Yes but it's the OTL situation. I'm talking about possible mozarabic states (the plausibility of it isn't the topic), critically around Valencia and Algrave, making the Galaico-Portuguese to not be separated (or the Valentian dialect to not exist by exemple).
Answering your questions as best as I can:
I doubt a surviving Mozarabic would be considered as a part of a dialectal continuum with Spanish much more than Portuguese is: The Pyrenean-Mozarabic family was already a seperate family within the Ibero-Romance group (even if it's a related group), and it's easy to figure that the divergence would increase with time, more so given the very strong influence of Arabic in it, eventually both becoming mutually unintelligible.
Even in case it was kept as a minorized language (which could happen, as in Al-Andalus, Arabic was the prestige language), a similar problem would appear, as its evolution would be more conservative. As an example, much of the differences between Galician and Portuguese are not as much due to a phenomena of linguistic substitution (even if it played its part), as due to Galician being much closer to the old Galician-Portuguese. Well, Mozarabic in its latter stages was divergent enough respect to Modern Spanish as it was to be part of a continuum with Spanish in an archaizant form.
I don't know about the patterns of Arabic lexicon: Spanish has enough of them and I can't figure out a definite pattern, but the Mozarabic texts I've seen show even more prevalence of them, and not just latinized arabic words but direct borrowings, as you may know. Maybe those indicating status and new products and devices would have a preference.
Finally, if it was kept due to some taifa surviving or such, you may take into account the possibility of Aljamiado -Mozarabic being systematically written in Arabic script- ending up being the most prevalent. If not, it would be impossible to tell, as each Ibero-Romance language has its own grammar rules.