• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Screening the Relocation Strategies of Water Quality Monitoring Stations By Compromise Programming1

Posted on: Sunday, 23 October 2005, 03:01 CDT

By Ning, Shu-Kuang; Chang, Ni-Bin

ABSTRACT:

Water quality monitoring network designs historically have tended to use experience, intuition, and subjective judgment in siting monitoring stations only sporadically. Better design procedures for optimizing monitoring systems with respect to multiple criteria decision analysis had rarely been put into practice up front when the needs for intensive monitoring became critical. This paper describes a systematic relocation strategy that is organized to identify several significant planning objectives and consider a series of inherent constraints simultaneously. The planning objectives considered in this analysis are designed to enhance the detection possibility for lower compliance areas, reflect the emphasis for different attainable water uses at different locations, promote the potential detection for the lower degradation areas of pollutants, increase the protection degree of those areas with higher population density in the proximity of the river system, and strengthen the pre-warning capability of water quality for water intakes. The constraint set contains the limitations of budget, the equity implication, and the detection sensitivity in the water environment. A case study in the Kao-Ping River Basin, South Taiwan, demonstrates the application potential of this methodology based on a seamless integration between the optimization and the simulation models. It enables identification of the optimal locational pattern stepwise using the embedded screening and sequencing capacity in a compromise programming model. However, a well calibrated and verified water quality model is an indispensable tool in support of this multiobjective evaluation. Extra sampling procedures become necessary for the sites with sparse environmental information. Comparison of planning outcomes of compromise programming is made against previously achieved analyses by using weighted programming and fuzzy programming.

(KEY TERMS: rivers/streams; monitoring network; water quality; watershed management; compromise programming; multicriteria decision making.)

Ning, Shu-Kuang and Mi-Bin Chang, 2005. Screening the Relocation Strategies of Water Quality Monitoring Stations by Compromise Programming. Journal of the American Water Resources Association (JAWRA) 41(5):1039-1052.

INTRODUCTION

To learn the processes, interactions, and impacts of environmental changes on livelihood, a water quality monitoring network is one of the vital environmental monitoring systems that are often installed to address regional water quality concerns at numerous time scales in an aquatic environment, such as river, estuary, reservoir/lake, wetland, or ground water system. The monitoring network was often designed to facilitate sustained monitoring of regular and emergency release attributes with regard to ecosystem features, environmental quality indices, human health factors, and economic impacts. The data from these sampling stations may help the decision makers identify problems, document improvements, and demonstrate overall trends in water quality. Thus, it is crucial to review the current network design procedures and develop basic guidelines to be followed in the design, expansion, and relocation of surface water quality monitoring networks. In recent years, the adequacy of collected water quality data and the performance of existing monitoring networks have been seriously evaluated for two basic reasons (Harmancioglu et al., 1998). First, an efficient information system is required to satisfy the needs of water quality management plans and to aid in the decision making process (Harmancioglu et al., 1998). second, this system has to be realized under the constraints of limited financial resources, sampling and analysis facilities, and manpower (Harmancioglu et al., 1998). To ensure cost-effectiveness, an evaluation adopted for a water quality monitoring network should cover all relevant technical design features, including selection of sampling sites, sampling frequencies, variables to be monitored, sensors synergy, and sampling duration (Loftis and Ward, 1980).

