Evaluating Impacts of Irrigation and Drought on River, Groundwater and a Terminal Wetland in the Zayanderud Basin, Iran

Abstract

The Zayanderud Basin is an important agricultural area in central Iran. In the Basin, irrigation consumes more than 90 percent of the water used, which threatens both the downstream historical city of Isfahan and the Gavkhuni Wetland reserve—the final recipient of the river water. To analyze impacts of land use changes and the occurrence of metrological and hydrological drought, we used groundwater data from 30 wells, the standardized precipitation index (SPI) and the streamflow drought index (SDI). Changes in the wetland were analyzed using normalized difference water index (NDWI) values and water mass depletion in the Basin was also assessed with gravity recovery and climate experiment (GRACE)-derived data. The results show that in 45 out of studied 50 years, the climate can be considered as normal in respect to mean precipitation amount, but hydrological droughts exist in more than half of the recorded years. The hydrological drought occurrence increased after the 1970s when large irrigation schemes were introduced. In recent decades, the flow rate reached zero in the downstream part of the Zayanderud River. NDWI values confirmed the severe drying of the Gavkhuni Wetland on several occasions, when compared to in situ data. The water mass depletion rate in the Basin is estimated to be 30 (±5) mm annually; groundwater exploitation has reached an average of 365 Mm³ annually, with a constant annual drop of 1 to 2.5 meters in the groundwater level annually. The results demonstrate the connection between groundwater and surface water resources management and highlight that groundwater depletion and the repeated occurrence of the Zayanderud River hydrological drought are directly related to human activities. The results can be used to assess sustainability of water management in the Basin.

1. Introduction

Arid and semi-arid climatic regions are important agricultural zones, as they host 41.3 percent of the world’s population [1]. Today, 14.7 percent of the farmed areas are irrigated in arid and semi-arid climatic zones, an important contributor in crop production [2]. As fully irrigated areas have two to three times higher crop yields than the rainfed areas, fully irrigated areas are increasing in dry regions [3]. The introduction of irrigation has increased agricultural drought occurrence since the mid-20th century [4] and stream flow depletion has led to wetlands loss [5]. Due to irrigation, the annual global groundwater depletion is estimated to be 283 (±40) km³ [6].

Recently, in some arid regions, actions have been taken to reduce the water consumption and depletion [7,8,9,10]. For example, the groundwater abstraction in the Jordan River Basin has been lowered to 150,000 m³ annually in attempts to reduce the annual depletion from 40 to 5 Mm³ [11]. The South African government is revitalizing the small-holders farms into large irrigation schemes in order to decrease the water depletion in the Olifants Basin [12]. Maintaining a central irrigation control system in the Lerma Chapala Basin, Mexico, has not only increased the water usage efficiency, but also increased the water volume of the Lake Chapala from 1330 Mm³ to 4250 Mm³ [13]. New regulation in the Yellow River Basin in China has been implemented to decrease the high pollution rate in the river due to industrial and agricultural residue [14]. Indian agricultural ministry is supporting less water demanding crops and rainfed agriculture in an attempt to face the water depletion in the Krishna and Bharani water Basins in the south [15,16]. Similar attempts are underway in other major basins in the arid and semi-arid zones, like in the Tunisian Mediterranean Basin and the Great Ruaha River Basin in Tanzania [17,18].

The objective of the present study is to assess consequences of agriculture in an arid river basin in Iran and to outline ways to improve future water management. The Zayanderud Basin in Iran has doubled its irrigated areas in the last fifty years which has led to the depletion of the river flow [19]. This affects the cultural heritage and image of the Isfahan city with its strong reliance on water and ancient bridges, etc. The loss of water also influences the main wetland in the region, the Gavkhuni Wetland and its biodiversity and function [19]. In this study, we aim to better document the case and analyze the effect of irrigation using remotely sensed data and in situ data to monitor the water loss and its consequences. The approach builds on change detection using multitemporal imagery, which has previously been applied to analyze changes in agricultural lands, forests, urban areas and water bodies [20]. We also use data from gravity recovery and climate experiment (GRACE) for water storage monitoring to assess temporal and spatial records [21,22,23,24,25,26,27,28,29]. In this study we use: (i) satellite images to estimate the surface area changes of the Gavkhuni Wetland and (ii) GRACE data to estimate the water mass fluctuation in the Basin. We analyze the drought frequency in the Basin from 1963 to 2012 to understand what the human factor of the drought occurrences is, and to discuss sustainable use scenarios for the Basin.

