak structural development was

Model Description
The AQUACROP model, developed by the FAO Land and Water Division, is a water-driven model that can be used as a planning and support tool for management decisions in both irrigated and rainfed agriculture [20,58]. The model computes the soil water balance in the root zone according to Allen et al. [59,60], calculating the root zone depletion, D r , at the end of every day, as: where the subscripts i and i − 1 refer to the current day and to the previous day, respectively, R (mm) is the rainfall depth, RO (mm) is the runoff, V IRR (mm) is the irrigation depth, CR (mm) is the capillary rise, ET (mm) is the actual crop evapotranspiration and DP (mm) is deep percolation. The parametric equations described by Liu et al. [61] are used to calculate CR and DP. The water balance is highly affected by runoff, RO, [62], which in turn is highly influenced by the soil's hydrological characteristics [63,64]. AQUACROP estimates RO by using the curve number approach [65], and splits the maximum ET component, ET c , in maximum soil evaporation, E s (mm), and maximum crop transpiration, T c (mm), which are estimated as: where ET 0 (mm) is the reference evapotranspiration [59], CC (%) is the crop development value, K e is the maximum soil evaporation coefficient (dimensionless), K r is the evaporation reduction coefficient (0-1) and K cTr is the standard crop transpiration coefficient (dimensionless) when CC = 100%. The actual transpiration, T a (mm), is calculated as: where K s is a water stress coefficient ranging between 0 and 1 [59,66]. Further details on calculating T a and E s are provided by Raes et al. [67]. In AQUACROP the ground dry biomass, B (kg ha −1 ), is computed as the product of the season-cumulated ratio T a /ET 0 and the adjusted biomass (water) productivity BWP* (g m −2 ), that takes into account the actual atmospheric CO 2 concentration [67], whereas the crop yield, Y (kg ha −1 ), is predicted by B as: where HI 0 is the reference harvest index, which indicates the harvestable proportion of biomass, and f HI is an adjustment factor computed by the model and related to five water stress factors, i.e., inhibition of leaf growth, inhibition of stomata, reduction in green canopy duration due to senescence, reduction in biomass due to pre-anthesis stress and pollination failure. The main input data required by AQUACROP are: Climatic data-daily maximum and minimum air temperature, T max ( • C) and T min ( • C), daily rainfall (R), reference evapotranspiration (ET 0 ) and mean annual CO 2 (ppm) concentration in the atmosphere.
Crop data-crop calendar with reference to the emergence, maximum coverage, maturity and senescence stages; maximum value of K cTr ; minimum and maximum root depths and roots expansion shape factor; the parameters of the canopy cover (CC) curve; the canopy growth coefficient CGC (% degree day −1 ) and the canopy decline coefficient CDC (% degree day −1 ); the parameters used to compute yield, BWP* and HI 0 defined above; the water stress coefficients relative to canopy expansion, stomatal closure, early canopy senescence and aeration stress due to water logging.
Soil data-number of layers (maximum 5) and soil layer depth, d (m), and for each layer volumetric water content at saturation, θ s (m 3 m −3 ), field capacity, θ fc , and wilting point, θ wp . Irrigation-irrigation system; scheduling parameters (time and depth criteria), in case of irrigation scheduled by the model, or time and volume of each water supply in case of user-imposed irrigation depth. The irrigation system can be selected as "sprinkler", "surface" or "drip". In the latter case, an adjustment coefficient (%) for partial soil wetting can be used to take into account the actual irrigated area.
Field and management data-salinity, soil fertility, mulching and runoff reduction practices.

Data Collection and Simulation Scenarios
Daily rainfall (R) and daily maximum and minimum air temperature (T max and T min ), recorded from 1 Jan 2006 to 31 Dec 2007 at Al Buraimi weather station (24.233 • lat, 55.783 • long, 299 m a.s.l.), were used to calculate the reference evapotranspiration ET 0 [60], using the Hargreaves equation [68], and to carry out the AQUACROP simulations. Al Buraimi is located in Oman at the border with the UAE, and is near to the Al Foah area. Data from the Al Buraimi weather station were used because they fully covered the investigated period.  Irrigation-irrigation system; scheduling parameters (time and depth criteria), in case of irrigation scheduled by the model, or time and volume of each water supply in case of user-imposed irrigation depth. The irrigation system can be selected as "sprinkler", "surface" or "drip". In the latter case, an adjustment coefficient (%) for partial soil wetting can be used to take into account the actual irrigated area.