The design issues relating to surface water quality monitoring networks have received wide attention since the 1970s (Moore, 1973; Beckers and Chamberlain, 1974; Lettenmaier, 1978; Ward, 1979). Various attempts were made in the 1980s to improve the monitoring efficiency with regard to the basic design criteria (Skalski and Mackenzie, 1982), optimization analysis (Groot and Schilperoot, 1983), comparisons of the features of fixed stations versus intensive surveys (Van Belle and Hughes, 1983), consolidation of the network design (Lettenmaier et al., 1984), importance of data collection (Whitfeld, 1988), and interpretation of monitoring outcomes (Ellis, 1989). Earlier studies in the 1990s included the work done by Smith and McBride (1990), Loftis et al (1991), and Esterby et al. (1992) that had covered the fundamental principles and applications in siting the water quality monitoring stations. Later, more advanced analyses utilized integer programming (Hudak et al., 1995), multiobjective programming (Harmancioglu and Alpaslan, 1992; Cieniawski et al., 1995), Kriging theory (Lo et al., 1996), goal programming (Ning and Chang, 2002), and fuzzy programming (Ning and Chang, 2004) for assessing complex issues of surface and ground water monitoring networks. The design principles, guidelines, strategies, and implementation for water quality monitoring and assessment that cover a broad sense of chemical and biological impact assessment were reviewed (Dixon and Chiswell, 1996; Timmerman et al., 1997; Leeds et al., 1997). Dixon and Gardner (1998) further described how the United Kingdom Acid Waters Monitoring Network (AWMN) worked collaboratively within a seven-year time frame. In particular, Harmancioglu et al. (1998) summarized the technologies and solutions for planning and designing various water quality monitoring networks. Information retrieval from the monitoring networks has turned out to be critical, recently leading to develop a variety of visualization techniques that might be useful for demonstrating important aspects of a water quality monitoring network (Boyer et al., 2000).

In the US, several state governments, such as Florida and Idaho, have implemented the water quality monitoring network program that is normally composed of four tiers (Idaho Department of Environmental Quality, 2002). Tier I monitoring is designed for meeting national goals under the Clean Water Act. Tier II is comprised of status and trend monitoring designed to answer province or state wide to regional questions. Tier III monitoring includes basin assessments and monitoring required for total maximum daily loads (TMDLs) development. Tier IV includes all monitoring tied to regulatory permits issued by the government and is associated with evaluating the effectiveness of best management practices (BMPs) or TMDLs. This paper presents a multiobjective evaluation that is equivalent to a basin-wide assessment for water quality monitoring (i.e., a Tier III analysis). It also uniquely shows a holistic approach of how to integrate an optimization scheme of compromise programming with QUAL2E simulation analysis for water quality monitoring network assessment (Ning et al., 2001; Ning and Chang, 2002). Compromise programming is a relatively recent decision science methodology. It is based on the notion of distance from an ideal solution, which combines the best and most useful features of both linear multiobjective programming and goal programming and is not limited to the linear cases; it can be used for identifying nondominated solutions under the most general conditions. The model allows prespecified goals leading to provide an excellent base and a flexible tool for interactive programming (Zeleny, 1982).

The objective of this study was to search for the optimal relocation strategy of water quality monitoring stations to meet the goal of long term monitoring of water quality variations in the Kao- Ping River Basin, South Taiwan. As a companion study of Ning and Chang (2002), this analysis emphasized indirect tradeoffs between distance based scalable objective functions and direct tradeoffs using an earlier weighted method so as to enable one to pinpoint and entail easily the compensatory implications of decision making. For the indefinite characteristic of planning objectives, another previous study, using the fuzzy mathematical programming technique, accomplished the purpose of uncertainty analysis desired in decision making (Ning and Chang, 2004). Therefore, the needs for an updated systematic assessment to explore the possible relocation strategy and expansion sequence of a monitoring network in the Kao-Ping River Basin motivated the following analysis.

STUDY AREA

The Kao-Ping River in Taiwan flows approximately 140 kilometers and drains towards the south part of the Taiwan Strait. With an area of 3,256 square kilometers, including the major administrative regions of Kaohsiung and Pingtung counties, the main stream of the Kao-Ping River originates from four small tributaries: Chi-San River, Liao-Nung River, Cho-Kou River, and Ai-Liao Riv\er (see Figure 1). From the confluence to the union with those tributaries at the LiLing Bridge, the river carries the name Kao-Ping River. But Liao-Nung River and Kao-Ping River are generally regarded as the integrated main stream of the Kao-Ping River system in many governmental management practices.

Figure 1. The Present Pollution Source Distribution in the Kao- Ping River Basin.

