Ash detection: software development and analysis
@bandersen @justins This issue is still under development, but below is a good starting point. Justin can add details on existing software that may help.
The World Meteorological Organization (WMO) sponsored an international inter-comparison of satellite-derived volcanic ash products. Due to resource challenges, the inter-comparison activities ended in 2018, before all of the desired analyses could be completed. The goal of this task is to complete the analyses related to satellite-based ash detection (e.g. is ash present in a given satellite pixel?). The inter-comparison work plan includes a description of the underlying data sets and inter-comparison objectives. The satellite scenes chosen for analysis are listed in this Excel spreadsheet. Detailed information on each data set contributor and the associated algorithms can be found in this spreadsheet. In terms of ash detection, human expert analysis of satellite imagery is the upper bound of performance for automated algorithms. As such, the primary objective of this task is to compare each manual analysis data set to the corresponding automated algorithms for the same satellite and imager date/times. Comparisons can be quantified via statics such as POD, FAR, CSI, etc.
Important Logistical Information:
We only have 1 copy of the WMO inter-comparison results, so be sure not to delete or write over (@justins maybe change permissions to read only so none of us accidentally delete or write over?) RAL data was changed to read only, 2018_WMO_INTERCOMPARISON is running in a similar fashion, will take multiple hours.
Location of all submitted data sets:
/apollo/grain/common/ral_wmo_ftp/ftp.rsg.rl.ac.uk/incoming
Location of manual analysis reference data:
/apollo/grain/common/2018_WMO_INTERCOMPARISON/netcdf_manual
Location of VOLCAT data sets:
/apollo/grain/common/2018_WMO_INTERCOMPARISON