JasonAlongTrack: A reformatted version of the Integrated Multi-Mission Ocean Altimeter Data for Climate Research Version 5.1

JasonAlongTrack contains geo-registered along-track sea surface height anomalies with respect to the DTU15 mean sea surface at 1-second intervals from Jason-class altimeters, reformatted for convenience into a 3D array with dimensions of along-track direction by geographically sorted track number by cycle.
This is a reformatted version of Beckley et al.'s Integrated Multi-Mission Ocean Altimeter Data for Climate Research complete time series Version 5.1 dataset, available from https://podaac.jpl.nasa.gov/dataset/MERGED_TP_J1_OSTM_OST_ALL_V51.   
The changes are as follows. Altimeter passes are sorted according to their initial longitude, then split into descending and ascending potions with all descending tracks preceding all ascending tracks. Descending tracks are then flipped so that latitude increases in the alongtrack direction for all tracks. This leads to a 3373 x 254 matrix of observational locations, with the first dimension being the along-track location and the second dimension being the track index. Sea surface height anomaly, time, and flag values are then placed into their correct locations within this matrix, such that these three variables are all of size 3373 x 254 x K where K is the number of cycles, currently 1087. A very good approximation to the time at each of the 3373 x 254 x K observation points is constructed with a length K array of cycles times together with a 3373 x 254 array of time offsets. A median-based editing criterion in introduced to identify a small number of suspect data points.  These are set to a value of NaN in sla, but their positions and values are recorded in rejected_index and rejected_values, respectively.  The DTU15 mean dynamic topography (mdt) is included, in addition to the mean sea surface field already provided, interpolated onto the track locations using bicubic interpolation.  Finally, an estimate of the small-scale noise level, sigma, is produced using a wavelet transform filter.