Fog water harvesting potential and its use in supplementary irrigation of rainfed crops (winter wheat) in Abi-beyglu, Ardabil (Iran)

Fog water harvesting has been studied by numerous researchers thanks to its sustainability and low-cost features. In many foggy areas around the world where the rainfall is inadequate or not uniformly distributed, application of fog as an unconventional water resource can be seen as a suitable alternative for supplying water demands for drinking, farming, and/or forest rehabilitation. In drought-affected areas, provision of potable water is a major concern due to long distance from water resource, tough topography, and low quality of available resources or high cost of supplying water from alternative resources. In areas where water transmission is infeasible or rainfall deficiency/heterogeneity across different seasons is encountered, fog water collection can serve as an alternative for supplying the water demand of farmlands (Carrera-Villacrés et al. 2017).

Geographically speaking, fog water collection depends on such factors as height and topography. The topography and height play critical roles in the formation and guidance of fog. As a potential water resource, the fog is usually studied in high mountainous areas where the required conditions for fog formation are met (Klemm et al. 2012). As the height increases, air temperature drops followed by compaction of the air humidity content (Molina & Escobar 2008). This implies that the best position to install fog water collection facility is along the top line. Successful fog water collection projects have been installed at heights in the range of 400–1,000 m above mean sea level (MSL). Installation of collectors on high lands can enhance the water collection yield by up to 19 folds (Ritter et al. 2008). Another parameter to consider is the liquid water content (LWC), which refers to liquid water mass per unit volume. LWC is directly associated with the collected water. In general, sea fog or the radiation fog formed at low heights fails to exhibit adequate LWC. This implies the necessity of high mountains that can serve as a basis for the formation of high-LWC mountainous fog (Bruijnzeel et al. 2005). The fog frequency and duration are directly associated with the water collection yield. However, the fog is a seasonal and local phenomenon in most regions. In Chile, for example, fog time extends over virtually the entire year while Dhofar in Oman hosts the phenomenon for no more than 3 months a year (Schemenauer & Cereceda 1994b). Therefore, in order to plan and implement a fog water collection project, one needs to gather adequate data on the fog frequency, duration, and seasonal variations. This can be obtained from satellite images, synoptic stations, local reports, and, if none of this is available, interviews with local people (Fessehaye et al. 2015). Wind serves as the most important parameter contributing to successful implementation of fog water collection projects. As the wind speed increases, the fog flux passing over the collector surface increases, thereby increasing the water collection yield (Schemenauer & Joe 1989). When using flat collectors, both in research works and operating projects, the wind direction is of paramount importance as the collector surface must be normal to the wind direction. The wider the angular deviation of the collector surface out of the normal-to-wind direction, the lower the water yield, with a virtually zero yield when the collector surface is parallel to the wind direction (Estrela et al. 2009). Coastal areas are usually foggy but the wind speed may render inadequate to move the fog; that is, without adequate wind speed, a passive collector cannot function in practice. Therefore, before implementing a fog water collection project, one needs to evaluate local wind speed for successful operation of the system, configure the structures to minimize the loss across the system (Holmes et al. 2015), and consider dominant wind direction to determine installation direction of the collector (Schemenauer & Cereceda 1994a).

Although fog water harvesting is expanding in different countries around the world, its application has roots in the old age. In the ancient age, fog and dew water harvesting had been common in coastal/mountainous and desert areas, respectively. For example, honey comb structures and stone cones with heights of up to 10 m had been used to collect dew water in ancient Palestine and Crimea, respectively (Nelson 2003). For years, this has been the only way to supply drinking water for humans and animals in the Canary Islands, Antofagasta (Chile), and mountains of Oman (Schemenauer & Cereceda 1994b). In the recent years, numerous studies have been performed to investigate the water harvested from the fog around the world, a summary of which is presented in Table 1. In these studies, various collectors (e.g., SFC, LFC, QFC, Juvik, and Harp) were used under different topographic and height conditions to evaluate potential yield of fog water collection in different areas. According to the results of these studies, average water yield of different collectors ranges within 0.16–10 L/m²/day. In the meantime, yields of as high as 30 L/m²/day were recorded by an omnidirectional SFC in Oman (Schemenauer & Cereceda 1994b). Similar values have been reported for other collectors.

