This work present a predictive map of arsénico (As) in Santa Rosa basin, Catamarca Province, Argentina, in order to provide a tool for urban planning towards the reduction of disease risk in local and neighboring population.
Arsenic presence in the whole Chacopampean region and its effects on human health has been documented extensively. Because the illness prevalence presents particular characteristics in Argentina, the local medical community calls it HACRE (hidroarsenisismo cronico regional endemico) and makes necessary epidemiologic studies of the whole region, to prepare the settlers of the risks of its consumption. Several Catamarca regions posses natural arsenic values that overcome the limits established by the Argentina Food Safety Code (CAA) of 0,05 mg/l and for the World Health Organization (WHO) of 0,01 mg/l. The arsenic presence behavior is erratic, in the vertical and horizontal planes, imparing to directly observe distribution patterns. Hence the need for spatial interpolation techniques that allow represent the global and local trend of the element.
Groundwater sampling was carried out in 2014, with private and municipal sites. The dataset arises from 72 wells, to a máximum depht of about 250 m, distributed over the whole alluvial plain, that include the study area. Santa Rosa Basin is part of the Rio Sali-Dulce Basin. Belongs to Sierras Pampeanas geological Province. The outcropping in the study region correspond to metamorphic rocks and Precambrian - Paleozoic igneous intrusions and Tertiary and Quaternary sedimentary cover.
Heavy metals were analized by atomic absorption, using AA300 HGA 800 Perkin Elmer equipment. Datasets were statistically analyzed obtaining a minimum of 0,003 mg/l and a maximum of 0,4 mg/l of As. For interpolation, the geostatistical Kriging method applied to each site unsampled “s”, a predicted value Vk (s) which linearly combines the observed or available values {v (s1) ... v (sn)}, considering the random nature of geo-referenced data V (s) plus two characteristic aspects of spatial processes, these are, the tendency μ (s) and parameterized spatial autocorrelation by variogram function g (h), expressed through a model structural additive or V (s) = μ (s) + δ (s).
The semivariogram function Var 1/2 [δ (s + h) -δ (s)] = g (h), measures the variability increases of the variable (As), relative to the separation between sites. The advantage of Kriging, unlike deterministic interpolation criteria, is that considers the distances and directions in which it governs the influence, producing more exact representations of reality and contributes a minimal estimation error; thus, being able to realize a better planning of the use of safe groundwater resources bearing in mind the spatial distribution of naturally occurring pollutants.