Trouble reading this
report, click to
reverse
colour scheme
This section considers how a water property rights framework based on the concept of capacity sharing might be applied to a range of different classes of water supply systems. The water supply system is considered in isolation from the broader surface water network, with issues related to management of multiple connected supply systems, including inter-regional water trade, being discussed later.
The single storage system is the class of system usually envisaged when capacity sharing is being proposed. In such a system, each water user is allocated a right to a share of the storage, a share of inflows into the storage and a share of any relevant delivery capacity constraints. Marginal storage and delivery losses are applied to stored water and delivered water, respectively. Within the system, water trade involves trade in water at the point of storage (e.g. temporary trade), trade in shares of storage and inflow (e.g. permanent trade), and trade in shares of delivery capacity.
While in figure 3 a single water demand node is shown, in reality there are likely to be multiple water users spatially distributed along the network. To reflect this, delivery constraints and marginal delivery losses should vary with water user location. In practice there may be a degree of approximation/aggregation, where groups of water users are treated as a single demand node facing identical marginal losses and delivery constraints. Such an approach is adopted in St George and MacIntyre Brook in Queensland where a series of zones are defined within the irrigation areas, with higher delivery loss factors applying in zones further from the source (Hughes and Goesch 2009).
The principles outlined above can be transferred adequately to the case of an unregulated water supply system. Pott et al. (2005) and Pott et al. (2009) considered extending the concept of capacity sharing to systems dominated by unregulated flows, referring to this approach as fractional water allocation and capacity sharing (FCA-CS).
In an unregulated system, each user would be allocated a right to a share of system inflow at a defined point and a share of marginal delivery losses. The defined source node would potentially be a stream gauge, upstream of all the significant water users. Such a system could operate over various time scales depending on the variability of the system. Under a continuous system, the maximum pumping/diversion rate for each user would be defined by each user’s share of total stream flow, less marginal delivery losses. Under a discrete time system (e.g. hourly, daily etc.), a limit of water use per unit of time would be similarly defined.
Under this approach to property rights, the water resources of the system are fully allocated but never over allocated. In particular, the water available to users located near the bottom of the system would be independent of the actions of users at the top of the system. Under this approach, water trade (i.e. temporary trade) would involve trading shares of system inflows at the point of the source node, such that there would be no external effects associated with water trade between different users within the system.
As noted by Dudley (1990), the case of multiple storages in series, with no significant water users or tributary flows in between, may be adequately approximated as a single storage system. This is the approach taken in Queensland where the major storages, in both St George and MacIntyre Brook, and a number of smaller downstream weirs are aggregated and treated as a single conceptual storage. Water users are allocated a share of this total system storage capacity. The operation of the multiple storage system (transfers of water between storages would then remains the responsibility of the water management authority.
In general, there may be a need to define separate rights to individual storages (or to groups of storages) where some storages receive independent tributary flows, where there are significant differences in storage losses between storages, or where there are significant demand nodes located between storages. Where separate rights were defined, individual users would have control over the transfer of water between storages.
Systems with multiple storages not in series will likely require the definition of separate storage rights. The important distinction is the extent to which the inflows into the storages are independent and imperfectly correlated. For example, the case where multiple storages are in series, but the downstream storage receives significant inflows from an independent tributary, may also require definition of separate property rights.
Consider the multiple storage (not in series) system in figure 3. Water users would be allocated a right to a share of storage capacity in each of the storages and a right to the inflows in each storage. In effect, each user would have two separate water accounts and, therefore, two sources from which water could be obtained. Separate delivery losses, storage losses and delivery capacity constraints would also be defined, specific to the storage/source.
Under such an approach, water users would be collectively responsible for the management of the multiple storage system, including determining distribution of total storage reserves across the two storages and taking into account differences in losses and expected future inflows. Water trade would involve two separate commodities: water in storage 1 and water in storage 2. Given all water users have access to both storages, the market prices of these two commodities would be similar, although they could diverge as a result of differences in storage losses, delivery losses or delivery constraints.