2. Materials and Methods

2.1. Study Area and Data

The Zayanderud River is the largest river in central Iran and forms one of the most important agricultural, industrial and urban zones of Iran [30]. The Zayanderud River originates from the Zagros Mountains and ends in the Gavkhuni Wetland (Figure 1). The Gavkhuni Wetland is one of the most important aquatic ecosystems in Iran; it is registered in the Ramsar convention of 1975. The Basin has an area of 41,524 km² with elevation range of 1470 to 3974 meters above sea level with precipitation ranges from 50 to 1500 millimeters (mm), with an annual average of 130 mm [31]. The average temperature ranges between 3 and 30 degrees Celsius (°C). The annual potential evapotranspiration is 1500 mm [30].

The Zayanderud River length is 350 km; more than 4 million residents inhabit its basin. The existence of deep and fertile soils, consisting of silts and clay loams, has provided a basis for intensive agriculture along the river [32], making agriculture the largest single water user in the Basin. With low precipitation rate in central and eastern areas of the Basin, irrigation is essential for the crop cultivation, especially to compensate water shortage in drought times and uses 90 percent of the total consumed water [33]. About 260,000 hectares (ha) are farmed in the Basin, mostly for cultivating wheat, barley, silage plants, potatoes and cotton [34]. Until 1952 water resources development in the Basin was confined to small diversion structures that provided irrigation water for irrigation systems. Irrigation water is derived from springs, underground canals (qanats) and snowmelt, with minimal number of full irrigation systems [35]. The major hydraulic structure, the Zayanderud Dam, was constructed in 1972 with a capacity of 1500 Mm³ [36]. Three transbasin diversion tunnels were constructed in 1985, 2004 and 2007. These transbasin diversion tunnels transferred 152 Mm³ annually to cities in Zayanderud facing water stress [30].

The study includes in situ data collected from the Basin and data derived from satellite image processing and modeling. The precipitation, evapotranspiration and temperature daily data for a period of 50 years from January 1963 to December 2012 are from the Iranian Meteorological Organization [37]. The groundwater monthly levels of 35 observation wells (Figure 1) were used for the period from January 1981 to December 2015, which were provided by Iranian Water Resources Management Company [38]. The Zayanderud River daily flow discharge is obtained from the gauge station at Varzaneh, for the period from January 1963 to December 2012. Varzaneh station is where the Zayanderud River flows into the Gavkhuni Wetland, giving the accurate monthly inflow volume to the wetland. Monthly GRACE datasets for total water mass storage variations from the period of April 2002 to December 2015 were processed to examine the total water mass variation in the Zayanderud Basin. The data consist of monthly snapshots, which when analyzed reveal monthly anomalies in total water storage. The GRACE data were provided by the German Research Center for Geosciences (GFZ).

Analysis of data over a long-time interval is important for several reasons. The 50 years precipitation and stream flow data (1963–2012), helps in monitoring the meteorological and hydrological droughts on a longer scale. In addition, it monitors any occurrence frequency change of the meteorological and hydrological drought in the 1980 s, the period when the irrigated areas widely expanded. The surface and groundwater data availability from 1985 until 2015, helps monitor the relationship between irrigated areas expansion and depletion of water resources. The GRACE data availability from 2002 till 2015 is limited to the monthly snapshot availability.

2.2. Monitoring the Meteorological and Hydrological Drought in the Zayanderud Basin

The standardized precipitation index (SPI) is widely used for defining and monitoring meteorological droughts [39,40,41]. The positive SPI values reflect wet conditions, while the negative values indicate meteorological drought. The detailed state definition is given in

Table 1. Definition of states of meteorological drought with SPI.

