Agricultural Reservoir Operation Strategy Considering Climate and Policy Changes

Abstract

Agricultural water is affected by climate change and water management. Agricultural reservoirs are increasing demand on the environmental water supply because the Korean government has recently implemented an integrated water resource management policy. However, agricultural reservoirs are still in operation solely to supply agricultural water. To examine sustainable agricultural water management under climate change, we analyzed the strategy of operating regulations to efficiently distribute agricultural water as environmental water. We simulated the agricultural reservoir operation, analyzing its water supply capacity by applying operation regulations. The simulation predicted that future water supply capacity would decrease if the existing operation were maintained, and agricultural reservoir operation will be necessary in the future. The proposed reservoir operating strategy decreased the maximum water shortage and number of water shortage days compared with the existing operation with the required water supply. Our results can contribute to agricultural reservoir operation strategies and sustainable water management in response to climate change and provide decision-making guidance on water distribution for environmental use in response to water management policy changes.

1. Introduction

Water management is affected by climate change, as water stress is predicted to increase in central and southern Europe, and competition for water distribution and use in North America has been intensifying [1]. In South Korea, water resources are expected to vary with climate change in reservoir watersheds [2,3,4,5,6]. In Japan, for the Yagisawa Reservoir in the Tone River Basin, the need for release control has been suggested in accordance with the reduced inflow owing to climate change [7]. In Iran, for the Chadegan Reservoir in the Ayandeh-Rud River Basin, it has been predicted that the water level would decrease two months earlier than the expected time, and agricultural water would be insufficient owing to climate change [8,9]. In the future, there is a high possibility that hydrological systems will be damaged by water shortages due to climate change. Future climate impact assessments are being analyzed based on climate change information [10,11,12,13].

Water reservoir management plans were established using hydrology modeling in response to climate change [14,15]. In the Shihmen Reservoir, Taiwan, environmental water increased, while the domestic water supply decreased, as deduced through the application of a compact genetic algorithm [16]. In South Korea, operation of the Seomjin Reservoir had a more positive effect on river ecology when applying the normal water level than when applying a variable restricted water level [17]. Moreover, in the Philippines, a zone-based operation of the Angat Reservoir improved water supply and power generation as compared with the actual operation [18]. In Egypt, the imbalance between water supply and demand for the Aswan High Reservoir could be improved using a master–slave algorithm [19]. In Iran, the Karaj Reservoir could mitigate the adverse effects of climate change by changing its operating method from the standard operating policy to a modified linear decision rule [20]. Other studies [21,22,23,24,25] have analyzed improvements in the downstream flow regime by operating single reservoirs and reservoir groups. It thus appears necessary to evaluate reservoirs and establish a strategy for sustainable water management. In Pakistan, a study on the impact of several small-scale reservoirs on agricultural and groundwater development suggested strategies to improve land and water productivity by reservoir construction [26]. In Greece, a water management policy involving the construction and utilization of several reservoirs has been proposed in response to an increased water demand [27]. In Zambia, the importance of water management for the supply and demand of reservoirs required for sustainable water resource management in dry areas has been emphasized. In addition, it has been suggested that a conditioned rate of water use is necessary in preparation for late rainfall [28].

In South Korea, there are 17,106 agricultural reservoirs [29]. Agricultural reservoirs are single-purpose facilities for irrigation water supply during the irrigation period from April to September, and the inflow to the reservoirs is secured during the non-irrigation period [30]. As most agricultural reservoirs are located upstream, they negatively affect the ecology of small streams in rural areas [31]. In the future, this situation is expected to worsen the water shortage induced by climate change. In 2019, the Korean government implemented a policy for integrated water management [32], through which the environmental expansion of agricultural water use was brought to the fore. Thus, agricultural reservoir operations should consider the supply of environmental water downstream. However, to date, agricultural reservoirs have been operated based on the experience of facility managers without clear operating regulations. Therefore, agricultural reservoir operating regulations should be implemented to comply with the government policy and climate change for the efficient management of limited water resources.

We assumed that an operation strategy of agricultural reservoirs considering water policy is appropriate in the context of climate change. The objective of this study was to analyze the effects of the future water management of agricultural reservoirs by applying operation regulations on environmental water. The results contribute to the efficient management of water in agricultural reservoirs in preparation for future climate and water policy changes.