The water year in a hydrological sense can be divided into two seasons. The wet season generally covers the time period from May to October, and the remaining time period is the dry season. Although the mean annual rainfall in this river basin is close to 3,000 mm, over 90 percent occurs in the wet season. Summer is hot, with the temperature often reaching 36C. The period of high flow rate in the stream usually occurs in late spring and summer due to monsoons and typhoons. During the monsoon period, the Kao-Ping River flow increases to a level approximately 8 to 12 times higher than flow during the dry season. Uneven rainfall across seasons has resulted in issues of water resources redistribution in the winter and earlier spring that inevitably requires building more reservoirs for water storage.

The drainage area in the Kao-Ping River Basin is primarily used for agricultural production. Crops that are produced from the agricultural fields include: rice, sugar cane, pineapple, and a variety of vegetables. Livestock farming is an active agricultural activity. In addition, a number of small scale and medium scale industries are also found in the downstream region. In addition to meeting the water demand for agricultural production and industrial manufacturing processes, water pumped from the river system is also essential for drinking and personal hygiene in this area of Taiwan. Concern and attention in water resources management have been directed primarily to the uneven distribution of rainfall and stream flows over the dry and wet seasons. While four reservoirs are in operation in the neighboring Tseng-Wen River Basin for water storage, none exists in the KaoPing River Basin. Yet the siting and building of an offstream reservoir - Mei-Nung Reservoir - with its intake (i.e., Chiu-Chuang) located in the upper stream area of the Kao-Ping River system, have resulted in an intensive debate over its potential ecological impacts. At present, the transfer of water resources from the main stream of the Kao-Ping River to the Nan-Hua Reservoir, which is located in the Tseng-Wen River system, becomes an indispensable solution to improve the reliability of water supply for two big cities - Tainan and Kaohsiung - in the coastal region. Although the agriculture sector is always the largest user of water, the latest development of three large-scale industrial complexes in the Tseng-Wen River system requires more water transferred from the upper stream area of the Kao-Ping River in the wet season. Industrial water requirements will therefore increase substantially. For the Kao-Ping River system, which has a long history of higher biochemical oxygen demand (BOD) and ammonia-nitrogen (NH^sub 3^-N) due to inadequate disposal of manure from livestock farming, industrial effluents, and domestic wastewater discharges, proper utilization of river water as potable water has encountered a new challenge with regard to both technological and managerial requirements. Figures 2 and 3 describe the concentrations of BOD and NH^sub 3^-N in the Kaohsiung River Basin, where long term violation of water quality standards is salient. Continuous discharge of organic, degradable wastewaters into the middle and downstream areas of the Kao-Ping River system, where most water intakes are located in this region, has resulted in the need for a systematic policy for improving the water quality condition. Thus, a sound water quality monitoring system to support essential analyses has become an urgent demand.

SIMULATION ANALYSIS

Simulation analysis methods are an integral part of a complete monitoring program design, which should be considered in the initial planning stage. The simulation outputs that characterized the internal features of a river system will then largely dictate the later site selection process. The simulation model QUAL2E, which was used as a tool in this study, is a steady state model for tracing conventional pollutants in one-dimensional streams and well mixed ecosystems (Brown and Barnwell, 1987). It illustrates the important physical, biological, and chemical processes and their interactions for the particular water quality constituents of interest based on a set of partial differential equations. The governing equations of QUAL2E illustrate the effects of dispersion, advection, constituent reactions, and interactions among constituents. It allows multiple waste discharges, withdrawals, tributary flows, and incremental inflow and outflow. QUAL2E may consider up to 15 constituents, including a conservative mineral (C), algae (A), NH^sub 3^-N), nitrite-nitrogen (NO^sub 2^-N), nitrate-nitrogen (NO^sub 3^-N), organic-phosphorus (org-P), dissolvedphosphorus (dis-P), BOD, dissolved oxygen (DO), coliform, and radioactive material. This study is designed to be capable of predicting the variations of DO and the decay rate of BOD, total phosphorus, and NH^sub 3^-N along the river reaches.