Field and management data-salinity, soil fertility, mulching and runoff reduction practices.

Data Collection and Simulation Scenarios
Daily rainfall (R) and daily maximum and minimum air temperature (Tmax and Tmin), recorded from 1 Jan 2006 to 31 Dec 2007 at Al Buraimi weather station (24.233° lat, 55.783° long, 299 m a.s.l.), were used to calculate the reference evapotranspiration ET0 [60], using the Hargreaves equation [68], and to carry out the AQUACROP simulations. Al Buraimi is located in Oman at the border with the UAE, and is near to the Al Foah area. Data from the Al Buraimi weather station were used because they fully covered the investigated period.  Biochar, BC, from forest biomass (Abies alba M., Larix decidua Mill., Picea excelsa L., Pinus nigra A. and Pinus sylvestris L., mixed in equal relative proportions), pyrolyzed at 450 °C for 48 h, had been used to amend the DS and to measure the soil water retention [51]. The BC fractions considered by Baiamonte et al. [51] were: fBC = 0.0 (soil only), fBC = 0.091, 0.23, 0.33 and 1.00 (BC only), with the biochar fraction, fBC, defined as: where PBC (g) and PS (g) are the BC and DS weights, respectively. Table 1 reports the saturated water content, θs, together with the θfc and θwp values previously measured for the DS without BC and with BC addition [51] and used as input to simulate the soil water balance and irrigation scheduling under drip irrigation. These hydraulic parameters were determined by the High Energy Moisture Characteristic (HEMC) technique [69,70], and therefore took into account the structural changes occurring in the soil-BC mixtures after BC addition. The importance of using hydraulic parameters determined with measurement techniques detecting even small structural changes as input to physically-based simulation models, in order to obtain reliable model predictions, has been evidenced in previous investigations (17). This also means that the ET values predicted by AQUACROP using these hydraulic parameters can be considered reliable.
The crops chosen for the simulations were wheat, sorghum and tomato, which are of economic interest for many countries all over the world. Production of wheat is negligible in the UAE due to extreme heat, low rainfall and barren desert soil, but increasing the local production could be of interest for the UAE. Abu Dhabi has two large wheat farms at Al Ain, and experimental farms at Rawaya and Mazaid (near Al Ain) were designed to encourage local Bedouins to take up settled Biochar, BC, from forest biomass (Abies alba M., Larix decidua Mill., Picea excelsa L., Pinus nigra A. and Pinus sylvestris L., mixed in equal relative proportions), pyrolyzed at 450 • C for 48 h, had been used to amend the DS and to measure the soil water retention [51]. The BC fractions considered by Baiamonte et al. [51] were: f BC = 0.0 (soil only), f BC = 0.091, 0.23, 0.33 and 1.00 (BC only), with the biochar fraction, f BC , defined as: where P BC (g) and P S (g) are the BC and DS weights, respectively. Table 1 reports the saturated water content, θ s , together with the θ fc and θ wp values previously measured for the DS without BC and with BC addition [51] and used as input to simulate the soil water balance and irrigation scheduling under drip irrigation. These hydraulic parameters were determined by the High Energy Moisture Characteristic (HEMC) technique [69,70], and therefore took into account the structural changes occurring in the soil-BC mixtures after BC addition. The importance of using hydraulic parameters determined with measurement techniques detecting even small structural changes as input to physically-based simulation models, in order to obtain reliable model predictions, has been evidenced in previous investigations (17). This also means that the ET values predicted by AQUACROP using these hydraulic parameters can be considered reliable. Table 1. Experimental soil hydrological parameters [51].
The crops chosen for the simulations were wheat, sorghum and tomato, which are of economic interest for many countries all over the world. Production of wheat is negligible in the UAE due to extreme heat, low rainfall and barren desert soil, but increasing the local production could be of interest for the UAE. Abu Dhabi has two large wheat farms at Al Ain, and experimental farms at Rawaya and Mazaid (near Al Ain) were designed to encourage local Bedouins to take up settled farming. As part of a research project, about 100 varieties of wheat used for the production of flour to make bread were cultivated in Al Ain in 2013, and the initial results were positive [68,71].
