Estimation of water consumption volume and water efficiency in irrigated and rainfed agriculture based on the WaPOR database in Iran

Predictions related to surface warming and changes in precipitation indicate that these changes significantly reduce agricultural productivity in most Asian countries, leading to heat stress, severe droughts, and floods (Sileshi et al. 2023). Water demand for agriculture, particularly for irrigation, appears to be more sensitive to climate change than industrial and urban sectors (Kirby & Mainuddin 2022). This is because agriculture is susceptible to precipitation and temperature at the field level. Therefore, climate change can significantly impact water requirements and irrigation planning in agriculture (Lu et al. 2020). Recurring droughts increase the water demand, but depending on the amount of precipitation and its impact on soil moisture, it is possible for water availability to decrease due to climate change as well (Corwin 2021). The existing climate change scenario indicates that climate change affects agricultural crop growth and water consumption. This factor is important for long-term planning regarding supply and demand management (Anh et al. 2023). Therefore, climate change will affect not only the minimum and maximum temperatures but also the magnitude or intensity of these variables and other meteorological variables (Yoon et al. 2022). Estimating the amount of water reaching agricultural uses in irrigation schemes is done based on estimating and determining evapotranspiration rates (Li 2020). According to water accounting definitions, actual evapotranspiration refers to a portion of the water withdrawn for the agricultural sector that is beneficially consumed, irretrievably lost, and unavailable for reuse (Karimi et al. 2012). A research study found that CHIRPS satellite data are the best option for quantifying water consumption in the Betwa River basin, India, with cereals being identified as the largest water-consuming land use (Singh et al. 2022b). The WA+ framework can be successfully applied to analyze irrigated and rainfed crops' water consumption patterns, land productivity, and water productivity (WP) individually (Singh et al. 2022a). Evapotranspiration consists of two components: evaporation and transpiration (Liu et al. 2022). The evaporation component of evapotranspiration can be reduced close to zero or converted to transpiration by applying various techniques. Non-beneficial water consumption in agriculture includes a portion of the water withdrawn that, in addition to being used for agricultural purposes, is not utilized for crop production but is lost through deep percolation or surface runoff, becoming inaccessible (Jovanovic et al. 2020). A portion of non-beneficial water consumption can be reclaimed from groundwater sources or surface runoff and used beneficially. Some of it may join the underground aquifers of freshwater or saline water sources and become inaccessible. Therefore, water conservation can occur by reducing evapotranspiration or diverting water to deep aquifers or saline water sources (Molden et al. 2001). Actual evapotranspiration (ETa) is an important component in determining the water balance in any given region, which depends on factors such as solar energy availability, leaf area index (LAI), soil moisture, and meteorological parameters (Song et al. 2022). Numerous studies have been conducted to measure actual evapotranspiration (ETa) at the field level. In most of these studies, ETa is indirectly estimated by considering crop coefficients and growth stages based on specific calculation standards. However, this method does not consider diseases or nutrient status (Allen et al. 1998a). Direct measurement of actual evapotranspiration requires high costs and facilities (Zheng et al. 2020). In recent decades, various methods have been developed to estimate evapotranspiration using land surface and crop models at regional to global scales. Ground observations are expected to provide continuous spatiotemporal information about surface variables and parameters essential for large-scale evapotranspiration estimates (Pan et al. 2020). Actual evapotranspiration (ETa) is determined through surface energy balance and calculation of physical variables in a given region. Therefore, various remote sensing algorithms have been developed to estimate actual evapotranspiration (ETa) (Mul & Bastiaanssen 2019). These algorithms include SEBS, SEBAL, METRIC, ETLooK, and Py_SEBAL (Fakhar & Kaviani 2022a). Each of these algorithms calculates the evapotranspiration rate with different levels of accuracy. Energy balance models primarily depend on land surface temperature. These models provide information about energy fluxes in the regions (Gowda et al. 2008). Most energy balance models, such as METRIC, require thermal infrared band imagery derived from cloud-free and atmospherically corrected images for generating land surface temperature maps (Jia et al. 2009). Therefore, cloudy images can lead to a reduction in the accuracy of the results. However, the ETLooK algorithm utilizes soil moisture extracted from passive microwave sensors instead of land surface temperature. Microwave data can provide surface information even on cloudy days since they are less affected by cloud cover (Tsang et al. 1977). In recent years, the FAO has developed open access to water efficiency systems in regions such as Africa and the Middle East, facing water crises (Mul & Bastiaanssen 2019). Estimating actual evapotranspiration from WaPOR is one of the most important products of this system. The WaPOR database provides 10-day and annual spatial maps with a pixel size of 250 m, generated using the ETLook algorithm, covering the years from 2009 to the present. The equations and calculation methods for estimating evapotranspiration are extensively described in research conducted by Bastiaanssen (Bastiaanssen et al. 2012). This global model is designed for automated processing, as it considers soil moisture as one layer of input data (Mul & Bastiaanssen 2019). Various research studies have been conducted in this field, including the study by (Barideh et al. 2022), which focused on estimating actual evapotranspiration using the global WaPOR model and comparing it with the Penman–Monteith equation in the Urmia Lake basin during the years 2010–2020. The results demonstrated the high accuracy of this global model compared to the data obtained from the equation. Considering the availability of continuous data in this model and its global coverage, a study conducted by (Blatchford et al. 2020) evaluated the evapotranspiration product of WaPOR at 14 meteorological stations in irrigated and rainfed areas of Africa. The results indicated a general correlation of 0.71 between the actual evapotranspiration derived from the WaPOR model at the point scale and the station information obtained. The average root mean square error (RMSE) was 1.2 mm/day. Indeed, these results can be promising, with WaPOR being an open-access dataset that provides near-real-time continent-wide values. The WaPOR database consists of several data components related to WP, biomass production, evapotranspiration, land cover, and ancillary data layers (FAO 2020). In a study conducted by Hajirad et al. (2023), the investigation of evapotranspiration variations using the WaPOR database in the Maroon-Jarahi basin in 2017 was carried out. The analysis of the obtained results revealed that cropping patterns and irrigation system types are highly influential factors in assessing WP. It offers valuable information for researchers to assess and manage water consumption in various regions, including agriculture. Indeed, a lack of information about actual water consumption in the agricultural sector can lead to inefficient policies. Understanding the barriers to optimal water consumption in agriculture and the impact of each influential indicator can significantly affect current and, especially, future planning. This research aims to investigate the values of evapotranspiration from the WaPOR model and WP in various climates of Iran using this global model from 2009 to 2022.

