Water Policy, Economics, and Systems Analysis
Resolving the scale incompatibility dilemma in river basin management
 This study illustrates how integrated river basin management can conflict with our increased emphasis on decentralizing water resources decision making. For over a decade, water and environmental decision making in many countries has been shifting from national levels to state/province and local levels. At the same time we have increasingly found that it is critical to consider how individual water resource decisions impact the river basin. We provide detailed examples of this incompatibility dilemma from the United States and Turkey as well as smaller examples from Japan and Macedonia. We argue that new institutional models are required for effective river basin management and that implementation of such models can be evaluated through the use of transaction costs. This study concludes with examples of institutional arrangements that can help bridge the incompatibility gap.
 Water management is currently trapped between two conflicting paradigms: as science demonstrates the need to integrate water management both spatially and temporally, decision making is becoming more localized and site-specific. We now recognize that effective management of river basins and water resources requires that managers understand the implications of their actions throughout a larger area (e.g., river basin, catchment, watershed ecoregion). Conversely, managers must accept that decision making around the world increasingly is being decentralized and shifted to a more local scale. This latter paradigm is a result of increasing public participation in a more collaborative model of decision making being implemented in many countries worldwide. That is, science supports an integrated, river basin-scale approach that requires a coarse-scale view while politics, local interests, and individualized decision making are forcing a fine-scale view. The conflict between these two conditions characterizes what we term the scale incompatibility dilemma.
 For more than a decade, environmental decision making in the United States and other countries has been shifting in locus from national levels to state/province, regional/district, and local levels. Significantly, the U.S. Environmental Protection Agency (U.S. EPA) has changed policy from command and control to a more collaborative and cooperative approach and from direct federal government responsibility to federal oversight of local, state, and tribal responsibility [U.S. EPA, 1998]. The shifts came partially from congressional pressure for environmental regulatory reform and partially from local communities that became tired of environmental decisions being imposed on them by government and industry. A consensus among a range of decision makers decided that environmental quality can and should be upheld more effectively through local change and collaboration to prevent pollution rather than through traditional command-and-control or end-of-the-pipe approaches [Murdock and Sexton, 1999]. Local mistrust of top-down directives from federal and state agencies also forced the U.S. EPA to change, as local landowners demanded and received greater control over managing specific elements of the landscape (e.g., plots of land, stream side buffer zones, wetlands, and irrigation systems). Developing countries also have moved toward decentralized water management particularly in river basin management and irrigation projects. This was partly due to pressure from the World Bank and partly due to the rising cost of water management and the difficulty of getting farmers to pay for operations and maintenance or to participate in river basin management projects [Easter, 1993]. Both Mexico and Turkey are countries where the decentralization has been fairly rapid in their public irrigation projects. In Mexico the management turnover seems to have improved the collection of water charges and improved water use efficiency, while in Turkey it has improved irrigation water delivery but not drainage [Hearne, 2004; Scheumann, 2002].
2. Integrative Management
 During the last 30 years, the applied ecological approach to water resource management has shifted from point- and population-specific questions to ecosystem-level, landscape-level, and cumulative effects questions. Pioneering work in the 1970s increased our knowledge of the links between stream characteristics and the surrounding landscape [Likens and Borman, 1974; Hynes, 1975]. In the 1980s, many scientists showed that hierarchy theory could help us understand the structure and function of stream ecosystems [Allen and Starr, 1982; O'Neill et al., 1989; Vannote et al., 1980; Minshall et al., 1983; Frissell et al., 1986]. In the 1990s, further research showed that behavior of individual ecosystems is controlled by dominant processes (e.g., energy flux, nutrient availability, and geomorphology), but the process that is dominant in any given case changes as we move among spatial scales [Ehleringer and Field, 1993; Petersen and Parker, 1998; Brezonik et al., 1999]. This revelation provides an explanation of why “the whole is greater than the sum of the parts.”
 Advances in conceptual design (i.e., a more explicit attention to scale influences) improved our understanding of the spatially hierarchical organization of streams and rivers. Clearly, water resources integrate with and influence their landscape as hydrology drives sequential changes in downstream water resources. Therefore landscapes containing streams and lakes are equally hierarchical [Pahl-Wostl, 1998]. Variations in stream biological integrity and movement of water, sediment, and pollutants through a landscape are controlled at several spatial and temporal scales. In many cases today we are able to model those relationships through space by modeling changes across broad regions [e.g., Omernik and Bailey, 1997; Brezonik et al., 1999].