Sorghum is predominantly grown in semi-arid and arid agro-ecological areas and is considered one of the crops that can also be cultivated for large-scale production, despite the harsh climate conditions of the UAE. Sorghum cultivation is being investigated on a trial basis at the experimental farms of UAE University. Tomato is one of the most widely cultivated crops in the world, including UAE, where the government policies are supporting tomato cultivation, enhancing locally-grown production for the UAE consumers [71]. Table 2 reports the main default parameters implemented in AQUACROP and used as inputs to carry out the simulations for the three considered crops. The length of crop development stages, T LCD (day), i.e., the time from sowing/transplanting to emergence, to maximum rooting depth, to the beginning of senescence and to maturity, were adjusted in order to take into account the site-specific conditions [59].
For each crop, a soil water balance was performed daily by using a simulation period between sowing/transplanting and the end of maturity. Since the accuracy of the soil water balance depends on the soil water content (θ) at the starting day, information not available, the first rainy day following sowing/transplanting, characterized by a rainfall depth > 15 mm, was chosen as starting date. This made it possible to start the simulations for all the crops from the same initial soil conditions at the field capacity, θ fc . Simulations were carried out in order to investigate the effects (i) of the different f BC , and (ii) of the different options of BC incorporation into soil layers of different depths, (Z DS+BC ), on soil-water balance, on irrigation efficiency and on the crop response to water.
Incorporation of BC into the soil was assumed to be technically feasible if carried out with plowing, before sowing, as also shown by Moiwo et al. [72].
Although values of Z DS+BC equal to 30 or 40 cm are likely to be non-sustainable possibilities, because of the considerable amounts of BC that should be used to amend the soil with these options, these values were also considered as a theoretical possibility of incorporating BC into the deepest soil layers. In case of low-cost BC obtained in farms by waste residues, this incorporation could be done during tillage before sowing when soil preparation is carried out, and could be justified by the benefits in terms of the increase in soil fertility and related crop productivity, in the long term [49].
Simulations were carried out by assuming drip irrigation [8,73] under two different management scenarios. In the first scenario, a condition of "no water stress for the crop" (NWS) was simulated by an irrigation scheduling bringing the soil water content at θ fc , when 65% of the readily available water (RAW) was depleted from the soil. RAW was calculated as a p fraction of AW max equal to 0.65 [59]. Therefore, for each crop, this scenario determined T/T max ratios equal to 1, evidencing the effect of BC on the predicted irrigation water needs (V IRR ) and on the irrigation scheduling when the objective was to avoid any water stress for the crop, obtaining the maximum yield, Y.
Instead, the second scenario simulated a condition of water stress for the crops by scheduling a deficit irrigation (DI), supplying the soil-crop system with a total amount of irrigation water, V IRR , restoring 70% of the maximum crop evapotranspiration, ET C . Under this DI condition, V IRR did not vary, for the same crop, at changing f BC and Z DS+BC , and the effect of BC on crop yield, Y, and on the WUE could instead be evaluated for each crop. Table 3 reports some information related to the simulated scenarios. The DS option refers to soil water balance simulations carried out for the soil without BC, whereas the "Z BC+DS " and Z DS are the different options of incorporating the three BC fractions (f BC = 0.091, f BC = 0.23 and f BC = 0.33) into soil layers of different depths.
The performance of each scenario, with the options of BC incorporation, on the soil-crop system and on irrigation management, was assessed by considering the outputs provided by AQUACROP, as well as by calculating the irrigation water use efficiency index, IWUE (kg m −3 ), as: where Y (t ha −1 ) is the crop yield and V IRR (mm) is the total irrigation volume. For the NWS scenario, the dimensionless ratios V IRR /V IRR,DS and N IRR /N IRR,DS , obtained by dividing the total irrigation volumes, V IRR , and the number of irrigations, N IRR , by the corresponding values obtained simulating only sand without BC, were considered.  For the DI scenario, the ratios between the actual transpiration, T, the maximum transpiration, T max , the yield, Y, the maximum yield, Y max , the irrigation water use efficiency, IWUE, and the water use efficiency, WUE, were analyzed.

