Temperature Verification

The daily maximum and minimum temperatures from the AWS are routinely compared against a digital max/min check thermometer and also against official stations in the nearby BoM AWS Network. This page outlines our methods of comparison and presents the ongoing results.

Check Gauge Comparisons

The AWS thermo-hygrograph is housed in a solar powered fan aspirated radiation shield (FARS) that protects it from direct solar radiation and other sources of radiated heat. The digital max/min check thermometer is housed in a standard BoM Stevenson Screen, albeit with a replacement roof made from thicker panels of much denser wood than the original roof. For the purpose of these comparisons, the temperatures recorded by the check thermometer are regarded as the reference values.

A daily comparison is made by subtracting the temperature recorded by the AWS from the temperature recorded by the check thermometer. The daily results are collated then analysed, and are presented in the following table under the heading Mean Check Bias. If the bias is positive it means the check thermometer exhibits a warm bias compared to the AWS, and conversely if the bias is negative it means the check thermometer exhibits a cool bias compared to the AWS.

The aim of these comparisons is to achieve a negligible bias. This indicates a very good agreement between the co-located sensors performing under the same conditions. This is the most reliable measure of the accuracy of the AWS maximum and temperature temperature data for this location.

BoM Network Comparisons

The nearby BoM AWS Network consists of stations at Laverton RAAF, Avalon Airport, Point Cook RAAF, Melbourne Airport and Essendon Airport. The temperature sensors used in these AWS are highly accurate platinum resistance thermometers housed in a standard BoM Stevenson Screen that is located in a well maintained open area with good all around exposure in accordance with BoM specifications.

Melbourne Airport and Essenson Airport are both located further inland away from the direct influence of Port Phillip Bay, unlike Truganina and the other BoM stations, and may not give representative temperatures in all situations. They are therefore excluded from the daily comparisons, even though the data are collected and reviewed. It is also evident the temperatures recorded at Laverton, Avalon and Point Cook, hereafter referred to as the comparison stations, can vary quite considerably even under the same conditions.

Since no single station is a perfect buddy, the buddy system used in the daily comparisons includes all of the comparison stations. Every day for both maximum and minimum temperature, the highest and lowest temperatures across all comparison stations creates a daily temperature range. This range is then adjusted to take into account the variability associated with siting meteorological instrumentation in an urban environment. The adjustment is a nominal amount of 0.5° C that is added to the highest temperature and subtracted from the lowest temperature to create a daily buddy range.

A daily comparison is made by testing if the temperature recorded by the AWS is within the buddy range. The daily results are collated then analysed, and are presented in the following table under the headings Within Buddy Range and Outside Buddy Range.

The aim of these comparisons is to achieve a majority of the AWS temperatures within the buddy range. This indicates a high level of areal consistency with the nearby BoM AWS. For temperatures outside of the buddy range the aim is to have a small variance.

Results

Early results for maximum temperature show a negligible bias between the AWS and the check thermometer, plus a high percentage of temperatures within the buddy range. This implies the data are accurate for this location and generally compare favourably with the nearby BoM AWS. For minimum temperature there is a moderate warm bias between the AWS and the check thermometer, plus a high percentage of temperatures within the buddy range. Since the AWS minimum temperature data generally compare favourably with the nearby BoM AWS and the maximum temperature has a negligible bias, it suggests the warm bias evident in the minimum temperatures recorded by the check thermometer is due to a localised nighttime heating.