Regional-Scale Leaching Assessments for Tenerife: Effect of Data Uncertainties Ricardo Diaz-Diaz and Keith Loague* ABSTRACT This paper reports assessments of nonpoint-source (NPS) ground water vulnerability due to pesticide leaching for the Canary Island of Tenerife. Index-based estimates of contamination potential were made for the pesticides Carbofuran and Ethoprophos using the attenu- ation and retardation factors (AF and RF). The major objective this effort was to characterize the data-related uncertainties, using first-order uncertainty analysis, that are associated with the new pesti- cide leaching estimates for Tenerife. The crop, recharge, and soils data for Tenerife were overlaid in this regional-scale study using a geographicinformation system approach. The strategy was to identify the effect of supplementalinformation in reducing uncertainties in the island-wide vulnerability estimates. Data improvements were esti- mated as reductions in the AF and RF input parameter uncertainties. The greatest improvements in the Tenerife vulnerability estimates were found for reductions in the chemical data uncertainties. The newvulnerability assessments for Tenerife will be useful for future agrochemical regulation decisions andfor targeting critical areas that may require remediation. N ONPOINT-SOURCE ground water contamination is widely recognized as a serious environmental problem (Barbash and Resek, 1996; Corwin and Loague, 1996; Corwinet al., 1997, 1999; Loagueet al., 1998; National Research Council, 1993). The NPS ground water contamination problem is especially im- portant for insular systems. Typical of most insular sys- tems, the ground water resources on the Canary Island of Tenerife (Fig. la) are in the form of an easily affected flesh water lens floating on top of a saltwater body. Silviculture in Tenerife is dependent upon pesticides; therefore, precious ground water resources are vulnera- ble to contamination from leaching pesticides. It should be pointed out that there are no regulations on pesticide use in Tenerife (de Lifian, 1996). The economic consequences of NPS ground water contamination, even at concentrations near the detect- able limit, cannot be ignored in Tenerife. For example, Leon-Guerrero et al. (1994) considered wellhead treat- ment costs for ground water contaminated with pesti- cides in Hawaiiby factoring annual treatment costs into annual financial returns from pineapple [Ananas como- sus (L.) Merr.] production. The results from the Leon- Guerrero cost-benefit effort for Hawaii pointed to the prevention of ground water contamination as a more cost-effective measure than wellhead treatment. There- fore, it should be obvious that reliable assessments of NPSground water vulnerability, in Tenerife and else- Ricardo Diaz-Diaz, Dep. of Pedology and Geology, Univ. of La La- guna, 38204 La Laguna, Tenerife, Spain. Keith Loague, Dep. of Geo- logical and Environmental Sciences, Stanford Univ., Stanford, CA 94305. Received 8 July 1999. *Corresponding author (keith@pangea. stanford.edu). Published in J. Environ. Qual. 29:835-847 (2000). where, are extremely important in the decision manage- ment arena. Diaz-Diaz et al. (1998) provided the first regional- scale assessments of NPS ground water vulnerability for Tenerife using the attenuation and retardation factors. In characterizing the vulnerability of Tenerife ground water to contamination from four different pesticides, in areas of banana (Musa paradisiaca L.) silviculture, Diaz-Diaz et al. (1998) discussed irrigation effects, pesti- cide half-life effects, soil depth effects, and presented a best-estimate scenario of vulnerability. The preliminary assessments of Diaz-Diaz et al. (1998) did not, however, include any consideration for the uncertainty in the re- gional-scale assessments due to uncertainties in the soil, recharge, or chemical data. In this paper we combine soil, climatic, and chemical data in a geographic informa- tion system frameworkwith two simple pesticide leach- ing indices to generate ground water vulnerability as- sessments for all the agricultural areas across Tenerife for two of the most important agrochemicals currently in use. The major focus of this effort was to characterize the uncertainty in the Tenerife ground water vulnerabil- ity assessments resulting from data uncertainties. The work in this study builds upon the efforts of Diaz- Diaz et al. (1998) for Tenerife and Loague (1994) Hawaii. The major difference in the work reported here and the Diaz-Diaz et al. (1998) effort is that the effect of data uncertainties in the Tenerife ground water vul- nerability assessments were not considered in the earlier study. As well, the Diaz-Diaz et al. (1998) study focused only on areas of banana silviculture, where in this effort the focus is on all agricultural areas. The important differences in the work reported here and the Loague (1994) effort are: (i) the assessments of ground water vulnerability in Hawaii were made only for the Pearl Harbor basin (i.e., not all the agricultural areas on Oahu were included); (ii) the recharge estimates for the Pearl Harbor basin were average values (based upon soil or- ders) and did not faithfully capture the spatial variations associated with the regional-scale patterns of precipita- tion, irrigation, and evapotranspiration; and (iii) only the influence of supplemental soil property information was considered (i.e., the influence of soil pH on chemical half-life values was not considered for the Hawaii study). METHODS Mobility Indices The AF represents an index of the pesticide massemission from the vadose zoneas the fraction of total pesticide applied. The AF is defined as: Abbreviations: AF, attenuation factor; NPS, nonpoint-source; RF, retardation factor. 835 Published May, 2000