1 Introduction

In recent years, changes in precipitation, temperature, and evaporation, which are key climatic variables, have had significant negative effects on agriculture due to climate change and global warming. Among the various water consumers, agriculture is very vulnerable to effects of climate change. Growing population, increasing demand, and diet change increase freshwater consumption in the future through irrigation agriculture to produce more food (Foley et al. 2011; UNESCO 2012; Wada and Bierkens 2014). Although, developing irrigation to increase food production has serious effects on the environment because of the limited freshwater availability (Reilly and Schimmelpfennig 1999). According to the Food and Agriculture Organization of the United Nations (FAO) (1999) and Postel (1999), the irrigated areas have increased as much as six times over the word. Similarly, 40 % of the world’s food is produced from irrigated farms. On the other hand, the irrigated area has expanded 1 % per year, and based on Cai and Rosegrant (2002), the demand for irrigation will be increased almost 14 % by 2025. In dealing with areas with limited water resources area, decline in food productions has a direct effect on human life and increases poverty in the local scale (Olesen and Bindi 2002). Accordingly, increasing water security is very crucial in such areas, which is one of the sufficient ways of water management and sustainable development. In order to develop water management policies and sustainability of agriculture, it is essential to assess the impacts of climate change on agricultural water requirement.

Recently, many investigations have been carried out in this sector (Garrote et al. 2015; Multsch et al. 2015 and Riediger et al. 2016). A study by Elgaali et al. (2007) showed climate change effects on irrigation water demand in southeastern Colorado. They found an increase in irrigation water demands under climate change. Rodríguez Díaz et al. (2007) studied the climate change impact on water requirement in Guadalquivir river basin in Spain. According to their report, there was a 15–20 % increase in irrigation water requirement in 2050. Using the decision support system for agro-technology transfer (DSSAT)-Canegro model, Knox et al. (2010) reported that climate change in the 2050s could increase sugarcane irrigation water requirements in Swaziland more than 9 %, and sucrose yields about 15 %. A study by Trnka et al. (2011) showed that there will be increased irrigation requirement due to climate change in the Central Mediterranean and the Iberian Peninsula for both winter and spring crops. Shahid (2011) analyzed the impacts of climate change on irrigation water demand of dry-season Boro rice in northwest Bangladesh. The results indicated that there will be no significant changes in total irrigation water requirement under climate change, while there will be an increased daily use of water for irrigation. Savé et al. (2012) studied irrigation water demand under climate change in a Spanish basin. They reported an increased irrigation water requirement related to the crop pattern in the study area between 40 and 250 % at the end of the twenty-first century. Also, Gondim et al. (2012) reported future increase in irrigation water requirement under climate change in the Jaguaribe River basin, Brazil. In addition, Rehana and Mujumdar (2013) reported an increased irrigation requirement by investigating the regional impacts of climate change on irrigation water demands for paddy, sugarcane, permanent gardens, and semidry crops over the command area of Bhadra reservoir, India. Elliott et al. (2014) found that the CO2 effects on crop growth and transpiration are a significant source of uncertainty for the assessment of climate change impacts on crop yields. In another study, Lee and Huang (2014) studied changes in irrigation demand under climate change for the future period (2046–2065) and compared the results to the base period (2004–2011) in north of Taiwan. They reported some increases in temperature and precipitation, which consequently led to changes in effective rainfall and crop requirement. Their study suggested an insignificant difference (<2.5 %) between current and future irrigation water requirements. Ashofteh et al. (2014) investigated the effects of climate change on agricultural water demand by considering risk analysis in East Azerbaijan, Iran. The results of this study indicated that risk of changes in crop water requirement increased approximately 3, 17, and 33 % for 25, 50, and 75 % risk levels, respectively. Valverde et al. (2015) examined the effects of climate change on irrigated agriculture in Guadiana river basin, in the south of Portugal. The results showed that there was an increase in crop irrigation requirement, particularly for maize, pasture, and orchards. Woznicki et al. (2015) studied irrigation demand under climate change by considering uncertainty and developing adaptation scenario in Kalamazoo River basin, southwest Michigan, USA. The results indicated an uncertainty in irrigation water requirement; also, a decrease in demand has been reported as a result of the delayed planting.

