26. Data analysis and visualization in atmospheric sciences

Audience level:
March 11th 8:05 a.m. – 8:10 a.m.


This poster demonstrates the Python based data analysis and visualization in atmospheric sciences with particular focus on modeled and observational aspects of marine clouds.


The commercial programming languages such as Matlab and particularly IDL have been dominating the field of atmospheric sciences for the past few decades. However, the Python language and its scientific ecosystem have become a promising alternative owing to their open-source, easy-reproducible, and user-friendly natures. As a part of my dissertation, I study the early stages of rain formation, which is data intensive work, both from modeling and observational aspects. Analysis of observational data utilizes the netcdf4-python and PyNIO data accessing libraries, for data arithmetics and manipulation the NumPy package. Modeling work uses various functionality from the SciPy package, also benefits from the near 100X execution speed-ups of Cython hybrid language. Visualizations mostly in 2D form are created using the matplotlib plotting library and seldom using MayaVi if 3D analysis is required. High interactivity of IPython interpreter greatly eases the experimental investigation of the work. Single to multi-dimensional datasets from in-situ measurements, and remote radar/lidar/satellite retrievals of marine stratocumulus type clouds are at the center of this study. Analysis and visualization of numerous sensory information from these datasets help us to understand the linkages between particles afloat in the air and interactions taking place in clouds. Certainly, the data oriented use of Python has obvious advantages facilitating the computational life of atmospheric scientists.