{"title":"An assessment of GPM IMERG Version 7 rainfall estimates over the North West Himalayan region","authors":"Sreyasi Biswas, Charu Singh, Vidhi Bharti","doi":"10.1016/j.atmosres.2025.107910","DOIUrl":null,"url":null,"abstract":"The latest Global Precipitation Mission (GPM) IMERG V07B Final Run (Hereafter IMERG) has been validated for rainfall estimates in the North West Himalayan (NWH) region during the Indian Summer Monsoon (ISM) season (June–September) of 2000–2022. The validation is assessed against daily 0.25<ce:sup loc=\"post\">o</ce:sup> x 0.25<ce:sup loc=\"post\">o</ce:sup> India Meteorological Department (IMD) gridded rainfall data. We found that IMERG inherently underestimates rainfall, particularly low-intensity (< 50 mm day<ce:sup loc=\"post\">−1</ce:sup>) rainfall. However, an overestimation is evident in high-intensity rainfall (50 mm day<ce:sup loc=\"post\">−1</ce:sup>–200 mm day<ce:sup loc=\"post\">−1</ce:sup>). This is consistent across all elevation ranges, from <1000 m to >5000 m, with the most significant negative bias observed at lower elevations. The proportion of such underestimated rainfall events increases with elevation, while the proportion of overestimated rainfall events decreases. Conclusively, IMERG is negatively skewed (−0.94). The proportion of accurate estimation of rainfall intensity is low and lies between 3 mm day<ce:sup loc=\"post\">−1</ce:sup> to 21 mm day<ce:sup loc=\"post\">−1</ce:sup> for <1000 m. IMERG performs the best in classifying a ‘rain event’ in Uttarakhand (UK) and Himachal Pradesh (HP), which is evident from near-optimal values of categorical metrics like False Alarm Ratio (FAR) (0.19 and 0.30 respectively), Probability of Detection (POD) (0.86 and 0.86 respectively), and Critical Success Index (CSI) (0.71 and 0.62 respectively). The classification of a “no rain event” by IMERG exhibits relatively low accuracy in the two states (Probability of False Detection (POFD) = 0.54 and 0.65 respectively). Overall, the Accuracy (ACC) in classifying a ‘rainfall event’, irrespective of it being a ‘rain event’ or a ‘no rain event’, is fairly good in UK (ACC = 0.75) and HP (ACC = 0.67) including the estimation of ‘rain event’ (Frequency Bias Index FBI = 1.07 and 1.23 respectively). The manifestation of stratiform rainfall in UK and HP could account for the underestimation of rainfall intensity and the discrepancies in the categorical metrics, owing to them being unrecognized by satellite due to warm cloud top temperatures. IMERG estimates are moderate over Jammu and Kashmir (JK) (FAR = 0.40, ACC = 0.1, CSI = 0.53, POFD = 0.64), while a large uncertainty in the performance of IMERG exists over Ladakh (LD) due to the paucity of IMD rain gauges (FAR = 0.62, ACC = 0.49, CSI = 0.34, POFD = 0.65).","PeriodicalId":8600,"journal":{"name":"Atmospheric Research","volume":"3 1","pages":""},"PeriodicalIF":4.5000,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Atmospheric Research","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1016/j.atmosres.2025.107910","RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"METEOROLOGY & ATMOSPHERIC SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
The latest Global Precipitation Mission (GPM) IMERG V07B Final Run (Hereafter IMERG) has been validated for rainfall estimates in the North West Himalayan (NWH) region during the Indian Summer Monsoon (ISM) season (June–September) of 2000–2022. The validation is assessed against daily 0.25o x 0.25o India Meteorological Department (IMD) gridded rainfall data. We found that IMERG inherently underestimates rainfall, particularly low-intensity (< 50 mm day−1) rainfall. However, an overestimation is evident in high-intensity rainfall (50 mm day−1–200 mm day−1). This is consistent across all elevation ranges, from <1000 m to >5000 m, with the most significant negative bias observed at lower elevations. The proportion of such underestimated rainfall events increases with elevation, while the proportion of overestimated rainfall events decreases. Conclusively, IMERG is negatively skewed (−0.94). The proportion of accurate estimation of rainfall intensity is low and lies between 3 mm day−1 to 21 mm day−1 for <1000 m. IMERG performs the best in classifying a ‘rain event’ in Uttarakhand (UK) and Himachal Pradesh (HP), which is evident from near-optimal values of categorical metrics like False Alarm Ratio (FAR) (0.19 and 0.30 respectively), Probability of Detection (POD) (0.86 and 0.86 respectively), and Critical Success Index (CSI) (0.71 and 0.62 respectively). The classification of a “no rain event” by IMERG exhibits relatively low accuracy in the two states (Probability of False Detection (POFD) = 0.54 and 0.65 respectively). Overall, the Accuracy (ACC) in classifying a ‘rainfall event’, irrespective of it being a ‘rain event’ or a ‘no rain event’, is fairly good in UK (ACC = 0.75) and HP (ACC = 0.67) including the estimation of ‘rain event’ (Frequency Bias Index FBI = 1.07 and 1.23 respectively). The manifestation of stratiform rainfall in UK and HP could account for the underestimation of rainfall intensity and the discrepancies in the categorical metrics, owing to them being unrecognized by satellite due to warm cloud top temperatures. IMERG estimates are moderate over Jammu and Kashmir (JK) (FAR = 0.40, ACC = 0.1, CSI = 0.53, POFD = 0.64), while a large uncertainty in the performance of IMERG exists over Ladakh (LD) due to the paucity of IMD rain gauges (FAR = 0.62, ACC = 0.49, CSI = 0.34, POFD = 0.65).
期刊介绍:
The journal publishes scientific papers (research papers, review articles, letters and notes) dealing with the part of the atmosphere where meteorological events occur. Attention is given to all processes extending from the earth surface to the tropopause, but special emphasis continues to be devoted to the physics of clouds, mesoscale meteorology and air pollution, i.e. atmospheric aerosols; microphysical processes; cloud dynamics and thermodynamics; numerical simulation, climatology, climate change and weather modification.