Next Article in Journal
Research Progress and Perspectives on Wastewater-Based Epidemiology: A Bibliometric Analysis
Previous Article in Journal
Changes in the Hydrological Regime of the Volga River and Their Influence on Caspian Sea Level Fluctuations
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Water
Volume 16
Issue 12
10.3390/w16121746
water-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Impact Assessment of Climate Change on Water Supply to Hsinchu Science Park in Taiwan

by
Tsung-Yu Lee
*,
Yun-Pan Lai
,
Tse-Yang Teng
and
Chi-Cheng Chiu
Department of Geography, National Taiwan Normal University, Taipei 106, Taiwan
*
Author to whom correspondence should be addressed.
Water 2024, 16(12), 1746; https://doi.org/10.3390/w16121746
Submission received: 16 May 2024 / Revised: 12 June 2024 / Accepted: 18 June 2024 / Published: 20 June 2024
(This article belongs to the Section Water Resources Management, Policy and Governance)
Browse Figures
Versions Notes

Abstract

:
The Hsinchu Science Park (HSP) in Taiwan plays a vital role in the chain of semiconductor production, but water scarcity has been challenging semiconductor manufacturing. The Baoshan Reservoir (BS) and the Baoshan Second Reservoir (BSR) are two major sources of water supply to the HSP. However, the impacts of climate change on the water supply have not been analyzed. In this study, a hydrological model (i.e., SWAT) and an operation model of the BR and the BSR were coupled to assess the climate change impacts on the inflow, outflow, and water storage volume (WSV) of the reservoirs. The simulations were based on the weather data for the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios of AR5 for the Periods of 2021–2040, 2041–2060, 2061–2060, and 2081–2100 derived from up to 33 GCMs/EMSs. It is found that more intensified global warming would generally result in more apparent rainfall seasonality that is wetter in the wet season and dryer in the dry season and more magnified seasonality in river flow. During the hotspot period of water shortage in the HSP from February to May, future water scarcity is expected to worsen. Among the 16 combinations of scenarios and Periods, 13 indicate lower WSV in the future compared to the Baseline. The annual mean number of ten-day periods with WSV lower than the operation rule curve ranges from 4.84 to 6.95 ten-day periods, higher than the Baseline of 4.81 ten-day periods. Overall, RCP6.0 has the most significant impact on the study area, with the highest annual economic loss occurring during the 2041-2060 period, reaching USD 1 billion (~2.37% of the 2023 annual production value) for the HSP. This study also provides a three-month cumulative rainfall threshold as an operational warning indicator for the HSP. Our assessment results indicate that future water supply to the HSP should be a serious concern for stabilizing the manufacturing processes and hence the global semiconductor component supply.

1. Introduction

Humans have come to understand that business development relies heavily on the natural environment. Therefore, excessive extraction of natural resources or environmental degradation during business development will eventually affect the development of enterprises, known as nature-related risks. In the face of the deteriorating natural environment, the Taskforce on Nature-related Financial Disclosures [1] has emerged. According to the World Economic Forum, all of the top global risks for the next decade are nature-related and related to climate change [2], including climate action failure, failure of climate-change adaptation, and extreme weather. As these risks affect corporate strategy and reputation [3], stakeholders expect businesses to measure and report their nature-related risks to ensure their sustainable operation [1,4,5,6]. Taiwan’s Hsinchu Science Park (HSP), known as Taiwan’s Silicon Valley [7], is now the world’s largest semiconductor component center [8,9]. Due to its high water usage in production and the gradual impact of water shortage, it has attracted international media attention [10,11]. However, there is still a lack of relevant assessments on the impact of future water shortages on the water supply of the HSP.
Taiwan receives high annual rainfall of up to 2510 mm, which is 2.6 times the world average. However, there are significant seasonal differences in rainfall, with 78% of rainfall occurring during the wet season (May to October). Due to the short and steep nature of rivers in Taiwan, only about 18% of rainfall is retained annually, with reservoirs playing a crucial role in storing and distributing water resources [12]. To support the development of the HSP, the Baoshan Reservoir (BR) and the Baoshan Second Reservoir (BSR) were constructed to meet the needs of residential and industrial water use [13]. However, the impact of climate change seems to have begun. Taiwan experiences an average of 3–4 typhoons each year, which are a major source of rainfall during the wet season. However, from Typhoon Bailu on 26 August 2019 until Typhoon Haitang on 3 September 2023, no typhoons hit Taiwan, breaking records since observations began. This most severe drought in history has prompted the international community to recognize the potential impacts of climate change on the semiconductor supply chain, with Taiwan being of paramount importance [14]. For example, the Taiwan Semiconductor Manufacturing Company (TSMC), which also has factories in the HSP, produces nearly 25% of the world’s semiconductors and 92% of the most advanced chips used in products such as smartphones and automotive artificial intelligence [15]. The shortage of chips affects not only Taiwan’s economy but also the world’s economy [16]. Drought is undoubtedly one of the main causes of the semiconductor supply shortage [17].
In response to climate change, Taiwan has established the Taiwan Climate Change Projection Information and Adaptation Knowledge Platform (TCCIP) to track and analyze Taiwan’s climate variations and to provide Taiwan with downscaled meteorological data based on IPCC assessment reports for assessing the impacts of climate change on various aspects [18]. While there are many related studies showing regional differences, the overall main trends regarding rainfall include the following: no significant change in annual rainfall but a decrease in the number of rain days per year [19]; a decrease in events of low rainfall intensity but an increase in events of high rainfall intensity [20,21]; a decrease in spring rainfall (February to April) but an increase in Meiyu rainfall (May to June) [22,23]. As for river discharge, changes in rainfall patterns have also led to more significant seasonality of dry and wet periods [24] and inter-annual variation [25]. Magnified river discharge has been observed in the condition of intensified rainfall [26]. Fluvial sediment transport has been even more magnified due to non-linearly increased landslides triggered by intensified rainfall [25,27]. Both the variations in river discharge and sediment transport leading to sedimentation are detrimental to the storage of excess rainfall in reservoirs, directly impacting reservoir management [25,28]. Assessments of climate change generally align with observed trends in historical data but are more extreme [29,30,31]. Lee et al. (2023) conducted an assessment of water balance under climate change for 75 watersheds across Taiwan, finding that up to 64 watersheds may face more severe water scarcity than at present, including the Shangping River Watershed, which is the source of the BS and BSR [32]. However, past studies evaluating water resources have rarely integrated watershed and reservoir operation models for assessment [33,34,35,36].
Since 100% of the water supply for the HSP currently comes from the BR and BSR, the water storage volume of these reservoirs directly affects the water supply to the HSP. So far, no integrated hydrological model and reservoir operation model for assessing the impact of climate change on the inflow, outflow, and water storage volume of these two reservoirs has been developed. Therefore, the purpose of this study is (1) to establish a hydrological model of the Shangping River Watershed to rationally simulate the river discharge of the Shangping River; (2) to establish an operational model of the BS and BSR to simulate the reservoir’s inflow, outflow, and water storage volume; (3) to incorporate climate change scenarios into the models to project the changes in river discharge and reservoir inflow, outflow, and water storage volume until the end of the century; (4) to assess the number of ten-day periods when water demand is not fully supplied under climate change scenarios; and (5) to assess the potential impact on water supply to the HSP and its production value under climate change scenarios. In order to ensure the sustainable development of Taiwan’s Silicon Valley, the Taiwanese government has been promoting the Forward-looking Infrastructure Development Program since 2017. This program includes water environment projects such as inter-basin water diversion, expansion of seawater desalination plants, and wastewater recycling, all of which require a more comprehensive simulation tool to understand their benefits under climate change. The results of this study will serve as a fundamental tool for evaluating the effectiveness of climate change adaptation measures and optimization in reservoir management, which is not within the scope of the present study.

2. Material and Methods

2.1. Study Area

The Shangping River Watershed, with the Shangping flow gauge located at its outlet, has a drainage area of 222 km2 in Hsinchu County (Figure 1). The elevation within the watershed ranges from 211 to 2512 m above sea level. The Water Resource Agency (WRA) is responsible for monitoring river discharge. The river discharge rate is estimated by substituting consecutive water levels into the individual rating curve, which is calibrated by field measurements every year. The annual average discharge measured at the flow gauge is 14.03 cms. The monthly average discharge ranges from 4.60 cms in November to 31.65 cms in September. The lowest monthly average discharge occurs from November to January, averaging around 5 cms. There are three weather stations, i.e., Shei-Pa, Heng-Shan, and Mei-Hua stations, maintained by the Central Weather Administration around the watershed. The annual rainfall of the Shangping River Watershed is ~2310 mm, and ~73% of the rainfall occurs during the wet season (May to October) which is also the typhoon season. The mean daily air temperature is 18.4 °C, with an average of 12.0 °C in January and 23.5 °C in July. The Shangping River Watershed contains 4 types of land uses, including forest (92.68% of watershed area), agriculture (3.21%), building (1.46%), and bare land (1.77%). The soil mainly consists of Entisols (66.6%), followed by Inceptisols (31.8%). There are also small amounts of Inceptisols and Ultisols (1.6%) formed in areas with gentle slopes.
In order to support the development of the HSP, now a global center for chip manufacturing, which requires massive consumption of pure water in the process of cleaning and etching, the Taiwan government began planning and designing the BR in 1976. Construction commenced in April 1981 and was completed in June 1985, with a total storage capacity of 5.47 × 106 m3. The daily water supply for domestic, industrial, and public use was approximately 10 × 104 m3. In light of the rapid industrial and commercial development in the Hsinchu area, the demand for public and industrial water continued to increase. The WRA planned to build the BSR, which was officially completed in June 2006, with an effective storage capacity of 32.18 × 106 m3. Originally, BR and BSR were operated separately with separate operation rule curves. They were jointly operated starting from August 2018. This joint operation increased the daily water supply by 28.2 × 104 m3, allowing the two reservoirs to individually or jointly supply water to the Baoshan Water Purification Plant. Both reservoirs are off-site reservoirs, with water sourced from the upstream of the Shangping River diverted into the reservoirs through the Shangping Weir for storage (Figure 1). In the past decade, the daily water usage in the HSP was 14.2 ± 0.81 × 104 m3, which is still below the designed water supply capacity of the two reservoirs. However, the HSP has already begun to face frequent water shortages. According to past news reports from 1987 to 2021, the HSP experienced 20 months of water shortages, with 17 of the months being months between February and April and 3 of the months being May. The water resource management in the region is beginning to be impacted by the extreme rainfall patterns, increasing dryness in the dry season and wetness in the wet season, caused by climate change [37].