In addition to the importance of the climate change impacts on water resources and agriculture, it is essential to consider the uncertainty existing in this field. In the studies of the climate change, the general circulation model (GCM) output is a major source of uncertainty (Hawkins and Sutton 2009). Among the reasons behind this uncertainty, one can mention the lack of adequate knowledge in climate systems and computational limits in simulation of physical processes on small scales (Williams and Tselioudis 2007; Knutti and Hegerl 2008; Greasby and Sain 2011). In using the GCM output in the related studies, there are two general approaches including single GCMs (Jones and Thornton 2003; Guo et al. 2010; Teixeira et al. 2013; Ahmadi et al. 2015) and multi-model projection (Özdoğan 2011; Gohari et al. 2014). According to Lee et al. (2011) and Gohari et al. (2013), the first approach may cause an error in the estimation of climatic parameters. On the other hand, combined use of several climatic models helps to estimate a more reliable range of climate changes (Kloster et al. 2010). Accordingly, a multi-model ensemble of GCM outputs has been used in a large number of studies related to the effects of climate change on water resources and agriculture (e.g., Raje and Mujumdar 2010; Joyce et al. 2011; Zareian et al. 2014). Weighting method is one of the multi-model approaches that used to manage the uncertainty of GCM output (Tebaldi and Knutti 2007; Daccache et al. 2011). One type of weighting methods is the one based on the performance of the models in simulation of observed climatic parameters (Greene et al. 2006; Cai et al. 2009; Lee and Wang 2014), which is also based on a risk approach in this study.

In the present study, the effects of climate change on agricultural water requirement of network irrigation have been analyzed in Ramhormoz plain. A risk framework is used based on a multi-model ensemble of GCM output for a more sufficient management and better structuring of uncertainties. In this study, the full applied methodology covers two main stages: (a) risk assessment of climate change and (b) agricultural water demand calculation. The first stage includes (i) production of climate scenarios using 14 GCMs under two emission scenarios of A2 and B1; (ii) use of mean observed temperature-precipitation (MOTP) method for weighting GCMs, generation of discrete probability distribution functions (PDFs), developing cumulative distribution functions (CDFs), and then extracting probability percentiles (25, 50, and 75 %); and (iii) stochastic downscaling by using combined Long Ashton Research Station-Weather Generator (LARS-WG) and change factor (CF).

2 Material and methods

2.1 Case study

The study area is the irrigation network of Ramhormoz plain, which is about 22,470 ha located in southwest of Iran. In this plain, the average annual precipitation is about 310 mm, and the average annual temperature is about 26 °C. Recently, the Jarreh reservoir, with 261-MCM volume, has been constructed in outlet of Zard river basin to provide the agricultural water requirement in this plain. Zard river basin has an elevation of 324 to 3295 m above sea level, the length of main channel is about 70 km, and average slope is about 3 %. Agriculture is considered as a major and important job in the study area. The main crop pattern of this plain is wheat, corn, barley, and alfalfa. The crop pattern of plants and their area under cultivation are shown in Table 1, and Fig. 1 shows the location of the study area in Iran.

Table 1 Crop pattern and area under cultivation of each crop in the Ramhormoz network irrigation
Full size table
Fig. 1
figure 1

Location of the study area in Iran

Full size image

2.2 Assessment of climate change

2.2.1 Production of climate scenarios

As prominent models for simulating the effects of increasing greenhouse gas concentration on global climate system, GCMs are mathematical models that attempt to describe the full three-dimensional geometry of Earth’s climate system. Table 2 describes 14 GCMs in detail, which are employed in this study under two emission scenarios (A2 and B1), extracted from the Fourth Assessment Report (AR4) of Intergovernmental Panel on Climate Change (IPCC). Describing a very heterogeneous world, A2 storyline and scenario assume a world with a continuously increasing global population and a regionally oriented economic growth that is more fragmented and slower than the other emission scenarios. The B1 storyline and scenario assume a convergent world with a global population that reaches to a peak in mid-century and thereafter declines. It assumes rapid changes in structures of economic toward a service and information economy with reductions in intensity of material and the introduction of clean and resource-efficient technologies (IPCC 2007).

