2001). As expected from previous studies (Suursaar et al., 1995, Astok et al., 1999 and Raudsepp et al., 2011),

our results of cumulative fluxes also indicated an annual net outflow in the Suur Strait. The northward fluxes (on average approximately 60 km3 yr− 1) were somewhat larger than those calculated by Raudsepp et al. (2011) for 2008 (23 km3 yr− 1). The difference could have occurred for several reasons. Firstly, Raudsepp et al. (2011) admitted that their (single-point) measuring site, which was at the depth of 3.5 m on one side of the strait, might not fully represent the Alpelisib mouse whole cross section. Also, the wind stress from the HIRLAM (High Resolution Limited Area Model) could have underestimated the winds above the narrow strait, as the corresponding model cells probably included land surface properties. On the other hand, as indicated by the long-term average wind speed at Kihnu (5.66 m s− 1 in 1966–2011 vs.

4.15 m s− 1 at Virtsu), our forcing may have overestimated the winds above the Väinameri part of the model domain. Finally, unlike Raudsepp et al. (2011), our calculations included constant 32 km3 yr− 1 inflows from rivers into the Gulf of Riga. (The seasonal variations in discharges have been largely controlled by the Riga Hydroelectric Power Plant on the River CP-868596 solubility dmso Daugava since 1974.) Although the larger part of that discharge ought to ‘flow out’ through the Irbe Strait, no one has any certain knowledge Ribonucleotide reductase of the actual proportion. In general, the inflow through the Irbe Strait should mirror the outflow through the Suur Strait, but in the relatively wide Irbe Strait under certain conditions in- and outflow can take place simultaneously (Lilover et al. 1998). The question could probably be solved either by studying Lagrangian

particle tracks (like Zhurbas et al. (2010) did in the Baltic Proper), or water ‘age’ (see e.g. Andrejev et al. 2004). Summarizing the problem for the Gulf of Riga, the interannual proportions as well as climatological shifts should remain the same, even though the exact magnitude of flows is unknown. Being differently exposed (Kõiguste to SE, Matsi mostly to S-SW), the locations showed a rather different wave time series (Figure 10). According to formal linear trends, the average wave heights have probably decreased at both locations. While at the windward Matsi the overall linear trend decreased very slightly in 1966– 2011, the trend was a significantly falling one near Kõiguste (Figure 10a). However, on the basis of annual maxima and higher quantiles (90%, 99%), the trends increased near Matsi, but still decreased near Kõiguste (Figure 10c,d). Especially at Matsi, the wave heights showed some quasi-periodic cycles with high stages in 1980–1995 and again after about 2007. The cycles basically followed those in atmospheric processes (Figure 9; Jaagus et al. 2008).