 Thirty years ago, river basin and water management focused on achieving one or a few goals such as sustained yield of a fish population, reduced soil loss, or increased irrigation. Since then, many countries have begun to reorient their goals and policies. In fact, after the International Conference on Water and the Environment (Dublin, 1992) and the Earth Summit (Rio de Janeiro, 1992), many countries signed postconference declarations supporting this approach. Those countries now frequently express their goals as integrated river basin water and water management as recommended by World Bank . This implies understanding multiple connections between natural resources within a hydrologic unit and recognizing that achievement of multiple goals in such a unit requires attention to those links. River basin and water management are interdisciplinary sciences that unite biophysical and socioeconomic disciplines and strive to optimize the societal goals for a landscape and water resource within an integrated framework [Brooks et al., 2003; Perry and Vanderklein, 1996; Berger and Ringler, 2002].
3. Decentralized Decision Making
 In most parts of the world in the last 30 years, there has been a move toward increasing local-level decision making. In this new paradigm, strategies for integrated water and river basin management become a summation of and are constrained by lower-level management practices, in contrast to “the whole is greater than the sum of the parts” as described in section 2. The goals of local governments, citizen groups, or firms are often based on short-run considerations and localized gains. These goals frequently are translated directly into management practices that become de facto policy. Because of the de facto nature of that decision making, there is a need to develop institutional arrangements to evaluate these decisions and their influence in the larger context (i.e., the river basin). In this context, effective implementation of river basin management principles requires models and institutional arrangements that can help resolve the scale incompatibility dilemma and guide local-level decision making toward a more integrated approach. These models and institutional arrangements incorporate knowledge about the structure and function of river basins or catchments.
 The shift in management paradigm from top-down and command-and-control to citizen involvement and local decision making is a significant but not unexpected result of the trends in national level policy that have been evident in the United State, Canada, and western Europe since the late 1960s [Fri, 1999]. Prior to the 1960s, environmental goals focused primarily on control of point source discharges to air and water. Standards and their enforcement varied from state to state and province to province. Environmental advocates feared that competition for economic development would drive future standards to the level of the lowest common denominator (i.e., the least water quality protection tolerated by the most people) [Fri, 1999]. Establishment of federal control in the U.S. Environmental Protection Agency and national, enforceable environmental standards in the United States and other western countries offered a solution to this problem. Water quality in these nations has improved as a result of subsequent federal clean water initiatives and regulation of large point sources of lake and river pollution.
 However, point source water quality impacts comprise only one element of the various forces that determine local and national environmental quality. Nonpoint sources, which are more difficult to regulate, often represent >85% of the water quality problems in a given river basin [Perry and Vanderklein, 1996]. Throughout the last 20 years, individuals across the United States increasingly have raised concern about local environmental problems; these local concerns are often idiosyncratic and endemic to small areas. Increased public involvement and less autocratic approaches have resulted in dramatic increases in the influence of local decision making [Gerlach, 1993].
 Local-scale environmental decision making differs in practice and philosophy from regional or river basin decision making. Every day, agencies and individuals make choices about the local environment at the scale of farm, small catchment, town, shire, or township. Typical environmental issues at this level include water quality, protection of critical natural areas, waste management, growth, and infrastructure concerns and historic or cultural preservation [English and Dale, 1999]. Local environmental decisions are made within the context of state and federal laws, local laws, and incentives, while cultural values are intended to balance the needs of local self-interests.
 The impact of local activities may be difficult to predict at the coarse scale because one does not know what other local actions will be taken. In addition, the cost of preventing local environmental damage appears right away, while local environmental impacts as well as the downstream (basin scale) impacts can occur much later [Gregersen and Easter, 1994; Goolsby et al., 1999]. The environmental damage created by one farm's practices often appears very small. This is especially true in the short term, when the cost of prevention is immediate and may be prohibitive for an individual farmer [Gregersen and Easter, 1994]. Yet at the coarse scale of a river basin, cumulative impacts of many small-sized polluters can cause major downstream damage [Goolsby et al., 1999]. In contrast, the cumulative loss in production from altering production practices will be small at the state or national level because of our general food abundance and our link to international markets.
