DEVELOPMENT OF MODELS FOR ECONOMIC EVALUATION OF INTEGRATED MANAGEMENT OF QUANTITY AND QUALITY OF IRRIGATION WATER WITHIN RIVER CATCHMENTS
1043/1/04
February 2004

EXECUTIVE SUMMARY

Motivation

Past efforts to protect the quality of South Africa's water resources were concentrated mainly on the control of effluents from point sources. Despite these efforts an apparent deterioration in the water quality of the country's surface waters is being observed (Pegram, Quibell and Görgens, 1997); reasons being that in many catchments there are zones where nonpoint source (NPS) contributions are significant or even dominant. Quibell (2000) argued that a lack of legislative and regulatory authority on the one hand and poorly defined linkages between implementable management actions and the processes that lead to NPS pollution on the other hand hampered the management of NPS pollution sources in the past.

The National Water Act (Act36 of 1998) now for the first time provides the legislative means to target NPS pollution with specific source-directed measures (Quibell, 2000). No specific differentiation is made between point and NPS pollution in the National Water Act which allows for the development of source specific procedures that address both point and NPS pollution from the source. The waste discharge charge system is one way in which the Department of Water Affairs and Forestry is implementing the National Water Act. The basis for this system is the polluter pays principle. The theory behind polluters paying pollution charges was that the individuals must pay for the cost incurred as a result of their pollution (Taviv, Herold, Forster, Roth and Clement, 1999). However, in order to institute a system of waste charges the relevant authority must be able to identify who caused the pollution and precisely how much of it. The last mentioned is a necessary condition for any charge system based on the polluter pays principle (Taviv et. al, 1999). Due to the unique characteristics of agricultural NPS pollution it is not straightforward to quantify exactly who has caused the pollution and how much of it. In part this is due to the complex relationship between agricultural production and damages from water pollution involving physical, biological and economic links. How well NPS pollution control policy performed often depends on how well these links are understood (Ribaudo, Horan, and Smith, 1999).

Problem statement and formulation of objectives

Several processes govern NPS pollution. The first process is concerning the production of pollution emissions, and therefore characterises the link between agricultural production practices and the movement of pollutants in the field. Emissions from each field interact with emissions from other fields and the surrounding ecosystem as they move to water resources, thereby altering pollution loads. Once agricultural pollutants are discharged into water resources it might undergo dilution or some form of chemical, physical and biological alteration depending on the assimilative power of the system to yield a final water quality at a specific location. Thus, in-stream water quality, measurements are dependent not only on the quantity of emissions, but also on the spatial interdependencies between different fields and sub-catchments relative to each other as well as the in-stream processes occurring as the pollutants are transported downstream.

A prerequisite for the meaningful formulation of any policy to efficiently control pollution within an integrated catchment management approach is information regarding economic environmental tradeoffs of alternative policy instruments to control NPS pollution. Previous efforts to quantify these tradeoffs involved the reduction of pollution emissions using aggregated field-scale emissions, thereby ignoring all the other processes that yield the final water quality. The implication of ignoring these processes was that areas with high field-scale emissions might not contribute significantly to water pollution at a specific location. Ignoring the interrelated linkages between agricultural production practices and pollution damage, as it has been done often in the past, might therefore stem relative comparisons between alternative policies inappropriate. Cost-effective NPS pollution abatement policy can only advance if economic environmental tradeoffs of alternative policy instruments to control NPS pollution are quantified, taking the spatial interdependencies between alternative pollution sources into account.

The main objective of this research was to develop a spatial decision-support system capable of quantifying economic environmental tradeoffs of alternative NPS pollution abatement instruments.

Specific objectives include:

• Development of procedures to integrate a catchment level NPS pollution simulation model with an economic optimisation model.
• Development of a spatial economic optimisation model capable of linking the spatial use of alternative management practices to a total catchment outlet water quality standard.
• Evaluation of the cost-effectiveness of alternative NPS pollution abatement instruments through the quantification of economic environmental tradeoff curves.

The research was conducted in the Gamtoos river catchment. The Gamtoos River was formed by the confluence of the Kouga and Groot Rivers. The drainage area of the 70 km long Gamtoos River, which is surrounded by the Baviaanskloof Mountains, constitutes an area of 1 357 km². About 7 400 ha were utilised to produce citrus, potatoes and other vegetable crops by 242 irrigators using micro-, drip- and centre pivot irrigation.

At the beginning of the project the project team envisaged the use of field-scale pollution loads aggregated over representative farms to quantify economic environmental tradeoffs of lowering nitrate NPS pollution. However, the research methodology changed significantly after the principal researcher had visited the Virginia Polytechnic Institute and State University, USA, in the second year of the research. After his visit the emphasis of the research was shifted towards catchment level analyses using a catchment level NPS pollution model to account for the spatial interdependencies that effect pollution. The Steering Committee agreed that the research team should focus on using a catchment approach when quantifying economic environmental tradeoffs of alternative pollution abatement policies.