Criterion	Description of State
SPI ≥ 2.0	Extremely Wet
1.5 ≤ SPI < 2.0	Severely Wet
1.0 ≤ SPI < 1.5	Moderately Wet
0 ≤ SPI < 1.0	Normally Wet
−1.0 ≤ SPI < 0	Normally Dry
−1.5 ≤ SPI < −1.0	Moderately Dry
−2.0 ≤ SPI < −1.5	Severely Dry
SPI < −2.0	Extremely Dry

as suggested by Mckee [42]. The gamma distribution of the monthly precipitation time series is calculated as suggested by Lloyd and Saunders [43]:

$$\mathrm{SPI}=-(\mathrm{t}-\frac{{\mathrm{c}}_0+{\mathrm{c}}_1\mathrm{t}+{\mathrm{c}}_2{\mathrm{t}}^2}{1+{\mathrm{d}}_1\mathrm{t}+{\mathrm{d}}_2{\mathrm{t}}^2+{\mathrm{d}}_3{\mathrm{t}}^3}) \mathrm{for}0<{\mathrm{H}}_{(\mathrm{x})}\le 0.5$$

(1)

$$\mathrm{SPI}=+(\mathrm{t}-\frac{{\mathrm{c}}_0+{\mathrm{c}}_1\mathrm{t}+{\mathrm{c}}_2{\mathrm{t}}^2}{1+{\mathrm{d}}_1\mathrm{t}+{\mathrm{d}}_2{\mathrm{t}}^2+{\mathrm{d}}_3{\mathrm{t}}^3}) \mathrm{for}0.5<{\mathrm{H}}_{(\mathrm{x})}<1$$

(2)

where c₀ = 2.515517, c₁ = 0.802853, c₂ = 0.010328, d₁ = 1.432788, d₂ = 0.189269 and d₃ = 0.001308. H(x) is the cumulative probability function for gamma distribution, and t is average precipitation per the scale parameter. The SPI calculation is performed using the SPI Generator application provided by the National Drought Mitigation Center (University of Nebraska, Lincoln, NE, USA) [44]

The hydrological drought index (SDI) uses standardized annual streamflow volumes to define values lower than the mean streamflow by at least one standard deviation value. Based on SDI values, the state of hydrological droughts is defined, which is close to that of SPI (

Table 2. Definition of states of hydrological drought with SDI.

Criterion	Description of State
SDI ≥ 0.0	No Drought
−1.0 ≤ SDI < 0.0	Mild Drought
−1.5 ≤ SDI < −1.0	Moderate Drought
−2.0 ≤ SDI < −1.5	Severe Drought
SDI < −2.0	Extreme Drought

). Nalbantis and Tsakiris [45] suggested that the SDI can be calculated using the following:

$${\mathrm{SDI}}_{(\mathrm{i})}=\frac{{\mathrm{F}}_{\mathrm{i}}-\overline{\mathrm{F}}}{{\mathrm{S}}_{\mathsf{\sigma }}}$$

(3)

Where F_i is the annual stream flow for a given period (i), $\overline{\mathrm{F}}$ and ${\mathrm{S}}_{\mathsf{\sigma }}$ are the mean and the standard deviation of the cumulative streamflow for the whole study period.

2.3. Monitoring Water Mass Depletion in the Zayanderud Basin

GRACE provides monthly terrestrial water storage anomalies, from observing Earth’s time dependent gravity field. The data represent monthly anomalies in total water storage, stored in the terrestrial environment either as soil moisture, groundwater, snow or surface water. Monthly snapshots from April 2002 to December 2015 were used to examine the total water mass variation in the Zayanderud Basin. The GRACE water mass change is often represented in the form of equivalent water height in mm. In order to identify water storage changes from the spherical harmonic coefficients provided by (GFZ), technical suggestions of Cheng, Sweson, Wahr and Tourian were followed [46,47,48,49,50,51]. These technical notes suggest methods to remove the errors and calibrate the data, image destriping filters were used in order to reach the total water mass variation volume in the Basin.

2.4. Using Normalized Difference Water Index (NDWI) to Monitor the Gavkhuni Wetland Water Mass Fluctuation

Monthly Landsat surface reflectance images for the Gavkhuni Wetland were downloaded for the period from January 1985 to December 2015. Most of selected images were cloud-free or nearly cloud-free, meaning the cloud cover is less than 10 percent of the snapshot area. The snapshots had been georeferenced with the topographic maps. In the electromagnetic spectrum, water absorbs almost all of incident radiation in the near and mid-infrared bands, while shows a strong reflectance in visible bands. Based on the spectral analysis, NDWI is used to extract the water body by [52]. It is a band ratio index between the green and near infrared (NIR) spectral bands, which would not only enhance water features, but also depress vegetation, bare lands and other environmental information [53]. NDWI is defined as:

$$\mathrm{NDWI}=\frac{{\mathrm{R}}_{\mathrm{Green}}-{\mathrm{R}}_{\mathrm{Nir}}}{{\mathrm{R}}_{\mathrm{Green}}+{\mathrm{R}}_{\mathrm{Nir}}}$$

