Generally, an annual time series for the Aegean SST

is used to explain the Eastern Mediterranean Transient (EMT) phenomenon. Much colder winters (winter mean temperature minus winter standard deviation < 14.5 °C) occurred in 1983, 1996, 1989, 2006 and 1993 (the years are ordered according to winter temperature starting from the lowest one), whereas much warmer winters (winter mean temperature plus winter standard deviation > 15.4 °C) occurred in 2005, 2011 and 2010 (the years are ordered according to winter temperature starting from the highest one) (data not shown). The cold winter in 1993 may explain the initiation of the EMT over the southern Aegean Sea in the early 1990s. selleck compound The EMT (Klein et al. 1999) formed because the Aegean Sea salinity increased from 1987 to 1991, followed by cold winters in 1992 and 1993. In this section, the SSTs up to 2100 projected using the RCP26, RCP45, RCP60 and RCP80 scenarios are investigated using CMIP5 ensemble means. Table 3 shows the performance of various CMIP5 ensemble mean AZD2014 scenarios used to estimate SST values. The SSTs obtained are compared directly with AVHRR SST annual and monthly data. The results in Table 3 are subjected to the t-test to determine whether the SSTs obtained using

CMIP5 ensemble means are significantly lower or higher than the AVHRR SST values. The annual CMIP5 ensemble mean scenarios significantly underestimate SST over the study area, most markedly over the Adriatic sub-basin (≈ 3.1 °C) and in June.

The Mediterranean Sea and the AAM sub-basin display significant monthly variation in the lower SST estimates, especially in January. In contrast, the Black Sea displays higher monthly SST estimates in cold months and lower monthly SST estimates in hot months. Generally, CMIP5 ensemble mean scenarios result in estimated SSTs that are much lower than those observed from AVHRR satellite images during the examined control period, indicating that the study area SST may be much higher than that projected for the end of the current century using CMIP5 ensemble means. All CMIP5 ensemble means used for SST scenarios indicate a significant warming over the 2000–2100 period in the study area, most (least) markedly using the RCP85 (RCP26) scenario, as seen in Figure 7a and Table 4, Florfenicol i.e. in the AAM sub-basin (0.3–1.6°C), Mediterranean Sea (0.5–2.6°C) and Black Sea (0.5–2.6°C). The AAM sub-basin displays the weakest warming trend in the current century, weaker than those of the Black and Mediterranean Seas. The Mediterranean Sea displays spatial variability in warming trends between its various sub-basins, the maximum (minimum) warming trend occurring in the Ionian and Levantine sub-basins (Alboran sub-basin), as seen in Table 4. All eight Mediterranean Sea sub-basins are projected to warm significantly, most (least) pronounced over the Levantine (Alboran) sub-basin. The Levantine (Alboran) sub-basin is projected to warm during the 21st century by amounts ranging from 0.