Earth observation data is often analyzed image by image since this is the natural format for most data captured by earth observation satellites. There are however numerous applications, ranging from hydrology to climate change analysis that require a detailed understanding of the temporal behaviour of geophysical processes. This is were complex time series analysis is often necessary. Current tools are often not optimized for that kind of data access since in the past datasets from high resolution satellites were sparse in time. But with the current ramp up of the Sentinel program also high resolution SAR and optical data will be available with dense temporal sampling.
This lack of suitable software frameworks makes it necessary for each scientist or group to write custom code for data conversion and time series based analysis. Often the output of the analysis is then only available in custom ASCII or proprietary storage formats suggested by the specific programming environment.
To confront these shortcomings we have developed a Python toolbox including the following features:
- core support of the NetCDF CF (Climate and Forecast) Metadata Conventions, especially Discrete Sampling Geometries.
- conversion of supported image data/orbit swaths into a format optimized for time series access.
- easy handling of data defined on discrete global grids or collections of in-situ stations.
- an API for implementing support of new, non standardized datasets.
- run custom or predefined time series based calculations or comparisons (validation) on multiple datasets utilizing parallelization through IPython.
- analysis results are automatically stored as CF-compliant netCDF files.
- visualization of CF-compliant netCDF files in a web browser through a WMS-T compatible visualization tool.
The toolbox is structured into modules which are hosted on Github and licensed through a permissive open source license. Each release of the software has a DOI making it citeable and improving chances for reproducible research.