Rt Hon Lord Lilley

    Lord Lilley: To ask Her Majesty’s Government, further to the report Prolonged Siberian heat of 2020, published on 15 July, whether the Met Office’s analysis assumes that all the months of June during 1926–2020 have the same statistical distribution for their daily temperature maxima at Verkhoyansk; and if so, why.

    Lord Callanan: The analysis does not assume that all the months of June during 1926-2020 have the same statistical distribution for the highest of their daily temperature maxima at Verkhoyansk. Instead, the method involves modelling secular changes in the data by a covariance with smoothed Global Mean Surface Temperature (GMST) and first removing this to create a set of residuals that may be assumed to be stationary and to which a Generalized Extreme Value (GEV) fit is then made. This method, based on peer-reviewed literature, is set out in the linked methods document which accompanies the report.

    Lord Lilley: To ask Her Majesty’s Government, further to the report Prolonged Siberian heat of 2020, published on 15 July, whether they will place a copy of a quantile-quantile plot of a GEV distribution against the distribution of June maximum temperatures at Verkhoyansk during 1926–2020 in the Library of the House.

    Lord Callanan: I have arranged for a copy to be placed in the Libraries of the House, along with explanatory text and supporting documents.

    Lord Lilley: To ask Her Majesty’s Government, further to the report Prolonged Siberian heat of 2020, published on 15 July, what assessment they have made of whether the report’s sample size is large enough to justify the use of a GEV asymptotic approximation when analysing the Verkhoyansk temperatures.

    Lord Callanan: The report’s sample size is large enough to justify the fitting method used in the analysis as daily temperature data at approximately one-year intervals are not significantly correlated from year to year over continental regions. The maximum daily temperature in a particular month or season in a particular year in a continental region does not serve as a good predictor of the maximum daily temperature in the same month or season in the following year (over and above the long-term effect of climate change) due to natural variability of the climate system. The June-July daily maximum temperature data over timescales from 1 day upwards shows no significant autocorrelation (correlation with itself across a period of time) above 1 week timescales and confirms that the 94 data points used in the report are independent, a sufficiently large number to justify the fitting method used in the analysis. The analysis in the report is based on peer reviewed methodology as set out in a paper by Van der Wiel et al. referenced in the report, which applied this fitting method to a dataset of a comparable size.

    Share.