Occasionally, problems with weather station metering equipment result in missing average daily temperatures. We denote missing data using a “–99” flag. Depending on the level of accuracy needed, the missing temperatures can be approximated as the mean of the daily temperatures recorded immediately before and after the missing data. For more accuracy, the user may find other temperature resources. For example, the mean of the daily minimum and maximum temperatures for many US cities dating back to about 1994 can be found http://www.wunderground.com/.
Please note that the “average” daily temperatures on our site are calculated as the mean of 24 hourly readings, rather than the mean of the daily minimum and maximum temperatures. Thus, our 24-hour average daily temperatures will not be exactly the same as the mean of the daily minimum and maximum temperatures. However, our research has shown that the standard deviation between the two methods of calculating the average is small (about 1F) with no significant bias. Thus, the error from filling missing temperatures in our data sets with “(Min+Max)/2” averages is small.