The applications of fog water mainly include the needs related to drinking and rehabilitation of forest areas. In the meantime, it has been used for agricultural applications in limited cases. The reason behind the limited use of fog water for agricultural applications is the relatively large water demand of the agricultural sector. Nevertheless, the fog water can be used for complementary irrigation of rainfed lands. Estrela et al. (2009) used the water collected by a passive large fog collector (LFC) to rehabilitate a forest area in the mountains of Valencia. They could collect water at an average rate of 3.3 L/m²/day. Results showed that micro irrigation of a segment of the forest rehabilitation seedlings for a short period of time enhances the survival rate and yield of the planted species. Another piece of research was done at the Marsa Matrouh Agricultural Research Center in Egypt by Harb et al. (2016) who investigated the effects of using the collected water from two fog water collectors (51 m² single-layer and bilayer propylene meshes with shading coefficients of 50 and 70%, respectively) in drip irrigation. Acknowledging the higher water captured from the bilayer propylene mesh, the resultant peanut yield and yield components were significantly different from those of the other collector. Carrera-Villacrés et al. (2017) used a fog water collection system to supply a part of the water demand for agricultural crops (especially the corn). They reported fog water yields of 5–20/m²/day for different days in a year. Results showed that the collected water can be devised to supply some 5% of the water demand for agricultural crops. In their studies in the San Cristóbal Island (Ecuador), Echeverria et al. (2020) installed two SFCs with shading coefficients of 50 and 35% at a height of 600 m. The water yield by the two SFCs was measured at 7.9 and 5.9 mm/day. They then presented the required number of collectors for satisfying 25 and 15% of the irrigation water deficiency in normal and dry years, respectively.

Inappropriate management of fresh water resources, especially the groundwater, drought and climate change (in the recent years) have led to a reduction in water resources, as reflected by the significant decline of groundwater table, which is the most prominent source of water. The urban population of the world who suffer from water scarcity is projected to increase from 933 million (i.e., one-third of total urban population in the world) in 2016 to 1.693–2.373 billion individuals in 2050 (some 50% of the world population) (He et al. 2021). The Middle East has been constantly hit by drought since after 1998, defining the worst drought period in the past 900 years, according to NASA (NASA 2016). Having such a vision in mind, competitions on water resource governance in these regions will be definitely boosted, thereby weakening the agriculture, worsening social constructs, encouraging urbanization of rural communities, increasing insecurity, contributing to instability, and finally triggering water wars. Due to low rainfall and inappropriate temporal and spatial distribution of the rainfall, Iran has been listed under arid and semi-arid regions, making it one of the most vulnerable areas to climate change (Al-Mandhari 2019). Given the population growth, expanded urbanization, and development of industrial and agricultural sectors, the demand for water is escalating in Iran. Preservation of existing water resources and search for alternative ones (e.g., air humidity, fog, etc.) are important measures of water resource management, which have been regarded by practitioners during the recent past. Considering the geographic position of the Ardabil Plain, which is overwhelmed by highly humid winds from the Caspian Sea that pull down stratus clouds (low-altitude) toward the ground surface, thereby lowering the air temperature, fog formation is a common phenomenon in the Abi-beyglu area (where the fog from the Caspian Sea enters the Ardabil Plain). This highlights the necessity of paying attention to fog water collection for various purposes. Obviously, one can store large amounts of water by collecting fog as it passes the region, with the stored water then used to supply a part of the farming or even drinking water demands for local villagers.

With an approximate area of 140,000 ha, the Ardabil Plain is delineated by 48°10′E–48°30′E and also 38°7′N to 38°23′N, Ardabil Province, Iran, at an average height of 1,330 m from MSL. Because of its geographical location, the plain exhibits special meteorological features including long freezing winters coupled with relatively moderate summers. According to long-term stats acquired at Ardabil Synoptic Station (39-year data), the area has experienced a minimum, maximum, and average temperature of −33.8, 39.8, and 9.2 °C. Average relative humidity (RH) is 71% and average annual rainfall is reportedly 293.3 mm, of which only 7% falls in the summer when the agricultural water demand is maximal – a fact that characterizes the region as a semi-arid area. Located in the Ardabil Plain, the Abi-beyglu area is the earliest location through which the Caspian humidity front flows into the Ardabil Plain. It is dominated by rainfed agriculture. Information received from the Ardabil Agriculture Organization show that some 80% of the lands in this region are dedicated to rainfed farming of crops. Barley and lentils are the main dry-farmed agricultural products of the region.