Dudley (1990) suggested that, where the multiple storages have similar characteristics and highly correlated inflows (so no diversification benefits exist), users could each be allocated a share of only one of the storages. However, this suggestion was made in the context of urban water; in the context of rural water systems used primarily for irrigation, this approach is unlikely to be appropriate. This is because rural systems typically involve greater intertemporal variation in water use and storage levels, and greater heterogeneity in water user reliability preferences.
Figure 4 displays some examples of water supply systems with tributary flows. The same property rights principles discussed above can equally be applied to these types of systems. In system A in figure 4, water users would be allocated a right to a share of storage capacity, a share of inflows into the storage, and a share of flows from the tributary (at a defined source node). As with the case of multiple storages, each of these water sources would also have specific delivery loss factors and delivery constraints.
In system B, users at demand node 1 are located upstream of the confluence and so cannot take delivery of water in the unregulated tributary. Water users in this location would be allocated rights in the storage but not to unregulated tributary flows. While tributary flows would be entirely allocated to users at node 2, significant trade in water at the point of storage could still occur between the two demand nodes. This type of property rights framework, by embodying the concept of source tagging, would help to address the external effects that can be associated with water trade in systems with significant tributary flows. This is an issue that has been noted previously by Heaney et al. (2006) and Beare et al. (2005).
A simple static analysis of this situation is shown in figure 5. It is assumed that a single class of water entitlement is defined and water allocation and entitlement trade within the system is unrestricted. Problems arise where significant volumes of water entitlements (or allocations) are traded upstream of the tributary, from node 2 to node 1, specifically where users at node 2 trade more than their effective share of water in storage. Since it is not physically possible to deliver all of this water to users at node 1, a reduction in water allocations for all users becomes necessary. Under these conditions, even those water users who are not directly engaged in water trade may be adversely affected. In contrast, under the property right framework outlined above, users at node 2 could only trade their share of water that is physically in storage to users at node 1.
The property rights framework could equally be applied to systems such as C in figure 4, where there is significant water use occurring on the tributary itself. Physical network constraints prevent direct trade in water between node 1 and node 3. These constraints are embedded within the proposed property rights framework: only nodes that are connected to a common water source can engage in water trade (nodes 1 and 2 can trade water in storage, nodes 2 and 3 can trade in tributary flows). However, nodes 1 and 3 can, subject to water availability at each source node, engage in an indirect form of trade through trade with node 2.
SunWater pers. comm. (2008) have proposed a simple approximate approach for dealing with tributary flows, which involves adjusting (increasing) delivery loss factors to reflect availability of tributary flows which supplement storage releases. However, such an approach would only be appropriate where tributary flows are relatively small and constant over time.
The property rights framework outlined above can be applied more generally for any water supply system involving multiple regulated and unregulated water sources, multiple water demand nodes and multiple delivery capacity constraints. Consider the general water supply system of 
water sources, with 
less than regulated water sources, 
water demand nodes and 
relevant delivery capacity constraints. The key parameters involved in defining and managing (i.e. water accounting) this water property rights system are listed below:
 |
User |
 |
User |
 |
User |
 |
User |
 |
Indicator variable = 1, if water source |
An example of this framework is shown in table 1 for the system illustrated in figure 6.
1 Hypothetical parameter values for system C |
|||||||
|---|---|---|---|---|---|---|---|
| f | i=1 |
i=2 |
i=3 |
dlf |
i=1 |
i=2 |
i=3 |
| n=1 | 0.5 |
0.5 |
0 |
n=1 |
0.2 |
0.3 |
1 |
| n=2 | 0 |
0.5 |
0.5 |
n=2 |
1 |
0.2 |
0.1 |
| s | i=1 |
i=2 |
i=3 |
c |
i=1 |
i=2 |
i=3 |
| m=1 | 0.5 |
0.5 |
0 |
d=1 |
0.5 |
0.5 |
0 |
d=2 |
0 |
1 |
0 |
||||
d=3 |
0 |
0.5 |
0.5 |
||||
| ic | n=1 |
n=2 |
|||||
| d=1 | 1 |
0 |
|||||
| d=2 | 1 |
1 |
|||||
| d=3 | 0 |
1 |
|||||
| Note: Shaded cells are fixed hydrological constraints, non-shaded can vary with trade | |||||||
The above parameters are specified for a fixed point in time. In practice, storage, flow shares and delivery constraint shares may change over time as a result of trade between users. Further delivery loss factors may be specified in a state dependent way (dependent on prevailing weather conditions etc.). The indicator variable ic is necessary in systems with tributary flows because multiple water sources may delivery water through the same delivery constraint.
