VNP14IMG is the short-name for an experimental Level-2 (L2) VIIRS 375 meters active fire product, which provides day and nighttime active fire detections over land and water globally; detections over water are typically associated with gas flares. This product’s algorithm is based on a similar, evolving Earth Observing System’s (EOS) MODIS baseline product (Thermal Anomalies and Fire) that continues to build a 1-km active fire data record since 2000. Maintaining close similarity in their product algorithms ensures the ability to support data continuity between the products derived from the MODIS and VIIRS missions.
In contrast to its ~5-minutes of observation in the MODIS version, the VIIRS L2 375 meters active fire product contains ~6-minute orbital segments from multiple scans. Each scan contains a fixed number of rows (total of 32), with each row representing one detector. VNP14IMG’s image dimensions are 6400 columns by 6464 or 6496 rows (total rows reflect the number of scans, which may vary in order to better accommodate ~6-minute orbital segments), and covers a ground swath width of approximately 3060 kilometers.
The experimental nature of this product stems from the fact that it leverages the 375-m (I-channels) data that were not originally designed to support active fire detection, and therefore is a relatively recent addition to the VIIRS land product suite. The VNP14IMG algorithm includes special handling of abnormal radiometric conditions in the mid-infrared emissive channel (I4) due to frequent pixel saturation over large and/or intense fires. By adding modifications to the original EOS MODIS algorithm to accommodate the higher spatial resolution, at 375 m, the VNP14IMG product provides better response over fires of relatively small footprints, and also an improved mapping of larger fires.
The VNP14IMG 375 m active fire product uses input data from all five 375 m I-channels (I1 – I5), and the dual-gain 750 m mid-infrared, M13, channel. Based on the heritage EOS–MODIS algorithm, the VIIRS active fire detection process uses a multispectral contextual algorithm to identify sub-pixel fires and other thermal anomalies in the Level-1 input data. Active fires are detected based on a combination of fixed and contextual tests for both daytime and nighttime observations.
The L2 VNP14IMG product offers several pieces of information. They include the following: a full-blown image classification product (two-dimensional fire mask), fire pixel’s latitude and longitude, fire pixel’s I4 and I5 brightness temperatures (BT), background’s I4 and I5 BTs, background’s I4 and I5 BT mean absolute deviation, background’s I4-I5 BT difference mean absolute deviation, fire pixel’s M13 radiance, background’s M13 radiance, mean background I4-I5 BT difference, fire radiative power (FRP), number of adjacent cloud pixels, number of adjacent water pixels, background window size, detection confidence. The L2 VNP14IMG product is offered in the netCDF4–HDF5 hybrid format that leverages several individual as well as collective strengths of the two data/file storage formats. VNP14IMG is currently available only through the LAADS Web.