2.2. Data, Calibration, and Validation of SWAT (Soil and Water Assessment Tool)

SWAT, developed by the Agricultural Research Service of the United States Department of Agriculture, is a process-based, semi-distributed, and basin-scale model. SWAT has been widely used globally [38] and applied to other watersheds in Taiwan to successfully simulate daily river discharge [39]. In this study, SWAT version 2012 was applied. There are four main types of data that SWAT requires: digital elevation model (DEM) data, land use data, soil data, and weather data. DEMs at a resolution of 30 m were retrieved from the Advanced Spaceborne Thermal Emission and Reflect Radiometer (ASTER). Land use data were from the National Land Surveying and Mapping Centre, Ministry of the Interior, Taiwan, distributed in 2005. Soil maps were from the Council of Agriculture of the Executive Yuan, Taiwan. Weather data were from weather stations maintained by the Central Weather Administration, including daily precipitation, maximum and minimum air temperature, wind speed, relative humidity, and solar radiation. The hydrological response units (HRUs) are the smallest spatial units for the land-based portion of the simulated hydrologic cycle. HRUs are created by overlaying soil, land use, and slope (calculated based on DEM data) classes within each sub-basin based on user-defined thresholds [40,41]. The Shangping Watershed has 9 sub-basins and 657 HRUs, with multiple HRU definition methods in zero thresholds.
SWAT-CUP, officially released in 2015, is a powerful computer program serving for sensitivity analysis, calibration, validation, and uncertainty analysis of SWAT [42]. SUFI2 (Sequential Uncertainty Fitting version 2) was selected because of its high calibration efficiency and accuracy, considering the uncertainty of input data [43]. The 95% prediction uncertainty (95PPU) was calculated at the 2.5% and 97.5% levels of the cumulative distribution of an output variable obtained through Latin hypercube sampling [44,45]. The efficiency of model calibration and uncertainty is determined by the r-factor and p-factor [46]. The range of the p-factor (observations bracketed by the prediction uncertainty) varies from 0 to 1. The r-factor (achievement of small uncertainty band) is the average width of the 95PPU band divided by the standard deviation of the measured variable, and it ranges from 0 to 1 [47]. Good model uncertainty is expressed by a higher value of the p-factor (towards 100%) and a lower value of the r-factor (towards 0).
The parameters selected for calibration were based on previous studies [39,48,49,50]. The surface runoff process was represented by CN2 and SURLAG. CN2 is the SCS curve number for moisture condition II, and SURLAG controls the fraction of total available water allowed to enter the reach on any one day. For groundwater simulation, the following six parameters were calibrated: GW_DELAY controls water retention in the soil before entering the shallow aquifer. ALPHA_BF is a direct index of groundwater flow response to the changes in recharge. GWQMN is the flow threshold depth of water in a shallow aquifer required for return flow to occur. REVAPMN is the threshold depth of water in the shallow aquifer for water moving from the shallow aquifer to the overlying unsaturated zone. RCHRG_DP controls the fraction of percolation from the root zone, which recharges the deep aquifer. ESCO is the soil evaporation compensation factor, accounting for the effects of capillary action, crusting, and cracks.
In this study, daily river discharge was the simulation target. SUFI2 was applied by using the Nash–Sutcliffe efficiency coefficient (NSE) as the major objective function in the calibration and validation processes. The NSE value, as defined in Equation (1), ranged from −∞ to 1.
N S E = 1 − ∑ i = 1 n ( O i − S i ) 2 ∑ i = 1 n ( O i − O a v e ) 2
where Si is the ith simulated daily discharge [m3/s], Oi is the ith observed daily discharge [m3/s], Oave is the mean of observed value, and n is the total number of observations. Discharge data at the Shangping flow gauge from 2012 to 2014 were used for calibration, and data from 2015 to 2017 were used for validation. The parameters were calibrated by running SUFI2 for three complete iterations (3000 simulations). Please refer to the work of Abbaspour [42,44,45] for detailed descriptions regarding theories and procedures of applying the SWAT-CUP to facilitate calibration and validation of SWAT. For the SWAT application cases in Taiwan, please refer to the work of Lin, G. Z. et al. [39].

2.3. Future Weather Data

The Taiwan Climate Change Projection Information and Adaptation Knowledge Platform (TCCIP) produced statistically downscaled and grid-based (at a resolution of 5 km) daily rainfall/temperature data from various general circulation models (GCMs) and earth system models (ESMs) in the AR5, including the Baseline data (1960–2005) and the future projection data (2021–2100) [18]. In order to achieve the most comprehensive analysis of climate change impacts, four RCP scenarios were used for the analysis. There were 21, 30, 17, and 33 GCMs/ESMs available for RCP2.6, RCP4.5, RCP6.0, and RCP8.5, respectively. To analyze and present results associated with short- to long-term impacts of climate changes, future projection data were divided into four Periods, i.e., 2021–2040, 2041–2060, 2061–2080, and 2081–2100 with codes of “S”, “M1”, “M2”, and “L”, respectively. In the following text, for instance, RCP8.5 in 2081–2100 will be denoted as RCP8.5L. In addition, the capitalized “Period” is particularly used while referring to the Baseline, S, M1, M2, and L Periods.
Although the TCCIP provides gridded rainfall data, in the context of water resource management, hydrological models are often built using rainfall data from weather stations within the watershed. Similarly, in this study, SWAT was calibrated and validated with rainfall at three weather stations and discharge at the Shangping flow gauge. Therefore, in practical operation, rainfall data were only extracted from the grids where the weather stations are located and inputted into the verified SWAT model to simulate the discharge for both the Baseline and future Periods. Since gridded data represent the average rainfall within the grid, they differ from the meaning represented by weather station data. When GCMs/ESMs simulate historical rainfall, there are gaps compared to gauged rainfall in terms of rain days and rainfall amounts. Therefore, a bias correction method proposed by the TCCIP was applied [51]. A quantile mapping empirical cumulative distribution function (ECDF) procedure was the basis of this method. In addition, the wet-day threshold of the GCM/ESM data was applied to fit the probability of gauged rainfall. Afterward, a bias correction function was established for each GCM/ESM for each month, using observed daily rainfall at weather stations and gridded daily rainfall at the corresponding grid points during the Baseline Period. Then, this function was applied to the future rainfall data of each GCM/ESM before they were input into the verified SWAT to obtain the discharge under corresponding rainfall conditions. Consistent with the observational period of weather stations, this study utilized observational data ranging from 1987 to 2005 as the Baseline Period. In addition, the bias correction method was not applied to the temperature data. Please refer to the work of Teng et al. [51] for more detailed procedures regarding the use of gridded rainfall data, published by the TCCIP, in hydrological simulations.

2.4. Rules of Reservoir Operation

Generally, reservoirs have operation rules that specify the outflow standards under different storage conditions. Since the adoption of joint operations by BR and BSR after August 2018, there has been no clear indication of their operation rules. It is only stated that when the water storage volume (WSV) is below the operation rule curve (ORC), the reservoir may reduce water supply. And there is only one curve change every ten days (as shown by the grey curve in Figure 2c). Therefore, this study utilized historical operation records, from August 2018 to December 2021, of the two reservoirs (i.e., inflow, water storage volume, and outflow), HSP water usage, and river discharge at Shangping flow gauge to summarize and establish an operation model based on ten-day periods. For detailed information, please refer to the results in Section 3.1.

2.5. Procedure for Climate Change Impact Assessment

The procedure for assessing the impact of climate change was as follows: (1) Verify that the reservoir operation rules can reasonably simulate the ten-day WSV of both BR and BSR. (2) Confirm that SWAT can reasonably simulate the daily discharge of Shangping River. (3) Calculate the bias-corrected rainfall for each GCM/ESM for the Baseline Period before inputting it into SWAT to simulate the corresponding daily discharge, and then input the daily discharge into the reservoir operation model to obtain the corresponding ten-day WSV. (4) Calculate the bias-corrected rainfall for each GCM/ESM for the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios before inputting it into SWAT to simulate the corresponding daily discharge and then inputting the daily discharge into the reservoir operation model to obtain the corresponding ten-day WSV. (5) Compare the differences in rainfall, discharge, and WSV for each GCM/ESM between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios.

