The dataset is organized in a folder (pngs) that includes all the 2100 processed SAR images in PNG format, a text file (tagsInfo.txt) that records all the tags assigned by the five experts for each image, and another text file (readme.txt) that provides an overall description of the dataset.

In the tagsInfo.txt file, each image and its tags are organized as follows:
ImageName,longitude,latitude
Expert1,WS,WS>MC,WS~MC,WS<MC,MC,NV,UN,AB,GF,CP,RC,LW
Expert2,WS,WS>MC,WS~MC,WS<MC,MC,NV,UN,AB,GF,CP,RC,LW
Expert3,WS,WS>MC,WS~MC,WS<MC,MC,NV,UN,AB,GF,CP,RC,LW
Expert4,WS,WS>MC,WS~MC,WS<MC,MC,NV,UN,AB,GF,CP,RC,LW
Expert5,WS,WS>MC,WS~MC,WS<MC,MC,NV,UN,AB,GF,CP,RC,LW

---->Description of the file header 
ImageName: file name of png image. It can be used to trace back to the S-1 WV SAR SLC products.
longitude: central longitude 
latitude: central latitude
Expert1: DV (Doug Vandemark)
Expert2: RF (Ralph Foster)
Expert3: JS (Justin Stopa)
Expert4: CW (Chen Wang)
Expert5: AA (Alex Ayet)
WS,WS>MC,WS~MC,WS<MC,MC,NV,UN,AB,GF,CP,RC,LW: binary (1=present or 0=not present) indicator of tags

---->Details of the defined tags
WS: Wind streaks 
WS>MC: more WS than MC
WS~MC: equally WS and MC
WS<MC: less WS than MC
MC: Micro convective cells
NV: Negligible atmospheric variability
UN: Uniformity 
AB: Atmospheric mass boundary
GF: Gust front
CP: Cold pool
RC: Rai cell
LW: Low winds