To gain a deeper understanding of the water quality condition in river reaches, a large-scale sampling campaign was carried out during dry and wet seasons (Ning et al., 2001). Each of the dry and wet sampling programs consisted of a 96-hour survey during August 1998 and February 1999. There was a total of 45 sampling sites covering the river system from its origin to the estuary. After accomplishing the field and laboratory measurements through a quality assurance and quality control procedure, environmental databases were integrated with a hydraulic database and applied for model (QUAL2E) calibration and verification (Ning et al., 2001). In the simulation analysis, the total study length of 170 km in the KaoPing River system was discretized into nine river reaches that consisted of 85 computational elements. The calibrated QUAL2E model was applied directly for deriving a new parameter - the half-life distance, representing the distance required for the decay of half of the concentration for the constitutes of concern. The condition for calculation of half-life distance is set on the critical scenario of river discharge, which was the average flow rate in the dry season (i.e., October to March). The planning alternatives may vary over the design flow rate in the river, and an applicable water quality model is necessary for the projection of half-life distance. The assumption for dry season steady state condition has to be maintained during the whole modeling process. Once the decay versus distance associated with each designated constituent can be identified via the simulation analysis, aggregate detection sensitivity of all constituents of concern in relation to each candidate site may become available for use in the subsequent optimization analysis. Moreover, the QUAL2E model is also responsible for simulating the water quality situation of concern in candidate monitoring stations that have no records. Figure 1 indicates the 21 candidate monitoring stations that were identified in the river system, in which seven of them are current. Obviously, most of the existing stations are located in the downstream area close to the higher populated region.

Figure 2. Water Quality Monitoring of BOD in the Kao-Ping River Basin.

Figure 3. Water Monitoring of Amonia-Nitrogen in the Kao-Ping River Basin.

FORMULATION OF COMPROMISE PROGRAMMING MODEL

Kwiatkowski (1991) pointed out some essential objectives that are common to many large water-quality monitoring-networks. They were designed to: (1) provide information on the location, severity, regional or volumetric extent, frequency, and duration of noncompliance of variables of concern; (2) support information for measuring site specific or whole network responses to increase anthropogenic inputs or control measures, using trend analyses or cause and effect relationships and determine the presence of new or hitherto undetected problems, leading to proactive, rather than reactive pollution control measures; (3) supply information for development and application of predictive models for assessing the impact of new pollution sources and assessing various enforcement and management strategies; and (4) determine ecosystem health and identify significant changes from normal succession or expected sequential changes that occur naturally in aquatic ecosystems.

Five planning objectives were considered in this analysis. They were designed to protect: (1) the monitoring sites for lower compliance areas of water quality, (2) important locations with regard to attainable water uses, (3) lower degradation areas of specific pollutants, (4) regions of higher population density, and (5) upstream districts of potable water intakes. The constraint set contained the budget limitations, equity implications, and detection sensitivity in the water environment, eventually leading to a multiobjective, mixed integer programming model having five objectives to be compromised, subject to several inherent constraints. Since the difference in streamflow rate between the wet and dry seasons is substantial, only the dry season condition is a critical consideration in the design procedure. Sampling sites will be selected from a set of candidate sites for relocation and expansion. The following compromise programming looks for a set of compromise solutions instead of an optimal solution based on a multi objective evaluation framework.

Objective Function

The planning objectives are designed to increase the detection possibility for those lower compliance areas, to emphasize potential uses of the water body at different locations, to incre\ase the potential detection sensitivity by applying better location strategy, to address the degree of protection of those areas with higher population density in the proximity of the river system, and to monitor the water quality in the areas upstream of all water intakes. The formulation of the objective functions is illustrated as follows.

TABLE 1. The Environmental Database of Candidate Stations.

DECISION WEIGHT ANALYSIS

Considering multiple objectives requires further decision analysis to determine the preference weights so as to aid in the tradeoff procedure in the multiobjective, decision making process. It may require an optimization study to extract such information and evaluate the overall effectiveness of the locating strategy. In a previous study, Ning and Chang (2002) had proposed a new method for evaluating the preference weights via questionnaire investigation and goal programming analysis. Five main river basins in Taiwan were selected as candidates for assessment of how well those monitoring stations were sited in each river basin with respect to the five designated planning objectives. With a questionnaire designed explicitly to elicit the performance of existing networks, 15 experts in the field of environmental management were invited to form an evaluation committee and, sequentially, grade the monitoring system situated in each river basin. The possible range of the grade was defined in advance to have a set of more consistent outputs. Until all similar monitoring systems can be evaluated and graded in terms of a series of criteria, all experts on the evaluation committee may proceed with an overall ranking with regard to the integrated performance of each system. Then, using a preemptive goal programming analysis may derive final weighting factors. This approach may provide more meaningful outputs rather than taking a straightforward mean value over all individual grades subject to the constraint that confirms the normalization of weighting factors directly (Ning and Chang, 2002).