 Managing river basins (as well as watersheds) comprehensively while still incorporating the unique requirements of local communities is an inherently challenging process. It requires society to institutionalize interdependence based upon biophysical connections from local to river basin scales. Careful river basin and water resource management recognizes and is based upon ecological interdependence: the actions of land managers in one part of the basin are reflected in downstream water quality and hydrologic conditions. In large river systems, widely separate ecosystems (e.g., the western U.S. Cornbelt Plains and the Gulf of Mexico) are linked through the transport of water, sediment, nutrients, organisms, and contaminants [Goolsby et al., 1999; Sparks, 1995]. This same perspective must be interlaced into local decision-making processes.
4. Case of the Minnesota River Basin as a U.S. Example
 The Minnesota River drains an intensively agricultural landscape in southern and western Minnesota; 92% of its 38,419 km2 basin is intensively cultivated [Minnesota Pollution Control Agency, 1994]. The river carries significant loads of sediment, nitrogen, phosphorus, pesticides, and pathogens. The river originates in eastern South Dakota and flows 539 km to its confluence with the Mississippi River. This is a classic lowland prairie river; average slope over the 539 km is 1.5% [Senjem, 1997]. Historically, the basin had extensive wetlands. More than 85% of those have been drained to increase agricultural productivity. Agricultural productivity interests dominated decision making in the basin until the early 1990s. More recently, a wide range of interest groups has emerged and the local, regional, and state agencies are attempting to envision and implement a variety of more integrated approaches to basin-scale and regional-scale approaches to management and decision making.
 The Minnesota Department of Natural Resources (MN DNR) and the Minnesota Pollution Control Agency enforce water use and water quality regulations developed for statewide application. Many Minnesota River Basin residents feel these regulations do not allow adequate flexibility to deal with local level conditions, and set unrealistic goals and limitations. Conversely, other critics argue that regulations (e.g., Best Management Practice requirements) are designed for site-specific problems and are not intended to address problems caused by the cumulative effects of many local management actions [MN DNR, 1998].
 Recent actions by a group of landowners in southern Minnesota exemplify the scale-incompatibility dilemma. Their actions demonstrate how implementation of local management practices can be incompatible with basin-scale management goals. A group of farmers in a small catchment nested within the Minnesota River Basin petitioned their county board in 1997 for permission to construct a drainage improvement project. Some residents protested about loss of wetlands and impacts to environmental quality, resulting in a court battle [Dovciak and Perry, 2000]. Reactions to this proposal show a deep-seated conflict of interest within the area between farming and environmental groups, both of whom reside in the basin.
 Many of the lands in the Minnesota River Basin are or historically were hydrologically saturated: in the late 1700s, thousands of wetlands dominated the landscape. Tile drainage lines removed the “excess” water and increased agricultural production on the land. Decisions about drainage (e.g., permission to expand drainage projects) are made at the county level. Drainage management is jointly controlled by state and federal law and theoretically designed to incorporate protection of wetlands and downstream property owners. Yet the drainage law clearly is intended to serve agricultural interests, while providing relatively minimal protection for environmental quality.
 Maintaining this balance between economic and environmental impacts creates challenges because of the negative effects of drainage on the environment. Drainage causes water quality problems through rapid transport of nitrogen, pesticides, and sediment from fields to streams; it causes hydrologic problems such as increased stream velocity and stream bank erosion. Drainage is responsible for biological problems through changes in water chemistry, habitat, and stream ecology. Those impacts are cumulative through space, but the changes are not linear. Downstream conditions are often more severely impacted than predicted by the accumulated upstream impacts. For example, major sources of nitrogen are the rivers and streams in southern Minnesota, Iowa, Illinois, Indiana, and Ohio, which help cause the hypoxia problem in the Gulf of Mexico. The Minnesota River alone is estimated to contribute ∼3.5% of the nitrogen load to the Gulf of Mexico [Goolsby et al., 1999].
