If it was the right type of oil, it would increase the likelihood, as it's the wrong type, the likelihood is unchanged
Bayes' rule doesn't work like that.
The posterior likelihood of an event is always affected by dependent evidence of a binary nature which itself has a prior probability which is neither 1 or 0.
The likelihood of this being the correct search area prior to the oil analysis was balanced somewhere between the likelihood given a positive oil test and the likelihood given a negative oil test. Where it lay along that line is dependent on the prior likelihood that the oil test would be positive.
That likelihood is unknown of course, but 1>P>0.