I dug around a bit to see whether I could find anything useful with regard to the question about the 737 wing TAI operation while on the ground. I am not rated on the airplane, so my comments must be so considered.
In the October 1977 issue of Airliner magazine, the following was stated:
During flight testing of a 737, unsymmetrical formations of clear ice occurred on leading edge devices during ground operations involving use of thrust reversers in light snow conditions with crosswinds. These formations, which were difficult to see, occurred in temperature conditions at or moderately below freezing; it appears that snow melted due to hot engine gasses and then refroze on contact with the cold leading edge devices, which are not anti-iced during ground operations. The presence of ice resulted in a tendency to roll at high angles of attack during subsequent takeoffs.
In the fall of 1981, Boeing reported the results of an extensive flight test program to evaluate the effects of frost and a heavier “corn” ice on the leading edge slats of the 737. One of the outputs was an increase in stall speed, over the clean wing 1g speed, of approximately 10 knots with frost, and 14 knots with corn ice, while in the flaps 1 configuration. These stall speed degradations were significantly improved at flaps 40 (approximately 4 knots for frost and 8 knots for corn ice). This is attributed to the differences in airflow between the “sealed” configuration between the slats and the main wing at flaps 1, and the “gapped” configuration at flaps 40.
In the fall of 1983, none other than Alan Mullaly reported in Airliner on an extensive flight test program using two variations of simulated frost. This was carried out on the 737 (similar testing was later extended to the 757 and 767). The entire upper surface was coated, as well as the portions of the slats and flaps that would be exposed during ground operations with the flaps retracted. Increases in the stall speed above the clean, 1g speed were in the 8 to 12 knot range for all three airplanes with flaps 5. The utility of the wing TAI system in ground operations was also tested, by using frost simulant applied to the slats based on the wing TAI predicted capability at both 0F and 20F. This may be the genesis of the 737 procedure.
Although most of this work was done with respect to the takeoff icing threat, in 1992 Gene Hill described testing with only the slats coated with a frost simulant. The results showed, approximately, a 15% loss in maximum lift capability for the 737 with flaps 5, compared to about 19% for the same airplane and flap position with frost simulant applied to both the slats and the entire upper wing surface. These results can be applied to both the takeoff and the inflight icing situations.
I am aware of only three large jet icing accidents involving ice (other than takeoff events). One is the 767 tailstrike at Halifax, the second a DC-8 landing accident at Edmonton, and the third is an Airbus upset near West Palm Beach. All had more than one contributing cause, but in each case the wing TAI was not operated. I am also aware of a handful of buffet encounters and a couple of stick shakers apparently due to a failure to operate the wing TAI. The concern that I have had with respect to larger jets has been the go-around with a contaminated wing; the lift degradations and stall speed increases reported in the Boeing work suggest a quite a bit of margin erosion is possible, and that go-around is where an erosion of margins may show up.
Regarding the statement that wing anti-icing is less effective with slats extended, we have an open question on another thread concerning the Airbus procedure, and I have no information to add to that discussion yet. In the case of the 757/767, for example, the slats are equipped with telescoping ducting for bleed air supply to the protected surfaces, and the 757 AMM states that “Bleed air can be fed to the leading edge slats regardless of slat position.” I don't believe there is any difference in thermal performance regardless of slat position, but there is the above-mentioned difference between sealed and gapped airflows.
If anyone is interested in the Airliner articles and does not have access to them, drop me a note and I can email them to you.