3. Results

3.1. Establishment of Reservoir Operation Rules

Table 1 presents the operation rules established for the BR and BSR using the information provided in Section 2.4. It is found that when WSV ≤ ORC, the inflow to the BR in the tth ten-day period (Qin,BR,t) had a linear relationship with the logarithmic value of river discharge at the same timestep (Qt) with R2 = 0.78 (Equation (4)), indicating an attempt to increase WSV as much as possible. When WSV > ORC, inflow is kept as high as possible (when Qt < 107 m3, R2 = 0.48) or at a constant water volume (when Qt ≥ 107 m3) (Equation (3)). As for the inflow to the BSR (Qin,BSR,t), when WSV ≤ ORC, Qin,BR,t had a linear relationship with Qt (Equation (7), R2 = 0.93). When WSV > ORC, inflow is kept as high as possible (when Qt < 107 m3, R2 = 0.36) or at a constant water volume (Qt ≥ 107 m3) (Equation (6)). As for the outflow from the BR in the tth ten-day period (Qout,BR,t), outflow had a linear relationship with WSV in the (t − 1)th ten-day period (VBR,t−1) with R2 = 0.51 (Equation (2)), indicating an attempt to supply as much water as possible. However, the outflow from the BSR (Qout,BSR,t) seemed to be affected by the joint operation. The total outflow from BR and BSR in the tth ten-day period was linearly correlated to the total WSV of BR and BSR in the (t − 1)th ten-day period and HSP daily water usage. Therefore, the total outflow was first estimated, and then Qout,BR,t (as estimated by Equation (2)) was subtracted to obtain Qout,BSR,t when it was larger than WSV in the (t − 1)th ten-day period (VBSR,t−1) (Equation (5), R2 = 0.53). Otherwise, Qout,BSR,t = VBSR,t−1 because the outflow cannot exceed the WSV. It must be acknowledged that the processes of establishing operation rules may not be perfect and may even be arbitrary, although these rules were validated using the observed WSV for the BR and BSR. The relationship between the simulated WSV based on the operation rules established in this study and the observed WSV is shown in Figure 2. Although the R2 values for the established rules were not excellent, Figure 2 demonstrates that the established operation rules are capable of reasonably describing the changes in WSV for both reservoirs every ten days.
Water 16 01746 g002
Figure 2. Observed (blue curve) and simulated (red) water storage volume for (a) Baoshan Reservoir (BR), (b) Baoshan Second Reservoir (BSR), and (c) both reservoirs. Owing to the joint operation of the two reservoirs, there is only one operation rule curve (grey) drawn in panel (c).
Figure 2. Observed (blue curve) and simulated (red) water storage volume for (a) Baoshan Reservoir (BR), (b) Baoshan Second Reservoir (BSR), and (c) both reservoirs. Owing to the joint operation of the two reservoirs, there is only one operation rule curve (grey) drawn in panel (c).
Water 16 01746 g002

3.2. SWAT Verification

Figure 3 illustrates the observed and simulated daily discharge at the Shangping flow gauge. The results showed that SWAT could simulate daily discharge very well, with the NSE at 0.84/0.72 for the calibration/validation, which was more than satisfactory compared to worldwide discharge simulations [52]. The p-factor and r-factor indices were used to judge the strength of calibration and validation. The p-factor/r-factor was 0.97/0.15 for the calibration and 0.70/0.26 for the validation period, which was fairly good as well [53]. For the study watershed and all the watersheds in Taiwan, the three-order magnitude variation of daily discharge, mostly owing to typhoon invasion, increased the complexities of simulating watershed hydrology. However, the high p-factor, low r-factor, and high NSE in this study indicated that SWAT model uncertainties fell within the permissible limits. Hence, the verified model can be used for further applications, such as the assessment of the impact of climate change on river discharge in the Shangping Watershed.
The best-fit parameter values and the final ranges corresponding to the 95PPU obtained from SWAT-CUP are shown in Table 2. The final ranges of parameter values obtained in this study are very close to those obtained by our research team in other watersheds in Taiwan [39]. This indicates the general characteristics of Taiwan’s watersheds, including higher infiltration, quicker recharge, and more groundwater discharge to the river. These characteristics are reflected in lower values of CN2, GW_DELAY, and GWQMN, as well as higher values of ALPHA_BF, compared to the values studied in flat-terrain watersheds [54,55]. Jasechko et al. [56] have demonstrated that the steeper the watershed is, the more groundwater dominates annual river discharge, based on the measurements of oxygen and hydrogen isotopes of global stream water rather than modeling work.

3.3. Climate Change Impact on Rainfall and Discharge

The original rainfall and discharge data are both on a daily scale. To demonstrate the changes in rainfall and discharge under climate change, the daily rainfall and discharge data from each scenario (RCPs) and GCMs/ESMs were aggregated into monthly total rainfall and total discharge. The monthly average rainfall and discharge for each Period (Baseline, S, M1, M2, and L) were then calculated. Using the Baseline values as a reference, the change rates [%] of S, M1, M2, and L values relative to the Baseline were calculated, expressed as (Future estimates − Baseline estimates)/(Baseline estimates) × 100%. Figure 4 illustrates the median values, among the GCMs/ESMs, of the change rates in monthly rainfall and discharge between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. Except for May and from October to January of the following year, where the changes in rainfall under climate change are not significant (among GCMs/ESMs) and lack consistency (among Periods), the rainfall in February to April shows <0% in almost all scenarios and Periods, indicating that more than half of the GCMs/ESMs predict a decrease in spring rainfall in the future. On the other hand, from June to August, the values are almost >0%, indicating that more than half of the GCMs/ESMs predict an increase in summer rainfall in the future. The HSP has been suffering from water shortages, mostly from February to April. In addition, typhoons affect the study area mainly from June to September, which is also one of the major challenges for reservoir operations. Therefore, this paper only provides a concise description of the rainfall and discharge changes from February to April and from June to September.
Since the trends in rainfall changes are similar for all weather stations, only the change rates averaged from the three weather stations are presented (Figure 4). For rainfall, the mean change rates for February to April with the worsening scenarios range from −14.6% to 4.5% (RCP2.6), −23.1% to 4.9% (RCP4.5), −25.2% to −2.8% (RCP6.0), and −25.5% to 4.1% (RCP8.5) among all the Periods. For June to September, the change rates range from −10.5% to 47.9% (RCP2.6), −4.8% to 43.0% (RCP4.5), −1.0% to 31.9% (RCP6.0), and 6.5% to 55.7% (RCP8.5). Generally, as the scenarios worsen (from RCP2.6 to RCP8.5), there is a decreasing trend in rainfall for February to April. While there is an increase in rainfall for June to September, the trend is less significant with the worsening scenarios. As for the results of river discharge, the change rates for February to April with the worsening scenarios range from −24.4% to 3.6% (RCP2.6), −25.8% to 5.4% (RCP4.5), −34.5% to −5.2% (RCP6.0), and −35.5% to −3.3% (RCP8.5) among all the Periods. For June to September, the change rates range from −12.4% to 33.1% (RCP2.6), −6.9% to 56.3% (RCP4.5), −9.8% to 48.9% (RCP6.0), and 4.4% to 65.3% (RCP8.5). Similar to findings from a previous study in Taiwan, variations in rainfall lead to greater variations in discharge [25,26]. The decrease in discharge for February to April is magnified compared to the decrease in rainfall, while the increase in discharge for June to September exceeds the increase in rainfall.