RESULTS AND DISCUSSIONS

The optimal relocation strategy of water quality monitoring networks based on the compromise programming analysis that can be realized as the optimization model was developed by the software LINDO. The outputs of planning scenarios are divided into three cases, depending on the value of the distance parameter s, as shown in Table 2. It is observed that conflict and compromise between these five planning objectives are obvious since the second objective tries to emphasize the importance of conserving the higher water quality regions in the upper stream area but the first and fourth objectives focus on the protection of those nonattainment areas close to the estuary region where the population density is higher. Besides, the first and second objectives exhibit higher impacts in the tradeoff process since the severe pollution areas are located downstream, with the source of potable water usually situated upstream. Equity concerns make the final selection of candidate sites those that are uniformly distributed in all tributaries.

No matter which value is selected for the distance parameter s, five candidate sites are always favored in the Liao-Nung River system when tradeoffs exist among objectives (Table 2). On the other hand, both existing monitoring stations (i.e., Y^sub 16^ and Y^sub 17^) that are located at the downstream area of this tributary were excluded in all runs due to lower detection sensitivity and potential for different attainable water uses. Only one station (i.e., Y^sub 21^) that is located in the upstream area of the Ai- Liao River system was selected as a monitoring station due to the recognition of the importance of use attainability and the sensitivity for the lower compliance condition. Nevertheless, varying the parameter value of s would generate quite a different siting strategy in the Chi-San River system due to the actual tradeoffs among objectives. The tradeoff mechanisms among differing objectives turn out to be more influential when considering those monitoring stations in the Chi-San River and Kao-Ping River simultaneously. Both requirements of increasing the potential for different attainable water uses and improving the protection of higher populated regions would dominate the final choice. This is evident in Table 2 when comparing the output conditions of s = 1 and s = 2. In the Kao-Ping River system, higher population density and worse water quality would drive at least four out of the five candidate stations to be selected when the total number of candidate sites remains the same.

Comparative studies for the relocation of a water quality monitoring network using different multiobjective evaluation approaches, such as the weighted approach and fuzzy approach, seem appealing in systems analysis. Table 2 suggests that using the fuzzy approach may result in quite different planning alternatives in comparison to those generated by either the weighted approach or compromise approach (Ning and Chang, 2002, 2004). The main discrepancies arise from the siting patterns in the middle sections of the Ai-Liao River and the Chi-San River. Yet there is no obvious difference between the planning alternatives produced by the weighted and compromise approaches. This does imply that the inclusion of decision weights assigned by decision makers would not disturb the weak tradeoff between objectives no matter whether those alternatives are selected within the compromise set or not.

Overall, the parameter value s was not sensitive enough to alter the major siting patterns as evidenced in these three planning scenarios. In the decision analysis, it appears that the effect of compensation between objectives is not so conspicuous in the optimization process. Finally, based on one of the planning scenarios in relation to the compromise programming framework (i.e., s = 1), the expansion sequence of selected monitoring stations could be derived by a step-wise procedure. The spatial priority analysis, as listed in Table 3, clearly provides a stepwise choice of expansion sequence in terms of the preference order of all candidate sties in the river system. It shows that Y^sub 11^ and Y^sub 31^, which are close to the water intakes in the upstream and middle reaches of the river, or Y^sub 41^, located in the populous area, are the most important sites to be built in the water quality monitoring system. The information of expansion sequence would be useful once the governmental budget becomes critical.

TABLE 2. The Optimal Planning Programs of Water Quality Monitoring Network.

