"Refuge," as defined in Webster's Ninth New Collegiate Dictionary, is
1) "A place that provides shelter or protection from danger or distress"2) "Something to which one has recourse in difficulty."
"Refugium," in contrast, is defined as "an area of relatively unaltered climate that is inhabited by plants and animals during a period of continental climatic change (as a glaciation) and remains as a center of relict forms from which a new dispersion and speciation may take place after climatic readjustment."
Therefore, the work conservation biologists do to retain natural habitats may be considered preservation of refugia. However, the construction of artificial habitats by conservation biologists (within a refugium) creates refuges.
With this in mind, here in the southwest, conservation biologists may consider the Owens Valley of California, where the Owens pupfish (Cyprinodon radiosus) resides, to be a refugium for the pupfish and other native plants and animals. Within this area conservation biologists have constructed the Owens Valley Native Fish Sanctuary , which is a refuge designed to protect native fishes and other, related taxa.
Similarly, the Ash Meadows National Wildlife Refuge east of Death Valley National Park serves to protect the refugium of Ash Meadows. Anything that the Ash Meadows National Wildlife Refuge might do artificially to protect or preserve the native fish fauna (such as the Amargosa Refuge) is just that: a refuge.
How did the terms come to be confused by conservation biologists? Probably because in the early days of our preservation work conservation biologists were so obsessed with protecting things that they never really looked up the definition of the words we were using. Besides, the term "Refugium" somehow seems to be more sophisticated and "cool," and therefore more likely to gain the attention of people who fund refuges.
In the case of the Pahrump poolfish (Empetrichthys latos), this fish no longer exists in its native habitat. It is therefore necessary (legally) for conservation biologists to construct artificial habitats (refuges) to keep them in existence. If and when their native habitat (Manse Spring in Pahrump Valley, Nevada) should again begin to flow (an optimistic assumption at this point in time), it would then be possible for conservation biologists to reintroduce them from their refuges back into the refugium that is their evolutionary home.
[Adapted from a correspondence between Edwin P. (Phil) Pister Executive Secretary, Desert Fishes Council , concerning ideas reaised by Dr. Gary Meffe, the editor of Conservation Biology and Edited by Matthew Huffine firstname.lastname@example.org]
 See http://www.thesierraweb.com/sightseeing/slough.html,