 In the context of these cumulative impacts the farmers mentioned earlier [Dovciak and Perry, 2000] requested and received permission to “improve” a drainage system (i.e., to increase the efficiency with which water is removed from the landscape). Their design called for the increased water volume to enter a small tributary that flowed into a trout stream whose temperature and water quality were protected under MN DNR (i.e., state) regulations. Local and regional environmental groups and one resident in the catchment protested this change on the grounds that increased water volume would degrade stream conditions.
 The conflict between the drainage interests and environmental interests characterizes the sharply divergent viewpoints of economic gain and environmental protection. Environmental groups did not view the drainage project as an isolated act of misguided management. Rather, they believed the project exemplified and contributed directly to cumulative effects in the Minnesota River and Upper Mississippi with significant impacts on the Gulf of Mexico. The farmers who proposed the project saw this as a neighborhood and community decision and felt that outside groups did not have a basis for concern or intervention. Furthermore, some local people argued that they as farmers were being asked to fund changes and actions valued by urban people (i.e., increased agricultural costs to serve the value set of urban and suburban people).
5. Case of the Seyhan River Basin as a Turkish Example
 These two paradigms (i.e., local control and integrated management) also are in conflict in other countries. One good example is found in Turkey's Lower Seyhan irrigation project in which a concerted effort has been made to turn over more of the irrigation and drainage responsibilities to water user organizations (WUOs). The Lower Seyhan irrigation system was designed to irrigate 175,000 ha and is bordered in the north by the Taurus Mountains and in the south by the Mediterranean Sea. The region is one of the most fertile areas of Turkey with a Mediterranean climate throughout the year, a climate favorable to many crops including cotton, citrus, cereals, and vegetables. The topography is very flat with a maximum elevation of 60 m in the north, dropping to 0–0.4 m in the south. An impermeable soil layer results in accumulated groundwater near the surface. Consequently, a superficial water horizon has been established near the ground surface and sizable areas become waterlogged during winter. This is followed in summer by evaporation, which lowers the groundwater level, leaving behind detrimental salts. Since natural conditions do not provide sufficient natural drainage, ∼105,000 ha in the proposed irrigation project were slightly too highly saline. Therefore a significant part of the area would have needed an artificial drainage system to improve agriculture even without the irrigation project [Scheumann, 2002].
 The Seyhan Dam was completed in 1956, and the volume of water was sufficient to irrigate the total project area. Between 1963 and 1987, during stages I, II, and III, irrigation and drainage facilities were constructed on 133,000 ha, or ∼70% of the total project. Although the last stage was postponed, ∼30,000 farm families are currently served by the project with two thirds of them owning land. The capacity of the delivery system is designed to serve the total project area on a 24-hour basis, and a vast network of tertiary, secondary, and main drains discharge agricultural effluents into the Mediterranean Sea [Scheumann, 2002].
 Before the operation and maintenance of the irrigation delivery system were turned over to the farmers, the lower sections of the system had increasing salinity and water-logging problems due to irrigation. Once operations and maintenance (OM) of the irrigation system were released to the water user groups during 1993–1994, both operations and maintenance of the system improved, and water supplies were more predictable. Yet the salinity and water-logging problems continued because there was no incentive mechanism to encourage farmers or WUOs to consider the impact of their irrigation and drainage decisions on downstream farmers and WUOs. When government had “full” control, it could, at least, limit water use and maintain drainage ditches. Currently, there is little incentive for upstream irrigators to do either [Scheumann, 2002].
 Scheumann  finds the policy decisions by the Turkish government have been too narrow and have failed to establish the necessary incentives so that WUOs will invest in farm drains and OM of the major drainage ditches. The large size of the irrigation project and the high costs of monitoring and enforcement meant that the state irrigation agency could not secure collective action on drainage and water-logging issues even though they had the support of WUOs. As she goes on to point out, a new level of collective action through cooperation among WUOs is needed when there are interdependent relationships among organizations and when the actions of one organizational unit, such as a WUO, are contingent on the actions of others [Scheumann, 2002].
