Spatial average aquifer recharge through atmospheric chloride mass balance and its uncertainty in continental Spain


  • Francisco J. Alcalá,

    Corresponding author
    1. Geo-Systems Centre (CVRM-IST), Technical University of Lisbon, Lisbon, Portugal
    2. Estación Experimental de Zonas Áridas, Consejo Superior de Investigaciones Científicas, Almería, Spain
    • Correspondence to: Francisco J. Alcalá, Technical University of Lisbon, Geo-Systems Centre (CVRM-IST), Av. Rovisco Pais, 1, 1049–001 Lisbon, Portugal.


    Search for more papers by this author
  • Emilio Custodio

    1. Department of Geo-Engineering/International Centre for Groundwater Hydrology (FCIHS), Technical University of Catalonia (UPC), Barcelona, Spain
    2. Royal Academy of Sciences of Spain (RAC), Madrid, Spain
    Search for more papers by this author


The atmospheric chloride mass balance (CMB) method allows spatial evaluations of the average diffuse aquifer recharge by rainfall (inline image) in large and varied territories when long-term steady conditions can be assumed. Often, the distributed average CMB variables necessary to calculate inline image have to be estimated from the available variable-length data series, which may be of suboptimal quality and spatial coverage. This paper explains the use of these data and the reliability of the results in continental Spain, chosen as a large and varied territory. The CMB variables have been regionalized by ordinary kriging at the same 4976 nodes of a 10 km × 10 km grid. Nodal inline image values vary from 14 to 810 mm year–1, 90% ranging from 30 to 300 mm year–1. The recharge-to-precipitation ratios vary from 0.03 in low-permeability formations and semiarid areas to 0.65 in some carbonate massifs. Integrated average results for the whole of continental Spain yield a potential aquifer recharge of 64 km3 year−1, the net recharge over permeable formations (40% of the territory) being 32 km3 year−1. Two main sources of uncertainty affecting inline image (given by the coefficient of variation, CV), induced by the inherent natural variability of the variables (CVR) and from mapping (inline image), have been segregated. The average CVR is 0.13 and could be improved with longer data series. The average inline image is 0.07 and may be decreased with better data coverage. The inline image estimates were compared with other regional and local recharge estimates, inline image being 4% and 1% higher, respectively. Copyright © 2012 John Wiley & Sons, Ltd.