CONCLUSIONS

Planning a sound water quality monitoring system in a river basin in response to the needs of a national or regional water quality monitoring goal is a complex and challenging task. This paper presents a multiobjective programming model with a view to optimizing the water quality monitoring network in a river basin. Extra sampling procedures become necessary for the sites with sparse environmental information. With the aid of the calibrated QUAL2E water quality simulation model, simulation analysis in the first stage is capable of determining the half-life distance along the river reaches for each candidate site based on the predicted environmental assimilative capacity. Such an application makes final assessment of the relocation alternatives achievable. To make successful simulation runs and formulate the representative optimization model, GIS performs the data archival, data analysis, and essential spatial analysis to aid in the entire analytical framework. Compromise programming technique provides a useful means to fulfill the flexible design of a water quality monitoring network via an unbiased and trustworthy decision procedure. However, decision makers still can emphasize the specific intention via the adjustment of weighting factors associated with the planning objectives considered in this analysis. Such analytical framework eventually leads to a successful selection of the compromise relocation strategy of water quality monitoring systems in the Kao- ping River Basin, South Taiwan, where the water resources and water quality management systems have to be intimately linked together to meet the goals of regional economic development. It can be concluded that part of the existing monitoring stations that were not sited at those strategic locations should be removed in the long run. Once the governmental budget runs into a critical situation, spatial priority analysis may clearly provide a step-wise choice of expansion sequence based on the preference order of all candidate stations in the network.

TABLE 3. Spatial Priority Analysis of All Candidate Sites.

1 Paper No. 04003 of the Journal of the American Water Resources Association (JAWRA) (Copyright 2005). Discussions are open until April 1, 2006.

LITERATURE CITED

Deckers C.V. and S.G. Chamberlain,1974. Design of Cost-Effective Water Quality Surveillance Systems. US EPA-600\1-74-404, Washington D. C.

Boyer, J.N., P. Sterling, and R.D. Jones, 2000. Maximizing Information From a Water Quality Monitoring Network Through Visualization Techniques. Estuarine, Coastal, and Shelf Science 50:39-48.

Brown, L.C. and T.O. Barnwell, 1987. The Enhanced Stream Water Quality Model QUAL2E and QUAL2E-UNCAS: Document and User Manual. EPA/ 600/3-87/007, U.S. Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia.

Cieniawski, S.E., J.W. Eheart, and S. Ranjithan, 1995. Using Genetic Algorithms to Solve a Multi-Objective Ground-Water Monitoring Problem. Water Resources Research 31(2):399-409.

Dixon, W and B. Chiswell, 1996. Review of Aquatic Monitoring Program Design. Water Research 30(9): 1935-1948.

Dixon, E. and M. Gardner, 1998. Analytical Quality Control for the UK Acid Waters Monitoring Network. The Science of the Total Environment 216:113-119.

Ellis J.C., 1989. Handbook on the Design and Interp\retation of Monitoring Programs. Publication NS 29, Water Research Centre, Medmenham, Finland.

ESRI, 1998. User's Manual for ArcView Version 3.1. Environmental Systems Research Institute Inc., Redlands, California.

Esterby S.R., A.H. El-Shaarawi, and H.O. Block, 1992. Detection of Water Quality Changes Along a River System. Environmental Monitoring and Assessment 23:219-242.

Groot S. and T. Schilperoot, 1983. Optimization of Water Quality Monitoring Networks. Water Science and Technology 16:275287.

Harmancioglu, N.B. and N. Alpaslan, 1992. Water Quality Monitoring Network Design: A Problem of Multi-Objective Decision Marking. Water Resources Bulletin 28(1):179-192.

Harmancioglu, N.B., O. Fistikoglu, S.D. Ozkul, V.P. Singh, and M.N. Alpaslan, 1998. Water Quality Monitoring Network Design. Kluwer Academic Publishers, Dorbrecht, The Netherlands.

Hudak, P.F., H.A. Loaiciga, and M.A. Marino, 1995. Regional- Scale Ground-Water Quality Monitoring Via Integer Programming. Journal of Hydrology 164:153-170.

Idaho Department of Environmental Quality, 2002. Surface Water Monitoring Strategy - Draft. Surface Water Quality Program. Available at http://www.epa.gov/nheerl/arm/documents/id_ monitoringstrategy.pdf. Accessed in May 2005.