(4)

Where R_Green is the top of atmospheric reflectance of the green band and R_Nir represents that of the near-infrared band, which corresponds to band 2 and band 4, respectively, in Landsat surface reflectance. The NDWI value ranges from −1 to 1. McFeeters [52] set zero as the threshold, where the cover type is water if NDWI > 0 and it is non-water if NDWI ≤ 0. In order to indicate the best NDWI value reflecting the Gavkhuni Wetland surface area, 365 are defined with values 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8 and 0.9, respectively, for the positive NDWI values. The higher NDWI value, the higher the reflection rate is, indicating higher aquatic bodies’ probability. For different sub-threshold values, the corresponding wetland area was obtained by image processing using a MATLAB code. The wetland volume monthly variation is calculated for each sub-threshold from the water balance equation:

ΔV = V_i − V_i−1 = A_i(P_i − E_i) + Q_iΔ_t

(5)

Where A_i is surface area equivalent to ($\frac{{\mathrm{A}}_{\mathrm{n}}+{\mathrm{A}}_{\mathrm{n}-1}}{2}$) where n is any given month and is expressed in km². P_i, E_i and Q_iΔ_t are the monthly precipitation, evapotranspiration and inflow volume, respectively. Both P_i and E_i are in mm, A_i(P_i – E_i) in m³ and Q_iΔ_t is in m³ per month. Using Equation (5) with every sub-threshold, nine different values of Δ_V are calculated for each value of A_i.

The monthly variation of the wetland water level can be expressed as:

$${\mathsf{\Delta }}_{\mathrm{H}}=\frac{{\mathsf{\Delta }}_{\mathrm{V}}}{{\mathrm{A}}_{\mathrm{i}}}$$

(6)

Again, using Equation (5) with every sub-threshold, nine different values of Δ_V are calculated for each value of A_i. Δ_H and can also be expressed as:

$${\mathsf{\Delta }}_{\mathrm{H}}=({\mathrm{P}}_{\mathrm{i}}-{\mathrm{E}}_{\mathrm{i}})+\frac{{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}{{\mathrm{A}}_{\mathrm{i}}}$$

(7)

having the values of Δ_H, P_i, E_i and A_i calculated, the value of $\frac{{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}{{\mathrm{A}}_{\mathrm{i}}}$ in Equation (5) can be expressed in mm for the nine different values of A_i. Having the real time value of Q_iΔ_t measured from the gauge station, it is compared with the nine different Q_iΔ_t values calculated. This can be done by using the coefficient of determination as follows:

$${\mathrm{R}}^2={\{\frac{\mathrm{n}(\sum {{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}_{\mathrm{obs}}\times {{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}_{\mathrm{mod}})-(\sum {{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}_{\mathrm{obs}})(\sum \mathrm{y})}{\sqrt{\left[\mathrm{n}\sum {({\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}})}_{\mathrm{obs}}^2-{(\sum {{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}_{\mathrm{obs}})}^2\right]\left[\mathrm{n}\sum {({\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}})}_{\mathrm{mod}}^2-{(\sum {{\mathrm{Q}}_{\mathrm{i}}{\mathsf{\Delta }}_{\mathrm{t}}}_{\mathrm{mod}})}^2\right]}}\}}^2$$

where n is the number of observations, Q_iΔ_{t obs} is the real time monthly flow volume and Q_iΔ_{t mod} is the flow volume calculated from Equation (4). The coefficient of determination R² ranges from 0 to 1, with 0 meaning that the calculated Q_iΔ_t is totally independent from the measured Q_iΔ_t and 1 meaning both values are related without errors. The sub-thresholds with the highest value here means that its area best fits the flow data available. This sub-threshold area is thus considered the most identical area to the real time wetland surface area. From the obtained data, an area (A_i) versus volume (Δ_V) relation can be obtained. This mathematical relation will help fill in the missed wetland area data. Data availability was restricted by climatic reasons, such as clouds and fogs and the blurriness of images downloaded.

