Environinformatics In Ecological Risk Assessment Of Agroecosystems Pollutant Leaching

Stoch Environ Res Risk Assess (2005) 19: 292–300 DOI 10.1007/s00477-005-0233-9

O R I GI N A L P A P E R

Nazzareno Diodato Æ Michele Ceccarelli

Environinformatics in ecological risk assessment of agroecosystems pollutant leaching

Published online: 5 April 2005
Springer-Verlag 2005

Abstract In this paper, a novel approach for mapping leaching risk at large sub-regional scale under limited information is presented, with acronym environinformatics in ecological risk assessment. The problem consists into quantifying the exchange frequency of the plant available soil water (Narula et al. in J Geogr Inform Decision Anal 7(1)32–46, 2002), this frequency can be adopted as a measure indicating the nutrient and contaminant leaching risk for a site. Our approach is based on integrating soil water balance with spatial analysis tools. However, any decision involved in scientific risk evaluation requires the accurate quantification of the degree of uncertainty arising from sampling, modelling and interpolation errors. The non-parametric geostatistical procedure of Indicator Kriging enables to circumvent this problem by estimating the probability that the true value exceed a set of threshold values. The transformation of leaching data to a binary response variable, known as ‘‘indicator’’, can lead to a soft description of leaching. Such soft description can mitigate the uncertainty in exchange frequency estimates of the plant available soil water. The approach was applied to a test site in Beneventan agroecosystem (South Italy) by using a long-term hydrological water balance acquired in a 40-years period. In this way, about 400 km2 (25%) of the total 2,000 km2 of the Benevento province were classified as areas sensitive to nutrient and contaminant leaching.

N. Diodato (&) Monte Pino Research Observatory on Climate and Landscape (Italy), GTOS\TEMS Network–Terrestrial Ecosystem Monitoring Sites (FAO–United Nations), 82100 Benevento, Italy E-mail: [email protected] URL: http://www.fao.org/gtos/tems M. Ceccarelli Research Center on Software Technologies-RCOST, University of Sannio, via Port’Arsa, 11, 82100 Benevento, Italy

Keywords Soil-water-balance Æ Pollutant leaching Æ Geostatistics Æ Probability maps Æ Risk assessment Æ Southern Italy

Introduction In some environmental conditions which occur in wet periods of the year, water surplus is the most important hydrological reservoir, especially in agricultural lands. However, in this context, water can be considered both a resource and a disturbance factor for agroecosystems, because some nutrient pollution that enter the water in a diffuse manner are carried mostly from rainfall events. Then, contamination of watercourses can seriously disturb aquatic eco-systems and may then contaminate drinking water supplies. As referred by Wit et al. (2002), the principal cause of agricultural nutrient pollution in Europe is the input of nutrients to agricultural land (fertilisers and manure) exceeding the output of nutrients from agricultural land (crop yield). Water resource quality is a useful descriptor of agricultural sustainability during the planning stage (Smith and McDonald 1998), and a defensible indicator of sustainable land management (Ochola and Kerkides 2003). Greater emphasis is being placed on the use of process-based mathematical models for the prediction of contaminant transport and to evaluate the risk of leaching and groundwater quality (Klimas 1996; Wingle et al. 1999; Alapati and Kabala 2000; Cepuder and Schmitten 2000; Violette et al. 2000; Dubus and Janssen 2003; Matjı´ ek et al. 2003; Bellin et al. 2004; Kunkel et al. 2004). A comprehensive description of approaches to flow and contaminant transport modeling used for the environmental risk assessment can be found in Chile`s and de Marsily (1993), and Dubus et al. (2002). Depending on the amount of information available to the modeller, primary data might include site characteristics (e.g. longitude, latitude, elevation), soil