He found that [PDF], for those three days, the average range between highs and lows at more than 4,000 weather stations across the US was 1 degree C wider than normal. In other words, contrails seemed to raise nighttime temperatures and lower daytimes ones.
But the real effect was in daytime highs, which were much higher. That would seem to indicate that, contrary to prevailing thinking, contrails might have a net cooling effect.
Certain areas seemed particularly sensitive to the absence of contrails. Because of unique climatic conditions in the atmosphere in these regions — chiefly, moisture-laden air — the Pacific Northwest and the Midwest are often covered by contrails. But when planes stopped flying right after 9-11, Travis also found that these areas saw the most dramatic increase in daytime highs.
Some have suggested that these results were skewed because unusually clear weather prevailed that week in September 2001. In other words, it was natural variation, not the absence of contrails, that led to the large temperature differential immediately following 9-11.
But there's no doubt that whether the net effect is to cool or warm, contrails can quite dramatically change cloud cover.