The required meteorological data (including temperature, RH, wind speed and prevailing direction, evaporation, and rainfall) during 1997–2021 from the Abi-beyglu weather station and during 2003–2021 from the Ardabil Airport synoptic station (such as monthly foggy and sunny days) were prepared. Noteworthily, 3-h average values of the considered parameters over the research period (April to December 2021) were also retrieved from the Ardabil Airport Synoptic Station and used in the analyses.

A collector can never harvest the entire LWC of the flowing air. In order to evaluate the fraction of the passing flux that is absorbed by a fog water collector, one can use the collector efficiency (⁠⁠), which has been defined as the ratio of the collected water per unit surface area of the collector to the passing flux across the cross section.

Next, the effects of the collected water on increasing yield and water productivity of the wheat were assessed by AquaCrop software. The AquaCrop presents a simple yet powerful model that can be used for a wide spectrum of different weather conditions and crops, including vegetables and cereals. Being prepared based on the revised version of the FAO paper No. 33, this model simulates the plant growth and yield as functions of input data including meteorological parameters, soil characteristics, and applied managerial measures (i.e., mulching, irrigation, fertilization, etc.) (Steduto et al. 2009). The AquaCrop model has been used in many investigations around the world, with its ability to simulate the growth and yield of crops proven through multiple calibration and validation studies (Todorovic et al. 2009).

To simulate the AquaCrop model, irrigation data, soil characteristics, crop parameters, farm conditions and management parameters were used based on the calendar-day mode. Soil data were obtained from local information and daily measured data of temperature, humidity, wind speed and solar radiation at the Abi-beyglu meteorological station were used to determine daily evapotranspiration of the crop. The parameterized AquaCrop model was used to simulate soil water content, CC, biomass and grain yield for wheat.

Simultaneously with the measurement of the collected water, possible occurrence of fog was recorded by the operator. Investigating the data on fog occurrence at the site of collectors and controlling them against the data received from the Ardabil Airport Synoptic Station, it was figured out that the two sites are completely identical in terms of fog occurrence and meteorological properties. Knowing that the collected fog water is a function of meteorological factors, we began by analyzing the meteorological parameters affecting the fog water collection at the research site followed by investigating the relationship between the collected water and these parameters. Afterwards, water demand of a major rainfed crop in the region (namely wheat) was calculated and yield improvement was investigated under various scenarios involving complementary irrigation with the collected water.

The Abi-beyglu Weather Station is the closest of the kind to the research site, where considered meteorological parameters are measured and recorded. This station is located at 38°16′54″N, 48°33′29″E. Table 1 presents monthly average values of the meteorological parameters at this station. As observed, maximum percent rainfalls occur in May and November (14.4%), while the minimum occurs in August (2%). Moreover, the highest and lowest seasonal rainfalls occur in fall and summer (34.2 and 11.6%, respectively). The table further presents the number of foggy days at the Ardabil Airport Synoptic Station. Accordingly, the highest fog frequency was observed in September followed by October and then December (12.7, 11.7, and 11.6%, respectively), while the lowest fog frequency was seen in June followed by July and then August (5, 5.6, and 6.1%, respectively). On the other hand, seasonally speaking, the highest fog frequency (i.e., 48%) occurred in fall. According to the results, on average, the study area experiences a total of 145 foggy days a year (covering more than one-third of the year). Therefore, it can be said that the average number of foggy days that occurs during agricultural activities is 83 days. The fog frequency indicates the inflow of the humidity front from the Caspian Sea toward the study area clearly. It is worth noting that the fog frequency data are based on the stats recorded at the Ardabil Airport Synoptic Station, and the fact that the study area is where the humidity front from the Caspian Sea flows into the region coupled with local observations indicate that the fog frequency is higher at the research site rather than the Ardabil Airport Synoptic Station.

In order to compile wind information across the region, long-term data on wind speed and direction at the Ardabil Airport Synoptic Station was retrieved and analyzed.

We further plotted seasonal wind rose diagrams of the data at the Ardabil Airport Synoptic Station, which is not presented in this paper to keep it neat. Investigation of the seasonal wind rose diagrams showed that the east wind retains its prevalence across the region in all seasons. Indeed, the east wind comprises 95, 90, 60, and 55% of the wind flows in spring, summer, fall, and winter, respectively, making it the prevailing wind system in the region. On the other hand, the east winds exhibit significantly higher speeds in the summer rather than the spring, fall, and winter. In the winter, the south and southwest winds blow at higher speeds than the east winds. Considering what was mentioned above, maximum utility of the humidity from the Caspian Sea, as the main source of humidity in the region, occurs in spring and fall.

