Several questions arise from this definition:
One definition (Williams 2006, Biology of Temporary Waters) states that all temporary waters are, for the most part, natural bodies of water which experience a recurrent dry phase (often predictable both in its time of onset and duration). The defining element is the cyclical nature of the drought, as some permanent waterbodies may dry up in exceptionally dry years. In New Zealand, the recurrence of drying may be frequent (e.g., 1-5x/year) or infrequent (once every 5 years) and may be aseasonal in some areas but seasonal in other areas.
It is likely that for ephemeral streams in New Zealand, a combination of hydrologic and geomorphologic indices may be useful in for definition and delineation. For example, a well defined channel (something greater than zero order), continuous flow for less than 99% and greater than 1% of the time for most years?
Comments, edits, additions welcome...