In Section 1, we review study cases on reservoir operation and changes in water policy. The research methods and data are presented in Section 2. In Section 3, we present the reservoir simulation results based on climate change and reservoir operation strategy. The results are discussed in Section 4. Finally, we provide a summary, highlight implications of our results, and state research limitations.

2. Materials and Methods

2.1. Study Area

The study area was the Hasan Reservoir located in central South Korea. Figure 1 shows the location of the reservoir watershed area (36°56′ N, 127°25′ E). Constructed upstream of the Jangyang Stream in 1959, this reservoir supplies water to an irrigation area of 519.5 ha, with a reservoir water storage capacity of 1676.0 × 10³ m³. The agricultural water from the reservoir flows to the Miho stream. The height and length of the reservoir are 18.8 and 169 m, respectively. The area of the reservoir watershed is 980 ha. The forest area in this watershed accounts for the largest share, at 81%, followed by agricultural land at 15.9%, grassland at 1.2%, and urban areas at 1%. In the past 30 years (1991–2020), the average rainfall was 1223 mm, and 69% of the total rainfall was concentrated from June to September. The average temperature in the area is 13.6 °C; August is the hottest month, with 25.8 °C, and January is the coolest month, with −2.3 °C. The average relative humidity is 67.9%, ranging from a minimum of 57.8% to a maximum of 76.6%. The average wind speed is 1.8 m/s, with a maximum of 2.2 m/s in April and a minimum of 1.4 m/s in October.

2.2. Data

Table 1. Meteorological, hydrological, and climate change data.

Data Set	Description	Location	Period
Meteorological data	Jincheonyeojung rainfall	36°38′ N, 127°27′ E	1970–2009
	Cheongju weather	36°38′ N, 127°26′ E	1970–2009
Hydrological data	Reservoir water storage	36°56′ N, 127°25′ E	1994–2007
Climate change data	RCP8.5 by HadGEM3-RA	36°57′ N, 127°23′ E	2010–2100

shows information about meteorological, hydrological, and climate change data used for modeling. Meteorological data used to model the operation of the Hasan Reservoir were collected from the Jincheonyeojung Rainfall Station by the Ministry of Environment and from the Cheongju Station by the Korea Meteorological Administration. We also collected daily data of these locations between 1970 and 2009 from the Water Resources Management Information System [33] and Weather Data Opening Portal [34]. Hydrological data, including the water storage area by reservoir elevation and water storage rate for calibration and validation of the simulation results, were obtained from the Rural Agricultural Water Resource Information System [35].

For the climate change scenario, i.e., future climate predictions, we used the daily data of the Korean Peninsula of the Representative Concentration Pathway (RCP). The RCP is a greenhouse gas concentration trajectory in which anthropogenic activities determine greenhouse gas (GHG) concentrations by adopting the IPCC 5th Assessment Report [36]. “Representative” means one radiative forcing with many socio-economic scenarios, and “Pathways” means changes over time in the greenhouse gas emission. The RCP has four scenarios: RCP2.6, RCP4.5, RCP6, and RCP8.5. The RCP numbers refer to radiative forcing values (2.6, 4.5, 6.0, and 8.5 W/m², respectively) in the year 2100. In RCP8.5, the current GHG emission trend is followed; in RCP6.0, GHG reduction policies are realized to a certain extent; in RCP4.5, GHG reduction policies are realized to a considerable extent; and in RCP2.6, GHG reduction policies start immediately. A high forcing figure indicates severe climate change in the future. Based on the abnormal climate report [37], the average annual temperature in Korea has been steadily rising over the past ten years, and the number of heatwave days (above 33 °C) increased from an average of ten (2000s) to 15.5 (2010s). Moreover, the average annual temperature in the study area has increased by 1.1 °C, while the average rainfall has decreased by 4.5% from the 2000s to the 2010s. It is expected that the vulnerability to water scarcity will increase in the future. Thus, we selected the RCP8.5 scenario, which conservatively approaches reservoir water management in response to climate change.

2.3. Agricultural Reservoir Operating Regulation

In Korea, operating regulations for agricultural reservoirs have not been established [38], but in Japan, agricultural reservoir operating regulations have been proposed by Senga [39]. The agricultural situation in Korea is similar to that in Japan. In our study, we suggested the agricultural reservoir operating regulations as the original operation, improving the supply of environmental water (Figure 2). This regulation is divided into the target line (TL) for normal operation, and restrictive release lines (RRL). If the reservoir is above the TL, discharge from the agricultural reservoir is promoted; below the TL, the water supply is limited based on the current reservoir water storage. The TL is defined as required water storage to ensure a certain probability that storage never empties.

