<January 5, 2008->
Llanystumdwy Rainfall and Temperature Observations
Rainfall and temperature records have been kept from 6 February 1994 to the present from observations in the garden of a house in the village. At 13.00 hours each day the temperature and rain accumulated in the previous 24 hours were recorded manually.
Temperature
The following chart shows the annual temperature cycle based on 13 years of observations.
Although these readings were not made with a maximum and minimum thermometer, they show a very close correspondence with the maximums recorded at Llansadwrn weather station about 37 km away.
We were interested to discover if any evidence of climate change could be detected in the temperature recordings. Accordingly, all 4,713 temperature observations were plotted against the date on which they were made and a linear trend line fitted through the points. The results are shown on the following chart.
The line shows the temperature having risen from 11.66 to 12.61 degrees during the period, i.e. a rise of 0.94 degrees. From R-squared and the number of degrees of freedom (4711), we calculate the t-value for the trend line to be 3.585 which is significant at 0.1% level, or has a less than one in a thousand probability of being due to an accident of sampling.
We treated maximum values recorded at Llansadwrn station in the same way. The trend line was y = 0.000301x + 2.336788 and R2 = 0.003252 with 3,285 degrees of freedom. It is close to the corresponding data for Llanystumdwy, i.e. for 1998 to 2006 whose trend line is y = 0.0003227x + 0.7039481 with R2 = 0.0031369. The value of t for the Llansadwrn data is 3.274 which is significant at less than 1%. The reduced significance compared with Llanystumdwy is attributable to the smaller number of degrees of freedom resulting from data for 9 rather than 13 years. We conclude that the rising air temperatures component of climate change is reflected in temperature observations at Llanystumdwy and Llansadwrn.
Rainfall
The chart below shows results for Llanystumdwy compared with Llansadwrn weather station. Over the whole period, Llansadwrn records 2.6% more rain than Llanystumdwy. The stations agree quite well on the variations between years, presumably because they are relatively close to each other (37 km).
A plot of the average of the annual totals for the two stations gave a trend line that suggests it is getting wetter by about 25 mm per year, but this was not statistically significant. The monthly distribution of rainfall is shown in the chart below.
June and July are our driest months and October and November our wettest.
As there were 13 years of observations available, we considered it worthwhile to calculate daily average rainfalls from the data set of 4,713 observations to see what happens. The results in the chart below show that despite 13 years of data collection there is not a great deal of smoothing out of fluctuations achieved by averaging.
However, there seem to be clusters of high and low values, so these were explored by calculating 7-day moving averages of the daily averages. For example, the value plotted for 4 January would be the average of the 7 values for 1 to 7 January inclusive; 5 January include data for 2-8 Jan and so on. To ensure that values at the ends had the same treatment, values at the opposite end of the year were used. Thus the 1 January value averages data for 29-31 Dec and 1-4 Jan, i.e. the year was treated as a repeating cycle. The resulting smoothing effect on the daily averages was a lot less than expected. Indeed, the chart below seems to show that there is a more or less rhythmic sequence of peaks and troughs throughout the year.
We have also compiled a plot of this chart, only for 1998 onwards, with both Llanystumdwy and Llansadwrn data. The latter shows the same effect with almost all the peaks and troughs coincident on those of Llanystumdwy. We are currently investigating whether the above finding is just a statistical anomaly or whether there is a real weather phenomenon behind the fluctuations. All lunar cycles and perturbations seem unlikely causes as they do not synchronise with calendar years. Comments towards clarifying what the above chart means would be welcome.
Llanystumdwy, 8 January 2006
The above table shows interesting trends in our mean January temperatures in Llanystumdwy from 1994 to 2007 (analysed January 2008). I quote David Heaf's email with the above attachment.
'I've just finished processing my rain/temp data
for 2007. Strangely it
comes out with the warmest average annual temp for 14 years of data
collection -- 13.3 deg C.
And the attachment vividly shows how our Januarys have warmed since
1994:
5.75 to 7.4 deg C.
Get planting your vineyard on that sunny slope!'