The northern and northwestern areas of the country typically experience temperatures below freezing during winter, while they have relatively humid weather in other seasons throughout the year (Shahabfar & Eitzinger 2013). The average rainfall in the northern parts of the Alborz mountain range is over 1,000 mm. The Alborz and Zagros mountains receive rainfall ranging from 600 to 800 mm and 500 to 600 mm, respectively. However, the central regions experience a significant decrease in precipitation, which has led to large areas, such as the Dasht-e Kavir and Dasht-e Lut, being classified as dry and desert lands based on land use maps due to aridity and water scarcity conditions (Figure 1(d)). Some regions in central Iran have not received significant rainfall for several years (Rahimi et al. 2013). According to reports, due to inefficient policies during the drought conditions in 2001, Iran lost 520 million dollars in agricultural production (Iranian Ministry of Energy 2018).

In which λ represents the latent heat of evaporation (J/kg), ET denotes evapotranspiration (kg/m²s), R_n is the net radiation (W/m²), G represents soil heat flux (W/m²), ρ_a represents air density (kg/m³), C_p denotes specific heat of dry air (J/kg·K), e_a represents actual vapor pressure (Pa), e_s represents saturation vapor pressure (Pa) which is a function of air temperature, Δ represents the slope of the saturation vapor pressure curve concerning temperature (Pa/K), r_a represents aerodynamic resistance (s/m), and r_s means surface resistance (s/m).

The two equations differ concerning the net available radiation (R_n, _soil, and R_n, _canopy) and the aerodynamic and surface resistance (r_n, _soil, r_a, _soil, r_n, _awning, r_a, _canopy). Furthermore, the soil heat flux (G) is not considered for transpiration. The other parameters of the equation are not considered further, as these are constants or variables that can be derived directly from mathematical relationships.

The soil is coupled with the underground soil moisture content or the root zone for transpiration. However, this relationship is established through the surface soil moisture content for evaporation. Water retained by vegetation cover is a process in which rainfall is intercepted by the leaves, preventing it from reaching the ground, and it evaporates directly from the leaves using energy that is not available for transpiration. A brief description of the sensors used in the WaPOR platform is mentioned in Table 1.

GBWP represents the total biomass production in kg/ha, E is evaporation, T is transpiration, and I is water retained by vegetation cover. All parameters in the denominator are measured in millimeters. Unlike GBWP, NBWP is particularly useful in monitoring the effective utilization of water by vegetation cover for beneficial biomass development (FAO 2020).

In which ET₀ represents reference evapotranspiration (mm/day), Rn denotes net radiation at the crop surface (MJ/m²/day). G represents soil heat flux density (MJ/m²/day). T represents the average daily air temperature at 2 m height (°C). u₂ represents the average daily wind speed at 2 m height (m/s). e_s represents the saturation vapor pressure (kPa). e_a represents the actual vapor pressure (kPa). (e_s–e_a) represents the vapor pressure deficit (kPa). Δ denotes the slope of the vapor pressure curve (kPa/°C). γ represents the psychrometric constant (kPa/°C). To assess the accuracy of the WaPOR model, the obtained data were compared with the values derived from the FAO-56 method.

