Deriving Long-term Dynamics of Land Surface Temperature over Europe: Towards a Daytime normalized AVHRR LST Product

Kurzfassung

Land Surface Temperature (LST) is recognized as one of the Essential Climate Variables by the World Meteorological Organization. It is a key parameter for climate models and a direct indicator of global warming. The Advanced Very High-Resolution Radiometer (AVHRR) is the only sensor that has been providing spatially and temporally continuous daily measurements for 40 years. In the TIMELINE project, consistent LST products were developed from AVHRR over Europe. However, the different overpass times and the orbital drift effect hide actual trends and anomalies in LST. Several methods exist to account for this effect of varying acquisition times on LST time series, which can be classified into physical and statistical methods. An important requirement for these methods is that they preserve the actual trends in LST. However, especially for the period before the year 2000 no independent LST datasets exist to validate these methods. As an approximation, historical measurements of near surface air temperature (Ta) at various stations across Europe can be used. Despite the known differences between LST and Ta at short time scales, it is expected, that long term trends correspond in these two variables. In this study, a representant of a physical and a statistical daytime correction model, respectively, are applied to the TIMELINE LST data and their performance is analyzed at different sites with different land cover across Europe. The physical model builds on information about the daily temperature circle at the given location, which is taken from SERVIRI geostationary LST data. With this model all LSTs are modelled to 13.30h true solar time. The statistical model uses the regression between the LST anomalies and the corresponding SZA anomalies for each day of the year throughout the time series. This allows to remove the orbit drift effect for each sensor. An additional offset is used to adjust the observation times of the sensors. Both models are validated against SERVIRI LSTs at the respective sites. In a second step, trends of the normalized LST are compared to trends of Ta for various sites and landcovers across Europe. The first analysis shows promising results regarding both models. This study is taking another step towards a 40-year daytime normalized AVHRR LST product, allowing a wide variety of applications in the context of climate and land surface change.