Impact of urban development on streamflow regime of a Portuguese peri-urban Mediterranean catchment

Purpose

Relatively little is known in the Mediterranean environment about changes in streamflow during urban development in partially urbanized peri-urban catchments. This paper explores the modification of streamflow regime as a consequence of the construction of an enterprise park, a major road, and expansion of residential areas, leading to urban areas increase from 32 to 40 % in a small catchment (6.2 km²), located in the periphery of one of the main cities in central mainland Portugal.

Materials and methods

The study was carried out over five hydrological years (October 2008–September 2009 to October 2012–September 2013), including two initial years of pre- and three following years of post-additional urban development. Streamflow was recorded by a V-notch weir at the catchment outlet at 5-min intervals. Rainfall was recorded at a weather station 0.5 km north of the catchment and by five tipping-bucket raingauges installed in January 2011 within the study catchment. Streamflow was converted into runoff and split into baseflow and stormflow components by applying a mathematical low-pass digital filter. Streamflow differences were investigated through changes in (i) annual runoff coefficients, (ii) annual baseflow index, (iii) seasonal baseflow index and stormflow coefficient, and (iv) storm event analysis.

Results and discussion

Annual runoff coefficient ranged from 14 to 21 % and storm runoff coefficient from 9 to 14 %, both between the driest 2011/12 and wettest 2012/13. Although these differences were influenced by inter-annual weather variability, a comparison between years with similar rainfall before and after additional urban development showed a 43 % increase in storm runoff. Impacts on streamflow were also noticed through changes on hydrograph: (i) regression lines of storm runoff against rainstorm parameters exhibited higher vertical positions in 2012/13 than 2008/09, (ii) gradual increase in peak flow but with a clear distance between pre- and post- additional urbanization, (iii) quicker response time from 60–75 min to 40–45 min between both periods, and (iv) decrease in recession time from 21–29 h to 7–9 h for the same periods.

Conclusions

The dispersed urban pattern and permeable soils provide many overland flow sinks, favouring relatively low storm runoff of the catchment. Nevertheless, the enlargement of impervious surfaces (from 12.8 to 17.0 %) and particularly the storm drainage system installed in new urban areas led to great changes on rainfall–runoff event responses. Urban planning should consider the landscape mosaic of peri-urban areas in order to maximize water infiltration and minimize the impacts on streamflow regime.