Pergamon Phys. Chem. Earth (C), Vol. 26, No. 10-12, pp. 717-722,200l 0 2001 Elsevier Science Ltd All rights reserved 1464-1917/01/%-see front matter PII: Sl464-1909(00)00102-2 Field Analysis of the Water Film Dynamics on a Road Pavement I. Becchi, E. Caporali, F. Castelli and C. Lorenzini Dipartimento di Ingegneria Civile, Universita degli Studi di Firenze - Via S. Marta, 3, 50139 Firenze, Italy Received 7 August 2000; accepted 27 October 2000 Abstract the amount of tyre in contact with the road. The analysis of water film dynamics on road pavement during high intensity rainfall events is fundamental in the study of problems of different nature, from road safety to urban drainage. The influence of precipitation time- variability and road pavement geometry on the water film dynamics is investigated using an experimental monitoring station installed on a mountain road which is prone to heavy storms. The monitoring station is composed by a road portion, a rain gauge and a video-camera automatically activated during intense rainfall. The investigation is based on the digital analysis of the recorded images between summer and autumn 1999. The water film surface irregularities, due to the raindrop impact with the surface and to the presence of roll waves, are studied. The first results are relative to the analysis of the roll-waves dynamics, as they are prevalent in terms of visible effects during the periods of intense rainfall. The digital analysis of the image sequences allows estimating direction and speed of the wave propagation with variable precision due to different light conditions. Using empirical relationships derived from laboratory experiments (Caporali et al., 2OOOb), and a high-resolution Digital Terrain Model (DTM) of the road surface, depth and discharge of the water film in the maximum local slope direction are investigated. In the worst case, with a relatively smooth road and limited tyre tread, the tyre may completely loose contact with the road, a condition known as aquaplaning (TRL, 1997). Many measuring methods have been proposed for water film depth on a real road surface but they show big limitations because of their interference with the observed phenomenon (Laganier, 1977). In this work a not-intrusive field method is proposed to estimate water film depth starting from sensing (remotely) the kinematic parameters of its irregularities. Particular attention has been posed on the roll waves phenomenon (Kenyon, 1998) for which a mono-dimensional scheme has been adopted. A detailed morphological analysis supports the procedure (Caporali et al., 2OOOa). Laboratory experiments have been designed ad hoc to test the applied methodology (Lorenzini and Becchi, 2000). 2 The monitoring station The monitoring station has been installed on a mountain road of the Versilia region on the Apuanian Alps (north of Tuscany, Italy) and it has been working since April 1999. 0 2001 Elsevier Science Ltd. All rights reserved 2.1 Measurement apparatus 1 Introduction Water film depth measurement on road pavements during intense rainfall events could be crucial in evaluating road safety. The presence of water has a lubricating effect on a wet road: friction, known in this context as skidding resistance, is markedly reduced compared with the dry value. If the water level increases beyond the point where the tyre tread and surface texture can disperse it, the thickness of the water film in front of the tyre will build up and begin to penetrate the contact patch, reducing further Correspondence to Enrica Caporali The station consists of a road portion and an experimental monitoring system mounted on the top of a 5 m pole on the edge of the road, at a distance of approximately 7 m from the centre of the target area. The monitoring system is composed of a tipping-bucket rain gauge with 0.2 mm resolution recording on EEPROM and a 8-mm video camera PAL system (576 lines resolution, 25 frames/s) automatically activated by means of an experimental rainfall-weighting scale system. The video camera is activated when rainfall intensity exceeds approximately 20 mm/hour (* 10 mm/hour) for at least 1 minute. The camera viewing cone is pointed on a 1x1 m square painted on the road pavement, which is used for geometrical referencing of video clips, and it covers a total 717