Results and Discussion
With reference to the three crops and to the NWS scenario, for each of the considered f BC (f BC = 0.091, 0.23 and 0.33), The main outputs of the soil water balance simulated by AQUACROP for the different Z DS+BC , including the results obtained for the DS, without BC, are reported in Table 4. The reference evapotranspiration, ET 0 (mm), during the simulation period, the total irrigation volume, V IRR (mm), predicted by the model, the actual crop transpiration, T (mm), which was equal to the maximum (T = T max ), the actual soil evaporation, E (mm), the number of irrigations, #IRR, and the average irrigation water volume, h IRR (mm), the crop biomass, B mass (t ha −1 ), the cumulative Harvest Index, H i (%), the crop yield, Y (t ha −1 ), the water use efficiency, WUE (kg m −3 ), and the irrigation water use efficiency, IWUE (kg m −3 ), with the latter calculated per Equation (7), are also reported in Table 4.
As can be seen in Table 4, BC positively affected the soil water balance, by decreasing the amount of V IRR necessary to fully meet the crop water requirement and to obtain the maximum Y, with V IRR values that decreased at increasing f BC and Z DS+BC , compared to DS.
The positive effects of BC on soil water balance were the consequence of the improved retention characteristics, which in turn were the consequence of the improved structural and aggregation soil properties due to BC addition. BC increased the number of pores with diameter = 9.09 µm (storage pores), improved the aggregate stability and increased the number of internal pores [51]. Since AW max increased at increased f BC , indicating a greater ability of the soil to store water as a consequence of BC addition, this explains why lower V IRR values were predicted when BC was applied at higher f BC and application depths (Z DS+BC ).
These results indicated that the addition of BC to the DS can be considered a water saving strategy. The decrease in the V IRR values was the consequence of the increase in soil porosity and water retention due to BC [51], as can be illustrated by analyzing the temporal variation of the volumetric water content, θ, and of the irrigation scheduling obtained by AQUACROP for the DS without BC and after BC addition. Figure 2a-c shows that, for the three crops, when the soil water balance was simulated for the DS, without BC, θ was almost constant during the entire simulation period and close to the value of 0.046, corresponding to the DS field capacity, θ fc (Table 1). Instead, when BC was applied to the DS at f BC = 0.23 and Z DS+BC = 30 cm (Figure 2d-f), a larger variability occurred in the θ values predicted before and after irrigation. This was the consequence of the increased field capacity, θ fc (from 0.046 to 0.152 cm 3 ), wilting point, θ wp (from 0.027 to 0.049 cm 3 ), and AW max , determined by the BC application ( Table 1).
The increase in AW max from 19 mm/m to 103 mm/m after BC addition at f BC = 0.23 indicated a considerably enhanced soil ability to act as a reservoir for water and to provide a higher amount of water to the crops. This explains why the simulation carried out with AW max = 103 mm/m, compared to that carried out with AW max = 19 mm/m, determined an irrigation scheduling predicting an amount of V IRR equal to 224.4 mm, for wheat, against the V IRR of 251.3 mm obtained for the DS. For sorghum, the predicted V IRR was equal to 308.8 mm, against the value of 343.8 mm, and for tomato the predicted V IRR was equal to 367.0 mm, against the value of 394.3 mm. Thus, for the NWS scenario, and for f BC = 0.23 and Z DS+BC = 30 cm, the reductions in the V IRR values were equal to 10.7%, 10.2% and 6.9% for wheat, sorghum and tomato, respectively, with the minimum effect on tomato and comparable effects on wheat and sorghum. These results indicated that, under the climatic conditions explored in this investigation, the addition of BC to the DS made it possible to save water, showing the positive role of BC in improving irrigation management for the considered DS and for the three crops.
