Mapping impervious surface change from remote sensing for hydrological modeling

J Dams, J Dujardin, Reggers R, I Bashir, F Canters, Okke Batelaan

    Research output: Contribution to journalArticlepeer-review

    85 Citations (Scopus)


    This paper describes a method to estimate changes in impervious surface fraction from widely available medium resolution remote sensing data. The potential of the impervious surface fraction data for hydrological modeling is illustrated by a case study for the Kleine Nete catchment, Belgium. Impervious surface fraction data for the Kleine Nete were simulated based on a Landsat TM image acquired in 1986 and a Landsat ETM+ of 2003. The paper assesses the impact of change in impervious surface cover between 1986 and 2003 on the vertical water balance and groundwater system. The results show that impervious surface fractions can be obtained from medium resolution remote sensing data with reasonable accuracy, the mean average error is around 18% and the bias close to zero for both the recent and historical image. Comparing the impervious surface fractions of 1986 and 2003 indicates that most urbanization occurred by densification of existing urban areas. The average impervious fraction of the urban cells increases from 25.4% to 29.2% between 1986 and 2003. Urban runoff increases by 9.5% between 1986 and 2003 due to the increase in impervious surfaces, while the evapotranspiration and recharge in urban areas decrease by 1.5% and 1.6% respectively. For the study area a land-use class based parameterization method underestimates groundwater recharge and head decreases in the most dense urban area and overestimates those in some of the smaller urban centers.

    Original languageEnglish
    Pages (from-to)84-95
    Number of pages12
    JournalJournal of Hydrology
    Publication statusPublished - 2 Apr 2013


    • Impervious surface
    • Remote sensing
    • Urbanization
    • Water balance


    Dive into the research topics of 'Mapping impervious surface change from remote sensing for hydrological modeling'. Together they form a unique fingerprint.

    Cite this