Historical variability in sea surface temperature (SST) in the North Atlantic (NA) is examined using trend and Empirical Orthogonal Function (EOF) analyses of annual and summer means from three interpolated monthly datasets: Hadley Centre Sea Ice and Sea Surface Temperature (HadISST1), Extended Reconstruction of SST (ERSST), and Centennial in situ Observation-Based Estimates (COBE). Comparisons with time series of upper-ocean temperature from four monitoring sites off Atlantic Canada reveal substantial similarity in the interannual to multi-decadal variability but notable differences in the longer-term trends. The magnitude of decadal-scale variability is comparable to, or greater than, the long-term changes in all of the datasets; together with the trend discrepancies, this needs to be considered in climate change applications. Averaged over the NA, the annual means have a long-term increasing trend and a pronounced multi-decadal variation, resembling those in global mean (land-ocean) surface temperature and the Atlantic Multi-decadal Oscillation (AMO). There is remarkable similarity in the spatial and temporal variability of the three leading EOF modes from the different gridded datasets, with the first highly correlated with the AMO, the second modestly correlated with the winter North Atlantic Oscillation, and the third apparently related to ocean circulation variability. Trends since 1981 are generally two to three times larger than those since 1900 and 1950, which is at least partly related to the phase of the AMO. Trends in the summer means are generally larger than in the annual means. Overall, the results provide support for both anthropogenic global warming and decadal-scale natural variations making important contributions to ocean climate variability in the Northwest Atlantic.

Alnus cremastogyne Burk. is a nonleguminous, nitrogen-fixing tree species. It is also the most important endemic species of Alnus Mill. in China, possessing important ecological functions. This study investigated population genetic variation in A. cremastogyne using 175 trees sampled from 14 populations native to Sichuan Province with 25 simple sequence repeat (SSR) markers. Our analysis showed that A. cremastogyne has an average of 5.83 alleles, 3.37 effective alleles, an expected heterozygosity of 0.63, and an observed heterozygosity of 0.739, indicating a relatively high level of genetic diversity. The A. cremastogyne populations in Liangshan Prefecture (Meigu, Mianning) showed the highest level of genetic diversity, whereas the Yanting population had the lowest. Our analysis also showed that the average genetic differentiation of 14 A. cremastogyne populations was 0.021. Analysis of molecular variance (AMOVA) revealed that 97% of the variation existed within populations; only 3% was among populations. Unweighted pair-group method with arithmetic means (UPGMA) clustering and genetic structure analysis showed that the 14 A. cremastogyne populations could be clearly divided into three clusters: Liangshan Prefecture population, Ganzi Prefecture population, the other population in the mountain area around the Sichuan Basin and central Sichuan hill area, indicating some geographical distribution. Further analysis using the Mantel test showed that this geographical distribution was significantly correlated with elevation.

Corylus yunnanensis (Franch.) A. Camus is a deciduous shrub, native to the Hengduan Mountain of Qinghai–Tibetan Plateau, and is an economically and ecologically important woody crop species. In the present study, one hundred and fifty trees sampled from ten populations of C. yunnanensis in Sichuan Province were investigated to assess the population genetic variation using nine SSR markers. The results revealed that C. yunnanensis has an average value of 12.111 alleles, 3.376 effective alleles, an expected heterozygosity of 0.648, and an observed heterozygosity of 0.630, presenting a relatively high level of genetic diversity. The C. yunnanensis populations in Maoxian and Wenchuan of Aba Prefecture expressed the highest value of genetic diversity, whereas the Hanyuan and Muli populations showed the lowest. Moreover, the genetic differentiation of ten C. yunnanensis populations averaged to 0.106. Correspondingly, AMOVA revealed that 87% of the total variance was accounted for the variation within populations, and only 13% was among the populations. Both UPGMA and Bayesian STRUCTURE clustering suggested that the ten C. yunnanensis populations could fall into three clusters: the Aba Prefecture population, the Ya’an population, and the population of Ganzi and Liangshan Prefecture, indicating a significant geographic distribution, which was also confirmed by the Mantel test. Our study could provide a better understanding of population genetic diversity, and serve valuable information for the genetic improvement of C. yunnanensis.