the predicted VIRR was equal to 308.8 mm, against the value of 343.8 mm, and for tomato the predicted VIRR was equal to 367.0 mm, against the value of 394.3 mm. Thus, for the NWS scenario, and for fBC = 0.23 and ZDS+BC = 30 cm, the reductions in the VIRR values were equal to 10.7%, 10.2% and 6.9% for wheat, sorghum and tomato, respectively, with the minimum effect on tomato and comparable effects on wheat and sorghum. These results indicated that, under the climatic conditions explored in this investigation, the addition of BC to the DS made it possible to save water, showing the positive role of BC in improving irrigation management for the considered DS and for the three crops.  (Table 4). However, also in these other cases, reductions in the VIRR and #IRR values due to BC addition were always maximal for wheat, followed by sorghum and by tomato. For wheat, the reductions in VIRR ranged from 18% (for fBC = 0.33 and ZDS+BC = 40 cm) to 3.1% (for fBC = 0.091 and ZDS+BC = 10 cm), for sorghum from 13.5% (fBC = 0.33 and ZDS+BC = 40 cm) to 3.7% (fBC = 0.091 and ZDS+BC = 10 cm) and for tomato from 7.5% (fBC = 0.33 and ZDS+BC = 40 cm) to 2.8% (fBC = 0.091 and ZDS+BC = 10 cm). These results indicated that i) the maximum BC benefit was found at the maximum considered depth of 40 cm, agreeing with the experimental results of Moiwo et al. [72], who found an increasing BC benefit with increasing application depth, ii) wheat was the crop showing the maximum benefit from BC application and tomato the crop having the  (Table 4). However, also in these other cases, reductions in the V IRR and #IRR values due to BC addition were always maximal for wheat, followed by sorghum and by tomato. For wheat, the reductions in V IRR ranged from 18% (for f BC = 0.33 and Z DS+BC = 40 cm) to 3.1% (for f BC = 0.091 and Z DS+BC = 10 cm), for sorghum from 13.5% (f BC = 0.33 and Z DS+BC = 40 cm) to 3.7% (f BC = 0.091 and Z DS+BC = 10 cm) and for tomato from 7.5% (f BC = 0.33 and Z DS+BC = 40 cm) to 2.8% (f BC = 0.091 and Z DS+BC = 10 cm). These results indicated that (i) the maximum BC benefit was found at the maximum considered depth of 40 cm, agreeing with the experimental results of Moiwo et al. [72], who found an increasing BC benefit with increasing application depth, (ii) wheat was the crop showing the maximum benefit from BC application and tomato the crop having the minimum benefit, with sorghum in the middle, and (iii) for wheat, lower f BC than those required for sorghum and tomato would be enough to obtain comparable V IRR reductions, indicating that BC was more efficiently used by wheat.
BC increased the irrigation water use efficiency (IWUE) for the three crops, with the maximum IWUE obtained for f BC = 0.33 and Z DS+BC = 40 cm (Table 4). However, the effect of BC on IWUE was maximal for wheat (from 2.77 kg m −3 to 3.38 kg m −3 ) and negligible for sorghum and tomato. This result is in agreement with the previously observed maximum effect of BC on the V IRR values observed for wheat.
For each crop and for the different f BC and Z DS+BC options, Figure 3 illustrates the dimensionless ratios V IRR /V IRR,DS (Figure 3a-c) and #IRR/#IRR DS (Figure 3d-f), obtained by dividing the V IRR and the #IRR by the corresponding DS without BC, as a function of the considered Z DS+BC . The lowest V IRR /V IRR,DS and #IRR/#IRR DS indicate more efficient water and energy saving solutions, either in terms of f BC or of Z DS+BC . These figures can be useful to select different possible f BC and Z DS+BC options even for f BC and Z DS+BC values not explored in the simulations.
The fBC and ZDS+BC options could be further considered in terms of feasibility and sustainability according to cost-benefit analyses taking into account site-specific factors, such as the cost and the availability of BC, the cost of water and of energy needed for irrigation, as well as the benefits of amending the DS by BC in terms of VIRR, #IRR and IWUE.
In the UAE, the addition to sandy soils of BC obtained from palm dates significantly improved plant growth in maize and soil productivity [74], since there are about 40 million palm trees and each tree generates annually about 15 kg of biomass waste, with a total of 600 million kg of green waste. The use of on-farm produced BC as amendment could be a sustainable practice enhancing the soil water retention [75] and increasing the IWUE [76], thus helping to save water.
If BC can be produced from locally abundant feedstock, the low cost of BC could justify the amount of BC to be used to amend the soil with higher fBC and/or ZDS+BC. An example of the use of BC from date palms as a strategy improving the soil water retention of desert sand in Saudi Arabia, increasing the yield of wheat, was recently reported by Alotaibi and Shoenau [49].  The f BC and Z DS+BC options could be further considered in terms of feasibility and sustainability according to cost-benefit analyses taking into account site-specific factors, such as the cost and the availability of BC, the cost of water and of energy needed for irrigation, as well as the benefits of amending the DS by BC in terms of V IRR , #IRR and IWUE.
In the UAE, the addition to sandy soils of BC obtained from palm dates significantly improved plant growth in maize and soil productivity [74], since there are about 40 million palm trees and each tree generates annually about 15 kg of biomass waste, with a total of 600 million kg of green waste. The use of on-farm produced BC as amendment could be a sustainable practice enhancing the soil water retention [75] and increasing the IWUE [76], thus helping to save water.
