I am doing a food availability survey for nectar eating animals which mostly eat nectar from a particular tree species. To do this I wanted to find an approximate way to quantify the number of flowers available each year and at different times of the year (It's fairly obvious that sometimes there are many animals and many flowers, and sometimes when there are no flowers there are also no animals but need to quantify this a bit more accurately).
I arbitrarily selected two 5 ha (100m x 500m) rectangular areas where the trees were fairly common and where the animals were known to forage, then split these into five 1 ha (100m x 100m) sections which were in turn divided into 5m x 5m quadrats. Within each 1 ha section four of the smaller quadrats were randomly selected using a random number generator. I counted the number of trees in the selected quadrat. Then I counted flowers on the closest tree to the chosen quadrats' centre.
So overall, in each 5 ha area I had 20 randomly selected locations where I took the closest tree and counted the flowers (and also counted the number of trees within 20 randomly selected quadrats). This could then be used to get some idea of the flowers available per hectare within the area, and I could compare this with the abundance of the animal species that I'm studying in the area.
Just wondering, is there any reason why I can't just reuse the same randomly chosen locations in consecutive years and/or seasons? i.e. just look at the exact same trees each year and count the flowers? Do I need to redo the random numbers and re-select random (probably different) trees each year? As far as I can see I think re-selecting the same trees would be fine as by counting the flowers I'm not affecting the flowering of the trees in any way, and I am comparing between years and seasons rather than trying to get a very accurate average number of flowers in a single year but perhaps I have missed something?
All help is appreciated,