This paper presents an attempt to summarize various sparse proxy series into continuous and exhaustive climatic data. Freeze-up and break-up dates, early meteorological records and tree-ring data have been combined for the Hudson Bay region and 22 continuous proxy series extending from 1700 to 1979 have been deduced. These new series in term provided the basis for a regressive reconstruction of six seasonal temperature series. Verification tests are successful mainly for the high frequencies components. The low frequencies variability is better estimated by a best analogues method. Both kinds of reconstructions have been combined to improve the results. The main characteristic of the reconstructions is a warming trend beginning at the end of the 19th century. Evidence for a beat wave resulting from 22-year solar and 18.6-year lunar nodal tidal cycles is presented. A phase analysis showed results consistent with other studies of summer temperature variability: temperature maxima correspond to sunspot minima ending an even cycle and are emphasized by the lunar maxima. Different phenomena are pointed out for autumn and winter temperatures: their maxima coincide to sunspot even maxima amplified by lunar minima. In spring, the transition season, these signals are not apparent.
展开▼