If BC can be produced from locally abundant feedstock, the low cost of BC could justify the amount of BC to be used to amend the soil with higher f BC and/or Z DS+BC . An example of the use of BC from date palms as a strategy improving the soil water retention of desert sand in Saudi Arabia, increasing the yield of wheat, was recently reported by Alotaibi and Shoenau [49]. Figure 3 also shows that, in general, similar V IRR /V IRR,DS and #IRR/#IRR DS ratios were obtained for Z DS+BC = 10, 20 cm, indicating that the application of BC into the first 10 cm could be sufficient in order to gain some benefits, except for wheat (at f BC = 0.091), sorghum (at f BC = 0.33) and tomato (at f BC = 0.23 and 0.33), for which Z DS+BC = 20 cm was necessary to further reduce V IRR .
For the three crops, Table 4 also shows that in the NWS scenario, no effects of BC occurred on the biomass, B mass , on the H i index or on the crop yield, Y. These expected results are due to the fact that no water stress was induced on the crops, and as a consequence the yield, Y, was always maximal. Therefore, negligible effects were also observed on the WUE, because of the slight influence of the evaporation component, E (Table 4), on ET, obtained by varying f BC and Z DS+BC .
With reference to the "deficit irrigation", DI, Table 5 reports the main outputs of the soil-water balance simulated by AQUACROP for the three considered crops and for the previously considered f BC and Z DS+BC values.
In particular, DI was applied by supplying to each crop an irrigation volume, V IRR , calculated as 0.7 ET C , and setting a number of irrigations (#IRR) of fixed depth, h IRR (mm), uniformly distributing V IRR in the simulation period. According to this irrigation scheduling (Table 3), the pairs (#IRR, h IRR ) satisfying the values of the pre-calculated V IRR were equal to (38,5), (33,8) and (37,8) for wheat, sorghum and tomato, respectively. The supplied V IRR were equal to 190 mm, 264 mm and 296 mm for wheat, sorghum and tomato, respectively.
As can be seen in Table 5, the application of BC increased the T/T max ratio, at increasing f BC and Z DS+BC , with the T/T max calculated by considering only the T/T max < 1 values. The lowest T/T max were obtained for wheat (average of all the values equal to 0.356), and the highest for sorghum (0.614), with tomato in the middle (0.595). These results indicated that a greater water stress (WS) condition occurred for wheat compared to sorghum and tomato.
For the three crops, the productivity parameter, B mass , increased at increasing f BC and Z DS+BC , with the highest B mass obtained for sorghum, followed by wheat and tomato. However, the H i values were maximum for tomato (H i between 61.8% and 54.6%), followed by wheat (between 46.2% and 25.7%). Instead, for sorghum, H i values were much lower (between 26.2% and 18.1%) than those obtained for wheat and tomato, thus indicating that the effect of water stress (WS) negatively impacted sorghum in terms of H i .
The low H i values obtained for sorghum can be explained by the shorter H i building up period (Table 2), during which T/T max lower than 1 occurred with a higher frequency than for wheat and tomato, as can be seen by considering the ratio (t ws /t LCD ) between the number of days, t ws , during which T/T max < 1, with respect to the length of the crop development stage, t LCD (Table 2). Indeed, as reported in Table 5, the t ws /t LCD ratios obtained for sorghum were higher than those for wheat and tomato, thus explaining the lowest H i and Y values obtained for sorghum, compared to wheat and tomato, notwithstanding its highest T/T max .
As can be seen in Table 5, for the three crops, the maximum positive effect of BC occurred when BC at f BC = 0.33 was considered to be applied up to the maximum considered depth, Z DS+BC = 40 cm. In this case, for wheat, Y increased from 1.5 t/ha (for the DS) to 6.1 t/ha, with this latter value being comparable to that of the NWS scenario (Y = 6.96 t/ta) and indicating that BC mitigated the yield reduction due to water stress. For tomato, Y increased from 4.3 t/ha (for the DS) to 7 t/ha, against a value of 7.84 t/ha for the NWS condition, indicating a similar beneficial effect of BC application. Instead, for sorghum, Y increased from 1.4 t/ha (for the DS) to 3.9 t/ha, against a value of 7.1 t/ha for the NWS scenario, indicating less benefits of BC application than for wheat and for tomato. Figure 4a-c illustrate the ratios Y/Y max , as a function of Z DS+BC . For wheat (Figure 4a), BC increased the Y/Y max values at increasing f BC and Z DS+BC , thus decreasing the Y reduction (1−Y/Y max ) from 78.4% for the DS to values ranging between 12.3% (for f BC = 0.33 and Z DS+BC = 40 cm) and 58.3% (for f BC = 0.091 and Z DS+BC = 10 cm).
