<b>Daily and monthly precipitation data</b><b>set</b><b>at 0.01</b><b>°</b><b>for the s</b><b>outhwest China </b><b>h</b><b>ighland </b><b>c</b><b>anyon </b><b>a</b><b>rea</b><b> </b><b>(</b><b>2017 </b><b>-</b><b>2022</b><b>)</b>

To enhance the southwest China highland canyon area (SCHCA), this study constructed fused datasets at both monthly and daily scales for the period 2017-2022 based on GPM-IMERG and meteorological station observed precipitation data, combined with the MGWR(multiscale geographically weighted regression), TL-MGWR(time-lagged MGWR)and geographical difference analysis. Findings demonstrate that the fused precipitation dataseteffectively alleviated the overestimation issue of the original GPM measurements on both monthly and daily timescales, whileprovidinga more detailed spatial distribution. Moreover, the correction process significantly improved the precision of the downscaling results. Therefore, the fused precipitation data has good application potential in the southwestern mountainous canyon area.为优化西南高原峡谷区（southwest China highland canyon area, SCHCA）的降水数据支撑能力，本研究以2017-2022年的GPM-IMERG降水数据与气象站点实测降水数据为基础，结合多尺度地理加权回归（multiscale geographically weighted regression, MGWR）、时滞多尺度地理加权回归（time-lagged MGWR, TL-MGWR）以及地理差异分析方法，构建了月尺度与日尺度的融合降水数据集。研究结果表明，该融合降水数据集在月、日两个时间尺度上均有效缓解了原始GPM测量产品的高估偏差，同时可呈现更为精细的空间分布特征。此外，该校正过程显著提升了降尺度结果的精度。因此，该融合降水数据在西南山地峡谷区具备良好的应用潜力。