3. Results

3.1. Meteorological and Hydrological Drought Occurrence Frequencies in the Zayanderud Basin

The occurrence of meteorological droughts is rare in the Zayanderud Basin and less frequent than the stream flow droughts. The SPI calculations show that in more than half of the years, 54 percent of the occurrence, the climate conditions are considered wet, with SPI higher than 0 (Figure 2A). The “normal dry” years, with SPI value between −1 and 0, occur in 36 percent of the years between 1963 and 2012. Normally dry years are not considered drought years. Moderately drought years with SPI value between −1.5 and −1 occur in only 10 percent of the years. None of the years in the study period was recorded to be “severely or extremely” dry. For the period of 50 years, SPI values showed a statistically insignificant positive trend, meaning that the wet and dry years are statistically distributed throughout the study period. On the other hand, due to irrigation and reduced river flow, hydrological drought was recorded in 60% of the year in the Zayanderud River flow, SDI value less than 0.

The intensity and frequency of the hydrological drought increased in the 1980s. For the periods between 1980 and 2012, the occurrence of the hydrological drought was recorded in 75 percent of the years, compared to 35 percent before the years 1980 (Figure 2B). Overall, over the 50 years, the SDI values showed a negative trend, with a constant decrease in the Zayanderud River flow volume. This is directly corelated to the overexploitation of the surface water resources, as irrigated areas increased from 25,000 to 40,000 hectares [19]. Statistically, the hydrological droughts before the 1980s were directly related to the precipitation volume, (Figure 2C), as their occurrence was restricted to the years that are considered dry. Figure 2C, shows clearly that the (SPI–SDI) value has a constant positive value in the 1990s and 2000s. The positive value of the (SPI–SDI) value indicates the occurrence of hydrological droughts in years that are considered to be either wet or normally dry. Examples on these can be seen for year 1993, which is one of the most wet years in the study period, witnessed a mild hydrological drought. In addition, while the years 1998 and 1999 are the years with the most severe hydrological drought during the study period, they were normally wet and dry years on the SPI scale.

3.2. Water Mass and Groundwater Depletion in the Zayanderud Basin

The GRACE derived data shows that for the period between 2002 and 2015, the annual water mass depletion height in the Zayanderud Basin is 32.6 mm (Figure 3). The average water mass depletion volume in this period is 1350 Mm³ annually. This includes surface and groundwater exploitation volume and evapotranspiration rate. It is observed that the water mass volume is at a constant depletion after the year 2007 even that those years are considered to be normally wet years based on precipitation volume.

Figure 3 also shows the groundwater level depletion in 12 major aquifers for the period between 1981 and 2015, the in situ data shows that the average annual groundwater depletion volume is 365 Mm³, of which 90 percent is for irrigation, and the average decrease in the groundwater table is 0.6 meters annually, contributing in total water mass depletion in the Basin. This indicates that the groundwater exploitation is unsustainable. The most sever groundwater level depletion is recorded in North and South Mahyar, Damaneh, Isfahan, Murchehkhort and Ardestan (Figure 1). All these areas are considered major agricultural zones, where groundwater is the main water sources for irrigation [38]. On average each irrigated hectare needs 5200 m³ of water annually [38]. Even with some recharge rate in the years 1986, 1993 and 2004, as these years are considered the most wet years, all aquifers witnessed constant decrease in the groundwater table level. The major depletion period occurs in the 1990s and 2000s and is aligned with hydrological drought recorded in the area for the same period (Figure 2). It is also noticed that in the years that are considered dry and moderately dry, the groundwater exploitation increases to match the evapotranspiration rate.

3.3. Gavkhuni Wetland Water Mass Fluctuation

The results show that the Gavkhuni Wetland water surface area is directly related to the inflow rate from the Zayanderud River. The river inflow rate is usually affected by the precipitation rate in winter and the snowmelt in spring. With the increase in the Zayanderud River water exploitation, the flow decreased alarmingly. By comparing the in situ flow volume changes and the area retrieved from the NDWI calculations, results show that threshold (0.5) derived surface area had the highest coefficient of determination (0.75) with the real time in situ flow data. This indicates that the threshold (0.5) derived area data are congruent in 75 percent of the times with the real wetland area.