The Soil and Water Assessment Tool (SWAT)(Neitsch, Arnold, Kiniry and Williams, 2001) was selected to simulate inputs for the spatial optimisation models that were used to quantify the economic and environmental tradeoffs of alternative policy instruments to combat NPS pollution. Since SWAT was fully integrated with GIS technology, the principal researchers had to learn basic GIS processing commands to convert different spatial data sources into the right format and to facilitate the configuration of the model. Spatial variability in the Gamtoos catchment was taken into account by delineating 22 sub-catchments and 129 hydrological response units of which 53 were used for irrigation purposes. The effects of 229 alternative crop, water and fertiliser input combinations on nitrate pollution parameters were thereafter simulated with SWAT.

The General Algebraic Modelling System (GAMS) (Brooke, Kendrick, Meeraus and Raman, 1998) was used to integrate SWAT with the spatial optimisation models developed in this research. GAMS is a very powerful modelling system that enables the modeller to import data files, manipulate large data sets, calculate input parameters, generate the optimisation matrix, optimise the model through an automatic link to several solvers, read the output from the solver and to generate customised output files. The integration was done through the development of procedures to extract the vast amount of output data which was differentiated by sub-catchment, hydrological response unit, land type, crop, planting date, fertiliser application rate, irrigation level, year of simulation and month of the year from the SWAT output files and to convert it to GAMS readable input files. Excel's macro-language was used to extract the data and to generate the GAMS readable input files. Special GAMS programming features were then used to import about 900 000 lines of output data from SWAT into the modelling system.

Once all the necessary SWAT data was imported into the modelling system, the data were combined with economic data to calculate the necessary technical coefficients for all the activities in the different spatial optimisation models. Due to problems simulating the pollution impact of citrus, potatoes and cabbage were chosen to approximate the impact of pollution control policies on abatement cost in the Gamtoos catchment. Baseline pollution levels were established using an optimisation model without any constraints on pollution. Several optimisation models were then used to model the economic environmental tradeoffs of abating baseline pollution levels. Qiu (1996) developed procedures to link the spatial use of alternative management systems to total catchment water quality through the use of pollution contribution factors, making the assumption that the pollution contribution factors did not change with shifts in landuse changes. These research efforts were improved upon in this research through the development of a non-linear spatial programming model with endogenous pollution contribution rates that were dependent on the spatial use of alternative management actions in the whole catchment. The spatial programming model which consisted of 677 constraints, 12 085 variables and 94 871 technical coefficients was used to establish the cost-effective allocation of management practices that achieved the water quality standard at minimum cost. Results from this model were used as benchmark to compare and determine the cost-effectiveness of increasing water and fertiliser cost as policy instruments to comply with a specific water quality standard. The impact of these policy instruments on landuse changes and associated gross margins was optimised using a programming model without any constraints on pollution. The impact of these policy instruments on pollution abatement was determined exogenously using the optimised landuse, given the policy instrument evaluated and the pollution output from SWAT. Similar procedures were used to determine the impact of a water market on pollution.

Results and conclusions

• Results from the baseline analysis indicated that significant variability exists between different sub-catchments with respect to gross margins per unit emitted nitrate. However, it could not be concluded that sub-catchments with relatively high values would have higher pollution abatement costs when the aim was to abate pollution at the catchment outlet, because each of the sub-catchments was contributing differently to the pollution problem at the catchment outlet.
• The ability of the non-linear spatial programming model to determine cost-effective economic environmental tradeoffs was clearly demonstrated. Significant tradeoffs were modelled at the sub-catcment level due to the existence of both positive and negative tradeoffs, while little trade-off was modelled at the catchment level up to a 20% pollution abatement level. Thus, choice of pollution abatement levels based on catchment level tradeoffs may not be socially acceptable since it may make profitable farming for some farmers impossible.
• Pollution abatement cost was not proportionally related to sub-catchment pollution abatement, indicating the importance of the pollution contribution factors in determining the abatement costs of a sub-catchment.
• Controlling the spatial use of alternative management options (crop, planting date, soil and input use) and a pollution tax were proposed as methods to achieve cost-effective pollution abatement. However, these options were very difficult to implement due to the large number of entities that needed to be controlled and the additional cost of obtaining the necessary water quality data at the sub-catchment level.
• The cost-effectiveness of increasing fertiliser and water cost were evaluated relative to achieving the water quality standard at the catchment outlet cost-effectively. Although increasing fertiliser cost had proved to be more cost-effective than increasing water cost, the cost-effectiveness of the former was still very low. The cost-effectiveness of a fertiliser tax was low because a uniform tax was applied to the catchment that was characterised by considerable spatial variability.
• Another important result was that increasing water cost would not necessarily improved water quality, especially if farmers were using water more efficiently without decreasing the areas irrigated. Increasing water use efficiency would reduce streamflow via reduced return flows. Reducing the amount of water available to dilute the pollution emissions would increase the concentration of the pollutants, thereby decreasing the quality of the water.
• A water market would encourage more efficient use of water, but would reduce irrigation water return flows and therefore the dilution capacity of the water system. To result in a water quality improvement, the landuse resulting from more efficient use of water must also result in a more than proportional decrease in pollution emissions.
• Results from the research not only showed that the developed procedures were suitable to quantify economic environmental tradeoffs necessary for NPS pollution abatement policy, but also suggest important policy implications:
• To improve the cost-effectiveness of pollution taxes, taxes should be based on the pollution contribution of each source to the pollution problem at a specific location and not on reduced emissions loads or pollution concentrations at the source.
• The cost-effectiveness of taxes on input use was very low.
• Water conservation policies with the aim of increasing water application efficiencies would increase water quality problems if farmers were allowed to increase areas irrigated using conserved water.

