The main inherent uncertainty sources in trend estimation based on satellite remote sensing data

Satellite data have been extensively used to characterize the trends in environmental and ecological parameters in response to climate change and anthropogenic activities due to the long-term and continuous data record from a global perspective. Nevertheless, trends estimated from different satellite data can be different and even contradict to each other, resulting in inconsistent scientific conclusions about Earth system dynamics. Previous studies were mainly focused on detecting trends or their spatial distribution characteristics, with much less attention to the factors leading up to uncertainties in the results. To address this problem, this review provides a comprehensive analysis of the main factors leading up to the uncertainty of results. The spatial-scale, temporal-scale, platform/sensor-related characteristics, and trend detection algorithms were considered. Their effects as well as the causes of such effects were discussed. And the way to minimize these effects was also suggested. Based upon above, the future development direction was pointed out to obtain more reliable trend estimates from satellite observations. This paper aims to deepen the consciousness and understanding of the detected trends which are not related to actual changes of the Earth surface but caused by perspectives taken for data selection and analysis. Although our discussion on these factors is very preliminary, efforts made here shall provide a first-step guideline that future work can be built up on, with the overarching aim of capturing the real trend of environmental changes using satellite remote sensing.