I believe some re-analysis will prove that 91L is in fact the original after all, and the recently emerged TW from Africa confused the maelstrom and made it difficult to determine for approximately 36 hours, which entity was dominate. First we had a soon-to-be 90L with a -80C top region, and some sufficient enough banding to warrant "yellow" coding. Then, the wave comes off Africa and bullies it's way into the region, and seeds of 90L seemed to yield to that arrival... getting temporarily lost, such that when 90L finally got around to being assigned, it was done so further E - perhaps inappropriately associated with a proving to be non-genesis TW.

The reason I speculate that is because the convection associated with that follow-up wave appeared to wane actually, and then this new persistent convective region re-materialized further W with what originally sparked the yellow coding - I believe in a discrete sense that should be deemed the actual 90L in conserving entities, because that is what has really survived the interaction.

That said, the region is also kissed in general by an impressive monsoonal ITCZ trough that is somewhat below normal surface pressure in that particular region ... acquiring cyclonic banding as we type. There does not appear to be any direct inhibitors to development; I thought easterly shear may have an impact but that does not seem to be evolving.

The longer term steering field appears to be fairly locked in the Global Numerical Models (GFS/GGEM...etc), and is understandable. There is a positive phase state in the North Atlantic Oscillation currently underway, and forecast to remain that way in the 21 member GFS ensemble mean. What this teleconnection supports is stronger rather than weaker subtropical ridge presence over the next 5-7days, across the breadth of the N Atlantic. Barring any anomalies relative to, a generalized W to WNW track motion of "91L" (which may very well be a depression sometime over the next 2 days) right out through any meaningful time range of prediction (obviously beyond D4 we have to take even a tightly clustered modeled solution with a degree of incredulity) should be expected.

Still, populations in the NE Caribbean through the Puerto Rico archipelago and up through the Bahamas should monitor this over the next few days.

It should be noted that the 00z ECMWF solution offered an intense hurricane solution entering the SE Bahama archipelago in 6 days.