6. Institutional Options
 Given the basic conflict between these two paradigms, what is needed are institutional arrangements that internalize the externalities that are created because of the hydrological interdependence of a river basin. Institutional arrangements set the ground rules for resource use. Institutes are “ordered relationships among people which define rights, exposure to rights of others, privileges and responsibilities” [Schmid, 1972]. So far, the emphasis has been on developing institutions that help decentralize management. What is now needed is an emphasis on developing institutional arrangements that help integrate these decentralized decisions. In other words, we need institutional arrangements that give local decision makers, community organizations, and user groups incentives to take into account the external, downstream impacts of their decisions. Also, these institutional arrangements should not have high transaction cost to develop and implement.
 In designing and developing institutional arrangements, one needs mechanisms to evaluate their likely effectiveness and cost. One means of making such an evaluation is to compare their relative transaction costs which include administrative costs and any cost required to enact the new institutions as well as the costs of monitoring and enforcement [McCann et al., 2004]. These costs can be significant and clearly favor existing institutional arrangements. There are other ways one could evaluate institutional arrangements, but for this study, we feel that transaction cost is the appropriate method to use.
 What types of institutional arrangements have helped internalize the external impacts of decentralized decisions at reasonable transaction costs? We offer several examples at spatial scales ranging from the river basin to the local landowner. We also suggest how high the transaction costs might be to install and implement some of the arrangements. One example is the requirement that a river basin authority or state board review certain local water decisions. One of these institutional arrangements is the requirement in the state of Colorado that permanent water transfers must be advertised, and if they are challenged by a third party, the trade must be reviewed by Colorado's water court which will determine if damage to third parties is important enough to cancel the water sale. In the states of Utah and New Mexico the Office of the State Engineer reviews all proposed water rights transfers to determine if external impacts are possible and if they are likely to be large enough to require the trade to be modified or rejected. Of these two options the water court has involved much higher transaction costs than the review by the State Engineer's Office has [Howe, 1998].
 A second institutional arrangement that has worked to get individual decision makers to coordinate their activities and reduce the negative stock externalities is the establishment of groundwater districts and associations of water users in those districts. The associations are then given authority to monitor district groundwater levels and restrict pumping rates when extraction levels exceed recharge levels by some specified amount. In some cases the reduction of extractions is controlled by pumping limits; in others, it is controlled by a tax per cubic meter of water extracted. If a tax is used, the tax revenue could be used to fund projects that would increase recharge rates and increase the stock of groundwater. Blomquist  describes such a water district in the southern part of California, and it is clear that the transaction costs are likely to be high in establishing and operating such districts. Thus we will not expect to see groundwater districts in areas that do not have serious groundwater problems.
 A third example of institutional arrangements to correct for off-site effects involves the problem of third-party economic impacts. One of the concerns in water trading, or transferring to other regions, is that there may be an economic loss in the region selling water and that water charges for operating the irrigation system will be much higher for each of the remaining water users. This is particularly true when those farmers selling water shift to dry land farming or fallow their land. The same total cost of operation must then be divided among fewer farmers.
 An example of an effective response to one of these third-party concerns is found in the Laganera region of Mexico, where an off-site water sale has occurred. To help offset the potential increase in cost of system operations for each of the remaining farmers, water buyers were required to credit 70% of the water charges for operations and maintenance in the new region back to the original water users association (seller's association) and 30% to the new association (buyer's association) to which the water was transferred. In the cases where municipalities are concerned about losing tax revenues when water rights are traded or transferred to other regions, the buyer of the water rights could be required to pay property taxes in the region from which the water rights are sold or make a lump sum payment to municipalities in the exporting region [Thobani, 1998]. This lump sum payment could also be set up as a direct payment to those downstream who are damaged by the reduced return flows and resulting change in streamflow patterns. These types of arrangement should not have high transaction costs and could easily be included or added to the law governing water trading. Enforcement of the payments may involve some added transaction cost but can be handled by existing water user organizations.
 Another example is the introduction of a point-nonpoint source pollution trading program that has the potential of reducing the cost of improving water quality. The program allows point sources of pollution to trade for pollution credit with nonpoint source (primarily agriculture). The saving occurs when it is cheaper for the nonpoint sources to reduce water pollution than it is for point sources. In the case of the Minnesota River the two trades that took place also cut overall pollution discharges from the combined point and nonpoint sources. Thus environmentalists, at the basin level, were pleased and point sources, at the local level, were happy since they were able to expand production. The transaction costs were high because of the cost of arranging and monitoring the trades with nonpoint sources [Fang and Easter, 2003]. A large share of the transaction costs were covered by the Minnesota Pollution Control Agency. For such a program to operate effectively in the future these transaction will need to be lowered since the pollution control agency is not likely to want to continue paying most of the transaction costs.
