A 1M pot will sweep the entirety of 1M regardless of taper. The issue is the distribution of resistance along the wiper's sweep. (This is what people mean when they say that a control "bunches up" at one part of the rotation.) In the graph below (line 2), see how the log pot allows for much more nuance in the beginning of the rotation before moving to a much more dramatic sweep when compared to the linear pot (line 1). This is appropriate for audio controls since humans do not perceive increases in volume linearly. This also works for different circuit elements due to their construction.
In the Low Tide circuit, that pot is adjusting the SLEW (damp) parameter. If you find that you need more nuanced control there, then swapping it out might be worth it. If you're happy with how that control responds as is, then I would leave it as is.
The pot essentially controls the signal conditioning of the randomizer output. First, the pot adjusts the series resistance leaving the randomizer to form a LPF with R63 and C36. The signal then passes through a unity gain buffer before an inverting op amp stage. The second half of the pot sets the gain for the op amp stage. This construction (with the capacitor) forms an
integrator. This circuit block essentially smooths the step transitions from the randomizer. The output is fed (via the DEPTH control) to the VCO input (i.e., the clock for the BBD).
You'll have the same control over the behavior of this circuit block with the linear pot, albeit differently. You'll have less fine control over the damping effect of the step transitions and will have to be more precise with your knob adjustment(s).
So, if you're happy with the control you have on the parameter now, I would not advise changing it; if you are not, then you might want to consider swapping it out.