I think it might if we design spatial replicates i.e. instead of visiting the same transect in a particular grid daily for 15 days you can make a long transect in the same grid, maybe of 10 km but divide it into several 1 km segments which you call replicates and survey each of them differently. However i am not sure if transect survey is a popular method in surveying amphibians or not?
I agree with Yadav. I've seen a study (that I can't find at the moment) that uses several transects within a broader grid as the replicates, which are used to derive the detection function for occupancy modelling.
If you are using timed searches or transects, you may be able to compartmentalize the the search/transect into psuedo-replicates. Jim Hines wrote a good paper on this approach. You could also consider a time-until-detection detection function. Attached are some good papers to look at.