Quality of collected water depends on local environmental conditions, and possible contamination of the collector surface with dust, insects, and birds' droppings. Although the fog water was expected to be pure, of high quality, and free of any chemical contamination, qualitative assessment of the water was performed through chemical analysis of the collected fog water at Central Laboratory of University of Mohaghegh Ardabili and comparing the results to the standard levels set by WHO (WHO 1993). Table 3 shows a comparison of the qualitative parameters of the fog water. The collected water was found to be highly pure with very low concentrations of sulfate, nitrate, phosphate, sodium, calcium, magnesium, and potassium. Moreover, the very low TDS levels indicated the absence of any industrial activity in the vicinity of the research site.

As it blows to hit the collector surface, the flow of fog is affected by a number of factors including the meteorological factors. In order to investigate the meteorological factors affecting the water yield, correlational analyses were performed between different meteorological parameters and the water yield in SPSS 26. Table 4 reports the correlation coefficient of the water yield of the Raschel-mesh collectors to different meteorological parameters.

Investigation of the correlation coefficient of the extracted water to meteorological parameters show that, among different parameters, significant correlations were observed between the extracted water and visibility (Visibility), temperature (T), rainfall (Rainfall), RH, evaporation (Evaporation), fog time (Fog time), and temperature difference between the dew point and dry air (T – T_dew) at a significance level of 0.01. On the other hand, no significant correlation was observed between the extracted water and the dew point and wind speed and direction. Visibility, evaporation, and temperature difference between the dew point and dry air exhibited negative correlations to the water yield, while rainfall, RH, and fog time were found to be positively correlated to the water yield. Of the pool of different parameters, the rainfall and evaporation exhibited the highest and lowest significant correlations to the water yield, respectively. An important point to note was the insignificance of the correlation of the wind speed and direction to the extracted water. As we know, wind is a crucial parameter affecting the fog water collection yield, with an expectedly direct association between them. However, as is seen from the results, they exhibit no significant correlation. The same outcome has been reported by Ritter et al. (2015) and Carrera-Villacrés et al. (2020). With increasing wind speed up to about 4 m/s, the collection efficiency increases non-linearly, but at higher speeds, the collection efficiency does not change significantly (Regalado & Ritter 2016). In this research, in more than 77% of the conditions, the wind has occurred with a speed higher than 4 m/s (Figure 2), which does not have a significant effect on the water collection efficiency. This issue was also identified in the statistical correlation between wind speed and the amount of collected water.

According to Figure 5(a), for the most part, fog is formed at relative humidities above 90%, which shows that the air is nearly saturated with humidity when the fog forms. Visibility is another parameter affecting the fog water collection yield. That is, as the visibility decreases, LWC increases. Increased LWC because of increased flux of fog per unit area results in increased fog water collection yield. Figure 5(b) shows the relationship between these two parameters. As observed, water collection by the collector occurs when the horizontal visibility is below 10,000 m. Of course, it should be noted that a fog-free rainfall may not reduce the horizontal visibility significantly but introduces some error into the measurement data as the rainfall and fog contributions were not separated in this study. Wind direction with respect to the fog collector affects the water yield. As mentioned earlier, the collectors were installed in the direction of the prevailing wind in the region. According to Figure 5(c), when the wind direction was perpendicular to the surface of the collector (or the deviation angle was small), the water yield was relatively high. In the cases where the water yield was high despite a large deviation between the wind direction and the collector direction, the main contributor to the collected water was the rainfall.

At the weather station, fog time was recorded once every three hours. Accordingly, the corresponding plot (Figure 5(d)) demonstrates the collected water versus fog time in multiples of 3 h. Except for a few exceptions (corresponding to simultaneous occurrence of fog and rainfall), higher water yields were observed with longer fog times. The fog time affects the water yield directly. Obviously, the longer the fog time the longer the available time for water collection, leading to a higher water yield.

