Improving the predictive models that identify trends underlying climate change
To effectively mitigate the worst effects of global warming will require an understanding on the underlying processes involved and at present climate forecast models are the only available method for predicting the future evolution of climate change. These simulations are underpinned by data derived from remote sensors, located either on the ground or on orbital satellites. However, changes in the ambient temperature and pressure between these two locations can introduce significant result variability. To differentiate between the natural changes of the climate system and ‘anthropogenic’ man-made signals, the accuracy and traceability of remote-sensor derived climate data must improve ten-fold. The European Space Agency (ESA) has also highlighted the need for climatologists - who are responsible for using climate forecast models - to better understand possible sources of experimental data spread and their corresponding effects on climate trend analysis. Climatologists must also be aware of the importance of the data’s SI traceability and be able to assess and calculate uncertainties in the Earth Observation (EO) traceability chain. In line with this need, the EMRP project European metrology for earth observation and climate delivered a training course in 2015 on 'Uncertainty Analysis for Earth Observation' to scientists from across Europe. The course generated interest world-wide from companies working in weather, climate, space and academia - highlighting the relevance and high demand for this training. However, due to global interest in this subject, an online e-learning course was required to enable wide-spread knowledge transfer and ‘upskill’ more individuals, providing them with the freedom to access this training ‘on demand’.
This EMPIR project aimed to maximise the impact and potential of the initial EMRP project, transferring valuable metrology skills to the EO community and disseminating project outputs more widely. It developed a scalable, freely available, e-learning training course to introduce a structured approach to uncertainty analysis. Comprised of three modules and supporting materials, the full course is available and compatible across multiple devices, operating systems and browsers - allowing users easy remote access from anywhere in the world at any time. Module 1 introduces some fundamental concepts and procedural approaches relating to uncertainty analysis in EO, and Module 2 applies this to a real-life practical example of radiometric instrument calibration. Module 3 examines the APEX spectral instrument that uses solar reflectance measurements to study natural phenomena aboard research flights, such as vegetation cover or air quality. The module also explores the application of uncertainty analysis to post-launch radiometric calibration techniques. Furthermore, the scalable nature of the course allows for the development of new modules, such as for surface temperature or surface albedo analysis. The course has already been used in analysing data on the radiance effects of single trees on their surroundings.
Climatologists’ improved knowledge of EO uncertainty calculations will have significant societal and environmental impact, ultimately providing more trustworthy climate forecasts and increasing confidence in climate adaptation and mitigation policies.