During the study period from 1985 to 2015, the area of the Gavkhuni Wetland fluctuated significantly. Using the water balance equation (Equation (5)), the missing surface area data of the Gavkhuni Wetland is interpolated (Figure 4B). The missing surface area here indicates images that could not be obtained, due to blurring or unavailability. This result shows that the NDWI derived data along with the inflow volume data can identify the surface water bodies’ area change. Results show that in the study period, 1985 till 2015, the wetland area changed rapidly and can be related to variation in the amount of precipitation. The highest wetland area, of average 480 km², was recorded in the years 1993 and 1994. This is related to the high precipitation in year 1993 (145 mm) (Figure 5). In the years 1988 and 1989 the wetland area was also high with an average area of 350 km², as the precipitation in this year was 218 mm. The wetland recorded its lowest area during the study period of an area of 30 km² in late 2008 (Figure 5). The precipitation rate in 2008 was the lowest recorded, with 24 mm of precipitation. In the Gavkhuni case, the precipitation rate change is not the only indicator for area change. Confirming this, (Figure 4A) shows the flow volume in million cubic meters into the Gavkhuni Wetland annually. Even with SPI confirming that 90 percent of the years in the period between 1963 and 2012 are wet, results show a depletion in the inflow rate. The average inflow is reduced from 226 Mm³ in the 1960s to 5 Mm³ in 2010. The major volume depletion happened in the 1980s and afterwards, as the average depletion volume reached 100 Mm³ each decade.

4. Discussion

Since the introduction of irrigation in the 1970s and 1980s, the river discharge and groundwater levels in Zayanderud Basin has decreased considerably. From the water source of Zargos Mountains, the flow decreased into the Zayanderud River, and on many occasions, the flow does not reach the downstream Gavkhuni Wetland. As the inflow reached zero, the Gavkhuni Wetland dried severely, and a minimum area was recorded in the winter of 2008/2009. The average water mass depletion derived from GRACE, in the river Basin is estimated 1350 Mm³ annually. The demand, which is largely generated by the expansion of the irrigation schemes, always exceeded the supply despite the water brought by reservoir and inter-basins transfer. An inter-basin transfer of 590 Mm³ of water is currently diverted from the Kuhrang River to the Zayanderud Basin [54]. The completion of the Behesht-Abad tunnel will bring additional 700 Mm³ downstream from the Chadgen Dam [55]. However, so far, the increase in the water availability has also increased the water demand, resulting in increased drying of the river and the wetland, e.g., increasing the irrigation efficiency, will lead to increase the irrigation in other parts of the Basin.

Large irrigation systems were constructed in the Zayanderud Basin in the 1970s and 1990s and were designed to cover an area of 200,000 hectares [54]. The irrigated areas in Zayanderud Basin are currently estimated to be 41,500 hectares, as lack of water resources limited the irrigation expansion. This expansion in the irrigated areas had a direct impact in increasing the hydrological drought intensity values. The hydrological drought occurrence increased 40 percent after the year 1980. Negative (SPI–SDI) values which indicate a stream flow higher than the precipitation volume, increased 40 percent in the 1970s and 20 percent in the 1980s. The hydrological drought occurred in 60 percent of the study years, and its intensity increased to 72 percent after the year 1980. Hydrological droughts have both environmental and ecological impacts.

The expansion in the agricultural areas has increased the stress significantly on groundwater, as irrigated schemes developed focused on groundwater extraction rather than formerly used underground canals (qanats). The groundwater exploitation increased as the well permits are granted by central administration, with limited knowledge of local hydrology. Data from the past 40 years have shown that the groundwater levels are dropping in all areas, and in some areas, they are dropping dramatically. The most severe groundwater depletion was recorded in Najafabad, because of large conversion of undeveloped lands which increased the groundwater depletion to 2.5 meters annually. In Isfahan, North Mayhar and Damaneh, the annual depletion is recorded to be between 1 and 1.5 meters. Annually in situ data shows that 365 Mm³ of groundwater resources are pumped to provide water for irrigation. This shows that the aquifers have a crucial role in compensating the deficiency of the surface water. The groundwater role can be maintained if the aquifers are replenished annually with renewable water resources. The constant depletion in the groundwater level, and the over exploitation rate will result in water scarcity in all major aquifers in the Basin.

