Abstract
In the water management literature both the normal and log-normal distribution are commonly used to model stochastic water pollution. The normality assumption is usually motivated by the central limit theorem, while the log-normality assumption is often motivated by the need to avoid the possibility of negative pollution loads. We utilize the truncated normal distribution as an alternative to these distributions. Using probabilistic constraints in a cost-minimization model for the Baltic Sea, we show that the distribution assumption bias is between 1% and 60%. Simulations show that a greater difference is to be expected for data with a higher degree of truncation. Using the normal distribution instead of the truncated normal distribution leads to an underestimation of the true cost. On the contrary, the difference in cost when using the normal versus the log-normal can be positive as well as negative.
Similar content being viewed by others
Notes
Kampas and Adamidis [18] also develop a deterministic equivalent for normal–log-normal distribution.
With shape we imply varying values of skewness and kurtosis. Skewness is a measure of symmetry, i.e. how a variable is distributed to the left and right of the center point. Kurtosis measures the peakedness of a distribution and it reflects the probability of extreme events. Data sets with high kurtosis have heavy tails (the uniform distribution is an extreme example).
Xu et al. [28] investigate the normal and log-normal distribution assumptions for sediment yields that affect returns from farming. Gren et al. [13] adopted a similar approach but in the context of water pollution. Kampas and White [17] compare the normal and log-normal distributional assumptions to a non-parametric assumption.
For the percent point function, we start with the probability and compute the corresponding Q for the cumulative distribution function. Hence, note that Q is defined by the left-hand-side of Eq. 7.
Hence, the mean load of the log-normally distributed pollution will be characterized by Q*. The generality of Eqs. 10–13 makes it equally applicable for the normal as well as the log-normal distribution. Hence, \( \hat Q \) from Eq. 10 equals \( Q* - {K_\alpha }{\left[ {{\text{Var}}(Q)} \right]^{0.5}} \)from Eq. 7, assuming normal distribution.
The choice between log-normal and truncated normal depends on the shape of the stochastic load data.
Either one of Eqs. 13 and 15 can be used to obtain the percent point function. Giving a numerical example; suppose that v = 0, E(Q) = 0, Var(Q) = 1, and α = 0.5, than \( {K_\alpha } = {\Phi^{ - 1}}\left( {0.75} \right) = 0.67 \) applying Eq. 15. Applying Eq. 13 we have \( {K_\alpha } = F_{\text{TN}}^{ - 1}\left( {0.5} \right) = 0.67 \).
The marginal cost of wetland restoration is assumed to be 5,050 SEK/ha for all countries besides Poland, Russia, Latvia, Lithuania, and Estonia, where a 60% lower cost is assumed, as suggested by Gren et al. [11]. Moreover, it has been assumed that 5% of arable land can be converted to wetland. Finally, each hectare of wetland is assumed to reduce the amount of nitrogen by 0.150 tonnes.
In principle, it would be interesting to compare the implications of different distributional assumptions for nitrogen to those for phosphorus, as both nutrients contribute to eutrophication of the Baltic Sea. Therefore, we carried out similar calculations for phosphorus, but could establish that the probability of erroneously predicting negative phosphorus loads using the normal distribution was also zero. Therefore, for phosphorus, the difference between a normal and truncated normal distribution would also be zero.
The average stochastic pollution can be reduced to zero at most. For sizable standard deviations compared to Q*, this could restrict the possibility to state that \( P\left( {Q \leqslant {Q^*}} \right) \geqslant \alpha \) with a high level of reliability.
References
Andersson, C., & Destouni, G. (2001). Groundwater transport in environmental risk and cost analysis: Role of random spatial variability and sorption kinetics. Ground Water, 39, 35–48.
Blackwood, L. G. (1992). The lognormal distribution, environmental data, and radiological monitoring. Environmental Monitoring and Assessment, 21, 193–210.
Bouraoui, F., Grizzetti, B., Granlund, K., Rekolainen, S., & Bidoglio, G. (2004). Impact of climate change on the water cycled and nutrient losses in a Finnish catchment. Climate Change, 66, 109–126.
Brooke, A., Kendrick, D., & Meeraus, A. (1988). GAMS—a user’s guide. San Francisco: Scientific.
Byström, O. (1999). Wetlands as a nitrogen sink—estimation of costs in Laholm bay. In M. Boman, R. Brännlund & B. Kriström (Eds.), Topics in environmental economics. Dordrecht: Klüwer Academic.
Charnes, A., & Cooper, W. W. (1963). Deterministic equivalents for optimizing and satisfying under chance constraints. Operational Research, 11(1), 18–39.
Cooper, W. W., Hemphill, H., Huang, Z., Li, S., Lelas, V., & Sullivan, D. W. (1997). Survey of mathematical programming models in air pollution management. European Journal of Operational Research, 96(1), 1–35.