3.4. Climate Change Impact on Water Storage Volume

The projected daily discharge for each scenario in each Period (i.e., Baseline, S, M1, M2, and L) can be input into the reservoir operation model. This involves simulating the WSV for each ten-day period according to the operation rules in Table 1. Taking into account the influence of the initial WSV setting on the simulation results, the results from the first year of the 20-year simulation are removed before calculating the average for each ten-day period over the remaining 19 years for any given GCM/ESM and any Period. Each GCM/ESM in each Period will produce an aforementioned result, and the median of all GCM/ESM results in each Period will be used as the representative, which will then be compared with the Baseline.
Figure 5 shows the average year-round total WSV of BS and BSR under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios for the Baseline, S, M1, M2, and L Periods. The Baseline curve is drawn in all small panels as a reference along with the projected WSV. Under the RCP2.6 scenario, the projected WSV is very similar to the Baseline, except for the M1 Period when the projected WSV (yellow curve in Figure 5a-2) is obviously lower than the Baseline before the 21st ten-day period. Under the RCP4.5 scenario, the lower projected WSV before the 21st ten-day period prevails for all the Periods, with an unchanged projected WSV for the rest of the ten-day periods. Under RCP6.0, the lower projected WSVs expand from before the 21st ten-day period to the entire year as the scenario worsens, which is particularly obvious in the M1 Period. Under the worst RCP8.5 scenario, the projected WSVs reveal different patterns compared to the Baseline among the Periods. For the S Period, the projected WSV is generally higher than the Baseline. For the M1 Period, the projected WSV is lower before the 21st ten-day period and remains similar to the Baseline afterward. However, for the M2 and L Periods, the projected WSVs are generally lower than the Baseline before the 21st ten-day period but higher afterward. In terms of the time of occurrence of the maximum and minimum WSVs, there is no significant change between the Baseline and all scenarios for all the Periods. The maximum WSV mostly occurs from late August to mid-September (24th to 26th ten-day periods), while the minimum WSV mostly occurs from mid-January to the end of February (2nd to 6th ten-day periods).
In terms of change rates [%] of WSV, compared to the Baseline, in the wet (May to October) and dry (November to April) seasons, under the RCP2.6 scenario in the dry season for all the Periods, the change rates are −0.01% (S), −1.45% (M1), 0.21% (M2), and −0.22% (L). In the wet season, the change rates are −0.33% (S), −1.06% (M1), 0.20% (M2), and 0.31% (L), showing an increasing trend toward the century’s end. Under the RCP4.5 scenario in the dry season, the change rates are −1.06% (S), −1.69% (M1), −1.65% (M2), and −2.54% (L) over time, indicating a gradual reduction in WSV. For the wet season, the change rates are −1.10% (S), −1.17% (M1), −0.66% (M2), and −1.00% (L). Under the RCP6.0 scenario in the dry season, the change rates are −3.68% (S), −7.57% (M1), −4.73% (M2), and −3.21% (L). For the wet season, the change rates are −1.84% (S), −4.62% (M1), −2.77% (M2), and −2.15% (L). Under the RCP8.5 scenario, the change rates in the dry season are 2.56% (S), −2.31% (M1), −1.32% (M2), and −1.86% (L); in the wet season, the change rates are 0.64% (S), −1.01% (M1), −1.17% (M2), and −1.40% (L).
It is found that there is no clear positive or negative relationship between the change rates in WSV during the dry and wet seasons over time and under different scenarios. However, it can be observed that the majority of the change rates are less than 0, indicating that more than 50% of the rainfall projected by GCMs/ESMs will lead to a general decrease in future WSV. This trend is particularly severe in the case of the RCP6.0 scenario. As shown in Figure 5, the WSVs simulated based on GCMs/ESMs vary, with some curves above the Baseline and others below. Only in the case of RCP6.0, almost all GCM/ESM results (grey curves in Figure 5c) are below the Baseline, making it the scenario with the greatest impact on the study area under climate change. However, as scenarios worsen and time progresses, the variability in projected results among GCMs/ESMs increases (i.e., RCP8.5L results vary the most), indicating the uncertainty of climate change. Undoubtedly, this further poses a greater challenge to water resource management as well.

3.5. Water Stress under Climate Change

We further analyzed the number of ten-day periods over the 19 years in each Period when WSV was below the ORC, indicating water demand is not fully satisfied, for each GCM/ESM under each scenario. The annual mean number of ten-day periods when WSV < ORC was then calculated and plotted as shown in Figure 6. Each box in Figure 6 represents the results of all GCMs/ESMs for a specific scenario and Period. For the Baseline Period (Figure 6a), the median value among all the GCM/ESM results is 4.81 ten-day periods, which a line is shown across all the panels as a reference. During 2021-2040 (S Period), the median values of the annual mean number of WSV < ORC ten-day period among all the GCMs/ESMs are 5.21 (RCP2.6), 4.84 (RCP4.5), 6.95 (RCP6.0), and 4.05 (RCP8.5) with the worsening scenarios (Figure 6b). For the period of 2041-2060 (M1), the values are 4.68 (RCP2.6), 5.34 (RCP4.5), 6.84 (RCP6.0), and 5.68 (RCP8.5) ten-day periods (Figure 6c). For 2061-2080 (M2), the values are 4.84 (RCP2.6), 5.37 (RCP4.5), 6.00 (RCP6.0), and 5.37 (RCP8.5) ten-day periods (Figure 6d). For 2081-2100 (L), the values are 4.63 (RCP2.6), 5.84 (RCP4.5), 6.11 (RCP6.0), and 5.39 (RCP8.5) (Figure 6e). Generally, within the same Period, the worse the scenario, the more ten-day periods in which WSV is lower than ORC. However, RCP6.0 still has relatively pessimistic results. Overall, there are only three projected results facing less water stress compared to the Baseline, i.e., RCP2.6M1, RCP2.6L, and RCP8.5S, in terms of the median value among all the GCM/ESM results. Among all the results for the S Period (Figure 6b), it can be observed that RCP8.5S appears to alleviate water stress. However, the rebound actually begins as early as RCP8.5M1 (Figure 6c). From the perspective of human development, which seems to follow RCP8.5 [57], one should not underestimate the overly optimistic perception of water stress relief brought by RCP8.5S and its cascading effects.

4. Discussion

4.1. Effects of Water Shortage

As the world’s center for semiconductor manufacturing, the HSP in Taiwan has attracted significant media attention due to the wide-ranging impacts of its water shortages [10,11]. A search of newspaper reports from 1987 to 2021 reveals that since its establishment, the HSP has experienced 20 months of water shortages, all occurring between February and May (the 6th to the 15th ten-day period), which coincides with the period of most significant reservoir WSV decline under climate change (Figure 5). Future water shortages are expected to expand, with the duration potentially extending into June and July (before the 21st ten-day period). While there is a trend of increasing WSV after the 21st ten-day period under the RCP8.5 scenario, the effects are mostly limited to influencing the period before January of the following year (before the 3rd ten-day period), thereby providing limited relief for the upcoming water shortage hotspot period beginning in February. The results for scenarios other than RCP8.5 are similar to the Baseline after the 21st ten-day period. Several studies have shown that Taiwan is experiencing a decrease in spring rainfall (February to April), a trend that is expected to worsen under climate change, which aligns with the findings of this study [37]. Overall, the outlook for water supply under climate change scenarios is serious. Of note, water shortages occur continuously and without interruption, one ten-day period after another. The results for all Periods and scenarios indicate that once the WSV falls below the ORC, the shortage will persist for 6.79 to 8.69 ten-day periods. Finding stable and continuous alternative water sources during shortages will be crucial, as the current method of emergency water transfer using water trucks from elsewhere will not be sufficient to meet future demands.
The production capacity of the HSP will be directly impacted by water shortages. According to the Taiwan Ministry of Economic Affairs project report in 2020, the indirect production value of each ton (i.e., m3) of water used in the HSP is approximately USD 425 [58]. Li et al. [59] used data from 2015 to 2021 to address that each ton of water from the Shangping River Watershed provides ecosystem services to the HSP worth between approximately USD 245 and USD 450. Based on the reservoir operation rules, it is estimated that, relative to the Baseline scenario, the annual outflow, collectively from the BR and BSR, will generally decrease except for in RCP2.6M2 and RCP8.5S, leading to general economic losses in semiconductor production, as shown in Table 3. If each ton of water shortage is estimated to result in an economic loss of USD 450, the water shortage will lead to the highest production value losses (RCP6.0M1) for the HSP reaching USD 1036 million (2.37% of the 2023 annual production value which is USD 43.78 billion), impacting the international integrated circuit supply chain and causing even greater economic losses for the downstream products. In the study, only water volume supply was used as the control factor for production value, and the capacity of machinery and equipment was not considered. The production values during periods of abundant water may have been overestimated, while the impact on production values during water shortages is more reliable. Therefore, the estimate in this study is conservative. Moreover, the water usage in the HSP is continuously growing, from 12.8 × 104 m3 per day in 2014 to 15.4 × 104 m3 per day in 2023, increasing steadily at an average rate of 0.26 × 104 m3 per day. Additionally, the production of 2-nanometer chips is expected to begin in 2025, which will require even more manufacturing water. This study estimated the outflow of the BSR based on an average water usage of 14.7 × 104 m3 per day from August 2018 to December 2021 (Table 1) and evaluated the impact of this water usage level under climate change. It can be imagined that the water shortage will be more severe than the currently estimated results, and the corresponding economic losses will also be much greater with increasing manufacturing costs and inflation. On top of that, the soil erosion and turbid river water caused by heavy rainfall and sedimentation in reservoirs will further exacerbate water shortage [25,27].

4.2. Early Warning Indicator

As the water supply infrastructure has not been fully developed, many factories in the HSP currently rent or dispatch water trucks to bring water back to the factory area when there is a shortage. In the case of limited water truck quantities and reserve water capacity, determining how to seize the opportunity to obtain a priority order for water intake has become a competitive issue among various factories. To provide advance warning to HSP businesses about water shortages and to make it easy for non-water resource professionals in the HSP to operate, this study used rainfall data from the MH weather station to provide an advance warning indicator. It first calculates the Standardized Precipitation Index (SPI) for 1 month, 3 months, and 6 months, i.e., SPI-1, SPI-3, and SPI-6 [60]. The results show that all 20 months of water shortage events in the HSP in the past occurred when SPI-3 in the previous month was less than 0. For example, if a water shortage occurred in February, then the SPI-3 for January must be less than 0, indicating that the cumulative rainfall from November to January was below the long-term average. Conversely, when the SPI-3 for a certain month is less than 0, and the WSV at the end of that month is close to or lower than the ORC, a water shortage is highly likely the next month. For the MH weather station, the rainfall thresholds corresponding to SPI-3 equal to zero for January, February, March, and April at the MH weather station are 175, 261, 367, and 450 mm, respectively. This means that when the cumulative rainfall from November to January is less than 175 mm, and the WSV continues to decline and is close to the ORC, the chance of a water shortage in February is very likely. This serves as a simple and actionable warning indicator for HSP staff.
This study further evaluates the change rates in cumulative rainfall every three months compared to the Baseline under different scenarios and Periods of climate change at the MH weather station (Figure 7). It uses the rainfall data from the grid at the location of the MH weather station after bias correction (Section 2.3). The results for all scenarios and Periods show that the cumulative rainfall every three months starting from October begins to decrease compared to the Baseline. The most significant decreases in three-month cumulative rainfall occur from December to February, January to March, and February to April. By the century end of the RCP8.5 scenario, the rainfall is projected to change by −9.4%, −16.0%, and −16.9%, respectively, for these periods, significantly increasing the risk of water shortage from March to May.