Kwiatkowski, R.E., 1991. Statistical Need in National Water Quality Monitoring Programs. Environment Monitoring Assessment, 17(2- 3):253-271.

Leeds, G.J.L., C. Neal, H.P. Jarvie, H. Vasey, and D.V. Leach, 1997. The LOIS River Monitoring Network: Strategy and Implementation. The Science of the Total Environment 194:101-109.

Lettenmaier D.P., 1978. Design Considerations for Ambient Stream Quality Monitoring. Water Resources Bulletin 14(4): 884-902.

Lettenmaier D.P., D.E. Anderson, and R.N. Brenner, 1984. Consolidation of a Stream Quality Monitoring Network. Water Resources Bulletin 20(4):473-481.

Lo, S.L., J.T. Kao, and S.M. Wang, 1996. Water Quality Monitoring Network Design of Keelung River, Northern Taiwan. Water Science and Technology 34(12):49-57.

Loftis J.C., G.B. McBride, and J.C. Ellis, 1991. Considerations for Scale in Water Quality Monitoring and Data Analysis. Water Resources Research 27(2):255-264.

Loftis, J.C. and R.C. Ward, 1980. Cost Effective Selection of Sampling Frequencies for Regulatory Water Quality Monitoring. Environment International 3:297-302.

Moore S.F., 1973. Estimation Theory Applications to Design of Water Quality Monitoring Systems. Journal of Hydraulics Division, ASCE, pp. 815-831.

Ning, S.K. and N.B. Chang, 2002. Multi-Objective Decision-Based Assessment of a Water Quality Monitoring Network in a River System. Journal of Environmental Monitoring 4:121-126.

Ning, S.K. and N.B. Chang, 2004. Optimal Expansion of Water Quality Monitoring Network by Fuzzy Optimization Approach. Environmental Monitoring and Assessment 91:145-170.

Ning, S.K., H.W. Chen, H.Y. Hsu, L. Yang, and N.B. Chang, 2001. Assessing Pollution Prevention Program by QUAL2E Simulation Analysis for Water Quality Management in the Kao-Ping River Basin, Taiwan. Journal of Environmental Management 61(1):61-76.

Schrage, L., 1991. User's Manual for Linear, Integer, and Quadratic Programming With LINDO. Release 5.0, The Scientific Press, South San Francisco, California.

Skalski J.R. and D.H. Mackenzie, 1982. A Design for Aquatic Monitoring Programs. Journal of Environmental Management 14(3):237- 251.

Smith D.G. and G.B. McBride, 1990. New Zealand's National Water Quality Monitoring Network Design and 1st Year's Operation. Water Resources Bulletin 26(5):767-775.

Timmerman, J.G., M. Adriaanse, R.M.A. Breukel, M.C.M. Van Oirschot, and J.J. Ottens, 1997. Guidelines for Water Quality Monitoring and Assessment of Transboundary Rivers. European Water Pollution Control 7(5):21-30.

Van Belle, G. and J.P. Hughes, 1983. Monitoring for Water Quality: Fixed Stations Versus Intensive Surveys. Journal of Water Pollution Control Federation 55(4):400-404.

Whitfeld PH., 1988. Goals and Data Collection Designs for Water Quality Monitoring. Water Resources Bulletin 24(4):775-780.

Ward R.C., 1979. Regulatory Water Quality Monitoring: A Systems Perspective. Water Resources Bulletin 15(2):369-380.

Zeleny, M., 1982. Multiple Criteria Decision Making. McGraw- Hill, New York, New York, pp. 315-325.

Shu-Kuang Ning and Ni-Bin Chang2

2 Respectively, Associate Professor, Department of Civil and Environmental Engineering, National University of Kaohsiung, No. 700, Kaohsiung University Road, Nan-Tzu District, 811, Kaohsiung, Taiwan; and Professor, Department of Civil and Environmental Engineering, University of Central Florida, Orlando, Florida 32816 (E-Mail/Chang: nchang@mail.ucf.edu).

Copyright American Water Resources Association Oct 2005

Source: Journal of the American Water Resources Association

More News in this Category