Jerome, Nicholas TanChilingaryan, SurenKopmann, AndreasWieser, AndreasKarsten Rink and Ariane Middel and Dirk Zeckzer and Roxana Bujack2017-06-122017-06-122017978-3-03868-040-6https://doi.org/10.2312/envirvis.20171097https://diglib.eg.org:443/handle/10.2312/envirvis20171097With Doppler wind lidar producing significant amounts of data, providing means to extract relevant information from the data that describes atmospheric phenomena such as rain and low-level clouds is of vital importance. However, a Doppler wind lidar with a 10 Hz sampling rate produces large-scale of data at approximately ten million data items per day; therefore, introducing challenges in perceptual and interactive scalability. We present an interactive web-based visualisation system that provides summary displays of the heterogeneous lidar data. Our system applies the client-server paradigm, where our server extracts information and encodes primary lidar attributes into image's colour channels. Then, we load these encoded images and show lidar data in multiple forms at the client-side. In contrast to script-based tools such as Matlab and Ferret, our system allows researchers to begin analysing the extensive data using a more top-down methodological approach. In particular, we implemented features like zooming, multivariate filtering, and hourly variance heat map, in which GPU shaders filter data according to specific attributes. With the encoded images readily stored at the server, researchers can browse through the vast amounts of data interactively.I.3.3 [Computer Graphics]Picture/Image GenerationMultivariate Encoded ImageI.3.8 [Computer Graphics]ApplicationsDoppler Wind Lidar Visualization10.2312/envirvis.201710977-11