Using the power of information of sparse data for soil improvement management

Abstract

Listen

Translate article

2 Introduction 2 Data 4 Methods 7 Analysis of spatial dependency 7 Ordinary Kriging 8 Conditional Simulation 10 Uncertainty Assessment 11 Results 15 Spatial Estimation 15 Uncertainty Assessment 17 Consequences 19 Conclusion 20 Uncertainty Estimation 20 Decision Relevance 21 Acknowledgements 22 References 23 1 Working Group: Environmental Decision-Making, Environmental Evaluation and Modeling Ute Schnabel, Olaf Tietje, & Roland W. Scholz 2 July 2002 Abstract For a sustainable management of natural resources, such as soil, the spatial distribution of the environmental impacts is a basic need for decision-making. However, for spatial interpolation in most cases only few data with a skewed distribution and uncertain information about soil contamination are available whereas decisions with high ”correctness” are required. In order to assess the power of information of sparse data a site of 15 square km with 76 soil samples was investigated. The soil was cadmium contaminated predominantly due to airborne emissions from a metal smelter. A lognormal probability distribution was found to appropriately estimate the probabilistic distribution of the contaminant. The spatial interpolation compares lognormal anisotropic kriging and conditional simulation. The resulting overall uncertainty from data sampling, sampling preparation, analytical measurement, interpolation and numerical representation has been investigated. The uncertainty of spatial interpolation was analyzed as the major component due to coarse data sampling and spatial heterogeneity. It is shown that the uncertainty can be efficiently estimated by calculating the percentiles of the lognormal probabilistic distribution function. This procedure also allows the calculation of the local probability of exceeding legal threshold values. Although the estimated uncertainty of the local prediction of the cadmium concentration is rather high, the procedure yields what is often required by decision making: a qualified rough estimate of the contamination. Conclusively, predicting the probability of exceeding a site-specific threshold can be used to roughly delineate prior areas for soil improvement, remediation, or restricted area use, based on the decision makers probability requirement.