Is there any correlation between population obesity and local temperature distribution? That is, are Alaskans generally thinner than Hawaiians and Swedish people thinner than Tunisians (correcting for other factors to the extent that is possible)?