add user_id to observations for auditing purposes

[michaelbyrd]

I don't think it is totally necessary, but I think it could be very helpful at a later time

[ahirschberg]

Scientists should also be able to view/edit their own observations. Also related to #52