Answers are grounded in this paper and cite their sources.
Monthly minimum and maximum air temperatures recorded at Rothamsted in Hertfordshire, England, from 1901‐1978 are analysed using the generalized extreme value distribution. Temperatures with return periods of 50 and 100 years, with their associated confidence limits, are produced for each month. The very low temperatures recorded in December 1981 and January 1982 are considered in the light of these results. The probability of air frost for each month is evaluated. From the distributions of monthly extremes the distributions of annual minima and maxima are derived. The assumptions implicit in the analysis, independence and stationarity, are examined. It is found that the data can be considered stationary in all cases except for the October minima, which appear to have been increasing over recent years. Correlations between monthly extremes are generally low with little discernible pattern, but confidence limits have not been determined.
No takes yet. Share an insight, caveat, or question.
Challenor et al. (1983) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: