Road runoff is a linear diffuse pollution source, that can cause significant impacts in the receiving environment, usually to its cumulative nature. This environmental problematic received an increasing concern by the scientific-technical community, proved by the increasing number of research studies.

In a case study in Portugal, carried out in 2003/2004, on a section of A25 highway near the sea, specificities not yet mentioned in the international literature were detected. Among other features, it was verified a high salinity associated with salt deposition of marine aerosols.

The present thesis aimed to evaluate the effects of salt deposition in the quality characteristics of road runoff in coastal areas, basing the research on a case study in the mentioned highway section.

Between 2008 and 2010, a systematic monitoring of road runoff, including the development of an integrated methodology for evaluating the salt deposition rate and its effect on resulting road runoff was performed. Through the three monitoring campaigns, 215 samples were collected (and analyzed for various water quality parameters) on 30 rainfall events, under different weather, salt deposition and traffic conditions.

The monitoring results where subjected to a statistical treatment and a exploratory statistical analysis, including the variables related to weather conditions, road section and salt deposition.

It was found that the Site Mean Concentrations (SMC) of the three campaigns had a high coefficient of variation. This result demonstrates the need for a extensive road monitoring for a better knowledge of the SMC of road runoff from a road site.

During several rainfall events the TSS, COD, BOD5 and Fe concentrations exceeded the emission limit values for wastewater discharge established in Decree Law 236/98 (related to point sources, but applied as a reference), witch confirms the environmental importance of this type of pollution.

The salt deposition rate results showed their decrease with the sea distance, as well a characteristic variation, with peaks in January and May, while significantly lower values are characteristic of the summer months.

It was confirmed the high salinity in the runoff, expressed in conductivity in a large part of the research methodology, with average values of conductivity of 325μS/cm and maximum values of 1288μS/cm. The computed Pearson correlation coefficients (r) between conductivity and chloride (r=0,80) where significant. Organic matter was also highlighted in correlation with the conductivity, with r=0,83 for TOC, r=0,66 for COD and r=0,71 for BOD5. For the other parameters, the best correlation was Cu, although less significant (r=0,54). Other exploratory data analysis techniques applied also showed these and other correlations.

The concentrations of the relevant pollutants in this investigation (related to salinity and organic matter), are correlated with variables characterizing the meteorological conditions, emphasizing the antecedent dry period to rainfall events and the average wind speed during this period (e.g, respectively, r= 0,56 and r =-0,49 with the conductivity).

Based on these results, the modeling of the main pollutants concentrations, using multivariate linear regression and considering three groups of independent variables (parameters related to weather conditions, road section and salt deposition) was analyzed. Some pollutants were also included as independent variables, what able to increase the determination coefficients of the models.

This methodology allowed to propose predictive equations for the concentrations of major pollutants (i.e. TSS, chlorides, COD, TOC, Fe, Zn e Cu), based on simple and easy measurement parameters, in real time, which is a relevant and innovative contribution for monitoring studies and the design and control of pollution retention systems.

To proposed model was tested and a internal validation was performed, including the technique of cross-validation, having obtained good results.

In future, it would be interesting to validate the developed model in different case studies by conducting new campaigns, including the installation of devices that, in real time, measures the independent variables. After verifying the need of specific calibrations to each site, this model, expeditiously and economically, can in real time evaluate the concentration and mass loads of relevant pollutants in road runoff, being a breakthrough for the knowledge and operation of these systems, providing a major contribution to the environment protection from this source of pollution.