Assessment of the bias introduced by the automatisation of climate records combining climatological and metrological approaches

Identifier:  TDX:2504

Handle: https://hdl.handle.net/20.500.11797/TDX2504

Authors:  Gilabert Gallart, Alba

Abstract:
During the last third of the 20th century, Automatic Weather Stations (AWS) have gradually replaced manual observations systems in many places around the world. This transition has the potential to introduce a systematic bias (AWS bias) into climate time-series when combined with measurements taken from manual (MAN) observing systems and constitutes one of the most striking problems which the climate community faces, as the quality and homogeneity of most global records is compromised. The objective of this work is the statistical description bias and its minimisation from derived climate series. This study can be divided into four parts. In the first parts the transition to AWS is examined with the exceptionally long parallel records of Ebro Observatory (Roquetes, Spain), Fabra Observatory (Barcelona, Spain) and Murcia Observatory (Murcia, Spain) at the daily scale. For this part is found that the main characteristic of the AWS bias is highly dependent on the AWS installed. The second part demonstrates how the introduction of metrological calibration procedures can reduce or smooth the transition to AWS and give traceability and improve the quality and reliability of the temperature time-series. To achieve this, three field trials are installed. As result is concluded that the introduction of these procedures reduces the bias and improves the quality of the temperature time-series. In the third part, the bias is analysed at the hourly scale. Here it is noted that the calibrated AWS (AWSc) reduces the bias at this scale, but other factors like the different response time of both sensors has a high effect on the differences AWSc – MAN. In the fourth part, the traceability of the adjusted temperature series by estimating the combined homogenisation plus instrumental uncertainty is estimated. This study shows that the consideration of these uncertainties also has an effect on the long-term temperature trend. As described in the previous paragraphs, this PhD thesis attempts to combine three disciplines -metrology, climatology and meteorology- to deal with the complex problem of automatic weather station transition, as a contribution to the global climate community.