Our calculations were performed using reservoir inflow (RI) and water demand data, as in Equation (1). In the original equation, operating regulations reflect only the demand for irrigation water (IW), but here, we also considered environmental water (EW). In addition, the TL and RRL calculation methods were retained. The irrigation period in this operating regulation was set from April to October, reflecting the irrigation period in Korea, and the environmental water is supplied during the non-irrigation period. V is the required water storage, which was calculated by accumulating QD (difference between reservoir inflow and water demand) values retroactively from the end of the irrigation period, as in Equation (2). The TL was drawn as the sequential V values corresponding to a drought, occurring once in ten years; s is zero. RRL was calculated based on the TL by applying a series limit in the range of 0 to 50% at 10% intervals; s is from 0.9 to 0.5 at 0.1 intervals. Environmental water can be supplied when there is a suppliable amount of water in the reservoir. It was calculated in a range in which the maximum constructed TL values do not exceed the storage capacity of the reservoir.

$$QD=RI-\left(1-s\right)\times \left(IW+EW\right)$$

(1)

$$V=\mathrm{previous}\mathrm{day}’\mathrm{s}V-QD,ifV<0thenV=0$$

(2)

2.4. Agricultural Reservoir Modeling

Agricultural reservoir modeling is a process of interpreting the changes in reservoir water volume using information on the amount of water flowing into and out of the reservoir. As the purpose of agricultural reservoirs is to supply water, the amount of water exceeding the full water level is naturally discharged. For agricultural reservoir modeling in Korea, the simplified water balance equation is based on the equation by Fowe et al. [40]. The change in reservoir storage was calculated using Equation (3). The water storage is commonly less than the reservoir water storage capacity; a value of zero indicates that the reservoir is depleted. Surface evaporation is the amount of water consumed from the water surface of the reservoir and is affected by evaporation and the water level. In addition, the overflow for a storage exceeding the full-water level is expressed by Equation (4):

$$\mathrm{VS}=\mathrm{previous}\mathrm{day}’\mathrm{s}\mathrm{VS}+\mathrm{RI}-\mathrm{RS}-\mathrm{RQ}$$

(3)

$$\mathrm{OV}=\mathrm{VS}-\mathrm{FS}$$

(4)

where, VS is water storage, RI is reservoir inflow, RS is the surface evaporation, RQ is the discharge quantity, OV is the overflow, and FS is water storage at full-water level.

Reservoir discharge is irrigation water, in which evapotranspiration, infiltration, and effective rainfall in paddy fields are considered. We calculated irrigation water by the water balance of paddy fields as shown in Equation (5) [41,42], with a water loss of 15% [43]. The paddy depth indicates the height of the paddy field, in the range of 60 to 80 mm. Paddy depth is generally set as 80 mm when designing reservoirs. Effective rainfall refers to natural rainfall that directly affects paddy fields at less than the paddy depth value. The paddy infiltration varies in the range of 4 to 9 mm, depending on the region. A value of 6 mm was used for infiltration based on a field survey. The reservoir inflow was calculated by an empirical formula based on multipurpose dam inflow data [44]. The runoff was calculated based on a nonlinear change relationship depending on the soil moisture storage, which ranges from 20 to 300 mm, as shown in Equation (6).

$$\mathrm{AW}=\mathrm{PD}+\mathrm{ER}-\mathrm{ET}-\mathrm{Inf}$$

(5)

$$\mathrm{Q}={\left(1-{\mathrm{e}}^{-0.003\times \mathrm{SM}}\right)}^{\left(0.2+{\mathrm{e}}^{-0.001\times \mathrm{SM}\times 4.5}\right)}\times \mathrm{SM}$$

(6)

where, PD is the paddy depth, ER is the effective rainfall, ET is evapotranspiration, Inf is infiltration, AW is the agricultural water, Q is the runoff, and SM is the soil moisture storage.

