We observed that there was a large range in the correlation coefficient of BC and PM2.5 (0.31e0.95) (Table 2), as well as in the BC/PM2.5 ratio (0.006e0.028). Various combustion conditions due to different stove types, human activities (cooking and adding fuels), and moisture content of fuel most likely have led to this result. A previous study reports that chimney stoves and simple stoves differ in air supply to the fuel and heat transfer rate, which leads to different combustion conditions (MacCarty et al., 2010).We also found evidence that the air circulation within households was different depending on the existence of a chimney and construction of the tent/house, which could affect the observed values at the
sampling site. Moreover, we observed a change in combustion from smoldering to flaming after residents added fuel and vice versa when the fuel was burning out. Additionally, we found from our measurements that the moisture content of yak dung also likely plays an important role in determining combustion conditions (Yang et al., 2003). Tibetan residents collected yak dung in a pile and left it uncovered in the sun to dry. Thus, some dung was not completely dried before it was burned and precipitation may have significantly increased the moisture content of dung.