4.3. Current Adaptation Measures

Under the stress of water shortage, determining how to more efficiently allocate and use water resources will be a key issue for both the government and companies to address. In terms of water resource policy, the government enacted the Forward-looking Infrastructure Development Program in 2017, which includes projects such as inter-basin water diversion, seawater desalination plants, and water recycling plants that may increase the water supply to the HSP [58,61]. Additionally, water rights can be allocated accordingly. Huang and Chang pointed out that effectively transferring agricultural water supply to local industrial and domestic water use can effectively alleviate water shortage problems [62]. To ensure the sustainable operation of the HSP, the government has begun raising the spillway of the BSR by 1.35 m, increasing the WSV by an additional 1.92 × 106 m3. This would increase the water supply benefit to at least 0.9 × 104 m3 per day. Regarding enterprises, recycling, reuse, and restoration are water strategies of semiconductor corporations [63], which have significantly increased wastewater recycling rates and consider wastewater recycling as another water source. TSMC, the company with the largest market share, can be taken as an example; according to its sustainability report [64], its water resource recycling volume reached 216 × 106 m3 in 2022, with an average manufacturing water recovery rate of 85.7%. Its goal includes reducing unit product water consumption by 30% by 2040. Further reducing water resource consumption and increasing recovery rates will enable better adaptation to future low water supply conditions. However, the evaluation of adaptation measures under climate change is not the focus of the study. Although these measures can be effectively evaluated within the current framework of this study, including the adjustment of the ORC, the results for this issue await further research.

5. Conclusions

Semiconductor components serve as the brain and heart of most modern technological applications. Artificial intelligence, 5G, the Internet of Things, and everyday devices like phones, TVs, computers, and various home appliances all rely on semiconductor components. Taiwan’s Hsinchu Science Park (HSP) plays a crucial role in the global semiconductor supply chain. The production of semiconductors requires a large amount of water. Despite the construction of the Baoshan Reservoir (BR) and Baoshan Second Reservoir (BSR) to maintain the HSP’s production capacity, it still faces the risk of water shortage due to climate variability, attracting international attention. This study simulates the impact of climate change on reservoir water storage volume (WSV) until the end of the century through the establishment of hydrological and reservoir operation models. The results are not optimistic. Overall, regardless of the scenario, the current hotspot period of severe water shortage from February to May will become more severe, and the WSV before July will be lower than the current situation. The outflow from the reservoirs will be less than it is now, directly impacting semiconductor production capacity and manufacturers’ profits. Each water shortage event may last more than 67 days, and emergency backup water sources may not be able to meet the shortfall, requiring stable alternative water sources. The government is actively addressing this through the Forward-looking Infrastructure Development Program, including inter-basin water diversion and desalination of seawater. Companies are also reducing water usage and increasing the use of recycled water to mitigate the water shortage risk caused by climate change. The aforementioned adaptation measures can be subsequently evaluated using the models developed in this study to assess their effectiveness. Before the infrastructure is fully developed, this study provides a water shortage indicator that non-water resource professionals can use. During the period from November to April, when the three-month cumulative rainfall at the Mei-Hua weather station is below the long-term average and the reservoir WSV continues to decline, there may be a water shortage the following month, providing a reference indicator for companies to prepare for emergency water supply. Under the TNFD initiative, it is emphasized that the development of enterprises depends on the ecosystem services provided by nature, such as the river water relied upon by the semiconductor industry. Enterprises have a responsibility to disclose the impact of natural resource changes under climate change on their operations. This includes integrating natural factors into financial and business decision-making, allowing companies and financial institutions to manage these risks and interact with nature in a sustainable manner. This study provides the changes in the water environment under climate change scenarios, allowing HSP enterprises to conduct follow-up assessments accordingly.

Author Contributions

Conceptualization, T.-Y.L.; Methodology, T.-Y.L.; Validation, T.-Y.L.; Formal analysis, Y.-P.L. and C.-C.C.; Resources, T.-Y.L.; Data curation, T.-Y.L. and C.-C.C.; Writing—original draft, T.-Y.L., Y.-P.L., T.-Y.T. and C.-C.C.; Writing—review & editing, T.-Y.L.; Visualization, T.-Y.L., Y.-P.L., T.-Y.T. and C.-C.C.; Supervision, T.-Y.L.; Project administration, T.-Y.L.; Funding acquisition, T.-Y.L.. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Taiwan’s National Science and Technology Council [MOST 110-2410-H-003-122-MY3] and Academia Sinica [AS-SS-111-04].

Data Availability Statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical restrictions.