Table 2 Description of the GCMs from IPCC’s Fourth Assessment Report (AR4)
Full size table

2.2.2 MOTP method and probabilistic assessment

In this step, after extracting monthly climatic variables from GCMs under A2 and B1 scenarios, difference and relativity of long-term monthly average values of temperature and precipitation for the base period (1971–2000) were compared with the future period (2025–2054) using the following structure:

$$ \varDelta {T}_i=\left({\overline{T}}_{GCM,fu{t}_i}\kern0.5em -{\overline{T}}_{GCM,bas{e}_i}\right) $$
(1)
$$ \varDelta {P}_i=\left(\frac{{\overline{P}}_{GCM,fu{t}_i}}{{\overline{P}}_{GCM,bas{e}_i}}\right) $$
(2)

where ΔT i and ΔP i are average long-term monthly temperature and precipitation change for i th month, respectively. \( {\overline{T}}_{GCM,fu{t}_i} \)and \( {\overline{P}}_{GCM,fu{t}_i} \) are average long-term monthly temperature and precipitation for i th month, simulated by GCM under related scenario in future period, respectively. \( {\overline{T}}_{GCM,bas{e}_i} \) and \( {\overline{P}}_{GCM,bas{e}_i} \)represent average long-term monthly temperature and precipitation for i th month, simulated by GCM in baseline period, respectively. This approach has been applied by Wilby and Harris (2006), Morid and Bavani (2010), and Zareian et al. (2014).

In the next step, to weigh each of the 14 GCMs, mean observed temperature–precipitation (MOTP) method is used based on Eq. (3). The most important feature of this method is considering the ability of the models to simulate the observed climate variables, i.e., the difference between the simulated long-term variables in each month in the baseline period and the corresponding observed values.

$$ {W}_{ij}=\frac{\left(\frac{1}{\varDelta {d}_{ij}}\right)}{{\displaystyle \sum_{j=1}^{14}\left(\frac{1}{\varDelta {d}_{ij}}\right)}} $$
(3)

where W ij is the weight of j th GCM in i th month and Δd ij is the difference between average temperature or precipitation simulated by j th GCM in i th month of base period and the corresponding observed value (Gohari et al. 2013).

In the following, monthly temperatures and relative changes of precipitation with the weights of corresponding GCMs were used to construction of PDFs for each month. Based on the prototype of discrete PDFs and similar studies (Pindyck 2012; Teutschbein and Seibert 2012; Gohari et al. 2013), Gamma distribution with two parameters has been chosen to construct continuous distributions as follows:

$$ f(x)=\frac{1}{\varGamma (k){b}^k}{x}^{k-1}{e}^{\kern0.5em -x/b};\begin{array}{cc}\kern2em & \kern2em \end{array}x\ge 0 $$
(4)
$$ \varGamma (k)={\displaystyle \underset{0}{\overset{\infty }{\int }}{x}^{k-1}}{e}^{-x}dx $$
(5)

where x is the variable, k and b are shape and scale parameters, and Γ(k) equals the incomplete Gamma function. It should be mentioned that the values of k and b are changed to get the best fit based on the maximum likelihood estimation method using sum of squared error (SSE), which functions as follows:

$$ \mathrm{S}\mathrm{S}\mathrm{E}={\displaystyle \sum_{i=1}^n{\left({y}_i-{y}_i^{\ast}\right)}^2} $$
(6)

where y i is the data point, y i * is the estimation of gamma function, and n is the number of data points.

