The Global Precipitation Measurement Mission: Performance Evaluation across Multiple Spatio-Temporal Scales

MSc. Rajani Kumar Pradhan, MSc

Doctoral thesis

Abstract:

Precipitation is probably one of the most important water cycle variables to measure. Considering its high spatio-temporal variability, satellite-based techniques emerge as the most reliable method for accurate measurements with near-global coverage. However, being the indirect measurement method, satellite estimates are subject to many assumptions and thus suffer from several limitations. Therefore, it is essential to thoroughly evaluate satellite estimates and comprehend their uncertainties prior to their utilization in scientific endeavors. This thesis aims to comprehensively evaluate the GPM IMERG precipitation estimates at the global level. Initially, it assesses the IMERG product's performance based on published studies between 2016 - 2019. This quantitative review lays down the foundation for the two other important research objectives. Firstly, the evaluation of IMERG precipitation estimates across the tropical oceans. Secondly, the assessment of IMERG alongside other satellites and reanalysis estimates at sub-daily scales, focusing on their ability to depict the diurnal cycle at the global level. The results reveal that IMERG effectively captures the spatial-temporal variation in precipitation across diverse geographic and climatic conditions, often outperforming TRMM products. However, IMERG exhibits certain limitations, particularly during heavy and extreme precipitation events. Despite its potential for detecting and tracking extreme events, IMERG underestimates the high precipitation intensities, especially over the core of typhoons and hurricanes. Its performance shows a relative improvement during summer compared to winter and is more effective in areas characterized by high-intensity and humid conditions than in drier regions with lower precipitation. In hydrological applications, it has shown comparatively better performance than TRMM, though its performance varies with the region, model, calibration datasets, etc. In terms of IMERG performance over the ocean, it represents well the precipitation spatial patterns across the tropical oceans, similar to the buoys. However, IMERG tends to overestimate the precipitation in the high precipitation regions, such as over the Indian and West Pacific Oceans. IMERG detection ability is best over these high-precipitation regions compared to the low-precipitation regions (i.e., East Pacific and Atlantic regions). The false alarms and positive hit bias are the major components that contribute to the total error. Finally, IMERG, along with other satellite estimates (GSMaP, CMORPH, PERSIANN), and the reanalysis product (ERA5), agree on producing major global diurnal features with an afternoon peak over land and an early morning peak over the ocean. However, discrepancies do exist among the estimates in terms of the exact time of peak, amount, frequency, and intensities. In particular, ERA5 suffers from overestimation (underestimation) of precipitation frequency (intensity) throughout the globe. Among the satellite precipitation estimates, IMERG and CMORPH are in close agreement in most instances, while GSMaP and PERSIANN agree in a few instances. The thesis contributes to a comprehensive understanding of the uncertainties and inaccuracies inherent in IMERG across a variety of geographical and climatic conditions at the global level. This not only provides valuable insights for algorithm developers to identify strengths and limitations but also empowers end-users and the scientific community to select datasets accordingly. As a result, this enhanced understanding improves the quality of outcomes and facilitates more informed decision-making processes. Given the widespread application of IMERG precipitation data in various fields such as weather forecasting, cyclone tracking, flash flood and landslide forecasting, elucidating the uncertainties associated with IMERG datasets holds paramount importance. Ultimately, this information leads to improved overall results an

Abstract:

d decision-making, thereby mitigating the societal impact of adverse events.

Keywords

GPM-IMERG, Precipitation, Evaluation, Performance, Diurnal cycle

Language used: English

Date on which the thesis was submitted / produced: 11. 4. 2024

Thesis defence

• Supervisor: doc. Mgr. Ing. Ioannis Markonis, Ph.D.
• Reader: David Moravec, Ph.D., Ing., Christoforos Pappas, externi, Nikolaos Mamassis, externi

Czech University of Life Sciences Prague

Faculty of Environmental Sciences

Doctoral programme / field:
Landscape Engineering / Environmental Modelling