We simulated reservoir modeling considering operating regulations. When the operating regulation was applied to the reservoir modeling, the irrigation and instream waters were fully supplied when the water storage was larger than TL, as shown in Equation (7). However, when the water storage was smaller than the TL, the water supply was saved according to the restriction ratio, as shown in Equation (8), following RRL.

$$\begin{array}{ccccccc}\mathrm{RO}=\left(\mathrm{AW}+\mathrm{IW}\right)\times 100\hfill & & & & & & \mathrm{if}\mathrm{WL}\ge \mathrm{TL}\hfill \end{array}$$

(7)

$$\begin{array}{ccc}\mathrm{RO}=\left(\mathrm{AW}+\mathrm{IW}\right)\times \left(1-\mathrm{s}\right)/100\hfill & & \mathrm{if}\mathrm{WL}<\mathrm{TL}\hfill \end{array}$$

(8)

where, RO, AW, IW, and s represent the reservoir discharge, agricultural water, instream flow, and restriction rate, respectively. WL is the current reservoir water level.

2.5. Evaluation of the Reservoir Operating Regulation

Past modeling results were evaluated before applying the climate change scenario data. The calibration and validation periods were set to 1994–2000 and 2001–2007, respectively. We selected the coefficient of determination (R²) [45] and relative error (RE) as indicators [46] based on the observed and simulated storage data using Equations (9) and (10):

$$R^2=({\displaystyle \sum }\left(\mathrm{O}-\overline{\mathrm{O}}\right)(\mathrm{S}-\overline{\mathrm{S}})/\sqrt{{\displaystyle \sum }{\left(\mathrm{O}-\overline{\mathrm{O}}\right)}^2}\sqrt{{\displaystyle \sum }{\left(\mathrm{S}-\overline{\mathrm{S}}\right)}^2}{)}^2$$

(9)

$$RE=\left(\overline{\mathrm{O}}-\overline{\mathrm{S}}\right)/\overline{\mathrm{O}}\times 100$$

(10)

where O and S are the observed and simulated storages, respectively; and $\overline{\mathrm{O}}$ and $\overline{\mathrm{S}}$ are the average observed and simulated storages, respectively.

The impact of climate change was evaluated from 2011 to 2100, compared to the baseline (1971–2010). The future period was divided into three sections (2025s; 2011–2040, 2055s; 2041–2070, and 2085s; 2071–2100). To evaluate the safety of the water supply capacity in the hydrology system, we selected previously suggested vulnerability and reliability values [47,48]. Vulnerability (vul) is an index that indicates the severity of water shortage when water supply fails, as expressed by Equation (11).

$$vul=max\left(vul\left(j\right)\right)$$

(11)

where, vul is vulnerability and vul(j) is the maximum water shortage during the drought period.

Furthermore, reliability (rel) is the probability of a normal water supply during the planned period and can be expressed by Equation (12). In South Korea, we set the reliability to 90% based on a drought frequency of ten years:

$$rel=\left(1-Tn/TN\right)\times 100$$

(12)

where, rel, TN, and Tn represent the water supply safety (%), total years of analysis, and number of years with insufficient water supply, respectively.

3. Results

3.1. Calibration and Validation of the Reservoir Modeling

Table 2. Results of calibration and validation, and reservoir modeling parameters.

Calibration		Validation		Parameter
RE	R2	RE	R2
−2.52	0.75	−7.56	0.78	Paddy depth: 80 mm
				Infiltration: 6 mm
				Initial soil moisture storage: 20 mm,
				Initial reservoir water storage: 100%,
				Initial environmental water: 0

shows the results of calibration and validation, and the reservoir modeling parameters. The RE and R² of the reservoir simulation model were −2.52% and 0.75 for the calibration period (1994–2000) and −7.56% and 0.78 for the validation period (2001–2007), respectively. Figure 3 shows the calibration and validation periods of the reservoir operation modeling from 1994 to 2007. Furthermore, the daily simulated water storages were similar to the observed values of the reservoir. The reliability of the reservoir modeling results was confirmed before the application of climate and reservoir operation changes.

3.2. Agricultural Reservoir Operating Regulation Considering Environmental Water

Figure 4 shows the results of the operating regulation of the agricultural reservoir considering environmental water. From the water storage of the reservoir, the appropriate amount to supply environmental water, excluding the priority-allocated irrigation water supply, was calculated as 2900 m³ per day for the entire period. The water supply strategy of the operating regulation varies depending on the current daily water storage standards, such as the TL and RRL. When the reservoir water level is maintained above the TL, both planned irrigation water and environmental water are supplied. When the level is below the TL, an operating strategy is implemented to limit irrigation and environmental water using a restriction rate for each water level section.