Accordingly, the developed discrete PDFs are used to construction of CDFs. More details about this method can be found in Gohari et al. (2013). In the next step, three different probabilistic scenarios are produced at three different percentiles (25, 50, and 75 %). The probability percentile represents low changes in temperature called “ideal scenario (IS)” and high changes in temperature called “critical scenario (CS),” and the probability percentile represents moderate changes in temperature called “moderate scenario (MS).”

2.2.3 Stochastic downscaling

As a downscaling technique, stochastic weather generator (WG) is used to produce daily site-specific climate scenarios (Barrow and Semenov 1995; Semenov 2007). These models are able to simulate synthetic time series of daily weather, which are statistically similar to observed weather (Richardson and Wright 1984 and Semenov 2007). In the present study, daily time series has been generated by LARS-WG (Semenov and Barrow 2002) from monthly climate change scenarios and historical daily climate data. The best distribution function for the observed daily time series for baseline period (1971–2000) has been selected, and then, synthetic future (2025–2054) daily time series are have been generated by LARS-WG based on the extracted climate change scenarios for 25, 50, and 75 % probability percentiles under A2 and B1 emission scenarios. As well as, the model was run to generate 300 years of future daily time series for each climate change probability percentile so as to deal with uncertainty in the outputs of LARS-WG. Accordingly, through averaging, the 300-year time series was divided into ten plausible daily time series for a 30-year period, and then, the average daily values of these ten time series are calculated. This approach was used by Gohari et al. (2013) to cut down the uncertainty of WG in generated climatic variables.

2.3 Agricultural water demand calculation

2.3.1 Crop evapotranspiration and effective precipitation calculation

Crop water requirement calculation is an important sector in agricultural water management. The FAO-24 methodology (Doorenboos and Pruitt 1977) is considered as the standard method for crop water requirement calculation and has been used in this study as follows:

$$ E{T}_{C_{j,t}}={K}_{C_{j,t}}\times E{T}_{0_t} $$
(7)

where \( {K}_{C_{j,t}} \)is the coefficient of the j th crop in the t th month, \( E{T}_{0_t} \)is the reference crop evapotranspiration in the t th month, and \( E{T}_{C_{j,t}} \)represents evapotranspiration for the j th crop in t th month. The input parameters for Eq. (7) are the length of four individual stages (initial season, growth season, mid-season, and late season) during the growing season and related crop coefficients (Kc). These coefficients are defined as the ratio between ETo and ETc for each part of the growing season (Multsch et al. 2015). Similarly, reference crop evapotranspiration can be derived based on the FAO Penman-Monteith method:

$$ E{T}_0=\frac{0.408\varDelta \left({R}_n-G\right)+\gamma \frac{900}{T+273}{u}_2\left({e}_s-{e}_a\right)}{\varDelta +\gamma \left(1+0.34{u}_2\right)} $$
(8)

where ET0 = reference evaporation [mm day−1], R n  = net radiation at the crop surface [MJ m−2 day−1], G = soil heat flux density [MJ m−2 day−1], T = mean daily air temperature at 2-m height [°C], u 2 = wind speed at 2-m height [m s−1], e s  = saturation vapor pressure [kPa], e a  = actual vapor pressure [kPa], e s  − e a  = saturation vapor pressure deficit [kPa], Δ = slope vapor pressure curve [kPa °C−1], and γ = psychrometric constant [kPa °C−1].

In this study, the USDA Soil Conservation Service (SCS) method has been employed to calculate effective rainfall using the CROPWAT model (Smith 1992; Clarke et al. 2000) based on the following:

$$ {PE}_{\mathrm{eff}}=\frac{\kern1em {P}_{\mathrm{tot}}\left(125-0.2{P}_{\mathrm{tot}}\right)\kern1em }{125}\kern.5em for{P}_{\mathrm{tot}}\le {250}^{\mathrm{mm}} $$
(9)
$$ {PE}_{\mathrm{eff}}=125+0.1{P}_{\mathrm{tot}}\begin{array}{ccc}\kern2em & \kern1em & for{P}_{\mathrm{tot}}\succ {250}^{\mathrm{mm}}\kern1em \end{array} $$
(10)

where PEeff is the effective precipitation (mm) and P tot represents total precipitation (mm).

