Aotearoa New Zealand Initial Ceilometer Analysis

Executive summary
Cloud cover significantly impacts the Earth’s energy balance by reflecting incoming solar energy and trapping infrared radiation. Accurate representation of clouds in climate models is crucial for predicting climate change. However, biases in cloud simulations remain, particularly in high latitudes, as evidenced by the World Climate Research Programme’s CMIP3 models. These biases often stem from inadequate cloud cover representation and difficulties in simulating super-cooled liquid water clouds, leading to errors in shortwave radiation estimates. This study aims to address these biases by analysing cloud data from ceilometers across Aotearoa New Zealand, focusing on the proportion of super-cooled liquid water clouds.
Data was collected from a network of 16 Vaisala CL31 ceilometers operated by MetService at various airports across Aotearoa New Zealand, covering both North and South Islands and Chatham Island. The ceilometers, primarily used for detecting cloud base height and vertical visibility, provided data from July 2021 to February 2023. The uncalibrated backscatter data was converted to NetCDF format using cl2nc software and processed with the Automatic Lidar and Ceilometer Framework (ALCF) for consistent noise removal, calibration, and cloud detection.
The default backscatter threshold for cloud detection, developed using LIDAR data from Antarctica, was found to be too low for mid-latitudes, leading to misclassification of surface aerosols and instrument artifacts as clouds. A higher threshold was selected to improve cloud detection accuracy. This adjustment significantly reduced the misclassification of low-altitude clouds while maintaining the detection of higher altitude clouds.
Rainfall, distinguishable in backscatter plots as diagonal bands, can interfere with cloud detection. The backscatter of rain can sometimes be as strong as clouds, complicating the accuracy of the automatic cloud identification mask generated by ALCF. Future analyses will incorporate rain gauge data to filter out periods of rainfall more effectively.
Hourly climatologies of cloud occurrence showed minimal variation between sites, with a slight increase in higher altitude clouds between 1100 and 1700 at most Te Ika-a-Māui / North Island sites. Most sites had a maximum cloud base height of around 2000 metres although some were higher at around 2500 metres, and some lower around 1500 metres. This could be due to weather patterns or surrounding terrain; further investigation is required. The available data was insufficient for developing monthly climatologies.
Future steps will include developing a rain mask, dividing the data based on the prevailing south westerly wind patterns and further analysing cloud heights concerning wind direction. Advanced models capable of classifying super-cooled liquid water clouds, such as G22-Davis and G22-Christchurch, will be used to refine climatologies and improve the accuracy of cloud representation in climate models.