Main content area

Estimation of Urban Impervious Fraction from Satellite Images and Its Impact on Peak Discharge Entering a Storm Sewer System

Ravagnani, Fabrizio, Pellegrinelli, Alberto, Franchini, Marco
Water resources management 2009 v.23 no.10 pp. 1893-1915
basins, image analysis, land cover, rain, reading, satellites, sewer systems, storms, surveys, towns, urban runoff, watersheds, Italy
The runoff coefficient in an urban basin is highly influenced by the impervious and pervious surface fractions since they affect the entity of rainwater entering a storm sewer system. Today these fractions can be estimated by reading high resolution satellite images that are readily available at a relatively low cost. However, this approach involves a certain margin of error when it comes to identifying the various types of cover and hence the total extent of impervious and pervious surfaces. The first problem addressed in this paper thus lies in assessing to what degree the error in the estimation of the two fractions--as derived from a reading of satellite images of the area taken into consideration--may impact the estimation of peak discharge which will be used in turn as a basis for designing or verifying a storm sewer system. A further aspect affecting the entity of rainwater that flows into a storm sewer system is the manner in which the impervious and pervious fractions are connected to the system itself. This type of information may not be deduced from satellite images, but only from an extensive field survey. However, such surveys are feasible only for areas of limited size, whereas they become prohibitive in terms of time and cost in the case of large catchments. An investigation was thus made into whether disregarding the type of connection would significantly affect the peak discharge taken as reference for the design or verification of a storm sewer system. The two above-mentioned problems were addressed in reference to a real case, represented by the town of Codigoro (Ferrara, Italy), in which five small basins with different types of land cover were selected. The results of our analysis show that an estimation of impervious and pervious fractions from high resolution satellite images is sufficiently precise and acceptable for estimating the peak discharge entering the sewer system. By contrast, disregarding the information on how the different areas are connected to the sewer system may lead to a marked overestimation of discharges.