2.3.2 Net water requirement and water demand volume estimation

Following the previous steps, the net water requirement and water demand volume are calculated based on Eqs.(11) and (12), respectively (Ashofteh et al. 2013; Multsch et al. 2015).

$$ {NWR}_{t,j}= \max \left({ET}_{c_{t,j}}-{PE}_{{\mathrm{eff}}_t},0\right) $$
(11)
$$ {WDV}_{t,j}=\left(\frac{NWR_{t,j}\times {A}_j}{100,000}\right) $$
(12)

where NWR t,j is the net water requirement (mm) for the j th crop in t th month, WDV t,j is the water demand volume (×106 m3) for the j th crop in t th month, and A j equals the area under cultivation (ha) for j th crop. The steps that are mentioned above are shown in Fig. 2.

Fig. 2
figure 2

Flowchart of the methodology

Full size image

3 Results and discussion

The box plots of changes of average annual precipitation and temperature in the study area, which are outputs from 14 GCMs [extracted from the data distribution center of IPCC (DDC: http://www.ipcc-data.org)], are shown in Fig. 3 for the future period (2025–2054) under two emission scenarios (A2 and B1). Based on the results, changes in average annual precipitation under A2 scenario range from 35 % decrease to 15 % increase, which can be changeable, and results of B2 scenario depicted a range from 37 % decrease to 18 % increase for the future period compared to the base period. Unlike this expected irregular process in precipitation changes, there is an expected regular increase of more than 1 °C in average annual temperature in all GCMs. The increasing rate of temperature between the two scenarios is different, and more increase is expected under A2 scenario. The highest increase of average annual simulated temperature by GCMs is 2.58 °C under scenario A2, and the least increase is 1.18 °C. The values under B1 scenario are 1.95 and 1.09 °C, respectively. In the outputs of all models, an increasing trend is seen in simulated temperature for future period compared to base period. According to these results, an uncertainty in simulation and outputs of different GCMs is clearly shown.

Fig. 3
figure 3

Box plots of average annual precipitation and temperature changes under emission scenarios (2025–2054)

Full size image

In the present study, the MOTP method of GCMs has been used in order to decrease the effects of this uncertainty. The results gained from GCM weight giving in simulated precipitation and temperatures in the base period are shown in Tables 3 and 4. The models HadCM3 and ECHO-G have the best performance (most weight) in simulated precipitation in the base period. Similarly, the models GISS-ER and ECHAM5-OM have the highest weight in simulating temperature in the base period. After weighting GCMs, the discrete PDFs and CDFs of monthly precipitation and temperature have been developed, as shown in Fig. 4.

Table 3 Weights (%) for 14 GCMs of precipitation simulation in the base period
Full size table
Table 4 Weights (%) for 14 GCMs of temperature simulation in the base period
Full size table
Fig. 4
figure 4

Constructed CDFs for relative precipitation and temperature changes (Mar)

Full size image

Figures 5 and 6 show monthly precipitation and temperature changes at different probability levels of 25, 50, and 75 % under emission scenarios (A2, B1). The change amount of monthly temperature under both scenarios is increasing in all months. This increasing amount of temperature differs in different months. Based on the results, the increases of expected average annual temperature under A2 scenario for probability levels of 75, 50, and 25 % were equal to 2.09, 1.89, and 1.59 °C, respectively, and 1.88, 1.61, and 1.36 °C for B1 scenario. It can be concluded that under A2 scenario, more increase in temperature is expected. Also, precipitation under climate change is expected to be very changeable and irregular under both scenarios, and most of the changes are observed in August and September. For example, at the probability level of 75 % under A2 scenario, the changes are observed with the ratios of 1.43 and 1.28 in August and September compared to changes in June with a ratio of 0.79. For B1 scenario, the ratio of changes has been from 1.45 in July to 0.84 in October at the probability level of 75 %.

