Temporal scale effects on trend estimates for solar radiation, thermal and snow conditions, and their feedbacks: the case from China

The uncertainty in trend estimates caused by temporal scale effect has rarely been studied. In response to this challenge, this study aims to explore temporal scale effects on trend detection based on in situ measurements throughout China given that they provide full temporal details of surface elements. Three common temporal scales (daily, monthly, and yearly) were used to present findings. Two different trend detection methods (i.e., simple linear regression and breaks for additive seasonal and trend) were employed to check the dependence of temporal scale effects on methods. It was found that temporal scale effects were dependent on the type of elements, the time span of datasets, and trend detection methods. They are more significant for elements with fast changes (e.g., snow depth) compared to those with gradual changes (e.g., albedo, air temperature). And temporal scale effects from daily to monthly scale are generally larger than those from monthly to yearly scale. The former is almost independent of trend detection methods, while the latter shows a clear dependence (i.e., dealing with seasonality or not). Trend estimates of surface albedo and longwave net radiation are mainly affected by temporal scale effects, while those of snow depth, shortwave net radiation, and air temperature are affected by both temporal scale effects and the choice of trend detection methods. The feedbacks between these elements show a clear dependence on temporal scale, which are generally stronger on monthly scale but weaker on yearly scale.