For sorghum (Figure 4b), the Y/Y max values were lower than those obtained for wheat, with reductions in Y decreasing from 80.3% for the DS to values ranging between 45.1% (for f BC = 0.33 and Z DS+BC = 40 cm) and 69.0% (for f BC = 0.091 and Z DS+BC = 10 cm).
Tomato (Figure 4c) was the crop less affected by water stress under DI conditions, with a reduction in Y equal to 45.1% for the DS, compared to the NWS scenario, much lower than those obtained for wheat (78.4%) and for sorghum (80.2%). For tomato, 1−Y/Y max decreased from 45.1% (DS) to 10.7% (for f BC = 0.33 and Z DS+BC = 40 cm) and to 26.0% (for f BC = 0.091 and Z DS+BC = 10 cm). These results demonstrated that with reference to the mitigating effect of BC on the Y reductions, wheat was the crop receiving the maximum benefits from BC application, followed by tomato, with sorghum showing the lowest benefits. Tomato (Figure 4c) was the crop less affected by water stress under DI conditions, with a reduction in Y equal to 45.1% for the DS, compared to the NWS scenario, much lower than those obtained for wheat (78.4%) and for sorghum (80.2%). For tomato, 1−Y/Ymax decreased from 45.1% (DS) to 10.7% (for fBC = 0.33 and ZDS+BC = 40 cm) and to 26.0% (for fBC = 0.091 and ZDS+BC = 10 cm). These results demonstrated that with reference to the mitigating effect of BC on the Y reductions, wheat was the crop receiving the maximum benefits from BC application, followed by tomato, with sorghum showing the lowest benefits.  The maximum IWUEs were obtained for wheat, followed by tomato, with lower IWUEs for sorghum even at the highest fBC and ZDS+BC.
When the WUE was considered (Table 5, Figure 4g-i), the maximum WUE was obtained for wheat and tomato, and the minimum for sorghum. These results depended on the lower actual ET predicted for wheat (average value 209.5 mm) compared to tomato (average value 309.8 mm) and to sorghum (average value 257.1 mm), with maximum Y obtained for tomato, followed by the minimum for sorghum. These results confirmed that, under DI, wheat was the crop providing the maximum WUE, and tomato the second possible best option, while cultivation of sorghum slightly benefited from BC application and could be inadvisable under deficit irrigation.
These outcomes could be of interest considering the economic importance of cultivating wheat and tomato in the UAE and in many other countries of the world.
For the three crops, Figure 5a,b illustrate the comparison between the WUE and IWUE obtained with the DI and NWS scenarios, either without BC and after BC addition, at fBC = 0.091 and ZDS+BC = 10  The maximum IWUEs were obtained for wheat, followed by tomato, with lower IWUEs for sorghum even at the highest f BC and Z DS+BC .
When the WUE was considered (Table 5, Figure 4g-i), the maximum WUE was obtained for wheat and tomato, and the minimum for sorghum. These results depended on the lower actual ET predicted for wheat (average value 209.5 mm) compared to tomato (average value 309.8 mm) and to sorghum (average value 257.1 mm), with maximum Y obtained for tomato, followed by the minimum for sorghum. These results confirmed that, under DI, wheat was the crop providing the maximum WUE, and tomato the second possible best option, while cultivation of sorghum slightly benefited from BC application and could be inadvisable under deficit irrigation.
These outcomes could be of interest considering the economic importance of cultivating wheat and tomato in the UAE and in many other countries of the world.
For the three crops, Figure 5a,b illustrate the comparison between the WUE and IWUE obtained with the DI and NWS scenarios, either without BC and after BC addition, at f BC = 0.091 and Z DS+BC = 10 cm, which is the option among those explored, involving the minimum amount of BC. Values of f BC lower than 0.091 were not explored in the simulations because the results previously obtained showed no effects on porosity, soil water retention and AW max of the DS when f BC = 0.014 was tested [51].
