There seems to be a lot of conflicting research on nuclear winter theory, especially since the 1990s. Carl Sagan famously predicted that the Kuwaiti oil well fires from Operation Desert Storm would result in a small global winter, but the effects turned out to be more localized and less severe than estimated. The aftermath of the Desert Storm fires caused nuclear winter theorists to admit flaws with the models and revise their estimates, especially since it was thought that petroleum fields and related facilities would be a major contributor to the cooling effect. However, even in the revision the theorists went from talking about temperature reductions of 15 to 25 degrees centigrade (27 to 45 degrees Fahrenheit) to temperature reductions of 10 to 20 degrees centigrade (18 to 36 degrees Fahrenheit), which still seems like a major reduction (source). More recently some research has indicated that even a small scale nuclear water would cause more widespread cooling and possibly major ozone layer damage (summaries here). The earlier research seems to make it look like nuclear winter would be more of a significant short term weather change lasting a few weeks or months, while more recent research seems to point towards it producing smaller but longer lasting climatic changes. Also, it seems that the total volume of materials burned wouldn't be very large when compared to what occurs during an average year. For example, the estimates of total material burned here range from 1,475 to 5,075 terragrams of material (equivalent to 1.475 to 5.075 billion metric tons), but the annual world consumption of coal is already on par with even the high end estimate. What would make the combustion of that material worse than typical fossil fuel consumption?