Acknowledgments

We wish to express our gratitude to the Water Resources Agency for providing the streamflow data, the Central Weather Administration for providing weather data, and the Taiwan Climate Change Projection Information and Adaptation Knowledge Platform (TCCIP) for providing projected weather data.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. TNFD. Recommendations of the Taskforce on Nature-Related Financial Disclosures. 2023. Available online: https://tnfd.global/publication/recommendations-of-the-taskforce-on-nature-related-financial-disclosures (accessed on 4 May 2024).
  2. World Economic Forum. The Global Risks Report 2022, 17th ed.; 2022; Available online: https://www3.weforum.org/docs/WEF_The_Global_Risks_Report_2022.pdf (accessed on 2 May 2024).
  3. Dobler, M.; Lajili, K.; Zéghal, D. Environmental performance, environmental risk and risk management. Bus. Strategy Environ. 2014, 23, 1–17. [Google Scholar] [CrossRef]
  4. ISSB. IFRS S1 General Requirements for Disclosure of Sustainability-Related Financial Information. 2023. Available online: https://www.ifrs.org/issued-standards/ifrs-sustainability-standards-navigator/ifrs-s1-general-requirements/ (accessed on 4 May 2024).
  5. TNFD. The TNFD Nature-Related Risk and Opportunity Management and Disclosure Framework, Beta v0.4. 2023. Available online: https://tnfd.global/publication/tnfd-v0-4-integrated-framework (accessed on 3 May 2024).
  6. World Economic Forum. Nature Risk Rising: Why the Crisis Engulfing Nature Matters for Business and the Economy. New Nature Economy Series. 2020. Available online: https://www3.weforum.org/docs/WEF_New_Nature_Economy_Report_2020.pdf (accessed on 3 May 2024).
  7. Mathew, J.A. A Silicon Valley of the East: Creating Taiwan’s semiconductor industry. Calif. Manag. Rev. 1997, 4, 26–54. [Google Scholar] [CrossRef]
  8. Chang, M.F.; Lin, C.; Shen, C.H.; Wang, S.W.; Chang, K.C.; Chang, R.C.H.; Yeh, W.K. The role of government policy in the building of a global semiconductor industry. Nat. Electron. 2021, 4, 230–233. [Google Scholar] [CrossRef]
  9. Voas, J.; Kshetri, N.; DeFranco, J.F. Scarcity and global insecurity: The semiconductor shortage. IT Prof. 2021, 23, 78–82. [Google Scholar] [CrossRef]
  10. BBC. Why the World Should Pay Attention to Taiwan’s Drought. 2021. Available online: https://www.bbc.com/news/world-asia-56798308 (accessed on 21 April 2021).
  11. CNN. How Island’s Historic Drought Could Threaten the Global Economy. 2021. Available online: https://www.cnn.com/2021/06/09/business/taiwan-drought-semiconductors-intl-hnk/index.html (accessed on 9 June 2021).
  12. Chen, R.S.; Tsai, C.M. Development of an evaluation system for sustaining reservoir functions—A case study of Shiwen Reservoir in Taiwan. Sustainability 2017, 9, 1387. [Google Scholar] [CrossRef]
  13. Huang, W.C.; Chang, T.H.; Yang, F.T. Water supply evaluation of Taiwan’s Silicon Valley. J. Am. Water Resour. Assoc. 2001, 37, 1279–1289. [Google Scholar] [CrossRef]
  14. Aviso, K.; Chien, C.-F.; Lim, M.K.; Tan, R.; Tseng, M.-L. Taiwan Drought was a Microcosm of Climate Change Adaptation Challenges in Complex Island Economies. Process Integr. Optim. Sustain. 2021, 5, 317–318. [Google Scholar] [CrossRef]
  15. Narvaez, L.; Janzen, S.; Eberle, C.; Sebesvari, Z. Technical Report: Taiwan Drought; Interconnected Disaster Risks 2021/2022; United Nations University-Institute for Environment and Human Security (UNU-EHS): Bonn, Germany, 2022. [Google Scholar]
  16. Kamasa, J. Microchips: Small and Demanded. CSS Anal. Secur. Policy 2021, 295, 1–4. [Google Scholar]
  17. Frost, K.; Hua, I. A spatially explicit assessment of water use by the global semiconductor industry. In Proceedings of the 2017 IEEE Conference on Technologies for Sustainability (SusTech), Phoenix, AZ, USA, 12–14 November 2017; IEEE: Piscataway, NJ, USA, 2017; pp. 1–5. [Google Scholar]
  18. Lin, L.Y.; Lin, C.T.; Chen, Y.M.; Cheng, C.T.; Li, H.C.; Chen, W.B. The Taiwan Climate Change Projection Information and Adaptation Knowledge Platform: A Decade of Climate Research. Water 2022, 14, 358. [Google Scholar] [CrossRef]
  19. Liu, S.C.; Wang, C.H.; Shiu, C.J.; Chang, H.W.; Hsiao, C.K.; Liaw, S.H. Reduction in Sunshine Duration over Taiwan: Causes and Implications. Terr. Atmos. Ocean. Sci. 2002, 13, 523–545. [Google Scholar] [CrossRef]
  20. Tung, Y.S.; Wang, C.Y.; Weng, S.P.; Yang, C.D. Extreme index trends of daily gridded rainfall dataset (1960–2017) in Taiwan. Terr. Atmos. Ocean. Sci. 2022, 33, 8. [Google Scholar] [CrossRef]
  21. Liu, S.C.; Fu, C.; Shiu, C.J.; Chen, J.P.; Wu, F. Temperature dependence of global precipitation extremes. Geophys. Res. Lett. 2009, 36, L17702. [Google Scholar] [CrossRef]
  22. Tung, Y.S.; Wang SY, S.; Chu, J.L.; Wu, C.H.; Chen, Y.M.; Cheng, C.T.; Lin, L.Y. Projected increase of the East Asian summer monsoon (Meiyu) in Taiwan by climate models with variable performance. Meteorol. Appl. 2020, 27, e1886. [Google Scholar] [CrossRef]
  23. Chou, C.; Chen, W.T.; Lo, M.H.; Lee, M.A.; Hsu, H.H.; Hong, C.C.; Cheng, C.T. Climate Change in Taiwan: Scientific Report 2017—Physical Phenomena and Mechanism; National Science and Technology Center for Disaster Reduction: New Taipei, Taiwan, 2017. Available online: https://tccip.ncdr.nat.gov.tw/upload/book/20180410112426.pdf (accessed on 6 November 2023).
  24. Chen, Y.J.; Chu, J.L.; Tung, C.P.; Yeh, K.C. Climate Change Impacts on Streamflow in Taiwan Catchments Based on Statistical Downscaling Data. Terr. Atmos. Ocean. Sci. 2016, 27, 741–755. [Google Scholar] [CrossRef]
  25. Kao, S.J.; Huang, J.C.; Lee, T.Y.; Liu, C.C.; Walling, D.E. The changing rainfall–runoff dynamics and sediment response of small mountainous rivers in Taiwan under a warming climate. In Proceedings of the Sediment Problems and Sediment Management in Asian River Basins, Hyderabad, India, 6–12 September 2011. [Google Scholar]
  26. Huang, J.C.; Lee, T.Y.; Lee, J.Y. Observed magnified runoff response to rainfall intensification under global warming. Environ. Res. Lett. 2014, 9, 034008. [Google Scholar] [CrossRef]
  27. Lee, T.Y.; Huang, J.C.; Lee, J.Y.; Jien, S.H.; Zehetner, F.; Kao, S.J. Magnified sediment export of small mountainous rivers in Taiwan: Chain reactions from increased rainfall intensity under global warming. PLoS ONE 2015, 10, e0138283. [Google Scholar] [CrossRef] [PubMed]
  28. Lee, L.C.; Hsu, T.C.; Lee, T.Y.; Shih, Y.T.; Lin, C.Y.; Jien, S.H.; Hein, T.; Zehetner, F.; Shiah, F.K.; Huang, J.C. Unusual roles of discharge, slope and SOC in DOC transport in small mountainous rivers, Taiwan. Sci. Rep. 2019, 9, 1574. [Google Scholar] [CrossRef] [PubMed]
  29. Tsai, I.C.; Hsieh, P.R.; Cheng, C.T.; Tung, Y.S.; Lin, L.Y.; Hsu, H.H. Impacts of 2 and 4 °C global warmings on extreme temperatures in Taiwan. Int. J. Climatol. 2023, 43, 702–719. [Google Scholar] [CrossRef]
  30. Chen, Y.J.; Lin, H.J.; Liou, J.J.; Cheng, C.T.; Chen, Y.M. Assessment of flood risk map under climate change RCP8.5 scenarios in Taiwan. Water 2022, 14, 207. [Google Scholar] [CrossRef]
  31. Huang, W.R.; Chang, Y.H.; Hsu, H.H.; Cheng, C.T.; Tu, C.Y. Dynamical downscaling simulation and future projection of summer rainfall in Taiwan: Contributions from different types of rain events. J. Geophys. Res. Atmos. 2016, 121, 13–973. [Google Scholar] [CrossRef]
  32. Lee, T.Y.; Chiu, C.C.; Chen, C.J.; Lin, C.Y.; Shiah, F.K. Assessing future availability of water resources in Taiwan based on the Budyko framework. Ecol. Indic. 2023, 146, 109808. [Google Scholar] [CrossRef]
  33. Chen, Y.H.R.; Tseng, H.W.; Hsu, K.C.; Chen, S.Y.; Ke, C.C.; Chiang, L.C. Evaluation of hydrological responses to climate change for a data-scarce mountainous watershed in Taiwan. J. Water Clim. Chang. 2023, 14, 1447–1465. [Google Scholar] [CrossRef]
  34. Lee, C.H.; Lin, S.H.; Kao, C.L.; Hong, M.Y.; Mr, P.C.H.; Shih, C.L.; Chuang, C.C. Impact of climate change on disaster events in metropolitan cities-trend of disasters reported by Taiwan national medical response and preparedness system. Environ. Res. 2020, 183, 109186. [Google Scholar] [CrossRef] [PubMed]
  35. Chang, I.C. A case study for identifying the potential challenges of water resources in the Yilan area of Taiwan: Using an adaptive water footprint approach. Environ. Sci. Pollut. Res. 2020, 27, 12725–12745. [Google Scholar] [CrossRef]
  36. Hsu, Y.J.; Fu, Y.; Bürgmann, R.; Hsu, S.Y.; Lin, C.C.; Tang, C.H.; Wu, Y.M. Assessing seasonal and interannual water storage variations in Taiwan using geodetic and hydrological data. Earth Planet. Sci. Lett. 2020, 550, 116532. [Google Scholar] [CrossRef]
  37. NSTC. National Climate Change Science Report 2024; National Science and Technology Center: New Taipei City, Taiwan, 2024. Available online: https://tccip.ncdr.nat.gov.tw/ScientificReport2024/ (accessed on 4 June 2024).
  38. Gassman, P.W.; Sadeghi, A.M.; Srinivasan, R. Applications of the SWAT model special section: Overview and insights. J. Environ. Qual. 2014, 43, 1–8. [Google Scholar] [CrossRef]