For the three crops, when the NWS scenario was considered, the effect of this f BC and Z DS+BC on IWUE and on WUE was negligible. Instead, under DI, BC increased the WUE and the IWUE, compared to the values obtained for the DS without BC, counterbalancing the lower irrigation amount, V IRR , supplied with the DI. The maximum increase in IWUE and WUE as an effect of BC addition was found for wheat, with lower effects on tomato than on wheat, and negligible effects on sorghum.
These results showed that even at the lowest considered f BC , among those considered, BC addition could be a useful strategy when DI is practiced for cultivation of wheat and tomato, under the extremely arid conditions of the considered study area and/or similar climatic areas. Our results agree with those obtained by Bagheri et al. [77] in Iran, indicating that the use of biochar can be a successful strategy for improving water use efficiency and reducing the effects of drought stress under extremely arid climates and for soils characterized by very poor soil water retention, as the desert sand considered in this investigation.
Although these results could be considered reliable, because based on accurately measured hydraulic properties as input of the hydrological AQUACROP component, and on suitable crop parameters, a field investigation aimed at validating some of the best options provided by the model should be the next step to be pursued. cm, which is the option among those explored, involving the minimum amount of BC. Values of fBC lower than 0.091 were not explored in the simulations because the results previously obtained showed no effects on porosity, soil water retention and AWmax of the DS when fBC = 0.014 was tested [51].
For the three crops, when the NWS scenario was considered, the effect of this fBC and ZDS+BC on IWUE and on WUE was negligible. Instead, under DI, BC increased the WUE and the IWUE, compared to the values obtained for the DS without BC, counterbalancing the lower irrigation amount, VIRR, supplied with the DI. The maximum increase in IWUE and WUE as an effect of BC addition was found for wheat, with lower effects on tomato than on wheat, and negligible effects on sorghum.
These results showed that even at the lowest considered fBC, among those considered, BC addition could be a useful strategy when DI is practiced for cultivation of wheat and tomato, under the extremely arid conditions of the considered study area and/or similar climatic areas. Our results agree with those obtained by Bagheri et al. [77] in Iran, indicating that the use of biochar can be a successful strategy for improving water use efficiency and reducing the effects of drought stress under extremely arid climates and for soils characterized by very poor soil water retention, as the desert sand considered in this investigation.
Although these results could be considered reliable, because based on accurately measured hydraulic properties as input of the hydrological AQUACROP component, and on suitable crop parameters, a field investigation aimed at validating some of the best options provided by the model should be the next step to be pursued. Figure 5. For the three considered crops, for BC fraction, fBC = 0.091, and for the depth of BC application, ZDS+BC = 10 cm, (a) water use efficiency, WUE, and (b) irrigation water use efficiency, IWUE, for the two considered irrigation management scenarios, no water stress (NWS) and deficit irrigation (DI), without BC (DS) and by BC application to DS (DS + BC). Table 4. Soil layering for different biochar fractions, fBC, and main outputs obtained using AQUACROP for the three considered crops. Data refer to a soil-water balance simulation with a "no water stress irrigation" management (NWS).     Table 5. Soil layering for different biochar fractions, f BC , and main outputs obtained using AQUACROP for the three considered crops. Data refer to a soil-water balance simulation with a "deficit irrigation" management (DI).

Conclusions
This investigation showed that the application of biochar to a desert sand for three crops, i.e., wheat, sorghum and tomato, positively affected the soil water balance and the irrigation scheduling simulated by the AQUACROP model, using as input accurately measured soil hydraulic parameters representing the physical and structural condition of the soil without and with BC addition.
The results indicated that the application of biochar could be proposed as a strategy mitigating water and energy saving consumptions when no water stress conditions were considered, in order to obtain the maximum crop yield. Wheat was the crop showing the maximum benefits from biochar application in terms of reduction in the amount of irrigation volume, of increase in the irrigation water use efficiency and of decrease in the number of predicted irrigations. Biochar showed beneficial effects also for sorghum and tomato but to a lesser extent than for wheat.
When a deficit irrigation was simulated, the application of biochar counterbalanced the lower amount of irrigation volume supplied under deficit irrigation, partially mitigating the crop yield reduction due to the water stress conditions, especially for wheat and for tomato, with a less pronounced effect for sorghum. Under deficit irrigation, the application of biochar could also be a useful management option to promote the cultivation of wheat and tomato in the UAE, as well as in other countries with similar climatic and soil conditions. This could help to increase the local production, thus reducing the impact of water shortage on food security and the dependency on food import.