 A fifth alternative, at the subregional scale, involves a well-functioning land market and downstream organizations willing to make use of that market. Before 1920, Japanese irrigation associations and municipalities which were located downstream in the river basin were very active in trying to reduce water pollution and erosion in their basin. Several different approaches were used. First, they purchased land in critical areas in the upper basin and cost shared on management of upland forested areas. Later, municipalities and power companies shared the cost of purchasing upland forested areas. Finally, with the increase in water use, the prefectural government assumed more responsibility and leased privately owned land in the upper basin and planted trees. This was done in cooperation with the downstream water users and included a cost-sharing arrangement between the water users and the government in terms of both costs and revenue [Nickum and Easter, 1990]. Existing irrigation and municipal government helped hold down transaction costs. However, to do the same thing in another area might involve high transaction costs if existing local organizations are not interested or are inactive in dealing with river basin problems. In such cases the transaction cost of educational programs and organization efforts can be high.
 Two specific examples at this scale are the Kitikami Basin in eastern Japan and the Strezevo Irrigation District in Macedonia. In the Kitikami Basin, residents and administrators have developed and implemented an actual example of integrated river basin management [Shirai, 1990]. Residents of the lower basin are involved in upper basin decision making through educational exchanges. The prefecture maintains an educational web site that is intended to stress upstream-downstream relationships for better community development (http://www.iwate.thr.mlit.go.jp/kids_english/cyousatai_e/). The Strezevo irrigation district in southern Macedonia (J. A. Perry, unpublished data, 1998) is earlier in its institutional development but seeks the same goals. Downstream water users have subsidized upstream land management practices to reduce sediment loads. The irrigation district has held community meetings to increase awareness and share perspectives. However, there is not yet effective economic or water resource monitoring nor are there effective economic exchanges in the river basin.
 A final example, at the finest scale, is to develop rules that require payment for damages created in a stream or lake by landowners or operators. In some cases a management agency may require an outright ban on certain activities that cause large external impacts, such as using septic tank systems near lakes or rivers or having farmers hook their home septic system to the tile drains in their fields. In both of these cases the goal is to reduce effluent from being discharged directly into a water source used by others. If such prohibitions are put in place, an effective means for monitoring and enforcement will be required for the prohibition to be effective. An alternative approach would be to impose a tax on the owners of the septic systems and set the tax equal to the damages caused. This again would require monitoring as well as tax collections. In both cases the monitoring and enforcement should not impose high transaction cost. In contrast, collection of the tax may be difficult and costly.
 In all of these cases we stress the need for integrated approaches that simultaneously consider the perspective of a range of users. We suggest that the scale incompatibility dilemma is complicating water resource decision making but that does not have to be the case. Paying more careful attention to scale issues and to communicating across scales, especially among those interested in different spatial scales, will reduce the influence of the dilemma. In addition, we need to look at the economic incentives for resource use in the basin to make sure they do not encourage activities that cause damages downstream. Institutional arrangements can be devised to encourage better land use practices.
 We have to be more concerned about the incentives current government policies give local communities and individual decision makers. What incentives do local water or land users have to take into account the possible positive or negative impacts they may have on others within the river basin? Clearly, education can help to show users these interconnections; however, in many cases it will take more than education alone. Can we be innovative enough to develop institutional arrangements that allow downstream water users to help pay for land use practices that improve the flow of “clean” water from upstream users? It all comes down to incentives. Decision makers at each spatial scale must clearly see how an integrated, cross-scale approach is in the best interests of all those in the river basin. In some cases this may involve direct government action, such as subsidies for resource conservation practices, while in others it may involve setting up a pollution permit trading program. However, in other cases it may take new institutional arrangements that allow water users to organize their own collective action as has happened in Japan. In all cases it will require innovation in institutional design and significant investment in cross-scale communication.