1620 1700 1800 »00 1970
FIGURE I 1.2 (upper diagram): Relationship between the number of days with a canal frozen between Haarlem and Leiden (Holland) per winter and average winter temperatures.The best-fit linear regression line is shown, (lower diagram): Reconstructed winter temperatures at De Bile, Holland, expressed as departures from the long-term average (+2 °C) based on regression shown and historical records of canal freezing frequencies (van den Dool et a/., 1978).
of rain days in parts of Japan and August temperature (based on instrumental data). He then used this to reconstruct August temperature from 1770-1840 using the historical records of rainfall frequency (Fig. 11.3).
Some observations may not need direct calibration if recent comparable contemporary observations are available. This applies to such things as rain/snow frequency, dates of first and last snowfall, river freeze/thaw dates, etc., providing urban heat island effects, or technological changes (such as river canalization) have not resulted in a non-homogeneous record.
The most widely recorded meteorological phenomenon in historical documents is snow, in terms of the date of the first and last snowfalls of each year, or the number of days with snow on the ground or the frequency of days with snowfall. Some of the earliest snowfall observations are from Hangchow, China, for the period A.D.
Was this article helpful?