  39. Lin, G.Z.; Hsu, S.Y.; Ho, C.C.; Chen, C.F.; Huang, J.C.; Lee, T.Y. Application of Soil and Water Assessment Tool (SWAT) to evaluate the fates of nitrogenous fertilizer in subtropical mountainous watershed tea farms. Environ. Monit. Assess. 2022, 194, 213. [Google Scholar] [CrossRef] [PubMed]
  40. Arnold, J.G.; Kiniry, J.R.; Srinivasan, R.; Williams, J.R.; Haney, E.B.; Neitsch, S.L. Soil and Water Assessment Tool Input/Output Documentation Version 2012a; Texas Water Resources Institute: College Station, TX, USA, 2012; p. 7. [Google Scholar]
  41. Neitsch, S.L.; Arnold, J.G.; Kiniry, J.R.; Williams, J.R. Soil and Water Assessment Tool Theoretical Documentation Version 2009; Texas Water Resources Institute: College Station, TX, USA, 2011. [Google Scholar]
  42. Abbaspour, K.C. SWAT-CUP2: SWAT Calibration and Uncertainty Programs—A User Manual; Department of Systems Analysis; Integrated Assessment and Modelling (SIAM), Eawag, Swiss Federal Institute of Aquatic Science and Technology: Duebendorf, Switzerland, 2008. [Google Scholar]
  43. Akhavan, S.; Abedi-Koupai, J.; Mousavi, S.F.; Afyuni, M.; Eslamian, S.S.; Abbaspour, K.C. Application of SWAT model to investigate nitrate leaching in Hamadan-Bahar Watershed. Iran. Agric. Ecosyst. Environ. 2010, 139, 675–688. [Google Scholar] [CrossRef]
  44. Abbaspour, K.C. User Manual for SWAT-CUP, SWAT Calibration and Uncertainty Analysis Programs; Swiss Federal Institute of Aquatic Science and Technology, Eawag: Duebendorf, Switzerland, 2007; p. 93. [Google Scholar]
  45. Abbaspour, K.C. SWAT-CUP 2012: SWAT Calibration and Uncertainty Programs—A User Manual; Eawag: Dübendorf, Switzerland, 2013; p. 103. [Google Scholar]
  46. Arnold, J.G.; Moriasi, D.N.; Gassman, P.W.; Abbaspour, K.C.; White, M.J.; Srinivasan, R.; Jha, M.K. SWAT: Model use, calibration, and validation. Trans. ASABE 2012, 55, 1491–1508. [Google Scholar] [CrossRef]
  47. Yang, J.; Reichert, P.; Abbaspour, K.C.; Xia, J.; Yang, H. Comparing uncertainty analysis techniques for a SWAT application to the Chaohe Basin in China. J. Hydrol. 2008, 358, 1–23. [Google Scholar] [CrossRef]
  48. Lam, Q.D.; Schmalz, B.; Fohrer, N. Modelling point and diffuse source pollution of nitrate in a rural lowland catchment using the SWAT model. Agric. Water Manag. 2010, 97, 317–325. [Google Scholar] [CrossRef]
  49. Yen, H.; Bailey, R.T.; Arabi, M.; Ahmadi, M.; White, M.J.; Arnold, J.G. The role of interior watershed processes in improving parameter estimation and performance of watershed models. J. Environ. Qual. 2014, 43, 1601–1613. [Google Scholar] [CrossRef]
  50. Guse, B.; Reusser, D.E.; Fohrer, N. How to improve the representation of hydrological processes in SWAT for a lowland catchment–temporal analysis of parameter sensitivity and model performance. Hydrol. Process. 2014, 28, 2651–2670. [Google Scholar] [CrossRef]
  51. Teng, T.Y.; Liu, T.M.; Tung, Y.S.; Cheng, K.S. Converting Climate Change Gridded Daily Rainfall to Station Daily Rainfall—A Case Study at Zengwen Reservoir. Water 2021, 13, 1516. [Google Scholar] [CrossRef]
  52. Moriasi, D.N.; Gitau, M.W.; Pai, N.; Daggupati, P. Hydrologic and water quality models: Performance measures and evaluation criteria. Trans. ASABE 2015, 58, 1763–1785. [Google Scholar]
  53. Abbaspour, K.C.; Rouholahnejad, E.; Vaghefi, S.; Srinivasan, R.; Yang, H.; Kløve, B. A continental-scale hydrology and water quality model for Europe: Calibration and uncertainty of a high-resolution large-scale SWAT model. J. Hydrol. 2015, 524, 733–752. [Google Scholar] [CrossRef]
  54. Santra, P.; Das, B.S. Modeling runoff from an agricultural watershed of western catchment of Chilika lake through ArcSWAT. J. Hydro-Environ. Res. 2013, 7, 261–269. [Google Scholar] [CrossRef]
  55. Sorando, R.; Comín, F.A.; Jiménez, J.J.; Sánchez-Pérez, J.M.; Sauvage, S. Water resources and nitrate discharges in relation to agricultural land uses in an intensively irrigated watershed. Sci. Total Environ. 2019, 659, 1293–1306. [Google Scholar] [CrossRef] [PubMed]
  56. Jasechko, S.; Kirchner, J.W.; Welker, J.M.; McDonnell, J.J. Substantial proportion of global streamflow less than three months old. Nat. Geosci. 2016, 9, 126–129. [Google Scholar] [CrossRef]
  57. Schwalm, C.R.; Glendon, S.; Duffy, P.B. RCP8.5 tracks cumulative CO2 emissions. Proc. Natl. Acad. Sci. USA 2020, 117, 19656–19657. [Google Scholar] [CrossRef]
  58. MEA. Forward-Looking Infrastructure Design—Water Environment Construction: Taoyuan-Hsinchu Backup Pipeline Project Plan; Ministry of Economic Affairs: New Taipei City, Taiwan, 2020. Available online: https://www-ws.wra.gov.tw/Download.ashx?u=LzAwMS9VcGxvYWQvNDAzL3JlbGZpbGUvMC8xODM2OS8xYzA3ZDI5ZS1lNmU0LTQxYzktOTM4Zi1hMDdmN2UwMDRiY2MucGRm&n=5qGD56u5566h57ea56ysMeasoeS%2fruaguOWumuacrC5wZGY%3d (accessed on 28 April 2024).
  59. Li, C.E.; Lin, Z.E.; Dong, W.H.; Kuo, N.W.; Chiueh, P.T. Evaluating the Impact of Water Scarcity on Industrial Output Value in Hsinchu Area from the Perspective of Ecosystem Provisioning Services. J. Technol. Manag. 2023, 28, 101–132. (In Chinese) [Google Scholar]
  60. Svoboda, M.; Hayes, M.; Wood, D. Standardized Precipitation Index User Guide; WMO-No.1090; World Meteorological Organization: Geneva, Standardized, 2012. [Google Scholar]
  61. NRWRB. Hsinchu Desalination Plant Project; Northern Region Water Resources Branch, Water Resources Agency: Taichung City, Taiwan, 2020. Available online: https://www.wranb.gov.tw/cp.aspx?n=37866 (accessed on 4 May 2024).
  62. Huang, W.C.; Chang, C.W. Water shortage risk in Taiwan’s Silicon Valley. J. Chin. Inst. Eng. 2022, 45, 513–520. [Google Scholar] [CrossRef]
  63. Wang, Q.; Huang, N.; Cai, H.; Chen, X.; Wu, Y. Water strategies and practices for sustainable development in the semiconductor industry. Water Cycle 2023, 4, 12–16. [Google Scholar] [CrossRef]
  64. TSMC. Sustainability Report. 2022. Available online: https://esg.tsmc.com/download/file/2022_sustainabilityReport/english/e-all.pdf (accessed on 4 May 2024).
Water 16 01746 g001
Figure 1. The landscape of the Shangping River Watershed, including weather stations, flow gauge, and land use patterns. Water from the Shangping River is diverted into the Baoshan Reservoir and Baoshan Second Reservoir through the Shangping Weir to supply the demand of Hsinchu Science Park. The three weather stations are the Shei-Pa (SP), Heng-Shan (HS), and Mei-Hua (MH) stations.
Figure 1. The landscape of the Shangping River Watershed, including weather stations, flow gauge, and land use patterns. Water from the Shangping River is diverted into the Baoshan Reservoir and Baoshan Second Reservoir through the Shangping Weir to supply the demand of Hsinchu Science Park. The three weather stations are the Shei-Pa (SP), Heng-Shan (HS), and Mei-Hua (MH) stations.
Water 16 01746 g001
Water 16 01746 g003
Figure 3. Daily observed (blue curve) and simulated (red) discharge at the Shangping flow gauge. The dashed line separates the data into two subsets for calibration and validation. To keep the figure clear, the 95PPU, which is very narrow and almost overlaps with the simulated discharge, is not drawn on the graph.
Figure 3. Daily observed (blue curve) and simulated (red) discharge at the Shangping flow gauge. The dashed line separates the data into two subsets for calibration and validation. To keep the figure clear, the 95PPU, which is very narrow and almost overlaps with the simulated discharge, is not drawn on the graph.
Water 16 01746 g003
Water 16 01746 g004
Figure 4. The change rates [%] in monthly rainfall and discharge between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. The calculation of the percentage in the figure is defined as (Future estimates − Baseline estimates)/(Baseline estimates) × 100%. The number in each cell of the figure, for the discharge panel, represents the median of all GCM/ESM results for a specific scenario, month, and Period. For the rainfall panel, the average of the results from the three weather stations is shown. The color of the cell represents the degree of change rate: white indicates 0%, blue indicates >0% (the darker the blue, the more positive the change), and red indicates <0% (the darker the red, the more negative the change). The historical average monthly flow rates are 4.88 (Jan.), 9.49 (Feb.), 12.12 (Mar.), 13.79 (Apr.), 15.64 (May), 23.04 (Jun.), 18.08 (Jul.), 31.65 (Aug.), 22.55 (Sep.), 11.31 (Oct.), 4.60 (Nov.), and 5.12 (Dec.) cms, respectively, as a reference.
Figure 4. The change rates [%] in monthly rainfall and discharge between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. The calculation of the percentage in the figure is defined as (Future estimates − Baseline estimates)/(Baseline estimates) × 100%. The number in each cell of the figure, for the discharge panel, represents the median of all GCM/ESM results for a specific scenario, month, and Period. For the rainfall panel, the average of the results from the three weather stations is shown. The color of the cell represents the degree of change rate: white indicates 0%, blue indicates >0% (the darker the blue, the more positive the change), and red indicates <0% (the darker the red, the more negative the change). The historical average monthly flow rates are 4.88 (Jan.), 9.49 (Feb.), 12.12 (Mar.), 13.79 (Apr.), 15.64 (May), 23.04 (Jun.), 18.08 (Jul.), 31.65 (Aug.), 22.55 (Sep.), 11.31 (Oct.), 4.60 (Nov.), and 5.12 (Dec.) cms, respectively, as a reference.