The normalized difference water index proved to be a reliable tool for calculating the surface water body’s area. Threshold (0.5) was congruent in 75 percent of the times with the real inflow data, helping in creating a monthly area variation of the wetland for the period of 30 years. The results show that even with the wetland having a maximum area of 480 km², the average area was 350 km² in the period of 30 years. As most of the Zayanderud River water is used along the stream flow, minimum water is reaching the wetland located in the downstream. According to latter indicator—and involving the evaporation rate—the volume of the water requirements of the Gavkhuni Wetland is estimated to be 244 Mm³ annually [56]. The Varanze hydrometric station located on the entrance of the Gavkhuni Wetland shows that for the study period between 1985 and 2015, the average inflow to the wetland is 117 Mm³ annually. As a result, the Gavkhuni swamp, which is considered an important wetland for migratory birds and registered as a Ramsar site, became degraded. On average the swamp area has been constantly decreasing and was totally dry in 2009.

The main lesson learnt from this study is that even in arid and semi-arid climatic zones, the water scarcity currently observed is related mainly to human related activities rather than climatic reasons. Zayanderud Basin is a good example of basins facing total “water bankruptcy” as outlined by Klein [57], due to resources overexploitation. Similar observation were made in different basins, for example: the Urmia and Bakhtegan Basins on local level [27,58] and the Colorado and Tucson Basins in the United States [59,60], the Heihe Basin in China [61] and the Lancing–Mekong Basin in Cambodia [62]. All these basins share the fact that the water demand in the basins is high compared to resources that are available, and careful planning is needed for the agricultural sector. It is recommended in such climatic areas, water should be better allocated to agriculture contributing to high-value crops, or/and crops with higher yield per cubic meter, so that agriculture can compete against other water using economic sectors. Even with irrigated crops having higher yields, the rainfed crops, such as wheat, maize and other cereals must be encouraged as their water requirement is minimal. The groundwater resources usage must be restricted to a certain volume per well, as recharge volume is much less than the pumped volume.

Definitely the highest priority is providing potable water in rural areas. The other industrial and agricultural sector water consumption must be ranked from highest to least important in the field of water consumption. As more water availability from basin transfer means more water usage, sustainability must be the key factor in such a basin. Water productivity improvement will have greater impact on small and medium sized farms than increasing the water use efficiency. Usually recovering the water flow in the river is a great challenge for the arid and semi-arid basins. For example, despite water diversion into the Zayanderud Basin, the river flow has been reduced, and so water efficiency and water saving is important. Promoting water management techniques at the farm level should be made in ways to benefit farmers economical and improve the agricultural sustainability. In the region, water management should also consider issues with salinization, water quality, etc. Recent studies show that collaboration between farmers could lead to a more sustainable use of water in similar regions [63].

5. Conclusions

Zayanderud water resources are declining due to overexploitation of the surface and groundwater resources mainly due to expanding agricultural land. The constant reoccurrence of the hydrological droughts is directly related to the human activities in the Basin. After the 1980s and due to expanding the irrigated areas, hydrological drought became more frequent, even that 90 percent of the study years are considered normal in the means of precipitation volume. The hydrological drought occurrence frequency increased to 75 percent of the years in the Zayanderud River and there have been almost no flow to the Gavkhuni Wetland, especially in the summer month when the water demand is the highest for agriculture. The groundwater depletion, due to overexploitation, limited the surface groundwater interaction, thus reducing the wetland recharge. With a constant depletion of 0.6 meters annually, the decrease in the groundwater table level reached 50 meter in some aquifers. It can be said that the water mass depletion in Zayanderud is directly related to human activities rather than climatic changes. A basin-level survey is required to know the total number of wells and their depth, quality and drilling date. These data will help optimize combining the usage of surface and groundwater, and the current detailed groundwater situation. Wells that reached alarming levels or quality must be closed—especially in drought years. Information must be given to farmers on means of crop, land and water management under different circumstances, especially farmers with access to poor water quality decreasing their yield rate and increasing soil salinity. As April is usually the month with the highest flow rate, water annual allowance calendars must be set, depending on the precipitation rate in the previous months and expected evapotranspiration rate in summer. Water allowance must decrease in drought years. In drought years deficit irrigation must be applied only to overcome the drought direct effects on crops. Water institutes must classify the water priorities in the time of drought between different economic sectors. In closed basins, like the Zayanderud, all the renewable water resources are easily used. Transbasin water transition is seen as a solution today for water scarcity in the Basin, but this has not improved the water flow in Isfahan city or the Gavkhuni Wetland.