In this model:
Biomass ~ Species + (1|Site / Species)
Biomass and Species are continuous variables, and Site is categorical.
We can see that Biomass is the model's response while Species is the predictor. What's more, a random effect is added to the model's intercept, because there is a (1 | R) term.
But, what does Site / Species mean?
The term (1|Site/Species) means that you have 2 levels of clustering. So you have measurements of Biomass of individuals, these individuals are clustered within Species, and Species are clustered/nested within Site. This means that any particular Species "belongs" to only one Site. If you have the same Species within different Sites, then you have crossed, not nested, random effects, and probably what you should use (if you have sufficient Sites and Species) is (1|Site) + (1|Species).
The model will specify random intercepts for both Site and Species, and so Species should not also be a fixed effect.
See here for more information about nested and crossed random effects.