Water 16 01746 g004
Water 16 01746 g005
Figure 5. The average year-round total water storage volume of Baoshan Reservoir and Baoshan Second Reservoir under (a) RCP2.6, (b) RCP4.5, (c) RCP6.0, and (d) RCP8.5 scenarios for the Baseline, S(-1), M1(-2), M2(-3), and L(-4) Periods. Blue curves indicate the results of the Baseline, remaining identical in every panel as a reference. Grey curves are the results of all the GCMs/ESMs for the corresponding scenario and Period. Green, yellow, orange, and red curves represent the median of the grey curves for the Periods of S, M1, M2, and L, respectively.
Figure 5. The average year-round total water storage volume of Baoshan Reservoir and Baoshan Second Reservoir under (a) RCP2.6, (b) RCP4.5, (c) RCP6.0, and (d) RCP8.5 scenarios for the Baseline, S(-1), M1(-2), M2(-3), and L(-4) Periods. Blue curves indicate the results of the Baseline, remaining identical in every panel as a reference. Grey curves are the results of all the GCMs/ESMs for the corresponding scenario and Period. Green, yellow, orange, and red curves represent the median of the grey curves for the Periods of S, M1, M2, and L, respectively.
Water 16 01746 g005
Water 16 01746 g006
Figure 6. The number of ten-day periods when water storage volume (WSV) was below the operation rule curve (ORC) in the (a) Baseline, (b) S, (c) M1, (d) M2, and (e) L Periods under Baseline, RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. Each box contains the annual mean number of ten-day periods over the 19 years when WSV < ORC derived from all the GCMs/ESMs for a specific scenario and Period. A grey line, representing the condition for the Baseline Period, is drawn as a reference. Values exceeding 1.5 times the interquartile range are defined as outliers (cross symbols).
Figure 6. The number of ten-day periods when water storage volume (WSV) was below the operation rule curve (ORC) in the (a) Baseline, (b) S, (c) M1, (d) M2, and (e) L Periods under Baseline, RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. Each box contains the annual mean number of ten-day periods over the 19 years when WSV < ORC derived from all the GCMs/ESMs for a specific scenario and Period. A grey line, representing the condition for the Baseline Period, is drawn as a reference. Values exceeding 1.5 times the interquartile range are defined as outliers (cross symbols).
Water 16 01746 g006
Water 16 01746 g007
Figure 7. The change rates [%] in three-month cumulative rainfall between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. The calculation of the percentage in the figure is defined as (Future estimates − Baseline estimates)/(Baseline estimates) × 100%. The number in each cell of the figure represents the median of all GCM/ESM results for a specific scenario, 3-month period, and Period. The color of the cell represents the degree of change rate: white indicates 0%, blue indicates >0% (the darker the blue, the more positive the change), and red indicates <0% (the darker the red, the more negative the change). The three-month cumulative rainfall levels corresponding to SPI-3 equal to zero, calculated by using June 1987 to December 2017 data for the MH weather station, are 175 mm (Nov.–Jan.), 261 mm (Dec.–Feb.), 367 mm (Jan.–Mar.), 450 mm (Feb.–Apr.), 565 mm (Mar.–May), 663 mm (Apr.–Jun.), 702 mm (May–Jul.), 799 mm (Jun.–Aug.), 768 mm (Jul.–Sep.), 627 mm (Aug.–Oct.), 317 mm (Sep.–Nov.), and 163 mm (Oct.–Dec.).
Figure 7. The change rates [%] in three-month cumulative rainfall between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. The calculation of the percentage in the figure is defined as (Future estimates − Baseline estimates)/(Baseline estimates) × 100%. The number in each cell of the figure represents the median of all GCM/ESM results for a specific scenario, 3-month period, and Period. The color of the cell represents the degree of change rate: white indicates 0%, blue indicates >0% (the darker the blue, the more positive the change), and red indicates <0% (the darker the red, the more negative the change). The three-month cumulative rainfall levels corresponding to SPI-3 equal to zero, calculated by using June 1987 to December 2017 data for the MH weather station, are 175 mm (Nov.–Jan.), 261 mm (Dec.–Feb.), 367 mm (Jan.–Mar.), 450 mm (Feb.–Apr.), 565 mm (Mar.–May), 663 mm (Apr.–Jun.), 702 mm (May–Jul.), 799 mm (Jun.–Aug.), 768 mm (Jul.–Sep.), 627 mm (Aug.–Oct.), 317 mm (Sep.–Nov.), and 163 mm (Oct.–Dec.).
Water 16 01746 g007
Table 1. The established operation rules of Baoshan Reservoir (BR) and Baoshan Second Reservoir (BSR) every ten days.
Table 1. The established operation rules of Baoshan Reservoir (BR) and Baoshan Second Reservoir (BSR) every ten days.
ReservoirReservoir Operation Established Operation Rules R 2
BRQout,BR,t 0.178·VBR,t−1 + 1.637(2)0.51
Qin,BR,t>ORCIf Qt < 1000, then Qin,BR,t = 3.84 + 0.015·Qt + 0.17·VBR,t−1; otherwise, Qin,BR,t = 125(3)0.48 *
≤ORC57.166·ln(Qt) − 260.71(4)0.78
BSRQout,BSR,tIf VBSR,t−1 > 9.535·Average daily HSP water usage + 0.04·(VBR,t−1 + VBSR,t−1)+54.31 − Qout,BR,t, then Qout,BSR,t = 9.535·Average daily HSP water usage+0.04·(VBR,t−1 + VBSR,t−1)+54.31 − Qout,BR,t; otherwise, Qout,BSR,t = VBSR,t−1(5)0.53
Qin,BSR,t>ORCIf Qt < 1000, then Qin,BSR,t = 242.1 + 0.36·Qt − 0.062·VBSR,t−1; otherwise, Qin,BSR,t = 250(6)0.36 *
≤ORC0.9067·Qt − 38.376(7)0.93
Notes: ORC: operation rule curve; Qout: outflow from the reservoir; Qin: inflow to the reservoir; Q: summation of every ten-day river discharge at Shangping flow gauge; V: water storage volume in the reservoir; average daily HSP water usage = 14.67 (data from August 2018 to December 2021); the unit for all the variables is [104 m3]; *: R2 is only for the regression equation.
Table 2. Calibration results of parameters in the SWAT for the Shangping River Watershed, including the best-fit parameter values and the final ranges (within the Min. and Max. values) for the 95% prediction uncertainty (95PPU).
Table 2. Calibration results of parameters in the SWAT for the Shangping River Watershed, including the best-fit parameter values and the final ranges (within the Min. and Max. values) for the 95% prediction uncertainty (95PPU).
ParameterDescriptionBest FitMin. ValueMax. Value
CN2Curve number for moisture condition II [%]−0.36−0.38−0.19
ALPHA_BFBase flow alpha factor [-]0.540.450.76
GW_DELAYDelay time for aquifer recharge [day]1.160.0022.8
GWQMNThreshold depth of water in shallow aquifer for return flow to occur [mm]1777546.81849
GW_REVAPGroundwater revap coefficient [-]0.110.080.25
RCHRG_DPDeep aquifer percolation coefficient [-]0.17−0.240.27
REVAPMNThreshold depth of water in the shallow aquifer for revap to occur [mm]6.720.00181.6
SOL_AWCAvailable water capacity of the soil layer [mm]−0.06−0.20−0.02
ESCOSoil evaporation compensation factor [-]0.060.000.30
SURLAGSurface runoff lag coefficient [-]15.714.424.3
Table 3. The annual economic losses [USD 1 million] in semiconductor production estimated from the difference in annual reservoir outflow between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. The economic loss is defined as (Future estimate − Baseline estimate) × 450 USDs/m3, where the estimate is the annual outflow of BR and BSR. The number in each cell of the table represents the median of all GCM/ESM results for a specific scenario and Period.
Table 3. The annual economic losses [USD 1 million] in semiconductor production estimated from the difference in annual reservoir outflow between the Baseline Period and the S, M1, M2, and L Periods under the RCP2.6, RCP4.5, RCP6.0, and RCP8.5 scenarios. The economic loss is defined as (Future estimate − Baseline estimate) × 450 USDs/m3, where the estimate is the annual outflow of BR and BSR. The number in each cell of the table represents the median of all GCM/ESM results for a specific scenario and Period.
SM1M2L
RCP2.6−122−511104−30
RCP4.5−38−349−238−410
RCP6.0−481−1036−753−484
RCP8.5171−385−289−686
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Lee, T.-Y.; Lai, Y.-P.; Teng, T.-Y.; Chiu, C.-C. Impact Assessment of Climate Change on Water Supply to Hsinchu Science Park in Taiwan. Water 2024, 16, 1746. https://doi.org/10.3390/w16121746

AMA Style

Lee T-Y, Lai Y-P, Teng T-Y, Chiu C-C. Impact Assessment of Climate Change on Water Supply to Hsinchu Science Park in Taiwan. Water. 2024; 16(12):1746. https://doi.org/10.3390/w16121746

Chicago/Turabian Style

Lee, Tsung-Yu, Yun-Pan Lai, Tse-Yang Teng, and Chi-Cheng Chiu. 2024. "Impact Assessment of Climate Change on Water Supply to Hsinchu Science Park in Taiwan" Water 16, no. 12: 1746. https://doi.org/10.3390/w16121746

APA Style

Lee, T. -Y., Lai, Y. -P., Teng, T. -Y., & Chiu, C. -C. (2024). Impact Assessment of Climate Change on Water Supply to Hsinchu Science Park in Taiwan. Water, 16(12), 1746. https://doi.org/10.3390/w16121746

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop