The coastal zone and the Mount Lebanon chain are characterized by their closeness, except in the north and the south of the country. The study area (Figure 1) includes the Lebanese coastline which extends about 230 kilometers in length from the northeast border at Aarida to the southwest border at Naqoura [11] . Based on a study done by the CDR in 2005, the coastal zones represent 8% of the total Lebanese surface area which is approximately 840 km² of the Lebanese territories. The study area includes 15 rivers [11] . These rivers transport the manmade wastes to the sea. The annual mean temperature in the coastal zone varies between 13.5˚C and 27˚C with an average annual rainfall of 600 mm [12] [13] . The coastal zone has a very high population density; it was estimated that more than 1.5 million inhabitants in the whole coastal area, which repre- sents 55 percent of the total population [11] . The major six cities that are highly popu-

lated and located directly on the coast line are from North to South: Tripoli, Jbeil, Jounieh, Beirut, Saida, and Tyre (Figure 1) [13] .

Landsat satellite images are characterized by medium spatial resolution with good capability to discriminate between different seawater aspects: normal versus polluted sea water [13] [14] . Two consecutive years were studied in 2014 and 2015 using Land sat OLI 8 satellite images. This approach aims to identify the pollution sources and types. The analyzed images are taken monthly for each year. This study focuses on monitoring change in pollution sources in space and time as well as the evaluation of the persistence of the identified sources. The goal of making monthly maps is to: compare the pollution changes and propagation monthly and yearly in order to monitor the changes in the types of pollutants; if the pollutants dumped into the sea have increased or decreased; and to detect new polluted sites, permanent polluted sites and to identify the critical, at risk areas where pollution needs to be stopped. Data can be downloaded (at no charge) from Glo Vis, Earth Explorer, or via the Land sat Look Viewer within 24 hours after the acquisition. In this study, Earth Explorer was used for download. All data images selected for analysis require several processing steps. The goal of these processing steps is to increase the accuracy, the clarity and interpretability of the digital data during image analysis. In this study, images were processed using ENVI5.2 and ArcGIS 10.3.1 software. After image processing, identification of marine pollutants was performed usual visual interpretation: color, tone, texture, flow pattern, and temperature differences [13] [14] .

Based on the dimension of the anomaly in Akkar, it is found that during the dry season such as June 2014, May 2015 and August 2015, the anomaly had a bigger width than during the wet season. This might be caused by the lack of water movement which allows the anomaly to be seen clearly. In Akkar area, an anomaly is visible every month. The anomalous could be explained as follow [19] :

Based on a research done by Shaban (2010) titled as “Geothermal Water in Lebanon: An Alternative Energy Source”, it was mentioned that in Akkar region along the Syrian border, there is a large basaltic rock plateau, and several observations for hot or warm water from springs and seepages at this area have been found. Field tests show that the average temperature of the springs ranges between 50˚C - 65˚C over various time periods. Furthermore, the hydrologic phenomenon of hot and warm water is not limited to the terrestrial environment, but extends into the marine environment. It showed the presence of geothermal water in Lebanon in four major geothermal domains especially in Akkar region. The presence of geothermal water in Akkar might explain the appearance of anomalous features on the maps. Another explanation might cause the presence of these anomalies such as the pollution from al Oustowan river that flows into this area [19] .

During February and November 2014, the amount of pollution was high compared to other months. These also corresponded with wet months, thus, river flow was greater and may have dispersed the plume. This cannot be said for 2015 when Nahr el Jaouz did not demonstrate an anomaly during the wet season but a large polluted area was visible in May 2015 and bigger than that of 2014. A feasible explanation might be the excessive dumping of wastewater.

During the wet season of 2014, Nahr Ibrahim has presented the appearance of anomaly and not during the dry season 2014. This might be due to the intensive water discharge during the wet season and its dryness during the summer. During the summer 2015, Nahr Ibrahim presented some anomalies on the maps. Similarly to Nahr el Jaouz, this might be due to the release of wastewater in the dry river.

An anomaly is visible in the thermal maps, meaning there is a differentiation in the water temperature. This might be caused by the submarine groundwater discharge present in the Jounieh area or there might be presence of warm water coming from ships that is contributing to the temperature differentiation. This anomaly only appeared in few months in 2014, but it was present during all months in 2015. This means that during 2015, Jounieh was more frequently polluted than 2014.

The following Table 2 presents the ranking of anomalies based on their frequencies, going from the presumed most polluted site to the least polluted site. The determinations are based on the appearance and dimension of anomalies on a monthly and annual basis.

Overall sea current dynamics in the area are directed from the south to the north, which control the spreading regime of pollution and divert it to the north shoreline [20] . This explains why the pollution direction is directed to the north in the maps. During April 2014 and June 2014, the current was directed to the south in northern parts (Figure 2). This might be explained in a number of ways: 1) Due to the diversion of the water current caused by the river flow; 2) springs in the sea water; 3) by pollutant discharge and dispersion; 4) By short term wind.

Further considerations include the dry and wet seasons. Anomalies are most clear during the wet season associated with intensive river discharge creating turbidity. Some focus, therefore, was placed on when an anomaly appears during the dry season. These

cases suggest this might be caused by a certain type of pollution as opposed to turbidity and discharge [21] . Examples are provided in Figures 3 through 8 below.

Figure 3 shows the pollution in red and yellow. Areas of interest are labeled (1), (2), (3), (4).

(1) and (2) are the areas of potential pollution; (3) marks possible oil slicks; (4) is the passage of a ship.

In Figure 4 it can be observed that the areas (1) and (2) of potential pollution extracted from Figure 3 continue to exist however the ship and oil slicks are no longer present, possibly explaining the color differences.

A Principal Component Analysis (PCA) was performed to enhance all uncorrelated bands and maximize any differences in ocean waters. The purpose is to emphasize potential pollutants. The same areas of interest are la-beled and clearly identifiable. This map gave the same information as the 421 band combination. The color contrasts are more defined and no new sources of pollution are apparent.

Again it is clear that the areas of potential pollution from Figure 5 (1, 2) still exist and although they are represented in different shades between the 2 dates in Figure 6.

It is feasible that different pollutants are being dumped between the two locations or that ocean circulation is causing them to be redistributed.

The spectral curves of all Landsat bands were examined for a profile straddling the area of pollution in site 1 in order to determine areas with strong absorption. The premise was that the band with the most absorption could be divided by the 2 adjacent bands using the equation:

to emphasize unique properties pertaining to pollution. It is possible that a zone with concentrated pollutants is emphasized by the light blue plume (1).

The plots in Figure 7(b) were examined for absorption features for the polluted areas. B₁ and B₂ was identified as having absorption centered over the feature to be detected (red lines), thus is used to create the simple ratio image Figure 7(a).

This time the blue plume in Figure 8 has dispersed and there is no longer a spike for these pollutants in site 1. This would be a good site to monitor on an operational basis for periodic dumping of polluting material. For this date, signs of pollution are reduced.