Fig. 5
figure 5

Monthly expected precipitation changes (ratio) at different probability levels under emission scenarios (2025–2054)

Full size image
Fig. 6
figure 6

Monthly expected temperature changes (°C) at different probability levels under emission scenarios (2025–2054)

Full size image

After evaluating and checking the performance of LARC-WG model in the study area, the expected monthly changes of temperature (minimum and maximum) and precipitation at different probability levels are downscaled for 300-year time series. In the next step, the series of 300 years are broken to 30-year time series, and finally, the daily average of these series is calculated. The seasonal and annual climatic variable changes are shown in Table 5, which were downscaled at probability levels of 25, 50, and 75 %. Based on the results, the increasing rates of seasonal and annual average temperature under A2 scenario are more than those under B1 scenario. The highest increase in temperature is expected in autumn and summer. In addition, the precipitation changes differ in seasons and it can be, in some way, expected to observe a decrease in average precipitation in spring and summer under both scenarios. Also, the most expected decrease of mean annual and seasonal (spring) precipitation was predicted to occur at the probability level of 25 %.

Table 5 Seasonal and annual downscaled variable changes at different probability levels under emission scenarios (2025–2054)
Full size table

Accordingly, the monthly averages at probability scenarios of CS, MS, and IS have been used to estimate the amount of ET0 in the future period. In all months, the results show an increase in ET0. The most significant increase of ET0 is expected under A2-CS belonging to March and August. Considering the amount of different KC in each month, which is related to each crop based on the FAO journal, local studies, experiences, and growth periods for each crop (supposing the area under cultivation for each crop is unchangeable), the amount of ETc has been calculated in different months affected by climate change. The results of calculated annual ETc under climate change using probability scenarios are shown in Table 6. Results indicated an increase in ETc in future period compared to the base period. This increase under A2 scenario is expected to be more than B1’s, which, compared to the higher increase in simulated temperature, seems natural. Also, the increase in ETc at probability scenarios of CS is more than that of MS and IS. After calculating the effective precipitation based on the Eqs. (9) and (10), net water requirement and water demand volume are calculated for the future period, which are shown in Tables 7 and 8. Under both emission scenarios in the amount of water required for all crops, an increase is expected in all three probability scenarios. The amount of annual net water requirement for crops in irrigation network has increased approximately 10, 7, and 6 % for probability scenarios of CS, MS, and IS under emission scenario of A2. The numbers for B1 scenario are equal to 7, 5, and 4 %. The highest changes of net water requirement increase are expected to be for the crops of bean, barley, clover, and wheat. The ratio of changes in annual net water requirement in the period of 2025–2054 compared to the base period for these crops is 1.68, 1.32, 1.26, and 1.21, respectively.

Table 6 Annual ETc (mm) at different probability levels under emission scenarios (2025–2054)
Full size table
Table 7 Annual NWR (mm) at different probability levels under emission scenarios (2025–2054)
Full size table
Table 8 Annual WDV (×106 m3) at different probability levels under emission scenarios (2025–2054)
Full size table

Table 8 shows the results gained through calculating the water demand volume in the future period compared to the base period. Considering the area under cultivation, the most increase in water demand volume in the future period is expected to be for wheat, barley, and alfalfa. The amounts of annual changes in expected water demand volume for all crops in future period compared to the base period under A2 scenario for probability scenarios of CS, MS, and IS are 17.82, 13.04, and 10.84 (×106 m3). These numbers are, respectively, 12.56, 9.25, and 7.11 (×106 m3) for that under B1 scenario. The ratios of annual water demand volume for specified crops under emission scenarios of A2 and B1 are shown in Fig. 7. Based on the results, at CS and under both A2 and B1 scenarios, the most significant changes are seen in ratio of water demand volume for plants. The most meaningful changes in CS are for bean, barley, and clover, with ratios of 1.67, 1.32, and 1.26, respectively. In this regard, the values of B1 scenario are 1.47, 1.22, and 1.18, respectively.

