Gridded population projections consistent with the shared socioeconomic pathways (SSPs) are critical for the studies of climate change impacts and their mitigation. Existing gridded population projections under the SSPs have relatively coarse resolution and issue of overestimation in populated areas, which further bias the analysis of climate change impacts. In this study, we proposed a scheme by integrating high‐resolution historical population maps and machine learning models to predict future built‐up land and population distributions, which were rendered consistent with the SSPs. Using this proposed method in China, we generated a set of 100‐m SSPs population maps for China from 2015 to 2050 at 5‐year intervals. Our projections revealed different spatial structures for the population distribution at the grid level and three modes of provincial population change across the five SSPs from 2015 to 2050. By applying the 100‐m SSPs population grids, we showed that, from 2015 to 2050, exposure to extreme heat in China will increase by 121–136% and 164–191% under the representative concentration pathways 4.5 and 8.5, respectively. We also found a severe spatial bias in the existing 1/8 ° SSPs population grids, i.e., 30–43% of the estimated population is wrongly allocated in cropland, forest, and pastureland. This bias results in substantial underestimation of extreme heat exposure in high‐density metropolitan areas and overestimation in medium and low‐density areas. Plain Language Summary In this study, we proposed a scheme by integrating high‐resolution historical population maps and machine learning models to predict future built‐up land and population distributions, which were rendered consistent with the SSPs. Using this proposed method, we generated a set of 100‐m SSPs population maps for China from 2015 to 2050 at 5‐year intervals. Our projections revealed different spatial structures for the population distribution at the grid level and three modes of provincial population change across the five different SSPs from 2015 to 2050. By applying the 100‐m SSPs population grids, we showed that, from 2015 to 2050, exposure to extreme heat in China will increase by 121–136% and 164–191% under the representative concentration pathways 4.5 and 8.5, respectively. We further compared our projections with the existing 1/8 ° SSPs population grids and we found a severe spatial bias in the 1/8 ° SSPs population grids: 30–43% of the estimated population is wrongly allocated in cropland, forest, and pastureland. This bias results in substantial underestimation of extreme heat exposure in high‐density metropolitan areas and overestimation in medium and low‐density areas.
展开▼
