This paper does seem to conclude in fairly certain terms however that efficiency is reduced on a lean burn engine (without taking compensating measures) by increases in ambient humidity. NOx is also reduced by increases in humidity.
http://www.cimac.com/cimac_cms/uploa...l_20090325.pdf
Though I'm at a loss at why the summer months are the least efficient in the graph, that is when I get my best numbers.