Electronic – Is it safe to connect a small solar panel to a lead acid with only a diode

Should work. Cheap charge controllers are a bang bang MOSFET switch, so either the battery is straight on the solar cell, or not. MPPT controllers have a switch mode buck converter; two of them on one panel will get very confused.

Once the battery is full, the charge controller switch will be mostly off, so the other charge controller can use the solar cell.

The danger might be that if both batteries are low, both controllers will come on together, shorting two batteries together and damaging the controllers. You could check for diodes or add them. They must go before the controllers, not between the controller and the battery, as it needs to sense the battery voltage.

The open questions are 1) are there already diodes in the charge controller, and 2) what does the controller do when the solar voltage collapses to ~14V (when the other one shorts the panel to the battery). It might panic and not charge.
It's worth asking the supplier about both of these, I'm sure they've had these questions before

After the discussion in the comments above and asking some electrical engineers, I decided to go with the following setup:

• leave the 48V battery-motor setup as it is
• wire three 18V panels in parallel. While this is not optimal, the change in voltage due to shading or orientation should be much lower than the loss in current, so the total power output should be higher than wiring them in series
• use a MPPT boost solar charge controller that will handle the panel output, boost the voltage to the level needed for battery charging and prevent overcharging of the battery

This is the low-cost version that will hopefully still yield good results. After construction and testing, I'll update you how well it works (or not).