"Three boxes were defined in Regions A and D to
focus additional acquisition of HiRISE images within
possible ellipse boxes. THEMIS predawn thermal
infrared image data, which proved very valuable in
finding low rock abundance areas across the entire
latitudinal girdle, were used in concert with hand generated
and automated rock counts in the boxes to
estimate rock densities beyond the HiRISE coverage.
Further, fine-scale (down to 2 m length scale) topography
and slopes were generated from HiRISE stereo
images. Longer length scale slopes were estimated
from MOLA data. All of the data were used to select
the highest priority box (Box 1) (Fig. 2,3) within Region
D for detailed assessments and final ellipse
placements. Rocks abundances are low and slopes are
typically <3° even at fine scales. Work is now proceeding
on defining final ellipse placements before
launch in August 2007. Additional HiRISE images of
the candidate landing sites will be scheduled when
atmospheric and lighting conditions allow in early
2008."
OVERVIEW OF MARS EXPLORATION PROGRAM 2007 PHOENIX MISSION LANDING SITE
SELECTION, p.2

It is possible that the volcanoes or gypsum deposits lie as well within Region D.
The primary selection criterion is lack of rocks to maximize safety of the mission. However, this sometimes conflicts with the area thought to offer the best science return.
Ray Arvidson mentions in this article that Region D was initially ruled out because other regions offered better science return:

Planetary News: Phoenix (2007)
A Green Valley for Phoenix
By Emily Lakdawalla
January 25, 2007
http://www.planetary.org/news/2007/0125_A_Green_Valley_for_Phoenix.html

I'll ask mission team members to see if Region D is firm as to the possible area for landing.

Bob Clark