Achieving set objectives and value of results

The main objective of this research was achieved through the development and application of an integrated modelling system consisting of a catchment level NPS pollution model and a spatial optimisation model to evaluate the cost-effectiveness of alternative policy instruments to abate pollution at the Gamtoos catchment outlet.

More specifically the following objectives were achieved.

• Procedures were developed to integrate SWAT with the spatial optimisation model through the use of macros and special GAMS programming features. The procedures developed in this research can be used to successfully integrate data transfer between any simulation model and GAMS.
• Linking the spatial use of alternative management practices to a catchment water quality standard was accomplished using the concept of pollution contribution factors. A major contribution was made by allowing the contribution factors to vary as the landuse in the catchment changed, resulting in more realistic tradeoffs.
• The greatest value of this research lies in the application of the integrated decision-support system to quantify economic environmental tradeoffs of alternative policy instruments to achieve NPS pollution abatement.

Only through the application of the decision-support system developed in this research would policy makers be convinced about the relative effectiveness of alternative policy instruments to control NPS pollution. Application of these models would further enhance the understanding of the interaction between water legislation, water policy administration, technology, hydrology, NPS pollution and human value systems necessary to advance water policy.

Further research proposals

NPS pollution control policy can only advance through a better understanding of the interrelated linkages between pollution-generating activities at field scale and the resulting damages caused at the catchment scale. Application of catchment scale NPS models may assist in understanding these linkages. However, it is important that these models should be able to quantify the impact of alternative management practices implemented at the field scale on water quality indicators measured at the catchment scale. Thus, there is a clear need for more detailed models that may complement modelling procedures used in South Africa to target problem areas. Due to the complexity of determining economic and environmental tradeoffs, application of these models should take place within multidisciplinary research teams.

More detailed NPS modelling to quantify economic and environmental tradeoffs should recognise that although catchments and sub-catchments are logical management units from an environmental viewpoint, it is decisions made at the farm level that determine pollution abatement cost. Further research is therefore necessary to more closely reflect the farm as management unit when quantifying economic environmental tradeoffs of improving water quality, thereby quantifying the impact of pollution abatement policies on irrigation farming profitability.

To advance the application of more detailed economically linked catchment scale NPS models the development of appropriate GIS data bases cannot be overemphasised. Although GIS soils information is available from the Institute of Soil Climate and Water (ISCW) the data contained in the database are not suited for direct hydrological modelling. Furthermore, most troublesome is the absence of detailed landuse information within the boundaries of water user associations (WUA) necessary for the evaluation of alternative management practices on pollution loads. In many instances the WUAs only know how much water has been distributed to specific farmers without knowing consumptive use patterns. Thus, basic spatial information on temporal cropping patterns and water use is not available in most catchments. A GIS database of temporal cropping patterns and irrigation technology is essential for model validation, and if available may prove to be of benefit to WAUs, given they will have to submit water use plans in future proving efficient use of their water supplies.

The mathematical programming model can be developed further to recognise that pollutant loads are inherently stochastic and that the risky environment in which farmers produce and market their crops may have a significant effect on pollution abatement cost. The procedure used to link the spatial use of different management options to the nitrate water quality indicator using a dynamic pollution contribution factor should be enhanced through the incorporation of pollutant decay rates.

In this research alternative policy instruments were evaluated to reduce NPS pollution assuming transactions cost is zero. Transactions cost include, among other things, the costs associated with implementing, administering, and enforcing policies, as well as the costs of obtaining information to design policies. Future research should also include transaction cost when evaluating alternative NPS control policies.