Considering potential occurrence of the fog in the region and the achieved average water yield by the Raschel water collectors, supplying water for complementary irrigation of the wheat was feasible. This was done under two irrigation scenarios. In scenario I, two complementary irrigations were applied to supply the wheat with 60 mm of water during the flowering stage in dry condition. In scenario II, however, one complementary irrigation was applied to supply the wheat with 30 mm of water during the flowering stage in normal condition. Complementary irrigation of the wheat with 30 and 60 mm of water in normal- and dry-condition scenarios could supply the crop with 26 and 34% of its water deficiency, respectively. According to the average collected water (3.6 L/m²/day), by installing 36 LFCs (with dimensions of 48 m²), the water needed for supplementary irrigation in the dry scenario can be supplied. Of course, as mentioned earlier, by installing the collectors at higher altitudes, the collection efficiency can be increased up to 19 times. Nevertheless, assuming a 5-fold increase in water collection efficiency, the number of LFCs required will be 7.

In order to investigate the effect of complementary irrigation in the two mentioned scenarios on the yield and yield components of wheat, AquaCrop model was utilized. Calibration of the AquaCrop model was conducted on the basis of the data collected from the region, in-person interviews with farming experts, and the data extracted from test farms, indicating that this model can well simulate the wheat yield and yield components. The conservative and non-conservative parameters of the crop are shown in Table 5.

Table 6 lists the parameters related to wheat production under the two scenarios. Dry yield increased by 1.7 and 0.6 ton/ha under the dry and normal scenarios, respectively, highlighting the positive effects of complementary irrigation on wheat production in the region. This improvement was also evident in water productivity levels, which increased by 0.38 and 0.08 kg/m³ under the dry and normal scenarios, respectively.

The potential of fog water collection varies depending on the meteorological parameters and the design of the harvesting system. Meteorological parameters cannot be controlled. Therefore, for good design to maximize collection capabilities, it is necessary to measure them more accurately. Air temperature and RH can be mentioned among the meteorological parameters effective in the formation of fog. As the RH increases and the temperature decreases to the dew point temperature, the water vapor in the atmosphere condenses and fog is formed. The number of foggy days indicates the suitable conditions of temperature and RH in the region. In addition to the number of foggy days, wind speed and its direction are also considered key parameters in water collection. So that if there is no wind, the fog drops will not be transported. If the wind speed is not enough, the collection efficiency will also decrease.

The wind increases the collected water in two ways: (1) It causes the transfer of suspended fog drops toward the collector. Therefore, as the wind speed increases, the fog flux entering the collector also increases, and according to Equation (1), the amount of water collected also increases. (2) It causes the fog droplets to collide with the collecting plate, and as a result, with the condensation of the droplets, water collection is done according to the main collection mechanism (inertia mechanism).

Although the design and construction of LFC type collectors is easier and less expensive than multi-faceted collectors, but in order to maximize the efficiency of water extraction, this type of collector must be placed perpendicular to the wind direction. After installing the collector perpendicular to the wind direction, the relative stability of the wind direction is important in increasing the water extraction efficiency, and its violation will introduce uncertainties in the results. The predominant wind direction of the studied area is from the east side, which can carry the moisture from the Caspian Sea to the studied plain (Figure 2).

Air quality and particles that get stuck on the collecting net affect the quality of the collected water. Fog is known as a solvent for bicarbonates, calcium, sodium chloride, heavy metals, nitrates, organic carbon and other ions. Depending on the concentration of dissolved chemical species in the air, the quality of collected water may be affected.

The implementation of fog water harvesting projects in a region can be aimed at providing drinking water or agricultural uses. But the most important thing is to consider the potential of that area in supplying the required amount of water. Of course, it should be noted that the fog water collected cannot meet all the needs of a populous society or the entire needs of the agricultural sector.

The fog collector system has been of interest especially in poor communities due to its simplicity in design, construction and operation, as well as no need for energy, with a relatively low cost. Investigations carried out in the studied area showed that all the necessary equipment such as collecting metal structure, polyethylene pipe and polypropylene mesh can be easily obtained from urban stores in this area. According to the estimates, the cost of building an LFC collector is around 50 million Rials (approximate 100 US $).

The data on the average water collected in different parts of the world (Table 1) show that the amount of extracted water varies in different regions. This is despite the fact that none of these projects were in the same conditions, such as the height of the collector installation, the type and dimensions of the collector, the measurement period, and the weather conditions. The average amount of water collected mentioned in this research (3.6 L/m²/day) is only related to the period with agricultural activities. Despite the mentioned conditions, in case of using LFC with dimensions of 40 m² and supplementary irrigation of 60 mm, 35 LFC devices will be needed for 1 ha of wheat field. This number of collectors is acceptable compared to the projects implemented in Chile (with 50 LFC devices) to meet the water needs of a small village.

