Most prior paleoclimate studies (discussed below) either lumped together the entire section at HAFO, or divided the Glenns Ferry Formation into only three portions.
The seasonal range in temperatures derived from oxygen isotopic analysis of a horse tooth from the HAFO (mistakenly referred to as Pliohippus; Kohn et al., 2002:155) was estimated at 12[degrees]C (Kohn et al., 2002), a 9[degrees]C difference from the estimate based on the fossil fish ototlith (Smith and Patterson, 1994).
These two studies used material from HAFO, but because the temporal intervals between data points are longer than the entire Pliocene interval at HAFO, they lack the resolution to examine fauna change within the Glenns Ferry Formation.
Ostracodes also suggest wetter-than-today conditions throughout the portion of the Glenns Ferry Formation at HAFO (Forester, 1991).
Pollen data from HAFO suggest pine woodland or open forest vegetation with steppe taxa (Leopold and Wright, 1985).
Therefore isotopic changes in snail shells from HAFO may reflect the growth and shrinking of Pliocene Lake Idaho.
The Glenns Ferry Formation at HAFO is divided into three informal units described as representing upper and lower floodplain environments and an intermediate marshy interval (Zakrzewski, 1969; Bjork, 1970; "members" sensu Lee et al., 1995).
Significant thicknesses of carbonaceous shales at HAFO occur for about 25 m stratigraphically above the middle member.
The terrestrial paleoclimatic data reviewed above lack the detailed resolution needed to examine the effect of climate change on the mammalian assemblage at HAFO. In order to compare stratigraphic changes in faunas at HAFO, climate patterns are needed at much more tightly constrained temporal intervals.
Global circulation models (GCMs) have estimated repeatedly the mean annual temperature for the HAFO area in the Pliocene at about 3.5[degrees]C warmer than today (Raymo et al., 1990; Covey et al., 1991; Crowley et al., 1994; Sloan et al., 1996; Haywood et al., 2000, 2001; Jiang et al., 2005), although Crowley (1991) suggested no difference in temperature.
Although GCMs do not provide the resolution necessary to examine the faunal shifts at HAFO in light of environmental changes, they can be used to suggest which areas changed in concert with southern Idaho in the Pliocene.
For the interval from 3.0 to 4.2 Ma, which encompasses the Glenns Ferry Formation deposits at HAFO, each dataset was rescaled so that the total temperature variation was the same for all.