The influence of Low Impact Development on rainfall-runoff relationships at catchment scale

Abstract

With rapid urbanization, Low Impact Development (LID) is promoted as an alternative to Conventional Drainage (CD), seeking a natural solution for current urban water problems. The positive effects of LID were the main theme of recent LID researches, but this project aims to deeply explore the hidden troubles about the extreme peak runoffs influenced by LID city on catchment scale.
In this research, the SUPERFLEX conceptual model was adapted to a rural-urban semi-distributed model to simulate the current rainfall-runoff relationship of the catchment where San Antonio city is located in. Besides, the model expressions of 4 representative LID practices (bioretention cells, vegetated swales, green roofs, and permeable pavements) were devised under SUPERFLEX framework. To deal with the prediction uncertainty, three urban development scenarios in 2040 and five LID implementation scenarios were designed for San Antonio city. And their influences on the basin peak runoffs will be quantitively studied.
Research result shows that, the urban runoff tended to swing between extreme flood and extreme drought in the reference situation; And next, the infill urban development strategy, which means developing the vacant or undeveloped land within an existing community, was more helpful on peak runoff and total runoff volume control than sprawl urban development strategy with the same population growth; Thirdly, the bioretention cells, vegetated swales, and permeable pavements had similar good performance on peak runoff reduction, which can be mainly ascribe to the stormwater infiltration process. As for the retention of total runoff volume, the bioretention cells, permeable pavements, and green roofs perform better than vegetated swales since the rapid water transportation character of vegetated swales decrease the water residence time for infiltration;
The runoff reduction function of LID practices performs effective on the large peaks in dry and normal seasons, but it will be restrained significantly in flood season. According to model result, the rural peak runoffs happened 6.5 to 15.5 hours after the urban peaks. And for 4 LID implementation scenarios in which 15% of urban grey areas is covered by LID practices, the urban peaks are delayed between 0.5 and 2.5 hours. And for the scenario with the LID cover areas as 50% of the urban grey areas, the time lag of urban peaks varies from 0.5 to 6.5 hours. For this scenario, since the obvious time delay of urban peaks, more stack of urban and rural peaks is caused by the time approaching of urban and rural peaks, which causes the increases of two total basin peaks in flood season from 3.57 to 3.65 mm/d and from 6.35 to 6.47 mm/d respectively.
In conclusion we may say that the stacking effect of LID implementation on total basin runoff is limited in the case of San Antonio basin, partly due to the fact that only a small part of this basin is urbanized.