3.3. Future Water Supply Capacity by Existing Operation

The variations in the future water supply capacity were analyzed by applying the existing operation for irrigation water only, which does not supply environmental water. The simulated operation results for the past three decades (1981–2010) under climate change are summarized in

Table 3. Results of reservoir modeling based on existing operation under climate change.

Classification	Temperature (°C)	Inflow (103 m3)	Irrigation Water (103 m3)	Water Storage (103 m3)	Overflow (103 m3)
Baseline	12.4	6703.5	2293.9	1338.5	4473.5
2025s	+1.3	5870.4 (−12.4%)	2540.4 (+10.7%)	1224.7 (−8.5%)	3454.0 (−22.8%)
2055s	+2.9	5835.8 (−12.9%)	2676.2 (+16.7%)	1219.4 (−8.9%)	3351.6 (−25.1%)
2085s	+4.8	6796.3 (+1.4%)	2680.4 (+16.9%)	1232.4 (−7.9%)	4382.5 (−2.0%)

. Without reduction in GHG emissions, the reservoir watershed temperature increased by 1.3, 2.9, and 4.9 °C in the 2025s, 2055s, and 2085s, respectively, compared with the baseline temperature of 12.4 °C. Compared with the baseline level, the future annual average reservoir inflow decreased by 12.4% and 12.9% in the 2025s and 2055s, respectively, but increased by 1.4% in the 2085s. The irrigation water supply in the future increased by 10.7–16.9%. For this scenario, the future storage decreased by 7.9–8.9%, and the decreasing range of water storage showed a similar trend. The overflow was predicted to decrease by 22.8, 25.1, and 2.0%, respectively, and the overflow of the 2085s was reduced compared with that of the 2025s and 2055s.

Figure 5 shows the results for the future water supply capacity of the agricultural reservoir through the application of the existing operation. The water supply reliability decreased in the 2025s, 2055s, and 2085s by 10, 23.3, and 16.7% compared with the baseline (100%), respectively. The water supply vulnerability increased in the 2025s, 2055s, and 2085s by 94,200, 52,500, and 94,300 m³ compared with the baseline (no water shortage), respectively. Thus, the water supply capacity decreased for the future periods when implementing an existing operation with no water management strategy under the climate change scenario without reduction in GHG emissions.

3.4. Water Management Effect Analysis of Reservoir Operating Regulations

The variations in the future water supply of the reservoir by applying the operating regulation were analyzed and compared with those under the existing operation.

Table 4. Summary of reservoir modeling results with reservoir operation regulation (unit: 10³ m³).

Factors	2025s		2055s		2075s
	Existing Operation	Operating Strategy	Existing Operation	Operating Strategy	Existing Operation	Operating Strategy
Irrigation water	2540.4	2445.5 (−3.7%)	2676.2	2514.2 (−6.1%)	2680.4	2474.3 (−7.7%)
Environmental water	0	262.0	0	259.8	0	259.0
Water storage	1224.7	1176.5 (−3.9%)	1219.4	1178.6 (−3.3%)	1232.4	1216.3 (−1.3%)
Overflow	3454.0	3220.0 (−6.8%)	3351.6	3131.8 (−6.6%)	4382.5	4139.0 (−5.6%)

depicts a summary of the reservoir modeling results using reservoir operation regulation. With the implementation of regulations, irrigation water in the 2025s, 2055s, and 2085s decreased by 3.7, 6.1, and 7.7%, respectively, compared with that under the existing operation (2540.4 × 10³ m³). The annual average environmental water supplied was 262,000–259,800 m³, which amounted to 10.3–10.7% of the total water supply. Moreover, reservoir water storages in future periods decreased by 1.3–3.9%, and the overflow decreased by 5.6–6.8% more as compared to that for the existing operation.

Figure 6 shows the comparison results of the future water supply capacity with the implementation of reservoir operation regulations. In the 2025s, both the existing and regulated operation results showed the same reliability (90%), but in the 2055s and 2085s, the reliability of the operating results was 6.6 and 6.7% higher, respectively, than that under the existing method. Thus, reliability improved under reservoir operation regulation. The reservoir vulnerability decreased by 72,700, 31,000, and 72,700 m³ for operating regulations in the 2025s, 2055s, and 2085s, respectively, compared with that under the existing operation. The regulated reservoir operation was analyzed for future periods, and a higher water supply reliability was found, indicating a lower vulnerability as compared to that of the existing operation conditions.