In this study, an attempt was made to compare the WaPOR model with the FAO-56 method. The FAO-56 method, used by the United Nations to estimate evapotranspiration, is recognized as a standard and reliable approach. On the other hand, the WaPOR model, developed by FAO, is an advanced model based on remote sensing data. Considering the vastness of Iran and the impracticality of comparing evapotranspiration for all provinces, 16 provinces from four different climatic regions were initially examined. The classification details of the provinces based on climate are indicated in Table 2. To this end, using meteorological data, 10-day evapotranspiration estimates were obtained over one year using the FAO-56 method for four climatic regions in Iran. A numerical comparison was then made between the values obtained from the FAO-56 method and those derived from the WaPOR model.

Moreover, the analysis of evaporation levels during the study period indicates that evaporation has fluctuated over the past 14 years. The magnitude of these fluctuations was relatively lower from 2009 to 2015, characterized by a gentler slope of changes. However, during the years 2015–2022, the intensity of changes has increased. These changes are particularly noticeable in the southern regions of the Zagros foothills and the coastal strip along the Caspian Sea, where a higher percentage of changes is observed.

The spatial distribution of actual transpiration across the country from 2009 to 2022 is depicted in Figure 4(b). The highest transpiration has occurred on the northern coasts, especially in the eastern Mazandaran and central regions of Golestan, which can be attributed to the region's climate and land use. In the significant areas of Golestan, which have a semi-arid climate, transpiration levels were close to 2,000 mm from 2009 to 2015. In contrast, in the provinces of Gilan and Mazandaran, with a humid environment, transpiration levels reached around 1,500 mm in 2016. However, from 2017 to 2022, actual transpiration gradually decreased. In 2022, transpiration levels dropped to less than 1,000 mm in the eastern regions of Mazandaran and the central parts of Golestan. Climate change in these areas can be considered as one of the reasons for the decrease in transpiration. As temperatures increase, evaporation rates from the surface also increase. Therefore, one contributing factor to the reduction in transpiration is rising temperatures. Another factor is the decrease in precipitation in these regions, as reduced rainfall reduces the available water in the soil and water sources, consequently resulting in decreased transpiration in these areas.

The conducted studies in the country's central regions also indicate that in a vast area of the east and central parts of the country, transpiration rates have been reported to be between zero and less than 10 mm. Figure 6(b), analyzing the annual average transpiration in different land areas, reveals that water is artificially supplied to the land in areas under irrigation network coverage. Consequently, transpiration rates are higher in these areas. When plants have abundant water, they can grow and transpire more water. Additionally, water is directly supplied to the plant roots in irrigated regions, increasing plant water consumption and enhancing transpiration rates.

Moreover, water is usually provided in larger quantities and more consistently in irrigated areas, ensuring a continuous water supply to the plants. This factor always allows plants access to water, leading to increased growth and higher transpiration rates. Overall, when considering the average transpiration across the entire country, the climatic diversity, variation in elevations, and plant diversity contribute to the similarity between the average transpiration of the whole country and the irrigated lands (Figure 6(b)).

Evapotranspiration is the process by which water molecules change from liquid to vapor from the surface of evaporation and plant stomata. This value includes transpiration from leaf stomata (ET_r) and evaporation from the soil and other moist surfaces (ET_a). The agricultural sector is recognized as the most significant water consumer. Therefore, considering the limited water resources, improving water efficiency based on water demand is necessary. In the WaPOR model, this feature is incorporated by estimating actual and reference evapotranspiration values daily, monthly, and yearly. The evapotranspiration rate also depends on the soil moisture percentage, and if water is available in sufficient quantities, evapotranspiration occurs at its maximum potential (Song et al. 2019). The difference between reference and actual evapotranspiration lies in the fact that reference evapotranspiration represents the amount of water a plant would consume if sufficient water were available. This value is defined based on specific conditions such as temperature, relative humidity, and solar radiation. On the other hand, evapotranspiration represents the amount of water a plant naturally consumes, which includes reference evapotranspiration and additional factors specific to the plant and its surroundings.

