Thanks for the links. Interestingly the authors in one of the reports don't find the dust flow model by itself to be enough at least not for this case. They suggest a release of CO2 gas may help the flow.

DARK SLOPE STREAKS AND ASSOCIATED LAYERED DEPOSITS ON THE SOUTHWESTERN FLOOR OF CASSINI IMPACT BASIN, MARS.
"Slope streaks in Cassini basin may have been
initiated by the collapse of selected portions of the
steep-walled mesa unit. We suggest that small-scale
terrain collapse may have contributed to the release of
small quantities of regolith CO2 gas, which when
mixed in with dust particles, served as a likely sup-porting medium for a “dusty” density flow."
www.lpi.usra.edu/meetings/lpsc2001/pdf/1380.pdf

It is also possible that liquid water might help the flow. For dry dust flows there should be an angle of repose at which dry dust should be stable. However, it is known that if there is liquid water mixed in then the angle of repose can decrease to approach 0
degrees:

Angle of Repose,
http://www.geo.mtu.edu/department/classes/ge404/avalanche/angle.html

NAME: Slope failure (landslides)
http://www.gcrio.org/geo/slope.html

LANDSLIDES, SLOPE FAILURE, AND OTHER MASS WASTING PROCESSES,
http://wapi.isu.edu/EnvGeo/EG4_mass_wasting/landslide2.html

Now consider the shallow slopes in these image:

TERMINUS OF LANDSLIDE FROM OLYMPUS WEST BASAL SCARP,
http://ida.wr.usgs.gov/html/orb_0263/26303.html

Bars in Mangala Valles side channel,
http://ida.wr.usgs.gov/html/fha011/fha01100.html

Bob Clark