It goes without saying that if the project is implemented on the ridge, the efficiency of water collection can increase up to 19 times (Ritter et al. 2008). Assuming a 5-fold increase in efficiency in high places, the number of LFCs required will also decrease to 8. Local observations showed that fog events occur at higher altitudes much more often than at lower altitudes. In addition, high wind speed and the absence of obstacles compared to low points are among the advantages of installing collectors on the ridge.

As a result, considering that agriculture is the main source of income for the people of this region, and considering the potential of this region for the implementation of water extraction projects from fog, investing in the implementation of such projects will minimize the damage caused by the decrease in rainfall and lack of irrigation water in the agricultural sector of the region. It should be mentioned that the income from the increase in the product can compensate the cost spent for the implementation of the project.

On average, the SFCs installed at the Abi-beyglu Area could collect fog water at 3.6 L/m²/day, which is comparable to the results of similar studies around the world. According to Table 1, average water yield of the SFC has been 6 L/m²/day, with a maximum level of 30 L/m²/day in Oman (Schemenauer & Cereceda 1994b) and a minimum level of 0.16 L/m²/day in Chile (Carvajal et al. 2022). It is obvious that the amount of extracted water was different from one place to another. Quantities of extracted water in Chile (Schemenauer & Cereceda 1994b; Carvajal et al. 2022), South Africa (Olivier 2002; Olivier & De Rautenbach 2002), Namibia (Shanyengana et al. 2002), United States (Ruiz 2005), Colombia (Molina & Escobar 2008; Escobar et al. 2010), Saudi Arabia (Al-Hassan 2009; Gandhidasan & Abualhamayel 2012) and Ecuador (Carrera-Villacrés et al. 2017) were reported less than the average, all of which used the SFC type collector. On the other hand, the water extraction values in Peru (Schemenauer & Cereceda 1994b), Oman (Schemenauer & Cereceda 1995), Nepal (Mac Quarrie et al. 2001), Spain (Marzol 2008), and Morocco (Marzol & Sánchez 2008) with the same collectors have been obtained higher than the average. There are many reasons that cause differences in the amount of extraction water, including the type and height of the collector installation and meteorological factors such as wind speed and direction.

The fact that average water yield in the study area was below the global average could be attributed to suboptimal location of the fog collectors, which were slightly deviated from the highlands over which the fog front blows into the Ardabil Plain. The collectors were installed at an elevation of 1,340 m, while the mentioned highlands are estimated to be as high as 1,700 m. Accordingly, should the collectors be installed at higher elevations, one may expect higher water yields. Another point to note is the fog time. According to historical data, the study area is foggy for an annual average of 145 days (i.e., more than one-third of a year). Therefore, the relatively long fog time in the study area highlights its large potentials for fog water collection.

In order to evaluate the potential of fog water collection in Abi-beyglu, northwestern Iran, a project was implemented where three fog water collectors were installed, and measurements were performed during April–December 2021. Accordingly, the volume of water collected in bilayer Raschel-mesh collectors was measured at different days. The results were then compared to similar projects and their correlations with meteorological parameters were assessed. Based on the results of this research, average fog water collection yield during the fog time (3.6 L/m²/day) was acceptable in comparison to similar projects in other countries. The relatively high number of foggy days in the study area (i.e., more one-third of a year) and distribution of the foggy days in different seasons provide the required basis for increased collection of fog water, which can be considered as an alternative water resource to supply local demand for drinking water and complementary irrigation of rainfed crops. Investigation of the contribution of collected water to supply water demand for wheat farming in the region showed that the collected fog water represents a proper resource for supplying a part of the water demand for dry-farming in the study area. This can be done through one or two complementary irrigations in sensitive growth stage(s) (e.g., flowering stage). Application of the proposed strategy on dry-farming of wheat in the study area brought about increased crop yield and water productivity.

This research was supported by the Vice-Chancellor's Office for Research of University of Mohaghegh Ardabili, contract number 99-d-9-19958.The authors also acknowledge Ardabil Regional Water Company for making available the meteorological data set used in this study.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.