Fig. 7
figure 7

Annual WDV changes (ratio) at different probability levels under emission scenarios (2025–2054)

Full size image

Generally, results showed an increase in water demand volume in future period compared to the base period, which could cause problems in agriculture management and food provision security of the study area considering the dry and semi-dry climates of this region and limited water resources. In addition, on results of the changes in climatic variables that caused an increase in irrigation water requirement, runoff from the Zard basin and Jarreh reservoir inflow quantity will surely be changed as well. Therefore, applying an adaptive management in water resources systems can be used as alternative solution for declining system vulnerability in the studied area.

4 Conclusion

Considering the impacts of agriculture on human life, an investigation of the effects of climate change on agriculture can better help water management and sustainable development. This issue seems to be more crucial in dry and semi-dry regions like Iran. Accordingly, in the present study, the effects of climate change on agricultural water requirement under different probability levels in Ramhormoz plain in southwest of Iran have been investigated. A risk framework has been used to manage uncertainties of GCM output and emission scenarios. In this framework, 14 GCMs are used under two emission scenarios of B1 and A2 through the MPOT method and a probabilistic approach. This methodology can be a proper solution to manage the uncertainty caused by the usage of GCMs in the assessment of climate change studies. The other aspects of uncertainty in irrigation demand under climate change are suggested to be addressed in future studies.

The results of examining the changes in monthly temperature indicated an increasing trend in all months and under both emission scenarios. Unlike the increasing and regular changes of temperature in the study area, an irregular change in precipitation (either decreasing or increasing) is expected in the future period compared to the base period. In addition, results of downscaled temperature under probability scenarios of CS, MS, and IS using LARS-WG model indicated an increase approximate to 1.67, 1.5, and 1.28 °C for A2 scenario and 1.48, 1.29, and 1.1 °C for B1 scenario, respectively, for the average annual temperature of the future period (2025–2054). It should be noted that the most expected decrease in precipitations under the studied emission scenarios is attributed to the probability scenario of CS with a value of approximately 10 %.

Calculation results related to ETc showed a monthly increase under both emission scenarios, as well as an increase in the amount of net water requirement and water demand volume. The increase in water demand volume and expected net water requirement in CS under both emission scenarios (A2, B1) is higher compared to the other probability scenarios (MS and IS) and shows a more severe condition. The amounts of expected increase of water requirement for probability scenarios of CS, MS, and IS are equal to 1082, 797, and 663 mm under emission scenario of A2 and 780, 565, and 443 mm for B1 scenario, respectively. Regarding to the area under cultivation for each crop, the increasing amounts of annual water demand volume in the study area in probability scenarios of CS, MS, and IS are 13, 10, and 8 % under A2 scenario and 9, 7, and 5 % under B1 scenario, respectively. This increasing amount in A2 scenario, considering its critical assumptions, is more than B1 scenario. Similarly, an increase in irrigation demand under climate change has been reported by Fischer et al. (2007), Gohari et al. (2013), Ashofteh et al. (2014), and Riediger et al. (2016).

Technical studies of climate change impacts on local-regional polices such as the construction of dams and hydraulic structures are very important. These strategies can enhance the performance of water resources systems and reduce the negative effects of climate change in the future. For example, according to the reports of Khuzestan water and power consulting, the construction studies of Abolabbas dam located upstream of the Jarreh reservoir (as a source of water supply to Ramhormoz plain) and Abolfares dam on A’alla river (as the other source of water supply) are currently under study. Therefore, an increase in net water requirement and water demand volume for the studied area (Ramhormoz plain) as was concluded in the present study can be useful for planning of these projects, and the authors suggest to consider this issue in the related studies.

Finally, given the limited water resources in dry and semi-dry regions like Iran and the role of agriculture in Iran’s economy, the development of adaptation scenario polices such as reduction of water loss, improving the efficiency of irrigation system, optimal irrigation planning, enhancing water use efficiency, applying the water price polices, changes in cropping patterns, etc. can reduce vulnerability in this sector.