Moreover, the intensity of precipitation reduction and climate change has been increasing in recent years. Consequently, farmers rely more on groundwater sources to meet their water needs. The excessive and unsustainable use of these valuable resources has led to a severe decline in groundwater levels in recent years. Therefore, the area under irrigation network coverage has significantly decreased due to limited water resources in these areas. Regions with an extra arid climate also show a high reduction proportion, with a 45% decrease in the area under irrigation network coverage. The trend of changes in land under irrigation network coverage in the semi-arid and humid climates shows a reduction of 7 and 2%, respectively. In the analysis of drylands, it is also observed that the highest level of reduction occurs in semi-arid lands. This can be attributed to changes in precipitation patterns for crop growth in these areas, leading to decreased cultivated area and crop yield. They consider crop growth sensitive to spatial and temporal factors such as climate, rainfall, and temperature (Armstrong et al. 2009). The environment can significantly influence crop performance, especially in dry conditions where spatial variability in soil structure can impact moisture scarcity and have a notable effect on plant growth (Milewski et al. 2022). In humid and arid regions, the area under dryland coverage has increased by 3 and 33%, respectively. A specific study found that lands once covered in vegetation in South Punjab had been transformed into urban areas over a span of 20 years. This was due to the expansion of infrastructure and increased commercialization (Hu et al. 2023). Overall, the analysis of changes in the drylands area also indicates that regions with a humid and very arid climate have experienced insignificant percentage changes from 2009 to 2022. The reason for this could be the stability of natural conditions in these areas.

On the other hand, in rainfed lands, due to the increased risk of plants being subjected to water stress conditions, optimal conditions for growth and production are not provided. Additionally, fertilizers and pesticides are usually not used in rainfed cultivation due to the absence of an irrigation schedule during the crop cultivation period. Therefore, this factor can be another reason for reducing biomass volume in rainfed lands. During the years under study, it is observed that the biomass variations do not follow a specific increasing or decreasing trend. From 2009 to 2013, the biomass volume showed a relatively growing trend, with a 13.8% increase in biomass variations in 2013 compared to 2009. The increase in the area under cultivation in irrigated lands may be the reason behind this increase. In the analysis of biomass levels in dry lands, it is also noted that the amount of biomass produced in these lands is highly dependent on the rainfall received during the plant growth season. Therefore, in areas with limited rainfall, interseasonal irrigation in dry lands can increase biomass. This is because the use of intermittent irrigation can help preserve soil moisture levels, which can lead to improved plant growth.

In this research, evapotranspiration values were estimated using the WaPOR model in 16 provinces located in four climatic regions. The analysis of the obtained outputs showed that the WaPOR model had the highest correlation, with R² = 0.91, in the semi-arid climate. In addition to examining and comparing evapotranspiration values, this study also focused on parameters such as biomass and other factors related to water management in agricultural lands. The correlation between the accuracy of biomass estimation and actual evapotranspiration in different climatic regions of Iran was also investigated in this study. The results indicated that the highest correlation between evapotranspiration, biomass, and actual evapotranspiration was observed in the semi-arid areas, with a correlation coefficient of R² = 0.91. Meanwhile, the lowest correlation, with a coefficient of R² = 0.47, was observed in humid regions. This difference may be due to the presence of cloud cover in humid areas, which could affect the accuracy of the WaPOR model in those regions. Furthermore, the analysis of biomass variations based on WaPOR model data showed a relatively increasing trend from 2009 to 2013. The biomass level in 2013 compared to 2009 increased by 13.8%, which could be attributed to the expansion of irrigated lands. On the other hand, from 2016 to 2020, despite no increase in the cultivated area under irrigation, the average biomass production in Iran as a whole showed a slight increase. This increase can be attributed to factors such as the modernization of irrigation equipment and optimal management practices in agricultural farms. The analysis of evapotranspiration changes indicated that since 2017, the average evapotranspiration in areas under irrigation coverage in the country has shown a relative increase compared to 2009 and 2010. Changes in economic patterns, food security, and a focus on increasing income in agriculture may be reasons for this increase, leading to higher water consumption and evapotranspiration. Therefore, the potential for reducing evapotranspiration in irrigated and rainfed agricultural lands can enhance the importance of investing in absolute water consumption reduction in the agricultural sector. Such investment helps save water, preserve water resources, improve the quality of agricultural products, protect the environment, and increase agricultural production efficiency.

• 1. Future research could focus on assessing the effectiveness of different water management strategies in enhancing water efficiency and agricultural productivity, building upon the findings of this study. This assessment could involve evaluating the impact of various irrigation techniques, cropping patterns, and water conservation measures on water consumption and agricultural productivity levels.

• 2. Given the expected consequences of climate change on agriculture, future research should prioritize developing adaptation strategies and policies aimed at mitigating adverse effects and bolstering the resilience of the agricultural sector. This could involve conducting studies to explore the potential benefits of climate-resilient crop varieties, precision irrigation technologies, and climate-smart agricultural practices. By focusing on these areas, researchers can contribute to formulating comprehensive and practical approaches to address the challenges posed by climate change in agriculture.

• 3. Furthermore, future research could incorporate economic analysis to assess the cost-effectiveness of different water management interventions and evaluate water consumption and efficiency. This financial analysis would provide valuable insights for policymakers and stakeholders, helping them make informed decisions regarding resource allocation and investments in the agricultural sector. By considering the economic aspect, researchers can contribute to developing sustainable and financially viable water management strategies.

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

The authors declare there is no conflict.