Implementation of capacity sharing involves a range of practical challenges; many of these are discussed at length in Hughes and Goesch (2009a, 2009b). Some of the challenges specifically associated with complex water supply systems are discussed below.
One important issue is the initialisation of shares: the process of converting existing water entitlements into shares of various water sources. Ideally, newly created bundles of water source shares will provide relatively equivalent claims to water as previous user entitlements. It is important to emphasise that this is a distributional issue and not an efficiency issue. In single storage systems, with a single water entitlement class, this process is less challenging. For example, as in St George and MacIntyre Brook, water users can be allocated a share of system storage and inflows equal to their share of total entitlement volume. However, even in these systems the introduction of marginal delivery losses has the potential to induce some distributional effects, as was observed in MacIntyre Brook (Hughes and Goesch 2009b).
The initialisation process is likely to be more difficult in complicated systems with multiple water sources. In particular, the process is likely to be challenging in systems where different users have access to different subsets of the available water sources. A static example is given in figure 7. Under a capacity sharing system (without trade), user water availability levels can, at best, only match those achieved under the entitlement system in one of the two states. Any (hydrologically feasible) allocation of water between the two users can be achieved as a result of trade, so there are no efficiency implications, but there are distributional implications.
This simple static framework does not necessarily tell the complete story. For example, where users have the ability to maintain varying levels of storage reserves, this would provide more flexibility, which would potentially allow previous entitlement outcomes to be matched more closely. While a dynamic analysis of the above scenario remains a subject for future research, the point remains that the initialisation of shares is likely to be a more difficult task in complex systems.
One area in which complex systems may involve some additional cost is information or decision-making burdens being shifted to water users. Traditionally, in systems with multiple water sources, much of the decision-making complexity is centrally managed and hidden from water users. Under the approach proposed above water users may be required to actively manage multiple water sources, which could involve some additional information, learning and decision-making costs. In particular, irrigators would need to be able to obtain and interpret information on expected inflows and expected losses for multiple water sources. In contrast, in simple systems capacity sharing is less likely to impose significant additional information costs. For example, Hughes and Goesch (2009b) found no evidence that the adoption of capacity sharing exposed irrigators at St George or MacIntyre Brook to any significant time or information costs.
The potential for information and decision-making costs raises the issue of implementing capacity sharing at an aggregated level, such as the water utility level, water demand node level or for smaller groups of irrigators, which is a possibility regularly put forward by Dudley (Dudley 1992). For example, groups of water users in a similar location with similar water needs could potentially manage a large capacity share cooperatively, or alternatively appoint a manager on their behalf. The appropriate aggregation would then depend, among other things, on relative information costs. As noted by Dudley (1992) it is in this sense that capacity sharing is consistent with the concept of common property as advocated by Quiggin (1988).
Another issue to consider is the effect of this type of property rights system on transaction costs of water trade. It might be argued that with a greater number of different types of water rights there would be more transactions and greater transaction costs. However, as Hughes and Goesch (2009a) noted, greater flexibility to manage water and storage resources may actually act to limit seasonal trade requirements. Further, it might be argued that the incorporation of hydrological constraints into the property rights framework could lower transaction costs. By reducing the potential for external effects associated with trade, less government oversight may be required, which may act to streamline trading processes.
Finally, there is the issue of the timing of water delivery. In regulated systems users would be required to order water, which could operate (as in Queensland) on a daily time scale (Hughes and Goesch 2009b). In an unregulated system, users could potentially extract their share of source inflow in real time (i.e. there need not be a time delay applied), as long as sufficient water is present. In highly variable systems or in times of high water scarcity there might be a case for implementing a system of time delays (for both regulated and unregulated sources), where water extraction would be delayed by the approximate time it takes water to be delivered between source and demand nodes.