Dr Carter suggests "that slight changes in the patterns of oceanic and atmospheric circulation may be more significant in affecting the patterns of urban temperature trends than the growth of urban areas". However, Dronia [1 ], by comparing 67 urban areas with 67 adjacent rural areas in the northern atmosphere, has demonstrated that the urban heating effect is the dominant factor affecting urban temperature trends during the last 90 years. If we had ignored this effect in our paper on 'Global Cooling?' [2], a critic could have argued that we observed no overall cooling in the southern hemisphere during the last three solar cycles because the urban heating effect had masked hemispheric cooling. Therefore, we tested the data for an urban heating effect and found that five of eight cities with populations greater than 750,000 showed a warming trend with a net change for the eight cities of +0.2~ significant at the 92.5% confidence level [3]. However, as progressively smaller cities are included [2, Fig. 3], the effect becomes less significant with an insignificant net change of only +0.02~ for 15 cities of greater than 250,000 inhabitants. Net hemispheric averages for 'non-urban' stations (defined variously for our purposes as < 250,000, < 500,000, < 750,000, < 106 inhabitants) in no case show a significant temperature change. The net hemispheric warming trend for stations in areas with less than 250,000 populations (+0.05~ was not significant by our criteria. There are regional variations but, in light of the above arguments, and by paraphrasing Dr Carter's suggestion, we conclude that the urban heating effect is real and significant but partly masked by "slight changes in the patterns of oceanic and atmospheric circulation".
Contrary to Dr Carter's statement, our case for a high latitude warming trend is not based, in large part, on four stations that lie between 51~ to 61~ We explicitly state that "the four stations (stations 1 to 4) at latitudes above 45~ (51 ~ to 61~ show a pronounced warming trend of 0.3~ but the number of stations is insufficient to establish confidence that the trend is representative of the entire zonal area" [2, p. 450]. Our evidence for a high latitude warming trend was based on 14 stations above 45~ that show a net increase of +0.37~ between the 1960-1964 pentad and the 1970-1974 pentad which is significant at the 95% confidence level [3]. Van Loon and Williams [4] found a similar average change from 1956 to 1972 within the latitudinal belt between 65~ to 95~ This change, weighted for area, was +0.6~ in winter and +0.2~ in Climatic Change 1 (1978) 387-389. All Rights Reserved.