Crow, E. L., & Shimizu, K. (eds). (1988). Lognormal distributions: Theory and applications. New York: Marcel Dekker.
Elofsson, K. (2000). Cost efficient reductions of stochastic nutrient loads to the Baltic Sea. Working Paper Series 6. Uppsala: Swedish University of Agricultural Sciences, Department of Economics.
Elofsson, K. (2003). Cost-effective reductions of stochastic agricultural loads to the Baltic Sea. Ecological Economics, 47, 13–31.
Gren, I-M., Jannke, P. & Elofsson, K. (1995), Cost of Nutrient Reductions to the Baltic Sea, Technical Report. Beijer Discussion Papers Series No. 70. Stockholm: Beijer International Institute of Ecological Economics, Royal Swedish Academy of Sciences.
Gren, I.-M., Destouni, G., & Scharin, H. (2000). Cost effective management of stochastic coastal water pollution. Environmental Modeling and Assessment, 5(4), 193–203.
Gren, I.-M., Destouni, G., & Tempone, R. (2002). Cost effective policies for alternative distributions of stochastic water pollution. Journal of Environmental Management, 66, 145–157.
Guo, P., Huang, G. H., He, L., & Sun, B. W. (2008). ITSSIP: Interval-parameter two-stage stochastic semi-infinite programming for environmental management under uncertainty. Environmental Modelling and Software, 23, 1422–1437.
Hald, A. (1952). Statistical theory with engineering applications. New York: Wiley.
Huang, G. H., Sae-Lim, N., Liu, L., & Chen, Z. (2001). An interval-parameter fuzzy-stochastic programming approach for municipal solid waste management and planning. Environmental Modeling and Assessment, 6, 271–283.
Kampas, A., & White, B. (2003). Probabilistic programming for nitrate pollution control: Comparing different probabilistic constraint approximations. European Journal of Operational Research, 147(1), 217–228.
Kampas, A., & Adamidis, K. (2005). Discussion of the paper “Cost effective policies for alternative distributions of stochastic water pollution” by Gren, Destouni and Tempone. Journal of Environmental Management, 66, 145–157.
Li, Y. P., Huang, G. H., & Nie, S. L. (2006). An interval-parameter multi-stage stochastic programming model for water resources management under uncertainty. Advances in Water Resources, 29(5), 776–789.
Maqsood, I., Huang, G. H., & Yeomans, J. S. (2005). An interval-parameter fuzzy two-stage stochastic program for water resources management under uncertainty. European Journal of Operational Research, 167(1), 208–225.
Milon, J. W. (1987). Optimizing nonpoint source controls in water quality regulation. Water Resources Bulletin, 23(3), 387–396.
Ott, W. R. (1990). A physical explanation of the log-normality of pollutant concentrations. Journal of the Air and Waste Management Association, 40(10), 1378–1383.
Paris, Q., & Easter, C. D. (1985). A programming model with stochastic technology and prices: The case of Australian agriculture. American Journal of Agricultural Economics, 67(1), 120–129.
Reimann, C., & Filzmoser, P. (2000). Normal and lognormal data distribution in geochemistry: Death of a myth. Consequences for the statistical treatment of geochemical and environmental data. Environmental Geology, 39, 1001–1014.
Sengupta, J. K., & Gruver, G. (1971). Chance-constrained linear programming under truncation and varying sample sizes. The Swedish Journal of Economics, 73(2), 184–203.
Van Buren, M. A., Watt, W. E., & Marsalek, J. (1997). Applications of the lognormal and normal distributions to stormwater quality parameters. Water Research, 3(1), 95–104.
Willett, K., Zhang, T., McTernan, W. F., Sharda, R., & Rossman, E. J. (1997). Regulation of pesticide discharge into surface and groundwater under uncertainty: A model for risk-profitability trade-offs and policy selection. Environmental Modeling and Assessment, 2, 211–220.
Xu, F., Prato, T., & Zhu, M. (1996). Effects of distribution assumptions for sediment yields on farm returns in a chance-constrained programming model. Review of Agricultural Economics, 18, 53–64.
Zhu, M., Taylor, D., Sarin, S., & Kramer, R. (1994). Chance constrained programming models for risk-based economic and policy analysis of soil conservation. Agricultural and Resource Economics Review, 23, 58–65.
Acknowledgments
The authors want to thank Clas Eriksson, Karin Larsen, Monica Campos, Yves Surry, and Erik Ansink for valuable comments. The usual disclaimer applies. Funding from Baltic Nest Institute, Aarhus University, Denmark, and the BONUS program is gratefully acknowledged.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Rights and permissions
About this article
Cite this article
Kataria, M., Elofsson, K. & Hasler, B. Distributional Assumptions in Chance-Constrained Programming Models of Stochastic Water Pollution. Environ Model Assess 15, 273–281 (2010). https://doi.org/10.1007/s10666-009-9205-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10666-009-9205-7