4. Discussion

Climate change inhibits the implementation of water management strategies for agricultural reservoirs. Our study predicted reservoir water shortage for a continued increase in irrigation water, as shown in Table 3. Even if precipitation remained the same during the irrigation period, as temperatures rise, the water demand for irrigation would increase. The future reservoir storage would decrease correspondingly due to a decreased inflow to the reservoir watershed and the increased irrigation water supply. Kim et al. [49] expected water shortage in the Han River watershed in Korea over time, which is similar to our results. Another case study in China [50] predicted that irrigation needs would increase in most areas of the mid-to-lower reaches of the Yangtze River. However, in the southern regions, irrigation demands have been shown to decrease. The changes in water demand differ by region in the same country. This is similar to the results of a previous study [51] in which the spatial distribution of potential evaporation varied by temperature. There have also been cases where the demand for irrigation water has changed. Climate change was found to reduce the water supply capacity of the agricultural reservoir, as shown in Figure 3. However, climate change showed a varying hydrological change for each specific period, rather than a constant trend over time. The existing reservoir operation is currently supplying all the necessary irrigation water. The future water supply capacity of the reservoir, however, is expected to get worse under the existing operation conditions. These results indicate that a reservoir operation strategy is necessary in response to climate change. The proposed reservoir operation regulations were adjusted based on the current water storage condition to prepare for the possibility of water shortage in the reservoir.

By implementing the integrated water resource management policy, the agricultural water management of South Korea has changed from empirical to scientific water resource management. A strategy should be prepared to supply environmental water in agricultural reservoirs based on the change in the national water management policy. The current reservoir operation merely supplies the required quantity of water. We believe that reservoir operations also need to regulate the water supply. The proposed strategy of regulating the operation of agricultural reservoirs is to promote water supply when the amount of stored water is above the TL; below the TL, the water supply is limited according to each water level zone. Alimohammadi et al. [19] suggested that existing operation methods are suitable if there is sufficient water in the reservoir, which is similar to the results for the TL in our study. RLL has similar characteristics in that the operation of the reservoir reduces the water supplied, similar to the hedging rule [52]. Our reservoir operation differs from these hedging operations, as it has different hedging regulations for each water level section. The proposed operating regulation is effective in agricultural reservoir operation, including the supply of environmental water considering policy changes. Chang et al. [53] have shown that all trends induced by the reservoir operation improved the reservoir water supply capacity. However, Jin and Lee [54] reported that the application of operation regulations improved the water supply reliability, while the vulnerability increased. Our study also showed that the maximum water shortage will decrease under the operating regulation compared with that under the existing operation (Figure 6b). However, the 2025s is a period without an improvement in water supply reliability, as shown in Figure 6a. Although the reliability of applying operation regulations increased compared with that under the existing operation in the 2055s, it did not satisfy the 90% design standard of reliability. Thus, the water management strategy of reservoir operating regulations does not indicate an unconditional improvement of the water supply capacity. Nevertheless, it can be inferred that operation regulations offer a more efficient and controlled water supply with environmental water for a sustainable agricultural water management than the existing operation.

The findings of this study have to be considered in light of its limitations. Sedimentation caused by biotic/abiotic factors [55,56] reduces the water storage capability of reservoirs, causing difficulties in future water management. Reservoirs are designed considering sedimentation, and maintenance includes the removal of sediments. Nevertheless, this study did not include such a maintenance scenario, but focused on solving the sustainable water resource management problem.

5. Conclusions

Agricultural reservoirs are used with the sole purpose of supplying irrigation water. In accordance with climate and policy changes, a usability review is being conducted for the use of agricultural water as environmental water. However, there are no proper reservoir operating regulations for the distribution between irrigation and environmental water. We evaluated the applicability of existing operating regulations for the supply of agricultural water in response to climate change using agricultural reservoir modeling, assuming no reduction in GHG emissions in the future. The effects of the reservoir in supplying environmental water were analyzed in comparison with the existing operation conditions. The results indicated that the water supply capacity will decrease in the future if the existing operation conditions are maintained, and reservoir operation regulation is necessary when considering future changes. Regulation of reservoir operations tended to decrease the maximum water shortage and the number of water shortage events compared with the existing operation; this was achieved by controlling the supply with agricultural and environmental water. Our results can contribute to agricultural reservoir operation strategies for a sustainable water management in response to climate change and provide guidance for decision-making with respect to water distribution for agricultural use with environmental supply in response to water management policy changes.